innerlee

5 ms·

1.
by innerlee 1y ago
The singular defects (or high-norm tokens) [1] may be related to attention sinks. It is interesting that the direction of all high-norm tokens share the same direction. Maybe the theory behind is not very complex and the issue can be fixed
2.
by innerlee 5y ago
For more text classification baselines (CRNN, NRTR, RubustScanner, SAR, SegOCR), checkout https://github.com/open-mmlab/mmocr They are reproducible, customizable.
3.
by innerlee 5y ago
We also expored ways of presenting hundreds [1] of AI models in a better way. Recently we updated the model zoo so that it is arranged according to benchmark datasets [2] and papers [3]. Do you think it satisfies the needs? [1] https:&#x2F
4.
OCR Baselines Based on PyTorch(github.com)
1 points by innerlee 5y ago|1 comments
5.
by innerlee 5y ago
One of the maintainer here. We just now open-sourced a new ocr library. It contains 5 text detection algorithm, 4 text recognition algorithms, and 1 key information extraction algorithm (details see [1]). It is used internally for some time
6.
by innerlee 6y ago
Checkout the openmmlab project [1], where messy research codes in computer vision are rewritten an reorganized into a coherent whole. If there are more researchers join this, then not only research are fully reproducible, much more accessib
7.
by innerlee 6y ago
Sure, checkpoints, configs and detailed training logs all are available at modelzoo https://mmaction2.readthedocs.io/en/latest/recognition_model... The single RGB stream top1 goes up to 73.48% with resnet50, and
8.
by innerlee 6y ago
shameless ad: try mmaction2, where every result is reproducible https://github.com/open-mmlab/mmaction2 . Modelzoo: https://mmaction2.readthedocs.io/en/latest/modelzoo.html
9.
Tasks and Algorithms Supported by MMPose(mmpose.readthedocs.io)
1 points by innerlee 6y ago|0 comments
10.
MMAction2 v0.8.0 Released(github.com)
1 points by innerlee 6y ago|0 comments
11.
by innerlee 6y ago
Here is a demo video generated by mmpose: https://www.youtube.com/watch?v=pIJpQg8mXUU
12.
MMPose: Pose Estimation in PyTorch(github.com)
2 points by innerlee 6y ago|1 comments
13.
by innerlee 6y ago
MMPose is an open-source toolbox for pose estimation based on PyTorch. It is a part of the OpenMMLab project. In V0.7.0, it adds support for the 3D human shape recovery task, adds support for WholeBody human pose estimation task, and the 2D
14.
MMPose v0.7.0 Released(github.com)
2 points by innerlee 6y ago|1 comments