kalviny/msdnet-pytorch

MSDNet

msdnet-pytorch's Issues

这个程序在 i5-8300H 12G内存 GTX 1050上得跑多长时间啊？

请问为什么在运行的时候为什么会报 AttributeError: module 'models' has no attribute 'resnet' 这个错误？

D:\software\ana\envs\env1\python.exe C:/Users/Administrator/Desktop/黄高/MSDNet-PyTorch-master/main.py
Traceback (most recent call last):
File "C:/Users/Administrator/Desktop/黄高/MSDNet-PyTorch-master/main.py", line 352, in
main()
File "C:/Users/Administrator/Desktop/黄高/MSDNet-PyTorch-master/main.py", line 62, in main
model = getattr(models, args.arch)(args)
AttributeError: module 'models' has no attribute 'resnet'

Process finished with exit code 1

Reproduce budget batch classification results for CIFAR100

I am trying to reproduce the MSDNet with dynamic evaluation curve of Figure 7b of the paper (budgeted batch classification for CIFAR-100).

I am training the MSDNet as mentioned in the README file.
I then evaluate by:
python main.py --data-root /path/to/cifar100/ --data cifar100 --save /path/to/out/ --arch msdnet --batch-size 64 --epochs 300 --nBlocks 5 --stepmode lin_grow --step 1 --base 1 --nChannels 16 --evalmode dynamic --evaluate-from /path/to/out/save_models/checkpoint_299.pth.tar --use-valid -j 1 --gpu 0

Output log

building network of steps:
[1, 2, 3, 4, 5] 15
 ********************** Block 1  **********************
|		inScales 3 outScales 3 inChannels 16 outChannels 6		|

 ********************** Block 2  **********************
|		inScales 3 outScales 3 inChannels 22 outChannels 6		|

|		inScales 3 outScales 3 inChannels 28 outChannels 6		|

 ********************** Block 3  **********************
|		inScales 3 outScales 3 inChannels 34 outChannels 6		|

|		inScales 3 outScales 3 inChannels 40 outChannels 6		|

|		inScales 3 outScales 2 inChannels 46 outChannels 6		|
|		Transition layer inserted! (max), inChannels 52, outChannels 26	|

 ********************** Block 4  **********************
|		inScales 2 outScales 2 inChannels 26 outChannels 6		|

|		inScales 2 outScales 2 inChannels 32 outChannels 6		|

|		inScales 2 outScales 2 inChannels 38 outChannels 6		|

|		inScales 2 outScales 2 inChannels 44 outChannels 6		|

 ********************** Block 5  **********************
|		inScales 2 outScales 1 inChannels 50 outChannels 6		|
|		Transition layer inserted! (max), inChannels 56, outChannels 28	|

|		inScales 1 outScales 1 inChannels 28 outChannels 6		|

|		inScales 1 outScales 1 inChannels 34 outChannels 6		|

|		inScales 1 outScales 1 inChannels 40 outChannels 6		|

|		inScales 1 outScales 1 inChannels 46 outChannels 6		|

---------------------
FLOPs: 6.86M, Params: 0.30M
---------------------
FLOPs: 14.35M, Params: 0.65M
---------------------
FLOPs: 27.54M, Params: 1.02M
---------------------
FLOPs: 41.71M, Params: 1.49M
---------------------
FLOPs: 58.48M, Params: 2.08M
building network of steps:
[1, 2, 3, 4, 5] 15
 ********************** Block 1  **********************
|		inScales 3 outScales 3 inChannels 16 outChannels 6		|

 ********************** Block 2  **********************
|		inScales 3 outScales 3 inChannels 22 outChannels 6		|

|		inScales 3 outScales 3 inChannels 28 outChannels 6		|

 ********************** Block 3  **********************
|		inScales 3 outScales 3 inChannels 34 outChannels 6		|

|		inScales 3 outScales 3 inChannels 40 outChannels 6		|

|		inScales 3 outScales 2 inChannels 46 outChannels 6		|
|		Transition layer inserted! (max), inChannels 52, outChannels 26	|

 ********************** Block 4  **********************
|		inScales 2 outScales 2 inChannels 26 outChannels 6		|

|		inScales 2 outScales 2 inChannels 32 outChannels 6		|

|		inScales 2 outScales 2 inChannels 38 outChannels 6		|

|		inScales 2 outScales 2 inChannels 44 outChannels 6		|

 ********************** Block 5  **********************
|		inScales 2 outScales 1 inChannels 50 outChannels 6		|
|		Transition layer inserted! (max), inChannels 56, outChannels 28	|

|		inScales 1 outScales 1 inChannels 28 outChannels 6		|

|		inScales 1 outScales 1 inChannels 34 outChannels 6		|

|		inScales 1 outScales 1 inChannels 40 outChannels 6		|

|		inScales 1 outScales 1 inChannels 46 outChannels 6		|

!!!!!! Load train_set_index !!!!!!
*********************
/home/damaskin/MSDNet-PyTorch/adaptive_inference.py:28: UserWarning: torch.range is deprecated in favor of torch.arange and will be removed in 0.5. Note that arange generates values in [start; end), not [start; end].
  probs = torch.exp(torch.log(_p) * torch.range(1, args.nBlocks))
valid acc: 59.020, test acc: 61.500, test flops: 7.09M
*********************
valid acc: 59.960, test acc: 62.080, test flops: 7.44M
*********************
valid acc: 60.700, test acc: 62.870, test flops: 7.90M
*********************
valid acc: 61.660, test acc: 63.500, test flops: 8.36M
*********************
valid acc: 62.700, test acc: 64.050, test flops: 8.93M
*********************
valid acc: 63.900, test acc: 64.850, test flops: 9.66M
*********************
valid acc: 65.040, test acc: 65.670, test flops: 10.53M
*********************
valid acc: 65.760, test acc: 66.400, test flops: 11.32M
*********************
valid acc: 66.920, test acc: 67.130, test flops: 12.34M
*********************
valid acc: 67.740, test acc: 67.800, test flops: 13.32M
*********************
valid acc: 68.600, test acc: 68.560, test flops: 14.48M
*********************
valid acc: 69.500, test acc: 69.330, test flops: 15.67M
*********************
valid acc: 69.880, test acc: 69.980, test flops: 16.88M
*********************
valid acc: 70.420, test acc: 70.310, test flops: 18.25M
*********************
valid acc: 70.820, test acc: 70.770, test flops: 19.61M
*********************
valid acc: 71.460, test acc: 71.150, test flops: 20.90M
*********************
valid acc: 71.840, test acc: 71.530, test flops: 22.28M
*********************
valid acc: 72.260, test acc: 71.970, test flops: 23.62M
*********************
valid acc: 72.860, test acc: 72.150, test flops: 25.03M
*********************
valid acc: 72.880, test acc: 72.360, test flops: 26.24M
*********************
valid acc: 72.880, test acc: 72.530, test flops: 27.46M
*********************
valid acc: 72.960, test acc: 72.610, test flops: 28.95M
*********************
valid acc: 73.080, test acc: 72.650, test flops: 30.14M
*********************
valid acc: 72.840, test acc: 72.900, test flops: 31.26M
*********************
valid acc: 72.780, test acc: 72.920, test flops: 32.32M
*********************
valid acc: 72.840, test acc: 73.070, test flops: 33.44M
*********************
valid acc: 72.640, test acc: 73.020, test flops: 34.26M
*********************
valid acc: 72.720, test acc: 73.160, test flops: 35.17M
*********************
valid acc: 72.740, test acc: 72.940, test flops: 36.09M
*********************
valid acc: 72.660, test acc: 73.030, test flops: 36.87M
*********************
valid acc: 72.700, test acc: 72.980, test flops: 37.61M
*********************
valid acc: 72.600, test acc: 72.920, test flops: 38.28M
*********************
valid acc: 72.580, test acc: 72.870, test flops: 39.01M
*********************
valid acc: 72.460, test acc: 72.680, test flops: 39.59M
*********************
valid acc: 72.400, test acc: 72.660, test flops: 40.20M
*********************
valid acc: 72.300, test acc: 72.660, test flops: 40.80M
*********************
valid acc: 72.220, test acc: 72.620, test flops: 41.28M
*********************
valid acc: 72.160, test acc: 72.580, test flops: 41.80M
*********************
valid acc: 72.160, test acc: 72.580, test flops: 42.16M

The final [test accuracy, budget] pair is [72.58, 0.4 * 10^8] which is not consistent with the results of Figure 7b (as it should have been [~74, 0.4 * 10^8]).
What are the parameters of the 3 MSDNets used in Figure 7b ?

the pretrained model

The tar files inside step=4 and step=7 folder seem corrupt. Could you please check it.

Test when evalmode is dynamic

Sorry, I couldn't understand the function dynamic_evaluate(model, test_loader, val_loader, args) when testing.

if args.evalmode == 'anytime':
            validate(test_loader, model, criterion)
        else:
            dynamic_evaluate(model, test_loader, val_loader, args)
        return

I have obtained models on my dataset. I want to test the model's performance on some images in a dir. However, these images don't have labels. Could I use dynamic (evalmode) or just anytime? Does the test data need labels when chosing dynamic?

在ImageNet上的预训练模型无法下载

Could you pls provide the ImageNet pretrained PyTorch model?

Always get "tar: This does not look like a tar archive" when extracting pre-trained models

Hi authors,

Thanks for your great work! I encountered a problem when trying to use your pretrained model to perform evaluations.
I downloaded the tar file from the dropbox link you provided. When I tried to extract it (using command tar -xvf msdnet-step\=4-block\=5.pth.tar, I alway got the following error:

tar: This does not look like a tar archive
tar: Skipping to next header
tar: Exiting with failure status due to previous errors

Could you tell me what I should do? Thanks a lot!

CIFAR-10 results

Hi Kalviny, thanks for your great work! I'm wondering where can I get the CIAFR-10 results of the MSDNet. Thank you!

please add a LICENSE

to enable other to re-use this source code.
Thank you!

Version of ImageNet dataset

Which version of ILSVRC (ImageNet) dataset was used for training the MSDNet in this repo? Was it the 2012 one (i.e. the same one as in the original MSDNet paper)?

could you tell me how to train my own image dataset rather than imagenet or cifar?

could you tell me how to train my own image dataset rather than imagenet or cifar?
Does it only need to edit the dataloader rather than edit the model code?

Thanks a lot!

Configure of step7

Hi, thanks for your work. This code is pretty clean and nice. I want to do get the result of imagenet step7. Can you add the inference command for step 7 imagenet?

The model technically never does 'early exit' ?

Hi, @kalviny thanks a lot for the excellent implementation! It has been extremely helpful for my work.

Reading through validate() in main.py and dynamic_evaluate() in adaptive_inference.py it seems to me that the model processes the entire input in both anytime prediction and budgeted batch classification setting:

In main.py in validate() function in Line 227
output = model(input_var)

In adaptive_inference.py in calc_logit() function in Line 53:
output = self.model(input_var)

which returns an array for results from each block. And then the accuracy calculations for early exit are done in hindsight considering whether the model would have been right or wrong if we consider the output from the intermediate classifiers. Could you please confirm that my interpretation is correct?

Pre_trained model

Your model folder have 2 file index.pth and msdnet-step=4-block=5.pth.tar. Which one I should load to "checkpoint = load_checkpoint(args)"?
I open index.pth, it may not the pretrained weight, while I can't extract msdnet-step=4-block=5.pth.tar.
Thank you

Problem with the meaning of arguments

What do you mean by arguments --step, --stepmode, --use-valid, -j and --base. Also please explain their functions in the formation of blocks of MSDNET ?

关于和论文的出入

请问这个实现了论文中所说的lazy evaluation吗

How to save the flops.pth of MSDNet in adaptive_inference.py?

Hi author, thanks for your work, but I do not understand the dynamic evaluation in origin paper and how does it works in the code. I can not fine the way to compute and save the flops in your code. Can you send me the code about this point? cause I think I can understand paper by code. Thank you very much.

Model initialization exception

Thanks for sharing your code. I find some anomalous code In msdnet.py#L322.
these three "if branch" will not be excuted in func _init_weights(self, m).

MSDNet-PyTorch/models/msdnet.py

Lines 242 to 264 in c16a62b

	for m in self.blocks:
	if hasattr(m, '__iter__'):
	for _m in m:
	self._init_weights(_m)
	else:
	self._init_weights(m)
	for m in self.classifier:
	if hasattr(m, '__iter__'):
	for _m in m:
	self._init_weights(_m)
	else:
	self._init_weights(m)
	def _init_weights(self, m):
	if isinstance(m, nn.Conv2d):
	n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
	m.weight.data.normal_(0, math.sqrt(2. / n))
	elif isinstance(m, nn.BatchNorm2d):
	m.weight.data.fill_(1)
	m.bias.data.zero_()
	elif isinstance(m, nn.Linear):
	m.bias.data.zero_()

About reproducing anytime experiment in paper

Hi, thank you very much for this repo! May I ask the performance of step=7/msdnet-step=7-block=5.pth.tar model on Imagenet anytime prediction task?

The result I got is

 * prec@1 62.896 prec@5 84.706
 * prec@1 70.316 prec@5 89.590
 * prec@1 73.000 prec@5 91.230
 * prec@1 73.594 prec@5 91.452
 * prec@1 74.592 prec@5 92.066

---------------------
FLOPs: 615.69M, Params: 8.76M
---------------------
FLOPs: 1436.39M, Params: 20.15M
---------------------
FLOPs: 2283.21M, Params: 31.73M
---------------------
FLOPs: 2967.42M, Params: 41.86M
---------------------
FLOPs: 3253.79M, Params: 62.31M

It is slightly lower than the performance provided in the figure of paper