Issues

• Is it possible to smooth FC2 input?

  mit-han-lab/smoothquant#108 · Filter by author MaxwellWjj opened on Dec 18, 2025
  1

• applying O3 quant config to llama

  mit-han-lab/smoothquant#107 · Filter by author cjm715 opened on Nov 26, 2025
  1

• How to perform real quantization model inference with SmoothQuant?

  mit-han-lab/smoothquant#106 · Filter by author wanto-rest opened on Sep 18, 2025

• SmoothQuant on t5?

  mit-han-lab/smoothquant#105 · Filter by author Manojbhat09 opened on Feb 8, 2025

• SmoothQuant kv-cache quantization

  mit-han-lab/smoothquant#104 · Filter by author wuguowuge opened on Feb 7, 2025

• Sensitivity analysis of model parameters

  mit-han-lab/smoothquant#102 · Filter by author shengqihailuo1 opened on Jan 16, 2025

• RuntimeError: “triu_tril_cuda_template“ not implemented for ‘BFloat16‘

  mit-han-lab/smoothquant#101 · Filter by author cuicui2023 opened on Jan 3, 2025

• error when getting the scales for my model

  mit-han-lab/smoothquant#100 · Filter by author rib12316 opened on Dec 19, 2024
  2

• How to use this for custom model?

  mit-han-lab/smoothquant#98 · Filter by author siddagra opened on Dec 7, 2024

• how to visual outier of activation

  mit-han-lab/smoothquant#97 · Filter by author harleyszhang opened on Nov 3, 2024
  2

• The upper and lower bounds seems to not 8 bits in some cases

  mit-han-lab/smoothquant#96 · Filter by author zhangyu68 opened on Oct 30, 2024

• Why only 4 layers?

  mit-han-lab/smoothquant#95 · Filter by author VincentXWD opened on Sep 7, 2024

• Support for Qwen2

  mit-han-lab/smoothquant#94 · Filter by author JiaXinLI98 opened on Jul 31, 2024

• How to quantize the out_proj and fc2 module in OPT model family

  mit-han-lab/smoothquant#93 · Filter by author yanchenmochen opened on Jul 30, 2024

• How to quantize llama3?

  mit-han-lab/smoothquant#92 · Filter by author jpyo0803 opened on Jul 22, 2024

• export_int8_model.py size issue

  mit-han-lab/smoothquant#91 · Filter by author jhy20k opened on Jul 11, 2024
  1

• quantify other models,

  mit-han-lab/smoothquant#90 · Filter by author AlexMa0 opened on Jul 9, 2024

• best Alpha value for Qwen 1.5 72B

  mit-han-lab/smoothquant#89 · Filter by author Riskin1999 opened on Jun 26, 2024
  1

• how to draw this result directly? is there any script?

  mit-han-lab/smoothquant#88 · Filter by author foreverpiano opened on Jun 5, 2024
  1

• Huggingface_Hub Issue

  mit-han-lab/smoothquant#87 · Filter by author faize5 opened on May 23, 2024
  2

• Can SmoothQuant be used on ViT models?

  mit-han-lab/smoothquant#86 · Filter by author n9s8a opened on Apr 24, 2024
  1

• Whether it can be supported stable diffusion

  mit-han-lab/smoothquant#85 · Filter by author songh11 opened on Apr 11, 2024

• Inquiry about Int8 BMM overflow

  mit-han-lab/smoothquant#84 · Filter by author luzai opened on Apr 9, 2024

• Error when running smoothquant_opt_real_int8_demo.ipynb

  mit-han-lab/smoothquant#83 · Filter by author kaijun924 opened on Apr 2, 2024

• how to use model.generate with smoothquant models

  mit-han-lab/smoothquant#82 · Filter by author haoyun-deng opened on Mar 31, 2024