Commit History

Fix: use merged_model_state.pt (LoRA baked into base weights, fixes missing FFN weights)
13790f5
verified

LisaMegaWatts commited on

Fix config: n_layers=8, n_monarch_heads=1, use progressive model
c0f06ec
verified

LisaMegaWatts commited on

Upload server.py with huggingface_hub
36991b4
verified

LisaMegaWatts commited on

Upload server.py with huggingface_hub
2576328
verified

LisaMegaWatts commited on