alpaca_convert/README.md

My personal repo to convert models from Lora to huggingface/ggml/gptq 4bit so I can run them in normal text-webui and llama.cpp

How do we do this?

1. lora -> hf
   • tloen/alpaca-lora/export_hf_checkpoint.py
2. hf -> 4bit
   • using GPTQ-for-LLaMa/llama.py CUDA_VISIBLE_DEVICES=0 python llama.py ./llama-hf/llama-7b c4 --wbits 4 --true-sequential --act-order --groupsize 128 --save llama7b-4bit-128g.pt
3. 4bit -> ggml
   • llama.cpp/convert-pth-to-ggml.py

TODO

• lora -> hf
  • test this
• hf -> 4bit
• 4bit to -> ggml
• test perplexity on llama and alpaca type prompts too! maybe use eluther evals

setup env

conda create -n textgen3 python=3.10.9
conda activate textgen3
mamba install pytorch torchvision torchaudio pytorch-cuda=11.7 cudatoolkit-dev==11.7  cudatoolkit=11.7 -c pytorch -c nvidia  -c conda-forge 
pip install -r requirements.txt
pip install -e .

download models

huggingface-cli login

# download base models
python scripts/download-model.py decapoda-research/llama-7b-hf
# python scripts/download-model.py decapoda-research/llama-13b-hf
# python scripts/download-model.py decapoda-research/llama-30b-hf

# download loras
python scripts/download-model.py tloen/alpaca-lora-7b
# python scripts/download-model.py chansung/alpaca-lora-13b
# python scripts/download-model.py chansung/alpaca-lora-30b

convert models

# convert
python scripts/export_hf_checkpoint.py ./data/models/llama-7b-hf -l ./data/loras/tloen_alpaca-lora-7b
# test
python scripts/test_01_delora.py models/tloen_alpaca-lora-7b-delorified

Links

• https://github.com/s4rduk4r/alpaca_lora_4bit_readme/blob/main/README.md