Fixes following #670 (#719)

这个提交包含在:
Simon Alibert
2025-02-12 12:53:55 +01:00
提交者 GitHub
父节点 90e099b39f
当前提交 e71095960f
共有 3 个文件被更改,包括 8 次插入 和 7 次删除

查看文件

@ -300,7 +300,7 @@ class PI0Policy(PreTrainedPolicy):
self._action_queue.extend(actions.transpose(0, 1)) self._action_queue.extend(actions.transpose(0, 1))
return self._action_queue.popleft() return self._action_queue.popleft()
def forward(self, batch: dict[str, Tensor], noise=None, time=None) -> dict[str, Tensor]: def forward(self, batch: dict[str, Tensor], noise=None, time=None) -> tuple[Tensor, dict[str, Tensor]]:
"""Do a full training forward pass to compute the loss""" """Do a full training forward pass to compute the loss"""
if self.config.adapt_to_pi_aloha: if self.config.adapt_to_pi_aloha:
batch[OBS_ROBOT] = self._pi_aloha_decode_state(batch[OBS_ROBOT]) batch[OBS_ROBOT] = self._pi_aloha_decode_state(batch[OBS_ROBOT])
@ -328,12 +328,12 @@ class PI0Policy(PreTrainedPolicy):
losses = losses[:, :, : self.config.max_action_dim] losses = losses[:, :, : self.config.max_action_dim]
loss_dict["losses_after_rm_padding"] = losses.clone() loss_dict["losses_after_rm_padding"] = losses.clone()
loss = losses.mean()
# For backward pass # For backward pass
loss_dict["loss"] = loss loss = losses.mean()
# For logging # For logging
loss_dict["l2_loss"] = loss.item() loss_dict["l2_loss"] = loss.item()
return loss_dict
return loss, loss_dict
def prepare_images(self, batch): def prepare_images(self, batch):
"""Apply Pi0 preprocessing to the images, like resizing to 224x224 and padding to keep aspect ratio, and """Apply Pi0 preprocessing to the images, like resizing to 224x224 and padding to keep aspect ratio, and

查看文件

@ -102,7 +102,7 @@ class WandBLogger:
self._wandb.log_artifact(artifact) self._wandb.log_artifact(artifact)
def log_dict(self, d: dict, step: int, mode: str = "train"): def log_dict(self, d: dict, step: int, mode: str = "train"):
if mode in {"train", "eval"}: if mode not in {"train", "eval"}:
raise ValueError(mode) raise ValueError(mode)
for k, v in d.items(): for k, v in d.items():
@ -114,7 +114,7 @@ class WandBLogger:
self._wandb.log({f"{mode}/{k}": v}, step=step) self._wandb.log({f"{mode}/{k}": v}, step=step)
def log_video(self, video_path: str, step: int, mode: str = "train"): def log_video(self, video_path: str, step: int, mode: str = "train"):
if mode in {"train", "eval"}: if mode not in {"train", "eval"}:
raise ValueError(mode) raise ValueError(mode)
wandb_video = self._wandb.Video(video_path, fps=self.env_fps, format="mp4") wandb_video = self._wandb.Video(video_path, fps=self.env_fps, format="mp4")

查看文件

@ -233,7 +233,7 @@ def train(cfg: TrainPipelineConfig):
logging.info(train_tracker) logging.info(train_tracker)
if wandb_logger: if wandb_logger:
wandb_log_dict = {**train_tracker.to_dict(), **output_dict} wandb_log_dict = {**train_tracker.to_dict(), **output_dict}
wandb_logger.log_dict(wandb_log_dict) wandb_logger.log_dict(wandb_log_dict, step)
train_tracker.reset_averages() train_tracker.reset_averages()
if cfg.save_checkpoint and is_saving_step: if cfg.save_checkpoint and is_saving_step:
@ -271,6 +271,7 @@ def train(cfg: TrainPipelineConfig):
logging.info(eval_tracker) logging.info(eval_tracker)
if wandb_logger: if wandb_logger:
wandb_log_dict = {**eval_tracker.to_dict(), **eval_info} wandb_log_dict = {**eval_tracker.to_dict(), **eval_info}
wandb_logger.log_dict(wandb_log_dict, step, mode="eval")
wandb_logger.log_video(eval_info["video_paths"][0], step, mode="eval") wandb_logger.log_video(eval_info["video_paths"][0], step, mode="eval")
if eval_env: if eval_env: