This is an awesome library, thanks @ddbourgin!!

Users might not know the best way to install this package and try it out. (I didn't, so I eventually just copied the source files.)
Neither the readme nor readthedocs have install instructions.

I couldn't find it on PyPi or Anaconda, and there doesn't appear to be a pyproject.toml, setup.cfg, setup.py, or conda recipe.

Moreover, the t

Please can you train ghostnet.
(i don't have the imagenet dataset)

When I delete the lib.utils. ,the test.py is right but the demo.py is wrong , however,when I use the lib.utils. , the demo.py is right but the test.py is wrong !!

why? why

In file binary segmentation (camvid).ipynb, block 5, there is:

# Lets look at data we have
dataset = Dataset(x_train_dir, y_train_dir, classes=['car', 'pedestrian'])
image, mask = dataset[5] # get some sample
visualize(
    image=image, 
    cars_mask=mask[..., 0].squeeze(),
    sky_mask=mask[..., 1].squeeze(),
    background_mask=mask[..., 2].squeeze(),
)

here, sky_mask

Using this issue to keep track of the state-of-the-art achieved using ResNeSt models:

Instance Segmentation

[![PWC](https://

When training, the augmentation RandomScaleCrop may downscale the image and the target label image. It then pads the image and the label with [self.fill][1] which is ZERO.
This is in contrast to the "ignore value" of the loss [that is set to 255][2].
This way the loss treats the padded region as valid "class 0" pixels and compute loss for it.

self.fill of the augmentation functions

with video_demo.py about 20% speed compared to your 1.0 repo. but thanks much for sharing!

The 2x down-sampling is one of the important operations in reference models. But, a convolution or a pooling with stride=2, padding='SAME' may result in different outputs over different deep learning libraries (e.g., TensorFlow, CNTK, Theano, Caffe, Torch, ...) due to their different padding behaviors.

For example (TensorNets syntax; but can be regarded as pseudo codes for other libraries),

In the LSTM documentation is specified that we should feed to the RNN inputs of shape (seq_len, batch, input_size) however it seems to me that we are feeding inputs with shape (batch, seq_len, input_size).

Therefore I believe that the parameter batch_first should be set True.

Describe the bug

Guide under Tensorflow.js heading needs an update, 'names' flag changed to 'name_policy ' according to latest commit?

Under supported layers heading, ConvTrsnpose2d==ConvTranspose2d?

https://github.com/yangxue0827/R2CNN_FPN_Tensorflow/blob/37da5d7e002792b6c73a0aefaed865ca1911b2d3/libs/box_utils/coordinate_convert.py#L43

See here,

Developer should keep in mind that the returned rotatedRect can contain negative indices when data is close to the containing Mat element bo

What is the Map in pascal voc 07 dataset using this code? As i read your code, i find that you do not fix the BN parameters. The batchsize is small, following the paper FPN or Mask RCNN, you need to set the training and trainbale to False for the slim.batch_norm op. I am testing whether this is the problem for the bad generalization for mutiple objects.