huggingface / huggingface-llama-recipes

huggingface/huggingface-llama-recipes Issues

How to run LLama8b/70b using FP8
Updated 9 days ago1
Llama-Vision FT Error: The number of images in each batch [1, 1] should be the same [1, 1] should be the same
Updated 13 days ago3
[BUG] missing `quantization_config` in fp8-405B.ipynb
Updated 13 days ago4
Unable to deep copy past_key_values
Updated 13 days ago3
Trying to run torch_compile_with_torchao cannot import name 'TorchAoConfig' from 'transformers'
Updated a month ago
quantization_config in peft finetune script is not use anywhere
Updated a month ago
Fail to access Meta-Llama-3.1-405B-Instruct-FP8 as a Pro huggingface user
Updated a month ago
Quantized Llama model for Question Answering
Updated 2 months ago
Running Llama 3.1 405B H100 Multi-Node
Updated 2 months ago
ValueError: `rope_scaling` must be a dictionary with two fields
Closed 2 months ago2
Running 4-bit on the fly quantization throws an error
Updated 2 months ago1
Llama 3.1 405B Cannot Run on 8*A100
Updated 2 months ago1
Missing dependency - accelerate
Closed 3 months ago