My Fine-tuned GPT-2 Medium

This repository contains a GPT-2 Medium model fine-tuned on custom instruction data using PyTorch and Hugging Face Transformers.

Model Details

• Base model: GPT-2 Medium (355M)
• Architecture: Standard GPT-2 (24 layers, 1024 embedding dim, 16 heads)
• Tokenizer: Uses the default GPT-2 tokenizer
• Format: Hugging Face Transformers-compatible (model.safetensors, config.json)

Usage

You can load and use this model directly with Hugging Face Transformers:

from transformers import GPT2LMHeadModel, GPT2Tokenizer

model = GPT2LMHeadModel.from_pretrained("sweatSmile/my-finetuned-gpt2-medium")
tokenizer = GPT2Tokenizer.from_pretrained("gpt2-medium")

prompt = "Below is an instruction that describes a task.\n\n### Instruction:\nWrite a poem about spring."
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=50)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Files

• model.safetensors: Model weights
• config.json: Model configuration
• generation_config.json: Generation parameters

License

This model is released under the Apache 2.0 license.

Author

• sweatSmile

---

For questions or issues, please open an issue on the Hugging Face repo.