Aaronhuang-778 BiLLM issues - Githubissues

Aaronhuang-778 / BiLLM

(ICML 2024) BiLLM: Pushing the Limit of Post-Training Quantization for LLMs

https://arxiv.org/abs/2402.04291

MIT License

198 stars 14 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Questions about how to reproduce the results in the paper

#19 LiuSiQi-TJ opened 1 week ago
1
ValueError: BuilderConfig 'allenai--c4' not found

#18 GreatGBL opened 3 months ago
0
为什么我在量化模型之后模型的体积几乎没有变化

#17 zxbjushuai closed 2 months ago
4
Does it support LLAMA3-8B-INSTUCT or QWEN2-7B-INSTRUCT?

#16 LiMa-cas opened 4 months ago
0
scale factor and bit storing calculation

#15 kaizizzzzzz opened 5 months ago
1
Issue with code replication

#14 Devy99 opened 5 months ago
2
Question about 1-bit compression with combined binary masks

#13 pprp opened 6 months ago
0
Any Plan for Multi-GPUs Support?

#12 pprp opened 7 months ago
0
question about weight storage when infer

#11 DamonsJ opened 7 months ago
0
Looking forward to supporting Mixtral_8x7b MoE

#10 Gierry opened 8 months ago
1
inference

#9 shyget opened 8 months ago
0
fix requirements.txt

#8 earsaxcs closed 8 months ago
0
Inference

#7 diff7 opened 9 months ago
0
I have a question about the paper.

#6 shampooooo closed 9 months ago
3
Possible Error in the Paper

#5 shampooooo closed 9 months ago
0
Do you quantize the LM head, embedding, and layernorms or just the weights?

#4 tsengalb99 opened 9 months ago
0
Update README.md

#3 eltociear closed 9 months ago
1
Model weight access?

#2 BarfingLemurs closed 9 months ago
1
Request: please consider evaluating pareto-optimality of BiLLM

#1 justheuristic opened 9 months ago
3