People say doing a stuff for 10,000 hours makes you an expert in it. This is my 10,000 hours.
- Convert existing to tensorflow and proceed in tensorflow and pytorch
- Do RNN for sent analysis on Imdb only, rest in transformers
RNN-GRU-LSTM Sequence-to-sequence Attention Mechanism Transformers Gato
Image:
4. YoLo
5. EfficientNet
6. U-Net
7. Nets on COCO Dataset
8. Mask-RCNN
9. ViT
NLP: \
- Transformers
- Text Generation / Summarization
- Machine Translation
- Handwriting recognition
- Image Captioning
- Tinyvoice (Whispernet)
- Generate Music?
Generation:
- AE
- VAE
- GANs
- Stable Diffusion
RL:
- Basic RL
- Advanced RL
- Super Advanced RL
Other Topics:
- Meta Learning
- ANML
- RL Data Generation
- Continual Learning
Image: \
- LeNet - 5
- AlexNet
- Residual Learning for Image Recognition
- Sentiment Analysis - IMDB - LSTM