qwopqwop200/GPTQ-for-LLaMa issues and pull requests

#25 - running build_exit error

Issue - State: closed - Opened by BenjaminHei over 1 year ago - 1 comment

#24 - More VRAM Efficient Attention

Issue - State: closed - Opened by MarkSchmidty over 1 year ago - 2 comments

#23 - Add benchmark results for 3 and 4 bit 33B

Pull Request - State: closed - Opened by ItsLogic over 1 year ago - 1 comment

#22 - Issue compiling in docker - No CUDA runtime is found

Issue - State: closed - Opened by TheTerrasque over 1 year ago - 5 comments

#21 - NameError: name 'quant_cuda' is not defined

Issue - State: closed - Opened by CyberTimon over 1 year ago - 2 comments

#20 - Add `.safetensors` support

Pull Request - State: closed - Opened by ghost over 1 year ago - 4 comments

#19 - Is it possible to reuse the GPTQ implementation to quant oasst-sft-1-pythia-12b into 4b?

Issue - State: closed - Opened by npk48 over 1 year ago - 2 comments

#18 - Tokenizer class LLaMATokenizer does not exist or is not currently imported.

Issue - State: closed - Opened by C0rn3j over 1 year ago - 12 comments

#17 - FP8 Quantization?

Issue - State: closed - Opened by philipturner over 1 year ago - 2 comments

#16 - Questions about group size

Issue - State: closed - Opened by DanielWe2 over 1 year ago - 7 comments

#15 - Are these errors expected ?

Issue - State: closed - Opened by USBhost over 1 year ago - 3 comments

#14 - Saving checkpoints?

Issue - State: closed - Opened by elephantpanda over 1 year ago - 1 comment

#13 - Model Quantization Instructions

Issue - State: closed - Opened by MarkSchmidty over 1 year ago - 3 comments

#12 - state_dict error on model load

Issue - State: closed - Opened by GamerUntouch over 1 year ago - 3 comments

#11 - Multiple errors while compiling the kernel

Issue - State: closed - Opened by athu16 over 1 year ago - 34 comments

#10 - Change ints to double in quant_cuda_kernel.cu?

Issue - State: closed - Opened by xiscoding over 1 year ago - 6 comments

#9 - Supports more than a single token

Pull Request - State: closed - Opened by clcarwin over 1 year ago - 1 comment

#8 - How to use for inference?

Issue - State: closed - Opened by DanielWe2 over 1 year ago - 5 comments

#7 - Does not compile on CUDA 12.0

Issue - State: closed - Opened by jtang613 over 1 year ago - 4 comments

#6 - CUDA kernel that supports more than a single token

Issue - State: closed - Opened by ahsima1 over 1 year ago - 2 comments

#5 - AttributeError: 'LLaMAModel' object has no attribute 'decoder'

Issue - State: closed - Opened by Minami-su over 1 year ago - 2 comments

#4 - Request: Optional non-CUDA version

Issue - State: closed - Opened by richardburleigh over 1 year ago - 8 comments

Ecosyste.ms: Issues

GitHub / qwopqwop200/GPTQ-for-LLaMa issues and pull requests