deepvoltaire / autoaugment Goto Github PK

hello,I want to know how to use autoaugment.py? thanks

How to use the policy given in the article to train our model? What should the training process be like? The test accuracy of the model trained according to my own thinking is very low，please answer

Hi, I meet a problem when I run AutoAugment and I can't find any solution by google.

My environment is different from the tested one. Well, I will feel lucky if it is not a complex compatibility issue.

My environment is as below:
python: 3.7
pytorch: 1.10
Pillow: 5.4.1

File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/site-packages/torch/utils/data/dataloader.py",
line 469, in init
w.start()
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/process.py", line 112, in start
self._popen = self._Popen(self)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/context.py", line 223, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/context.py", line 284, in _Popen
return Popen(process_obj)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/popen_spawn_posix.py", line 32,
in init
super().init(process_obj)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/popen_fork.py", line 20, in init
self._launch(process_obj)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/popen_spawn_posix.py", line 47,
in _launch
reduction.dump(process_obj, fp)
File "/home/dc2-user/anaconda3/envs/kaidi_env/lib/python3.7/multiprocessing/reduction.py", line 60, in dump
ForkingPickler(file, protocol).dump(obj)
AttributeError: Can't pickle local object "SubPolicy.init.<locals>.<lambda>"

Hi,

Thanks a lot for your kindly publishing the code base !!

I noticed that, in this line, the magnitude of translate operation is divided by 331. Would you please tell me the reason you implement this?

By the way, is the meanings of the magnitudes and their explanations in the paper of learning augs for object detection the same as in this paper? Could I simply modify the combination of the policies in this codebase to implement the augmentation method proposed in that paper?

I am looking for your generous guidance :)

How can I train my own AutoAugment for my custom dataset ?

When I test the demo AutoAugment_Exploration.ipynb, I always get the same error: "TypeError: transform() got an unexpected keyword argument 'fillcolor'". Anyone knows why this happens?

Hi, @DeepVoltaire

I tried your autoaugment policies to my image-classification learning(imagenet-pretrained fine tuning).

But, accuracy is not good than default augmentation.

What's wrong to me?

Autoaugmentation is not suitable for transfer learning or finetuning?

Thanks at any rate.

In autoaugment.py.
best Sub-policies.How to determine the parameters.Do these parameters also use my own data?

This is a great tool - but wondering how to go about installing it to run with my pytorch models?

Hi,
Thank you for the wonderful work!
Just wanna share a tiny concern. It seems to me that here the magitude should be magnitude * random.choice([-1, 1], otherwise the rotate will always be in one direction.

There are one or more calendar dates in the readme. They're written in a locale specific style, and this can lead to needless confusion.For example, July 11 can be confused with November 7. Please rewrite the dates in a neutral format, e.g. as in this comment.

I want to apply data enhancement on tensor instead of first data enhancement and then convert into tensor. Can this be done?

Magnitudes are in range [1-10], but based on https://pillow.readthedocs.io/en/3.0.x/reference/ImageOps.html#PIL.ImageOps.posterize

solarize: range is [0, 256]
posterize: range is [1, 8]

Shouldn't the magnitudes be normalized in https://github.com/DeepVoltaire/AutoAugment/blob/master/autoaugment.py#L210 ? For example, line 211 would be

"solarize": lambda img, magnitude: ImageOps.solarize(img, 256.0 / magnitude)

Why train a model directly with cifar10 policy given in the article, but the test accuracy of the model is very low ?

Hi,

I tried to train EfficientNet-B0 with the configuration of 'https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/main.py', but only obtain top-1 accuracy of 75%. The autoaugment paper reported a top-1 accuracy of 77.3%. The model script I used is from 'https://github.com/kakaobrain/fast-autoaugment/tree/master/FastAutoAugment/networks/efficientnet_pytorch'. Has anyone tried to train EfficientNet? I don't know what I did wrong.

Thanks

Hi, does anyone gets the performance on ImageNet with the provided autoaugment?
Here is my results with autoaugment using official implementation, compared to official results, no impressive improvements are got?
Results of ResNet50,101,152 in terms of top1/5 accuracy:
official without autoaugment: 76.15/92.87, 77.37/93.56, 78.31/94.06.
mine with autoaugment: 75.33/92.45, 77.57/93.78, 78.51/94.07.
Update: all the above results are tested with training epochs as 90, a longer one such as 270 used in the paper may help get the reported results.

TypeError: transform() got an unexpected keyword argument 'fillcolor'

Hello, how can I solve this problem?

Hi,

I notice that for autocontrast, the magnitude is not used, and for ImageOps.autocontrast, there is an args named cutoff, which controls the degree of contrast. If we set cutoff=0, the original image would be returned by this function. Does this mean that the ImageOps.autocontrast is not used at all in both the auto-ml search phase and the verification phase(experiments on imagenet dataset) ?

Hi,

Would it be possible to release the pretrained imagenet autoaugment models? In particular, the resnet50 one would be very helpful. I am looking to use it for further downstream work.

thanks

The readme says the following for RandomCrop with cifar10:

 [transforms.RandomCrop(32, padding=4, fill=128), # gray fill value is important bc of the color operations

However, torchvision.transforms.RandomCrop(size, padding=0, pad_if_needed=False) does not support fill as a valid parameter.

Did you run full training with the transformations?

This repo is just the result from the paper, which implements the policies instead of the RL agent controller.

Sub-policy 24(Equalize,0.8,8)(Equalize,0.6,3)

I noticed that the autocontrast operation didn't fill in the cutoff parameter. Is this normal? Without this parameter, it can only make a small change to the picture.

AttributeError: Can't pickle local object 'SubPolicy.init..' when used in ptorch distributed learning

When I used auto augment with my custom NN arch with CIFAR-10 dataset, I got more testing accuracy and less training accuracy. I tried many things and it happens when I implement autoaugmentation only, the under fitting problem, Can please explain why this happening?
Just looking for answers...

downgrading to 6.1 fixed the problem.

may want to leave notes in the readme that 7.0.0 doesn't work.

A link to your TDS blog post can be added to the readme.