Synchronizing local compiler cache.
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +58 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/97050f455a3dc7373088.json +58 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/2515e75e29085415e970.json +62 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/96f6efa1fe88a6a9d908.json +62 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/d5321525bfafe56c1fa7.json +62 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/unsloth/Llama-3.2-1B-Instruct/7aa61ecee499b16d2a75.json +63 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/unsloth/Llama-3.2-1B-Instruct/ac22ae3a40ff509f85b5.json +63 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama4_text/tiny-random/llama-4/c108820859fb593e97ef.json +81 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/mixtral/dacorvo/Mixtral-tiny/d27d3b7e6c219c39e0c6.json +58 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/phi3/yujiepan/phi-4-tiny-random/3211e8079a23f7cca17d.json +59 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/Qwen/Qwen2.5-0.5B/708e8d742ec0e959ffbf.json +82 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/Qwen/Qwen2.5-0.5B/e2a98d277483d98cbc4a.json +82 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/yujiepan/qwen2.5-128k-tiny-random/55d828b008dd8415c97f.json +64 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/57719fe1acd233b9fab8.json +87 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/f4eb8a30c219ffacbbb4.json +87 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/c7cf2a4267f41d17facc.json +65 -0
- neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/smollm3/HuggingFaceTB/SmolLM3-3B/7a707f573aa21c74857f.json +134 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.neff +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/wrapped_neff.hlo +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.neff +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/wrapped_neff.hlo +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.hlo_module.pb +3 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.neff +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/compile_flags.json +1 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/model.done +0 -0
- neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/model.hlo_module.pb +3 -0
.gitattributes
CHANGED
|
@@ -13369,3 +13369,61 @@ neuronxcc-2.21.18209.0+043b1bf7/MODULE_d7e4ab61410d7ed2f2fe+002646c9/wrapped_nef
|
|
| 13369 |
neuronxcc-2.21.33363.0+82129205/MODULE_ecafd5c328f53085c3ac+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13370 |
neuronxcc-2.21.33363.0+82129205/MODULE_ecafd5c328f53085c3ac+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13371 |
neuronxcc-2.21.33363.0+82129205/MODULE_f9c594cf85ceddc7f3ad+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13369 |
neuronxcc-2.21.33363.0+82129205/MODULE_ecafd5c328f53085c3ac+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13370 |
neuronxcc-2.21.33363.0+82129205/MODULE_ecafd5c328f53085c3ac+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13371 |
neuronxcc-2.21.33363.0+82129205/MODULE_f9c594cf85ceddc7f3ad+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13372 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13373 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13374 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13375 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13376 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_155d3059a636dbde2772+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13377 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_1752954cd84d208a194c+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13378 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_1752954cd84d208a194c+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13379 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_1df28a40cf4145bdfcae+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13380 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_254c67a8dd8bfb2e240d+ac10809c/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13381 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_254c67a8dd8bfb2e240d+ac10809c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13382 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_32f05041dda2b4148a17+d3bd0a6b/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13383 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_32f05041dda2b4148a17+d3bd0a6b/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13384 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_3e60ddce17a65ac4ee68+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13385 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_410ef55a3aca5cd80fc8+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13386 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_54e3a5446bba5bfa79af+fb4cc044/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13387 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_62abb2454c6da8d57ad5+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13388 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_62abb2454c6da8d57ad5+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13389 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_70488145aaecc47ba5cb+70994d97/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13390 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_70488145aaecc47ba5cb+70994d97/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13391 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_733744f5680ca67f7ff9+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13392 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_73587bc7122ba5a81758+80d05c3f/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13393 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_73587bc7122ba5a81758+80d05c3f/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13394 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_75329916c4da1120954c+6e4949b4/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13395 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_7c590ab98f45f8a8dd15+283df001/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13396 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_7c9d7dd1d913b7d0da1a+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13397 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_8cc67d6573976e582c95+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13398 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_8eedc4b0995b3a9b6f52+677eeb9d/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13399 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_913d465d9ff4ee2a67a8+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13400 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_913d465d9ff4ee2a67a8+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13401 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_934d9b0b9c6b99b58d80+e49a1423/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13402 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_934d9b0b9c6b99b58d80+e49a1423/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13403 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_93b30fb36d12200ba503+cb72af16/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13404 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_93b30fb36d12200ba503+cb72af16/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13405 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_944763b36d810f4af582+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13406 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_99a19dff67d566a541e8+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13407 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_99a19dff67d566a541e8+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13408 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_b23271513f9095bde4db+2dde74c7/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13409 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_b23271513f9095bde4db+2dde74c7/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13410 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_b5dfe3f2e102fa0dc843+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13411 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_b5dfe3f2e102fa0dc843+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13412 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_b6ed8374075f070c66d6+a32116a7/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13413 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_bdf25cde1a48d420319c+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13414 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_bdf25cde1a48d420319c+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13415 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_c30a746238d3804f9626+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13416 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_c4737893663f5ff66b53+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13417 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_c82782e8097520d499e5+e0ac078c/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13418 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_c82782e8097520d499e5+e0ac078c/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13419 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_ccb75cdf21908f3399fe+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13420 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_ccb75cdf21908f3399fe+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13421 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_ccd99eedd56f698b9cb3+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13422 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_cde3417e6918fabff4a7+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13423 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_cde3417e6918fabff4a7+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13424 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_d30bb77760f1638377ba+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13425 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_d30bb77760f1638377ba+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
| 13426 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_e0ce3525251d35c118ed+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13427 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_f52fe8e88daf6dc9619a+24129607/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13428 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_fef9240b8fbfaca895ba+a02c3a36/model.neff filter=lfs diff=lfs merge=lfs -text
|
| 13429 |
+
neuronxcc-2.21.33363.0+82129205/MODULE_fef9240b8fbfaca895ba+a02c3a36/wrapped_neff.hlo filter=lfs diff=lfs merge=lfs -text
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/granite/hf-internal-testing/tiny-random-GraniteForCausalLM/97050f455a3dc7373088.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"GraniteForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"attention_multiplier": 1.0,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"embedding_multiplier": 1.0,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 32,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 64,
|
| 17 |
+
"logits_scaling": 1.0,
|
| 18 |
+
"max_position_embeddings": 2048,
|
| 19 |
+
"mlp_bias": false,
|
| 20 |
+
"model_type": "granite",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 23 |
+
"batch_size": 1,
|
| 24 |
+
"capacity_factor": null,
|
| 25 |
+
"checkpoint_id": "hf-internal-testing/tiny-random-GraniteForCausalLM",
|
| 26 |
+
"checkpoint_revision": "c3074ebc0ac2fe545305f5e5f6cce2cc9b2aa0c5",
|
| 27 |
+
"continuous_batching": false,
|
| 28 |
+
"ep_degree": 1,
|
| 29 |
+
"fused_qkv": true,
|
| 30 |
+
"glu_mlp": true,
|
| 31 |
+
"local_ranks_size": 2,
|
| 32 |
+
"max_batch_size": 1,
|
| 33 |
+
"max_context_length": 1024,
|
| 34 |
+
"max_topk": 256,
|
| 35 |
+
"n_active_tokens": 1024,
|
| 36 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 37 |
+
"on_device_sampling": true,
|
| 38 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 39 |
+
"output_logits": false,
|
| 40 |
+
"pp_degree": 1,
|
| 41 |
+
"sequence_length": 1024,
|
| 42 |
+
"speculation_length": 0,
|
| 43 |
+
"start_rank_id": 0,
|
| 44 |
+
"target": "trn1",
|
| 45 |
+
"torch_dtype": "float32",
|
| 46 |
+
"tp_degree": 2
|
| 47 |
+
},
|
| 48 |
+
"num_attention_heads": 4,
|
| 49 |
+
"num_hidden_layers": 2,
|
| 50 |
+
"num_key_value_heads": 4,
|
| 51 |
+
"residual_multiplier": 1.0,
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_scaling": null,
|
| 54 |
+
"rope_theta": 10000.0,
|
| 55 |
+
"tie_word_embeddings": false,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 49152
|
| 58 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/2515e75e29085415e970.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "float16",
|
| 11 |
+
"head_dim": 4,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 16,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 64,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 2,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
| 24 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
| 25 |
+
"continuous_batching": true,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 2,
|
| 31 |
+
"max_context_length": 128,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 128,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 128,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn2",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/96f6efa1fe88a6a9d908.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "float16",
|
| 11 |
+
"head_dim": 4,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 16,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 64,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
| 24 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 1024,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 1024,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 1024,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/llamafactory/tiny-random-Llama-3/d5321525bfafe56c1fa7.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "llamafactory/tiny-random-Llama-3",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "float16",
|
| 11 |
+
"head_dim": 4,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 16,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 64,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "llamafactory/tiny-random-Llama-3",
|
| 24 |
+
"checkpoint_revision": "bf2a2e3bf199ad2ee96f02a3c00246c608db22a8",
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 512,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 512,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 512,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "float16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 4,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 4,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 8.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": false,
|
| 60 |
+
"use_cache": true,
|
| 61 |
+
"vocab_size": 128256
|
| 62 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/unsloth/Llama-3.2-1B-Instruct/7aa61ecee499b16d2a75.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "unsloth/Llama-3.2-1B-Instruct",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 64,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 2048,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 8192,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "unsloth/Llama-3.2-1B-Instruct",
|
| 24 |
+
"checkpoint_revision": null,
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": false,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 4096,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 4096,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": false,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 4096,
|
| 40 |
+
"speculation_length": 5,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "bfloat16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 32,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"num_key_value_heads": 8,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 32.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": true,
|
| 60 |
+
"unsloth_fixed": true,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"vocab_size": 128256
|
| 63 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama/unsloth/Llama-3.2-1B-Instruct/ac22ae3a40ff509f85b5.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "unsloth/Llama-3.2-1B-Instruct",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"LlamaForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 64,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 2048,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 8192,
|
| 16 |
+
"max_position_embeddings": 131072,
|
| 17 |
+
"mlp_bias": false,
|
| 18 |
+
"model_type": "llama",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "unsloth/Llama-3.2-1B-Instruct",
|
| 24 |
+
"checkpoint_revision": null,
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": false,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 4096,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 4096,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": false,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 4096,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "bfloat16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 32,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"num_key_value_heads": 8,
|
| 49 |
+
"pretraining_tp": 1,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_scaling": {
|
| 52 |
+
"factor": 32.0,
|
| 53 |
+
"high_freq_factor": 4.0,
|
| 54 |
+
"low_freq_factor": 1.0,
|
| 55 |
+
"original_max_position_embeddings": 8192,
|
| 56 |
+
"rope_type": "llama3"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 500000.0,
|
| 59 |
+
"tie_word_embeddings": true,
|
| 60 |
+
"unsloth_fixed": true,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"vocab_size": 128256
|
| 63 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/llama4_text/tiny-random/llama-4/c108820859fb593e97ef.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "tiny-random/llama-4",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_chunk_size": 128,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_scale": 0.1,
|
| 9 |
+
"attn_temperature_tuning": 4,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"dtype": "bfloat16",
|
| 12 |
+
"floor_scale": 8192,
|
| 13 |
+
"for_llm_compressor": false,
|
| 14 |
+
"head_dim": 32,
|
| 15 |
+
"hidden_act": "silu",
|
| 16 |
+
"hidden_size": 32,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"interleave_moe_layer_step": 2,
|
| 19 |
+
"intermediate_size": 64,
|
| 20 |
+
"intermediate_size_mlp": 128,
|
| 21 |
+
"layer_types": [
|
| 22 |
+
"chunked_attention",
|
| 23 |
+
"chunked_attention",
|
| 24 |
+
"chunked_attention",
|
| 25 |
+
"full_attention"
|
| 26 |
+
],
|
| 27 |
+
"max_position_embeddings": 1048576,
|
| 28 |
+
"model_type": "llama4_text",
|
| 29 |
+
"moe_layers": [
|
| 30 |
+
1,
|
| 31 |
+
3
|
| 32 |
+
],
|
| 33 |
+
"neuron": {
|
| 34 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 35 |
+
"batch_size": 1,
|
| 36 |
+
"capacity_factor": null,
|
| 37 |
+
"checkpoint_id": "tiny-random/llama-4",
|
| 38 |
+
"checkpoint_revision": "9e716f5d4d1ffe0a44a15f46f4a12b840439aba4",
|
| 39 |
+
"continuous_batching": false,
|
| 40 |
+
"ep_degree": 1,
|
| 41 |
+
"fused_qkv": false,
|
| 42 |
+
"glu_mlp": true,
|
| 43 |
+
"local_ranks_size": 2,
|
| 44 |
+
"max_batch_size": 1,
|
| 45 |
+
"max_context_length": 1024,
|
| 46 |
+
"max_topk": 256,
|
| 47 |
+
"n_active_tokens": 1024,
|
| 48 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 49 |
+
"on_device_sampling": true,
|
| 50 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 51 |
+
"output_logits": false,
|
| 52 |
+
"pp_degree": 1,
|
| 53 |
+
"sequence_length": 1024,
|
| 54 |
+
"speculation_length": 0,
|
| 55 |
+
"start_rank_id": 0,
|
| 56 |
+
"target": "trn1",
|
| 57 |
+
"torch_dtype": "bfloat16",
|
| 58 |
+
"tp_degree": 2
|
| 59 |
+
},
|
| 60 |
+
"no_rope_layers": [
|
| 61 |
+
1,
|
| 62 |
+
1,
|
| 63 |
+
1,
|
| 64 |
+
0
|
| 65 |
+
],
|
| 66 |
+
"num_attention_heads": 1,
|
| 67 |
+
"num_experts_per_tok": 1,
|
| 68 |
+
"num_hidden_layers": 4,
|
| 69 |
+
"num_key_value_heads": 1,
|
| 70 |
+
"num_local_experts": 8,
|
| 71 |
+
"output_router_logits": false,
|
| 72 |
+
"rms_norm_eps": 1e-05,
|
| 73 |
+
"rope_scaling": null,
|
| 74 |
+
"rope_theta": 500000.0,
|
| 75 |
+
"router_aux_loss_coef": 0.001,
|
| 76 |
+
"router_jitter_noise": 0.0,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_qk_norm": true,
|
| 80 |
+
"vocab_size": 202048
|
| 81 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/mixtral/dacorvo/Mixtral-tiny/d27d3b7e6c219c39e0c6.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "dacorvo/Mixtral-tiny",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"MixtralForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "float16",
|
| 10 |
+
"head_dim": 32,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 1024,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3584,
|
| 15 |
+
"max_position_embeddings": 1024,
|
| 16 |
+
"model_type": "mixtral",
|
| 17 |
+
"neuron": {
|
| 18 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 19 |
+
"batch_size": 1,
|
| 20 |
+
"capacity_factor": null,
|
| 21 |
+
"checkpoint_id": "dacorvo/Mixtral-tiny",
|
| 22 |
+
"checkpoint_revision": "c557ba205ddff6ea911f4719e0d543d6c08356b6",
|
| 23 |
+
"continuous_batching": false,
|
| 24 |
+
"ep_degree": 1,
|
| 25 |
+
"fused_qkv": false,
|
| 26 |
+
"glu_mlp": true,
|
| 27 |
+
"local_ranks_size": 2,
|
| 28 |
+
"max_batch_size": 1,
|
| 29 |
+
"max_context_length": 1024,
|
| 30 |
+
"max_topk": 256,
|
| 31 |
+
"n_active_tokens": 1024,
|
| 32 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 33 |
+
"on_device_sampling": false,
|
| 34 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 35 |
+
"output_logits": false,
|
| 36 |
+
"pp_degree": 1,
|
| 37 |
+
"sequence_length": 1024,
|
| 38 |
+
"speculation_length": 0,
|
| 39 |
+
"start_rank_id": 0,
|
| 40 |
+
"target": "trn1",
|
| 41 |
+
"torch_dtype": "float16",
|
| 42 |
+
"tp_degree": 2
|
| 43 |
+
},
|
| 44 |
+
"num_attention_heads": 32,
|
| 45 |
+
"num_experts_per_tok": 2,
|
| 46 |
+
"num_hidden_layers": 2,
|
| 47 |
+
"num_key_value_heads": 8,
|
| 48 |
+
"num_local_experts": 8,
|
| 49 |
+
"output_router_logits": false,
|
| 50 |
+
"rms_norm_eps": 1e-05,
|
| 51 |
+
"rope_theta": 10000.0,
|
| 52 |
+
"router_aux_loss_coef": 0.001,
|
| 53 |
+
"router_jitter_noise": 0.0,
|
| 54 |
+
"sliding_window": 4096,
|
| 55 |
+
"tie_word_embeddings": false,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 32000
|
| 58 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/phi3/yujiepan/phi-4-tiny-random/3211e8079a23f7cca17d.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "yujiepan/phi-4-tiny-random",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Phi3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"auto_map": {},
|
| 11 |
+
"dtype": "bfloat16",
|
| 12 |
+
"embd_pdrop": 0.0,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 16,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 32,
|
| 17 |
+
"max_position_embeddings": 16384,
|
| 18 |
+
"model_type": "phi3",
|
| 19 |
+
"neuron": {
|
| 20 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 21 |
+
"batch_size": 1,
|
| 22 |
+
"capacity_factor": null,
|
| 23 |
+
"checkpoint_id": "yujiepan/phi-4-tiny-random",
|
| 24 |
+
"checkpoint_revision": "18a9a1168dc97ac6d128f811925670c275610f5a",
|
| 25 |
+
"continuous_batching": false,
|
| 26 |
+
"ep_degree": 1,
|
| 27 |
+
"fused_qkv": true,
|
| 28 |
+
"glu_mlp": true,
|
| 29 |
+
"local_ranks_size": 2,
|
| 30 |
+
"max_batch_size": 1,
|
| 31 |
+
"max_context_length": 1024,
|
| 32 |
+
"max_topk": 256,
|
| 33 |
+
"n_active_tokens": 1024,
|
| 34 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 35 |
+
"on_device_sampling": true,
|
| 36 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 37 |
+
"output_logits": false,
|
| 38 |
+
"pp_degree": 1,
|
| 39 |
+
"sequence_length": 1024,
|
| 40 |
+
"speculation_length": 0,
|
| 41 |
+
"start_rank_id": 0,
|
| 42 |
+
"target": "trn1",
|
| 43 |
+
"torch_dtype": "bfloat16",
|
| 44 |
+
"tp_degree": 2
|
| 45 |
+
},
|
| 46 |
+
"num_attention_heads": 2,
|
| 47 |
+
"num_hidden_layers": 2,
|
| 48 |
+
"num_key_value_heads": 1,
|
| 49 |
+
"original_max_position_embeddings": 16384,
|
| 50 |
+
"partial_rotary_factor": 1.0,
|
| 51 |
+
"resid_pdrop": 0.0,
|
| 52 |
+
"rms_norm_eps": 1e-05,
|
| 53 |
+
"rope_scaling": null,
|
| 54 |
+
"rope_theta": 250000,
|
| 55 |
+
"sliding_window": null,
|
| 56 |
+
"tie_word_embeddings": false,
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 100352
|
| 59 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/Qwen/Qwen2.5-0.5B/708e8d742ec0e959ffbf.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen2.5-0.5B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen2ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 896,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 4864,
|
| 14 |
+
"layer_types": [
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention"
|
| 39 |
+
],
|
| 40 |
+
"max_position_embeddings": 32768,
|
| 41 |
+
"max_window_layers": 24,
|
| 42 |
+
"model_type": "qwen2",
|
| 43 |
+
"neuron": {
|
| 44 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 45 |
+
"batch_size": 1,
|
| 46 |
+
"capacity_factor": null,
|
| 47 |
+
"checkpoint_id": "Qwen/Qwen2.5-0.5B",
|
| 48 |
+
"checkpoint_revision": "060db6499f32faf8b98477b0a26969ef7d8b9987",
|
| 49 |
+
"continuous_batching": false,
|
| 50 |
+
"ep_degree": 1,
|
| 51 |
+
"fused_qkv": false,
|
| 52 |
+
"glu_mlp": true,
|
| 53 |
+
"local_ranks_size": 2,
|
| 54 |
+
"max_batch_size": 1,
|
| 55 |
+
"max_context_length": 4096,
|
| 56 |
+
"max_topk": 256,
|
| 57 |
+
"n_active_tokens": 4096,
|
| 58 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 59 |
+
"on_device_sampling": true,
|
| 60 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 61 |
+
"output_logits": false,
|
| 62 |
+
"pp_degree": 1,
|
| 63 |
+
"sequence_length": 4096,
|
| 64 |
+
"speculation_length": 0,
|
| 65 |
+
"start_rank_id": 0,
|
| 66 |
+
"target": "trn1",
|
| 67 |
+
"torch_dtype": "bfloat16",
|
| 68 |
+
"tp_degree": 2
|
| 69 |
+
},
|
| 70 |
+
"num_attention_heads": 14,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"num_key_value_heads": 2,
|
| 73 |
+
"rms_norm_eps": 1e-06,
|
| 74 |
+
"rope_scaling": null,
|
| 75 |
+
"rope_theta": 1000000.0,
|
| 76 |
+
"sliding_window": null,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_mrope": false,
|
| 80 |
+
"use_sliding_window": false,
|
| 81 |
+
"vocab_size": 151936
|
| 82 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/Qwen/Qwen2.5-0.5B/e2a98d277483d98cbc4a.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen2.5-0.5B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen2ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 896,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 4864,
|
| 14 |
+
"layer_types": [
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention"
|
| 39 |
+
],
|
| 40 |
+
"max_position_embeddings": 32768,
|
| 41 |
+
"max_window_layers": 24,
|
| 42 |
+
"model_type": "qwen2",
|
| 43 |
+
"neuron": {
|
| 44 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 45 |
+
"batch_size": 2,
|
| 46 |
+
"capacity_factor": null,
|
| 47 |
+
"checkpoint_id": "Qwen/Qwen2.5-0.5B",
|
| 48 |
+
"checkpoint_revision": "060db6499f32faf8b98477b0a26969ef7d8b9987",
|
| 49 |
+
"continuous_batching": true,
|
| 50 |
+
"ep_degree": 1,
|
| 51 |
+
"fused_qkv": false,
|
| 52 |
+
"glu_mlp": true,
|
| 53 |
+
"local_ranks_size": 2,
|
| 54 |
+
"max_batch_size": 2,
|
| 55 |
+
"max_context_length": 1024,
|
| 56 |
+
"max_topk": 256,
|
| 57 |
+
"n_active_tokens": 1024,
|
| 58 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 59 |
+
"on_device_sampling": false,
|
| 60 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 61 |
+
"output_logits": false,
|
| 62 |
+
"pp_degree": 1,
|
| 63 |
+
"sequence_length": 1024,
|
| 64 |
+
"speculation_length": 0,
|
| 65 |
+
"start_rank_id": 0,
|
| 66 |
+
"target": "trn1",
|
| 67 |
+
"torch_dtype": "bfloat16",
|
| 68 |
+
"tp_degree": 2
|
| 69 |
+
},
|
| 70 |
+
"num_attention_heads": 14,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"num_key_value_heads": 2,
|
| 73 |
+
"rms_norm_eps": 1e-06,
|
| 74 |
+
"rope_scaling": null,
|
| 75 |
+
"rope_theta": 1000000.0,
|
| 76 |
+
"sliding_window": null,
|
| 77 |
+
"tie_word_embeddings": true,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"use_mrope": false,
|
| 80 |
+
"use_sliding_window": false,
|
| 81 |
+
"vocab_size": 151936
|
| 82 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen2/yujiepan/qwen2.5-128k-tiny-random/55d828b008dd8415c97f.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "yujiepan/qwen2.5-128k-tiny-random",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen2ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"dtype": "bfloat16",
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 8,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 16,
|
| 14 |
+
"layer_types": [
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention"
|
| 17 |
+
],
|
| 18 |
+
"max_position_embeddings": 32768,
|
| 19 |
+
"max_window_layers": 1,
|
| 20 |
+
"model_type": "qwen2",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 23 |
+
"batch_size": 1,
|
| 24 |
+
"capacity_factor": null,
|
| 25 |
+
"checkpoint_id": "yujiepan/qwen2.5-128k-tiny-random",
|
| 26 |
+
"checkpoint_revision": "c8296d4ca3f87782876d2382fbb6481d1beb8ef0",
|
| 27 |
+
"continuous_batching": false,
|
| 28 |
+
"ep_degree": 1,
|
| 29 |
+
"fused_qkv": false,
|
| 30 |
+
"glu_mlp": true,
|
| 31 |
+
"local_ranks_size": 2,
|
| 32 |
+
"max_batch_size": 1,
|
| 33 |
+
"max_context_length": 1024,
|
| 34 |
+
"max_topk": 256,
|
| 35 |
+
"n_active_tokens": 1024,
|
| 36 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 37 |
+
"on_device_sampling": true,
|
| 38 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 39 |
+
"output_logits": false,
|
| 40 |
+
"pp_degree": 1,
|
| 41 |
+
"sequence_length": 1024,
|
| 42 |
+
"speculation_length": 0,
|
| 43 |
+
"start_rank_id": 0,
|
| 44 |
+
"target": "trn1",
|
| 45 |
+
"torch_dtype": "bfloat16",
|
| 46 |
+
"tp_degree": 2
|
| 47 |
+
},
|
| 48 |
+
"num_attention_heads": 4,
|
| 49 |
+
"num_hidden_layers": 2,
|
| 50 |
+
"num_key_value_heads": 2,
|
| 51 |
+
"rms_norm_eps": 1e-06,
|
| 52 |
+
"rope_scaling": {
|
| 53 |
+
"factor": 4.0,
|
| 54 |
+
"original_max_position_embeddings": 32768,
|
| 55 |
+
"rope_type": "yarn",
|
| 56 |
+
"type": "yarn"
|
| 57 |
+
},
|
| 58 |
+
"rope_theta": 1000000.0,
|
| 59 |
+
"sliding_window": null,
|
| 60 |
+
"tie_word_embeddings": false,
|
| 61 |
+
"use_cache": true,
|
| 62 |
+
"use_sliding_window": false,
|
| 63 |
+
"vocab_size": 152064
|
| 64 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/57719fe1acd233b9fab8.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 128,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_types": [
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention",
|
| 44 |
+
"full_attention"
|
| 45 |
+
],
|
| 46 |
+
"max_position_embeddings": 32768,
|
| 47 |
+
"max_window_layers": 28,
|
| 48 |
+
"model_type": "qwen3",
|
| 49 |
+
"neuron": {
|
| 50 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 51 |
+
"batch_size": 4,
|
| 52 |
+
"capacity_factor": null,
|
| 53 |
+
"checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 54 |
+
"checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
|
| 55 |
+
"continuous_batching": false,
|
| 56 |
+
"ep_degree": 1,
|
| 57 |
+
"fused_qkv": true,
|
| 58 |
+
"glu_mlp": true,
|
| 59 |
+
"local_ranks_size": 2,
|
| 60 |
+
"max_batch_size": 4,
|
| 61 |
+
"max_context_length": 1024,
|
| 62 |
+
"max_topk": 256,
|
| 63 |
+
"n_active_tokens": 1024,
|
| 64 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 65 |
+
"on_device_sampling": false,
|
| 66 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 67 |
+
"output_logits": false,
|
| 68 |
+
"pp_degree": 1,
|
| 69 |
+
"sequence_length": 1024,
|
| 70 |
+
"speculation_length": 0,
|
| 71 |
+
"start_rank_id": 0,
|
| 72 |
+
"target": "trn1",
|
| 73 |
+
"torch_dtype": "bfloat16",
|
| 74 |
+
"tp_degree": 2
|
| 75 |
+
},
|
| 76 |
+
"num_attention_heads": 16,
|
| 77 |
+
"num_hidden_layers": 28,
|
| 78 |
+
"num_key_value_heads": 8,
|
| 79 |
+
"rms_norm_eps": 1e-06,
|
| 80 |
+
"rope_scaling": null,
|
| 81 |
+
"rope_theta": 1000000,
|
| 82 |
+
"sliding_window": null,
|
| 83 |
+
"tie_word_embeddings": true,
|
| 84 |
+
"use_cache": true,
|
| 85 |
+
"use_sliding_window": false,
|
| 86 |
+
"vocab_size": 151669
|
| 87 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3/Qwen/Qwen3-Embedding-0.6B/f4eb8a30c219ffacbbb4.json
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"head_dim": 128,
|
| 12 |
+
"hidden_act": "silu",
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_types": [
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention",
|
| 44 |
+
"full_attention"
|
| 45 |
+
],
|
| 46 |
+
"max_position_embeddings": 32768,
|
| 47 |
+
"max_window_layers": 28,
|
| 48 |
+
"model_type": "qwen3",
|
| 49 |
+
"neuron": {
|
| 50 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 51 |
+
"batch_size": 6,
|
| 52 |
+
"capacity_factor": null,
|
| 53 |
+
"checkpoint_id": "Qwen/Qwen3-Embedding-0.6B",
|
| 54 |
+
"checkpoint_revision": "c54f2e6e80b2d7b7de06f51cec4959f6b3e03418",
|
| 55 |
+
"continuous_batching": false,
|
| 56 |
+
"ep_degree": 1,
|
| 57 |
+
"fused_qkv": true,
|
| 58 |
+
"glu_mlp": true,
|
| 59 |
+
"local_ranks_size": 2,
|
| 60 |
+
"max_batch_size": 6,
|
| 61 |
+
"max_context_length": 1024,
|
| 62 |
+
"max_topk": 256,
|
| 63 |
+
"n_active_tokens": 1024,
|
| 64 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 65 |
+
"on_device_sampling": false,
|
| 66 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 67 |
+
"output_logits": false,
|
| 68 |
+
"pp_degree": 1,
|
| 69 |
+
"sequence_length": 1024,
|
| 70 |
+
"speculation_length": 0,
|
| 71 |
+
"start_rank_id": 0,
|
| 72 |
+
"target": "trn1",
|
| 73 |
+
"torch_dtype": "bfloat16",
|
| 74 |
+
"tp_degree": 2
|
| 75 |
+
},
|
| 76 |
+
"num_attention_heads": 16,
|
| 77 |
+
"num_hidden_layers": 28,
|
| 78 |
+
"num_key_value_heads": 8,
|
| 79 |
+
"rms_norm_eps": 1e-06,
|
| 80 |
+
"rope_scaling": null,
|
| 81 |
+
"rope_theta": 1000000,
|
| 82 |
+
"sliding_window": null,
|
| 83 |
+
"tie_word_embeddings": true,
|
| 84 |
+
"use_cache": true,
|
| 85 |
+
"use_sliding_window": false,
|
| 86 |
+
"vocab_size": 151669
|
| 87 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/qwen3_moe/optimum-internal-testing/tiny-random-qwen3_moe/c7cf2a4267f41d17facc.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Qwen3MoeForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"decoder_sparse_step": 2,
|
| 11 |
+
"dtype": "float32",
|
| 12 |
+
"head_dim": 32,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 64,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 128,
|
| 17 |
+
"max_position_embeddings": 40960,
|
| 18 |
+
"max_window_layers": 1,
|
| 19 |
+
"mlp_only_layers": [],
|
| 20 |
+
"model_type": "qwen3_moe",
|
| 21 |
+
"moe_intermediate_size": 128,
|
| 22 |
+
"neuron": {
|
| 23 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 24 |
+
"batch_size": 1,
|
| 25 |
+
"capacity_factor": null,
|
| 26 |
+
"checkpoint_id": "optimum-internal-testing/tiny-random-qwen3_moe",
|
| 27 |
+
"checkpoint_revision": "e0230be2839556b44b7400a233c73c74b4abb7af",
|
| 28 |
+
"continuous_batching": false,
|
| 29 |
+
"ep_degree": 1,
|
| 30 |
+
"fused_qkv": false,
|
| 31 |
+
"glu_mlp": true,
|
| 32 |
+
"local_ranks_size": 2,
|
| 33 |
+
"max_batch_size": 1,
|
| 34 |
+
"max_context_length": 1024,
|
| 35 |
+
"max_topk": 256,
|
| 36 |
+
"n_active_tokens": 1024,
|
| 37 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 38 |
+
"on_device_sampling": true,
|
| 39 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 40 |
+
"output_logits": false,
|
| 41 |
+
"pp_degree": 1,
|
| 42 |
+
"sequence_length": 1024,
|
| 43 |
+
"speculation_length": 0,
|
| 44 |
+
"start_rank_id": 0,
|
| 45 |
+
"target": "trn1",
|
| 46 |
+
"torch_dtype": "float32",
|
| 47 |
+
"tp_degree": 2
|
| 48 |
+
},
|
| 49 |
+
"norm_topk_prob": true,
|
| 50 |
+
"num_attention_heads": 2,
|
| 51 |
+
"num_experts": 8,
|
| 52 |
+
"num_experts_per_tok": 2,
|
| 53 |
+
"num_hidden_layers": 2,
|
| 54 |
+
"num_key_value_heads": 1,
|
| 55 |
+
"output_router_logits": false,
|
| 56 |
+
"rms_norm_eps": 1e-06,
|
| 57 |
+
"rope_scaling": null,
|
| 58 |
+
"rope_theta": 1000000.0,
|
| 59 |
+
"router_aux_loss_coef": 0.001,
|
| 60 |
+
"sliding_window": null,
|
| 61 |
+
"tie_word_embeddings": true,
|
| 62 |
+
"use_cache": true,
|
| 63 |
+
"use_sliding_window": false,
|
| 64 |
+
"vocab_size": 151936
|
| 65 |
+
}
|
neuronxcc-2.21.33363.0+82129205/0_REGISTRY/0.4.3.dev2/smollm3/HuggingFaceTB/SmolLM3-3B/7a707f573aa21c74857f.json
ADDED
|
@@ -0,0 +1,134 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_entry_class": "SingleModelCacheEntry",
|
| 3 |
+
"_model_id": "HuggingFaceTB/SmolLM3-3B",
|
| 4 |
+
"_task": "text-generation",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"SmolLM3ForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"attention_bias": false,
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"dtype": "bfloat16",
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 11008,
|
| 15 |
+
"layer_types": [
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention",
|
| 42 |
+
"full_attention",
|
| 43 |
+
"full_attention",
|
| 44 |
+
"full_attention",
|
| 45 |
+
"full_attention",
|
| 46 |
+
"full_attention",
|
| 47 |
+
"full_attention",
|
| 48 |
+
"full_attention",
|
| 49 |
+
"full_attention",
|
| 50 |
+
"full_attention",
|
| 51 |
+
"full_attention"
|
| 52 |
+
],
|
| 53 |
+
"max_position_embeddings": 65536,
|
| 54 |
+
"max_window_layers": 28,
|
| 55 |
+
"mlp_bias": false,
|
| 56 |
+
"model_type": "smollm3",
|
| 57 |
+
"neuron": {
|
| 58 |
+
"_serialized_key": "NxDNeuronConfig",
|
| 59 |
+
"batch_size": 1,
|
| 60 |
+
"capacity_factor": null,
|
| 61 |
+
"checkpoint_id": "HuggingFaceTB/SmolLM3-3B",
|
| 62 |
+
"checkpoint_revision": "a07cc9a04f16550a088caea529712d1d335b0ac1",
|
| 63 |
+
"continuous_batching": false,
|
| 64 |
+
"ep_degree": 1,
|
| 65 |
+
"fused_qkv": true,
|
| 66 |
+
"glu_mlp": true,
|
| 67 |
+
"local_ranks_size": 2,
|
| 68 |
+
"max_batch_size": 1,
|
| 69 |
+
"max_context_length": 1024,
|
| 70 |
+
"max_topk": 256,
|
| 71 |
+
"n_active_tokens": 1024,
|
| 72 |
+
"neuronxcc_version": "2.21.33363.0+82129205",
|
| 73 |
+
"on_device_sampling": true,
|
| 74 |
+
"optimum_neuron_version": "0.4.3.dev2",
|
| 75 |
+
"output_logits": false,
|
| 76 |
+
"pp_degree": 1,
|
| 77 |
+
"sequence_length": 1024,
|
| 78 |
+
"speculation_length": 0,
|
| 79 |
+
"start_rank_id": 0,
|
| 80 |
+
"target": "trn1",
|
| 81 |
+
"torch_dtype": "bfloat16",
|
| 82 |
+
"tp_degree": 2
|
| 83 |
+
},
|
| 84 |
+
"no_rope_layer_interval": 4,
|
| 85 |
+
"no_rope_layers": [
|
| 86 |
+
1,
|
| 87 |
+
1,
|
| 88 |
+
1,
|
| 89 |
+
0,
|
| 90 |
+
1,
|
| 91 |
+
1,
|
| 92 |
+
1,
|
| 93 |
+
0,
|
| 94 |
+
1,
|
| 95 |
+
1,
|
| 96 |
+
1,
|
| 97 |
+
0,
|
| 98 |
+
1,
|
| 99 |
+
1,
|
| 100 |
+
1,
|
| 101 |
+
0,
|
| 102 |
+
1,
|
| 103 |
+
1,
|
| 104 |
+
1,
|
| 105 |
+
0,
|
| 106 |
+
1,
|
| 107 |
+
1,
|
| 108 |
+
1,
|
| 109 |
+
0,
|
| 110 |
+
1,
|
| 111 |
+
1,
|
| 112 |
+
1,
|
| 113 |
+
0,
|
| 114 |
+
1,
|
| 115 |
+
1,
|
| 116 |
+
1,
|
| 117 |
+
0,
|
| 118 |
+
1,
|
| 119 |
+
1,
|
| 120 |
+
1,
|
| 121 |
+
0
|
| 122 |
+
],
|
| 123 |
+
"num_attention_heads": 16,
|
| 124 |
+
"num_hidden_layers": 36,
|
| 125 |
+
"num_key_value_heads": 4,
|
| 126 |
+
"pretraining_tp": 2,
|
| 127 |
+
"rms_norm_eps": 1e-06,
|
| 128 |
+
"rope_scaling": null,
|
| 129 |
+
"rope_theta": 5000000.0,
|
| 130 |
+
"sliding_window": null,
|
| 131 |
+
"use_cache": false,
|
| 132 |
+
"use_sliding_window": false,
|
| 133 |
+
"vocab_size": 128256
|
| 134 |
+
}
|
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=2 --vectorize-strided-dma ", "-O2", "--lnc=1", "--logfile=/tmp/nxd_model/token_generation/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f0c908a5ddbf03970315b73168cfd6d3b14f39328f2b700c5d1e9c4144323034
|
| 3 |
+
size 91147
|
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9022b1945dd2e7ef8b45eadbd2dc20dfb74b1add40ec7aad07e57bad2a500b25
|
| 3 |
+
size 277504
|
neuronxcc-2.21.33363.0+82129205/MODULE_0cdc152aff6a7cbf36f6+a02c3a36/wrapped_neff.hlo
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95021fb9ec8e64ac096a6e7f993b7f066b620b290e2aa82be5be7695581db3f0
|
| 3 |
+
size 289031
|
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1", "--enable-saturate-infinity", "--auto-cast=none", "--model-type=transformer", "-O1", "--logfile=/tmp/nxdi_test_373770ad-59f8-4200-be26-fe3a56e83168/compiler_workdir/DecoderLayerWrapper/_tp0_bk0/log-neuron-cc.txt", "--enable-internal-neff-wrapper"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:76dc525043c8d0fc6b46382b17659e6a2f2e981788c635021a8a61aa832d2866
|
| 3 |
+
size 14480
|
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/model.neff
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:816f3156b088f29ab7d5de607261ccb476e608025bee91f0c33a986b785ca1b6
|
| 3 |
+
size 267264
|
neuronxcc-2.21.33363.0+82129205/MODULE_1015ebb8c8fe4c3252c9+3287a374/wrapped_neff.hlo
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b1dc8581fcc1d833e79312ef56cd444de96a45e0adda2e969d09f0a90d9fcd3
|
| 3 |
+
size 272962
|
neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68d800848ea22ac4d6f5cf0e1e6088da78460e16b5340dd0f956b5b5bf0c1834
|
| 3 |
+
size 1122
|
neuronxcc-2.21.33363.0+82129205/MODULE_10217061096959125489+e30acd3a/model.neff
ADDED
|
Binary file (31.7 kB). View file
|
|
|
neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0113611968c8ed0cbeaea5ff3bba2ce4eecd5fc989e2592c6b8ac5fdf09c91c2
|
| 3 |
+
size 1562
|
neuronxcc-2.21.33363.0+82129205/MODULE_10244305442015770634+e30acd3a/model.neff
ADDED
|
Binary file (31.7 kB). View file
|
|
|
neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a35335d3e96ee86956cdf93862bf42ae6f83017e9b331b77531ad6942a62366d
|
| 3 |
+
size 1269
|
neuronxcc-2.21.33363.0+82129205/MODULE_10606948783918825529+e30acd3a/model.neff
ADDED
|
Binary file (31.7 kB). View file
|
|
|
neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:153fdddade7e1b284f8d78df84523bd332eccc7ddfc7a0698ad8653d560c55ca
|
| 3 |
+
size 1124
|
neuronxcc-2.21.33363.0+82129205/MODULE_10645643398657092095+e30acd3a/model.neff
ADDED
|
Binary file (31.7 kB). View file
|
|
|
neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7434cf1ede5dc6759a8ad5cdc472fb8e1668b1a8bd8dd7e7e08b5c09401b8065
|
| 3 |
+
size 1562
|
neuronxcc-2.21.33363.0+82129205/MODULE_10661660426924300837+e30acd3a/model.neff
ADDED
|
Binary file (31.7 kB). View file
|
|
|
neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/compile_flags.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
["--target=trn1"]
|
neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/model.done
ADDED
|
File without changes
|
neuronxcc-2.21.33363.0+82129205/MODULE_10746122569655005679+e30acd3a/model.hlo_module.pb
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7000c9195e6c3c9abfcbac09928468ef4c204488094a046ec7abfd1b7ea8657
|
| 3 |
+
size 1563
|