Instructions to use EleutherAI/pythia-70m-deduped with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use EleutherAI/pythia-70m-deduped with Transformers:

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

pipe = pipeline("text-generation", model="EleutherAI/pythia-70m-deduped")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pythia-70m-deduped")
model = AutoModelForCausalLM.from_pretrained("EleutherAI/pythia-70m-deduped")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use EleutherAI/pythia-70m-deduped with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "EleutherAI/pythia-70m-deduped"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EleutherAI/pythia-70m-deduped",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/EleutherAI/pythia-70m-deduped

SGLang

How to use EleutherAI/pythia-70m-deduped 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 "EleutherAI/pythia-70m-deduped" \
    --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": "EleutherAI/pythia-70m-deduped",
		"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 "EleutherAI/pythia-70m-deduped" \
        --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": "EleutherAI/pythia-70m-deduped",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use EleutherAI/pythia-70m-deduped with Docker Model Runner:
```
docker model run hf.co/EleutherAI/pythia-70m-deduped
```

Model init does not correspond to paper's (and pre-trained weights') initialization scheme

by jacobpfau - opened Jun 7, 2023

Discussion

jacobpfau

Jun 7, 2023

•

edited Jun 7, 2023

The Eleuther trained models seem to rely on initialization schemes defined here. Meanwhile, the HF version does initialization here -- these are very different. In practice, the Eleuther initialization scheme seems far better on my data. It would be great if the HF version could be updated to correspond to the model's intended initialization scheme.

The significance of this difference can be seen by comparing model performance between:
model = GPTNeoXForCausalLM(config).to("cuda")
and
model = GPTNeoXForCausalLM.from_pretrained(
'EleutherAI/pythia-70m',
revision="step0",
).to("cuda")

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment