facebookresearch/kill-the-bits

Issues

• Inference bug.

  #41 opened 4 years ago by zhangkui669
  1

• Reproducing memory of MaskRCNN and semisupervised ResNet-50

  #39 opened 4 years ago by una-dinosauria
  4

• About the results on semi-supervised ResNet-50

  #38 opened 4 years ago by una-dinosauria
  2

• How does dynamic X determine fixed Q(W)?

  #37 opened 4 years ago by bigdan12
  1

• CentroidSGD in centroid_sgd.py

  #36 opened 4 years ago by cscftao
  1

• Why the activation is the input?

  #35 opened 4 years ago by yw155
  1

• 已压缩模型进行指定图片测试

  #34 opened 5 years ago by findlazygirl
  2

• Question about open code and different accuracy when infer by using quantized pth file

  #33 opened 5 years ago by Eunhui-Kim
  1

• RuntimeError: svd_cuda: the updating process of SBDSDC did not converge

  #13 opened 5 years ago by alanMachineLeraning
  3

• RuntimeError for resnet50 small&large

  #32 opened 5 years ago by bupt-renpei
  1

• How to illustrate 'in-domain' and 'out-of-domain'?

  #31 opened 5 years ago by yw155
  2

• cannot reproduce the results

  #30 opened 5 years ago by minhson
  0

• What should I do if I want to quantize Mask-RCNN by myself

  #29 opened 5 years ago by gitarya
  1

• The val set of imagenet is not stored by its "label folder"!

  #5 opened 5 years ago by talenz
  5

• I got a bad performance with "mask_r_cnn.pth" .How can I improve it, pelase ?

  #28 opened 5 years ago by xscjun
  7

• =

  #27 opened 5 years ago by EdwardLiuYu
  2

• About multi-gpu training(i can't reopen the former issue)

  #26 opened 5 years ago by Fight-hawk
  0

• About multi-gpu training

  #25 opened 5 years ago by Fight-hawk
  6

• How to implement multi-gnu training?

  #20 opened 5 years ago by yw155
  3

• Compression scheduler

  #24 opened 5 years ago by rafikg
  1

• processing time

  #23 opened 5 years ago by rafikg
  1

• small typo in section 4.2

  #22 opened 5 years ago by rafikg
  1

• footprint memory cost!

  #21 opened 5 years ago by rafikg
  1

• How to use '_register_hooks'?

  #19 opened 5 years ago by yw155
  2

• Why the second GPU has memory consumption?

  #18 opened 5 years ago by yw155
  4

• The definition of 'quantizer' in quantize.py

  #17 opened 5 years ago by yw155
  1

• The process of testing MaskRCNN was killed by system due to over RAM.

  #16 opened 5 years ago by mainguyenanhvu
  3

• I can't run python inference.py --model maskrcnn_resnet50_fpn --state-dict-compressed models/compressed/mask_r_cnn.pth --device cuda --data-path YOUR_COCO_PATH

  #15 opened 5 years ago by mainguyenanhvu
  5

• The variable 'in_activations_current' becomes 0 valued at the 9th layer of ResNet50

  #14 opened 5 years ago by yw155
  3

• train acc is about 60%, but test acc is only 8%

  #12 opened 5 years ago by xiaodao2049
  2

• counting ops

  #11 opened 5 years ago by shenyann
  1

• RuntimeError: svd_cuda: the updating process of SBDSDC did not converge (error: 8)

  #10 opened 5 years ago by alanMachineLeraning
  0

• About small datasets training

  #6 opened 5 years ago by wnma3mz
  8

• suspicious that the code of the compression model saved in the published source code is wrong.

  #9 opened 5 years ago by alanMachineLeraning
  8

• TypeError: 'module' object is not callable

  #8 opened 5 years ago by alanMachineLeraning
  1

• Empty cluster still remain

  #7 opened 5 years ago by tuanho27
  5

• how to quantize Mask R-CNN?

  #2 opened 5 years ago by 92ypli
  2

• Does this Quantization model accuracy is calculated on fixed point or floating point?

  #4 opened 5 years ago by tiru1930
  3

• inference time？

  #1 opened 5 years ago by shiyongde
  1