-
Revert to last working version (verbose logs + VAE + simple MSE loss — before v3/v4 broke training)
-
Revert to v4 epoch-based loop + fix streaming repeat + fix scheduler + remove data std norm
-
v4.1: Fix training loop — step-based (not epochs), auto-cycle dataset, fix streaming exhaust, fix scheduler warning
-
v4: Add Min-SNR-γ + velocity direction loss + CCA + multi-scale loss + gate bias (DeepSeek/FasterDiT/DiCo/DiMR research)
-
v3: Add large anime/art datasets, cosine-with-restarts schedule, 2x LR, smart warmup, streaming support, resume training
-
Add verbose training logs: ETA, loss trend, speed, VRAM, grad norm, epoch summaries
-
Optimize notebook: 40% faster CfC blocks, simplified spatial mix
-
Optimize: remove redundant 7x7 convs from CfC heads, simplify spatial mix (40% faster CfC, 60% fewer large convs)
-
Update README with VAE integration and verified datasets
-
v2: Add VAE latent training, fix datasets, streaming support
-
Update README with VAE and verified datasets
-
V2: Add VAE, fix datasets, streaming, precaching
-
Upload liquid_diffusion/trainer.py
-
Upload validate.py
-
Upload README.md
-
Upload LiquidDiffusion_Training.ipynb
-
Add comprehensive README
-
Add Colab training notebook
-
Fix GroupNorm for arbitrary channel counts
-
Upload test_model.py
-
Upload liquid_diffusion/__init__.py
-
Upload liquid_diffusion/trainer.py
-
Upload liquid_diffusion/model.py
-
initial commit
