-
v2: Updated notebook with real dataset training
-
v2: Updated README with Real Mamba SSM and real dataset info
-
v2: Training with real datasets, Mamba SSM, push-to-hub
-
v2: Real Mamba SSM backbone (pure PyTorch), fixes torch._utils error
-
v1.4: replace SSM scan with parallel causal linear attn, batch wavelet subbands, use F.scaled_dot_product_attention, fix AMP on CPU
-
v1.4: replace SSM scan with parallel causal linear attn, batch wavelet subbands, use F.scaled_dot_product_attention, fix AMP on CPU
-
v1.3: vectorized SSM scan, fix NaN, all 5 stage cells — ArtFlow_Training.ipynb
-
v1.3: vectorized SSM scan, fix NaN, all 5 stage cells — artflow_model.py
-
Update README for v1.2 — clean repo structure
-
Remove old scripts (superseded by artflow_train.py)
-
Rebuild notebook: imports from repo scripts, no code duplication
-
Add training script with modern APIs (no deprecations)
-
Update README with complete training notebook and stability mechanisms
-
Add standalone training stability test (100-step validation)
-
Add complete Colab-friendly training notebook with stability mechanisms
-
Add Stage 1 training script (Colab/Kaggle compatible)
-
Add validated PyTorch prototype implementation
-
Add complete architecture specification (1000+ lines)
-
Add ArtFlow architecture specification
-
initial commit