facebookresearch/kill-the-bits issues and pull requests

#41 - Inference bug.

Issue - State: closed - Opened by zhangkui669 over 3 years ago - 1 comment

#40 - Address semisup R50 part of #38 and #39

Pull Request - State: closed - Opened by una-dinosauria almost 4 years ago
Labels: CLA Signed

#39 - Reproducing memory of MaskRCNN and semisupervised ResNet-50

Issue - State: closed - Opened by una-dinosauria about 4 years ago - 4 comments

#38 - About the results on semi-supervised ResNet-50

Issue - State: closed - Opened by una-dinosauria about 4 years ago - 2 comments

#37 - How does dynamic X determine fixed Q(W)?

Issue - State: closed - Opened by sfbby about 4 years ago - 1 comment

#36 - CentroidSGD in centroid_sgd.py

Issue - State: closed - Opened by cscftao about 4 years ago - 1 comment

#35 - Why the activation is the input?

Issue - State: closed - Opened by yw155 over 4 years ago - 1 comment

#34 - 已压缩模型进行指定图片测试

Issue - State: closed - Opened by findlazygirl over 4 years ago - 2 comments

#33 - Question about open code and different accuracy when infer by using quantized pth file

Issue - State: closed - Opened by Eunhui-Kim over 4 years ago - 1 comment

#32 - RuntimeError for resnet50 small&large

Issue - State: closed - Opened by bupt-renpei over 4 years ago - 1 comment

#31 - How to illustrate 'in-domain' and 'out-of-domain'?

Issue - State: closed - Opened by yw155 almost 5 years ago - 2 comments

#30 - cannot reproduce the results

Issue - State: closed - Opened by minhson almost 5 years ago

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

Issue - State: closed - Opened by gitarya almost 5 years ago - 1 comment

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

Issue - State: closed - Opened by xscjun almost 5 years ago - 7 comments

#27 - =

Issue - State: closed - Opened by EdwardLiuYu about 5 years ago - 2 comments

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

Issue - State: closed - Opened by Fight-hawk about 5 years ago

#25 - About multi-gpu training

Issue - State: closed - Opened by Fight-hawk about 5 years ago - 6 comments

#24 - Compression scheduler

Issue - State: closed - Opened by rafikg about 5 years ago - 1 comment

#23 - processing time

Issue - State: closed - Opened by rafikg about 5 years ago - 1 comment

#22 - small typo in section 4.2

Issue - State: closed - Opened by rafikg about 5 years ago - 1 comment

#21 - footprint memory cost!

Issue - State: closed - Opened by rafikg about 5 years ago - 1 comment

#20 - How to implement multi-gnu training?

Issue - State: closed - Opened by yw155 about 5 years ago - 3 comments

#19 - How to use '_register_hooks'?

Issue - State: closed - Opened by yw155 about 5 years ago - 2 comments

#18 - Why the second GPU has memory consumption?

Issue - State: closed - Opened by yw155 about 5 years ago - 4 comments

#17 - The definition of 'quantizer' in quantize.py

Issue - State: closed - Opened by yw155 about 5 years ago - 1 comment

#16 - The process of testing MaskRCNN was killed by system due to over RAM.

Issue - State: closed - Opened by mainguyenanhvu about 5 years ago - 3 comments

#15 - 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

Issue - State: closed - Opened by mainguyenanhvu about 5 years ago - 5 comments

#14 - The variable 'in_activations_current' becomes 0 valued at the 9th layer of ResNet50

Issue - State: closed - Opened by yw155 about 5 years ago - 3 comments

#13 - RuntimeError: svd_cuda: the updating process of SBDSDC did not converge

Issue - State: closed - Opened by alanMachineLeraning about 5 years ago - 3 comments

#12 - train acc is about 60%, but test acc is only 8%

Issue - State: closed - Opened by xiaodao2049 about 5 years ago - 2 comments

#11 - counting ops

Issue - State: closed - Opened by shenyann about 5 years ago - 1 comment

#10 - RuntimeError: svd_cuda: the updating process of SBDSDC did not converge (error: 8)

Issue - State: closed - Opened by alanMachineLeraning about 5 years ago

#9 - suspicious that the code of the compression model saved in the published source code is wrong.

Issue - State: closed - Opened by alanMachineLeraning about 5 years ago - 8 comments

#8 - TypeError: 'module' object is not callable

Issue - State: closed - Opened by alanMachineLeraning about 5 years ago - 1 comment

#7 - Empty cluster still remain

Issue - State: closed - Opened by tuanho27 over 5 years ago - 5 comments

#6 - About small datasets training

Issue - State: closed - Opened by wnma3mz over 5 years ago - 8 comments

#5 - The val set of imagenet is not stored by its "label folder"!

Issue - State: closed - Opened by talenz over 5 years ago - 5 comments

#4 - Does this Quantization model accuracy is calculated on fixed point or floating point?

Issue - State: closed - Opened by tiru1930 over 5 years ago - 3 comments

#3 - Fix version specification

Pull Request - State: closed - Opened by take-cheeze over 5 years ago - 2 comments
Labels: CLA Signed

#2 - how to quantize Mask R-CNN?

Issue - State: closed - Opened by 92ypli over 5 years ago - 2 comments

#1 - inference time？

Issue - State: closed - Opened by shiyongde over 5 years ago - 1 comment

GitHub / facebookresearch/kill-the-bits issues and pull requests