[RFC] MindCV 0.2 Roadmap - New Feature/Model Request

[SamitHuang]

We are open to collecting valuable new models and algorithms from users!

You are welcome to leave a comment to suggest a new feature/model/algorithm for MIndCV. You can also vote 👍 or dis-vote 👎 a suggested feature. If you are interested in implementing one of the suggested features, please leave a comment and let us know

We are looking forward to your voice :)

[Ash-Lee233]

Models

• [ ] XCiT - #399
• [ ] beit
• [ ] cait
• [ ] coat
• [ ] convmixer
• [ ] convnext
• [ ] nest
• [ ] crossvit
• [ ] dpn
• [ ] efficientnet
• [ ] efficientnetv2
• [ ] ghostnet
• [x] googlenet - #397
• [ ] halonet
• [ ] hrnet
• [ ] inception_v3 - #397
• [ ] inception_v4 - #397
• [ ] levit
• [ ] mae
• [ ] mlpmixer
• [ ] mobilevit
• [ ] mixnet
• [ ] nasnet
• [ ] nfnet
• [ ] pit
• [ ] pvt - #271
• [ ] pvtv2 - #327
• [ ] pnasnet
• [ ] repvgg
• [ ] resnest
• [ ] resnetv2
• [ ] resnext
• [ ] senet
• [ ] sknet
• [ ] swin_transformer
• [ ] swin_transformerv2
• [ ] vgg
• [ ] volo
• [ ] xception - #397

[Ash-Lee233]

Models

• [ ] beitv2
• [ ] CoAtNet
• [ ] convnextv2
• [ ] cspnet
• [ ] deit
• [ ] dla
• [ ] hardcorenas
• [ ] inception_resnet_v2
• [ ] maxvit
• [ ] mobilevit-v2
• [ ] simclr
• [ ] tnt
• [ ] tresnet
• [ ] twins
• [ ] CLIP
• [ ] ViTDet
• [ ] vovnet
• [ ] wideresnet

[Ash-Lee233]

Augmentation

• [ ] AugMix
• [ ] 3-Augment
• [ ] TrivialAugmentation
• [ ] Repeated Augment
• [ ] TokenMix

[Ash-Lee233]

Loss

• [ ] JSD Loss

[Ash-Lee233]

Regularization

• [x] Gradient Clip (https://github.com/mindspore-lab/mindcv/pull/469)
• [ ] Adaptive Gradient Clip

[Ash-Lee233]

Optimizer

• [ ] AdamP

[Ash-Lee233]

Scheduler

• [ ] cyclic warmup cosine annealing
• [ ] MultiStep learning rate decay