Instructions to use himanshu-dutta/pycoder-gpt2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use himanshu-dutta/pycoder-gpt2 with Transformers:

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

pipe = pipeline("text-generation", model="himanshu-dutta/pycoder-gpt2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("himanshu-dutta/pycoder-gpt2")
model = AutoModelForCausalLM.from_pretrained("himanshu-dutta/pycoder-gpt2")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use himanshu-dutta/pycoder-gpt2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "himanshu-dutta/pycoder-gpt2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "himanshu-dutta/pycoder-gpt2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/himanshu-dutta/pycoder-gpt2

SGLang

How to use himanshu-dutta/pycoder-gpt2 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 "himanshu-dutta/pycoder-gpt2" \
    --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": "himanshu-dutta/pycoder-gpt2",
		"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 "himanshu-dutta/pycoder-gpt2" \
        --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": "himanshu-dutta/pycoder-gpt2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use himanshu-dutta/pycoder-gpt2 with Docker Model Runner:
```
docker model run hf.co/himanshu-dutta/pycoder-gpt2
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

PyCoder is a tool to generate python code out of a few given topics and a description. It uses GPT-2 language model as its engine. Pycoder poses writing Python code as a conditional-Causal Language Modelling(c-CLM). It has been trained on millions of lines of Python code written by all of us. At the current stage and state of training, it produces sensible code with few lines of description, but the scope of improvement for the model is limitless.

Pycoder has been developed as a Command-Line tool (CLI), an API endpoint, as well as a python package (yet to be deployed to PyPI). This repository acts as a framework for anyone who either wants to try to build Pycoder from scratch or turn Pycoder into maybe a CPPCoder or JSCoder 😃. A blog post about the development of the project will be released soon.

To use Pycoder as a CLI utility, clone the repository as normal, and install the package with:

foo@bar:❯ pip install pycoder

After this the package could be verified and accessed as either a native CLI tool or a python package with:

foo@bar:❯ python -m pycoder --version

Or directly as:

foo@bar:❯ pycoder --version

On installation the CLI can be used directly, such as:

foo@bar:❯ pycoder -t pytorch -t torch -d "a trainer class to train vision model" -ml 120

The API endpoint is deployed using FastAPI. Once all the requirements have been installed for the project, the API can be accessed with:

foo@bar:❯ pycoder --endpoint PORT_NUMBER

Or

foo@bar:❯ pycoder -e PORT_NUMBER

Tech Stack

Tested Platforms

BibTeX

If you want to cite the framework feel free to use this:

@article{dutta2021pycoder,
  title={Pycoder},
  author={Dutta, H},
  journal={GitHub. Note: https://github.com/himanshu-dutta/pycoder},
  year={2021}
}

Downloads last month: 4