An Issues about RuntimeError

[hanninaa]

`

| time/ | |
| fps | 16 |
| iterations | 1 |
| time_elapsed | 200 |
| total_timesteps | 3355 |

Traceback (most recent call last):
File "scripts/training/train_text_generation.py", line 84, in
main(
File "scripts/training/train_text_generation.py", line 55, in main
trainer.train_and_eval()
File "/root/autodl-tmp/RL4LMs/rl4lms/envs/text_generation/training_utils.py", line 205, in train_and_eval
self._alg.learn(self._n_steps_per_iter)
File "/root/autodl-tmp/RL4LMs/rl4lms/algorithms/ppo/ppo.py", line 342, in learn
return super().learn(
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/stable_baselines3/common/on_policy_algorithm.py", line 267, in learn
self.train()
File "/root/autodl-tmp/RL4LMs/rl4lms/algorithms/ppo/ppo.py", line 215, in train
evaluation_output: EvaluateActionsOutput = self.policy.evaluate_actions(
File "/root/autodl-tmp/RL4LMs/rl4lms/envs/text_generation/policy/seq2seq_policy.py", line 242, in evaluate_actions
policy_outputs = self.forward_policy(obs=obs, actions=actions)
File "/root/autodl-tmp/RL4LMs/rl4lms/envs/text_generation/policy/seq2seq_policy.py", line 116, in forward_policy
past_model_kwargs = unwrap_model(
File "/root/autodl-tmp/RL4LMs/rl4lms/envs/text_generation/hf_generation_utils.py", line 534, in _prepare_encoder_decoder_kwargs_for_generation
model_kwargs["encoder_outputs"]: ModelOutput = encoder(
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/transformers/models/t5/modeling_t5.py", line 1033, in forward
layer_outputs = layer_module(
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/transformers/models/t5/modeling_t5.py", line 668, in forward
self_attention_outputs = self.layer[0](
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/transformers/models/t5/modeling_t5.py", line 574, in forward
attention_output = self.SelfAttention(
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "/root/miniconda3/envs/rl/lib/python3.8/site-packages/transformers/models/t5/modeling_t5.py", line 511, in forward
scores = torch.matmul(
RuntimeError: CUDA out of memory. Tried to allocate 768.00 MiB (GPU 0; 31.75 GiB total capacity; 28.32 GiB already allocated; 327.94 MiB free; 30.22 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

When` I try the example, all the environment has done, and I use V100-32G. It continues to show the error about RuntimeError. I changed the.yml file pyremeter but still see this bug. So I want to ask how to solve the problem, or have I done something wrong?