All
pi
Molmo
Newest ↓
↺
Scanning R2 checkpoints…
▶ Fine-tune π0.5
⋯
×
Delete dataset
×
How it trained
Training loss
Trained on
Delete episode
×
loading episode…
‹ ep
⏸
ep ›
Train π0.5
×
GPUs
Hyperparameters
(auto-set from GPU count — edit any to override)
batch size (global)
train steps
peak lr
warmup steps
save every
auto-stop (hrs)
Runs top → bottom · you advance to the next policy manually · loop blocks repeat a policy
×
Policies — drag in →
Chain
×
Cancel
Delete