Instructions to use jacobthebanana/galactica-30b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jacobthebanana/galactica-30b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="jacobthebanana/galactica-30b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jacobthebanana/galactica-30b")
model = AutoModelForCausalLM.from_pretrained("jacobthebanana/galactica-30b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use jacobthebanana/galactica-30b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jacobthebanana/galactica-30b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jacobthebanana/galactica-30b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/jacobthebanana/galactica-30b

SGLang

How to use jacobthebanana/galactica-30b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "jacobthebanana/galactica-30b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jacobthebanana/galactica-30b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "jacobthebanana/galactica-30b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jacobthebanana/galactica-30b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use jacobthebanana/galactica-30b with Docker Model Runner:
```
docker model run hf.co/jacobthebanana/galactica-30b
```

JAX weights converted from Torch checkpoint at facebook/galactica-30b.

(env) ubuntu@vm:~$ JAX_PLATFORM_NAME=cpu python3
>>> import jax
>>> print(jax.devices())   
[CpuDevice(id=0)]  # Ensure that model weights are loaded into CPU RAM, not accelerator memory.
>>> from transformers import FlaxOPTForCausalLM
>>> model = FlaxOPTForCausalLM.from_pretrained("facebook/galactica-30b", from_pt=True)
>>> model.push_to_hub(hf_model_repo)

Citation and Attribution

Citation from the original repo is reproduced below as per the cc-by-nc-4.0 licsense.

@inproceedings{GALACTICA,
    title={GALACTICA: A Large Language Model for Science},
    author={Ross Taylor and Marcin Kardas and Guillem Cucurull and Thomas Scialom and Anthony Hartshorn and Elvis Saravia and Andrew Poulton and Viktor Kerkez and Robert Stojnic},
    year={2022}
}

Research supported with Cloud TPUs from Google's TPU Research Cloud (TRC)

Downloads last month: 11