Instructions to use petals-team/StableBeluga2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use petals-team/StableBeluga2 with Transformers:

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

pipe = pipeline("text-generation", model="petals-team/StableBeluga2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("petals-team/StableBeluga2")
model = AutoModelForCausalLM.from_pretrained("petals-team/StableBeluga2")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use petals-team/StableBeluga2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "petals-team/StableBeluga2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "petals-team/StableBeluga2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/petals-team/StableBeluga2

SGLang

How to use petals-team/StableBeluga2 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 "petals-team/StableBeluga2" \
    --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": "petals-team/StableBeluga2",
		"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 "petals-team/StableBeluga2" \
        --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": "petals-team/StableBeluga2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use petals-team/StableBeluga2 with Docker Model Runner:
```
docker model run hf.co/petals-team/StableBeluga2
```

borzunov commited on Aug 23, 2023

Commit

0cc4f70

1 Parent(s): 69185c0

Explain changes and their meaning for Petals

Browse files

Files changed (1) hide show

README.md +7 -3

README.md CHANGED Viewed

@@ -14,9 +14,13 @@ pipeline_tag: text-generation
 This repository contains the model from the [stabilityai/StableBeluga2](https://huggingface.co/stabilityai/StableBeluga2) repository with the following changes:
-- Storing weights in `bfloat16` instead of `float32` (2x smaller files)
-- Storing weights in small shards (< 1 GiB each)
-- Added a [Safetensors](https://github.com/huggingface/safetensors) version
 We provide the original README below. Please refer there for model details and licensing information.

 This repository contains the model from the [stabilityai/StableBeluga2](https://huggingface.co/stabilityai/StableBeluga2) repository with the following changes:
+1. **Storing weights in `bfloat16` instead of `float32`.**
+    This leads to 2x smaller files and a small quality loss, which is not significant compared to the loss caused by NF4 quantization used in Petals by default.
+1. **Storing weights in small shards.**
+    Each transformer block is stored in its own shard (1.71 GB each). The input and output embeddings and adjacent layernorms are in a separate shard (1.05 GB) too.
+    This way, Petals clients and servers don't have to download any excess data besides the layers they actually use.
+1. **Using [Safetensors](https://github.com/huggingface/safetensors) instead of Pickle.**
+    This allows faster loading with smaller RAM requirements.
 We provide the original README below. Please refer there for model details and licensing information.