Skip to content

Issues: vllm-project/llm-compressor

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

25 Open 88 Closed
25 Open 88 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

How to speedup the process of quantization which takes almost 20 hours to quantize llama3-70B with w8a8 enhancement New feature or request
#968 opened Dec 11, 2024 by moonlightian
KeyError: 'model.layers.0.self_attn.k_scale' bug Something isn't working
#967 opened Dec 10, 2024 by wxsms
@dsikka
5
[Bug]: When I use llmcompressor to quantify the llama3 70b model to int8-a8w8,it shows ValueError: Failed to invert hessian due to numerical instability. bug Something isn't working
#966 opened Dec 10, 2024 by rexmxw02
@dsikka
1
The new version 0.3.0 takes a long time for quantization and eventually fails due to OOM bug Something isn't working
#965 opened Dec 10, 2024 by okwinds
@horheynm
Error when quantizing LLama 3.3 70b to FP8 bug Something isn't working
#963 opened Dec 6, 2024 by Syst3m1cAn0maly
2
Can I load the stage_quantization model using SparseAutoModelForCausalLM? bug Something isn't working
#962 opened Dec 6, 2024 by jiangjiadi
@dsikka
1
Gptj use gptq quant bug bug Something isn't working
#961 opened Dec 6, 2024 by yemyhdtrc6088
How to recover stage quantization from finetuning stage after an error bug Something isn't working
#957 opened Dec 5, 2024 by jiangjiadi
@dsikka
4
About lora finetuning of 2:4 sparse and sparse quant models enhancement New feature or request
#952 opened Dec 4, 2024 by arunpatala
3
Qwen2VL FP8_DYNAMIC Failed bug Something isn't working
#951 opened Dec 4, 2024 by LugerW-A
@kylesayrs
8
quantization + sparsification - model outputs zeros bug Something isn't working
#942 opened Nov 28, 2024 by nirey10
@rahul-tuli
8
Several wandb init bug Something isn't working
#934 opened Nov 26, 2024 by fzyzcjy
Got Error when I load a 2of4 model using vllm. bug Something isn't working
#926 opened Nov 19, 2024 by jiangjiadi
@dsikka
13
Finetuning in 2:4 sparsity w4a16 example fails with multiple GPUs bug Something isn't working
#911 opened Nov 13, 2024 by arunpatala
@dsikka
4
W8A8 quant for GPT-J failed bug Something isn't working
#909 opened Nov 12, 2024 by zhouyuan
3
[Usage] How to manually set calibration_function?
#886 opened Nov 1, 2024 by donpromax
@rahul-tuli
2
Any plan about W4A8 enhancement New feature or request
#873 opened Oct 29, 2024 by Arcmoon-Hu
2
Does llm-compressor support minicpm3 which is MLA architecture？ enhancement New feature or request
#860 opened Oct 22, 2024 by piamo
@kylesayrs
2
Is it possible to quantize to FP8 W8A16 without calibration data enhancement New feature or request
#858 opened Oct 21, 2024 by us58
1
Perplexity (ppl) Calculation of Local Sparse Model: NaN issue bug Something isn't working
#853 opened Oct 19, 2024 by HengJayWang
2
SmoothQuant doesn't respect ignored modules for VLMs bug Something isn't working
#687 opened Sep 26, 2024 by mgoin
@kylesayrs
3
KV Cache Quantization example cause problem bug Something isn't working
#660 opened Sep 25, 2024 by weicheng59
@markurtz
4
[USAGE] FP8 W8A8 (+KV) with LORA Adapters enhancement New feature or request
#164 opened Sep 11, 2024 by paulliwog
@markurtz
2
Yaml parsing fails with a custom mapping provided to SmoothQuantModifier recipe bug Something isn't working
#105 opened Aug 22, 2024 by aatkinson
@kylesayrs
8
Layers not skipped with ignore=[ "re:.*"] bug Something isn't working
#91 opened Aug 15, 2024 by horheynm
@kylesayrs
2
ProTip! Updated in the last three days: updated:>2024-12-09.