| GPTQ 4bit quantization of: https://huggingface.co/chavinlo/gpt4-x-alpaca | |
| Command: | |
| CUDA_VISIBLE_DEVICES=0 python llama.py ./models/chavinlo-gpt4-x-alpaca | |
| --wbits 4 | |
| --true-sequential | |
| --act-order | |
| --groupsize 128 | |
| --save gpt-x-alpaca-13b-native-4bit-128g.pt | |