fix(profiling): use bf16 dtype and gradient checkpointing for pi0/pi05

Enable --policy.dtype=bfloat16 and --policy.gradient_checkpointing=true
for pi0, pi0_fast, and pi05 profiling specs. Combined with use_amp=true,
this brings the 4B-param VLA models well within the 22GB GPU budget.

Made-with: Cursor
This commit is contained in:
Pepijn
2026-04-16 15:35:25 +02:00
parent dbe01b0444
commit a4544ffea7
+6
View File
@@ -58,8 +58,10 @@
"--dataset.episodes=[0]",
"--policy.path=lerobot/pi0_base",
"--policy.device=cuda",
"--policy.dtype=bfloat16",
"--policy.n_action_steps=30",
"--policy.use_amp=true",
"--policy.gradient_checkpointing=true",
"--batch_size=1",
"--rename_map={\"observation.images.front\": \"observation.images.base_0_rgb\", \"observation.images.wrist\": \"observation.images.left_wrist_0_rgb\"}"
]
@@ -71,8 +73,10 @@
"--dataset.episodes=[0]",
"--policy.path=lerobot/pi0fast-base",
"--policy.device=cuda",
"--policy.dtype=bfloat16",
"--policy.n_action_steps=30",
"--policy.use_amp=true",
"--policy.gradient_checkpointing=true",
"--batch_size=1",
"--rename_map={\"observation.images.front\": \"observation.images.base_0_rgb\", \"observation.images.wrist\": \"observation.images.left_wrist_0_rgb\"}"
]
@@ -84,8 +88,10 @@
"--dataset.episodes=[0]",
"--policy.path=lerobot/pi05_base",
"--policy.device=cuda",
"--policy.dtype=bfloat16",
"--policy.n_action_steps=30",
"--policy.use_amp=true",
"--policy.gradient_checkpointing=true",
"--batch_size=1",
"--policy.normalization_mapping={\"ACTION\": \"MEAN_STD\", \"STATE\": \"MEAN_STD\", \"VISUAL\": \"IDENTITY\"}",
"--rename_map={\"observation.images.front\": \"observation.images.base_0_rgb\", \"observation.images.wrist\": \"observation.images.left_wrist_0_rgb\"}"