Ctrl+K

instance run 2026-05-30: liger 0.8.0 workarounds + journal

14573d1 about 1 month ago

__init__.py

210 Bytes
v1.5 pivot: 152M (18L x 768d) hero config, ReLU2 FFN, final logit soft-cap. 350M kept as reference. about 1 month ago
checkpoint.py

4.05 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) about 1 month ago
config.py

5.16 kB
v1.5 pivot: 152M (18L x 768d) hero config, ReLU2 FFN, final logit soft-cap. 350M kept as reference. about 1 month ago
data.py

8.05 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) about 1 month ago
model.py

13.1 kB
instance run 2026-05-30: liger 0.8.0 workarounds + journal about 1 month ago
monitor.py

2.91 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) about 1 month ago
optim.py

10.5 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) about 1 month ago
train.py

12.2 kB
v2: 363M hero run (Muon hybrid, WSD, Liger, SmolLM 75/15/10 mix) about 1 month ago