Davide Macario

Generation error on MPS (Torch >= 2.2.0, MacOS 14.4)

8

When running `train_gpt2.py`, I get all 16 output tokens equal to "!" (token 0). Here is the complete output: ```text ❯ python3 train_gpt2.py using device: mps loading weights from pretrained...

Citing this project in research

4

Since I believe this project is being used a lot in academia/research, especially as a way to explain and showcase LLMs and the decoder-only Transformer architecture, I think it needs...

Torch >= 2.2.0 inference issues on MPS

3

When running ```python python sample.py --init_from=gpt2 --num_samples=2 --max_new_tokens=100 ``` having set `device = 'mps'` on my M1 Pro MacBook (MacOS 14.4), with Torch 2.2.1 and 2.2.0, I get this output:...

The file `/litgpt/prompts.py` was missing the mapping between the `llama3` model name and the `Llama3` PromptStyle class in the variable `prompt_styles`.