ModelTC/MQBench

Issues

• WARNING

  #269 opened 2 months ago by zfstr
  0

• ImportError: cannot import name 'ConvBNReLUFusion' from 'torch.quantization.fx.fusion_patterns'

  #268 opened 3 months ago by huangweiwh
  1

• AttributeError: 'NoneType' object has no attribute 'detach'

  #264 opened 4 months ago by shenxiaozheng
  8

• 关于混合精度mixed-precision量化的支持和bug问题

  #238 opened 5 months ago by Tlinn7
  4

• 请问，有对不同BenkendType之间转换的工具，比如Academic转到SNPE的后端？

  #267 opened 5 months ago by hewumars
  1

• how to config additional_function_type in yaml config?

  #263 opened 6 months ago by Alicture
  1

• MQBench是否有交流群

  #247 opened 6 months ago by UnendingGlory
  3

• 如何添加定制的conv

  #261 opened 7 months ago by iamweiweishi
  1

• model.zero_grad() ?

  #258 opened 7 months ago by jeannotes
  2

• why the scale and zero_point not update in Learnablefakequantize? I used different methods to quantize the yolov5

  #259 opened 7 months ago by buzhou1
  4

• Can't convert to deploy after use advanced_ptq

  #262 opened 7 months ago by buzhou1
  2

• 为什么vitis后端每层都量化？

  #257 opened 7 months ago by jeannotes
  3

• spring 安装失败

  #255 opened 7 months ago by zyzqmm
  2

• 关于模型中有控制流不能被追踪的问题

  #253 opened 8 months ago by wrxjt
  2

• The accuracy of mqbench QAT is inconsistent with that of running it using SNPE

  #254 opened 8 months ago by theNefelibata
  3

• Why is the "Fisher Diagnoal Matrix" mentioned in brecq not used in mqbench ?

  #252 opened 8 months ago by flytigerw
  1

• PTQ量化后如何获得一个纯净的pytorch模型

  #251 opened 8 months ago by Arthur-Ling
  1

• lsq 训练量化error： name 'grad_factor' is not defined

  #239 opened 9 months ago by wangshankun
  2

• About the quantization of recurrent neural network!

  #241 opened 9 months ago by larenzhang
  2

• 输入输出设置成16bit

  #244 opened 9 months ago by pixel-lt
  2

• Does it support on NV tensorrt

  #246 opened 9 months ago by lix19937
  2

• advanced PTQ accuracy too low

  #250 opened 9 months ago by taobiaoli
  8

• 使用convert_deploy导出的onnx进行推理的指标与在训练过程中的指标不一致，要差于训练过程，做了QAT量化导出和非QAT量化导出的实验对比，非QAT量化导出的onnx推理指标正常

  #248 opened 9 months ago by Jim2016713
  2

• Is the current repo no longer maintained

  #249 opened 9 months ago by lix19937
  2

• Pot scale 的伪量化模型和部署模型精度差异很大

  #235 opened 10 months ago by wangshankun
  6

• 模型转换onnx失败

  #245 opened 10 months ago by SunCrazy
  1

• 自定义算子量化

  #243 opened a year ago by YaqinLong
  1

• PTQ后转TVM失败

  #242 opened a year ago by Xrigok
  1

• vitis-ai-gpu:1.4.1难以建立，请问有什么可用的镜像吗？

  #260 opened a year ago by ranger-wang
  2

• ConvTranspose2d in Adaround

  #210 opened a year ago by jmliu206
  4

• when will you update to torch 2.0

  #256 opened a year ago by jeannotes
  1

• 如何使用weight_equalization 呢？

  #236 opened a year ago by LHH20000923
  1

• qat 训练 模型convert_deploy的问题

  #233 opened a year ago by wqn628
  2

• Comparison with the unquantized model on the OpenVINO backend

  #234 opened a year ago by audreyeternal
  1

• 关于DP 和 DDP 训练时报错 “symbolically traced variables cannot be used as inputs to control flow”

  #230 opened a year ago by shinianzhihou
  2

• onnx qnn 部署模型问题

  #231 opened a year ago by wangshankun
  7

• RuntimeError: shape '[64, 1, 1, 1]' is invalid for input of size 1

  #228 opened a year ago by Haleski47
  3

• Try to convert my model but get error

  #212 opened a year ago by szu-lisili
  3

• 关于使用ONNX-QNN在生成Deploy模型出现的问题

  #223 opened a year ago by Zhoukai1234
  2

• MQBench在pytorch1.11出现torch接口不兼容不支持

  #222 opened a year ago by meton-robean
  1

• The QAT top1@acc of mobilenet_v2 a4w4 LSQ cannot be reproduced as the paper shown 70.6%.

  #221 opened a year ago by LuletterSoul
  1

• 关于yolov5s进行PTQ量化出现TraceError问题

  #220 opened a year ago by xiaopengaia
  3

• 关于如何将conv和bn层进行合并的问题

  #218 opened a year ago by xiaopengaia
  3

• 关于量化模型不能进行保存的问题

  #216 opened a year ago by xiaopengaia
  3

• 使用MQBench进行PTQ量化不清楚是否量化成功？

  #214 opened a year ago by xiaopengaia
  4

• Does MQBench supports Advanced PTQ on the transformer model?

  #215 opened a year ago by hobbitlzy
  3

• ImportError: cannot import name 'builder' from 'google.protobuf.internal'

  #213 opened a year ago by xiaopengaia
  2

• Is dynamic batch supported?

  #211 opened a year ago by audreyeternal
  2

• DDP训练出错

  #240 opened a year ago by wmn931201
  1

• 不能够复现Application-imagenet_example ，基于tensorRT的evaluation结果精度下降40%，向您请教。

  #224 opened a year ago by YeRen123455
  3