Instructions to use nilq/baby-python-mistral-1L-tiny-TinyStories-ft with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nilq/baby-python-mistral-1L-tiny-TinyStories-ft with Transformers:

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

pipe = pipeline("text-generation", model="nilq/baby-python-mistral-1L-tiny-TinyStories-ft")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("nilq/baby-python-mistral-1L-tiny-TinyStories-ft")
model = AutoModelForCausalLM.from_pretrained("nilq/baby-python-mistral-1L-tiny-TinyStories-ft", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nilq/baby-python-mistral-1L-tiny-TinyStories-ft with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nilq/baby-python-mistral-1L-tiny-TinyStories-ft"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nilq/baby-python-mistral-1L-tiny-TinyStories-ft",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/nilq/baby-python-mistral-1L-tiny-TinyStories-ft

SGLang

How to use nilq/baby-python-mistral-1L-tiny-TinyStories-ft 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 "nilq/baby-python-mistral-1L-tiny-TinyStories-ft" \
    --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": "nilq/baby-python-mistral-1L-tiny-TinyStories-ft",
		"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 "nilq/baby-python-mistral-1L-tiny-TinyStories-ft" \
        --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": "nilq/baby-python-mistral-1L-tiny-TinyStories-ft",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use nilq/baby-python-mistral-1L-tiny-TinyStories-ft with Docker Model Runner:
```
docker model run hf.co/nilq/baby-python-mistral-1L-tiny-TinyStories-ft
```

baby-python-mistral-1L-tiny-TinyStories-ft / README.md

nilq

Update README.md

41d2718 verified almost 2 years ago

preview code

Raw

History Blame Contribute Delete

1.74 kB

metadata

base_model: nilq/baby-python-mistral-1L-tiny-base
tags:
  - generated_from_trainer
datasets:
  - roneneldan/TinyStories
metrics:
  - accuracy
model-index:
  - name: baby-python-mistral-1L-tiny-TinyStories-ft
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: roneneldan/TinyStories
          type: roneneldan/TinyStories
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5196418943843818

baby-python-mistral-1L-tiny-TinyStories-ft

This model is a fine-tuned version of nilq/baby-python-mistral-1L-tiny-base on the roneneldan/TinyStories dataset. This is the TinyStories model in the paper Tracking Universal Features Through Fine-Tuning and Model Merging. It achieves the following results on the evaluation set:

Loss: 2.0442
Accuracy: 0.5196

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 64
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 500
num_epochs: 1.0

Training results

Framework versions

Transformers 4.38.1
Pytorch 2.2.0+cu121
Datasets 2.17.1
Tokenizers 0.15.2