This is the gptq 4bit quantization of this model: https://huggingface.co/dvruette/llama-13b-pretrained-sft-do2
This quantization was made by using this repository: https://github.com/qwopqwop200/GPTQ-for-LLaMa/tree/triton
And I used the triton branch with all the gptq implementations available (true_sequential + act_order + groupsize 128)
CUDA_VISIBLE_DEVICES=0 python llama.py ./llama-13b-pretrained-sft-do2-4bit-128g-TRITON c4 --wbits 4 --true-sequential --act-order --groupsize 128 --save_safetensors llama-13b-pretrained-sft-do2-4bit-128g-TRITON.safetensors
To use the triton model on oobabooga's webui, you can refer to this post to get rid of all the errors you can encounter: https://github.com/oobabooga/text-generation-webui/issues/734
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.