antoinelouis_-_belgpt2-4bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

belgpt2 - bnb 4bits

• Model creator: https://huggingface.co/antoinelouis/
• Original model: https://huggingface.co/antoinelouis/belgpt2/

Original model description:

language:

• fr license:
• mit widget:
• text: Hier, Elon Musk a
• text: Pourquoi a-t-il
• text: Tout à coup, elle metrics:
• perplexity library_name: transformers pipeline_tag: text-generation

---

BelGPT-2

The 1st GPT-2 model pre-trained on a very large and heterogeneous French corpus (~60Gb).

Usage

You can use BelGPT-2 with 🤗 transformers:

import torch
from transformers import GPT2Tokenizer, GPT2LMHeadModel

# Load pretrained model and tokenizer
model = GPT2LMHeadModel.from_pretrained("antoiloui/belgpt2")
tokenizer = GPT2Tokenizer.from_pretrained("antoiloui/belgpt2")

# Generate a sample of text
model.eval()
output = model.generate(
            bos_token_id=random.randint(1,50000),
            do_sample=True,   
            top_k=50, 
            max_length=100,
            top_p=0.95, 
            num_return_sequences=1
)

# Decode it
decoded_output = []
for sample in output:
    decoded_output.append(tokenizer.decode(sample, skip_special_tokens=True))
print(decoded_output)

Data

Below is the list of all French copora used to pre-trained the model:

Documentation

Detailed documentation on the pre-trained model, its implementation, and the data can be found here.

Citation

For attribution in academic contexts, please cite this work as:

@misc{louis2020belgpt2,
  author = {Louis, Antoine},
  title = {{BelGPT-2: A GPT-2 Model Pre-trained on French Corpora}},
  year = {2020},
  howpublished = {\url{https://github.com/ant-louis/belgpt2}},
}