Hugging Face's logo

mindpadi
/
hybrid_classifier_suite

Text Classification
sentence-transformers
Joblib
Scikit-learn
English
intent-classification
emotion-detection
mental-health
lstm
Model card Files
xet
Community
onisj commited on
Commit
bb92ae3
·
verified ·
1 Parent(s): 3f03384

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +3 -0
  2. bert_embeddings.npy +3 -0
  3. distilgpt2/config.json +46 -0
  4. distilgpt2/generation_config.json +6 -0
  5. distilgpt2/merges.txt +0 -0
  6. distilgpt2/tokenizer.json +0 -0
  7. distilgpt2/vocab.json +0 -0
  8. emotion_classifier/checkpoint-119517/scheduler.pt +3 -0
  9. emotion_classifier/checkpoint-119517/training_args.bin +3 -0
  10. emotion_classifier/checkpoint-39839/model.safetensors +3 -0
  11. emotion_classifier/checkpoint-39839/rng_state.pth +3 -0
  12. emotion_classifier/checkpoint-39839/scheduler.pt +3 -0
  13. emotion_classifier/checkpoint-39839/training_args.bin +3 -0
  14. emotion_classifier/checkpoint-79678/rng_state.pth +3 -0
  15. emotion_classifier/checkpoint-79678/scheduler.pt +3 -0
  16. emotion_classifier/checkpoint-79678/training_args.bin +3 -0
  17. emotion_classifier/model.safetensors +3 -0
  18. fine_tuned_distilgpt2_lora/.DS_Store +3 -0
  19. fine_tuned_distilgpt2_lora/adapter_model.safetensors +3 -0
  20. fine_tuned_distilgpt2_lora/checkpoint-1598250/optimizer.pt +3 -0
  21. fine_tuned_distilgpt2_lora/checkpoint-1598250/trainer_state.json +3 -0
  22. fine_tuned_distilgpt2_lora/checkpoint-1598264/adapter_model.safetensors +3 -0
  23. fine_tuned_distilgpt2_lora/checkpoint-1598264/optimizer.pt +3 -0
  24. fine_tuned_distilgpt2_lora/checkpoint-1598264/rng_state.pth +3 -0
  25. fine_tuned_distilgpt2_lora/checkpoint-1598264/trainer_state.json +3 -0
  26. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743370296.onisj.local.78999.0 +3 -0
  27. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743561793.onisj.local.62916.0 +3 -0
  28. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743789658.onisj.local.2551.0 +3 -0
  29. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743943210.onisj.local.11693.0 +3 -0
  30. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744027345.onisj.local.13791.0 +3 -0
  31. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744235031.onisj.local.14287.0 +3 -0
  32. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744425415.onisj.local.3156.0 +3 -0
  33. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744651592.onisj.local.35391.0 +3 -0
  34. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744799093.onisj.local.6118.0 +3 -0
  35. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744891112.onisj.local.34883.0 +3 -0
  36. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744987584.onisj.local.14707.0 +3 -0
  37. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745107869.onisj.local.23033.0 +3 -0
  38. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745170183.onisj.local.36505.0 +3 -0
  39. fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745408660.onisj.local.2317.0 +3 -0
  40. fine_tuned_gpt2/checkpoint-1000/config.json +39 -0
  41. fine_tuned_gpt2/checkpoint-1000/generation_config.json +6 -0
  42. fine_tuned_gpt2/checkpoint-1000/trainer_state.json +1533 -0
  43. fine_tuned_gpt2/checkpoint-1500/config.json +39 -0
  44. fine_tuned_gpt2/checkpoint-1500/generation_config.json +6 -0
  45. fine_tuned_gpt2/checkpoint-1500/trainer_state.json +2283 -0
  46. fine_tuned_gpt2/checkpoint-2000/config.json +39 -0
  47. fine_tuned_gpt2/checkpoint-2000/generation_config.json +6 -0
  48. fine_tuned_gpt2/checkpoint-2000/trainer_state.json +3033 -0
  49. fine_tuned_gpt2/checkpoint-2500/config.json +39 -0
  50. fine_tuned_gpt2/checkpoint-2500/generation_config.json +6 -0
.gitattributes CHANGED
@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ fine_tuned_distilgpt2_lora/.DS_Store filter=lfs diff=lfs merge=lfs -text
37
+ fine_tuned_distilgpt2_lora/checkpoint-1598264/trainer_state.json filter=lfs diff=lfs merge=lfs -text
38
+ fine_tuned_distilgpt2_lora/checkpoint-1598250/trainer_state.json filter=lfs diff=lfs merge=lfs -text
bert_embeddings.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:929ccff1583416e52fce954303e894472b886d779d6be5cce6770d129d47fa06
3
+ size 30720128
distilgpt2/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "distilgpt2",
3
+ "_num_labels": 1,
4
+ "activation_function": "gelu_new",
5
+ "architectures": [
6
+ "GPT2LMHeadModel"
7
+ ],
8
+ "attn_pdrop": 0.1,
9
+ "bos_token_id": 50256,
10
+ "embd_pdrop": 0.1,
11
+ "eos_token_id": 50256,
12
+ "id2label": {
13
+ "0": "LABEL_0"
14
+ },
15
+ "initializer_range": 0.02,
16
+ "label2id": {
17
+ "LABEL_0": 0
18
+ },
19
+ "layer_norm_epsilon": 1e-05,
20
+ "model_type": "gpt2",
21
+ "n_ctx": 1024,
22
+ "n_embd": 768,
23
+ "n_head": 12,
24
+ "n_inner": null,
25
+ "n_layer": 6,
26
+ "n_positions": 1024,
27
+ "reorder_and_upcast_attn": false,
28
+ "resid_pdrop": 0.1,
29
+ "scale_attn_by_inverse_layer_idx": false,
30
+ "scale_attn_weights": true,
31
+ "summary_activation": null,
32
+ "summary_first_dropout": 0.1,
33
+ "summary_proj_to_labels": true,
34
+ "summary_type": "cls_index",
35
+ "summary_use_proj": true,
36
+ "task_specific_params": {
37
+ "text-generation": {
38
+ "do_sample": true,
39
+ "max_length": 50
40
+ }
41
+ },
42
+ "torch_dtype": "float32",
43
+ "transformers_version": "4.49.0",
44
+ "use_cache": true,
45
+ "vocab_size": 50257
46
+ }
distilgpt2/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.49.0"
6
+ }
distilgpt2/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
distilgpt2/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
distilgpt2/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
emotion_classifier/checkpoint-119517/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f956e6aa91e4b40b252e821e4419f726f38a3dd1527f3e5ba9e02748126fbbe
3
+ size 1064
emotion_classifier/checkpoint-119517/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5d7a07b9baeaf2d2861a371db2a2874fad2ed4279f4f6090f2fd66dff8685f6
3
+ size 5368
emotion_classifier/checkpoint-39839/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2230ad442cbb1df6d6c1a9a3cf45b64a9e9a67bbd723f289d077cb36b24d5768
3
+ size 267838720
emotion_classifier/checkpoint-39839/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b182573f61d8bcf5eaefcbf8f98d8734b6db51b44ad36aed3a305c431539fa1
3
+ size 13990
emotion_classifier/checkpoint-39839/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f939aa4dea0ee2da0db573c3063761c71d749cf8b3163284a82833fb3a41df77
3
+ size 1064
emotion_classifier/checkpoint-39839/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5d7a07b9baeaf2d2861a371db2a2874fad2ed4279f4f6090f2fd66dff8685f6
3
+ size 5368
emotion_classifier/checkpoint-79678/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06e55bfc8723f269a626afca0be6f7def5753f3bb265436b94c5580b703cfcc7
3
+ size 13990
emotion_classifier/checkpoint-79678/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebe79402ff79dac5e3b3630d224890b66813f45fcdf401cc026efe2ee5c3778b
3
+ size 1064
emotion_classifier/checkpoint-79678/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5d7a07b9baeaf2d2861a371db2a2874fad2ed4279f4f6090f2fd66dff8685f6
3
+ size 5368
emotion_classifier/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a70113e18da0efe55f4a251403eaf2c6835ed53f912852136d1e2786436d2740
3
+ size 267838720
fine_tuned_distilgpt2_lora/.DS_Store ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2dae9519a8ff736077d1b8b10bf4ab5e54cdb67b555355050c383e1ac3836029
3
+ size 1257476
fine_tuned_distilgpt2_lora/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89df27fcd8c25b26a69ace477c641628b9488755f7f39d8e0c784b001d9ac66c
3
+ size 591352
fine_tuned_distilgpt2_lora/checkpoint-1598250/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb5a23a45428aecadd326d34876ac841a4071dbf1bd75c035ea61d26d471b3b0
3
+ size 1190010
fine_tuned_distilgpt2_lora/checkpoint-1598250/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3296bd356e5463d7a027c6259312aa23a67d9bf247304739b5912576e54315e
3
+ size 41046620
fine_tuned_distilgpt2_lora/checkpoint-1598264/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89df27fcd8c25b26a69ace477c641628b9488755f7f39d8e0c784b001d9ac66c
3
+ size 591352
fine_tuned_distilgpt2_lora/checkpoint-1598264/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc0ea97a74aaa2a7a8d97ca2f438469a2e90daf1f0c09e7bd5b693c82a754278
3
+ size 1190010
fine_tuned_distilgpt2_lora/checkpoint-1598264/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b182573f61d8bcf5eaefcbf8f98d8734b6db51b44ad36aed3a305c431539fa1
3
+ size 13990
fine_tuned_distilgpt2_lora/checkpoint-1598264/trainer_state.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55f8db5d2ee4b09661ac74ac27d94a5e556d229141cb898e46625e59d3492792
3
+ size 41046862
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743370296.onisj.local.78999.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7897cfaecdf199d6c32c117c223554ceaf26d3ecb644da0068f68e8aacf6a041
3
+ size 827758
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743561793.onisj.local.62916.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e0a379458c656fa84348b30d37a050dac37a57e4f8767674056a62f304a3a48
3
+ size 2348158
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743789658.onisj.local.2551.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2a3b8d000e26b6be05fffb5abf580ebae41f8452775439f56e315fb6518e4c4
3
+ size 2953630
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1743943210.onisj.local.11693.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d67d1cc912eec2602f7d42db58edebf1aaee4a11fee2dd3a2a4164e70da681f
3
+ size 2609230
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744027345.onisj.local.13791.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5105e93b369438ad7faab1f99283d7dfe79fe1330fbf46fa09e635d0d06a9fd0
3
+ size 2292718
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744235031.onisj.local.14287.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:928654377ffed2cecc7b2927d89d79dab894bfb4983a8b5d5b87fa7a8fee0d48
3
+ size 6902974
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744425415.onisj.local.3156.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7819a17aa9abd1f022f291e4a8fe287fe2cf04ecebde4eab10ddce5cdbd01af5
3
+ size 6997390
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744651592.onisj.local.35391.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4834932e0d9e9e708a32c6ed1226c853685dbbf232ab43d21bcd303e95c3c94a
3
+ size 5790814
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744799093.onisj.local.6118.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40daacaac8956b4ab1cb9245b79f077cfb4985b3152f374a7af4f4868bf89a22
3
+ size 4180702
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744891112.onisj.local.34883.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:558ce9a3c5be16eecba9df2dfca91e69d7fe4d8f4c59e5d7db6180d9b885f4bb
3
+ size 2421790
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1744987584.onisj.local.14707.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:015fa1146ade3c8533c22c2a268769e492a69615fcb81a4299d9b7413639bacd
3
+ size 3498670
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745107869.onisj.local.23033.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f3b9746ecabfffcd36e24841eb3a6e312db5bf43f9180ab0b8e43dbce47c264
3
+ size 2514526
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745170183.onisj.local.36505.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a62da4b4fc5514fa99ee298ce3c5e0abb87dfa552754a50fac319ab304bc69be
3
+ size 7136206
fine_tuned_distilgpt2_lora/runs/events.out.tfevents.1745408660.onisj.local.2317.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef2f9f6005ef4ab3f6300dd37c40b3f0d7457efcd5c1c974065191576796e4be
3
+ size 174767
fine_tuned_gpt2/checkpoint-1000/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/Users/user/projects/MindPadiBig/mindpadi/models/gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.49.0",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
fine_tuned_gpt2/checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.49.0"
6
+ }
fine_tuned_gpt2/checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,1533 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.9095934629440308,
3
+ "best_model_checkpoint": "/Users/user/projects/MindPadiBig/mindpadi/models/fine_tuned_gpt2/checkpoint-1000",
4
+ "epoch": 1.0,
5
+ "eval_steps": 10,
6
+ "global_step": 1000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "grad_norm": 126.21493530273438,
14
+ "learning_rate": 1.0000000000000002e-06,
15
+ "loss": 6.9352,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.01,
20
+ "eval_loss": 6.456012725830078,
21
+ "eval_runtime": 128.7117,
22
+ "eval_samples_per_second": 7.769,
23
+ "eval_steps_per_second": 1.942,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02,
28
+ "grad_norm": 112.57171630859375,
29
+ "learning_rate": 2.0000000000000003e-06,
30
+ "loss": 6.2877,
31
+ "step": 20
32
+ },
33
+ {
34
+ "epoch": 0.02,
35
+ "eval_loss": 6.174145698547363,
36
+ "eval_runtime": 133.4573,
37
+ "eval_samples_per_second": 7.493,
38
+ "eval_steps_per_second": 1.873,
39
+ "step": 20
40
+ },
41
+ {
42
+ "epoch": 0.03,
43
+ "grad_norm": 185.73190307617188,
44
+ "learning_rate": 3e-06,
45
+ "loss": 5.9569,
46
+ "step": 30
47
+ },
48
+ {
49
+ "epoch": 0.03,
50
+ "eval_loss": 5.594101428985596,
51
+ "eval_runtime": 136.3272,
52
+ "eval_samples_per_second": 7.335,
53
+ "eval_steps_per_second": 1.834,
54
+ "step": 30
55
+ },
56
+ {
57
+ "epoch": 0.04,
58
+ "grad_norm": 160.77322387695312,
59
+ "learning_rate": 4.000000000000001e-06,
60
+ "loss": 4.9783,
61
+ "step": 40
62
+ },
63
+ {
64
+ "epoch": 0.04,
65
+ "eval_loss": 4.95811128616333,
66
+ "eval_runtime": 123.0408,
67
+ "eval_samples_per_second": 8.127,
68
+ "eval_steps_per_second": 2.032,
69
+ "step": 40
70
+ },
71
+ {
72
+ "epoch": 0.05,
73
+ "grad_norm": 71.51310729980469,
74
+ "learning_rate": 5e-06,
75
+ "loss": 4.3123,
76
+ "step": 50
77
+ },
78
+ {
79
+ "epoch": 0.05,
80
+ "eval_loss": 4.30466890335083,
81
+ "eval_runtime": 117.1478,
82
+ "eval_samples_per_second": 8.536,
83
+ "eval_steps_per_second": 2.134,
84
+ "step": 50
85
+ },
86
+ {
87
+ "epoch": 0.06,
88
+ "grad_norm": 55.53121566772461,
89
+ "learning_rate": 6e-06,
90
+ "loss": 3.7047,
91
+ "step": 60
92
+ },
93
+ {
94
+ "epoch": 0.06,
95
+ "eval_loss": 3.8378841876983643,
96
+ "eval_runtime": 109.8285,
97
+ "eval_samples_per_second": 9.105,
98
+ "eval_steps_per_second": 2.276,
99
+ "step": 60
100
+ },
101
+ {
102
+ "epoch": 0.07,
103
+ "grad_norm": 16.639528274536133,
104
+ "learning_rate": 7.000000000000001e-06,
105
+ "loss": 4.1835,
106
+ "step": 70
107
+ },
108
+ {
109
+ "epoch": 0.07,
110
+ "eval_loss": 3.6265947818756104,
111
+ "eval_runtime": 107.5258,
112
+ "eval_samples_per_second": 9.3,
113
+ "eval_steps_per_second": 2.325,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.08,
118
+ "grad_norm": 11.000739097595215,
119
+ "learning_rate": 8.000000000000001e-06,
120
+ "loss": 3.0598,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.08,
125
+ "eval_loss": 3.5009496212005615,
126
+ "eval_runtime": 105.6005,
127
+ "eval_samples_per_second": 9.47,
128
+ "eval_steps_per_second": 2.367,
129
+ "step": 80
130
+ },
131
+ {
132
+ "epoch": 0.09,
133
+ "grad_norm": 18.876184463500977,
134
+ "learning_rate": 9e-06,
135
+ "loss": 3.0835,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.09,
140
+ "eval_loss": 3.402592182159424,
141
+ "eval_runtime": 106.8507,
142
+ "eval_samples_per_second": 9.359,
143
+ "eval_steps_per_second": 2.34,
144
+ "step": 90
145
+ },
146
+ {
147
+ "epoch": 0.1,
148
+ "grad_norm": 10.568647384643555,
149
+ "learning_rate": 1e-05,
150
+ "loss": 3.5607,
151
+ "step": 100
152
+ },
153
+ {
154
+ "epoch": 0.1,
155
+ "eval_loss": 3.3077728748321533,
156
+ "eval_runtime": 107.8277,
157
+ "eval_samples_per_second": 9.274,
158
+ "eval_steps_per_second": 2.319,
159
+ "step": 100
160
+ },
161
+ {
162
+ "epoch": 0.11,
163
+ "grad_norm": 6.663860321044922,
164
+ "learning_rate": 1.1000000000000001e-05,
165
+ "loss": 3.3173,
166
+ "step": 110
167
+ },
168
+ {
169
+ "epoch": 0.11,
170
+ "eval_loss": 3.221048355102539,
171
+ "eval_runtime": 104.8207,
172
+ "eval_samples_per_second": 9.54,
173
+ "eval_steps_per_second": 2.385,
174
+ "step": 110
175
+ },
176
+ {
177
+ "epoch": 0.12,
178
+ "grad_norm": 9.180710792541504,
179
+ "learning_rate": 1.2e-05,
180
+ "loss": 3.4318,
181
+ "step": 120
182
+ },
183
+ {
184
+ "epoch": 0.12,
185
+ "eval_loss": 3.1418848037719727,
186
+ "eval_runtime": 104.9529,
187
+ "eval_samples_per_second": 9.528,
188
+ "eval_steps_per_second": 2.382,
189
+ "step": 120
190
+ },
191
+ {
192
+ "epoch": 0.13,
193
+ "grad_norm": 7.3332414627075195,
194
+ "learning_rate": 1.3000000000000001e-05,
195
+ "loss": 3.1961,
196
+ "step": 130
197
+ },
198
+ {
199
+ "epoch": 0.13,
200
+ "eval_loss": 3.069470167160034,
201
+ "eval_runtime": 107.1455,
202
+ "eval_samples_per_second": 9.333,
203
+ "eval_steps_per_second": 2.333,
204
+ "step": 130
205
+ },
206
+ {
207
+ "epoch": 0.14,
208
+ "grad_norm": 11.190179824829102,
209
+ "learning_rate": 1.4000000000000001e-05,
210
+ "loss": 2.821,
211
+ "step": 140
212
+ },
213
+ {
214
+ "epoch": 0.14,
215
+ "eval_loss": 2.998840093612671,
216
+ "eval_runtime": 109.7303,
217
+ "eval_samples_per_second": 9.113,
218
+ "eval_steps_per_second": 2.278,
219
+ "step": 140
220
+ },
221
+ {
222
+ "epoch": 0.15,
223
+ "grad_norm": 10.532099723815918,
224
+ "learning_rate": 1.5e-05,
225
+ "loss": 2.8245,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.15,
230
+ "eval_loss": 2.927992820739746,
231
+ "eval_runtime": 112.3035,
232
+ "eval_samples_per_second": 8.904,
233
+ "eval_steps_per_second": 2.226,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 0.16,
238
+ "grad_norm": 9.649062156677246,
239
+ "learning_rate": 1.6000000000000003e-05,
240
+ "loss": 2.9165,
241
+ "step": 160
242
+ },
243
+ {
244
+ "epoch": 0.16,
245
+ "eval_loss": 2.8786165714263916,
246
+ "eval_runtime": 128.1716,
247
+ "eval_samples_per_second": 7.802,
248
+ "eval_steps_per_second": 1.951,
249
+ "step": 160
250
+ },
251
+ {
252
+ "epoch": 0.17,
253
+ "grad_norm": 9.107309341430664,
254
+ "learning_rate": 1.7000000000000003e-05,
255
+ "loss": 3.1418,
256
+ "step": 170
257
+ },
258
+ {
259
+ "epoch": 0.17,
260
+ "eval_loss": 2.8244595527648926,
261
+ "eval_runtime": 125.0759,
262
+ "eval_samples_per_second": 7.995,
263
+ "eval_steps_per_second": 1.999,
264
+ "step": 170
265
+ },
266
+ {
267
+ "epoch": 0.18,
268
+ "grad_norm": 13.698325157165527,
269
+ "learning_rate": 1.8e-05,
270
+ "loss": 3.0323,
271
+ "step": 180
272
+ },
273
+ {
274
+ "epoch": 0.18,
275
+ "eval_loss": 2.772019386291504,
276
+ "eval_runtime": 129.6329,
277
+ "eval_samples_per_second": 7.714,
278
+ "eval_steps_per_second": 1.929,
279
+ "step": 180
280
+ },
281
+ {
282
+ "epoch": 0.19,
283
+ "grad_norm": 11.200780868530273,
284
+ "learning_rate": 1.9e-05,
285
+ "loss": 2.9199,
286
+ "step": 190
287
+ },
288
+ {
289
+ "epoch": 0.19,
290
+ "eval_loss": 2.7285261154174805,
291
+ "eval_runtime": 105.2039,
292
+ "eval_samples_per_second": 9.505,
293
+ "eval_steps_per_second": 2.376,
294
+ "step": 190
295
+ },
296
+ {
297
+ "epoch": 0.2,
298
+ "grad_norm": 11.479965209960938,
299
+ "learning_rate": 2e-05,
300
+ "loss": 2.6733,
301
+ "step": 200
302
+ },
303
+ {
304
+ "epoch": 0.2,
305
+ "eval_loss": 2.695300340652466,
306
+ "eval_runtime": 102.9704,
307
+ "eval_samples_per_second": 9.712,
308
+ "eval_steps_per_second": 2.428,
309
+ "step": 200
310
+ },
311
+ {
312
+ "epoch": 0.21,
313
+ "grad_norm": 11.359460830688477,
314
+ "learning_rate": 2.1e-05,
315
+ "loss": 2.8493,
316
+ "step": 210
317
+ },
318
+ {
319
+ "epoch": 0.21,
320
+ "eval_loss": 2.66671085357666,
321
+ "eval_runtime": 101.0064,
322
+ "eval_samples_per_second": 9.9,
323
+ "eval_steps_per_second": 2.475,
324
+ "step": 210
325
+ },
326
+ {
327
+ "epoch": 0.22,
328
+ "grad_norm": 13.90817928314209,
329
+ "learning_rate": 2.2000000000000003e-05,
330
+ "loss": 2.8467,
331
+ "step": 220
332
+ },
333
+ {
334
+ "epoch": 0.22,
335
+ "eval_loss": 2.6324076652526855,
336
+ "eval_runtime": 102.7459,
337
+ "eval_samples_per_second": 9.733,
338
+ "eval_steps_per_second": 2.433,
339
+ "step": 220
340
+ },
341
+ {
342
+ "epoch": 0.23,
343
+ "grad_norm": 12.95048999786377,
344
+ "learning_rate": 2.3000000000000003e-05,
345
+ "loss": 2.5137,
346
+ "step": 230
347
+ },
348
+ {
349
+ "epoch": 0.23,
350
+ "eval_loss": 2.599924087524414,
351
+ "eval_runtime": 106.8131,
352
+ "eval_samples_per_second": 9.362,
353
+ "eval_steps_per_second": 2.341,
354
+ "step": 230
355
+ },
356
+ {
357
+ "epoch": 0.24,
358
+ "grad_norm": 10.900901794433594,
359
+ "learning_rate": 2.4e-05,
360
+ "loss": 2.5896,
361
+ "step": 240
362
+ },
363
+ {
364
+ "epoch": 0.24,
365
+ "eval_loss": 2.56974458694458,
366
+ "eval_runtime": 106.0351,
367
+ "eval_samples_per_second": 9.431,
368
+ "eval_steps_per_second": 2.358,
369
+ "step": 240
370
+ },
371
+ {
372
+ "epoch": 0.25,
373
+ "grad_norm": 11.728984832763672,
374
+ "learning_rate": 2.5e-05,
375
+ "loss": 3.0882,
376
+ "step": 250
377
+ },
378
+ {
379
+ "epoch": 0.25,
380
+ "eval_loss": 2.5507752895355225,
381
+ "eval_runtime": 107.1284,
382
+ "eval_samples_per_second": 9.335,
383
+ "eval_steps_per_second": 2.334,
384
+ "step": 250
385
+ },
386
+ {
387
+ "epoch": 0.26,
388
+ "grad_norm": 9.277604103088379,
389
+ "learning_rate": 2.6000000000000002e-05,
390
+ "loss": 2.3151,
391
+ "step": 260
392
+ },
393
+ {
394
+ "epoch": 0.26,
395
+ "eval_loss": 2.525682210922241,
396
+ "eval_runtime": 123.4713,
397
+ "eval_samples_per_second": 8.099,
398
+ "eval_steps_per_second": 2.025,
399
+ "step": 260
400
+ },
401
+ {
402
+ "epoch": 0.27,
403
+ "grad_norm": 10.388689041137695,
404
+ "learning_rate": 2.7000000000000002e-05,
405
+ "loss": 2.5188,
406
+ "step": 270
407
+ },
408
+ {
409
+ "epoch": 0.27,
410
+ "eval_loss": 2.499377489089966,
411
+ "eval_runtime": 127.0825,
412
+ "eval_samples_per_second": 7.869,
413
+ "eval_steps_per_second": 1.967,
414
+ "step": 270
415
+ },
416
+ {
417
+ "epoch": 0.28,
418
+ "grad_norm": 7.337620258331299,
419
+ "learning_rate": 2.8000000000000003e-05,
420
+ "loss": 2.4583,
421
+ "step": 280
422
+ },
423
+ {
424
+ "epoch": 0.28,
425
+ "eval_loss": 2.4751341342926025,
426
+ "eval_runtime": 129.7707,
427
+ "eval_samples_per_second": 7.706,
428
+ "eval_steps_per_second": 1.926,
429
+ "step": 280
430
+ },
431
+ {
432
+ "epoch": 0.29,
433
+ "grad_norm": 9.492632865905762,
434
+ "learning_rate": 2.9e-05,
435
+ "loss": 2.4552,
436
+ "step": 290
437
+ },
438
+ {
439
+ "epoch": 0.29,
440
+ "eval_loss": 2.449791669845581,
441
+ "eval_runtime": 127.0769,
442
+ "eval_samples_per_second": 7.869,
443
+ "eval_steps_per_second": 1.967,
444
+ "step": 290
445
+ },
446
+ {
447
+ "epoch": 0.3,
448
+ "grad_norm": 12.488316535949707,
449
+ "learning_rate": 3e-05,
450
+ "loss": 2.8118,
451
+ "step": 300
452
+ },
453
+ {
454
+ "epoch": 0.3,
455
+ "eval_loss": 2.4329984188079834,
456
+ "eval_runtime": 125.4816,
457
+ "eval_samples_per_second": 7.969,
458
+ "eval_steps_per_second": 1.992,
459
+ "step": 300
460
+ },
461
+ {
462
+ "epoch": 0.31,
463
+ "grad_norm": 10.057647705078125,
464
+ "learning_rate": 3.1e-05,
465
+ "loss": 2.5144,
466
+ "step": 310
467
+ },
468
+ {
469
+ "epoch": 0.31,
470
+ "eval_loss": 2.4146034717559814,
471
+ "eval_runtime": 122.883,
472
+ "eval_samples_per_second": 8.138,
473
+ "eval_steps_per_second": 2.034,
474
+ "step": 310
475
+ },
476
+ {
477
+ "epoch": 0.32,
478
+ "grad_norm": 9.721734046936035,
479
+ "learning_rate": 3.2000000000000005e-05,
480
+ "loss": 2.7473,
481
+ "step": 320
482
+ },
483
+ {
484
+ "epoch": 0.32,
485
+ "eval_loss": 2.4046664237976074,
486
+ "eval_runtime": 113.7917,
487
+ "eval_samples_per_second": 8.788,
488
+ "eval_steps_per_second": 2.197,
489
+ "step": 320
490
+ },
491
+ {
492
+ "epoch": 0.33,
493
+ "grad_norm": 4.945127487182617,
494
+ "learning_rate": 3.3e-05,
495
+ "loss": 2.46,
496
+ "step": 330
497
+ },
498
+ {
499
+ "epoch": 0.33,
500
+ "eval_loss": 2.380826473236084,
501
+ "eval_runtime": 111.3358,
502
+ "eval_samples_per_second": 8.982,
503
+ "eval_steps_per_second": 2.245,
504
+ "step": 330
505
+ },
506
+ {
507
+ "epoch": 0.34,
508
+ "grad_norm": 12.477653503417969,
509
+ "learning_rate": 3.4000000000000007e-05,
510
+ "loss": 2.6254,
511
+ "step": 340
512
+ },
513
+ {
514
+ "epoch": 0.34,
515
+ "eval_loss": 2.371356964111328,
516
+ "eval_runtime": 112.2082,
517
+ "eval_samples_per_second": 8.912,
518
+ "eval_steps_per_second": 2.228,
519
+ "step": 340
520
+ },
521
+ {
522
+ "epoch": 0.35,
523
+ "grad_norm": 10.182991027832031,
524
+ "learning_rate": 3.5e-05,
525
+ "loss": 2.1996,
526
+ "step": 350
527
+ },
528
+ {
529
+ "epoch": 0.35,
530
+ "eval_loss": 2.3486673831939697,
531
+ "eval_runtime": 107.6972,
532
+ "eval_samples_per_second": 9.285,
533
+ "eval_steps_per_second": 2.321,
534
+ "step": 350
535
+ },
536
+ {
537
+ "epoch": 0.36,
538
+ "grad_norm": 10.488390922546387,
539
+ "learning_rate": 3.6e-05,
540
+ "loss": 2.6718,
541
+ "step": 360
542
+ },
543
+ {
544
+ "epoch": 0.36,
545
+ "eval_loss": 2.3371098041534424,
546
+ "eval_runtime": 111.7846,
547
+ "eval_samples_per_second": 8.946,
548
+ "eval_steps_per_second": 2.236,
549
+ "step": 360
550
+ },
551
+ {
552
+ "epoch": 0.37,
553
+ "grad_norm": 7.945727825164795,
554
+ "learning_rate": 3.7e-05,
555
+ "loss": 2.5803,
556
+ "step": 370
557
+ },
558
+ {
559
+ "epoch": 0.37,
560
+ "eval_loss": 2.3198955059051514,
561
+ "eval_runtime": 107.6209,
562
+ "eval_samples_per_second": 9.292,
563
+ "eval_steps_per_second": 2.323,
564
+ "step": 370
565
+ },
566
+ {
567
+ "epoch": 0.38,
568
+ "grad_norm": 11.906257629394531,
569
+ "learning_rate": 3.8e-05,
570
+ "loss": 2.4184,
571
+ "step": 380
572
+ },
573
+ {
574
+ "epoch": 0.38,
575
+ "eval_loss": 2.3155248165130615,
576
+ "eval_runtime": 108.3746,
577
+ "eval_samples_per_second": 9.227,
578
+ "eval_steps_per_second": 2.307,
579
+ "step": 380
580
+ },
581
+ {
582
+ "epoch": 0.39,
583
+ "grad_norm": 11.00644302368164,
584
+ "learning_rate": 3.9000000000000006e-05,
585
+ "loss": 2.2515,
586
+ "step": 390
587
+ },
588
+ {
589
+ "epoch": 0.39,
590
+ "eval_loss": 2.3117854595184326,
591
+ "eval_runtime": 105.7771,
592
+ "eval_samples_per_second": 9.454,
593
+ "eval_steps_per_second": 2.363,
594
+ "step": 390
595
+ },
596
+ {
597
+ "epoch": 0.4,
598
+ "grad_norm": 8.327221870422363,
599
+ "learning_rate": 4e-05,
600
+ "loss": 2.4152,
601
+ "step": 400
602
+ },
603
+ {
604
+ "epoch": 0.4,
605
+ "eval_loss": 2.289562940597534,
606
+ "eval_runtime": 110.3598,
607
+ "eval_samples_per_second": 9.061,
608
+ "eval_steps_per_second": 2.265,
609
+ "step": 400
610
+ },
611
+ {
612
+ "epoch": 0.41,
613
+ "grad_norm": 12.397188186645508,
614
+ "learning_rate": 4.1e-05,
615
+ "loss": 2.2665,
616
+ "step": 410
617
+ },
618
+ {
619
+ "epoch": 0.41,
620
+ "eval_loss": 2.2749555110931396,
621
+ "eval_runtime": 105.9973,
622
+ "eval_samples_per_second": 9.434,
623
+ "eval_steps_per_second": 2.359,
624
+ "step": 410
625
+ },
626
+ {
627
+ "epoch": 0.42,
628
+ "grad_norm": 11.297379493713379,
629
+ "learning_rate": 4.2e-05,
630
+ "loss": 1.6996,
631
+ "step": 420
632
+ },
633
+ {
634
+ "epoch": 0.42,
635
+ "eval_loss": 2.2712552547454834,
636
+ "eval_runtime": 108.2952,
637
+ "eval_samples_per_second": 9.234,
638
+ "eval_steps_per_second": 2.309,
639
+ "step": 420
640
+ },
641
+ {
642
+ "epoch": 0.43,
643
+ "grad_norm": 6.590099811553955,
644
+ "learning_rate": 4.3e-05,
645
+ "loss": 2.0371,
646
+ "step": 430
647
+ },
648
+ {
649
+ "epoch": 0.43,
650
+ "eval_loss": 2.2652268409729004,
651
+ "eval_runtime": 159.4264,
652
+ "eval_samples_per_second": 6.272,
653
+ "eval_steps_per_second": 1.568,
654
+ "step": 430
655
+ },
656
+ {
657
+ "epoch": 0.44,
658
+ "grad_norm": 10.717235565185547,
659
+ "learning_rate": 4.4000000000000006e-05,
660
+ "loss": 2.0536,
661
+ "step": 440
662
+ },
663
+ {
664
+ "epoch": 0.44,
665
+ "eval_loss": 2.254389762878418,
666
+ "eval_runtime": 119.2716,
667
+ "eval_samples_per_second": 8.384,
668
+ "eval_steps_per_second": 2.096,
669
+ "step": 440
670
+ },
671
+ {
672
+ "epoch": 0.45,
673
+ "grad_norm": 12.450675964355469,
674
+ "learning_rate": 4.5e-05,
675
+ "loss": 2.3436,
676
+ "step": 450
677
+ },
678
+ {
679
+ "epoch": 0.45,
680
+ "eval_loss": 2.241189956665039,
681
+ "eval_runtime": 101.1513,
682
+ "eval_samples_per_second": 9.886,
683
+ "eval_steps_per_second": 2.472,
684
+ "step": 450
685
+ },
686
+ {
687
+ "epoch": 0.46,
688
+ "grad_norm": 10.726924896240234,
689
+ "learning_rate": 4.600000000000001e-05,
690
+ "loss": 2.2921,
691
+ "step": 460
692
+ },
693
+ {
694
+ "epoch": 0.46,
695
+ "eval_loss": 2.233902931213379,
696
+ "eval_runtime": 96.5389,
697
+ "eval_samples_per_second": 10.359,
698
+ "eval_steps_per_second": 2.59,
699
+ "step": 460
700
+ },
701
+ {
702
+ "epoch": 0.47,
703
+ "grad_norm": 10.492919921875,
704
+ "learning_rate": 4.7e-05,
705
+ "loss": 2.0704,
706
+ "step": 470
707
+ },
708
+ {
709
+ "epoch": 0.47,
710
+ "eval_loss": 2.223443031311035,
711
+ "eval_runtime": 94.3771,
712
+ "eval_samples_per_second": 10.596,
713
+ "eval_steps_per_second": 2.649,
714
+ "step": 470
715
+ },
716
+ {
717
+ "epoch": 0.48,
718
+ "grad_norm": 12.142420768737793,
719
+ "learning_rate": 4.8e-05,
720
+ "loss": 2.3475,
721
+ "step": 480
722
+ },
723
+ {
724
+ "epoch": 0.48,
725
+ "eval_loss": 2.20881986618042,
726
+ "eval_runtime": 88.7095,
727
+ "eval_samples_per_second": 11.273,
728
+ "eval_steps_per_second": 2.818,
729
+ "step": 480
730
+ },
731
+ {
732
+ "epoch": 0.49,
733
+ "grad_norm": 11.031810760498047,
734
+ "learning_rate": 4.9e-05,
735
+ "loss": 2.3123,
736
+ "step": 490
737
+ },
738
+ {
739
+ "epoch": 0.49,
740
+ "eval_loss": 2.193828582763672,
741
+ "eval_runtime": 90.0998,
742
+ "eval_samples_per_second": 11.099,
743
+ "eval_steps_per_second": 2.775,
744
+ "step": 490
745
+ },
746
+ {
747
+ "epoch": 0.5,
748
+ "grad_norm": 8.5560884475708,
749
+ "learning_rate": 5e-05,
750
+ "loss": 2.4333,
751
+ "step": 500
752
+ },
753
+ {
754
+ "epoch": 0.5,
755
+ "eval_loss": 2.1911978721618652,
756
+ "eval_runtime": 93.2158,
757
+ "eval_samples_per_second": 10.728,
758
+ "eval_steps_per_second": 2.682,
759
+ "step": 500
760
+ },
761
+ {
762
+ "epoch": 0.51,
763
+ "grad_norm": 9.033146858215332,
764
+ "learning_rate": 4.9800000000000004e-05,
765
+ "loss": 1.8733,
766
+ "step": 510
767
+ },
768
+ {
769
+ "epoch": 0.51,
770
+ "eval_loss": 2.1788506507873535,
771
+ "eval_runtime": 92.6001,
772
+ "eval_samples_per_second": 10.799,
773
+ "eval_steps_per_second": 2.7,
774
+ "step": 510
775
+ },
776
+ {
777
+ "epoch": 0.52,
778
+ "grad_norm": 12.906278610229492,
779
+ "learning_rate": 4.96e-05,
780
+ "loss": 2.4447,
781
+ "step": 520
782
+ },
783
+ {
784
+ "epoch": 0.52,
785
+ "eval_loss": 2.170750379562378,
786
+ "eval_runtime": 88.3981,
787
+ "eval_samples_per_second": 11.312,
788
+ "eval_steps_per_second": 2.828,
789
+ "step": 520
790
+ },
791
+ {
792
+ "epoch": 0.53,
793
+ "grad_norm": 13.015795707702637,
794
+ "learning_rate": 4.94e-05,
795
+ "loss": 1.848,
796
+ "step": 530
797
+ },
798
+ {
799
+ "epoch": 0.53,
800
+ "eval_loss": 2.161830186843872,
801
+ "eval_runtime": 91.5465,
802
+ "eval_samples_per_second": 10.923,
803
+ "eval_steps_per_second": 2.731,
804
+ "step": 530
805
+ },
806
+ {
807
+ "epoch": 0.54,
808
+ "grad_norm": 12.636075019836426,
809
+ "learning_rate": 4.92e-05,
810
+ "loss": 2.2895,
811
+ "step": 540
812
+ },
813
+ {
814
+ "epoch": 0.54,
815
+ "eval_loss": 2.1537909507751465,
816
+ "eval_runtime": 86.8374,
817
+ "eval_samples_per_second": 11.516,
818
+ "eval_steps_per_second": 2.879,
819
+ "step": 540
820
+ },
821
+ {
822
+ "epoch": 0.55,
823
+ "grad_norm": 7.789541244506836,
824
+ "learning_rate": 4.9e-05,
825
+ "loss": 1.8584,
826
+ "step": 550
827
+ },
828
+ {
829
+ "epoch": 0.55,
830
+ "eval_loss": 2.148059129714966,
831
+ "eval_runtime": 89.2466,
832
+ "eval_samples_per_second": 11.205,
833
+ "eval_steps_per_second": 2.801,
834
+ "step": 550
835
+ },
836
+ {
837
+ "epoch": 0.56,
838
+ "grad_norm": 10.41247844696045,
839
+ "learning_rate": 4.88e-05,
840
+ "loss": 2.2246,
841
+ "step": 560
842
+ },
843
+ {
844
+ "epoch": 0.56,
845
+ "eval_loss": 2.1307997703552246,
846
+ "eval_runtime": 88.7743,
847
+ "eval_samples_per_second": 11.265,
848
+ "eval_steps_per_second": 2.816,
849
+ "step": 560
850
+ },
851
+ {
852
+ "epoch": 0.57,
853
+ "grad_norm": 8.509031295776367,
854
+ "learning_rate": 4.86e-05,
855
+ "loss": 2.1681,
856
+ "step": 570
857
+ },
858
+ {
859
+ "epoch": 0.57,
860
+ "eval_loss": 2.117568016052246,
861
+ "eval_runtime": 90.3366,
862
+ "eval_samples_per_second": 11.07,
863
+ "eval_steps_per_second": 2.767,
864
+ "step": 570
865
+ },
866
+ {
867
+ "epoch": 0.58,
868
+ "grad_norm": 5.50130558013916,
869
+ "learning_rate": 4.8400000000000004e-05,
870
+ "loss": 1.9694,
871
+ "step": 580
872
+ },
873
+ {
874
+ "epoch": 0.58,
875
+ "eval_loss": 2.111440896987915,
876
+ "eval_runtime": 92.8312,
877
+ "eval_samples_per_second": 10.772,
878
+ "eval_steps_per_second": 2.693,
879
+ "step": 580
880
+ },
881
+ {
882
+ "epoch": 0.59,
883
+ "grad_norm": 10.277608871459961,
884
+ "learning_rate": 4.82e-05,
885
+ "loss": 2.3159,
886
+ "step": 590
887
+ },
888
+ {
889
+ "epoch": 0.59,
890
+ "eval_loss": 2.1107966899871826,
891
+ "eval_runtime": 91.0566,
892
+ "eval_samples_per_second": 10.982,
893
+ "eval_steps_per_second": 2.746,
894
+ "step": 590
895
+ },
896
+ {
897
+ "epoch": 0.6,
898
+ "grad_norm": 10.415202140808105,
899
+ "learning_rate": 4.8e-05,
900
+ "loss": 2.0298,
901
+ "step": 600
902
+ },
903
+ {
904
+ "epoch": 0.6,
905
+ "eval_loss": 2.1007392406463623,
906
+ "eval_runtime": 90.2452,
907
+ "eval_samples_per_second": 11.081,
908
+ "eval_steps_per_second": 2.77,
909
+ "step": 600
910
+ },
911
+ {
912
+ "epoch": 0.61,
913
+ "grad_norm": 9.051558494567871,
914
+ "learning_rate": 4.78e-05,
915
+ "loss": 2.1856,
916
+ "step": 610
917
+ },
918
+ {
919
+ "epoch": 0.61,
920
+ "eval_loss": 2.0893638134002686,
921
+ "eval_runtime": 97.6967,
922
+ "eval_samples_per_second": 10.236,
923
+ "eval_steps_per_second": 2.559,
924
+ "step": 610
925
+ },
926
+ {
927
+ "epoch": 0.62,
928
+ "grad_norm": 6.953254222869873,
929
+ "learning_rate": 4.76e-05,
930
+ "loss": 1.8901,
931
+ "step": 620
932
+ },
933
+ {
934
+ "epoch": 0.62,
935
+ "eval_loss": 2.0824646949768066,
936
+ "eval_runtime": 91.0278,
937
+ "eval_samples_per_second": 10.986,
938
+ "eval_steps_per_second": 2.746,
939
+ "step": 620
940
+ },
941
+ {
942
+ "epoch": 0.63,
943
+ "grad_norm": 7.948352813720703,
944
+ "learning_rate": 4.74e-05,
945
+ "loss": 2.3327,
946
+ "step": 630
947
+ },
948
+ {
949
+ "epoch": 0.63,
950
+ "eval_loss": 2.074693441390991,
951
+ "eval_runtime": 101.8312,
952
+ "eval_samples_per_second": 9.82,
953
+ "eval_steps_per_second": 2.455,
954
+ "step": 630
955
+ },
956
+ {
957
+ "epoch": 0.64,
958
+ "grad_norm": 9.384662628173828,
959
+ "learning_rate": 4.72e-05,
960
+ "loss": 2.3253,
961
+ "step": 640
962
+ },
963
+ {
964
+ "epoch": 0.64,
965
+ "eval_loss": 2.067964792251587,
966
+ "eval_runtime": 93.9189,
967
+ "eval_samples_per_second": 10.647,
968
+ "eval_steps_per_second": 2.662,
969
+ "step": 640
970
+ },
971
+ {
972
+ "epoch": 0.65,
973
+ "grad_norm": 6.1795759201049805,
974
+ "learning_rate": 4.7e-05,
975
+ "loss": 2.189,
976
+ "step": 650
977
+ },
978
+ {
979
+ "epoch": 0.65,
980
+ "eval_loss": 2.0557608604431152,
981
+ "eval_runtime": 97.6001,
982
+ "eval_samples_per_second": 10.246,
983
+ "eval_steps_per_second": 2.561,
984
+ "step": 650
985
+ },
986
+ {
987
+ "epoch": 0.66,
988
+ "grad_norm": 9.642065048217773,
989
+ "learning_rate": 4.6800000000000006e-05,
990
+ "loss": 2.146,
991
+ "step": 660
992
+ },
993
+ {
994
+ "epoch": 0.66,
995
+ "eval_loss": 2.0541083812713623,
996
+ "eval_runtime": 98.8714,
997
+ "eval_samples_per_second": 10.114,
998
+ "eval_steps_per_second": 2.529,
999
+ "step": 660
1000
+ },
1001
+ {
1002
+ "epoch": 0.67,
1003
+ "grad_norm": 6.987886428833008,
1004
+ "learning_rate": 4.660000000000001e-05,
1005
+ "loss": 1.8356,
1006
+ "step": 670
1007
+ },
1008
+ {
1009
+ "epoch": 0.67,
1010
+ "eval_loss": 2.053222417831421,
1011
+ "eval_runtime": 90.3869,
1012
+ "eval_samples_per_second": 11.064,
1013
+ "eval_steps_per_second": 2.766,
1014
+ "step": 670
1015
+ },
1016
+ {
1017
+ "epoch": 0.68,
1018
+ "grad_norm": 10.291470527648926,
1019
+ "learning_rate": 4.64e-05,
1020
+ "loss": 2.3633,
1021
+ "step": 680
1022
+ },
1023
+ {
1024
+ "epoch": 0.68,
1025
+ "eval_loss": 2.0410830974578857,
1026
+ "eval_runtime": 92.923,
1027
+ "eval_samples_per_second": 10.762,
1028
+ "eval_steps_per_second": 2.69,
1029
+ "step": 680
1030
+ },
1031
+ {
1032
+ "epoch": 0.69,
1033
+ "grad_norm": 8.183449745178223,
1034
+ "learning_rate": 4.6200000000000005e-05,
1035
+ "loss": 2.1429,
1036
+ "step": 690
1037
+ },
1038
+ {
1039
+ "epoch": 0.69,
1040
+ "eval_loss": 2.035935640335083,
1041
+ "eval_runtime": 106.6596,
1042
+ "eval_samples_per_second": 9.376,
1043
+ "eval_steps_per_second": 2.344,
1044
+ "step": 690
1045
+ },
1046
+ {
1047
+ "epoch": 0.7,
1048
+ "grad_norm": 7.329190254211426,
1049
+ "learning_rate": 4.600000000000001e-05,
1050
+ "loss": 2.2302,
1051
+ "step": 700
1052
+ },
1053
+ {
1054
+ "epoch": 0.7,
1055
+ "eval_loss": 2.033024311065674,
1056
+ "eval_runtime": 100.5784,
1057
+ "eval_samples_per_second": 9.942,
1058
+ "eval_steps_per_second": 2.486,
1059
+ "step": 700
1060
+ },
1061
+ {
1062
+ "epoch": 0.71,
1063
+ "grad_norm": 9.20626163482666,
1064
+ "learning_rate": 4.58e-05,
1065
+ "loss": 2.0641,
1066
+ "step": 710
1067
+ },
1068
+ {
1069
+ "epoch": 0.71,
1070
+ "eval_loss": 2.025780439376831,
1071
+ "eval_runtime": 112.3754,
1072
+ "eval_samples_per_second": 8.899,
1073
+ "eval_steps_per_second": 2.225,
1074
+ "step": 710
1075
+ },
1076
+ {
1077
+ "epoch": 0.72,
1078
+ "grad_norm": 11.579280853271484,
1079
+ "learning_rate": 4.5600000000000004e-05,
1080
+ "loss": 2.0106,
1081
+ "step": 720
1082
+ },
1083
+ {
1084
+ "epoch": 0.72,
1085
+ "eval_loss": 2.023242950439453,
1086
+ "eval_runtime": 100.8727,
1087
+ "eval_samples_per_second": 9.913,
1088
+ "eval_steps_per_second": 2.478,
1089
+ "step": 720
1090
+ },
1091
+ {
1092
+ "epoch": 0.73,
1093
+ "grad_norm": 9.62726879119873,
1094
+ "learning_rate": 4.5400000000000006e-05,
1095
+ "loss": 2.1251,
1096
+ "step": 730
1097
+ },
1098
+ {
1099
+ "epoch": 0.73,
1100
+ "eval_loss": 2.019562244415283,
1101
+ "eval_runtime": 98.6758,
1102
+ "eval_samples_per_second": 10.134,
1103
+ "eval_steps_per_second": 2.534,
1104
+ "step": 730
1105
+ },
1106
+ {
1107
+ "epoch": 0.74,
1108
+ "grad_norm": 8.502236366271973,
1109
+ "learning_rate": 4.52e-05,
1110
+ "loss": 2.1068,
1111
+ "step": 740
1112
+ },
1113
+ {
1114
+ "epoch": 0.74,
1115
+ "eval_loss": 2.0147552490234375,
1116
+ "eval_runtime": 101.3373,
1117
+ "eval_samples_per_second": 9.868,
1118
+ "eval_steps_per_second": 2.467,
1119
+ "step": 740
1120
+ },
1121
+ {
1122
+ "epoch": 0.75,
1123
+ "grad_norm": 8.522944450378418,
1124
+ "learning_rate": 4.5e-05,
1125
+ "loss": 2.3803,
1126
+ "step": 750
1127
+ },
1128
+ {
1129
+ "epoch": 0.75,
1130
+ "eval_loss": 2.0065298080444336,
1131
+ "eval_runtime": 115.4706,
1132
+ "eval_samples_per_second": 8.66,
1133
+ "eval_steps_per_second": 2.165,
1134
+ "step": 750
1135
+ },
1136
+ {
1137
+ "epoch": 0.76,
1138
+ "grad_norm": 9.323762893676758,
1139
+ "learning_rate": 4.4800000000000005e-05,
1140
+ "loss": 2.2126,
1141
+ "step": 760
1142
+ },
1143
+ {
1144
+ "epoch": 0.76,
1145
+ "eval_loss": 2.000988006591797,
1146
+ "eval_runtime": 131.4565,
1147
+ "eval_samples_per_second": 7.607,
1148
+ "eval_steps_per_second": 1.902,
1149
+ "step": 760
1150
+ },
1151
+ {
1152
+ "epoch": 0.77,
1153
+ "grad_norm": 7.243586540222168,
1154
+ "learning_rate": 4.46e-05,
1155
+ "loss": 2.2732,
1156
+ "step": 770
1157
+ },
1158
+ {
1159
+ "epoch": 0.77,
1160
+ "eval_loss": 1.9990744590759277,
1161
+ "eval_runtime": 117.2201,
1162
+ "eval_samples_per_second": 8.531,
1163
+ "eval_steps_per_second": 2.133,
1164
+ "step": 770
1165
+ },
1166
+ {
1167
+ "epoch": 0.78,
1168
+ "grad_norm": 10.664429664611816,
1169
+ "learning_rate": 4.44e-05,
1170
+ "loss": 2.2521,
1171
+ "step": 780
1172
+ },
1173
+ {
1174
+ "epoch": 0.78,
1175
+ "eval_loss": 1.988065242767334,
1176
+ "eval_runtime": 105.5882,
1177
+ "eval_samples_per_second": 9.471,
1178
+ "eval_steps_per_second": 2.368,
1179
+ "step": 780
1180
+ },
1181
+ {
1182
+ "epoch": 0.79,
1183
+ "grad_norm": 9.04030704498291,
1184
+ "learning_rate": 4.4200000000000004e-05,
1185
+ "loss": 1.9933,
1186
+ "step": 790
1187
+ },
1188
+ {
1189
+ "epoch": 0.79,
1190
+ "eval_loss": 1.9845080375671387,
1191
+ "eval_runtime": 101.4877,
1192
+ "eval_samples_per_second": 9.853,
1193
+ "eval_steps_per_second": 2.463,
1194
+ "step": 790
1195
+ },
1196
+ {
1197
+ "epoch": 0.8,
1198
+ "grad_norm": 9.38503646850586,
1199
+ "learning_rate": 4.4000000000000006e-05,
1200
+ "loss": 1.9956,
1201
+ "step": 800
1202
+ },
1203
+ {
1204
+ "epoch": 0.8,
1205
+ "eval_loss": 1.9831095933914185,
1206
+ "eval_runtime": 148.7473,
1207
+ "eval_samples_per_second": 6.723,
1208
+ "eval_steps_per_second": 1.681,
1209
+ "step": 800
1210
+ },
1211
+ {
1212
+ "epoch": 0.81,
1213
+ "grad_norm": 6.387115955352783,
1214
+ "learning_rate": 4.38e-05,
1215
+ "loss": 2.0059,
1216
+ "step": 810
1217
+ },
1218
+ {
1219
+ "epoch": 0.81,
1220
+ "eval_loss": 1.976269006729126,
1221
+ "eval_runtime": 129.9491,
1222
+ "eval_samples_per_second": 7.695,
1223
+ "eval_steps_per_second": 1.924,
1224
+ "step": 810
1225
+ },
1226
+ {
1227
+ "epoch": 0.82,
1228
+ "grad_norm": 8.890172004699707,
1229
+ "learning_rate": 4.36e-05,
1230
+ "loss": 2.2373,
1231
+ "step": 820
1232
+ },
1233
+ {
1234
+ "epoch": 0.82,
1235
+ "eval_loss": 1.974748969078064,
1236
+ "eval_runtime": 164.1513,
1237
+ "eval_samples_per_second": 6.092,
1238
+ "eval_steps_per_second": 1.523,
1239
+ "step": 820
1240
+ },
1241
+ {
1242
+ "epoch": 0.83,
1243
+ "grad_norm": 10.762927055358887,
1244
+ "learning_rate": 4.3400000000000005e-05,
1245
+ "loss": 2.186,
1246
+ "step": 830
1247
+ },
1248
+ {
1249
+ "epoch": 0.83,
1250
+ "eval_loss": 1.9657306671142578,
1251
+ "eval_runtime": 142.9199,
1252
+ "eval_samples_per_second": 6.997,
1253
+ "eval_steps_per_second": 1.749,
1254
+ "step": 830
1255
+ },
1256
+ {
1257
+ "epoch": 0.84,
1258
+ "grad_norm": 9.816293716430664,
1259
+ "learning_rate": 4.32e-05,
1260
+ "loss": 1.6506,
1261
+ "step": 840
1262
+ },
1263
+ {
1264
+ "epoch": 0.84,
1265
+ "eval_loss": 1.9642084836959839,
1266
+ "eval_runtime": 173.606,
1267
+ "eval_samples_per_second": 5.76,
1268
+ "eval_steps_per_second": 1.44,
1269
+ "step": 840
1270
+ },
1271
+ {
1272
+ "epoch": 0.85,
1273
+ "grad_norm": 7.931281089782715,
1274
+ "learning_rate": 4.3e-05,
1275
+ "loss": 1.7588,
1276
+ "step": 850
1277
+ },
1278
+ {
1279
+ "epoch": 0.85,
1280
+ "eval_loss": 1.9623171091079712,
1281
+ "eval_runtime": 158.121,
1282
+ "eval_samples_per_second": 6.324,
1283
+ "eval_steps_per_second": 1.581,
1284
+ "step": 850
1285
+ },
1286
+ {
1287
+ "epoch": 0.86,
1288
+ "grad_norm": 8.325611114501953,
1289
+ "learning_rate": 4.2800000000000004e-05,
1290
+ "loss": 1.8801,
1291
+ "step": 860
1292
+ },
1293
+ {
1294
+ "epoch": 0.86,
1295
+ "eval_loss": 1.9604990482330322,
1296
+ "eval_runtime": 156.6306,
1297
+ "eval_samples_per_second": 6.384,
1298
+ "eval_steps_per_second": 1.596,
1299
+ "step": 860
1300
+ },
1301
+ {
1302
+ "epoch": 0.87,
1303
+ "grad_norm": 8.10270881652832,
1304
+ "learning_rate": 4.26e-05,
1305
+ "loss": 1.8894,
1306
+ "step": 870
1307
+ },
1308
+ {
1309
+ "epoch": 0.87,
1310
+ "eval_loss": 1.9559086561203003,
1311
+ "eval_runtime": 165.5477,
1312
+ "eval_samples_per_second": 6.041,
1313
+ "eval_steps_per_second": 1.51,
1314
+ "step": 870
1315
+ },
1316
+ {
1317
+ "epoch": 0.88,
1318
+ "grad_norm": 8.666308403015137,
1319
+ "learning_rate": 4.24e-05,
1320
+ "loss": 2.1945,
1321
+ "step": 880
1322
+ },
1323
+ {
1324
+ "epoch": 0.88,
1325
+ "eval_loss": 1.953808307647705,
1326
+ "eval_runtime": 155.1083,
1327
+ "eval_samples_per_second": 6.447,
1328
+ "eval_steps_per_second": 1.612,
1329
+ "step": 880
1330
+ },
1331
+ {
1332
+ "epoch": 0.89,
1333
+ "grad_norm": 7.216370105743408,
1334
+ "learning_rate": 4.22e-05,
1335
+ "loss": 2.1958,
1336
+ "step": 890
1337
+ },
1338
+ {
1339
+ "epoch": 0.89,
1340
+ "eval_loss": 1.9536200761795044,
1341
+ "eval_runtime": 153.3145,
1342
+ "eval_samples_per_second": 6.523,
1343
+ "eval_steps_per_second": 1.631,
1344
+ "step": 890
1345
+ },
1346
+ {
1347
+ "epoch": 0.9,
1348
+ "grad_norm": 7.325380802154541,
1349
+ "learning_rate": 4.2e-05,
1350
+ "loss": 1.8137,
1351
+ "step": 900
1352
+ },
1353
+ {
1354
+ "epoch": 0.9,
1355
+ "eval_loss": 1.9444538354873657,
1356
+ "eval_runtime": 157.7008,
1357
+ "eval_samples_per_second": 6.341,
1358
+ "eval_steps_per_second": 1.585,
1359
+ "step": 900
1360
+ },
1361
+ {
1362
+ "epoch": 0.91,
1363
+ "grad_norm": 6.625280857086182,
1364
+ "learning_rate": 4.18e-05,
1365
+ "loss": 1.8925,
1366
+ "step": 910
1367
+ },
1368
+ {
1369
+ "epoch": 0.91,
1370
+ "eval_loss": 1.9409335851669312,
1371
+ "eval_runtime": 149.5155,
1372
+ "eval_samples_per_second": 6.688,
1373
+ "eval_steps_per_second": 1.672,
1374
+ "step": 910
1375
+ },
1376
+ {
1377
+ "epoch": 0.92,
1378
+ "grad_norm": 5.897461414337158,
1379
+ "learning_rate": 4.16e-05,
1380
+ "loss": 1.8322,
1381
+ "step": 920
1382
+ },
1383
+ {
1384
+ "epoch": 0.92,
1385
+ "eval_loss": 1.9331930875778198,
1386
+ "eval_runtime": 127.1983,
1387
+ "eval_samples_per_second": 7.862,
1388
+ "eval_steps_per_second": 1.965,
1389
+ "step": 920
1390
+ },
1391
+ {
1392
+ "epoch": 0.93,
1393
+ "grad_norm": 9.08010482788086,
1394
+ "learning_rate": 4.14e-05,
1395
+ "loss": 2.042,
1396
+ "step": 930
1397
+ },
1398
+ {
1399
+ "epoch": 0.93,
1400
+ "eval_loss": 1.9316186904907227,
1401
+ "eval_runtime": 118.4354,
1402
+ "eval_samples_per_second": 8.443,
1403
+ "eval_steps_per_second": 2.111,
1404
+ "step": 930
1405
+ },
1406
+ {
1407
+ "epoch": 0.94,
1408
+ "grad_norm": 9.264948844909668,
1409
+ "learning_rate": 4.12e-05,
1410
+ "loss": 1.7826,
1411
+ "step": 940
1412
+ },
1413
+ {
1414
+ "epoch": 0.94,
1415
+ "eval_loss": 1.9262688159942627,
1416
+ "eval_runtime": 113.7982,
1417
+ "eval_samples_per_second": 8.787,
1418
+ "eval_steps_per_second": 2.197,
1419
+ "step": 940
1420
+ },
1421
+ {
1422
+ "epoch": 0.95,
1423
+ "grad_norm": 10.431904792785645,
1424
+ "learning_rate": 4.1e-05,
1425
+ "loss": 1.7883,
1426
+ "step": 950
1427
+ },
1428
+ {
1429
+ "epoch": 0.95,
1430
+ "eval_loss": 1.9221816062927246,
1431
+ "eval_runtime": 107.3673,
1432
+ "eval_samples_per_second": 9.314,
1433
+ "eval_steps_per_second": 2.328,
1434
+ "step": 950
1435
+ },
1436
+ {
1437
+ "epoch": 0.96,
1438
+ "grad_norm": 5.923398017883301,
1439
+ "learning_rate": 4.08e-05,
1440
+ "loss": 1.5646,
1441
+ "step": 960
1442
+ },
1443
+ {
1444
+ "epoch": 0.96,
1445
+ "eval_loss": 1.9227478504180908,
1446
+ "eval_runtime": 121.0769,
1447
+ "eval_samples_per_second": 8.259,
1448
+ "eval_steps_per_second": 2.065,
1449
+ "step": 960
1450
+ },
1451
+ {
1452
+ "epoch": 0.97,
1453
+ "grad_norm": 8.440842628479004,
1454
+ "learning_rate": 4.0600000000000004e-05,
1455
+ "loss": 1.6706,
1456
+ "step": 970
1457
+ },
1458
+ {
1459
+ "epoch": 0.97,
1460
+ "eval_loss": 1.9174282550811768,
1461
+ "eval_runtime": 121.92,
1462
+ "eval_samples_per_second": 8.202,
1463
+ "eval_steps_per_second": 2.051,
1464
+ "step": 970
1465
+ },
1466
+ {
1467
+ "epoch": 0.98,
1468
+ "grad_norm": 6.066262245178223,
1469
+ "learning_rate": 4.0400000000000006e-05,
1470
+ "loss": 1.6543,
1471
+ "step": 980
1472
+ },
1473
+ {
1474
+ "epoch": 0.98,
1475
+ "eval_loss": 1.915424108505249,
1476
+ "eval_runtime": 111.8461,
1477
+ "eval_samples_per_second": 8.941,
1478
+ "eval_steps_per_second": 2.235,
1479
+ "step": 980
1480
+ },
1481
+ {
1482
+ "epoch": 0.99,
1483
+ "grad_norm": 7.314166069030762,
1484
+ "learning_rate": 4.02e-05,
1485
+ "loss": 2.1565,
1486
+ "step": 990
1487
+ },
1488
+ {
1489
+ "epoch": 0.99,
1490
+ "eval_loss": 1.9187248945236206,
1491
+ "eval_runtime": 110.4362,
1492
+ "eval_samples_per_second": 9.055,
1493
+ "eval_steps_per_second": 2.264,
1494
+ "step": 990
1495
+ },
1496
+ {
1497
+ "epoch": 1.0,
1498
+ "grad_norm": 3.6229448318481445,
1499
+ "learning_rate": 4e-05,
1500
+ "loss": 1.7919,
1501
+ "step": 1000
1502
+ },
1503
+ {
1504
+ "epoch": 1.0,
1505
+ "eval_loss": 1.9095934629440308,
1506
+ "eval_runtime": 103.6218,
1507
+ "eval_samples_per_second": 9.65,
1508
+ "eval_steps_per_second": 2.413,
1509
+ "step": 1000
1510
+ }
1511
+ ],
1512
+ "logging_steps": 10,
1513
+ "max_steps": 3000,
1514
+ "num_input_tokens_seen": 0,
1515
+ "num_train_epochs": 3,
1516
+ "save_steps": 500,
1517
+ "stateful_callbacks": {
1518
+ "TrainerControl": {
1519
+ "args": {
1520
+ "should_epoch_stop": false,
1521
+ "should_evaluate": false,
1522
+ "should_log": false,
1523
+ "should_save": true,
1524
+ "should_training_stop": false
1525
+ },
1526
+ "attributes": {}
1527
+ }
1528
+ },
1529
+ "total_flos": 261292032000000.0,
1530
+ "train_batch_size": 4,
1531
+ "trial_name": null,
1532
+ "trial_params": null
1533
+ }
fine_tuned_gpt2/checkpoint-1500/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/Users/user/projects/MindPadiBig/mindpadi/models/gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.49.0",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
fine_tuned_gpt2/checkpoint-1500/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.49.0"
6
+ }
fine_tuned_gpt2/checkpoint-1500/trainer_state.json ADDED
@@ -0,0 +1,2283 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.803325891494751,
3
+ "best_model_checkpoint": "/Users/user/projects/MindPadiBig/mindpadi/models/fine_tuned_gpt2/checkpoint-1500",
4
+ "epoch": 1.5,
5
+ "eval_steps": 10,
6
+ "global_step": 1500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "grad_norm": 126.21493530273438,
14
+ "learning_rate": 1.0000000000000002e-06,
15
+ "loss": 6.9352,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.01,
20
+ "eval_loss": 6.456012725830078,
21
+ "eval_runtime": 128.7117,
22
+ "eval_samples_per_second": 7.769,
23
+ "eval_steps_per_second": 1.942,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02,
28
+ "grad_norm": 112.57171630859375,
29
+ "learning_rate": 2.0000000000000003e-06,
30
+ "loss": 6.2877,
31
+ "step": 20
32
+ },
33
+ {
34
+ "epoch": 0.02,
35
+ "eval_loss": 6.174145698547363,
36
+ "eval_runtime": 133.4573,
37
+ "eval_samples_per_second": 7.493,
38
+ "eval_steps_per_second": 1.873,
39
+ "step": 20
40
+ },
41
+ {
42
+ "epoch": 0.03,
43
+ "grad_norm": 185.73190307617188,
44
+ "learning_rate": 3e-06,
45
+ "loss": 5.9569,
46
+ "step": 30
47
+ },
48
+ {
49
+ "epoch": 0.03,
50
+ "eval_loss": 5.594101428985596,
51
+ "eval_runtime": 136.3272,
52
+ "eval_samples_per_second": 7.335,
53
+ "eval_steps_per_second": 1.834,
54
+ "step": 30
55
+ },
56
+ {
57
+ "epoch": 0.04,
58
+ "grad_norm": 160.77322387695312,
59
+ "learning_rate": 4.000000000000001e-06,
60
+ "loss": 4.9783,
61
+ "step": 40
62
+ },
63
+ {
64
+ "epoch": 0.04,
65
+ "eval_loss": 4.95811128616333,
66
+ "eval_runtime": 123.0408,
67
+ "eval_samples_per_second": 8.127,
68
+ "eval_steps_per_second": 2.032,
69
+ "step": 40
70
+ },
71
+ {
72
+ "epoch": 0.05,
73
+ "grad_norm": 71.51310729980469,
74
+ "learning_rate": 5e-06,
75
+ "loss": 4.3123,
76
+ "step": 50
77
+ },
78
+ {
79
+ "epoch": 0.05,
80
+ "eval_loss": 4.30466890335083,
81
+ "eval_runtime": 117.1478,
82
+ "eval_samples_per_second": 8.536,
83
+ "eval_steps_per_second": 2.134,
84
+ "step": 50
85
+ },
86
+ {
87
+ "epoch": 0.06,
88
+ "grad_norm": 55.53121566772461,
89
+ "learning_rate": 6e-06,
90
+ "loss": 3.7047,
91
+ "step": 60
92
+ },
93
+ {
94
+ "epoch": 0.06,
95
+ "eval_loss": 3.8378841876983643,
96
+ "eval_runtime": 109.8285,
97
+ "eval_samples_per_second": 9.105,
98
+ "eval_steps_per_second": 2.276,
99
+ "step": 60
100
+ },
101
+ {
102
+ "epoch": 0.07,
103
+ "grad_norm": 16.639528274536133,
104
+ "learning_rate": 7.000000000000001e-06,
105
+ "loss": 4.1835,
106
+ "step": 70
107
+ },
108
+ {
109
+ "epoch": 0.07,
110
+ "eval_loss": 3.6265947818756104,
111
+ "eval_runtime": 107.5258,
112
+ "eval_samples_per_second": 9.3,
113
+ "eval_steps_per_second": 2.325,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.08,
118
+ "grad_norm": 11.000739097595215,
119
+ "learning_rate": 8.000000000000001e-06,
120
+ "loss": 3.0598,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.08,
125
+ "eval_loss": 3.5009496212005615,
126
+ "eval_runtime": 105.6005,
127
+ "eval_samples_per_second": 9.47,
128
+ "eval_steps_per_second": 2.367,
129
+ "step": 80
130
+ },
131
+ {
132
+ "epoch": 0.09,
133
+ "grad_norm": 18.876184463500977,
134
+ "learning_rate": 9e-06,
135
+ "loss": 3.0835,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.09,
140
+ "eval_loss": 3.402592182159424,
141
+ "eval_runtime": 106.8507,
142
+ "eval_samples_per_second": 9.359,
143
+ "eval_steps_per_second": 2.34,
144
+ "step": 90
145
+ },
146
+ {
147
+ "epoch": 0.1,
148
+ "grad_norm": 10.568647384643555,
149
+ "learning_rate": 1e-05,
150
+ "loss": 3.5607,
151
+ "step": 100
152
+ },
153
+ {
154
+ "epoch": 0.1,
155
+ "eval_loss": 3.3077728748321533,
156
+ "eval_runtime": 107.8277,
157
+ "eval_samples_per_second": 9.274,
158
+ "eval_steps_per_second": 2.319,
159
+ "step": 100
160
+ },
161
+ {
162
+ "epoch": 0.11,
163
+ "grad_norm": 6.663860321044922,
164
+ "learning_rate": 1.1000000000000001e-05,
165
+ "loss": 3.3173,
166
+ "step": 110
167
+ },
168
+ {
169
+ "epoch": 0.11,
170
+ "eval_loss": 3.221048355102539,
171
+ "eval_runtime": 104.8207,
172
+ "eval_samples_per_second": 9.54,
173
+ "eval_steps_per_second": 2.385,
174
+ "step": 110
175
+ },
176
+ {
177
+ "epoch": 0.12,
178
+ "grad_norm": 9.180710792541504,
179
+ "learning_rate": 1.2e-05,
180
+ "loss": 3.4318,
181
+ "step": 120
182
+ },
183
+ {
184
+ "epoch": 0.12,
185
+ "eval_loss": 3.1418848037719727,
186
+ "eval_runtime": 104.9529,
187
+ "eval_samples_per_second": 9.528,
188
+ "eval_steps_per_second": 2.382,
189
+ "step": 120
190
+ },
191
+ {
192
+ "epoch": 0.13,
193
+ "grad_norm": 7.3332414627075195,
194
+ "learning_rate": 1.3000000000000001e-05,
195
+ "loss": 3.1961,
196
+ "step": 130
197
+ },
198
+ {
199
+ "epoch": 0.13,
200
+ "eval_loss": 3.069470167160034,
201
+ "eval_runtime": 107.1455,
202
+ "eval_samples_per_second": 9.333,
203
+ "eval_steps_per_second": 2.333,
204
+ "step": 130
205
+ },
206
+ {
207
+ "epoch": 0.14,
208
+ "grad_norm": 11.190179824829102,
209
+ "learning_rate": 1.4000000000000001e-05,
210
+ "loss": 2.821,
211
+ "step": 140
212
+ },
213
+ {
214
+ "epoch": 0.14,
215
+ "eval_loss": 2.998840093612671,
216
+ "eval_runtime": 109.7303,
217
+ "eval_samples_per_second": 9.113,
218
+ "eval_steps_per_second": 2.278,
219
+ "step": 140
220
+ },
221
+ {
222
+ "epoch": 0.15,
223
+ "grad_norm": 10.532099723815918,
224
+ "learning_rate": 1.5e-05,
225
+ "loss": 2.8245,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.15,
230
+ "eval_loss": 2.927992820739746,
231
+ "eval_runtime": 112.3035,
232
+ "eval_samples_per_second": 8.904,
233
+ "eval_steps_per_second": 2.226,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 0.16,
238
+ "grad_norm": 9.649062156677246,
239
+ "learning_rate": 1.6000000000000003e-05,
240
+ "loss": 2.9165,
241
+ "step": 160
242
+ },
243
+ {
244
+ "epoch": 0.16,
245
+ "eval_loss": 2.8786165714263916,
246
+ "eval_runtime": 128.1716,
247
+ "eval_samples_per_second": 7.802,
248
+ "eval_steps_per_second": 1.951,
249
+ "step": 160
250
+ },
251
+ {
252
+ "epoch": 0.17,
253
+ "grad_norm": 9.107309341430664,
254
+ "learning_rate": 1.7000000000000003e-05,
255
+ "loss": 3.1418,
256
+ "step": 170
257
+ },
258
+ {
259
+ "epoch": 0.17,
260
+ "eval_loss": 2.8244595527648926,
261
+ "eval_runtime": 125.0759,
262
+ "eval_samples_per_second": 7.995,
263
+ "eval_steps_per_second": 1.999,
264
+ "step": 170
265
+ },
266
+ {
267
+ "epoch": 0.18,
268
+ "grad_norm": 13.698325157165527,
269
+ "learning_rate": 1.8e-05,
270
+ "loss": 3.0323,
271
+ "step": 180
272
+ },
273
+ {
274
+ "epoch": 0.18,
275
+ "eval_loss": 2.772019386291504,
276
+ "eval_runtime": 129.6329,
277
+ "eval_samples_per_second": 7.714,
278
+ "eval_steps_per_second": 1.929,
279
+ "step": 180
280
+ },
281
+ {
282
+ "epoch": 0.19,
283
+ "grad_norm": 11.200780868530273,
284
+ "learning_rate": 1.9e-05,
285
+ "loss": 2.9199,
286
+ "step": 190
287
+ },
288
+ {
289
+ "epoch": 0.19,
290
+ "eval_loss": 2.7285261154174805,
291
+ "eval_runtime": 105.2039,
292
+ "eval_samples_per_second": 9.505,
293
+ "eval_steps_per_second": 2.376,
294
+ "step": 190
295
+ },
296
+ {
297
+ "epoch": 0.2,
298
+ "grad_norm": 11.479965209960938,
299
+ "learning_rate": 2e-05,
300
+ "loss": 2.6733,
301
+ "step": 200
302
+ },
303
+ {
304
+ "epoch": 0.2,
305
+ "eval_loss": 2.695300340652466,
306
+ "eval_runtime": 102.9704,
307
+ "eval_samples_per_second": 9.712,
308
+ "eval_steps_per_second": 2.428,
309
+ "step": 200
310
+ },
311
+ {
312
+ "epoch": 0.21,
313
+ "grad_norm": 11.359460830688477,
314
+ "learning_rate": 2.1e-05,
315
+ "loss": 2.8493,
316
+ "step": 210
317
+ },
318
+ {
319
+ "epoch": 0.21,
320
+ "eval_loss": 2.66671085357666,
321
+ "eval_runtime": 101.0064,
322
+ "eval_samples_per_second": 9.9,
323
+ "eval_steps_per_second": 2.475,
324
+ "step": 210
325
+ },
326
+ {
327
+ "epoch": 0.22,
328
+ "grad_norm": 13.90817928314209,
329
+ "learning_rate": 2.2000000000000003e-05,
330
+ "loss": 2.8467,
331
+ "step": 220
332
+ },
333
+ {
334
+ "epoch": 0.22,
335
+ "eval_loss": 2.6324076652526855,
336
+ "eval_runtime": 102.7459,
337
+ "eval_samples_per_second": 9.733,
338
+ "eval_steps_per_second": 2.433,
339
+ "step": 220
340
+ },
341
+ {
342
+ "epoch": 0.23,
343
+ "grad_norm": 12.95048999786377,
344
+ "learning_rate": 2.3000000000000003e-05,
345
+ "loss": 2.5137,
346
+ "step": 230
347
+ },
348
+ {
349
+ "epoch": 0.23,
350
+ "eval_loss": 2.599924087524414,
351
+ "eval_runtime": 106.8131,
352
+ "eval_samples_per_second": 9.362,
353
+ "eval_steps_per_second": 2.341,
354
+ "step": 230
355
+ },
356
+ {
357
+ "epoch": 0.24,
358
+ "grad_norm": 10.900901794433594,
359
+ "learning_rate": 2.4e-05,
360
+ "loss": 2.5896,
361
+ "step": 240
362
+ },
363
+ {
364
+ "epoch": 0.24,
365
+ "eval_loss": 2.56974458694458,
366
+ "eval_runtime": 106.0351,
367
+ "eval_samples_per_second": 9.431,
368
+ "eval_steps_per_second": 2.358,
369
+ "step": 240
370
+ },
371
+ {
372
+ "epoch": 0.25,
373
+ "grad_norm": 11.728984832763672,
374
+ "learning_rate": 2.5e-05,
375
+ "loss": 3.0882,
376
+ "step": 250
377
+ },
378
+ {
379
+ "epoch": 0.25,
380
+ "eval_loss": 2.5507752895355225,
381
+ "eval_runtime": 107.1284,
382
+ "eval_samples_per_second": 9.335,
383
+ "eval_steps_per_second": 2.334,
384
+ "step": 250
385
+ },
386
+ {
387
+ "epoch": 0.26,
388
+ "grad_norm": 9.277604103088379,
389
+ "learning_rate": 2.6000000000000002e-05,
390
+ "loss": 2.3151,
391
+ "step": 260
392
+ },
393
+ {
394
+ "epoch": 0.26,
395
+ "eval_loss": 2.525682210922241,
396
+ "eval_runtime": 123.4713,
397
+ "eval_samples_per_second": 8.099,
398
+ "eval_steps_per_second": 2.025,
399
+ "step": 260
400
+ },
401
+ {
402
+ "epoch": 0.27,
403
+ "grad_norm": 10.388689041137695,
404
+ "learning_rate": 2.7000000000000002e-05,
405
+ "loss": 2.5188,
406
+ "step": 270
407
+ },
408
+ {
409
+ "epoch": 0.27,
410
+ "eval_loss": 2.499377489089966,
411
+ "eval_runtime": 127.0825,
412
+ "eval_samples_per_second": 7.869,
413
+ "eval_steps_per_second": 1.967,
414
+ "step": 270
415
+ },
416
+ {
417
+ "epoch": 0.28,
418
+ "grad_norm": 7.337620258331299,
419
+ "learning_rate": 2.8000000000000003e-05,
420
+ "loss": 2.4583,
421
+ "step": 280
422
+ },
423
+ {
424
+ "epoch": 0.28,
425
+ "eval_loss": 2.4751341342926025,
426
+ "eval_runtime": 129.7707,
427
+ "eval_samples_per_second": 7.706,
428
+ "eval_steps_per_second": 1.926,
429
+ "step": 280
430
+ },
431
+ {
432
+ "epoch": 0.29,
433
+ "grad_norm": 9.492632865905762,
434
+ "learning_rate": 2.9e-05,
435
+ "loss": 2.4552,
436
+ "step": 290
437
+ },
438
+ {
439
+ "epoch": 0.29,
440
+ "eval_loss": 2.449791669845581,
441
+ "eval_runtime": 127.0769,
442
+ "eval_samples_per_second": 7.869,
443
+ "eval_steps_per_second": 1.967,
444
+ "step": 290
445
+ },
446
+ {
447
+ "epoch": 0.3,
448
+ "grad_norm": 12.488316535949707,
449
+ "learning_rate": 3e-05,
450
+ "loss": 2.8118,
451
+ "step": 300
452
+ },
453
+ {
454
+ "epoch": 0.3,
455
+ "eval_loss": 2.4329984188079834,
456
+ "eval_runtime": 125.4816,
457
+ "eval_samples_per_second": 7.969,
458
+ "eval_steps_per_second": 1.992,
459
+ "step": 300
460
+ },
461
+ {
462
+ "epoch": 0.31,
463
+ "grad_norm": 10.057647705078125,
464
+ "learning_rate": 3.1e-05,
465
+ "loss": 2.5144,
466
+ "step": 310
467
+ },
468
+ {
469
+ "epoch": 0.31,
470
+ "eval_loss": 2.4146034717559814,
471
+ "eval_runtime": 122.883,
472
+ "eval_samples_per_second": 8.138,
473
+ "eval_steps_per_second": 2.034,
474
+ "step": 310
475
+ },
476
+ {
477
+ "epoch": 0.32,
478
+ "grad_norm": 9.721734046936035,
479
+ "learning_rate": 3.2000000000000005e-05,
480
+ "loss": 2.7473,
481
+ "step": 320
482
+ },
483
+ {
484
+ "epoch": 0.32,
485
+ "eval_loss": 2.4046664237976074,
486
+ "eval_runtime": 113.7917,
487
+ "eval_samples_per_second": 8.788,
488
+ "eval_steps_per_second": 2.197,
489
+ "step": 320
490
+ },
491
+ {
492
+ "epoch": 0.33,
493
+ "grad_norm": 4.945127487182617,
494
+ "learning_rate": 3.3e-05,
495
+ "loss": 2.46,
496
+ "step": 330
497
+ },
498
+ {
499
+ "epoch": 0.33,
500
+ "eval_loss": 2.380826473236084,
501
+ "eval_runtime": 111.3358,
502
+ "eval_samples_per_second": 8.982,
503
+ "eval_steps_per_second": 2.245,
504
+ "step": 330
505
+ },
506
+ {
507
+ "epoch": 0.34,
508
+ "grad_norm": 12.477653503417969,
509
+ "learning_rate": 3.4000000000000007e-05,
510
+ "loss": 2.6254,
511
+ "step": 340
512
+ },
513
+ {
514
+ "epoch": 0.34,
515
+ "eval_loss": 2.371356964111328,
516
+ "eval_runtime": 112.2082,
517
+ "eval_samples_per_second": 8.912,
518
+ "eval_steps_per_second": 2.228,
519
+ "step": 340
520
+ },
521
+ {
522
+ "epoch": 0.35,
523
+ "grad_norm": 10.182991027832031,
524
+ "learning_rate": 3.5e-05,
525
+ "loss": 2.1996,
526
+ "step": 350
527
+ },
528
+ {
529
+ "epoch": 0.35,
530
+ "eval_loss": 2.3486673831939697,
531
+ "eval_runtime": 107.6972,
532
+ "eval_samples_per_second": 9.285,
533
+ "eval_steps_per_second": 2.321,
534
+ "step": 350
535
+ },
536
+ {
537
+ "epoch": 0.36,
538
+ "grad_norm": 10.488390922546387,
539
+ "learning_rate": 3.6e-05,
540
+ "loss": 2.6718,
541
+ "step": 360
542
+ },
543
+ {
544
+ "epoch": 0.36,
545
+ "eval_loss": 2.3371098041534424,
546
+ "eval_runtime": 111.7846,
547
+ "eval_samples_per_second": 8.946,
548
+ "eval_steps_per_second": 2.236,
549
+ "step": 360
550
+ },
551
+ {
552
+ "epoch": 0.37,
553
+ "grad_norm": 7.945727825164795,
554
+ "learning_rate": 3.7e-05,
555
+ "loss": 2.5803,
556
+ "step": 370
557
+ },
558
+ {
559
+ "epoch": 0.37,
560
+ "eval_loss": 2.3198955059051514,
561
+ "eval_runtime": 107.6209,
562
+ "eval_samples_per_second": 9.292,
563
+ "eval_steps_per_second": 2.323,
564
+ "step": 370
565
+ },
566
+ {
567
+ "epoch": 0.38,
568
+ "grad_norm": 11.906257629394531,
569
+ "learning_rate": 3.8e-05,
570
+ "loss": 2.4184,
571
+ "step": 380
572
+ },
573
+ {
574
+ "epoch": 0.38,
575
+ "eval_loss": 2.3155248165130615,
576
+ "eval_runtime": 108.3746,
577
+ "eval_samples_per_second": 9.227,
578
+ "eval_steps_per_second": 2.307,
579
+ "step": 380
580
+ },
581
+ {
582
+ "epoch": 0.39,
583
+ "grad_norm": 11.00644302368164,
584
+ "learning_rate": 3.9000000000000006e-05,
585
+ "loss": 2.2515,
586
+ "step": 390
587
+ },
588
+ {
589
+ "epoch": 0.39,
590
+ "eval_loss": 2.3117854595184326,
591
+ "eval_runtime": 105.7771,
592
+ "eval_samples_per_second": 9.454,
593
+ "eval_steps_per_second": 2.363,
594
+ "step": 390
595
+ },
596
+ {
597
+ "epoch": 0.4,
598
+ "grad_norm": 8.327221870422363,
599
+ "learning_rate": 4e-05,
600
+ "loss": 2.4152,
601
+ "step": 400
602
+ },
603
+ {
604
+ "epoch": 0.4,
605
+ "eval_loss": 2.289562940597534,
606
+ "eval_runtime": 110.3598,
607
+ "eval_samples_per_second": 9.061,
608
+ "eval_steps_per_second": 2.265,
609
+ "step": 400
610
+ },
611
+ {
612
+ "epoch": 0.41,
613
+ "grad_norm": 12.397188186645508,
614
+ "learning_rate": 4.1e-05,
615
+ "loss": 2.2665,
616
+ "step": 410
617
+ },
618
+ {
619
+ "epoch": 0.41,
620
+ "eval_loss": 2.2749555110931396,
621
+ "eval_runtime": 105.9973,
622
+ "eval_samples_per_second": 9.434,
623
+ "eval_steps_per_second": 2.359,
624
+ "step": 410
625
+ },
626
+ {
627
+ "epoch": 0.42,
628
+ "grad_norm": 11.297379493713379,
629
+ "learning_rate": 4.2e-05,
630
+ "loss": 1.6996,
631
+ "step": 420
632
+ },
633
+ {
634
+ "epoch": 0.42,
635
+ "eval_loss": 2.2712552547454834,
636
+ "eval_runtime": 108.2952,
637
+ "eval_samples_per_second": 9.234,
638
+ "eval_steps_per_second": 2.309,
639
+ "step": 420
640
+ },
641
+ {
642
+ "epoch": 0.43,
643
+ "grad_norm": 6.590099811553955,
644
+ "learning_rate": 4.3e-05,
645
+ "loss": 2.0371,
646
+ "step": 430
647
+ },
648
+ {
649
+ "epoch": 0.43,
650
+ "eval_loss": 2.2652268409729004,
651
+ "eval_runtime": 159.4264,
652
+ "eval_samples_per_second": 6.272,
653
+ "eval_steps_per_second": 1.568,
654
+ "step": 430
655
+ },
656
+ {
657
+ "epoch": 0.44,
658
+ "grad_norm": 10.717235565185547,
659
+ "learning_rate": 4.4000000000000006e-05,
660
+ "loss": 2.0536,
661
+ "step": 440
662
+ },
663
+ {
664
+ "epoch": 0.44,
665
+ "eval_loss": 2.254389762878418,
666
+ "eval_runtime": 119.2716,
667
+ "eval_samples_per_second": 8.384,
668
+ "eval_steps_per_second": 2.096,
669
+ "step": 440
670
+ },
671
+ {
672
+ "epoch": 0.45,
673
+ "grad_norm": 12.450675964355469,
674
+ "learning_rate": 4.5e-05,
675
+ "loss": 2.3436,
676
+ "step": 450
677
+ },
678
+ {
679
+ "epoch": 0.45,
680
+ "eval_loss": 2.241189956665039,
681
+ "eval_runtime": 101.1513,
682
+ "eval_samples_per_second": 9.886,
683
+ "eval_steps_per_second": 2.472,
684
+ "step": 450
685
+ },
686
+ {
687
+ "epoch": 0.46,
688
+ "grad_norm": 10.726924896240234,
689
+ "learning_rate": 4.600000000000001e-05,
690
+ "loss": 2.2921,
691
+ "step": 460
692
+ },
693
+ {
694
+ "epoch": 0.46,
695
+ "eval_loss": 2.233902931213379,
696
+ "eval_runtime": 96.5389,
697
+ "eval_samples_per_second": 10.359,
698
+ "eval_steps_per_second": 2.59,
699
+ "step": 460
700
+ },
701
+ {
702
+ "epoch": 0.47,
703
+ "grad_norm": 10.492919921875,
704
+ "learning_rate": 4.7e-05,
705
+ "loss": 2.0704,
706
+ "step": 470
707
+ },
708
+ {
709
+ "epoch": 0.47,
710
+ "eval_loss": 2.223443031311035,
711
+ "eval_runtime": 94.3771,
712
+ "eval_samples_per_second": 10.596,
713
+ "eval_steps_per_second": 2.649,
714
+ "step": 470
715
+ },
716
+ {
717
+ "epoch": 0.48,
718
+ "grad_norm": 12.142420768737793,
719
+ "learning_rate": 4.8e-05,
720
+ "loss": 2.3475,
721
+ "step": 480
722
+ },
723
+ {
724
+ "epoch": 0.48,
725
+ "eval_loss": 2.20881986618042,
726
+ "eval_runtime": 88.7095,
727
+ "eval_samples_per_second": 11.273,
728
+ "eval_steps_per_second": 2.818,
729
+ "step": 480
730
+ },
731
+ {
732
+ "epoch": 0.49,
733
+ "grad_norm": 11.031810760498047,
734
+ "learning_rate": 4.9e-05,
735
+ "loss": 2.3123,
736
+ "step": 490
737
+ },
738
+ {
739
+ "epoch": 0.49,
740
+ "eval_loss": 2.193828582763672,
741
+ "eval_runtime": 90.0998,
742
+ "eval_samples_per_second": 11.099,
743
+ "eval_steps_per_second": 2.775,
744
+ "step": 490
745
+ },
746
+ {
747
+ "epoch": 0.5,
748
+ "grad_norm": 8.5560884475708,
749
+ "learning_rate": 5e-05,
750
+ "loss": 2.4333,
751
+ "step": 500
752
+ },
753
+ {
754
+ "epoch": 0.5,
755
+ "eval_loss": 2.1911978721618652,
756
+ "eval_runtime": 93.2158,
757
+ "eval_samples_per_second": 10.728,
758
+ "eval_steps_per_second": 2.682,
759
+ "step": 500
760
+ },
761
+ {
762
+ "epoch": 0.51,
763
+ "grad_norm": 9.033146858215332,
764
+ "learning_rate": 4.9800000000000004e-05,
765
+ "loss": 1.8733,
766
+ "step": 510
767
+ },
768
+ {
769
+ "epoch": 0.51,
770
+ "eval_loss": 2.1788506507873535,
771
+ "eval_runtime": 92.6001,
772
+ "eval_samples_per_second": 10.799,
773
+ "eval_steps_per_second": 2.7,
774
+ "step": 510
775
+ },
776
+ {
777
+ "epoch": 0.52,
778
+ "grad_norm": 12.906278610229492,
779
+ "learning_rate": 4.96e-05,
780
+ "loss": 2.4447,
781
+ "step": 520
782
+ },
783
+ {
784
+ "epoch": 0.52,
785
+ "eval_loss": 2.170750379562378,
786
+ "eval_runtime": 88.3981,
787
+ "eval_samples_per_second": 11.312,
788
+ "eval_steps_per_second": 2.828,
789
+ "step": 520
790
+ },
791
+ {
792
+ "epoch": 0.53,
793
+ "grad_norm": 13.015795707702637,
794
+ "learning_rate": 4.94e-05,
795
+ "loss": 1.848,
796
+ "step": 530
797
+ },
798
+ {
799
+ "epoch": 0.53,
800
+ "eval_loss": 2.161830186843872,
801
+ "eval_runtime": 91.5465,
802
+ "eval_samples_per_second": 10.923,
803
+ "eval_steps_per_second": 2.731,
804
+ "step": 530
805
+ },
806
+ {
807
+ "epoch": 0.54,
808
+ "grad_norm": 12.636075019836426,
809
+ "learning_rate": 4.92e-05,
810
+ "loss": 2.2895,
811
+ "step": 540
812
+ },
813
+ {
814
+ "epoch": 0.54,
815
+ "eval_loss": 2.1537909507751465,
816
+ "eval_runtime": 86.8374,
817
+ "eval_samples_per_second": 11.516,
818
+ "eval_steps_per_second": 2.879,
819
+ "step": 540
820
+ },
821
+ {
822
+ "epoch": 0.55,
823
+ "grad_norm": 7.789541244506836,
824
+ "learning_rate": 4.9e-05,
825
+ "loss": 1.8584,
826
+ "step": 550
827
+ },
828
+ {
829
+ "epoch": 0.55,
830
+ "eval_loss": 2.148059129714966,
831
+ "eval_runtime": 89.2466,
832
+ "eval_samples_per_second": 11.205,
833
+ "eval_steps_per_second": 2.801,
834
+ "step": 550
835
+ },
836
+ {
837
+ "epoch": 0.56,
838
+ "grad_norm": 10.41247844696045,
839
+ "learning_rate": 4.88e-05,
840
+ "loss": 2.2246,
841
+ "step": 560
842
+ },
843
+ {
844
+ "epoch": 0.56,
845
+ "eval_loss": 2.1307997703552246,
846
+ "eval_runtime": 88.7743,
847
+ "eval_samples_per_second": 11.265,
848
+ "eval_steps_per_second": 2.816,
849
+ "step": 560
850
+ },
851
+ {
852
+ "epoch": 0.57,
853
+ "grad_norm": 8.509031295776367,
854
+ "learning_rate": 4.86e-05,
855
+ "loss": 2.1681,
856
+ "step": 570
857
+ },
858
+ {
859
+ "epoch": 0.57,
860
+ "eval_loss": 2.117568016052246,
861
+ "eval_runtime": 90.3366,
862
+ "eval_samples_per_second": 11.07,
863
+ "eval_steps_per_second": 2.767,
864
+ "step": 570
865
+ },
866
+ {
867
+ "epoch": 0.58,
868
+ "grad_norm": 5.50130558013916,
869
+ "learning_rate": 4.8400000000000004e-05,
870
+ "loss": 1.9694,
871
+ "step": 580
872
+ },
873
+ {
874
+ "epoch": 0.58,
875
+ "eval_loss": 2.111440896987915,
876
+ "eval_runtime": 92.8312,
877
+ "eval_samples_per_second": 10.772,
878
+ "eval_steps_per_second": 2.693,
879
+ "step": 580
880
+ },
881
+ {
882
+ "epoch": 0.59,
883
+ "grad_norm": 10.277608871459961,
884
+ "learning_rate": 4.82e-05,
885
+ "loss": 2.3159,
886
+ "step": 590
887
+ },
888
+ {
889
+ "epoch": 0.59,
890
+ "eval_loss": 2.1107966899871826,
891
+ "eval_runtime": 91.0566,
892
+ "eval_samples_per_second": 10.982,
893
+ "eval_steps_per_second": 2.746,
894
+ "step": 590
895
+ },
896
+ {
897
+ "epoch": 0.6,
898
+ "grad_norm": 10.415202140808105,
899
+ "learning_rate": 4.8e-05,
900
+ "loss": 2.0298,
901
+ "step": 600
902
+ },
903
+ {
904
+ "epoch": 0.6,
905
+ "eval_loss": 2.1007392406463623,
906
+ "eval_runtime": 90.2452,
907
+ "eval_samples_per_second": 11.081,
908
+ "eval_steps_per_second": 2.77,
909
+ "step": 600
910
+ },
911
+ {
912
+ "epoch": 0.61,
913
+ "grad_norm": 9.051558494567871,
914
+ "learning_rate": 4.78e-05,
915
+ "loss": 2.1856,
916
+ "step": 610
917
+ },
918
+ {
919
+ "epoch": 0.61,
920
+ "eval_loss": 2.0893638134002686,
921
+ "eval_runtime": 97.6967,
922
+ "eval_samples_per_second": 10.236,
923
+ "eval_steps_per_second": 2.559,
924
+ "step": 610
925
+ },
926
+ {
927
+ "epoch": 0.62,
928
+ "grad_norm": 6.953254222869873,
929
+ "learning_rate": 4.76e-05,
930
+ "loss": 1.8901,
931
+ "step": 620
932
+ },
933
+ {
934
+ "epoch": 0.62,
935
+ "eval_loss": 2.0824646949768066,
936
+ "eval_runtime": 91.0278,
937
+ "eval_samples_per_second": 10.986,
938
+ "eval_steps_per_second": 2.746,
939
+ "step": 620
940
+ },
941
+ {
942
+ "epoch": 0.63,
943
+ "grad_norm": 7.948352813720703,
944
+ "learning_rate": 4.74e-05,
945
+ "loss": 2.3327,
946
+ "step": 630
947
+ },
948
+ {
949
+ "epoch": 0.63,
950
+ "eval_loss": 2.074693441390991,
951
+ "eval_runtime": 101.8312,
952
+ "eval_samples_per_second": 9.82,
953
+ "eval_steps_per_second": 2.455,
954
+ "step": 630
955
+ },
956
+ {
957
+ "epoch": 0.64,
958
+ "grad_norm": 9.384662628173828,
959
+ "learning_rate": 4.72e-05,
960
+ "loss": 2.3253,
961
+ "step": 640
962
+ },
963
+ {
964
+ "epoch": 0.64,
965
+ "eval_loss": 2.067964792251587,
966
+ "eval_runtime": 93.9189,
967
+ "eval_samples_per_second": 10.647,
968
+ "eval_steps_per_second": 2.662,
969
+ "step": 640
970
+ },
971
+ {
972
+ "epoch": 0.65,
973
+ "grad_norm": 6.1795759201049805,
974
+ "learning_rate": 4.7e-05,
975
+ "loss": 2.189,
976
+ "step": 650
977
+ },
978
+ {
979
+ "epoch": 0.65,
980
+ "eval_loss": 2.0557608604431152,
981
+ "eval_runtime": 97.6001,
982
+ "eval_samples_per_second": 10.246,
983
+ "eval_steps_per_second": 2.561,
984
+ "step": 650
985
+ },
986
+ {
987
+ "epoch": 0.66,
988
+ "grad_norm": 9.642065048217773,
989
+ "learning_rate": 4.6800000000000006e-05,
990
+ "loss": 2.146,
991
+ "step": 660
992
+ },
993
+ {
994
+ "epoch": 0.66,
995
+ "eval_loss": 2.0541083812713623,
996
+ "eval_runtime": 98.8714,
997
+ "eval_samples_per_second": 10.114,
998
+ "eval_steps_per_second": 2.529,
999
+ "step": 660
1000
+ },
1001
+ {
1002
+ "epoch": 0.67,
1003
+ "grad_norm": 6.987886428833008,
1004
+ "learning_rate": 4.660000000000001e-05,
1005
+ "loss": 1.8356,
1006
+ "step": 670
1007
+ },
1008
+ {
1009
+ "epoch": 0.67,
1010
+ "eval_loss": 2.053222417831421,
1011
+ "eval_runtime": 90.3869,
1012
+ "eval_samples_per_second": 11.064,
1013
+ "eval_steps_per_second": 2.766,
1014
+ "step": 670
1015
+ },
1016
+ {
1017
+ "epoch": 0.68,
1018
+ "grad_norm": 10.291470527648926,
1019
+ "learning_rate": 4.64e-05,
1020
+ "loss": 2.3633,
1021
+ "step": 680
1022
+ },
1023
+ {
1024
+ "epoch": 0.68,
1025
+ "eval_loss": 2.0410830974578857,
1026
+ "eval_runtime": 92.923,
1027
+ "eval_samples_per_second": 10.762,
1028
+ "eval_steps_per_second": 2.69,
1029
+ "step": 680
1030
+ },
1031
+ {
1032
+ "epoch": 0.69,
1033
+ "grad_norm": 8.183449745178223,
1034
+ "learning_rate": 4.6200000000000005e-05,
1035
+ "loss": 2.1429,
1036
+ "step": 690
1037
+ },
1038
+ {
1039
+ "epoch": 0.69,
1040
+ "eval_loss": 2.035935640335083,
1041
+ "eval_runtime": 106.6596,
1042
+ "eval_samples_per_second": 9.376,
1043
+ "eval_steps_per_second": 2.344,
1044
+ "step": 690
1045
+ },
1046
+ {
1047
+ "epoch": 0.7,
1048
+ "grad_norm": 7.329190254211426,
1049
+ "learning_rate": 4.600000000000001e-05,
1050
+ "loss": 2.2302,
1051
+ "step": 700
1052
+ },
1053
+ {
1054
+ "epoch": 0.7,
1055
+ "eval_loss": 2.033024311065674,
1056
+ "eval_runtime": 100.5784,
1057
+ "eval_samples_per_second": 9.942,
1058
+ "eval_steps_per_second": 2.486,
1059
+ "step": 700
1060
+ },
1061
+ {
1062
+ "epoch": 0.71,
1063
+ "grad_norm": 9.20626163482666,
1064
+ "learning_rate": 4.58e-05,
1065
+ "loss": 2.0641,
1066
+ "step": 710
1067
+ },
1068
+ {
1069
+ "epoch": 0.71,
1070
+ "eval_loss": 2.025780439376831,
1071
+ "eval_runtime": 112.3754,
1072
+ "eval_samples_per_second": 8.899,
1073
+ "eval_steps_per_second": 2.225,
1074
+ "step": 710
1075
+ },
1076
+ {
1077
+ "epoch": 0.72,
1078
+ "grad_norm": 11.579280853271484,
1079
+ "learning_rate": 4.5600000000000004e-05,
1080
+ "loss": 2.0106,
1081
+ "step": 720
1082
+ },
1083
+ {
1084
+ "epoch": 0.72,
1085
+ "eval_loss": 2.023242950439453,
1086
+ "eval_runtime": 100.8727,
1087
+ "eval_samples_per_second": 9.913,
1088
+ "eval_steps_per_second": 2.478,
1089
+ "step": 720
1090
+ },
1091
+ {
1092
+ "epoch": 0.73,
1093
+ "grad_norm": 9.62726879119873,
1094
+ "learning_rate": 4.5400000000000006e-05,
1095
+ "loss": 2.1251,
1096
+ "step": 730
1097
+ },
1098
+ {
1099
+ "epoch": 0.73,
1100
+ "eval_loss": 2.019562244415283,
1101
+ "eval_runtime": 98.6758,
1102
+ "eval_samples_per_second": 10.134,
1103
+ "eval_steps_per_second": 2.534,
1104
+ "step": 730
1105
+ },
1106
+ {
1107
+ "epoch": 0.74,
1108
+ "grad_norm": 8.502236366271973,
1109
+ "learning_rate": 4.52e-05,
1110
+ "loss": 2.1068,
1111
+ "step": 740
1112
+ },
1113
+ {
1114
+ "epoch": 0.74,
1115
+ "eval_loss": 2.0147552490234375,
1116
+ "eval_runtime": 101.3373,
1117
+ "eval_samples_per_second": 9.868,
1118
+ "eval_steps_per_second": 2.467,
1119
+ "step": 740
1120
+ },
1121
+ {
1122
+ "epoch": 0.75,
1123
+ "grad_norm": 8.522944450378418,
1124
+ "learning_rate": 4.5e-05,
1125
+ "loss": 2.3803,
1126
+ "step": 750
1127
+ },
1128
+ {
1129
+ "epoch": 0.75,
1130
+ "eval_loss": 2.0065298080444336,
1131
+ "eval_runtime": 115.4706,
1132
+ "eval_samples_per_second": 8.66,
1133
+ "eval_steps_per_second": 2.165,
1134
+ "step": 750
1135
+ },
1136
+ {
1137
+ "epoch": 0.76,
1138
+ "grad_norm": 9.323762893676758,
1139
+ "learning_rate": 4.4800000000000005e-05,
1140
+ "loss": 2.2126,
1141
+ "step": 760
1142
+ },
1143
+ {
1144
+ "epoch": 0.76,
1145
+ "eval_loss": 2.000988006591797,
1146
+ "eval_runtime": 131.4565,
1147
+ "eval_samples_per_second": 7.607,
1148
+ "eval_steps_per_second": 1.902,
1149
+ "step": 760
1150
+ },
1151
+ {
1152
+ "epoch": 0.77,
1153
+ "grad_norm": 7.243586540222168,
1154
+ "learning_rate": 4.46e-05,
1155
+ "loss": 2.2732,
1156
+ "step": 770
1157
+ },
1158
+ {
1159
+ "epoch": 0.77,
1160
+ "eval_loss": 1.9990744590759277,
1161
+ "eval_runtime": 117.2201,
1162
+ "eval_samples_per_second": 8.531,
1163
+ "eval_steps_per_second": 2.133,
1164
+ "step": 770
1165
+ },
1166
+ {
1167
+ "epoch": 0.78,
1168
+ "grad_norm": 10.664429664611816,
1169
+ "learning_rate": 4.44e-05,
1170
+ "loss": 2.2521,
1171
+ "step": 780
1172
+ },
1173
+ {
1174
+ "epoch": 0.78,
1175
+ "eval_loss": 1.988065242767334,
1176
+ "eval_runtime": 105.5882,
1177
+ "eval_samples_per_second": 9.471,
1178
+ "eval_steps_per_second": 2.368,
1179
+ "step": 780
1180
+ },
1181
+ {
1182
+ "epoch": 0.79,
1183
+ "grad_norm": 9.04030704498291,
1184
+ "learning_rate": 4.4200000000000004e-05,
1185
+ "loss": 1.9933,
1186
+ "step": 790
1187
+ },
1188
+ {
1189
+ "epoch": 0.79,
1190
+ "eval_loss": 1.9845080375671387,
1191
+ "eval_runtime": 101.4877,
1192
+ "eval_samples_per_second": 9.853,
1193
+ "eval_steps_per_second": 2.463,
1194
+ "step": 790
1195
+ },
1196
+ {
1197
+ "epoch": 0.8,
1198
+ "grad_norm": 9.38503646850586,
1199
+ "learning_rate": 4.4000000000000006e-05,
1200
+ "loss": 1.9956,
1201
+ "step": 800
1202
+ },
1203
+ {
1204
+ "epoch": 0.8,
1205
+ "eval_loss": 1.9831095933914185,
1206
+ "eval_runtime": 148.7473,
1207
+ "eval_samples_per_second": 6.723,
1208
+ "eval_steps_per_second": 1.681,
1209
+ "step": 800
1210
+ },
1211
+ {
1212
+ "epoch": 0.81,
1213
+ "grad_norm": 6.387115955352783,
1214
+ "learning_rate": 4.38e-05,
1215
+ "loss": 2.0059,
1216
+ "step": 810
1217
+ },
1218
+ {
1219
+ "epoch": 0.81,
1220
+ "eval_loss": 1.976269006729126,
1221
+ "eval_runtime": 129.9491,
1222
+ "eval_samples_per_second": 7.695,
1223
+ "eval_steps_per_second": 1.924,
1224
+ "step": 810
1225
+ },
1226
+ {
1227
+ "epoch": 0.82,
1228
+ "grad_norm": 8.890172004699707,
1229
+ "learning_rate": 4.36e-05,
1230
+ "loss": 2.2373,
1231
+ "step": 820
1232
+ },
1233
+ {
1234
+ "epoch": 0.82,
1235
+ "eval_loss": 1.974748969078064,
1236
+ "eval_runtime": 164.1513,
1237
+ "eval_samples_per_second": 6.092,
1238
+ "eval_steps_per_second": 1.523,
1239
+ "step": 820
1240
+ },
1241
+ {
1242
+ "epoch": 0.83,
1243
+ "grad_norm": 10.762927055358887,
1244
+ "learning_rate": 4.3400000000000005e-05,
1245
+ "loss": 2.186,
1246
+ "step": 830
1247
+ },
1248
+ {
1249
+ "epoch": 0.83,
1250
+ "eval_loss": 1.9657306671142578,
1251
+ "eval_runtime": 142.9199,
1252
+ "eval_samples_per_second": 6.997,
1253
+ "eval_steps_per_second": 1.749,
1254
+ "step": 830
1255
+ },
1256
+ {
1257
+ "epoch": 0.84,
1258
+ "grad_norm": 9.816293716430664,
1259
+ "learning_rate": 4.32e-05,
1260
+ "loss": 1.6506,
1261
+ "step": 840
1262
+ },
1263
+ {
1264
+ "epoch": 0.84,
1265
+ "eval_loss": 1.9642084836959839,
1266
+ "eval_runtime": 173.606,
1267
+ "eval_samples_per_second": 5.76,
1268
+ "eval_steps_per_second": 1.44,
1269
+ "step": 840
1270
+ },
1271
+ {
1272
+ "epoch": 0.85,
1273
+ "grad_norm": 7.931281089782715,
1274
+ "learning_rate": 4.3e-05,
1275
+ "loss": 1.7588,
1276
+ "step": 850
1277
+ },
1278
+ {
1279
+ "epoch": 0.85,
1280
+ "eval_loss": 1.9623171091079712,
1281
+ "eval_runtime": 158.121,
1282
+ "eval_samples_per_second": 6.324,
1283
+ "eval_steps_per_second": 1.581,
1284
+ "step": 850
1285
+ },
1286
+ {
1287
+ "epoch": 0.86,
1288
+ "grad_norm": 8.325611114501953,
1289
+ "learning_rate": 4.2800000000000004e-05,
1290
+ "loss": 1.8801,
1291
+ "step": 860
1292
+ },
1293
+ {
1294
+ "epoch": 0.86,
1295
+ "eval_loss": 1.9604990482330322,
1296
+ "eval_runtime": 156.6306,
1297
+ "eval_samples_per_second": 6.384,
1298
+ "eval_steps_per_second": 1.596,
1299
+ "step": 860
1300
+ },
1301
+ {
1302
+ "epoch": 0.87,
1303
+ "grad_norm": 8.10270881652832,
1304
+ "learning_rate": 4.26e-05,
1305
+ "loss": 1.8894,
1306
+ "step": 870
1307
+ },
1308
+ {
1309
+ "epoch": 0.87,
1310
+ "eval_loss": 1.9559086561203003,
1311
+ "eval_runtime": 165.5477,
1312
+ "eval_samples_per_second": 6.041,
1313
+ "eval_steps_per_second": 1.51,
1314
+ "step": 870
1315
+ },
1316
+ {
1317
+ "epoch": 0.88,
1318
+ "grad_norm": 8.666308403015137,
1319
+ "learning_rate": 4.24e-05,
1320
+ "loss": 2.1945,
1321
+ "step": 880
1322
+ },
1323
+ {
1324
+ "epoch": 0.88,
1325
+ "eval_loss": 1.953808307647705,
1326
+ "eval_runtime": 155.1083,
1327
+ "eval_samples_per_second": 6.447,
1328
+ "eval_steps_per_second": 1.612,
1329
+ "step": 880
1330
+ },
1331
+ {
1332
+ "epoch": 0.89,
1333
+ "grad_norm": 7.216370105743408,
1334
+ "learning_rate": 4.22e-05,
1335
+ "loss": 2.1958,
1336
+ "step": 890
1337
+ },
1338
+ {
1339
+ "epoch": 0.89,
1340
+ "eval_loss": 1.9536200761795044,
1341
+ "eval_runtime": 153.3145,
1342
+ "eval_samples_per_second": 6.523,
1343
+ "eval_steps_per_second": 1.631,
1344
+ "step": 890
1345
+ },
1346
+ {
1347
+ "epoch": 0.9,
1348
+ "grad_norm": 7.325380802154541,
1349
+ "learning_rate": 4.2e-05,
1350
+ "loss": 1.8137,
1351
+ "step": 900
1352
+ },
1353
+ {
1354
+ "epoch": 0.9,
1355
+ "eval_loss": 1.9444538354873657,
1356
+ "eval_runtime": 157.7008,
1357
+ "eval_samples_per_second": 6.341,
1358
+ "eval_steps_per_second": 1.585,
1359
+ "step": 900
1360
+ },
1361
+ {
1362
+ "epoch": 0.91,
1363
+ "grad_norm": 6.625280857086182,
1364
+ "learning_rate": 4.18e-05,
1365
+ "loss": 1.8925,
1366
+ "step": 910
1367
+ },
1368
+ {
1369
+ "epoch": 0.91,
1370
+ "eval_loss": 1.9409335851669312,
1371
+ "eval_runtime": 149.5155,
1372
+ "eval_samples_per_second": 6.688,
1373
+ "eval_steps_per_second": 1.672,
1374
+ "step": 910
1375
+ },
1376
+ {
1377
+ "epoch": 0.92,
1378
+ "grad_norm": 5.897461414337158,
1379
+ "learning_rate": 4.16e-05,
1380
+ "loss": 1.8322,
1381
+ "step": 920
1382
+ },
1383
+ {
1384
+ "epoch": 0.92,
1385
+ "eval_loss": 1.9331930875778198,
1386
+ "eval_runtime": 127.1983,
1387
+ "eval_samples_per_second": 7.862,
1388
+ "eval_steps_per_second": 1.965,
1389
+ "step": 920
1390
+ },
1391
+ {
1392
+ "epoch": 0.93,
1393
+ "grad_norm": 9.08010482788086,
1394
+ "learning_rate": 4.14e-05,
1395
+ "loss": 2.042,
1396
+ "step": 930
1397
+ },
1398
+ {
1399
+ "epoch": 0.93,
1400
+ "eval_loss": 1.9316186904907227,
1401
+ "eval_runtime": 118.4354,
1402
+ "eval_samples_per_second": 8.443,
1403
+ "eval_steps_per_second": 2.111,
1404
+ "step": 930
1405
+ },
1406
+ {
1407
+ "epoch": 0.94,
1408
+ "grad_norm": 9.264948844909668,
1409
+ "learning_rate": 4.12e-05,
1410
+ "loss": 1.7826,
1411
+ "step": 940
1412
+ },
1413
+ {
1414
+ "epoch": 0.94,
1415
+ "eval_loss": 1.9262688159942627,
1416
+ "eval_runtime": 113.7982,
1417
+ "eval_samples_per_second": 8.787,
1418
+ "eval_steps_per_second": 2.197,
1419
+ "step": 940
1420
+ },
1421
+ {
1422
+ "epoch": 0.95,
1423
+ "grad_norm": 10.431904792785645,
1424
+ "learning_rate": 4.1e-05,
1425
+ "loss": 1.7883,
1426
+ "step": 950
1427
+ },
1428
+ {
1429
+ "epoch": 0.95,
1430
+ "eval_loss": 1.9221816062927246,
1431
+ "eval_runtime": 107.3673,
1432
+ "eval_samples_per_second": 9.314,
1433
+ "eval_steps_per_second": 2.328,
1434
+ "step": 950
1435
+ },
1436
+ {
1437
+ "epoch": 0.96,
1438
+ "grad_norm": 5.923398017883301,
1439
+ "learning_rate": 4.08e-05,
1440
+ "loss": 1.5646,
1441
+ "step": 960
1442
+ },
1443
+ {
1444
+ "epoch": 0.96,
1445
+ "eval_loss": 1.9227478504180908,
1446
+ "eval_runtime": 121.0769,
1447
+ "eval_samples_per_second": 8.259,
1448
+ "eval_steps_per_second": 2.065,
1449
+ "step": 960
1450
+ },
1451
+ {
1452
+ "epoch": 0.97,
1453
+ "grad_norm": 8.440842628479004,
1454
+ "learning_rate": 4.0600000000000004e-05,
1455
+ "loss": 1.6706,
1456
+ "step": 970
1457
+ },
1458
+ {
1459
+ "epoch": 0.97,
1460
+ "eval_loss": 1.9174282550811768,
1461
+ "eval_runtime": 121.92,
1462
+ "eval_samples_per_second": 8.202,
1463
+ "eval_steps_per_second": 2.051,
1464
+ "step": 970
1465
+ },
1466
+ {
1467
+ "epoch": 0.98,
1468
+ "grad_norm": 6.066262245178223,
1469
+ "learning_rate": 4.0400000000000006e-05,
1470
+ "loss": 1.6543,
1471
+ "step": 980
1472
+ },
1473
+ {
1474
+ "epoch": 0.98,
1475
+ "eval_loss": 1.915424108505249,
1476
+ "eval_runtime": 111.8461,
1477
+ "eval_samples_per_second": 8.941,
1478
+ "eval_steps_per_second": 2.235,
1479
+ "step": 980
1480
+ },
1481
+ {
1482
+ "epoch": 0.99,
1483
+ "grad_norm": 7.314166069030762,
1484
+ "learning_rate": 4.02e-05,
1485
+ "loss": 2.1565,
1486
+ "step": 990
1487
+ },
1488
+ {
1489
+ "epoch": 0.99,
1490
+ "eval_loss": 1.9187248945236206,
1491
+ "eval_runtime": 110.4362,
1492
+ "eval_samples_per_second": 9.055,
1493
+ "eval_steps_per_second": 2.264,
1494
+ "step": 990
1495
+ },
1496
+ {
1497
+ "epoch": 1.0,
1498
+ "grad_norm": 3.6229448318481445,
1499
+ "learning_rate": 4e-05,
1500
+ "loss": 1.7919,
1501
+ "step": 1000
1502
+ },
1503
+ {
1504
+ "epoch": 1.0,
1505
+ "eval_loss": 1.9095934629440308,
1506
+ "eval_runtime": 103.6218,
1507
+ "eval_samples_per_second": 9.65,
1508
+ "eval_steps_per_second": 2.413,
1509
+ "step": 1000
1510
+ },
1511
+ {
1512
+ "epoch": 1.01,
1513
+ "grad_norm": 7.193479537963867,
1514
+ "learning_rate": 3.9800000000000005e-05,
1515
+ "loss": 1.5053,
1516
+ "step": 1010
1517
+ },
1518
+ {
1519
+ "epoch": 1.01,
1520
+ "eval_loss": 1.9059890508651733,
1521
+ "eval_runtime": 128.523,
1522
+ "eval_samples_per_second": 7.781,
1523
+ "eval_steps_per_second": 1.945,
1524
+ "step": 1010
1525
+ },
1526
+ {
1527
+ "epoch": 1.02,
1528
+ "grad_norm": 8.449844360351562,
1529
+ "learning_rate": 3.960000000000001e-05,
1530
+ "loss": 1.7817,
1531
+ "step": 1020
1532
+ },
1533
+ {
1534
+ "epoch": 1.02,
1535
+ "eval_loss": 1.9032129049301147,
1536
+ "eval_runtime": 118.3594,
1537
+ "eval_samples_per_second": 8.449,
1538
+ "eval_steps_per_second": 2.112,
1539
+ "step": 1020
1540
+ },
1541
+ {
1542
+ "epoch": 1.03,
1543
+ "grad_norm": 8.57231330871582,
1544
+ "learning_rate": 3.94e-05,
1545
+ "loss": 1.9459,
1546
+ "step": 1030
1547
+ },
1548
+ {
1549
+ "epoch": 1.03,
1550
+ "eval_loss": 1.9003936052322388,
1551
+ "eval_runtime": 148.1691,
1552
+ "eval_samples_per_second": 6.749,
1553
+ "eval_steps_per_second": 1.687,
1554
+ "step": 1030
1555
+ },
1556
+ {
1557
+ "epoch": 1.04,
1558
+ "grad_norm": 7.97957181930542,
1559
+ "learning_rate": 3.9200000000000004e-05,
1560
+ "loss": 1.9442,
1561
+ "step": 1040
1562
+ },
1563
+ {
1564
+ "epoch": 1.04,
1565
+ "eval_loss": 1.8994051218032837,
1566
+ "eval_runtime": 138.9543,
1567
+ "eval_samples_per_second": 7.197,
1568
+ "eval_steps_per_second": 1.799,
1569
+ "step": 1040
1570
+ },
1571
+ {
1572
+ "epoch": 1.05,
1573
+ "grad_norm": 6.236805438995361,
1574
+ "learning_rate": 3.9000000000000006e-05,
1575
+ "loss": 1.6471,
1576
+ "step": 1050
1577
+ },
1578
+ {
1579
+ "epoch": 1.05,
1580
+ "eval_loss": 1.8940227031707764,
1581
+ "eval_runtime": 125.7231,
1582
+ "eval_samples_per_second": 7.954,
1583
+ "eval_steps_per_second": 1.988,
1584
+ "step": 1050
1585
+ },
1586
+ {
1587
+ "epoch": 1.06,
1588
+ "grad_norm": 8.682119369506836,
1589
+ "learning_rate": 3.88e-05,
1590
+ "loss": 1.5088,
1591
+ "step": 1060
1592
+ },
1593
+ {
1594
+ "epoch": 1.06,
1595
+ "eval_loss": 1.9049922227859497,
1596
+ "eval_runtime": 123.3481,
1597
+ "eval_samples_per_second": 8.107,
1598
+ "eval_steps_per_second": 2.027,
1599
+ "step": 1060
1600
+ },
1601
+ {
1602
+ "epoch": 1.07,
1603
+ "grad_norm": 7.925257682800293,
1604
+ "learning_rate": 3.86e-05,
1605
+ "loss": 1.5251,
1606
+ "step": 1070
1607
+ },
1608
+ {
1609
+ "epoch": 1.07,
1610
+ "eval_loss": 1.900003433227539,
1611
+ "eval_runtime": 128.7457,
1612
+ "eval_samples_per_second": 7.767,
1613
+ "eval_steps_per_second": 1.942,
1614
+ "step": 1070
1615
+ },
1616
+ {
1617
+ "epoch": 1.08,
1618
+ "grad_norm": 10.475297927856445,
1619
+ "learning_rate": 3.8400000000000005e-05,
1620
+ "loss": 2.0858,
1621
+ "step": 1080
1622
+ },
1623
+ {
1624
+ "epoch": 1.08,
1625
+ "eval_loss": 1.8909506797790527,
1626
+ "eval_runtime": 130.1802,
1627
+ "eval_samples_per_second": 7.682,
1628
+ "eval_steps_per_second": 1.92,
1629
+ "step": 1080
1630
+ },
1631
+ {
1632
+ "epoch": 1.09,
1633
+ "grad_norm": 8.481929779052734,
1634
+ "learning_rate": 3.82e-05,
1635
+ "loss": 1.9986,
1636
+ "step": 1090
1637
+ },
1638
+ {
1639
+ "epoch": 1.09,
1640
+ "eval_loss": 1.8873074054718018,
1641
+ "eval_runtime": 124.38,
1642
+ "eval_samples_per_second": 8.04,
1643
+ "eval_steps_per_second": 2.01,
1644
+ "step": 1090
1645
+ },
1646
+ {
1647
+ "epoch": 1.1,
1648
+ "grad_norm": 7.110391616821289,
1649
+ "learning_rate": 3.8e-05,
1650
+ "loss": 2.0248,
1651
+ "step": 1100
1652
+ },
1653
+ {
1654
+ "epoch": 1.1,
1655
+ "eval_loss": 1.8794257640838623,
1656
+ "eval_runtime": 113.42,
1657
+ "eval_samples_per_second": 8.817,
1658
+ "eval_steps_per_second": 2.204,
1659
+ "step": 1100
1660
+ },
1661
+ {
1662
+ "epoch": 1.11,
1663
+ "grad_norm": 5.841686725616455,
1664
+ "learning_rate": 3.7800000000000004e-05,
1665
+ "loss": 1.7042,
1666
+ "step": 1110
1667
+ },
1668
+ {
1669
+ "epoch": 1.11,
1670
+ "eval_loss": 1.8803025484085083,
1671
+ "eval_runtime": 109.2435,
1672
+ "eval_samples_per_second": 9.154,
1673
+ "eval_steps_per_second": 2.288,
1674
+ "step": 1110
1675
+ },
1676
+ {
1677
+ "epoch": 1.12,
1678
+ "grad_norm": 4.138281345367432,
1679
+ "learning_rate": 3.76e-05,
1680
+ "loss": 1.749,
1681
+ "step": 1120
1682
+ },
1683
+ {
1684
+ "epoch": 1.12,
1685
+ "eval_loss": 1.8786567449569702,
1686
+ "eval_runtime": 109.7489,
1687
+ "eval_samples_per_second": 9.112,
1688
+ "eval_steps_per_second": 2.278,
1689
+ "step": 1120
1690
+ },
1691
+ {
1692
+ "epoch": 1.13,
1693
+ "grad_norm": 8.400279998779297,
1694
+ "learning_rate": 3.74e-05,
1695
+ "loss": 1.8222,
1696
+ "step": 1130
1697
+ },
1698
+ {
1699
+ "epoch": 1.13,
1700
+ "eval_loss": 1.882354736328125,
1701
+ "eval_runtime": 118.6402,
1702
+ "eval_samples_per_second": 8.429,
1703
+ "eval_steps_per_second": 2.107,
1704
+ "step": 1130
1705
+ },
1706
+ {
1707
+ "epoch": 1.1400000000000001,
1708
+ "grad_norm": 6.2840166091918945,
1709
+ "learning_rate": 3.72e-05,
1710
+ "loss": 2.0553,
1711
+ "step": 1140
1712
+ },
1713
+ {
1714
+ "epoch": 1.1400000000000001,
1715
+ "eval_loss": 1.8792822360992432,
1716
+ "eval_runtime": 120.4519,
1717
+ "eval_samples_per_second": 8.302,
1718
+ "eval_steps_per_second": 2.076,
1719
+ "step": 1140
1720
+ },
1721
+ {
1722
+ "epoch": 1.15,
1723
+ "grad_norm": 7.492580413818359,
1724
+ "learning_rate": 3.7e-05,
1725
+ "loss": 1.7812,
1726
+ "step": 1150
1727
+ },
1728
+ {
1729
+ "epoch": 1.15,
1730
+ "eval_loss": 1.875270962715149,
1731
+ "eval_runtime": 125.0379,
1732
+ "eval_samples_per_second": 7.998,
1733
+ "eval_steps_per_second": 1.999,
1734
+ "step": 1150
1735
+ },
1736
+ {
1737
+ "epoch": 1.16,
1738
+ "grad_norm": 7.373579978942871,
1739
+ "learning_rate": 3.68e-05,
1740
+ "loss": 1.8063,
1741
+ "step": 1160
1742
+ },
1743
+ {
1744
+ "epoch": 1.16,
1745
+ "eval_loss": 1.8711602687835693,
1746
+ "eval_runtime": 123.1041,
1747
+ "eval_samples_per_second": 8.123,
1748
+ "eval_steps_per_second": 2.031,
1749
+ "step": 1160
1750
+ },
1751
+ {
1752
+ "epoch": 1.17,
1753
+ "grad_norm": 8.983260154724121,
1754
+ "learning_rate": 3.66e-05,
1755
+ "loss": 2.0774,
1756
+ "step": 1170
1757
+ },
1758
+ {
1759
+ "epoch": 1.17,
1760
+ "eval_loss": 1.8704919815063477,
1761
+ "eval_runtime": 115.7789,
1762
+ "eval_samples_per_second": 8.637,
1763
+ "eval_steps_per_second": 2.159,
1764
+ "step": 1170
1765
+ },
1766
+ {
1767
+ "epoch": 1.18,
1768
+ "grad_norm": 8.41231918334961,
1769
+ "learning_rate": 3.6400000000000004e-05,
1770
+ "loss": 1.5289,
1771
+ "step": 1180
1772
+ },
1773
+ {
1774
+ "epoch": 1.18,
1775
+ "eval_loss": 1.8638745546340942,
1776
+ "eval_runtime": 108.5452,
1777
+ "eval_samples_per_second": 9.213,
1778
+ "eval_steps_per_second": 2.303,
1779
+ "step": 1180
1780
+ },
1781
+ {
1782
+ "epoch": 1.19,
1783
+ "grad_norm": 7.630991458892822,
1784
+ "learning_rate": 3.62e-05,
1785
+ "loss": 1.7938,
1786
+ "step": 1190
1787
+ },
1788
+ {
1789
+ "epoch": 1.19,
1790
+ "eval_loss": 1.8574650287628174,
1791
+ "eval_runtime": 127.6165,
1792
+ "eval_samples_per_second": 7.836,
1793
+ "eval_steps_per_second": 1.959,
1794
+ "step": 1190
1795
+ },
1796
+ {
1797
+ "epoch": 1.2,
1798
+ "grad_norm": 9.092334747314453,
1799
+ "learning_rate": 3.6e-05,
1800
+ "loss": 1.8305,
1801
+ "step": 1200
1802
+ },
1803
+ {
1804
+ "epoch": 1.2,
1805
+ "eval_loss": 1.8641369342803955,
1806
+ "eval_runtime": 102.0991,
1807
+ "eval_samples_per_second": 9.794,
1808
+ "eval_steps_per_second": 2.449,
1809
+ "step": 1200
1810
+ },
1811
+ {
1812
+ "epoch": 1.21,
1813
+ "grad_norm": 5.7870378494262695,
1814
+ "learning_rate": 3.58e-05,
1815
+ "loss": 1.6586,
1816
+ "step": 1210
1817
+ },
1818
+ {
1819
+ "epoch": 1.21,
1820
+ "eval_loss": 1.8618590831756592,
1821
+ "eval_runtime": 115.5522,
1822
+ "eval_samples_per_second": 8.654,
1823
+ "eval_steps_per_second": 2.164,
1824
+ "step": 1210
1825
+ },
1826
+ {
1827
+ "epoch": 1.22,
1828
+ "grad_norm": 6.22267484664917,
1829
+ "learning_rate": 3.56e-05,
1830
+ "loss": 1.7842,
1831
+ "step": 1220
1832
+ },
1833
+ {
1834
+ "epoch": 1.22,
1835
+ "eval_loss": 1.8590444326400757,
1836
+ "eval_runtime": 119.2331,
1837
+ "eval_samples_per_second": 8.387,
1838
+ "eval_steps_per_second": 2.097,
1839
+ "step": 1220
1840
+ },
1841
+ {
1842
+ "epoch": 1.23,
1843
+ "grad_norm": 6.128435134887695,
1844
+ "learning_rate": 3.54e-05,
1845
+ "loss": 1.6496,
1846
+ "step": 1230
1847
+ },
1848
+ {
1849
+ "epoch": 1.23,
1850
+ "eval_loss": 1.856571078300476,
1851
+ "eval_runtime": 115.6665,
1852
+ "eval_samples_per_second": 8.646,
1853
+ "eval_steps_per_second": 2.161,
1854
+ "step": 1230
1855
+ },
1856
+ {
1857
+ "epoch": 1.24,
1858
+ "grad_norm": 6.75801944732666,
1859
+ "learning_rate": 3.52e-05,
1860
+ "loss": 1.9095,
1861
+ "step": 1240
1862
+ },
1863
+ {
1864
+ "epoch": 1.24,
1865
+ "eval_loss": 1.8567980527877808,
1866
+ "eval_runtime": 117.0444,
1867
+ "eval_samples_per_second": 8.544,
1868
+ "eval_steps_per_second": 2.136,
1869
+ "step": 1240
1870
+ },
1871
+ {
1872
+ "epoch": 1.25,
1873
+ "grad_norm": 7.787381649017334,
1874
+ "learning_rate": 3.5e-05,
1875
+ "loss": 1.9586,
1876
+ "step": 1250
1877
+ },
1878
+ {
1879
+ "epoch": 1.25,
1880
+ "eval_loss": 1.8500583171844482,
1881
+ "eval_runtime": 112.5946,
1882
+ "eval_samples_per_second": 8.881,
1883
+ "eval_steps_per_second": 2.22,
1884
+ "step": 1250
1885
+ },
1886
+ {
1887
+ "epoch": 1.26,
1888
+ "grad_norm": 5.179034233093262,
1889
+ "learning_rate": 3.48e-05,
1890
+ "loss": 1.483,
1891
+ "step": 1260
1892
+ },
1893
+ {
1894
+ "epoch": 1.26,
1895
+ "eval_loss": 1.8572009801864624,
1896
+ "eval_runtime": 126.9259,
1897
+ "eval_samples_per_second": 7.879,
1898
+ "eval_steps_per_second": 1.97,
1899
+ "step": 1260
1900
+ },
1901
+ {
1902
+ "epoch": 1.27,
1903
+ "grad_norm": 9.003114700317383,
1904
+ "learning_rate": 3.46e-05,
1905
+ "loss": 1.5393,
1906
+ "step": 1270
1907
+ },
1908
+ {
1909
+ "epoch": 1.27,
1910
+ "eval_loss": 1.8491322994232178,
1911
+ "eval_runtime": 130.4481,
1912
+ "eval_samples_per_second": 7.666,
1913
+ "eval_steps_per_second": 1.916,
1914
+ "step": 1270
1915
+ },
1916
+ {
1917
+ "epoch": 1.28,
1918
+ "grad_norm": 9.390629768371582,
1919
+ "learning_rate": 3.4399999999999996e-05,
1920
+ "loss": 1.9644,
1921
+ "step": 1280
1922
+ },
1923
+ {
1924
+ "epoch": 1.28,
1925
+ "eval_loss": 1.8494079113006592,
1926
+ "eval_runtime": 133.2843,
1927
+ "eval_samples_per_second": 7.503,
1928
+ "eval_steps_per_second": 1.876,
1929
+ "step": 1280
1930
+ },
1931
+ {
1932
+ "epoch": 1.29,
1933
+ "grad_norm": 8.873686790466309,
1934
+ "learning_rate": 3.4200000000000005e-05,
1935
+ "loss": 1.8397,
1936
+ "step": 1290
1937
+ },
1938
+ {
1939
+ "epoch": 1.29,
1940
+ "eval_loss": 1.8387765884399414,
1941
+ "eval_runtime": 133.1956,
1942
+ "eval_samples_per_second": 7.508,
1943
+ "eval_steps_per_second": 1.877,
1944
+ "step": 1290
1945
+ },
1946
+ {
1947
+ "epoch": 1.3,
1948
+ "grad_norm": 8.92082691192627,
1949
+ "learning_rate": 3.4000000000000007e-05,
1950
+ "loss": 1.5783,
1951
+ "step": 1300
1952
+ },
1953
+ {
1954
+ "epoch": 1.3,
1955
+ "eval_loss": 1.8463449478149414,
1956
+ "eval_runtime": 132.1846,
1957
+ "eval_samples_per_second": 7.565,
1958
+ "eval_steps_per_second": 1.891,
1959
+ "step": 1300
1960
+ },
1961
+ {
1962
+ "epoch": 1.31,
1963
+ "grad_norm": 9.575182914733887,
1964
+ "learning_rate": 3.38e-05,
1965
+ "loss": 1.3814,
1966
+ "step": 1310
1967
+ },
1968
+ {
1969
+ "epoch": 1.31,
1970
+ "eval_loss": 1.8390209674835205,
1971
+ "eval_runtime": 112.5756,
1972
+ "eval_samples_per_second": 8.883,
1973
+ "eval_steps_per_second": 2.221,
1974
+ "step": 1310
1975
+ },
1976
+ {
1977
+ "epoch": 1.32,
1978
+ "grad_norm": 7.706173896789551,
1979
+ "learning_rate": 3.3600000000000004e-05,
1980
+ "loss": 1.9143,
1981
+ "step": 1320
1982
+ },
1983
+ {
1984
+ "epoch": 1.32,
1985
+ "eval_loss": 1.839473843574524,
1986
+ "eval_runtime": 133.4819,
1987
+ "eval_samples_per_second": 7.492,
1988
+ "eval_steps_per_second": 1.873,
1989
+ "step": 1320
1990
+ },
1991
+ {
1992
+ "epoch": 1.33,
1993
+ "grad_norm": 6.447656154632568,
1994
+ "learning_rate": 3.3400000000000005e-05,
1995
+ "loss": 1.8488,
1996
+ "step": 1330
1997
+ },
1998
+ {
1999
+ "epoch": 1.33,
2000
+ "eval_loss": 1.834652066230774,
2001
+ "eval_runtime": 124.1955,
2002
+ "eval_samples_per_second": 8.052,
2003
+ "eval_steps_per_second": 2.013,
2004
+ "step": 1330
2005
+ },
2006
+ {
2007
+ "epoch": 1.34,
2008
+ "grad_norm": 7.253933906555176,
2009
+ "learning_rate": 3.32e-05,
2010
+ "loss": 1.7947,
2011
+ "step": 1340
2012
+ },
2013
+ {
2014
+ "epoch": 1.34,
2015
+ "eval_loss": 1.830536127090454,
2016
+ "eval_runtime": 126.3069,
2017
+ "eval_samples_per_second": 7.917,
2018
+ "eval_steps_per_second": 1.979,
2019
+ "step": 1340
2020
+ },
2021
+ {
2022
+ "epoch": 1.35,
2023
+ "grad_norm": 7.001797199249268,
2024
+ "learning_rate": 3.3e-05,
2025
+ "loss": 1.9053,
2026
+ "step": 1350
2027
+ },
2028
+ {
2029
+ "epoch": 1.35,
2030
+ "eval_loss": 1.8331165313720703,
2031
+ "eval_runtime": 130.4927,
2032
+ "eval_samples_per_second": 7.663,
2033
+ "eval_steps_per_second": 1.916,
2034
+ "step": 1350
2035
+ },
2036
+ {
2037
+ "epoch": 1.3599999999999999,
2038
+ "grad_norm": 8.703651428222656,
2039
+ "learning_rate": 3.2800000000000004e-05,
2040
+ "loss": 1.8354,
2041
+ "step": 1360
2042
+ },
2043
+ {
2044
+ "epoch": 1.3599999999999999,
2045
+ "eval_loss": 1.8279739618301392,
2046
+ "eval_runtime": 110.2803,
2047
+ "eval_samples_per_second": 9.068,
2048
+ "eval_steps_per_second": 2.267,
2049
+ "step": 1360
2050
+ },
2051
+ {
2052
+ "epoch": 1.37,
2053
+ "grad_norm": 8.572402000427246,
2054
+ "learning_rate": 3.26e-05,
2055
+ "loss": 1.7067,
2056
+ "step": 1370
2057
+ },
2058
+ {
2059
+ "epoch": 1.37,
2060
+ "eval_loss": 1.8316876888275146,
2061
+ "eval_runtime": 107.9421,
2062
+ "eval_samples_per_second": 9.264,
2063
+ "eval_steps_per_second": 2.316,
2064
+ "step": 1370
2065
+ },
2066
+ {
2067
+ "epoch": 1.38,
2068
+ "grad_norm": 5.029556751251221,
2069
+ "learning_rate": 3.24e-05,
2070
+ "loss": 1.5889,
2071
+ "step": 1380
2072
+ },
2073
+ {
2074
+ "epoch": 1.38,
2075
+ "eval_loss": 1.8241760730743408,
2076
+ "eval_runtime": 119.602,
2077
+ "eval_samples_per_second": 8.361,
2078
+ "eval_steps_per_second": 2.09,
2079
+ "step": 1380
2080
+ },
2081
+ {
2082
+ "epoch": 1.3900000000000001,
2083
+ "grad_norm": 6.689691066741943,
2084
+ "learning_rate": 3.2200000000000003e-05,
2085
+ "loss": 1.8915,
2086
+ "step": 1390
2087
+ },
2088
+ {
2089
+ "epoch": 1.3900000000000001,
2090
+ "eval_loss": 1.8192516565322876,
2091
+ "eval_runtime": 124.5483,
2092
+ "eval_samples_per_second": 8.029,
2093
+ "eval_steps_per_second": 2.007,
2094
+ "step": 1390
2095
+ },
2096
+ {
2097
+ "epoch": 1.4,
2098
+ "grad_norm": 7.118290424346924,
2099
+ "learning_rate": 3.2000000000000005e-05,
2100
+ "loss": 1.8986,
2101
+ "step": 1400
2102
+ },
2103
+ {
2104
+ "epoch": 1.4,
2105
+ "eval_loss": 1.823815941810608,
2106
+ "eval_runtime": 113.9473,
2107
+ "eval_samples_per_second": 8.776,
2108
+ "eval_steps_per_second": 2.194,
2109
+ "step": 1400
2110
+ },
2111
+ {
2112
+ "epoch": 1.41,
2113
+ "grad_norm": 8.022662162780762,
2114
+ "learning_rate": 3.18e-05,
2115
+ "loss": 1.9665,
2116
+ "step": 1410
2117
+ },
2118
+ {
2119
+ "epoch": 1.41,
2120
+ "eval_loss": 1.8193540573120117,
2121
+ "eval_runtime": 125.65,
2122
+ "eval_samples_per_second": 7.959,
2123
+ "eval_steps_per_second": 1.99,
2124
+ "step": 1410
2125
+ },
2126
+ {
2127
+ "epoch": 1.42,
2128
+ "grad_norm": 6.133983135223389,
2129
+ "learning_rate": 3.16e-05,
2130
+ "loss": 1.6947,
2131
+ "step": 1420
2132
+ },
2133
+ {
2134
+ "epoch": 1.42,
2135
+ "eval_loss": 1.813571810722351,
2136
+ "eval_runtime": 116.6704,
2137
+ "eval_samples_per_second": 8.571,
2138
+ "eval_steps_per_second": 2.143,
2139
+ "step": 1420
2140
+ },
2141
+ {
2142
+ "epoch": 1.43,
2143
+ "grad_norm": 10.109593391418457,
2144
+ "learning_rate": 3.1400000000000004e-05,
2145
+ "loss": 1.8847,
2146
+ "step": 1430
2147
+ },
2148
+ {
2149
+ "epoch": 1.43,
2150
+ "eval_loss": 1.8171520233154297,
2151
+ "eval_runtime": 119.4118,
2152
+ "eval_samples_per_second": 8.374,
2153
+ "eval_steps_per_second": 2.094,
2154
+ "step": 1430
2155
+ },
2156
+ {
2157
+ "epoch": 1.44,
2158
+ "grad_norm": 6.10088586807251,
2159
+ "learning_rate": 3.12e-05,
2160
+ "loss": 2.0466,
2161
+ "step": 1440
2162
+ },
2163
+ {
2164
+ "epoch": 1.44,
2165
+ "eval_loss": 1.8149670362472534,
2166
+ "eval_runtime": 119.8555,
2167
+ "eval_samples_per_second": 8.343,
2168
+ "eval_steps_per_second": 2.086,
2169
+ "step": 1440
2170
+ },
2171
+ {
2172
+ "epoch": 1.45,
2173
+ "grad_norm": 6.327012538909912,
2174
+ "learning_rate": 3.1e-05,
2175
+ "loss": 1.6163,
2176
+ "step": 1450
2177
+ },
2178
+ {
2179
+ "epoch": 1.45,
2180
+ "eval_loss": 1.8084301948547363,
2181
+ "eval_runtime": 111.056,
2182
+ "eval_samples_per_second": 9.004,
2183
+ "eval_steps_per_second": 2.251,
2184
+ "step": 1450
2185
+ },
2186
+ {
2187
+ "epoch": 1.46,
2188
+ "grad_norm": 7.103956699371338,
2189
+ "learning_rate": 3.08e-05,
2190
+ "loss": 1.8625,
2191
+ "step": 1460
2192
+ },
2193
+ {
2194
+ "epoch": 1.46,
2195
+ "eval_loss": 1.8112941980361938,
2196
+ "eval_runtime": 133.3088,
2197
+ "eval_samples_per_second": 7.501,
2198
+ "eval_steps_per_second": 1.875,
2199
+ "step": 1460
2200
+ },
2201
+ {
2202
+ "epoch": 1.47,
2203
+ "grad_norm": 5.035983085632324,
2204
+ "learning_rate": 3.06e-05,
2205
+ "loss": 1.8754,
2206
+ "step": 1470
2207
+ },
2208
+ {
2209
+ "epoch": 1.47,
2210
+ "eval_loss": 1.8135524988174438,
2211
+ "eval_runtime": 128.5309,
2212
+ "eval_samples_per_second": 7.78,
2213
+ "eval_steps_per_second": 1.945,
2214
+ "step": 1470
2215
+ },
2216
+ {
2217
+ "epoch": 1.48,
2218
+ "grad_norm": 8.39537525177002,
2219
+ "learning_rate": 3.04e-05,
2220
+ "loss": 1.9341,
2221
+ "step": 1480
2222
+ },
2223
+ {
2224
+ "epoch": 1.48,
2225
+ "eval_loss": 1.80543851852417,
2226
+ "eval_runtime": 119.3552,
2227
+ "eval_samples_per_second": 8.378,
2228
+ "eval_steps_per_second": 2.095,
2229
+ "step": 1480
2230
+ },
2231
+ {
2232
+ "epoch": 1.49,
2233
+ "grad_norm": 8.173503875732422,
2234
+ "learning_rate": 3.02e-05,
2235
+ "loss": 2.0139,
2236
+ "step": 1490
2237
+ },
2238
+ {
2239
+ "epoch": 1.49,
2240
+ "eval_loss": 1.8039709329605103,
2241
+ "eval_runtime": 117.2862,
2242
+ "eval_samples_per_second": 8.526,
2243
+ "eval_steps_per_second": 2.132,
2244
+ "step": 1490
2245
+ },
2246
+ {
2247
+ "epoch": 1.5,
2248
+ "grad_norm": 8.277552604675293,
2249
+ "learning_rate": 3e-05,
2250
+ "loss": 1.7314,
2251
+ "step": 1500
2252
+ },
2253
+ {
2254
+ "epoch": 1.5,
2255
+ "eval_loss": 1.803325891494751,
2256
+ "eval_runtime": 109.3603,
2257
+ "eval_samples_per_second": 9.144,
2258
+ "eval_steps_per_second": 2.286,
2259
+ "step": 1500
2260
+ }
2261
+ ],
2262
+ "logging_steps": 10,
2263
+ "max_steps": 3000,
2264
+ "num_input_tokens_seen": 0,
2265
+ "num_train_epochs": 3,
2266
+ "save_steps": 500,
2267
+ "stateful_callbacks": {
2268
+ "TrainerControl": {
2269
+ "args": {
2270
+ "should_epoch_stop": false,
2271
+ "should_evaluate": false,
2272
+ "should_log": false,
2273
+ "should_save": true,
2274
+ "should_training_stop": false
2275
+ },
2276
+ "attributes": {}
2277
+ }
2278
+ },
2279
+ "total_flos": 391938048000000.0,
2280
+ "train_batch_size": 4,
2281
+ "trial_name": null,
2282
+ "trial_params": null
2283
+ }
fine_tuned_gpt2/checkpoint-2000/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/Users/user/projects/MindPadiBig/mindpadi/models/gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.49.0",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
fine_tuned_gpt2/checkpoint-2000/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.49.0"
6
+ }
fine_tuned_gpt2/checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,3033 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 1.7480344772338867,
3
+ "best_model_checkpoint": "/Users/user/projects/MindPadiBig/mindpadi/models/fine_tuned_gpt2/checkpoint-2000",
4
+ "epoch": 2.0,
5
+ "eval_steps": 10,
6
+ "global_step": 2000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.01,
13
+ "grad_norm": 126.21493530273438,
14
+ "learning_rate": 1.0000000000000002e-06,
15
+ "loss": 6.9352,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.01,
20
+ "eval_loss": 6.456012725830078,
21
+ "eval_runtime": 128.7117,
22
+ "eval_samples_per_second": 7.769,
23
+ "eval_steps_per_second": 1.942,
24
+ "step": 10
25
+ },
26
+ {
27
+ "epoch": 0.02,
28
+ "grad_norm": 112.57171630859375,
29
+ "learning_rate": 2.0000000000000003e-06,
30
+ "loss": 6.2877,
31
+ "step": 20
32
+ },
33
+ {
34
+ "epoch": 0.02,
35
+ "eval_loss": 6.174145698547363,
36
+ "eval_runtime": 133.4573,
37
+ "eval_samples_per_second": 7.493,
38
+ "eval_steps_per_second": 1.873,
39
+ "step": 20
40
+ },
41
+ {
42
+ "epoch": 0.03,
43
+ "grad_norm": 185.73190307617188,
44
+ "learning_rate": 3e-06,
45
+ "loss": 5.9569,
46
+ "step": 30
47
+ },
48
+ {
49
+ "epoch": 0.03,
50
+ "eval_loss": 5.594101428985596,
51
+ "eval_runtime": 136.3272,
52
+ "eval_samples_per_second": 7.335,
53
+ "eval_steps_per_second": 1.834,
54
+ "step": 30
55
+ },
56
+ {
57
+ "epoch": 0.04,
58
+ "grad_norm": 160.77322387695312,
59
+ "learning_rate": 4.000000000000001e-06,
60
+ "loss": 4.9783,
61
+ "step": 40
62
+ },
63
+ {
64
+ "epoch": 0.04,
65
+ "eval_loss": 4.95811128616333,
66
+ "eval_runtime": 123.0408,
67
+ "eval_samples_per_second": 8.127,
68
+ "eval_steps_per_second": 2.032,
69
+ "step": 40
70
+ },
71
+ {
72
+ "epoch": 0.05,
73
+ "grad_norm": 71.51310729980469,
74
+ "learning_rate": 5e-06,
75
+ "loss": 4.3123,
76
+ "step": 50
77
+ },
78
+ {
79
+ "epoch": 0.05,
80
+ "eval_loss": 4.30466890335083,
81
+ "eval_runtime": 117.1478,
82
+ "eval_samples_per_second": 8.536,
83
+ "eval_steps_per_second": 2.134,
84
+ "step": 50
85
+ },
86
+ {
87
+ "epoch": 0.06,
88
+ "grad_norm": 55.53121566772461,
89
+ "learning_rate": 6e-06,
90
+ "loss": 3.7047,
91
+ "step": 60
92
+ },
93
+ {
94
+ "epoch": 0.06,
95
+ "eval_loss": 3.8378841876983643,
96
+ "eval_runtime": 109.8285,
97
+ "eval_samples_per_second": 9.105,
98
+ "eval_steps_per_second": 2.276,
99
+ "step": 60
100
+ },
101
+ {
102
+ "epoch": 0.07,
103
+ "grad_norm": 16.639528274536133,
104
+ "learning_rate": 7.000000000000001e-06,
105
+ "loss": 4.1835,
106
+ "step": 70
107
+ },
108
+ {
109
+ "epoch": 0.07,
110
+ "eval_loss": 3.6265947818756104,
111
+ "eval_runtime": 107.5258,
112
+ "eval_samples_per_second": 9.3,
113
+ "eval_steps_per_second": 2.325,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.08,
118
+ "grad_norm": 11.000739097595215,
119
+ "learning_rate": 8.000000000000001e-06,
120
+ "loss": 3.0598,
121
+ "step": 80
122
+ },
123
+ {
124
+ "epoch": 0.08,
125
+ "eval_loss": 3.5009496212005615,
126
+ "eval_runtime": 105.6005,
127
+ "eval_samples_per_second": 9.47,
128
+ "eval_steps_per_second": 2.367,
129
+ "step": 80
130
+ },
131
+ {
132
+ "epoch": 0.09,
133
+ "grad_norm": 18.876184463500977,
134
+ "learning_rate": 9e-06,
135
+ "loss": 3.0835,
136
+ "step": 90
137
+ },
138
+ {
139
+ "epoch": 0.09,
140
+ "eval_loss": 3.402592182159424,
141
+ "eval_runtime": 106.8507,
142
+ "eval_samples_per_second": 9.359,
143
+ "eval_steps_per_second": 2.34,
144
+ "step": 90
145
+ },
146
+ {
147
+ "epoch": 0.1,
148
+ "grad_norm": 10.568647384643555,
149
+ "learning_rate": 1e-05,
150
+ "loss": 3.5607,
151
+ "step": 100
152
+ },
153
+ {
154
+ "epoch": 0.1,
155
+ "eval_loss": 3.3077728748321533,
156
+ "eval_runtime": 107.8277,
157
+ "eval_samples_per_second": 9.274,
158
+ "eval_steps_per_second": 2.319,
159
+ "step": 100
160
+ },
161
+ {
162
+ "epoch": 0.11,
163
+ "grad_norm": 6.663860321044922,
164
+ "learning_rate": 1.1000000000000001e-05,
165
+ "loss": 3.3173,
166
+ "step": 110
167
+ },
168
+ {
169
+ "epoch": 0.11,
170
+ "eval_loss": 3.221048355102539,
171
+ "eval_runtime": 104.8207,
172
+ "eval_samples_per_second": 9.54,
173
+ "eval_steps_per_second": 2.385,
174
+ "step": 110
175
+ },
176
+ {
177
+ "epoch": 0.12,
178
+ "grad_norm": 9.180710792541504,
179
+ "learning_rate": 1.2e-05,
180
+ "loss": 3.4318,
181
+ "step": 120
182
+ },
183
+ {
184
+ "epoch": 0.12,
185
+ "eval_loss": 3.1418848037719727,
186
+ "eval_runtime": 104.9529,
187
+ "eval_samples_per_second": 9.528,
188
+ "eval_steps_per_second": 2.382,
189
+ "step": 120
190
+ },
191
+ {
192
+ "epoch": 0.13,
193
+ "grad_norm": 7.3332414627075195,
194
+ "learning_rate": 1.3000000000000001e-05,
195
+ "loss": 3.1961,
196
+ "step": 130
197
+ },
198
+ {
199
+ "epoch": 0.13,
200
+ "eval_loss": 3.069470167160034,
201
+ "eval_runtime": 107.1455,
202
+ "eval_samples_per_second": 9.333,
203
+ "eval_steps_per_second": 2.333,
204
+ "step": 130
205
+ },
206
+ {
207
+ "epoch": 0.14,
208
+ "grad_norm": 11.190179824829102,
209
+ "learning_rate": 1.4000000000000001e-05,
210
+ "loss": 2.821,
211
+ "step": 140
212
+ },
213
+ {
214
+ "epoch": 0.14,
215
+ "eval_loss": 2.998840093612671,
216
+ "eval_runtime": 109.7303,
217
+ "eval_samples_per_second": 9.113,
218
+ "eval_steps_per_second": 2.278,
219
+ "step": 140
220
+ },
221
+ {
222
+ "epoch": 0.15,
223
+ "grad_norm": 10.532099723815918,
224
+ "learning_rate": 1.5e-05,
225
+ "loss": 2.8245,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.15,
230
+ "eval_loss": 2.927992820739746,
231
+ "eval_runtime": 112.3035,
232
+ "eval_samples_per_second": 8.904,
233
+ "eval_steps_per_second": 2.226,
234
+ "step": 150
235
+ },
236
+ {
237
+ "epoch": 0.16,
238
+ "grad_norm": 9.649062156677246,
239
+ "learning_rate": 1.6000000000000003e-05,
240
+ "loss": 2.9165,
241
+ "step": 160
242
+ },
243
+ {
244
+ "epoch": 0.16,
245
+ "eval_loss": 2.8786165714263916,
246
+ "eval_runtime": 128.1716,
247
+ "eval_samples_per_second": 7.802,
248
+ "eval_steps_per_second": 1.951,
249
+ "step": 160
250
+ },
251
+ {
252
+ "epoch": 0.17,
253
+ "grad_norm": 9.107309341430664,
254
+ "learning_rate": 1.7000000000000003e-05,
255
+ "loss": 3.1418,
256
+ "step": 170
257
+ },
258
+ {
259
+ "epoch": 0.17,
260
+ "eval_loss": 2.8244595527648926,
261
+ "eval_runtime": 125.0759,
262
+ "eval_samples_per_second": 7.995,
263
+ "eval_steps_per_second": 1.999,
264
+ "step": 170
265
+ },
266
+ {
267
+ "epoch": 0.18,
268
+ "grad_norm": 13.698325157165527,
269
+ "learning_rate": 1.8e-05,
270
+ "loss": 3.0323,
271
+ "step": 180
272
+ },
273
+ {
274
+ "epoch": 0.18,
275
+ "eval_loss": 2.772019386291504,
276
+ "eval_runtime": 129.6329,
277
+ "eval_samples_per_second": 7.714,
278
+ "eval_steps_per_second": 1.929,
279
+ "step": 180
280
+ },
281
+ {
282
+ "epoch": 0.19,
283
+ "grad_norm": 11.200780868530273,
284
+ "learning_rate": 1.9e-05,
285
+ "loss": 2.9199,
286
+ "step": 190
287
+ },
288
+ {
289
+ "epoch": 0.19,
290
+ "eval_loss": 2.7285261154174805,
291
+ "eval_runtime": 105.2039,
292
+ "eval_samples_per_second": 9.505,
293
+ "eval_steps_per_second": 2.376,
294
+ "step": 190
295
+ },
296
+ {
297
+ "epoch": 0.2,
298
+ "grad_norm": 11.479965209960938,
299
+ "learning_rate": 2e-05,
300
+ "loss": 2.6733,
301
+ "step": 200
302
+ },
303
+ {
304
+ "epoch": 0.2,
305
+ "eval_loss": 2.695300340652466,
306
+ "eval_runtime": 102.9704,
307
+ "eval_samples_per_second": 9.712,
308
+ "eval_steps_per_second": 2.428,
309
+ "step": 200
310
+ },
311
+ {
312
+ "epoch": 0.21,
313
+ "grad_norm": 11.359460830688477,
314
+ "learning_rate": 2.1e-05,
315
+ "loss": 2.8493,
316
+ "step": 210
317
+ },
318
+ {
319
+ "epoch": 0.21,
320
+ "eval_loss": 2.66671085357666,
321
+ "eval_runtime": 101.0064,
322
+ "eval_samples_per_second": 9.9,
323
+ "eval_steps_per_second": 2.475,
324
+ "step": 210
325
+ },
326
+ {
327
+ "epoch": 0.22,
328
+ "grad_norm": 13.90817928314209,
329
+ "learning_rate": 2.2000000000000003e-05,
330
+ "loss": 2.8467,
331
+ "step": 220
332
+ },
333
+ {
334
+ "epoch": 0.22,
335
+ "eval_loss": 2.6324076652526855,
336
+ "eval_runtime": 102.7459,
337
+ "eval_samples_per_second": 9.733,
338
+ "eval_steps_per_second": 2.433,
339
+ "step": 220
340
+ },
341
+ {
342
+ "epoch": 0.23,
343
+ "grad_norm": 12.95048999786377,
344
+ "learning_rate": 2.3000000000000003e-05,
345
+ "loss": 2.5137,
346
+ "step": 230
347
+ },
348
+ {
349
+ "epoch": 0.23,
350
+ "eval_loss": 2.599924087524414,
351
+ "eval_runtime": 106.8131,
352
+ "eval_samples_per_second": 9.362,
353
+ "eval_steps_per_second": 2.341,
354
+ "step": 230
355
+ },
356
+ {
357
+ "epoch": 0.24,
358
+ "grad_norm": 10.900901794433594,
359
+ "learning_rate": 2.4e-05,
360
+ "loss": 2.5896,
361
+ "step": 240
362
+ },
363
+ {
364
+ "epoch": 0.24,
365
+ "eval_loss": 2.56974458694458,
366
+ "eval_runtime": 106.0351,
367
+ "eval_samples_per_second": 9.431,
368
+ "eval_steps_per_second": 2.358,
369
+ "step": 240
370
+ },
371
+ {
372
+ "epoch": 0.25,
373
+ "grad_norm": 11.728984832763672,
374
+ "learning_rate": 2.5e-05,
375
+ "loss": 3.0882,
376
+ "step": 250
377
+ },
378
+ {
379
+ "epoch": 0.25,
380
+ "eval_loss": 2.5507752895355225,
381
+ "eval_runtime": 107.1284,
382
+ "eval_samples_per_second": 9.335,
383
+ "eval_steps_per_second": 2.334,
384
+ "step": 250
385
+ },
386
+ {
387
+ "epoch": 0.26,
388
+ "grad_norm": 9.277604103088379,
389
+ "learning_rate": 2.6000000000000002e-05,
390
+ "loss": 2.3151,
391
+ "step": 260
392
+ },
393
+ {
394
+ "epoch": 0.26,
395
+ "eval_loss": 2.525682210922241,
396
+ "eval_runtime": 123.4713,
397
+ "eval_samples_per_second": 8.099,
398
+ "eval_steps_per_second": 2.025,
399
+ "step": 260
400
+ },
401
+ {
402
+ "epoch": 0.27,
403
+ "grad_norm": 10.388689041137695,
404
+ "learning_rate": 2.7000000000000002e-05,
405
+ "loss": 2.5188,
406
+ "step": 270
407
+ },
408
+ {
409
+ "epoch": 0.27,
410
+ "eval_loss": 2.499377489089966,
411
+ "eval_runtime": 127.0825,
412
+ "eval_samples_per_second": 7.869,
413
+ "eval_steps_per_second": 1.967,
414
+ "step": 270
415
+ },
416
+ {
417
+ "epoch": 0.28,
418
+ "grad_norm": 7.337620258331299,
419
+ "learning_rate": 2.8000000000000003e-05,
420
+ "loss": 2.4583,
421
+ "step": 280
422
+ },
423
+ {
424
+ "epoch": 0.28,
425
+ "eval_loss": 2.4751341342926025,
426
+ "eval_runtime": 129.7707,
427
+ "eval_samples_per_second": 7.706,
428
+ "eval_steps_per_second": 1.926,
429
+ "step": 280
430
+ },
431
+ {
432
+ "epoch": 0.29,
433
+ "grad_norm": 9.492632865905762,
434
+ "learning_rate": 2.9e-05,
435
+ "loss": 2.4552,
436
+ "step": 290
437
+ },
438
+ {
439
+ "epoch": 0.29,
440
+ "eval_loss": 2.449791669845581,
441
+ "eval_runtime": 127.0769,
442
+ "eval_samples_per_second": 7.869,
443
+ "eval_steps_per_second": 1.967,
444
+ "step": 290
445
+ },
446
+ {
447
+ "epoch": 0.3,
448
+ "grad_norm": 12.488316535949707,
449
+ "learning_rate": 3e-05,
450
+ "loss": 2.8118,
451
+ "step": 300
452
+ },
453
+ {
454
+ "epoch": 0.3,
455
+ "eval_loss": 2.4329984188079834,
456
+ "eval_runtime": 125.4816,
457
+ "eval_samples_per_second": 7.969,
458
+ "eval_steps_per_second": 1.992,
459
+ "step": 300
460
+ },
461
+ {
462
+ "epoch": 0.31,
463
+ "grad_norm": 10.057647705078125,
464
+ "learning_rate": 3.1e-05,
465
+ "loss": 2.5144,
466
+ "step": 310
467
+ },
468
+ {
469
+ "epoch": 0.31,
470
+ "eval_loss": 2.4146034717559814,
471
+ "eval_runtime": 122.883,
472
+ "eval_samples_per_second": 8.138,
473
+ "eval_steps_per_second": 2.034,
474
+ "step": 310
475
+ },
476
+ {
477
+ "epoch": 0.32,
478
+ "grad_norm": 9.721734046936035,
479
+ "learning_rate": 3.2000000000000005e-05,
480
+ "loss": 2.7473,
481
+ "step": 320
482
+ },
483
+ {
484
+ "epoch": 0.32,
485
+ "eval_loss": 2.4046664237976074,
486
+ "eval_runtime": 113.7917,
487
+ "eval_samples_per_second": 8.788,
488
+ "eval_steps_per_second": 2.197,
489
+ "step": 320
490
+ },
491
+ {
492
+ "epoch": 0.33,
493
+ "grad_norm": 4.945127487182617,
494
+ "learning_rate": 3.3e-05,
495
+ "loss": 2.46,
496
+ "step": 330
497
+ },
498
+ {
499
+ "epoch": 0.33,
500
+ "eval_loss": 2.380826473236084,
501
+ "eval_runtime": 111.3358,
502
+ "eval_samples_per_second": 8.982,
503
+ "eval_steps_per_second": 2.245,
504
+ "step": 330
505
+ },
506
+ {
507
+ "epoch": 0.34,
508
+ "grad_norm": 12.477653503417969,
509
+ "learning_rate": 3.4000000000000007e-05,
510
+ "loss": 2.6254,
511
+ "step": 340
512
+ },
513
+ {
514
+ "epoch": 0.34,
515
+ "eval_loss": 2.371356964111328,
516
+ "eval_runtime": 112.2082,
517
+ "eval_samples_per_second": 8.912,
518
+ "eval_steps_per_second": 2.228,
519
+ "step": 340
520
+ },
521
+ {
522
+ "epoch": 0.35,
523
+ "grad_norm": 10.182991027832031,
524
+ "learning_rate": 3.5e-05,
525
+ "loss": 2.1996,
526
+ "step": 350
527
+ },
528
+ {
529
+ "epoch": 0.35,
530
+ "eval_loss": 2.3486673831939697,
531
+ "eval_runtime": 107.6972,
532
+ "eval_samples_per_second": 9.285,
533
+ "eval_steps_per_second": 2.321,
534
+ "step": 350
535
+ },
536
+ {
537
+ "epoch": 0.36,
538
+ "grad_norm": 10.488390922546387,
539
+ "learning_rate": 3.6e-05,
540
+ "loss": 2.6718,
541
+ "step": 360
542
+ },
543
+ {
544
+ "epoch": 0.36,
545
+ "eval_loss": 2.3371098041534424,
546
+ "eval_runtime": 111.7846,
547
+ "eval_samples_per_second": 8.946,
548
+ "eval_steps_per_second": 2.236,
549
+ "step": 360
550
+ },
551
+ {
552
+ "epoch": 0.37,
553
+ "grad_norm": 7.945727825164795,
554
+ "learning_rate": 3.7e-05,
555
+ "loss": 2.5803,
556
+ "step": 370
557
+ },
558
+ {
559
+ "epoch": 0.37,
560
+ "eval_loss": 2.3198955059051514,
561
+ "eval_runtime": 107.6209,
562
+ "eval_samples_per_second": 9.292,
563
+ "eval_steps_per_second": 2.323,
564
+ "step": 370
565
+ },
566
+ {
567
+ "epoch": 0.38,
568
+ "grad_norm": 11.906257629394531,
569
+ "learning_rate": 3.8e-05,
570
+ "loss": 2.4184,
571
+ "step": 380
572
+ },
573
+ {
574
+ "epoch": 0.38,
575
+ "eval_loss": 2.3155248165130615,
576
+ "eval_runtime": 108.3746,
577
+ "eval_samples_per_second": 9.227,
578
+ "eval_steps_per_second": 2.307,
579
+ "step": 380
580
+ },
581
+ {
582
+ "epoch": 0.39,
583
+ "grad_norm": 11.00644302368164,
584
+ "learning_rate": 3.9000000000000006e-05,
585
+ "loss": 2.2515,
586
+ "step": 390
587
+ },
588
+ {
589
+ "epoch": 0.39,
590
+ "eval_loss": 2.3117854595184326,
591
+ "eval_runtime": 105.7771,
592
+ "eval_samples_per_second": 9.454,
593
+ "eval_steps_per_second": 2.363,
594
+ "step": 390
595
+ },
596
+ {
597
+ "epoch": 0.4,
598
+ "grad_norm": 8.327221870422363,
599
+ "learning_rate": 4e-05,
600
+ "loss": 2.4152,
601
+ "step": 400
602
+ },
603
+ {
604
+ "epoch": 0.4,
605
+ "eval_loss": 2.289562940597534,
606
+ "eval_runtime": 110.3598,
607
+ "eval_samples_per_second": 9.061,
608
+ "eval_steps_per_second": 2.265,
609
+ "step": 400
610
+ },
611
+ {
612
+ "epoch": 0.41,
613
+ "grad_norm": 12.397188186645508,
614
+ "learning_rate": 4.1e-05,
615
+ "loss": 2.2665,
616
+ "step": 410
617
+ },
618
+ {
619
+ "epoch": 0.41,
620
+ "eval_loss": 2.2749555110931396,
621
+ "eval_runtime": 105.9973,
622
+ "eval_samples_per_second": 9.434,
623
+ "eval_steps_per_second": 2.359,
624
+ "step": 410
625
+ },
626
+ {
627
+ "epoch": 0.42,
628
+ "grad_norm": 11.297379493713379,
629
+ "learning_rate": 4.2e-05,
630
+ "loss": 1.6996,
631
+ "step": 420
632
+ },
633
+ {
634
+ "epoch": 0.42,
635
+ "eval_loss": 2.2712552547454834,
636
+ "eval_runtime": 108.2952,
637
+ "eval_samples_per_second": 9.234,
638
+ "eval_steps_per_second": 2.309,
639
+ "step": 420
640
+ },
641
+ {
642
+ "epoch": 0.43,
643
+ "grad_norm": 6.590099811553955,
644
+ "learning_rate": 4.3e-05,
645
+ "loss": 2.0371,
646
+ "step": 430
647
+ },
648
+ {
649
+ "epoch": 0.43,
650
+ "eval_loss": 2.2652268409729004,
651
+ "eval_runtime": 159.4264,
652
+ "eval_samples_per_second": 6.272,
653
+ "eval_steps_per_second": 1.568,
654
+ "step": 430
655
+ },
656
+ {
657
+ "epoch": 0.44,
658
+ "grad_norm": 10.717235565185547,
659
+ "learning_rate": 4.4000000000000006e-05,
660
+ "loss": 2.0536,
661
+ "step": 440
662
+ },
663
+ {
664
+ "epoch": 0.44,
665
+ "eval_loss": 2.254389762878418,
666
+ "eval_runtime": 119.2716,
667
+ "eval_samples_per_second": 8.384,
668
+ "eval_steps_per_second": 2.096,
669
+ "step": 440
670
+ },
671
+ {
672
+ "epoch": 0.45,
673
+ "grad_norm": 12.450675964355469,
674
+ "learning_rate": 4.5e-05,
675
+ "loss": 2.3436,
676
+ "step": 450
677
+ },
678
+ {
679
+ "epoch": 0.45,
680
+ "eval_loss": 2.241189956665039,
681
+ "eval_runtime": 101.1513,
682
+ "eval_samples_per_second": 9.886,
683
+ "eval_steps_per_second": 2.472,
684
+ "step": 450
685
+ },
686
+ {
687
+ "epoch": 0.46,
688
+ "grad_norm": 10.726924896240234,
689
+ "learning_rate": 4.600000000000001e-05,
690
+ "loss": 2.2921,
691
+ "step": 460
692
+ },
693
+ {
694
+ "epoch": 0.46,
695
+ "eval_loss": 2.233902931213379,
696
+ "eval_runtime": 96.5389,
697
+ "eval_samples_per_second": 10.359,
698
+ "eval_steps_per_second": 2.59,
699
+ "step": 460
700
+ },
701
+ {
702
+ "epoch": 0.47,
703
+ "grad_norm": 10.492919921875,
704
+ "learning_rate": 4.7e-05,
705
+ "loss": 2.0704,
706
+ "step": 470
707
+ },
708
+ {
709
+ "epoch": 0.47,
710
+ "eval_loss": 2.223443031311035,
711
+ "eval_runtime": 94.3771,
712
+ "eval_samples_per_second": 10.596,
713
+ "eval_steps_per_second": 2.649,
714
+ "step": 470
715
+ },
716
+ {
717
+ "epoch": 0.48,
718
+ "grad_norm": 12.142420768737793,
719
+ "learning_rate": 4.8e-05,
720
+ "loss": 2.3475,
721
+ "step": 480
722
+ },
723
+ {
724
+ "epoch": 0.48,
725
+ "eval_loss": 2.20881986618042,
726
+ "eval_runtime": 88.7095,
727
+ "eval_samples_per_second": 11.273,
728
+ "eval_steps_per_second": 2.818,
729
+ "step": 480
730
+ },
731
+ {
732
+ "epoch": 0.49,
733
+ "grad_norm": 11.031810760498047,
734
+ "learning_rate": 4.9e-05,
735
+ "loss": 2.3123,
736
+ "step": 490
737
+ },
738
+ {
739
+ "epoch": 0.49,
740
+ "eval_loss": 2.193828582763672,
741
+ "eval_runtime": 90.0998,
742
+ "eval_samples_per_second": 11.099,
743
+ "eval_steps_per_second": 2.775,
744
+ "step": 490
745
+ },
746
+ {
747
+ "epoch": 0.5,
748
+ "grad_norm": 8.5560884475708,
749
+ "learning_rate": 5e-05,
750
+ "loss": 2.4333,
751
+ "step": 500
752
+ },
753
+ {
754
+ "epoch": 0.5,
755
+ "eval_loss": 2.1911978721618652,
756
+ "eval_runtime": 93.2158,
757
+ "eval_samples_per_second": 10.728,
758
+ "eval_steps_per_second": 2.682,
759
+ "step": 500
760
+ },
761
+ {
762
+ "epoch": 0.51,
763
+ "grad_norm": 9.033146858215332,
764
+ "learning_rate": 4.9800000000000004e-05,
765
+ "loss": 1.8733,
766
+ "step": 510
767
+ },
768
+ {
769
+ "epoch": 0.51,
770
+ "eval_loss": 2.1788506507873535,
771
+ "eval_runtime": 92.6001,
772
+ "eval_samples_per_second": 10.799,
773
+ "eval_steps_per_second": 2.7,
774
+ "step": 510
775
+ },
776
+ {
777
+ "epoch": 0.52,
778
+ "grad_norm": 12.906278610229492,
779
+ "learning_rate": 4.96e-05,
780
+ "loss": 2.4447,
781
+ "step": 520
782
+ },
783
+ {
784
+ "epoch": 0.52,
785
+ "eval_loss": 2.170750379562378,
786
+ "eval_runtime": 88.3981,
787
+ "eval_samples_per_second": 11.312,
788
+ "eval_steps_per_second": 2.828,
789
+ "step": 520
790
+ },
791
+ {
792
+ "epoch": 0.53,
793
+ "grad_norm": 13.015795707702637,
794
+ "learning_rate": 4.94e-05,
795
+ "loss": 1.848,
796
+ "step": 530
797
+ },
798
+ {
799
+ "epoch": 0.53,
800
+ "eval_loss": 2.161830186843872,
801
+ "eval_runtime": 91.5465,
802
+ "eval_samples_per_second": 10.923,
803
+ "eval_steps_per_second": 2.731,
804
+ "step": 530
805
+ },
806
+ {
807
+ "epoch": 0.54,
808
+ "grad_norm": 12.636075019836426,
809
+ "learning_rate": 4.92e-05,
810
+ "loss": 2.2895,
811
+ "step": 540
812
+ },
813
+ {
814
+ "epoch": 0.54,
815
+ "eval_loss": 2.1537909507751465,
816
+ "eval_runtime": 86.8374,
817
+ "eval_samples_per_second": 11.516,
818
+ "eval_steps_per_second": 2.879,
819
+ "step": 540
820
+ },
821
+ {
822
+ "epoch": 0.55,
823
+ "grad_norm": 7.789541244506836,
824
+ "learning_rate": 4.9e-05,
825
+ "loss": 1.8584,
826
+ "step": 550
827
+ },
828
+ {
829
+ "epoch": 0.55,
830
+ "eval_loss": 2.148059129714966,
831
+ "eval_runtime": 89.2466,
832
+ "eval_samples_per_second": 11.205,
833
+ "eval_steps_per_second": 2.801,
834
+ "step": 550
835
+ },
836
+ {
837
+ "epoch": 0.56,
838
+ "grad_norm": 10.41247844696045,
839
+ "learning_rate": 4.88e-05,
840
+ "loss": 2.2246,
841
+ "step": 560
842
+ },
843
+ {
844
+ "epoch": 0.56,
845
+ "eval_loss": 2.1307997703552246,
846
+ "eval_runtime": 88.7743,
847
+ "eval_samples_per_second": 11.265,
848
+ "eval_steps_per_second": 2.816,
849
+ "step": 560
850
+ },
851
+ {
852
+ "epoch": 0.57,
853
+ "grad_norm": 8.509031295776367,
854
+ "learning_rate": 4.86e-05,
855
+ "loss": 2.1681,
856
+ "step": 570
857
+ },
858
+ {
859
+ "epoch": 0.57,
860
+ "eval_loss": 2.117568016052246,
861
+ "eval_runtime": 90.3366,
862
+ "eval_samples_per_second": 11.07,
863
+ "eval_steps_per_second": 2.767,
864
+ "step": 570
865
+ },
866
+ {
867
+ "epoch": 0.58,
868
+ "grad_norm": 5.50130558013916,
869
+ "learning_rate": 4.8400000000000004e-05,
870
+ "loss": 1.9694,
871
+ "step": 580
872
+ },
873
+ {
874
+ "epoch": 0.58,
875
+ "eval_loss": 2.111440896987915,
876
+ "eval_runtime": 92.8312,
877
+ "eval_samples_per_second": 10.772,
878
+ "eval_steps_per_second": 2.693,
879
+ "step": 580
880
+ },
881
+ {
882
+ "epoch": 0.59,
883
+ "grad_norm": 10.277608871459961,
884
+ "learning_rate": 4.82e-05,
885
+ "loss": 2.3159,
886
+ "step": 590
887
+ },
888
+ {
889
+ "epoch": 0.59,
890
+ "eval_loss": 2.1107966899871826,
891
+ "eval_runtime": 91.0566,
892
+ "eval_samples_per_second": 10.982,
893
+ "eval_steps_per_second": 2.746,
894
+ "step": 590
895
+ },
896
+ {
897
+ "epoch": 0.6,
898
+ "grad_norm": 10.415202140808105,
899
+ "learning_rate": 4.8e-05,
900
+ "loss": 2.0298,
901
+ "step": 600
902
+ },
903
+ {
904
+ "epoch": 0.6,
905
+ "eval_loss": 2.1007392406463623,
906
+ "eval_runtime": 90.2452,
907
+ "eval_samples_per_second": 11.081,
908
+ "eval_steps_per_second": 2.77,
909
+ "step": 600
910
+ },
911
+ {
912
+ "epoch": 0.61,
913
+ "grad_norm": 9.051558494567871,
914
+ "learning_rate": 4.78e-05,
915
+ "loss": 2.1856,
916
+ "step": 610
917
+ },
918
+ {
919
+ "epoch": 0.61,
920
+ "eval_loss": 2.0893638134002686,
921
+ "eval_runtime": 97.6967,
922
+ "eval_samples_per_second": 10.236,
923
+ "eval_steps_per_second": 2.559,
924
+ "step": 610
925
+ },
926
+ {
927
+ "epoch": 0.62,
928
+ "grad_norm": 6.953254222869873,
929
+ "learning_rate": 4.76e-05,
930
+ "loss": 1.8901,
931
+ "step": 620
932
+ },
933
+ {
934
+ "epoch": 0.62,
935
+ "eval_loss": 2.0824646949768066,
936
+ "eval_runtime": 91.0278,
937
+ "eval_samples_per_second": 10.986,
938
+ "eval_steps_per_second": 2.746,
939
+ "step": 620
940
+ },
941
+ {
942
+ "epoch": 0.63,
943
+ "grad_norm": 7.948352813720703,
944
+ "learning_rate": 4.74e-05,
945
+ "loss": 2.3327,
946
+ "step": 630
947
+ },
948
+ {
949
+ "epoch": 0.63,
950
+ "eval_loss": 2.074693441390991,
951
+ "eval_runtime": 101.8312,
952
+ "eval_samples_per_second": 9.82,
953
+ "eval_steps_per_second": 2.455,
954
+ "step": 630
955
+ },
956
+ {
957
+ "epoch": 0.64,
958
+ "grad_norm": 9.384662628173828,
959
+ "learning_rate": 4.72e-05,
960
+ "loss": 2.3253,
961
+ "step": 640
962
+ },
963
+ {
964
+ "epoch": 0.64,
965
+ "eval_loss": 2.067964792251587,
966
+ "eval_runtime": 93.9189,
967
+ "eval_samples_per_second": 10.647,
968
+ "eval_steps_per_second": 2.662,
969
+ "step": 640
970
+ },
971
+ {
972
+ "epoch": 0.65,
973
+ "grad_norm": 6.1795759201049805,
974
+ "learning_rate": 4.7e-05,
975
+ "loss": 2.189,
976
+ "step": 650
977
+ },
978
+ {
979
+ "epoch": 0.65,
980
+ "eval_loss": 2.0557608604431152,
981
+ "eval_runtime": 97.6001,
982
+ "eval_samples_per_second": 10.246,
983
+ "eval_steps_per_second": 2.561,
984
+ "step": 650
985
+ },
986
+ {
987
+ "epoch": 0.66,
988
+ "grad_norm": 9.642065048217773,
989
+ "learning_rate": 4.6800000000000006e-05,
990
+ "loss": 2.146,
991
+ "step": 660
992
+ },
993
+ {
994
+ "epoch": 0.66,
995
+ "eval_loss": 2.0541083812713623,
996
+ "eval_runtime": 98.8714,
997
+ "eval_samples_per_second": 10.114,
998
+ "eval_steps_per_second": 2.529,
999
+ "step": 660
1000
+ },
1001
+ {
1002
+ "epoch": 0.67,
1003
+ "grad_norm": 6.987886428833008,
1004
+ "learning_rate": 4.660000000000001e-05,
1005
+ "loss": 1.8356,
1006
+ "step": 670
1007
+ },
1008
+ {
1009
+ "epoch": 0.67,
1010
+ "eval_loss": 2.053222417831421,
1011
+ "eval_runtime": 90.3869,
1012
+ "eval_samples_per_second": 11.064,
1013
+ "eval_steps_per_second": 2.766,
1014
+ "step": 670
1015
+ },
1016
+ {
1017
+ "epoch": 0.68,
1018
+ "grad_norm": 10.291470527648926,
1019
+ "learning_rate": 4.64e-05,
1020
+ "loss": 2.3633,
1021
+ "step": 680
1022
+ },
1023
+ {
1024
+ "epoch": 0.68,
1025
+ "eval_loss": 2.0410830974578857,
1026
+ "eval_runtime": 92.923,
1027
+ "eval_samples_per_second": 10.762,
1028
+ "eval_steps_per_second": 2.69,
1029
+ "step": 680
1030
+ },
1031
+ {
1032
+ "epoch": 0.69,
1033
+ "grad_norm": 8.183449745178223,
1034
+ "learning_rate": 4.6200000000000005e-05,
1035
+ "loss": 2.1429,
1036
+ "step": 690
1037
+ },
1038
+ {
1039
+ "epoch": 0.69,
1040
+ "eval_loss": 2.035935640335083,
1041
+ "eval_runtime": 106.6596,
1042
+ "eval_samples_per_second": 9.376,
1043
+ "eval_steps_per_second": 2.344,
1044
+ "step": 690
1045
+ },
1046
+ {
1047
+ "epoch": 0.7,
1048
+ "grad_norm": 7.329190254211426,
1049
+ "learning_rate": 4.600000000000001e-05,
1050
+ "loss": 2.2302,
1051
+ "step": 700
1052
+ },
1053
+ {
1054
+ "epoch": 0.7,
1055
+ "eval_loss": 2.033024311065674,
1056
+ "eval_runtime": 100.5784,
1057
+ "eval_samples_per_second": 9.942,
1058
+ "eval_steps_per_second": 2.486,
1059
+ "step": 700
1060
+ },
1061
+ {
1062
+ "epoch": 0.71,
1063
+ "grad_norm": 9.20626163482666,
1064
+ "learning_rate": 4.58e-05,
1065
+ "loss": 2.0641,
1066
+ "step": 710
1067
+ },
1068
+ {
1069
+ "epoch": 0.71,
1070
+ "eval_loss": 2.025780439376831,
1071
+ "eval_runtime": 112.3754,
1072
+ "eval_samples_per_second": 8.899,
1073
+ "eval_steps_per_second": 2.225,
1074
+ "step": 710
1075
+ },
1076
+ {
1077
+ "epoch": 0.72,
1078
+ "grad_norm": 11.579280853271484,
1079
+ "learning_rate": 4.5600000000000004e-05,
1080
+ "loss": 2.0106,
1081
+ "step": 720
1082
+ },
1083
+ {
1084
+ "epoch": 0.72,
1085
+ "eval_loss": 2.023242950439453,
1086
+ "eval_runtime": 100.8727,
1087
+ "eval_samples_per_second": 9.913,
1088
+ "eval_steps_per_second": 2.478,
1089
+ "step": 720
1090
+ },
1091
+ {
1092
+ "epoch": 0.73,
1093
+ "grad_norm": 9.62726879119873,
1094
+ "learning_rate": 4.5400000000000006e-05,
1095
+ "loss": 2.1251,
1096
+ "step": 730
1097
+ },
1098
+ {
1099
+ "epoch": 0.73,
1100
+ "eval_loss": 2.019562244415283,
1101
+ "eval_runtime": 98.6758,
1102
+ "eval_samples_per_second": 10.134,
1103
+ "eval_steps_per_second": 2.534,
1104
+ "step": 730
1105
+ },
1106
+ {
1107
+ "epoch": 0.74,
1108
+ "grad_norm": 8.502236366271973,
1109
+ "learning_rate": 4.52e-05,
1110
+ "loss": 2.1068,
1111
+ "step": 740
1112
+ },
1113
+ {
1114
+ "epoch": 0.74,
1115
+ "eval_loss": 2.0147552490234375,
1116
+ "eval_runtime": 101.3373,
1117
+ "eval_samples_per_second": 9.868,
1118
+ "eval_steps_per_second": 2.467,
1119
+ "step": 740
1120
+ },
1121
+ {
1122
+ "epoch": 0.75,
1123
+ "grad_norm": 8.522944450378418,
1124
+ "learning_rate": 4.5e-05,
1125
+ "loss": 2.3803,
1126
+ "step": 750
1127
+ },
1128
+ {
1129
+ "epoch": 0.75,
1130
+ "eval_loss": 2.0065298080444336,
1131
+ "eval_runtime": 115.4706,
1132
+ "eval_samples_per_second": 8.66,
1133
+ "eval_steps_per_second": 2.165,
1134
+ "step": 750
1135
+ },
1136
+ {
1137
+ "epoch": 0.76,
1138
+ "grad_norm": 9.323762893676758,
1139
+ "learning_rate": 4.4800000000000005e-05,
1140
+ "loss": 2.2126,
1141
+ "step": 760
1142
+ },
1143
+ {
1144
+ "epoch": 0.76,
1145
+ "eval_loss": 2.000988006591797,
1146
+ "eval_runtime": 131.4565,
1147
+ "eval_samples_per_second": 7.607,
1148
+ "eval_steps_per_second": 1.902,
1149
+ "step": 760
1150
+ },
1151
+ {
1152
+ "epoch": 0.77,
1153
+ "grad_norm": 7.243586540222168,
1154
+ "learning_rate": 4.46e-05,
1155
+ "loss": 2.2732,
1156
+ "step": 770
1157
+ },
1158
+ {
1159
+ "epoch": 0.77,
1160
+ "eval_loss": 1.9990744590759277,
1161
+ "eval_runtime": 117.2201,
1162
+ "eval_samples_per_second": 8.531,
1163
+ "eval_steps_per_second": 2.133,
1164
+ "step": 770
1165
+ },
1166
+ {
1167
+ "epoch": 0.78,
1168
+ "grad_norm": 10.664429664611816,
1169
+ "learning_rate": 4.44e-05,
1170
+ "loss": 2.2521,
1171
+ "step": 780
1172
+ },
1173
+ {
1174
+ "epoch": 0.78,
1175
+ "eval_loss": 1.988065242767334,
1176
+ "eval_runtime": 105.5882,
1177
+ "eval_samples_per_second": 9.471,
1178
+ "eval_steps_per_second": 2.368,
1179
+ "step": 780
1180
+ },
1181
+ {
1182
+ "epoch": 0.79,
1183
+ "grad_norm": 9.04030704498291,
1184
+ "learning_rate": 4.4200000000000004e-05,
1185
+ "loss": 1.9933,
1186
+ "step": 790
1187
+ },
1188
+ {
1189
+ "epoch": 0.79,
1190
+ "eval_loss": 1.9845080375671387,
1191
+ "eval_runtime": 101.4877,
1192
+ "eval_samples_per_second": 9.853,
1193
+ "eval_steps_per_second": 2.463,
1194
+ "step": 790
1195
+ },
1196
+ {
1197
+ "epoch": 0.8,
1198
+ "grad_norm": 9.38503646850586,
1199
+ "learning_rate": 4.4000000000000006e-05,
1200
+ "loss": 1.9956,
1201
+ "step": 800
1202
+ },
1203
+ {
1204
+ "epoch": 0.8,
1205
+ "eval_loss": 1.9831095933914185,
1206
+ "eval_runtime": 148.7473,
1207
+ "eval_samples_per_second": 6.723,
1208
+ "eval_steps_per_second": 1.681,
1209
+ "step": 800
1210
+ },
1211
+ {
1212
+ "epoch": 0.81,
1213
+ "grad_norm": 6.387115955352783,
1214
+ "learning_rate": 4.38e-05,
1215
+ "loss": 2.0059,
1216
+ "step": 810
1217
+ },
1218
+ {
1219
+ "epoch": 0.81,
1220
+ "eval_loss": 1.976269006729126,
1221
+ "eval_runtime": 129.9491,
1222
+ "eval_samples_per_second": 7.695,
1223
+ "eval_steps_per_second": 1.924,
1224
+ "step": 810
1225
+ },
1226
+ {
1227
+ "epoch": 0.82,
1228
+ "grad_norm": 8.890172004699707,
1229
+ "learning_rate": 4.36e-05,
1230
+ "loss": 2.2373,
1231
+ "step": 820
1232
+ },
1233
+ {
1234
+ "epoch": 0.82,
1235
+ "eval_loss": 1.974748969078064,
1236
+ "eval_runtime": 164.1513,
1237
+ "eval_samples_per_second": 6.092,
1238
+ "eval_steps_per_second": 1.523,
1239
+ "step": 820
1240
+ },
1241
+ {
1242
+ "epoch": 0.83,
1243
+ "grad_norm": 10.762927055358887,
1244
+ "learning_rate": 4.3400000000000005e-05,
1245
+ "loss": 2.186,
1246
+ "step": 830
1247
+ },
1248
+ {
1249
+ "epoch": 0.83,
1250
+ "eval_loss": 1.9657306671142578,
1251
+ "eval_runtime": 142.9199,
1252
+ "eval_samples_per_second": 6.997,
1253
+ "eval_steps_per_second": 1.749,
1254
+ "step": 830
1255
+ },
1256
+ {
1257
+ "epoch": 0.84,
1258
+ "grad_norm": 9.816293716430664,
1259
+ "learning_rate": 4.32e-05,
1260
+ "loss": 1.6506,
1261
+ "step": 840
1262
+ },
1263
+ {
1264
+ "epoch": 0.84,
1265
+ "eval_loss": 1.9642084836959839,
1266
+ "eval_runtime": 173.606,
1267
+ "eval_samples_per_second": 5.76,
1268
+ "eval_steps_per_second": 1.44,
1269
+ "step": 840
1270
+ },
1271
+ {
1272
+ "epoch": 0.85,
1273
+ "grad_norm": 7.931281089782715,
1274
+ "learning_rate": 4.3e-05,
1275
+ "loss": 1.7588,
1276
+ "step": 850
1277
+ },
1278
+ {
1279
+ "epoch": 0.85,
1280
+ "eval_loss": 1.9623171091079712,
1281
+ "eval_runtime": 158.121,
1282
+ "eval_samples_per_second": 6.324,
1283
+ "eval_steps_per_second": 1.581,
1284
+ "step": 850
1285
+ },
1286
+ {
1287
+ "epoch": 0.86,
1288
+ "grad_norm": 8.325611114501953,
1289
+ "learning_rate": 4.2800000000000004e-05,
1290
+ "loss": 1.8801,
1291
+ "step": 860
1292
+ },
1293
+ {
1294
+ "epoch": 0.86,
1295
+ "eval_loss": 1.9604990482330322,
1296
+ "eval_runtime": 156.6306,
1297
+ "eval_samples_per_second": 6.384,
1298
+ "eval_steps_per_second": 1.596,
1299
+ "step": 860
1300
+ },
1301
+ {
1302
+ "epoch": 0.87,
1303
+ "grad_norm": 8.10270881652832,
1304
+ "learning_rate": 4.26e-05,
1305
+ "loss": 1.8894,
1306
+ "step": 870
1307
+ },
1308
+ {
1309
+ "epoch": 0.87,
1310
+ "eval_loss": 1.9559086561203003,
1311
+ "eval_runtime": 165.5477,
1312
+ "eval_samples_per_second": 6.041,
1313
+ "eval_steps_per_second": 1.51,
1314
+ "step": 870
1315
+ },
1316
+ {
1317
+ "epoch": 0.88,
1318
+ "grad_norm": 8.666308403015137,
1319
+ "learning_rate": 4.24e-05,
1320
+ "loss": 2.1945,
1321
+ "step": 880
1322
+ },
1323
+ {
1324
+ "epoch": 0.88,
1325
+ "eval_loss": 1.953808307647705,
1326
+ "eval_runtime": 155.1083,
1327
+ "eval_samples_per_second": 6.447,
1328
+ "eval_steps_per_second": 1.612,
1329
+ "step": 880
1330
+ },
1331
+ {
1332
+ "epoch": 0.89,
1333
+ "grad_norm": 7.216370105743408,
1334
+ "learning_rate": 4.22e-05,
1335
+ "loss": 2.1958,
1336
+ "step": 890
1337
+ },
1338
+ {
1339
+ "epoch": 0.89,
1340
+ "eval_loss": 1.9536200761795044,
1341
+ "eval_runtime": 153.3145,
1342
+ "eval_samples_per_second": 6.523,
1343
+ "eval_steps_per_second": 1.631,
1344
+ "step": 890
1345
+ },
1346
+ {
1347
+ "epoch": 0.9,
1348
+ "grad_norm": 7.325380802154541,
1349
+ "learning_rate": 4.2e-05,
1350
+ "loss": 1.8137,
1351
+ "step": 900
1352
+ },
1353
+ {
1354
+ "epoch": 0.9,
1355
+ "eval_loss": 1.9444538354873657,
1356
+ "eval_runtime": 157.7008,
1357
+ "eval_samples_per_second": 6.341,
1358
+ "eval_steps_per_second": 1.585,
1359
+ "step": 900
1360
+ },
1361
+ {
1362
+ "epoch": 0.91,
1363
+ "grad_norm": 6.625280857086182,
1364
+ "learning_rate": 4.18e-05,
1365
+ "loss": 1.8925,
1366
+ "step": 910
1367
+ },
1368
+ {
1369
+ "epoch": 0.91,
1370
+ "eval_loss": 1.9409335851669312,
1371
+ "eval_runtime": 149.5155,
1372
+ "eval_samples_per_second": 6.688,
1373
+ "eval_steps_per_second": 1.672,
1374
+ "step": 910
1375
+ },
1376
+ {
1377
+ "epoch": 0.92,
1378
+ "grad_norm": 5.897461414337158,
1379
+ "learning_rate": 4.16e-05,
1380
+ "loss": 1.8322,
1381
+ "step": 920
1382
+ },
1383
+ {
1384
+ "epoch": 0.92,
1385
+ "eval_loss": 1.9331930875778198,
1386
+ "eval_runtime": 127.1983,
1387
+ "eval_samples_per_second": 7.862,
1388
+ "eval_steps_per_second": 1.965,
1389
+ "step": 920
1390
+ },
1391
+ {
1392
+ "epoch": 0.93,
1393
+ "grad_norm": 9.08010482788086,
1394
+ "learning_rate": 4.14e-05,
1395
+ "loss": 2.042,
1396
+ "step": 930
1397
+ },
1398
+ {
1399
+ "epoch": 0.93,
1400
+ "eval_loss": 1.9316186904907227,
1401
+ "eval_runtime": 118.4354,
1402
+ "eval_samples_per_second": 8.443,
1403
+ "eval_steps_per_second": 2.111,
1404
+ "step": 930
1405
+ },
1406
+ {
1407
+ "epoch": 0.94,
1408
+ "grad_norm": 9.264948844909668,
1409
+ "learning_rate": 4.12e-05,
1410
+ "loss": 1.7826,
1411
+ "step": 940
1412
+ },
1413
+ {
1414
+ "epoch": 0.94,
1415
+ "eval_loss": 1.9262688159942627,
1416
+ "eval_runtime": 113.7982,
1417
+ "eval_samples_per_second": 8.787,
1418
+ "eval_steps_per_second": 2.197,
1419
+ "step": 940
1420
+ },
1421
+ {
1422
+ "epoch": 0.95,
1423
+ "grad_norm": 10.431904792785645,
1424
+ "learning_rate": 4.1e-05,
1425
+ "loss": 1.7883,
1426
+ "step": 950
1427
+ },
1428
+ {
1429
+ "epoch": 0.95,
1430
+ "eval_loss": 1.9221816062927246,
1431
+ "eval_runtime": 107.3673,
1432
+ "eval_samples_per_second": 9.314,
1433
+ "eval_steps_per_second": 2.328,
1434
+ "step": 950
1435
+ },
1436
+ {
1437
+ "epoch": 0.96,
1438
+ "grad_norm": 5.923398017883301,
1439
+ "learning_rate": 4.08e-05,
1440
+ "loss": 1.5646,
1441
+ "step": 960
1442
+ },
1443
+ {
1444
+ "epoch": 0.96,
1445
+ "eval_loss": 1.9227478504180908,
1446
+ "eval_runtime": 121.0769,
1447
+ "eval_samples_per_second": 8.259,
1448
+ "eval_steps_per_second": 2.065,
1449
+ "step": 960
1450
+ },
1451
+ {
1452
+ "epoch": 0.97,
1453
+ "grad_norm": 8.440842628479004,
1454
+ "learning_rate": 4.0600000000000004e-05,
1455
+ "loss": 1.6706,
1456
+ "step": 970
1457
+ },
1458
+ {
1459
+ "epoch": 0.97,
1460
+ "eval_loss": 1.9174282550811768,
1461
+ "eval_runtime": 121.92,
1462
+ "eval_samples_per_second": 8.202,
1463
+ "eval_steps_per_second": 2.051,
1464
+ "step": 970
1465
+ },
1466
+ {
1467
+ "epoch": 0.98,
1468
+ "grad_norm": 6.066262245178223,
1469
+ "learning_rate": 4.0400000000000006e-05,
1470
+ "loss": 1.6543,
1471
+ "step": 980
1472
+ },
1473
+ {
1474
+ "epoch": 0.98,
1475
+ "eval_loss": 1.915424108505249,
1476
+ "eval_runtime": 111.8461,
1477
+ "eval_samples_per_second": 8.941,
1478
+ "eval_steps_per_second": 2.235,
1479
+ "step": 980
1480
+ },
1481
+ {
1482
+ "epoch": 0.99,
1483
+ "grad_norm": 7.314166069030762,
1484
+ "learning_rate": 4.02e-05,
1485
+ "loss": 2.1565,
1486
+ "step": 990
1487
+ },
1488
+ {
1489
+ "epoch": 0.99,
1490
+ "eval_loss": 1.9187248945236206,
1491
+ "eval_runtime": 110.4362,
1492
+ "eval_samples_per_second": 9.055,
1493
+ "eval_steps_per_second": 2.264,
1494
+ "step": 990
1495
+ },
1496
+ {
1497
+ "epoch": 1.0,
1498
+ "grad_norm": 3.6229448318481445,
1499
+ "learning_rate": 4e-05,
1500
+ "loss": 1.7919,
1501
+ "step": 1000
1502
+ },
1503
+ {
1504
+ "epoch": 1.0,
1505
+ "eval_loss": 1.9095934629440308,
1506
+ "eval_runtime": 103.6218,
1507
+ "eval_samples_per_second": 9.65,
1508
+ "eval_steps_per_second": 2.413,
1509
+ "step": 1000
1510
+ },
1511
+ {
1512
+ "epoch": 1.01,
1513
+ "grad_norm": 7.193479537963867,
1514
+ "learning_rate": 3.9800000000000005e-05,
1515
+ "loss": 1.5053,
1516
+ "step": 1010
1517
+ },
1518
+ {
1519
+ "epoch": 1.01,
1520
+ "eval_loss": 1.9059890508651733,
1521
+ "eval_runtime": 128.523,
1522
+ "eval_samples_per_second": 7.781,
1523
+ "eval_steps_per_second": 1.945,
1524
+ "step": 1010
1525
+ },
1526
+ {
1527
+ "epoch": 1.02,
1528
+ "grad_norm": 8.449844360351562,
1529
+ "learning_rate": 3.960000000000001e-05,
1530
+ "loss": 1.7817,
1531
+ "step": 1020
1532
+ },
1533
+ {
1534
+ "epoch": 1.02,
1535
+ "eval_loss": 1.9032129049301147,
1536
+ "eval_runtime": 118.3594,
1537
+ "eval_samples_per_second": 8.449,
1538
+ "eval_steps_per_second": 2.112,
1539
+ "step": 1020
1540
+ },
1541
+ {
1542
+ "epoch": 1.03,
1543
+ "grad_norm": 8.57231330871582,
1544
+ "learning_rate": 3.94e-05,
1545
+ "loss": 1.9459,
1546
+ "step": 1030
1547
+ },
1548
+ {
1549
+ "epoch": 1.03,
1550
+ "eval_loss": 1.9003936052322388,
1551
+ "eval_runtime": 148.1691,
1552
+ "eval_samples_per_second": 6.749,
1553
+ "eval_steps_per_second": 1.687,
1554
+ "step": 1030
1555
+ },
1556
+ {
1557
+ "epoch": 1.04,
1558
+ "grad_norm": 7.97957181930542,
1559
+ "learning_rate": 3.9200000000000004e-05,
1560
+ "loss": 1.9442,
1561
+ "step": 1040
1562
+ },
1563
+ {
1564
+ "epoch": 1.04,
1565
+ "eval_loss": 1.8994051218032837,
1566
+ "eval_runtime": 138.9543,
1567
+ "eval_samples_per_second": 7.197,
1568
+ "eval_steps_per_second": 1.799,
1569
+ "step": 1040
1570
+ },
1571
+ {
1572
+ "epoch": 1.05,
1573
+ "grad_norm": 6.236805438995361,
1574
+ "learning_rate": 3.9000000000000006e-05,
1575
+ "loss": 1.6471,
1576
+ "step": 1050
1577
+ },
1578
+ {
1579
+ "epoch": 1.05,
1580
+ "eval_loss": 1.8940227031707764,
1581
+ "eval_runtime": 125.7231,
1582
+ "eval_samples_per_second": 7.954,
1583
+ "eval_steps_per_second": 1.988,
1584
+ "step": 1050
1585
+ },
1586
+ {
1587
+ "epoch": 1.06,
1588
+ "grad_norm": 8.682119369506836,
1589
+ "learning_rate": 3.88e-05,
1590
+ "loss": 1.5088,
1591
+ "step": 1060
1592
+ },
1593
+ {
1594
+ "epoch": 1.06,
1595
+ "eval_loss": 1.9049922227859497,
1596
+ "eval_runtime": 123.3481,
1597
+ "eval_samples_per_second": 8.107,
1598
+ "eval_steps_per_second": 2.027,
1599
+ "step": 1060
1600
+ },
1601
+ {
1602
+ "epoch": 1.07,
1603
+ "grad_norm": 7.925257682800293,
1604
+ "learning_rate": 3.86e-05,
1605
+ "loss": 1.5251,
1606
+ "step": 1070
1607
+ },
1608
+ {
1609
+ "epoch": 1.07,
1610
+ "eval_loss": 1.900003433227539,
1611
+ "eval_runtime": 128.7457,
1612
+ "eval_samples_per_second": 7.767,
1613
+ "eval_steps_per_second": 1.942,
1614
+ "step": 1070
1615
+ },
1616
+ {
1617
+ "epoch": 1.08,
1618
+ "grad_norm": 10.475297927856445,
1619
+ "learning_rate": 3.8400000000000005e-05,
1620
+ "loss": 2.0858,
1621
+ "step": 1080
1622
+ },
1623
+ {
1624
+ "epoch": 1.08,
1625
+ "eval_loss": 1.8909506797790527,
1626
+ "eval_runtime": 130.1802,
1627
+ "eval_samples_per_second": 7.682,
1628
+ "eval_steps_per_second": 1.92,
1629
+ "step": 1080
1630
+ },
1631
+ {
1632
+ "epoch": 1.09,
1633
+ "grad_norm": 8.481929779052734,
1634
+ "learning_rate": 3.82e-05,
1635
+ "loss": 1.9986,
1636
+ "step": 1090
1637
+ },
1638
+ {
1639
+ "epoch": 1.09,
1640
+ "eval_loss": 1.8873074054718018,
1641
+ "eval_runtime": 124.38,
1642
+ "eval_samples_per_second": 8.04,
1643
+ "eval_steps_per_second": 2.01,
1644
+ "step": 1090
1645
+ },
1646
+ {
1647
+ "epoch": 1.1,
1648
+ "grad_norm": 7.110391616821289,
1649
+ "learning_rate": 3.8e-05,
1650
+ "loss": 2.0248,
1651
+ "step": 1100
1652
+ },
1653
+ {
1654
+ "epoch": 1.1,
1655
+ "eval_loss": 1.8794257640838623,
1656
+ "eval_runtime": 113.42,
1657
+ "eval_samples_per_second": 8.817,
1658
+ "eval_steps_per_second": 2.204,
1659
+ "step": 1100
1660
+ },
1661
+ {
1662
+ "epoch": 1.11,
1663
+ "grad_norm": 5.841686725616455,
1664
+ "learning_rate": 3.7800000000000004e-05,
1665
+ "loss": 1.7042,
1666
+ "step": 1110
1667
+ },
1668
+ {
1669
+ "epoch": 1.11,
1670
+ "eval_loss": 1.8803025484085083,
1671
+ "eval_runtime": 109.2435,
1672
+ "eval_samples_per_second": 9.154,
1673
+ "eval_steps_per_second": 2.288,
1674
+ "step": 1110
1675
+ },
1676
+ {
1677
+ "epoch": 1.12,
1678
+ "grad_norm": 4.138281345367432,
1679
+ "learning_rate": 3.76e-05,
1680
+ "loss": 1.749,
1681
+ "step": 1120
1682
+ },
1683
+ {
1684
+ "epoch": 1.12,
1685
+ "eval_loss": 1.8786567449569702,
1686
+ "eval_runtime": 109.7489,
1687
+ "eval_samples_per_second": 9.112,
1688
+ "eval_steps_per_second": 2.278,
1689
+ "step": 1120
1690
+ },
1691
+ {
1692
+ "epoch": 1.13,
1693
+ "grad_norm": 8.400279998779297,
1694
+ "learning_rate": 3.74e-05,
1695
+ "loss": 1.8222,
1696
+ "step": 1130
1697
+ },
1698
+ {
1699
+ "epoch": 1.13,
1700
+ "eval_loss": 1.882354736328125,
1701
+ "eval_runtime": 118.6402,
1702
+ "eval_samples_per_second": 8.429,
1703
+ "eval_steps_per_second": 2.107,
1704
+ "step": 1130
1705
+ },
1706
+ {
1707
+ "epoch": 1.1400000000000001,
1708
+ "grad_norm": 6.2840166091918945,
1709
+ "learning_rate": 3.72e-05,
1710
+ "loss": 2.0553,
1711
+ "step": 1140
1712
+ },
1713
+ {
1714
+ "epoch": 1.1400000000000001,
1715
+ "eval_loss": 1.8792822360992432,
1716
+ "eval_runtime": 120.4519,
1717
+ "eval_samples_per_second": 8.302,
1718
+ "eval_steps_per_second": 2.076,
1719
+ "step": 1140
1720
+ },
1721
+ {
1722
+ "epoch": 1.15,
1723
+ "grad_norm": 7.492580413818359,
1724
+ "learning_rate": 3.7e-05,
1725
+ "loss": 1.7812,
1726
+ "step": 1150
1727
+ },
1728
+ {
1729
+ "epoch": 1.15,
1730
+ "eval_loss": 1.875270962715149,
1731
+ "eval_runtime": 125.0379,
1732
+ "eval_samples_per_second": 7.998,
1733
+ "eval_steps_per_second": 1.999,
1734
+ "step": 1150
1735
+ },
1736
+ {
1737
+ "epoch": 1.16,
1738
+ "grad_norm": 7.373579978942871,
1739
+ "learning_rate": 3.68e-05,
1740
+ "loss": 1.8063,
1741
+ "step": 1160
1742
+ },
1743
+ {
1744
+ "epoch": 1.16,
1745
+ "eval_loss": 1.8711602687835693,
1746
+ "eval_runtime": 123.1041,
1747
+ "eval_samples_per_second": 8.123,
1748
+ "eval_steps_per_second": 2.031,
1749
+ "step": 1160
1750
+ },
1751
+ {
1752
+ "epoch": 1.17,
1753
+ "grad_norm": 8.983260154724121,
1754
+ "learning_rate": 3.66e-05,
1755
+ "loss": 2.0774,
1756
+ "step": 1170
1757
+ },
1758
+ {
1759
+ "epoch": 1.17,
1760
+ "eval_loss": 1.8704919815063477,
1761
+ "eval_runtime": 115.7789,
1762
+ "eval_samples_per_second": 8.637,
1763
+ "eval_steps_per_second": 2.159,
1764
+ "step": 1170
1765
+ },
1766
+ {
1767
+ "epoch": 1.18,
1768
+ "grad_norm": 8.41231918334961,
1769
+ "learning_rate": 3.6400000000000004e-05,
1770
+ "loss": 1.5289,
1771
+ "step": 1180
1772
+ },
1773
+ {
1774
+ "epoch": 1.18,
1775
+ "eval_loss": 1.8638745546340942,
1776
+ "eval_runtime": 108.5452,
1777
+ "eval_samples_per_second": 9.213,
1778
+ "eval_steps_per_second": 2.303,
1779
+ "step": 1180
1780
+ },
1781
+ {
1782
+ "epoch": 1.19,
1783
+ "grad_norm": 7.630991458892822,
1784
+ "learning_rate": 3.62e-05,
1785
+ "loss": 1.7938,
1786
+ "step": 1190
1787
+ },
1788
+ {
1789
+ "epoch": 1.19,
1790
+ "eval_loss": 1.8574650287628174,
1791
+ "eval_runtime": 127.6165,
1792
+ "eval_samples_per_second": 7.836,
1793
+ "eval_steps_per_second": 1.959,
1794
+ "step": 1190
1795
+ },
1796
+ {
1797
+ "epoch": 1.2,
1798
+ "grad_norm": 9.092334747314453,
1799
+ "learning_rate": 3.6e-05,
1800
+ "loss": 1.8305,
1801
+ "step": 1200
1802
+ },
1803
+ {
1804
+ "epoch": 1.2,
1805
+ "eval_loss": 1.8641369342803955,
1806
+ "eval_runtime": 102.0991,
1807
+ "eval_samples_per_second": 9.794,
1808
+ "eval_steps_per_second": 2.449,
1809
+ "step": 1200
1810
+ },
1811
+ {
1812
+ "epoch": 1.21,
1813
+ "grad_norm": 5.7870378494262695,
1814
+ "learning_rate": 3.58e-05,
1815
+ "loss": 1.6586,
1816
+ "step": 1210
1817
+ },
1818
+ {
1819
+ "epoch": 1.21,
1820
+ "eval_loss": 1.8618590831756592,
1821
+ "eval_runtime": 115.5522,
1822
+ "eval_samples_per_second": 8.654,
1823
+ "eval_steps_per_second": 2.164,
1824
+ "step": 1210
1825
+ },
1826
+ {
1827
+ "epoch": 1.22,
1828
+ "grad_norm": 6.22267484664917,
1829
+ "learning_rate": 3.56e-05,
1830
+ "loss": 1.7842,
1831
+ "step": 1220
1832
+ },
1833
+ {
1834
+ "epoch": 1.22,
1835
+ "eval_loss": 1.8590444326400757,
1836
+ "eval_runtime": 119.2331,
1837
+ "eval_samples_per_second": 8.387,
1838
+ "eval_steps_per_second": 2.097,
1839
+ "step": 1220
1840
+ },
1841
+ {
1842
+ "epoch": 1.23,
1843
+ "grad_norm": 6.128435134887695,
1844
+ "learning_rate": 3.54e-05,
1845
+ "loss": 1.6496,
1846
+ "step": 1230
1847
+ },
1848
+ {
1849
+ "epoch": 1.23,
1850
+ "eval_loss": 1.856571078300476,
1851
+ "eval_runtime": 115.6665,
1852
+ "eval_samples_per_second": 8.646,
1853
+ "eval_steps_per_second": 2.161,
1854
+ "step": 1230
1855
+ },
1856
+ {
1857
+ "epoch": 1.24,
1858
+ "grad_norm": 6.75801944732666,
1859
+ "learning_rate": 3.52e-05,
1860
+ "loss": 1.9095,
1861
+ "step": 1240
1862
+ },
1863
+ {
1864
+ "epoch": 1.24,
1865
+ "eval_loss": 1.8567980527877808,
1866
+ "eval_runtime": 117.0444,
1867
+ "eval_samples_per_second": 8.544,
1868
+ "eval_steps_per_second": 2.136,
1869
+ "step": 1240
1870
+ },
1871
+ {
1872
+ "epoch": 1.25,
1873
+ "grad_norm": 7.787381649017334,
1874
+ "learning_rate": 3.5e-05,
1875
+ "loss": 1.9586,
1876
+ "step": 1250
1877
+ },
1878
+ {
1879
+ "epoch": 1.25,
1880
+ "eval_loss": 1.8500583171844482,
1881
+ "eval_runtime": 112.5946,
1882
+ "eval_samples_per_second": 8.881,
1883
+ "eval_steps_per_second": 2.22,
1884
+ "step": 1250
1885
+ },
1886
+ {
1887
+ "epoch": 1.26,
1888
+ "grad_norm": 5.179034233093262,
1889
+ "learning_rate": 3.48e-05,
1890
+ "loss": 1.483,
1891
+ "step": 1260
1892
+ },
1893
+ {
1894
+ "epoch": 1.26,
1895
+ "eval_loss": 1.8572009801864624,
1896
+ "eval_runtime": 126.9259,
1897
+ "eval_samples_per_second": 7.879,
1898
+ "eval_steps_per_second": 1.97,
1899
+ "step": 1260
1900
+ },
1901
+ {
1902
+ "epoch": 1.27,
1903
+ "grad_norm": 9.003114700317383,
1904
+ "learning_rate": 3.46e-05,
1905
+ "loss": 1.5393,
1906
+ "step": 1270
1907
+ },
1908
+ {
1909
+ "epoch": 1.27,
1910
+ "eval_loss": 1.8491322994232178,
1911
+ "eval_runtime": 130.4481,
1912
+ "eval_samples_per_second": 7.666,
1913
+ "eval_steps_per_second": 1.916,
1914
+ "step": 1270
1915
+ },
1916
+ {
1917
+ "epoch": 1.28,
1918
+ "grad_norm": 9.390629768371582,
1919
+ "learning_rate": 3.4399999999999996e-05,
1920
+ "loss": 1.9644,
1921
+ "step": 1280
1922
+ },
1923
+ {
1924
+ "epoch": 1.28,
1925
+ "eval_loss": 1.8494079113006592,
1926
+ "eval_runtime": 133.2843,
1927
+ "eval_samples_per_second": 7.503,
1928
+ "eval_steps_per_second": 1.876,
1929
+ "step": 1280
1930
+ },
1931
+ {
1932
+ "epoch": 1.29,
1933
+ "grad_norm": 8.873686790466309,
1934
+ "learning_rate": 3.4200000000000005e-05,
1935
+ "loss": 1.8397,
1936
+ "step": 1290
1937
+ },
1938
+ {
1939
+ "epoch": 1.29,
1940
+ "eval_loss": 1.8387765884399414,
1941
+ "eval_runtime": 133.1956,
1942
+ "eval_samples_per_second": 7.508,
1943
+ "eval_steps_per_second": 1.877,
1944
+ "step": 1290
1945
+ },
1946
+ {
1947
+ "epoch": 1.3,
1948
+ "grad_norm": 8.92082691192627,
1949
+ "learning_rate": 3.4000000000000007e-05,
1950
+ "loss": 1.5783,
1951
+ "step": 1300
1952
+ },
1953
+ {
1954
+ "epoch": 1.3,
1955
+ "eval_loss": 1.8463449478149414,
1956
+ "eval_runtime": 132.1846,
1957
+ "eval_samples_per_second": 7.565,
1958
+ "eval_steps_per_second": 1.891,
1959
+ "step": 1300
1960
+ },
1961
+ {
1962
+ "epoch": 1.31,
1963
+ "grad_norm": 9.575182914733887,
1964
+ "learning_rate": 3.38e-05,
1965
+ "loss": 1.3814,
1966
+ "step": 1310
1967
+ },
1968
+ {
1969
+ "epoch": 1.31,
1970
+ "eval_loss": 1.8390209674835205,
1971
+ "eval_runtime": 112.5756,
1972
+ "eval_samples_per_second": 8.883,
1973
+ "eval_steps_per_second": 2.221,
1974
+ "step": 1310
1975
+ },
1976
+ {
1977
+ "epoch": 1.32,
1978
+ "grad_norm": 7.706173896789551,
1979
+ "learning_rate": 3.3600000000000004e-05,
1980
+ "loss": 1.9143,
1981
+ "step": 1320
1982
+ },
1983
+ {
1984
+ "epoch": 1.32,
1985
+ "eval_loss": 1.839473843574524,
1986
+ "eval_runtime": 133.4819,
1987
+ "eval_samples_per_second": 7.492,
1988
+ "eval_steps_per_second": 1.873,
1989
+ "step": 1320
1990
+ },
1991
+ {
1992
+ "epoch": 1.33,
1993
+ "grad_norm": 6.447656154632568,
1994
+ "learning_rate": 3.3400000000000005e-05,
1995
+ "loss": 1.8488,
1996
+ "step": 1330
1997
+ },
1998
+ {
1999
+ "epoch": 1.33,
2000
+ "eval_loss": 1.834652066230774,
2001
+ "eval_runtime": 124.1955,
2002
+ "eval_samples_per_second": 8.052,
2003
+ "eval_steps_per_second": 2.013,
2004
+ "step": 1330
2005
+ },
2006
+ {
2007
+ "epoch": 1.34,
2008
+ "grad_norm": 7.253933906555176,
2009
+ "learning_rate": 3.32e-05,
2010
+ "loss": 1.7947,
2011
+ "step": 1340
2012
+ },
2013
+ {
2014
+ "epoch": 1.34,
2015
+ "eval_loss": 1.830536127090454,
2016
+ "eval_runtime": 126.3069,
2017
+ "eval_samples_per_second": 7.917,
2018
+ "eval_steps_per_second": 1.979,
2019
+ "step": 1340
2020
+ },
2021
+ {
2022
+ "epoch": 1.35,
2023
+ "grad_norm": 7.001797199249268,
2024
+ "learning_rate": 3.3e-05,
2025
+ "loss": 1.9053,
2026
+ "step": 1350
2027
+ },
2028
+ {
2029
+ "epoch": 1.35,
2030
+ "eval_loss": 1.8331165313720703,
2031
+ "eval_runtime": 130.4927,
2032
+ "eval_samples_per_second": 7.663,
2033
+ "eval_steps_per_second": 1.916,
2034
+ "step": 1350
2035
+ },
2036
+ {
2037
+ "epoch": 1.3599999999999999,
2038
+ "grad_norm": 8.703651428222656,
2039
+ "learning_rate": 3.2800000000000004e-05,
2040
+ "loss": 1.8354,
2041
+ "step": 1360
2042
+ },
2043
+ {
2044
+ "epoch": 1.3599999999999999,
2045
+ "eval_loss": 1.8279739618301392,
2046
+ "eval_runtime": 110.2803,
2047
+ "eval_samples_per_second": 9.068,
2048
+ "eval_steps_per_second": 2.267,
2049
+ "step": 1360
2050
+ },
2051
+ {
2052
+ "epoch": 1.37,
2053
+ "grad_norm": 8.572402000427246,
2054
+ "learning_rate": 3.26e-05,
2055
+ "loss": 1.7067,
2056
+ "step": 1370
2057
+ },
2058
+ {
2059
+ "epoch": 1.37,
2060
+ "eval_loss": 1.8316876888275146,
2061
+ "eval_runtime": 107.9421,
2062
+ "eval_samples_per_second": 9.264,
2063
+ "eval_steps_per_second": 2.316,
2064
+ "step": 1370
2065
+ },
2066
+ {
2067
+ "epoch": 1.38,
2068
+ "grad_norm": 5.029556751251221,
2069
+ "learning_rate": 3.24e-05,
2070
+ "loss": 1.5889,
2071
+ "step": 1380
2072
+ },
2073
+ {
2074
+ "epoch": 1.38,
2075
+ "eval_loss": 1.8241760730743408,
2076
+ "eval_runtime": 119.602,
2077
+ "eval_samples_per_second": 8.361,
2078
+ "eval_steps_per_second": 2.09,
2079
+ "step": 1380
2080
+ },
2081
+ {
2082
+ "epoch": 1.3900000000000001,
2083
+ "grad_norm": 6.689691066741943,
2084
+ "learning_rate": 3.2200000000000003e-05,
2085
+ "loss": 1.8915,
2086
+ "step": 1390
2087
+ },
2088
+ {
2089
+ "epoch": 1.3900000000000001,
2090
+ "eval_loss": 1.8192516565322876,
2091
+ "eval_runtime": 124.5483,
2092
+ "eval_samples_per_second": 8.029,
2093
+ "eval_steps_per_second": 2.007,
2094
+ "step": 1390
2095
+ },
2096
+ {
2097
+ "epoch": 1.4,
2098
+ "grad_norm": 7.118290424346924,
2099
+ "learning_rate": 3.2000000000000005e-05,
2100
+ "loss": 1.8986,
2101
+ "step": 1400
2102
+ },
2103
+ {
2104
+ "epoch": 1.4,
2105
+ "eval_loss": 1.823815941810608,
2106
+ "eval_runtime": 113.9473,
2107
+ "eval_samples_per_second": 8.776,
2108
+ "eval_steps_per_second": 2.194,
2109
+ "step": 1400
2110
+ },
2111
+ {
2112
+ "epoch": 1.41,
2113
+ "grad_norm": 8.022662162780762,
2114
+ "learning_rate": 3.18e-05,
2115
+ "loss": 1.9665,
2116
+ "step": 1410
2117
+ },
2118
+ {
2119
+ "epoch": 1.41,
2120
+ "eval_loss": 1.8193540573120117,
2121
+ "eval_runtime": 125.65,
2122
+ "eval_samples_per_second": 7.959,
2123
+ "eval_steps_per_second": 1.99,
2124
+ "step": 1410
2125
+ },
2126
+ {
2127
+ "epoch": 1.42,
2128
+ "grad_norm": 6.133983135223389,
2129
+ "learning_rate": 3.16e-05,
2130
+ "loss": 1.6947,
2131
+ "step": 1420
2132
+ },
2133
+ {
2134
+ "epoch": 1.42,
2135
+ "eval_loss": 1.813571810722351,
2136
+ "eval_runtime": 116.6704,
2137
+ "eval_samples_per_second": 8.571,
2138
+ "eval_steps_per_second": 2.143,
2139
+ "step": 1420
2140
+ },
2141
+ {
2142
+ "epoch": 1.43,
2143
+ "grad_norm": 10.109593391418457,
2144
+ "learning_rate": 3.1400000000000004e-05,
2145
+ "loss": 1.8847,
2146
+ "step": 1430
2147
+ },
2148
+ {
2149
+ "epoch": 1.43,
2150
+ "eval_loss": 1.8171520233154297,
2151
+ "eval_runtime": 119.4118,
2152
+ "eval_samples_per_second": 8.374,
2153
+ "eval_steps_per_second": 2.094,
2154
+ "step": 1430
2155
+ },
2156
+ {
2157
+ "epoch": 1.44,
2158
+ "grad_norm": 6.10088586807251,
2159
+ "learning_rate": 3.12e-05,
2160
+ "loss": 2.0466,
2161
+ "step": 1440
2162
+ },
2163
+ {
2164
+ "epoch": 1.44,
2165
+ "eval_loss": 1.8149670362472534,
2166
+ "eval_runtime": 119.8555,
2167
+ "eval_samples_per_second": 8.343,
2168
+ "eval_steps_per_second": 2.086,
2169
+ "step": 1440
2170
+ },
2171
+ {
2172
+ "epoch": 1.45,
2173
+ "grad_norm": 6.327012538909912,
2174
+ "learning_rate": 3.1e-05,
2175
+ "loss": 1.6163,
2176
+ "step": 1450
2177
+ },
2178
+ {
2179
+ "epoch": 1.45,
2180
+ "eval_loss": 1.8084301948547363,
2181
+ "eval_runtime": 111.056,
2182
+ "eval_samples_per_second": 9.004,
2183
+ "eval_steps_per_second": 2.251,
2184
+ "step": 1450
2185
+ },
2186
+ {
2187
+ "epoch": 1.46,
2188
+ "grad_norm": 7.103956699371338,
2189
+ "learning_rate": 3.08e-05,
2190
+ "loss": 1.8625,
2191
+ "step": 1460
2192
+ },
2193
+ {
2194
+ "epoch": 1.46,
2195
+ "eval_loss": 1.8112941980361938,
2196
+ "eval_runtime": 133.3088,
2197
+ "eval_samples_per_second": 7.501,
2198
+ "eval_steps_per_second": 1.875,
2199
+ "step": 1460
2200
+ },
2201
+ {
2202
+ "epoch": 1.47,
2203
+ "grad_norm": 5.035983085632324,
2204
+ "learning_rate": 3.06e-05,
2205
+ "loss": 1.8754,
2206
+ "step": 1470
2207
+ },
2208
+ {
2209
+ "epoch": 1.47,
2210
+ "eval_loss": 1.8135524988174438,
2211
+ "eval_runtime": 128.5309,
2212
+ "eval_samples_per_second": 7.78,
2213
+ "eval_steps_per_second": 1.945,
2214
+ "step": 1470
2215
+ },
2216
+ {
2217
+ "epoch": 1.48,
2218
+ "grad_norm": 8.39537525177002,
2219
+ "learning_rate": 3.04e-05,
2220
+ "loss": 1.9341,
2221
+ "step": 1480
2222
+ },
2223
+ {
2224
+ "epoch": 1.48,
2225
+ "eval_loss": 1.80543851852417,
2226
+ "eval_runtime": 119.3552,
2227
+ "eval_samples_per_second": 8.378,
2228
+ "eval_steps_per_second": 2.095,
2229
+ "step": 1480
2230
+ },
2231
+ {
2232
+ "epoch": 1.49,
2233
+ "grad_norm": 8.173503875732422,
2234
+ "learning_rate": 3.02e-05,
2235
+ "loss": 2.0139,
2236
+ "step": 1490
2237
+ },
2238
+ {
2239
+ "epoch": 1.49,
2240
+ "eval_loss": 1.8039709329605103,
2241
+ "eval_runtime": 117.2862,
2242
+ "eval_samples_per_second": 8.526,
2243
+ "eval_steps_per_second": 2.132,
2244
+ "step": 1490
2245
+ },
2246
+ {
2247
+ "epoch": 1.5,
2248
+ "grad_norm": 8.277552604675293,
2249
+ "learning_rate": 3e-05,
2250
+ "loss": 1.7314,
2251
+ "step": 1500
2252
+ },
2253
+ {
2254
+ "epoch": 1.5,
2255
+ "eval_loss": 1.803325891494751,
2256
+ "eval_runtime": 109.3603,
2257
+ "eval_samples_per_second": 9.144,
2258
+ "eval_steps_per_second": 2.286,
2259
+ "step": 1500
2260
+ },
2261
+ {
2262
+ "epoch": 1.51,
2263
+ "grad_norm": 9.082454681396484,
2264
+ "learning_rate": 2.98e-05,
2265
+ "loss": 1.8701,
2266
+ "step": 1510
2267
+ },
2268
+ {
2269
+ "epoch": 1.51,
2270
+ "eval_loss": 1.8039252758026123,
2271
+ "eval_runtime": 120.7697,
2272
+ "eval_samples_per_second": 8.28,
2273
+ "eval_steps_per_second": 2.07,
2274
+ "step": 1510
2275
+ },
2276
+ {
2277
+ "epoch": 1.52,
2278
+ "grad_norm": 8.467123031616211,
2279
+ "learning_rate": 2.96e-05,
2280
+ "loss": 1.9261,
2281
+ "step": 1520
2282
+ },
2283
+ {
2284
+ "epoch": 1.52,
2285
+ "eval_loss": 1.8020439147949219,
2286
+ "eval_runtime": 153.5563,
2287
+ "eval_samples_per_second": 6.512,
2288
+ "eval_steps_per_second": 1.628,
2289
+ "step": 1520
2290
+ },
2291
+ {
2292
+ "epoch": 1.53,
2293
+ "grad_norm": 5.499407768249512,
2294
+ "learning_rate": 2.94e-05,
2295
+ "loss": 1.8215,
2296
+ "step": 1530
2297
+ },
2298
+ {
2299
+ "epoch": 1.53,
2300
+ "eval_loss": 1.7952924966812134,
2301
+ "eval_runtime": 155.9237,
2302
+ "eval_samples_per_second": 6.413,
2303
+ "eval_steps_per_second": 1.603,
2304
+ "step": 1530
2305
+ },
2306
+ {
2307
+ "epoch": 1.54,
2308
+ "grad_norm": 5.463346004486084,
2309
+ "learning_rate": 2.9199999999999998e-05,
2310
+ "loss": 1.7999,
2311
+ "step": 1540
2312
+ },
2313
+ {
2314
+ "epoch": 1.54,
2315
+ "eval_loss": 1.8003672361373901,
2316
+ "eval_runtime": 121.61,
2317
+ "eval_samples_per_second": 8.223,
2318
+ "eval_steps_per_second": 2.056,
2319
+ "step": 1540
2320
+ },
2321
+ {
2322
+ "epoch": 1.55,
2323
+ "grad_norm": 6.3344244956970215,
2324
+ "learning_rate": 2.9e-05,
2325
+ "loss": 1.7531,
2326
+ "step": 1550
2327
+ },
2328
+ {
2329
+ "epoch": 1.55,
2330
+ "eval_loss": 1.7988600730895996,
2331
+ "eval_runtime": 146.8067,
2332
+ "eval_samples_per_second": 6.812,
2333
+ "eval_steps_per_second": 1.703,
2334
+ "step": 1550
2335
+ },
2336
+ {
2337
+ "epoch": 1.56,
2338
+ "grad_norm": 10.925795555114746,
2339
+ "learning_rate": 2.88e-05,
2340
+ "loss": 1.9658,
2341
+ "step": 1560
2342
+ },
2343
+ {
2344
+ "epoch": 1.56,
2345
+ "eval_loss": 1.7972372770309448,
2346
+ "eval_runtime": 139.3218,
2347
+ "eval_samples_per_second": 7.178,
2348
+ "eval_steps_per_second": 1.794,
2349
+ "step": 1560
2350
+ },
2351
+ {
2352
+ "epoch": 1.5699999999999998,
2353
+ "grad_norm": 5.8058271408081055,
2354
+ "learning_rate": 2.86e-05,
2355
+ "loss": 1.7174,
2356
+ "step": 1570
2357
+ },
2358
+ {
2359
+ "epoch": 1.5699999999999998,
2360
+ "eval_loss": 1.7915621995925903,
2361
+ "eval_runtime": 128.6971,
2362
+ "eval_samples_per_second": 7.77,
2363
+ "eval_steps_per_second": 1.943,
2364
+ "step": 1570
2365
+ },
2366
+ {
2367
+ "epoch": 1.58,
2368
+ "grad_norm": 5.098724365234375,
2369
+ "learning_rate": 2.84e-05,
2370
+ "loss": 1.5216,
2371
+ "step": 1580
2372
+ },
2373
+ {
2374
+ "epoch": 1.58,
2375
+ "eval_loss": 1.7929859161376953,
2376
+ "eval_runtime": 147.7,
2377
+ "eval_samples_per_second": 6.77,
2378
+ "eval_steps_per_second": 1.693,
2379
+ "step": 1580
2380
+ },
2381
+ {
2382
+ "epoch": 1.5899999999999999,
2383
+ "grad_norm": 5.371964454650879,
2384
+ "learning_rate": 2.8199999999999998e-05,
2385
+ "loss": 1.7362,
2386
+ "step": 1590
2387
+ },
2388
+ {
2389
+ "epoch": 1.5899999999999999,
2390
+ "eval_loss": 1.7911442518234253,
2391
+ "eval_runtime": 163.4413,
2392
+ "eval_samples_per_second": 6.118,
2393
+ "eval_steps_per_second": 1.53,
2394
+ "step": 1590
2395
+ },
2396
+ {
2397
+ "epoch": 1.6,
2398
+ "grad_norm": 11.085586547851562,
2399
+ "learning_rate": 2.8000000000000003e-05,
2400
+ "loss": 1.866,
2401
+ "step": 1600
2402
+ },
2403
+ {
2404
+ "epoch": 1.6,
2405
+ "eval_loss": 1.7892712354660034,
2406
+ "eval_runtime": 140.8786,
2407
+ "eval_samples_per_second": 7.098,
2408
+ "eval_steps_per_second": 1.775,
2409
+ "step": 1600
2410
+ },
2411
+ {
2412
+ "epoch": 1.6099999999999999,
2413
+ "grad_norm": 6.125006675720215,
2414
+ "learning_rate": 2.7800000000000005e-05,
2415
+ "loss": 1.6734,
2416
+ "step": 1610
2417
+ },
2418
+ {
2419
+ "epoch": 1.6099999999999999,
2420
+ "eval_loss": 1.7879585027694702,
2421
+ "eval_runtime": 151.484,
2422
+ "eval_samples_per_second": 6.601,
2423
+ "eval_steps_per_second": 1.65,
2424
+ "step": 1610
2425
+ },
2426
+ {
2427
+ "epoch": 1.62,
2428
+ "grad_norm": 5.7702202796936035,
2429
+ "learning_rate": 2.7600000000000003e-05,
2430
+ "loss": 1.5737,
2431
+ "step": 1620
2432
+ },
2433
+ {
2434
+ "epoch": 1.62,
2435
+ "eval_loss": 1.784743309020996,
2436
+ "eval_runtime": 149.681,
2437
+ "eval_samples_per_second": 6.681,
2438
+ "eval_steps_per_second": 1.67,
2439
+ "step": 1620
2440
+ },
2441
+ {
2442
+ "epoch": 1.63,
2443
+ "grad_norm": 8.718552589416504,
2444
+ "learning_rate": 2.7400000000000002e-05,
2445
+ "loss": 1.5957,
2446
+ "step": 1630
2447
+ },
2448
+ {
2449
+ "epoch": 1.63,
2450
+ "eval_loss": 1.7875487804412842,
2451
+ "eval_runtime": 131.254,
2452
+ "eval_samples_per_second": 7.619,
2453
+ "eval_steps_per_second": 1.905,
2454
+ "step": 1630
2455
+ },
2456
+ {
2457
+ "epoch": 1.6400000000000001,
2458
+ "grad_norm": 7.785655498504639,
2459
+ "learning_rate": 2.7200000000000004e-05,
2460
+ "loss": 1.7956,
2461
+ "step": 1640
2462
+ },
2463
+ {
2464
+ "epoch": 1.6400000000000001,
2465
+ "eval_loss": 1.7841577529907227,
2466
+ "eval_runtime": 141.0508,
2467
+ "eval_samples_per_second": 7.09,
2468
+ "eval_steps_per_second": 1.772,
2469
+ "step": 1640
2470
+ },
2471
+ {
2472
+ "epoch": 1.65,
2473
+ "grad_norm": 8.244903564453125,
2474
+ "learning_rate": 2.7000000000000002e-05,
2475
+ "loss": 1.5131,
2476
+ "step": 1650
2477
+ },
2478
+ {
2479
+ "epoch": 1.65,
2480
+ "eval_loss": 1.7864689826965332,
2481
+ "eval_runtime": 153.0649,
2482
+ "eval_samples_per_second": 6.533,
2483
+ "eval_steps_per_second": 1.633,
2484
+ "step": 1650
2485
+ },
2486
+ {
2487
+ "epoch": 1.6600000000000001,
2488
+ "grad_norm": 9.530441284179688,
2489
+ "learning_rate": 2.6800000000000004e-05,
2490
+ "loss": 1.5239,
2491
+ "step": 1660
2492
+ },
2493
+ {
2494
+ "epoch": 1.6600000000000001,
2495
+ "eval_loss": 1.7876701354980469,
2496
+ "eval_runtime": 154.2237,
2497
+ "eval_samples_per_second": 6.484,
2498
+ "eval_steps_per_second": 1.621,
2499
+ "step": 1660
2500
+ },
2501
+ {
2502
+ "epoch": 1.67,
2503
+ "grad_norm": 7.262298583984375,
2504
+ "learning_rate": 2.6600000000000003e-05,
2505
+ "loss": 1.7847,
2506
+ "step": 1670
2507
+ },
2508
+ {
2509
+ "epoch": 1.67,
2510
+ "eval_loss": 1.7802685499191284,
2511
+ "eval_runtime": 146.4116,
2512
+ "eval_samples_per_second": 6.83,
2513
+ "eval_steps_per_second": 1.708,
2514
+ "step": 1670
2515
+ },
2516
+ {
2517
+ "epoch": 1.6800000000000002,
2518
+ "grad_norm": 5.070998668670654,
2519
+ "learning_rate": 2.64e-05,
2520
+ "loss": 1.9444,
2521
+ "step": 1680
2522
+ },
2523
+ {
2524
+ "epoch": 1.6800000000000002,
2525
+ "eval_loss": 1.779894232749939,
2526
+ "eval_runtime": 135.7223,
2527
+ "eval_samples_per_second": 7.368,
2528
+ "eval_steps_per_second": 1.842,
2529
+ "step": 1680
2530
+ },
2531
+ {
2532
+ "epoch": 1.69,
2533
+ "grad_norm": 8.932817459106445,
2534
+ "learning_rate": 2.6200000000000003e-05,
2535
+ "loss": 1.7473,
2536
+ "step": 1690
2537
+ },
2538
+ {
2539
+ "epoch": 1.69,
2540
+ "eval_loss": 1.7891018390655518,
2541
+ "eval_runtime": 133.6805,
2542
+ "eval_samples_per_second": 7.481,
2543
+ "eval_steps_per_second": 1.87,
2544
+ "step": 1690
2545
+ },
2546
+ {
2547
+ "epoch": 1.7,
2548
+ "grad_norm": 7.791315078735352,
2549
+ "learning_rate": 2.6000000000000002e-05,
2550
+ "loss": 1.6432,
2551
+ "step": 1700
2552
+ },
2553
+ {
2554
+ "epoch": 1.7,
2555
+ "eval_loss": 1.7791037559509277,
2556
+ "eval_runtime": 156.5434,
2557
+ "eval_samples_per_second": 6.388,
2558
+ "eval_steps_per_second": 1.597,
2559
+ "step": 1700
2560
+ },
2561
+ {
2562
+ "epoch": 1.71,
2563
+ "grad_norm": 6.859830379486084,
2564
+ "learning_rate": 2.58e-05,
2565
+ "loss": 1.6596,
2566
+ "step": 1710
2567
+ },
2568
+ {
2569
+ "epoch": 1.71,
2570
+ "eval_loss": 1.7749369144439697,
2571
+ "eval_runtime": 156.7103,
2572
+ "eval_samples_per_second": 6.381,
2573
+ "eval_steps_per_second": 1.595,
2574
+ "step": 1710
2575
+ },
2576
+ {
2577
+ "epoch": 1.72,
2578
+ "grad_norm": 4.347317695617676,
2579
+ "learning_rate": 2.5600000000000002e-05,
2580
+ "loss": 1.3664,
2581
+ "step": 1720
2582
+ },
2583
+ {
2584
+ "epoch": 1.72,
2585
+ "eval_loss": 1.7769114971160889,
2586
+ "eval_runtime": 151.4431,
2587
+ "eval_samples_per_second": 6.603,
2588
+ "eval_steps_per_second": 1.651,
2589
+ "step": 1720
2590
+ },
2591
+ {
2592
+ "epoch": 1.73,
2593
+ "grad_norm": 8.920775413513184,
2594
+ "learning_rate": 2.54e-05,
2595
+ "loss": 1.7536,
2596
+ "step": 1730
2597
+ },
2598
+ {
2599
+ "epoch": 1.73,
2600
+ "eval_loss": 1.779599905014038,
2601
+ "eval_runtime": 171.3944,
2602
+ "eval_samples_per_second": 5.834,
2603
+ "eval_steps_per_second": 1.459,
2604
+ "step": 1730
2605
+ },
2606
+ {
2607
+ "epoch": 1.74,
2608
+ "grad_norm": 9.783912658691406,
2609
+ "learning_rate": 2.5200000000000003e-05,
2610
+ "loss": 1.9861,
2611
+ "step": 1740
2612
+ },
2613
+ {
2614
+ "epoch": 1.74,
2615
+ "eval_loss": 1.7734276056289673,
2616
+ "eval_runtime": 155.338,
2617
+ "eval_samples_per_second": 6.438,
2618
+ "eval_steps_per_second": 1.609,
2619
+ "step": 1740
2620
+ },
2621
+ {
2622
+ "epoch": 1.75,
2623
+ "grad_norm": 6.196943759918213,
2624
+ "learning_rate": 2.5e-05,
2625
+ "loss": 1.8841,
2626
+ "step": 1750
2627
+ },
2628
+ {
2629
+ "epoch": 1.75,
2630
+ "eval_loss": 1.7724976539611816,
2631
+ "eval_runtime": 158.7667,
2632
+ "eval_samples_per_second": 6.299,
2633
+ "eval_steps_per_second": 1.575,
2634
+ "step": 1750
2635
+ },
2636
+ {
2637
+ "epoch": 1.76,
2638
+ "grad_norm": 5.560904502868652,
2639
+ "learning_rate": 2.48e-05,
2640
+ "loss": 1.7785,
2641
+ "step": 1760
2642
+ },
2643
+ {
2644
+ "epoch": 1.76,
2645
+ "eval_loss": 1.7708581686019897,
2646
+ "eval_runtime": 143.0319,
2647
+ "eval_samples_per_second": 6.991,
2648
+ "eval_steps_per_second": 1.748,
2649
+ "step": 1760
2650
+ },
2651
+ {
2652
+ "epoch": 1.77,
2653
+ "grad_norm": 6.589326858520508,
2654
+ "learning_rate": 2.46e-05,
2655
+ "loss": 1.7063,
2656
+ "step": 1770
2657
+ },
2658
+ {
2659
+ "epoch": 1.77,
2660
+ "eval_loss": 1.7703443765640259,
2661
+ "eval_runtime": 145.0987,
2662
+ "eval_samples_per_second": 6.892,
2663
+ "eval_steps_per_second": 1.723,
2664
+ "step": 1770
2665
+ },
2666
+ {
2667
+ "epoch": 1.78,
2668
+ "grad_norm": 8.978352546691895,
2669
+ "learning_rate": 2.44e-05,
2670
+ "loss": 2.0151,
2671
+ "step": 1780
2672
+ },
2673
+ {
2674
+ "epoch": 1.78,
2675
+ "eval_loss": 1.7695786952972412,
2676
+ "eval_runtime": 151.0921,
2677
+ "eval_samples_per_second": 6.618,
2678
+ "eval_steps_per_second": 1.655,
2679
+ "step": 1780
2680
+ },
2681
+ {
2682
+ "epoch": 1.79,
2683
+ "grad_norm": 8.987360000610352,
2684
+ "learning_rate": 2.4200000000000002e-05,
2685
+ "loss": 1.5089,
2686
+ "step": 1790
2687
+ },
2688
+ {
2689
+ "epoch": 1.79,
2690
+ "eval_loss": 1.768951177597046,
2691
+ "eval_runtime": 156.8603,
2692
+ "eval_samples_per_second": 6.375,
2693
+ "eval_steps_per_second": 1.594,
2694
+ "step": 1790
2695
+ },
2696
+ {
2697
+ "epoch": 1.8,
2698
+ "grad_norm": 7.1296186447143555,
2699
+ "learning_rate": 2.4e-05,
2700
+ "loss": 1.7401,
2701
+ "step": 1800
2702
+ },
2703
+ {
2704
+ "epoch": 1.8,
2705
+ "eval_loss": 1.7716623544692993,
2706
+ "eval_runtime": 157.509,
2707
+ "eval_samples_per_second": 6.349,
2708
+ "eval_steps_per_second": 1.587,
2709
+ "step": 1800
2710
+ },
2711
+ {
2712
+ "epoch": 1.81,
2713
+ "grad_norm": 10.590949058532715,
2714
+ "learning_rate": 2.38e-05,
2715
+ "loss": 1.7703,
2716
+ "step": 1810
2717
+ },
2718
+ {
2719
+ "epoch": 1.81,
2720
+ "eval_loss": 1.7669039964675903,
2721
+ "eval_runtime": 142.8584,
2722
+ "eval_samples_per_second": 7.0,
2723
+ "eval_steps_per_second": 1.75,
2724
+ "step": 1810
2725
+ },
2726
+ {
2727
+ "epoch": 1.8199999999999998,
2728
+ "grad_norm": 7.991542339324951,
2729
+ "learning_rate": 2.36e-05,
2730
+ "loss": 2.0104,
2731
+ "step": 1820
2732
+ },
2733
+ {
2734
+ "epoch": 1.8199999999999998,
2735
+ "eval_loss": 1.7663322687149048,
2736
+ "eval_runtime": 145.2675,
2737
+ "eval_samples_per_second": 6.884,
2738
+ "eval_steps_per_second": 1.721,
2739
+ "step": 1820
2740
+ },
2741
+ {
2742
+ "epoch": 1.83,
2743
+ "grad_norm": 7.469418525695801,
2744
+ "learning_rate": 2.3400000000000003e-05,
2745
+ "loss": 1.9194,
2746
+ "step": 1830
2747
+ },
2748
+ {
2749
+ "epoch": 1.83,
2750
+ "eval_loss": 1.76361083984375,
2751
+ "eval_runtime": 139.8496,
2752
+ "eval_samples_per_second": 7.151,
2753
+ "eval_steps_per_second": 1.788,
2754
+ "step": 1830
2755
+ },
2756
+ {
2757
+ "epoch": 1.8399999999999999,
2758
+ "grad_norm": 8.684300422668457,
2759
+ "learning_rate": 2.32e-05,
2760
+ "loss": 1.7216,
2761
+ "step": 1840
2762
+ },
2763
+ {
2764
+ "epoch": 1.8399999999999999,
2765
+ "eval_loss": 1.764965295791626,
2766
+ "eval_runtime": 124.8219,
2767
+ "eval_samples_per_second": 8.011,
2768
+ "eval_steps_per_second": 2.003,
2769
+ "step": 1840
2770
+ },
2771
+ {
2772
+ "epoch": 1.85,
2773
+ "grad_norm": 6.893486976623535,
2774
+ "learning_rate": 2.3000000000000003e-05,
2775
+ "loss": 1.7436,
2776
+ "step": 1850
2777
+ },
2778
+ {
2779
+ "epoch": 1.85,
2780
+ "eval_loss": 1.7624765634536743,
2781
+ "eval_runtime": 151.3283,
2782
+ "eval_samples_per_second": 6.608,
2783
+ "eval_steps_per_second": 1.652,
2784
+ "step": 1850
2785
+ },
2786
+ {
2787
+ "epoch": 1.8599999999999999,
2788
+ "grad_norm": 7.313007354736328,
2789
+ "learning_rate": 2.2800000000000002e-05,
2790
+ "loss": 1.8161,
2791
+ "step": 1860
2792
+ },
2793
+ {
2794
+ "epoch": 1.8599999999999999,
2795
+ "eval_loss": 1.759124994277954,
2796
+ "eval_runtime": 155.9019,
2797
+ "eval_samples_per_second": 6.414,
2798
+ "eval_steps_per_second": 1.604,
2799
+ "step": 1860
2800
+ },
2801
+ {
2802
+ "epoch": 1.87,
2803
+ "grad_norm": 6.408918857574463,
2804
+ "learning_rate": 2.26e-05,
2805
+ "loss": 1.8378,
2806
+ "step": 1870
2807
+ },
2808
+ {
2809
+ "epoch": 1.87,
2810
+ "eval_loss": 1.7628072500228882,
2811
+ "eval_runtime": 132.3341,
2812
+ "eval_samples_per_second": 7.557,
2813
+ "eval_steps_per_second": 1.889,
2814
+ "step": 1870
2815
+ },
2816
+ {
2817
+ "epoch": 1.88,
2818
+ "grad_norm": 6.225015163421631,
2819
+ "learning_rate": 2.2400000000000002e-05,
2820
+ "loss": 1.7146,
2821
+ "step": 1880
2822
+ },
2823
+ {
2824
+ "epoch": 1.88,
2825
+ "eval_loss": 1.7633777856826782,
2826
+ "eval_runtime": 130.7812,
2827
+ "eval_samples_per_second": 7.646,
2828
+ "eval_steps_per_second": 1.912,
2829
+ "step": 1880
2830
+ },
2831
+ {
2832
+ "epoch": 1.8900000000000001,
2833
+ "grad_norm": 1.981947898864746,
2834
+ "learning_rate": 2.22e-05,
2835
+ "loss": 1.6352,
2836
+ "step": 1890
2837
+ },
2838
+ {
2839
+ "epoch": 1.8900000000000001,
2840
+ "eval_loss": 1.7599279880523682,
2841
+ "eval_runtime": 117.9616,
2842
+ "eval_samples_per_second": 8.477,
2843
+ "eval_steps_per_second": 2.119,
2844
+ "step": 1890
2845
+ },
2846
+ {
2847
+ "epoch": 1.9,
2848
+ "grad_norm": 7.133483409881592,
2849
+ "learning_rate": 2.2000000000000003e-05,
2850
+ "loss": 1.9264,
2851
+ "step": 1900
2852
+ },
2853
+ {
2854
+ "epoch": 1.9,
2855
+ "eval_loss": 1.7584002017974854,
2856
+ "eval_runtime": 109.1257,
2857
+ "eval_samples_per_second": 9.164,
2858
+ "eval_steps_per_second": 2.291,
2859
+ "step": 1900
2860
+ },
2861
+ {
2862
+ "epoch": 1.9100000000000001,
2863
+ "grad_norm": 3.5345520973205566,
2864
+ "learning_rate": 2.18e-05,
2865
+ "loss": 1.2704,
2866
+ "step": 1910
2867
+ },
2868
+ {
2869
+ "epoch": 1.9100000000000001,
2870
+ "eval_loss": 1.7551823854446411,
2871
+ "eval_runtime": 129.6678,
2872
+ "eval_samples_per_second": 7.712,
2873
+ "eval_steps_per_second": 1.928,
2874
+ "step": 1910
2875
+ },
2876
+ {
2877
+ "epoch": 1.92,
2878
+ "grad_norm": 8.16379451751709,
2879
+ "learning_rate": 2.16e-05,
2880
+ "loss": 1.5037,
2881
+ "step": 1920
2882
+ },
2883
+ {
2884
+ "epoch": 1.92,
2885
+ "eval_loss": 1.7550432682037354,
2886
+ "eval_runtime": 107.3423,
2887
+ "eval_samples_per_second": 9.316,
2888
+ "eval_steps_per_second": 2.329,
2889
+ "step": 1920
2890
+ },
2891
+ {
2892
+ "epoch": 1.9300000000000002,
2893
+ "grad_norm": 5.34440803527832,
2894
+ "learning_rate": 2.1400000000000002e-05,
2895
+ "loss": 1.5265,
2896
+ "step": 1930
2897
+ },
2898
+ {
2899
+ "epoch": 1.9300000000000002,
2900
+ "eval_loss": 1.7561466693878174,
2901
+ "eval_runtime": 106.8984,
2902
+ "eval_samples_per_second": 9.355,
2903
+ "eval_steps_per_second": 2.339,
2904
+ "step": 1930
2905
+ },
2906
+ {
2907
+ "epoch": 1.94,
2908
+ "grad_norm": 7.982861042022705,
2909
+ "learning_rate": 2.12e-05,
2910
+ "loss": 1.5862,
2911
+ "step": 1940
2912
+ },
2913
+ {
2914
+ "epoch": 1.94,
2915
+ "eval_loss": 1.7594772577285767,
2916
+ "eval_runtime": 101.5387,
2917
+ "eval_samples_per_second": 9.848,
2918
+ "eval_steps_per_second": 2.462,
2919
+ "step": 1940
2920
+ },
2921
+ {
2922
+ "epoch": 1.95,
2923
+ "grad_norm": 6.90481424331665,
2924
+ "learning_rate": 2.1e-05,
2925
+ "loss": 1.6048,
2926
+ "step": 1950
2927
+ },
2928
+ {
2929
+ "epoch": 1.95,
2930
+ "eval_loss": 1.7539196014404297,
2931
+ "eval_runtime": 105.056,
2932
+ "eval_samples_per_second": 9.519,
2933
+ "eval_steps_per_second": 2.38,
2934
+ "step": 1950
2935
+ },
2936
+ {
2937
+ "epoch": 1.96,
2938
+ "grad_norm": 5.775398254394531,
2939
+ "learning_rate": 2.08e-05,
2940
+ "loss": 1.572,
2941
+ "step": 1960
2942
+ },
2943
+ {
2944
+ "epoch": 1.96,
2945
+ "eval_loss": 1.750312328338623,
2946
+ "eval_runtime": 111.7724,
2947
+ "eval_samples_per_second": 8.947,
2948
+ "eval_steps_per_second": 2.237,
2949
+ "step": 1960
2950
+ },
2951
+ {
2952
+ "epoch": 1.97,
2953
+ "grad_norm": 6.477075099945068,
2954
+ "learning_rate": 2.06e-05,
2955
+ "loss": 1.8335,
2956
+ "step": 1970
2957
+ },
2958
+ {
2959
+ "epoch": 1.97,
2960
+ "eval_loss": 1.7535051107406616,
2961
+ "eval_runtime": 125.6752,
2962
+ "eval_samples_per_second": 7.957,
2963
+ "eval_steps_per_second": 1.989,
2964
+ "step": 1970
2965
+ },
2966
+ {
2967
+ "epoch": 1.98,
2968
+ "grad_norm": 6.374350547790527,
2969
+ "learning_rate": 2.04e-05,
2970
+ "loss": 1.8152,
2971
+ "step": 1980
2972
+ },
2973
+ {
2974
+ "epoch": 1.98,
2975
+ "eval_loss": 1.7487550973892212,
2976
+ "eval_runtime": 137.3756,
2977
+ "eval_samples_per_second": 7.279,
2978
+ "eval_steps_per_second": 1.82,
2979
+ "step": 1980
2980
+ },
2981
+ {
2982
+ "epoch": 1.99,
2983
+ "grad_norm": 6.992517948150635,
2984
+ "learning_rate": 2.0200000000000003e-05,
2985
+ "loss": 1.7704,
2986
+ "step": 1990
2987
+ },
2988
+ {
2989
+ "epoch": 1.99,
2990
+ "eval_loss": 1.7500554323196411,
2991
+ "eval_runtime": 145.872,
2992
+ "eval_samples_per_second": 6.855,
2993
+ "eval_steps_per_second": 1.714,
2994
+ "step": 1990
2995
+ },
2996
+ {
2997
+ "epoch": 2.0,
2998
+ "grad_norm": 6.812110424041748,
2999
+ "learning_rate": 2e-05,
3000
+ "loss": 1.967,
3001
+ "step": 2000
3002
+ },
3003
+ {
3004
+ "epoch": 2.0,
3005
+ "eval_loss": 1.7480344772338867,
3006
+ "eval_runtime": 152.2467,
3007
+ "eval_samples_per_second": 6.568,
3008
+ "eval_steps_per_second": 1.642,
3009
+ "step": 2000
3010
+ }
3011
+ ],
3012
+ "logging_steps": 10,
3013
+ "max_steps": 3000,
3014
+ "num_input_tokens_seen": 0,
3015
+ "num_train_epochs": 3,
3016
+ "save_steps": 500,
3017
+ "stateful_callbacks": {
3018
+ "TrainerControl": {
3019
+ "args": {
3020
+ "should_epoch_stop": false,
3021
+ "should_evaluate": false,
3022
+ "should_log": false,
3023
+ "should_save": true,
3024
+ "should_training_stop": false
3025
+ },
3026
+ "attributes": {}
3027
+ }
3028
+ },
3029
+ "total_flos": 522584064000000.0,
3030
+ "train_batch_size": 4,
3031
+ "trial_name": null,
3032
+ "trial_params": null
3033
+ }
fine_tuned_gpt2/checkpoint-2500/config.json ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/Users/user/projects/MindPadiBig/mindpadi/models/gpt2",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "GPT2LMHeadModel"
6
+ ],
7
+ "attn_pdrop": 0.1,
8
+ "bos_token_id": 50256,
9
+ "embd_pdrop": 0.1,
10
+ "eos_token_id": 50256,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "reorder_and_upcast_attn": false,
21
+ "resid_pdrop": 0.1,
22
+ "scale_attn_by_inverse_layer_idx": false,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50
33
+ }
34
+ },
35
+ "torch_dtype": "float32",
36
+ "transformers_version": "4.49.0",
37
+ "use_cache": true,
38
+ "vocab_size": 50257
39
+ }
fine_tuned_gpt2/checkpoint-2500/generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 50256,
4
+ "eos_token_id": 50256,
5
+ "transformers_version": "4.49.0"
6
+ }