Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / spcl/quarot issues and pull requests
#50 - How to save the quanted model
Issue -
State: open - Opened by cquxl 6 days ago
- 2 comments
#49 - GPTQ dequantization
Issue -
State: closed - Opened by JeevanBhoot 7 days ago
- 2 comments
#48 - H100 Support
Issue -
State: closed - Opened by carlguo866 12 days ago
- 1 comment
#47 - question about quantization group size
Issue -
State: closed - Opened by mxjmtxrm 16 days ago
- 1 comment
#46 - Question about Rotation
Issue -
State: closed - Opened by blgimagineb 18 days ago
#45 - Accuracy drop after rotating model
Issue -
State: closed - Opened by mxjmtxrm 20 days ago
- 1 comment
#44 - question about Hadamard dimension
Issue -
State: closed - Opened by mxjmtxrm 22 days ago
- 5 comments
#43 - Reproducing paper Table 8
Issue -
State: closed - Opened by mjyun01 about 1 month ago
- 1 comment
#42 - How is perplexity calculated with the KV cache?
Issue -
State: closed - Opened by tsengalb99 about 2 months ago
- 1 comment
#41 - [Q] Having not matched size Hadamard matrix
Issue -
State: closed - Opened by Coco58323 3 months ago
- 5 comments
#40 - apply_exact_had_to_linear for v_proj.bias if v_proj.bias is not None
Issue -
State: closed - Opened by dyou-dev 3 months ago
- 1 comment
#39 - questions about the rotate
Issue -
State: closed - Opened by Gloria2tt 3 months ago
- 1 comment
#38 - [Inference speed] Speed up on prefilling stage, slow down on decoding stage
Issue -
State: closed - Opened by ChenMnZ 3 months ago
- 3 comments
#37 - Inference
Issue -
State: closed - Opened by zhentingqi 3 months ago
- 2 comments
#36 - A question regarding the rotation matching pairs
Issue -
State: closed - Opened by Menace-Dragon 3 months ago
- 1 comment
#35 - Mistral support
Issue -
State: closed - Opened by DavidePaglieri 3 months ago
- 1 comment
#34 - Accuracy drop after `fuse_layer_norms`
Issue -
State: closed - Opened by Niko-zyf 4 months ago
- 1 comment
#33 - mlp_sizes seem wrong in qlinear_benchmark.py
Issue -
State: closed - Opened by yyfcc17 4 months ago
- 4 comments
#33 - mlp_sizes seem wrong in qlinear_benchmark.py
Issue -
State: closed - Opened by yyfcc17 4 months ago
- 4 comments
#33 - mlp_sizes seem wrong in qlinear_benchmark.py
Issue -
State: closed - Opened by yyfcc17 4 months ago
- 4 comments
#32 - When is online Hadamard applied during evaluation?
Issue -
State: closed - Opened by pavelgolikov 4 months ago
- 1 comment
#31 - args.distribute_model seems to be undefined
Issue -
State: closed - Opened by WeiMa01 4 months ago
- 3 comments
#31 - args.distribute_model seems to be undefined
Issue -
State: closed - Opened by WeiMa01 4 months ago
- 3 comments
#30 - Outputs of OPT models become different after fusing LayerNorm.
Issue -
State: closed - Opened by SShock92 4 months ago
- 3 comments
#30 - Outputs of OPT models become different after fusing LayerNorm.
Issue -
State: closed - Opened by SShock92 4 months ago
- 3 comments
#29 - opt model with layernorm, the input of layernorm can use hadamard transform?
Issue -
State: closed - Opened by JiangYongYu1 4 months ago
- 4 comments
#29 - opt model with layernorm, the input of layernorm can use hadamard transform?
Issue -
State: closed - Opened by JiangYongYu1 4 months ago
- 4 comments
#28 - Relations with SpinQuant?
Issue -
State: closed - Opened by RanchiZhao 4 months ago
- 3 comments
#28 - Relations with SpinQuant?
Issue -
State: closed - Opened by RanchiZhao 4 months ago
- 3 comments
#27 - Does QuaRot only support Llama and OPT style LLM?
Issue -
State: closed - Opened by NicoNico6 5 months ago
- 1 comment
#27 - Does QuaRot only support Llama and OPT style LLM?
Issue -
State: closed - Opened by NicoNico6 5 months ago
- 1 comment
#26 - Question about Hadamard transformation and outlier reduction
Issue -
State: closed - Opened by KimythAnly 5 months ago
- 2 comments
#26 - Question about Hadamard transformation and outlier reduction
Issue -
State: closed - Opened by KimythAnly 5 months ago
- 2 comments
#25 - Other quantization results of rotated model
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 8 comments
#25 - Other quantization results of rotated model
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 8 comments
#24 - How to get models with only offline rotation (or models for weight-only quantization)
Issue -
State: closed - Opened by Tracin 5 months ago
- 6 comments
#24 - How to get models with only offline rotation (or models for weight-only quantization)
Issue -
State: closed - Opened by Tracin 5 months ago
- 6 comments
#23 - Question about exact_had_to_linear
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 1 comment
#23 - Question about exact_had_to_linear
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 1 comment
#22 - accuracy of weight only quantization decrease significantly after weight rotation
Issue -
State: closed - Opened by luchangli03 5 months ago
- 12 comments
#22 - accuracy of weight only quantization decrease significantly after weight rotation
Issue -
State: closed - Opened by luchangli03 5 months ago
- 12 comments
#21 - Question about rotation.
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 3 comments
#21 - Question about rotation.
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 3 comments
#20 - How to deal with GQA?
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 2 comments
#20 - How to deal with GQA?
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 2 comments
#19 - multi GPU inference
Issue -
State: closed - Opened by hensiesp32 5 months ago
- 1 comment
#19 - multi GPU inference
Issue -
State: closed - Opened by hensiesp32 5 months ago
- 1 comment
#18 - How to get a fake quantized model?
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 1 comment
#18 - How to get a fake quantized model?
Issue -
State: closed - Opened by mxjmtxrm 5 months ago
- 1 comment
#17 - Fix LayerNorm fusion for tied embeddings
Pull Request -
State: closed - Opened by smpanaro 6 months ago
- 1 comment
#17 - Fix LayerNorm fusion for tied embeddings
Pull Request -
State: closed - Opened by smpanaro 6 months ago
- 1 comment
#16 - Wrong result obtained in case of w4a16 quantization?
Issue -
State: closed - Opened by hyx1999 6 months ago
- 2 comments
#16 - Wrong result obtained in case of w4a16 quantization?
Issue -
State: closed - Opened by hyx1999 6 months ago
- 2 comments
#15 - Questions related to Compile the QuaRot on CPU and Model Saving
Issue -
State: closed - Opened by HuangOwen 6 months ago
- 1 comment
#15 - Questions related to Compile the QuaRot on CPU and Model Saving
Issue -
State: closed - Opened by HuangOwen 6 months ago
- 1 comment
#14 - Question about reproducing Fig.1
Issue -
State: closed - Opened by xinghaow99 6 months ago
- 4 comments
#14 - Question about reproducing Fig.1
Issue -
State: closed - Opened by xinghaow99 6 months ago
- 4 comments
#13 - Can we directly load a QuaRot-GPTQ quantized model and do lm_eval evaluation?
Issue -
State: closed - Opened by Shuai-Xie 6 months ago
- 1 comment
#13 - Can we directly load a QuaRot-GPTQ quantized model and do lm_eval evaluation?
Issue -
State: closed - Opened by Shuai-Xie 6 months ago
- 1 comment
#12 - opt model ppl bug
Issue -
State: closed - Opened by zhsky2017 7 months ago
- 3 comments
#12 - opt model ppl bug
Issue -
State: closed - Opened by zhsky2017 7 months ago
- 3 comments
#11 - Questions on online quantization
Issue -
State: closed - Opened by lzhangzz 7 months ago
- 4 comments
#11 - Questions on online quantization
Issue -
State: closed - Opened by lzhangzz 7 months ago
- 4 comments
#10 - Online hadamard bug
Issue -
State: closed - Opened by nailimixaM 7 months ago
#10 - Online hadamard bug
Issue -
State: closed - Opened by nailimixaM 7 months ago
#9 - Some questions
Issue -
State: closed - Opened by catid 7 months ago
- 1 comment
#9 - Some questions
Issue -
State: closed - Opened by catid 7 months ago
- 1 comment
#8 - Question about whether it is necessary to fuse layernorm to linear
Issue -
State: closed - Opened by Oliver-ss 7 months ago
- 14 comments
#7 - [Small Bug] The embedding fusion is not necessary for LLaMA models.
Issue -
State: closed - Opened by ChenMnZ 7 months ago
- 6 comments
#6 - [question] Is it possible to quantize Mixtral?
Issue -
State: closed - Opened by accupham 7 months ago
- 3 comments
#5 - Do I need to use merge a hadamard matrix into W_v if I only want to do 4 bit KV caching?
Issue -
State: closed - Opened by YLGH 7 months ago
- 4 comments
#4 - Question about online hadamard transformation before down-proj and o_proj
Issue -
State: closed - Opened by ChenMnZ 7 months ago
- 1 comment
#3 - Questions about reproduction of weight-only quantization.
Issue -
State: closed - Opened by ChenMnZ 7 months ago
- 6 comments
#2 - Fix typo
Pull Request -
State: open - Opened by eltociear 7 months ago
#1 - Applying rotation to HuggingFace model
Issue -
State: closed - Opened by YLGH 7 months ago
- 12 comments