Instructions to use Guspard-ew/BeanSLM-Instruct-278M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Guspard-ew/BeanSLM-Instruct-278M with Transformers:

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

pipe = pipeline("text-generation", model="Guspard-ew/BeanSLM-Instruct-278M")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Guspard-ew/BeanSLM-Instruct-278M", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Guspard-ew/BeanSLM-Instruct-278M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Guspard-ew/BeanSLM-Instruct-278M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Guspard-ew/BeanSLM-Instruct-278M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Guspard-ew/BeanSLM-Instruct-278M

SGLang

How to use Guspard-ew/BeanSLM-Instruct-278M 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 "Guspard-ew/BeanSLM-Instruct-278M" \
    --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": "Guspard-ew/BeanSLM-Instruct-278M",
		"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 "Guspard-ew/BeanSLM-Instruct-278M" \
        --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": "Guspard-ew/BeanSLM-Instruct-278M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Guspard-ew/BeanSLM-Instruct-278M with Docker Model Runner:
```
docker model run hf.co/Guspard-ew/BeanSLM-Instruct-278M
```

Imrpovement Ideas

by Guspard-ew - opened 4 days ago

Discussion

Guspard-ew

Owner 4 days ago

•

edited 4 days ago

Making this model from scratch has been a long journey for me, starting at 20m up to 278m with many versions... This is the first version of its size and i will make other improved BeanSLM-Instruct-278M models with more pretraining and maybe one day more seqlen (256 rn i dont have compute), my pretraining loss did get stuck at 3.3 so i used that.... but yeah ill try to improve that (I used finewebedu) yeah... if any of you have any tips please reach out here also please dont hate on my model ( Im a kid and yeah my ais probably arent perfect)

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