pytorch-retinanet | Pytorch implementation of RetinaNet object detection | Computer Vision library

by yhenon Python Version: Current License: Apache-2.0

X-Ray Key Features Code Snippets Community Discussions(2)Vulnerabilities Install Support

kandi X-RAY | pytorch-retinanet Summary

pytorch-retinanet is a Python library typically used in Artificial Intelligence, Computer Vision, Deep Learning, Pytorch applications. pytorch-retinanet has no bugs, it has no vulnerabilities, it has a Permissive License and it has medium support. However pytorch-retinanet build file is not available. You can download it from GitHub.

Pytorch implementation of RetinaNet object detection.

Support

Quality

Security

License

Reuse

Support

pytorch-retinanet has a medium active ecosystem.

It has 1977 star(s) with 650 fork(s). There are 30 watchers for this library.

It had no major release in the last 6 months.

There are 135 open issues and 82 have been closed. On average issues are closed in 77 days. There are 9 open pull requests and 0 closed requests.

It has a neutral sentiment in the developer community.

The latest version of pytorch-retinanet is current.

Quality

pytorch-retinanet has 0 bugs and 0 code smells.

Security

pytorch-retinanet has no vulnerabilities reported, and its dependent libraries have no vulnerabilities reported.

pytorch-retinanet code analysis shows 0 unresolved vulnerabilities.

There are 0 security hotspots that need review.

License

pytorch-retinanet is licensed under the Apache-2.0 License. This license is Permissive.

Permissive licenses have the least restrictions, and you can use them in most projects.

Reuse

pytorch-retinanet releases are not available. You will need to build from source code and install.

pytorch-retinanet has no build file. You will be need to create the build yourself to build the component from source.

Installation instructions, examples and code snippets are available.

pytorch-retinanet saves you 656 person hours of effort in developing the same functionality from scratch.

It has 1521 lines of code, 90 functions and 14 files.

It has medium code complexity. Code complexity directly impacts maintainability of the code.

Top functions reviewed by kandi - BETA

kandi has reviewed pytorch-retinanet and discovered the below as its top functions. This is intended to give you an instant insight into pytorch-retinanet implemented functionality, and help decide if they suit your requirements.

Evaluate a COCO model
Evaluate a generator function
Compute the area of the PR curve
Get the annotations from the given generator
Detects the image of the given model
Loads the classes from the csv file
Draws a caption
Compute the anchors for the given image
Generate anchors
Shift coordinates by a given shape
Compute anchors for a given image
Compute image shape
Compute the loss function
Calculate the IOU curve
Read annotations from csv file
Parse the value
Freeze BatchNorm layers
Returns the number of classes

Get all kandi verified functions for this library.

pytorch-retinanet Key Features

No Key Features are available at this moment for pytorch-retinanet.

pytorch-retinanet Examples and Code Snippets

No Code Snippets are available at this moment for pytorch-retinanet.

Community Discussions

Trending Discussions on pytorch-retinanet

How are the FCN heads convolved over RetinaNet's FPN features?

modification from classification loss to regression loss

QUESTION

How are the FCN heads convolved over RetinaNet's FPN features?

Asked 2020-May-12 at 18:26

I've recently read the RetinaNet paper and I have yet to understood one minor detail:
We have the multi-scale feature maps obtained from the FPN (P2,...P7).
Then the two FCN heads (the classifier head and regessor head) are convolving each one of the feature maps.
However, each feature map has different spatial scale, so, how does the classifier head and regressor head maintain fixed output volumes, given all their convolution parameters are fix? (i.e. 3x3 filter with stride 1, etc).

Looking at this line at PyTorch's implementation of RetinaNet, I see the heads just convolve each feature and then all features are stacked somehow (the only common dimension between them is the Channel dimension which is 256, but spatially they are double from each other).
Would love to hear how are they combined, I wasn't able to understand that point.

...

ANSWER

Answered 2020-May-12 at 18:26

After the convolution at each pyramid step, you reshape the outputs to be of shape (H*W, out_dim) (with out_dim being num_classes * num_anchors for the class head and 4 * num_anchors for the bbox regressor). Finally, you can concatenate the resulting tensors along the H*W dimension, which is now possible because all the other dimensions match, and compute losses as you would on a network with a single feature layer.

Source https://stackoverflow.com/questions/61736928

QUESTION

modification from classification loss to regression loss

Asked 2020-Jan-29 at 18:03

General

I am following this repo for object detection https://github.com/yhenon/pytorch-retinanet

Motivation

Object detection networks usually perform 2 tasks.For every object in the image output a class confidence score and bounding box regression score.For my task along with these 2 outputs, for every object i want to output one more regression score which will be between 0-5.

Problem statement

The approach i have taken is that since the network already does classifcation i thought i would modify some of the parts and make it a regression loss.The loss used in the repo mentioned above is focal loss.Part of what the classification loss looks like is described below, i would like to modify this to be a regression loss.

...

ANSWER

Answered 2020-Jan-29 at 09:39

If I understand it correctly you want to add another regression parameter which should have values in [0-5].

Instead of trying to change the classification part you can just add it to the box regression part. So you add one more parameter to predict for every anchor.

Here is the loss calculation for the regression parameters. If you added one parameter to every anchor (probably by increasing the filter count in the last layer) you have to "extract" it here like with the other parameters. You also probably want to think about the activation for the value e.g. sigmoid so you get something between 0 and 1, you can rescale to 0-5 if you want.

After line 107 you want to add something like this

Source https://stackoverflow.com/questions/59909066

Community Discussions, Code Snippets contain sources that include Stack Exchange Network

Vulnerabilities

No vulnerabilities reported

Install pytorch-retinanet

Install the required packages:.
Clone this repo
Install the required packages:
Install the python packages:

Support

For any new features, suggestions and bugs create an issue on GitHub. If you have any questions check and ask questions on community page Stack Overflow .

Find more information at: