Skip to content
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 10 additions & 8 deletions xtuner/v1/train/rl_trainer.py
Original file line number Diff line number Diff line change
Expand Up @@ -1013,17 +1013,19 @@ def _train_one_batch(

# fit内部再做worker的onload,ppo onload critic, 其他的onload actor
with timer("training", step_timer_dict):
workers_log_item, data_info = self.train_controller.fit(
train_batch,
pack_max_length=self._train_worker_cfg.pack_max_length,
rollout_idx=train_step,
raw_rewards_sum=raw_rewards_sum,
raw_rewards_count=raw_rewards_count,
)
try:
workers_log_item, data_info = self.train_controller.fit(
train_batch,
pack_max_length=self._train_worker_cfg.pack_max_length,
rollout_idx=train_step,
raw_rewards_sum=raw_rewards_sum,
raw_rewards_count=raw_rewards_count,
)
finally:
self._release_train_batch_resources(train_batch)

data_info["invalid_train_groups"] = dropped_groups
data_info["invalid_train_samples"] = dropped_samples
self._release_train_batch_resources(train_batch)

return {
"data_info": data_info,
Expand Down
Loading