Instructions to use roneneldan/TinyStories-Instruct-33M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use roneneldan/TinyStories-Instruct-33M with Transformers:

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

pipe = pipeline("text-generation", model="roneneldan/TinyStories-Instruct-33M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("roneneldan/TinyStories-Instruct-33M")
model = AutoModelForCausalLM.from_pretrained("roneneldan/TinyStories-Instruct-33M", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use roneneldan/TinyStories-Instruct-33M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "roneneldan/TinyStories-Instruct-33M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "roneneldan/TinyStories-Instruct-33M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/roneneldan/TinyStories-Instruct-33M

SGLang

How to use roneneldan/TinyStories-Instruct-33M 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 "roneneldan/TinyStories-Instruct-33M" \
    --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": "roneneldan/TinyStories-Instruct-33M",
		"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 "roneneldan/TinyStories-Instruct-33M" \
        --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": "roneneldan/TinyStories-Instruct-33M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use roneneldan/TinyStories-Instruct-33M with Docker Model Runner:
```
docker model run hf.co/roneneldan/TinyStories-Instruct-33M
```

I fine-tuned this model on "tool dataset"

by nikitastaf1996 - opened May 16, 2023

Discussion

nikitastaf1996

May 16, 2023

As experiment I decided to fine-tune this model on medium-size-generated-tasks dataset.
The goal was to follow ReAct Langchain agent format. While using python_repl tool.
In my experience you need 13b or 30b model to do that.
It successfully follows the format and even tries to write some shitty code.
Given model size and time to fine-tune it's success.
Link:https://huggingface.co/nikitastaf1996/TinyStories-Instruct-33M-react-medium-tasks-dirty

nikitastaf1996 changed discussion title from I fine-tuned this on "tool dataset" to I fine-tuned this model on "tool dataset" May 16, 2023

nikitastaf1996 changed discussion status to closed May 16, 2023

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