Instructions to use Smilyai-labs-beta-testers/Smilyai-3rd-gen with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Smilyai-labs-beta-testers/Smilyai-3rd-gen with Transformers:

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

pipe = pipeline("text-generation", model="Smilyai-labs-beta-testers/Smilyai-3rd-gen")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Smilyai-labs-beta-testers/Smilyai-3rd-gen", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Smilyai-labs-beta-testers/Smilyai-3rd-gen with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Smilyai-labs-beta-testers/Smilyai-3rd-gen"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Smilyai-labs-beta-testers/Smilyai-3rd-gen",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Smilyai-labs-beta-testers/Smilyai-3rd-gen

SGLang

How to use Smilyai-labs-beta-testers/Smilyai-3rd-gen 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 "Smilyai-labs-beta-testers/Smilyai-3rd-gen" \
    --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": "Smilyai-labs-beta-testers/Smilyai-3rd-gen",
		"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 "Smilyai-labs-beta-testers/Smilyai-3rd-gen" \
        --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": "Smilyai-labs-beta-testers/Smilyai-3rd-gen",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Smilyai-labs-beta-testers/Smilyai-3rd-gen with Docker Model Runner:
```
docker model run hf.co/Smilyai-labs-beta-testers/Smilyai-3rd-gen
```

Do you guys have any small model tips?

by Bc-AI - opened Aug 29

Discussion

Bc-AI

Smilyai-labs-beta-testers org Aug 29

@Banaxi-Tech @vovaRL Do you guys have any tips for making good small models?

Banaxi-Tech

Smilyai-labs-beta-testers org Aug 29

@Banaxi-Tech @vovaRL Do you guys have any tips for making good small models?

These are my tips:

For data use:
https://huggingface.co/datasets/HuggingFaceFW/finephrase
https://huggingface.co/datasets/epfml/FineWeb-HQ
https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu
cosmopedia-v2 of https://huggingface.co/datasets/HuggingFaceTB/smollm-corpus
https://huggingface.co/datasets/mlfoundations/dclm-baseline-1.0

https://huggingface.co/datasets/AxiomicLabs/NPset-2-Python-Edu*
For SFT or DPO:
https://huggingface.co/datasets/HuggingFaceTB/smol-smoltalk

If you want a good coding score use
https://huggingface.co/datasets/AxiomicLabs/NPset-2-Python-Edu but then it will just get better at logprob not actual code.

If you want actual code not some pseudocode use https://huggingface.co/datasets/HuggingFaceCode/stack-v3-train

And if your model is bigger than 200M add
https://huggingface.co/datasets/HuggingFaceFW/finepdfs-edu

Datdanboi25

Smilyai-labs-beta-testers org Aug 30

Yeah this is a good setup

smilyai-large-team

Smilyai-labs-beta-testers org Aug 31

nice

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