Key errors when running with nuScenes dataset

[purewater0901]

Due to the change in the latest commit, I encountered the following errors. Is it okay to remove all_other_agents_curr_speed from the batch keys?

This is my error

Traceback (most recent call last):
  File "scripts/evaluate.py", line 428, in <module>
    run_evaluation(
  File "scripts/evaluate.py", line 147, in run_evaluation
    stats, info, renderings, adjust_plans, trace = rollout_episodes(
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/env_utils.py", line 286, in rollout_episodes
    action = policy.get_action(obs_torch, step_index=counter)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 282, in get_action
    agents_action, agents_action_info = self.agents_policy.get_action(
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 237, in get_action
    action, action_info = self.policy.get_action(obs, **kwargs)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 178, in get_action
    return self.model.get_action(obs, **self.action_kwargs, **kwargs)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 107, in get_action
    _, action_info = self.sampler.get_action(obs)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 64, in get_action
    actions_tiled, _ = self.controller.get_action(
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/algos/multiagent_algos.py", line 134, in get_action
    preds = self.model.forward_prediction(feats_tiled, obs_tiled, plan=plan_tiled)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/models/multiagent_models.py", line 334, in forward_prediction
    dyn_states = batch_utils().get_current_states_all_agents(
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/batch_utils.py", line 265, in get_current_states_all_agents
    state_all = trajdataBatchUtils.batch_to_raw_all_agents(batch, step_time)
  File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/batch_utils.py", line 209, in batch_to_raw_all_agents
    data_batch["all_other_agents_curr_speed"]
KeyError: 'all_other_agents_curr_speed'

[chenyx09]

Hi Yutaka,

It seems that the key was removed by accident in the last commit. Is your problem solved after you add it back?

Best, Yuxiao

On Sun, Oct 8, 2023 at 8:53 PM Yutaka Shimizu @.***> wrote:

Due to the change in the latest commit, I encountered the following errors. Is it okay to remove all_other_agents_curr_speed from the batch keys?

This is my error

Traceback (most recent call last): File "scripts/evaluate.py", line 428, in run_evaluation( File "scripts/evaluate.py", line 147, in run_evaluation stats, info, renderings, adjust_plans, trace = rollout_episodes( File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/env_utils.py", line 286, in rollout_episodes action = policy.get_action(obs_torch, step_index=counter) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 282, in get_action agents_action, agents_action_info = self.agents_policy.get_action( File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 237, in get_action action, action_info = self.policy.get_action(obs, kwargs) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 178, in get_action return self.model.get_action(obs, self.action_kwargs, **kwargs) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 107, in getaction , action_info = self.sampler.get_action(obs) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/policies/wrappers.py", line 64, in get_action actionstiled, = self.controller.get_action( File "/home/yutaka/nvidia_simulator/tbsim/tbsim/algos/multiagent_algos.py", line 134, in get_action preds = self.model.forward_prediction(feats_tiled, obs_tiled, plan=plan_tiled) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/models/multiagent_models.py", line 334, in forward_prediction dyn_states = batch_utils().get_current_states_all_agents( File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/batch_utils.py", line 265, in get_current_states_all_agents state_all = trajdataBatchUtils.batch_to_raw_all_agents(batch, step_time) File "/home/yutaka/nvidia_simulator/tbsim/tbsim/utils/batch_utils.py", line 209, in batch_to_raw_all_agents data_batch["all_other_agents_curr_speed"] KeyError: 'all_other_agents_curr_speed'

[image: image] https://user-images.githubusercontent.com/43805014/273494864-13809df7-1c3c-4d42-8345-d81f3e071d25.png

— Reply to this email directly, view it on GitHub https://github.com/NVlabs/traffic-behavior-simulation/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC5LEZWRM563RDOWUYAAGUDX6NYNDAVCNFSM6AAAAAA5YFZGFWVHI2DSMVQWIX3LMV43ASLTON2WKOZRHEZTEMRVGY3TSNA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[purewater0901]

Hello Yuxiao

Yes, after I added it manually, it works without any issues.