issues for ruunning train.py with TD3

[faker52]

nohup: ignoring input [robosuite WARNING] No private macro file found! (init.py:7) [robosuite WARNING] It is recommended to use a private macro file (init.py:8) [robosuite WARNING] To setup, run: python /home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/robosuite/scripts/setup_macros.py (init.py:9) No personal conf_private.py found. doodad not detected

------------- Running TD3 -------------- Params: variant: scripts/variantTD3Pnp.json

2023-11-07 18:43:56.892015 CST | Variant: 2023-11-07 18:43:56.892349 CST | { "algorithm": "TD3", "algorithm_kwargs": { "batch_size": 128, "eval_max_path_length": 500, "expl_max_path_length": 500, "min_num_steps_before_training": 3300, "num_epochs": 2000, "num_eval_steps_per_epoch": 2500, "num_expl_steps_per_train_loop": 2500, "num_trains_per_train_loop": 1000 }, "eval_environment_kwargs": { "control_freq": 20, "controller": "OSC_POSE", "env_name": "Lift", "hard_reset": false, "horizon": 500, "ignore_done": true, "reward_scale": 1.0, "robots": "['Panda']" }, "expl_environment_kwargs": { "control_freq": 20, "controller": "OSC_POSE", "env_name": "PickPlaceCan", "hard_reset": false, "horizon": 500, "ignore_done": true, "reward_scale": 1.0, "robots": "['Panda']" }, "policy_kwargs": { "hidden_sizes": [ 256, 256 ] }, "qf_kwargs": { "hidden_sizes": [ 256, 256 ] }, "replay_buffer_size": 1000000, "seed": 17, "trainer_kwargs": { "discount": 0.99, "reward_scale": 1.0 }, "version": "normal" } /home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/gym/spaces/box.py:127: UserWarning: WARN: Box bound precision lowered by casting to float32 logger.warn(f"Box bound precision lowered by casting to {self.dtype}") Traceback (most recent call last): File "scripts/train.py", line 131, in run_experiment() File "scripts/train.py", line 104, in run_experiment experiment(variant, agent=args.agent) File "/data1/qxx/RL/Change-maple/bseline/robosuite-benchmark/util/rlkit_utils.py", line 163, in experiment algorithm.train() File "/data1/qxx/RL/Change-maple/bseline/robosuite-benchmark/util/rlkit_custom.py", line 46, in train self._train() File "/data1/qxx/RL/Change-maple/bseline/robosuite-benchmark/util/rlkit_custom.py", line 213, in _train self.eval_data_collector.collect_new_paths( File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/samplers/data_collector/path_collector.py", line 42, in collect_new_paths path = rollout( File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/samplers/rollout_functions.py", line 112, in rollout a, agent_info = agent.get_action(o) File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/torch/networks.py", line 111, in get_action actions = self.get_actions(obs_np[None]) File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/torch/networks.py", line 115, in get_actions return eval_np(self, obs) File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/torch/core.py", line 18, in eval_np outputs = module(*torch_args, torch_kwargs) File "/home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "/home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl return forward_call(args, kwargs) File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/torch/networks.py", line 108, in forward return super().forward(obs, kwargs) File "/data1/qxx/RL/Change-maple/bseline/rlkit/rlkit/torch/networks.py", line 69, in forward h = fc(h) File "/home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl return self._call_impl(*args, *kwargs) File "/home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl return forward_call(args, kwargs) File "/home/qxx/anaconda3/envs/robosuite/lib/python3.8/site-packages/torch/nn/modules/linear.py", line 114, in forward return F.linear(input, self.weight, self.bias) RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x42 and 46x256)

[amandlek]

We need more information to reproduce the issue. Can you give us the precise training command, config being used, and information on the dataset?