BC #02: MMOneSphere, ntrain{100,500}, PNet2_eepose (4DoF demonstrator)
On the newer dataset now regressing straight to `eepose` so it's 6DoF axis angle, and then for test time we end up clearing out the x and z axis for the rotation so it's just the y axis. Looks bad which is good (well, assuming our other methods do better). NOTE: this hack could be an issue because the delta (amount of rotation) depends on the magnitude of the axis-angles. Of course, again the MSEs all had x and z as 0 so if it predicts something there it isn't learning correctly...
Created on April 28|Last edited on May 2
Comment
Will test 100 and 500 demo cases.
Eval Success Rate and MSEs
Run set
2
Example GIFs?
NOTE: number of eval episodes is 4 instead of 10 (oops sorry). Not a big deal here though. Seems like it's producing the same bad policies. It's also not really learning how to rotate at all. I actually wonder if this is a harder way to do rotations in favor of a more natural scooping parameterization or something that rotates just to bring the ball towards the tip?
First seed after 250 epochs.
Second seed -- seems to have crashed.
Third seed after 250 epochs.
Variant?
Remember this one has clipping of the y axis, which I feel like is going to be a problem. Yes this has translation_axis_angle ... and uses v04 of the demonstrator, yes. It does use a weighted MSE to increase the weight on the rotation part.
{"_hidden_keys": [],"act_type": "eepose","actor_lr": 0.0001,"agent": "bc","alg_policy": "ladle_algorithmic_v04","algorithm": "BC","batch_size": 24,"bc_data_dir": "/data/dseita/softgym_mm/data_demo/MMOneSphere_v01_BClone_filtered_ladle_algorithmic_v04_nVars_2000_obs_combo_act_translation_axis_angle","bc_data_filtered": true,"data_buffer_capacity": 1000000,"encoder_type": "pointnet","env_kwargs": {"action_mode": "translation","action_repeat": 8,"camera_name": "top_down","deterministic": false,"headless": true,"horizon": 100,"num_variations": 1000,"observation_mode": "cam_rgb","render": true,"render_mode": "fluid"},"env_kwargs_action_mode": "translation_axis_angle","env_kwargs_camera_height": 128,"env_kwargs_camera_width": 128,"env_kwargs_deterministic": false,"env_kwargs_num_variations": 2000,"env_kwargs_observation_mode": "point_cloud","env_name": "MMOneSphere","env_version": "v01","exp_name": "MMOneSphere_v01_BC_ntrain_0100_PCL_PNet2_eepose_2022_04_27_20_15_02_0001","hidden_dim": 1024,"lambda_pos": 1.0,"lambda_rot": 100.0,"log_interval": 1,"n_epochs": 250,"n_eval_episodes": 4,"n_train_demos": 100,"num_filters": 32,"num_layers": 4,"project_axis_ang_y": true,"save_freq": 10,"save_model": true,"save_video": true,"seed": 100,"wandb_entity": "mooey5775","wandb_project": "mixed_media","weighted_MSE": true}
Add a comment