GitHub - vwrj/CPC: PyTorch implementation of Data-Efficient Image Recognition with Contrastive Predictive Coding

Summary: CPC encodes the input into a sequence of representations, then predicts future observations from past ones.

Name		Name	Last commit message	Last commit date
Latest commit History 11 Commits
imgs		imgs
README.md		README.md
imagenette- frozen, 24_transform_supervised_ bn, val shuffle True.ipynb		imagenette- frozen, 24_transform_supervised_ bn, val shuffle True.ipynb
imagenette- frozen, batch_norm, validation shuffle True.ipynb		imagenette- frozen, batch_norm, validation shuffle True.ipynb
imagenette- frozen, bn, 2048 -> 512 -> 25 -> 50 -> 50 (7, bs 48).ipynb		imagenette- frozen, bn, 2048 -> 512 -> 25 -> 50 -> 50 (7, bs 48).ipynb
self_supervised imagenette, 2048 to 16, no normalization (no 0.485).ipynb		self_supervised imagenette, 2048 to 16, no normalization (no 0.485).ipynb
self_supervised imagenette, 2048 to 7 (bs 48) data_aug_split.ipynb		self_supervised imagenette, 2048 to 7 (bs 48) data_aug_split.ipynb
self_supervised imagenette, no normalization (no 0.485).ipynb		self_supervised imagenette, no normalization (no 0.485).ipynb
self_supervised.ipynb		self_supervised.ipynb

Provide feedback