Instructions to use bigscience/bigscience-small-testing with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bigscience/bigscience-small-testing with Transformers:

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

pipe = pipeline("text-generation", model="bigscience/bigscience-small-testing")

# Load model directly
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("bigscience/bigscience-small-testing")
model = AutoModel.from_pretrained("bigscience/bigscience-small-testing")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use bigscience/bigscience-small-testing with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/bigscience/bigscience-small-testing

SGLang

How to use bigscience/bigscience-small-testing 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 "bigscience/bigscience-small-testing" \
    --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": "bigscience/bigscience-small-testing",
		"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 "bigscience/bigscience-small-testing" \
        --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": "bigscience/bigscience-small-testing",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use bigscience/bigscience-small-testing with Docker Model Runner:
```
docker model run hf.co/bigscience/bigscience-small-testing
```

bigscience-small-testing

File size: 711 Bytes

{
  "_name_or_path": "/home/younes/Desktop/Work/data/megatron-debug/",
  "apply_residual_connection_post_layernorm": false,
  "architectures": [
    "BloomModel"
  ],
  "attention_dropout": 0.1,
  "bias_dropout_fusion": true,
  "bos_token_id": 0,
  "dtype": "bfloat16",
  "eos_token_id": 0,
  "hidden_dropout": 0.1,
  "hidden_size": 64,
  "initializer_range": 0.02,
  "layer_norm_epsilon": 1e-05,
  "masked_softmax_fusion": true,
  "model_type": "bloom",
  "n_head": 8,
  "n_inner": null,
  "n_layer": 2,
  "pretraining_tp": 2,
  "seq_length": 20,
  "skip_bias_add": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.18.0",
  "use_cache": false,
  "vocab_size": 250880,
  "slow_but_exact": true
}