Using a pre-trained model

[adhocmaster]

Hi there,

I would like to test the pre-trained model against pedestrian models that I have been developing. I see there is a weight folder but still not sure how to create an agent and load the weights. Can you guide me?

[dotchen]

Thank you for your interest in our project.

Please follow the instructions here on how to directly run an agent with pre-trained model weights. Please make sure to have installed git-lfs before cloning the repo.

[adhocmaster]

Thanks for the prompt reponse.

[varunjammula]

Hi @dotchen, when are you planning on releasing the Leaderboard weights for model evaluation? Thanks in advance.

[dotchen]

I will try to find sometime to refactor and release it. Hopefully around June.

[ricokoff]

Hi @dotchen,

Are there any updates on the leaderboard weights?

[xolovezarii]

Thank you for your interest in our project.

Please follow the instructions here on how to directly run an agent with pre-trained model weights. Please make sure to have installed git-lfs before cloning the repo.

Hi @dotchen I followed what said in the repo but still don't know what should I do with "wights" folder!? I run the project and it' output is like this:

Running the route ======[Agent] Wallclock_time = 2024-03-10 08:13:24.232159 / 0.0 / Sim_time = 0.05000000074505806 / 50.00000074505806x ======[Agent] Wallclock_time = 2024-03-10 08:13:25.015090 / 0.782931 / Sim_time = 0.10000000149011612 / 0.12756224908839695x

Stopping the route, the agent has crashed:

Not compiled with CUDA support

Traceback (most recent call last): File "/home/missakhbariun/LAV/leaderboard/leaderboard/scenarios/scenario_manager.py", line 152, in _tick_scenario ego_action = self._agent() File "/home/missakhbariun/LAV/leaderboard/leaderboard/autoagents/agent_wrapper.py", line 75, in call return self._agent() File "/home/missakhbariun/LAV/leaderboard/leaderboard/autoagents/autonomous_agent.py", line 115, in call control = self.run_step(input_data, timestamp) File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context return func(*args, kwargs) File "/home/missakhbariun/LAV/team_code_v2/lav_agent.py", line 318, in run_step pred_bev = self.lidar_model([lidar_points], [len(lidar_points)]) File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl return forward_call(*input, *kwargs) File "/home/missakhbariun/LAV/team_code_v2/models/lidar.py", line 37, in forward features = self.point_pillar_net(lidars, num_points) File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl return forward_call(input, kwargs) File "/home/missakhbariun/LAV/team_code_v2/models/point_pillar.py", line 114, in forward features = self.point_net(decorated_points, inverse_indices) File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl return forward_call(*input, *kwargs) File "/home/missakhbariun/LAV/team_code_v2/models/point_pillar.py", line 33, in forward feat_max = scatter_max(feat, inverse_indices, dim=0)[0] File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch_scatter/scatter.py", line 72, in scatter_max return torch.ops.torch_scatter.scatter_max(src, index, dim, out, dim_size) File "/home/missakhbariun/miniconda3/envs/LAV-env/lib/python3.7/site-packages/torch/_ops.py", line 442, in call return self._op(args, **kwargs or {}) RuntimeError: Not compiled with CUDA support

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "/home/missakhbariun/LAV/leaderboard/leaderboard_evaluator.py", line 352, in _load_and_run_scenario self.manager.run_scenario() File "/home/missakhbariun/LAV/leaderboard/leaderboard/scenarios/scenario_manager.py", line 136, in run_scenario self._tick_scenario(timestamp) File "/home/missakhbariun/LAV/leaderboard/leaderboard/scenarios/scenario_manager.py", line 159, in _tick_scenario raise AgentError(e) leaderboard.autoagents.agent_wrapper.AgentError: Not compiled with CUDA support

Stopping the route

========= Results of RouteScenario_25 (repetition 0) ------ FAILURE =========

╒═════════════════════════════════╤═════════════════════╕ │ Start Time │ 2024-03-07 22:05:05 │ ├─────────────────────────────────┼─────────────────────┤ │ End Time │ 2024-03-07 22:05:07 │ ├─────────────────────────────────┼─────────────────────┤ │ Duration (System Time) │ 1.97s │ ├─────────────────────────────────┼─────────────────────┤ │ Duration (Game Time) │ 0.1s │ ├─────────────────────────────────┼─────────────────────┤ │ Ratio (System Time / Game Time) │ 0.051 │ ╘═════════════════════════════════╧═════════════════════╛

╒═══════════════════════╤═════════╤═════════╕ │ Criterion │ Result │ Value │ ├───────────────────────┼─────────┼─────────┤ │ RouteCompletionTest │ FAILURE │ 0.0 % │ ├───────────────────────┼─────────┼─────────┤ │ OutsideRouteLanesTest │ SUCCESS │ 0 % │ ├───────────────────────┼─────────┼─────────┤ │ CollisionTest │ SUCCESS │ 0 times │ ├───────────────────────┼─────────┼─────────┤ │ RunningRedLightTest │ SUCCESS │ 0 times │ ├───────────────────────┼─────────┼─────────┤ │ RunningStopTest │ SUCCESS │ 0 times │ ├───────────────────────┼─────────┼─────────┤ │ InRouteTest │ SUCCESS │ │ ├───────────────────────┼─────────┼─────────┤ │ AgentBlockedTest │ SUCCESS │ │ ├───────────────────────┼─────────┼─────────┤ │ Timeout │ SUCCESS │ │ ╘═══════════════════════╧═════════╧═════════╛

Registering the route statistics Registering the global statistics

It looks like it did not use the pretrained wights and models.

what should I do??? I would be so thankful if you guys help me, too.@varunjammula @adhocmaster @ricokoff @philkr