Add model implementations #112
Description
Below are models that we still need to implement
Will implement
These are models for which we are actively seeking implementations/weights.
Model implementations
- VGG
- ResNet
- DenseNet
- SqueezeNet
- GoogLeNet
- Inception v3
- Inception v4 (see Implementation of
Inceptionv4,InceptionResNetv2andXception#170) - Inception ResNet v2 (see Implementation of
Inceptionv4,InceptionResNetv2andXception#170) - Xception (see Implementation of
Inceptionv4,InceptionResNetv2andXception#170) - ResNeXt
- MobileNet v1
- MobileNet v2
- MLPMixer
- Vision Transformer (see Implementation of base ViT model #105)
- EfficientNet (see Implement EfficientNet #113 Implement EfficientNet #171)
- EfficientNet v2
- CaiT
- LeViT
- Swin
- MobileViT
- ConvNeXt (see Implementation for ConvNeXt #119)
- UNet (see Adding UNet Model #210)
- DeiT
Pre-trained weights
- VGG (see Add pre-trained models from HuggingFace and add accuracy testing #164)
- ResNet (see Add pre-trained models from HuggingFace and add accuracy testing #164)
- DenseNet
- SqueezeNet
- GoogLeNet
- Inception v3
- Inception v4
- Inception ResNet v2
- Xception
- ResNeXt
- MobileNet v1
- MobileNet v2
- MLPMixer
- Vision Transformer
- EfficientNet
- EfficientNet v2
- CaiT
- LeViT
- Swin
- MobileViT
- ConvNeXt
- UNet
- DeiT
Will consider implementing
These are models that aren't necessarily must haves, but we are happy to accept contributions that add them.
- ESRGAN (see Implement ESRGAN #118).