single-shot-robot

Josselin's group workspace

Group: soft_modularization

1-10

of 10

Tags

Notes

Author

josselin

State

Finished

Start time

June 4th, 2023 12:26:31 AM

Runtime

1h 51m 2s

Tracked hours

1h 50m 57s

Run path

single-shot-robot/PTSL_MTRL10-200k/kghkm5mn

Linux-5.15.0-1031-aws-x86_64-with-glibc2.10

Python version

3.8.0

Git repository

git clone https://github.com/JosselinSomervilleRoberts/mtrl.git

Git state

git checkout -b "soft_modularization_10" ab657e27077c3651fb35e14544c2f98c68907237

Command

main.py setup=metaworld env=metaworld-mt10 agent=state_sac experiment.num_eval_episodes=1 experiment.num_train_steps=200000 setup.seed_ref=1 setup.num_seeds=10 setup.name=soft_modularization replay_buffer.batch_size=1280 agent.multitask.num_envs=10 agent.multitask.should_use_disentangled_alpha=True agent.multitask.should_use_task_encoder=True agent.encoder.type_to_select=feedforward agent.multitask.actor_cfg.should_condition_model_on_task_info=True agent.multitask.actor_cfg.should_condition_encoder_on_task_info=False agent.multitask.actor_cfg.should_concatenate_task_info_with_encoder=False agent.multitask.actor_cfg.moe_cfg.should_use=True agent.multitask.actor_cfg.moe_cfg.mode=soft_modularization agent.multitask.should_use_multi_head_policy=False agent.encoder.feedforward.hidden_dim=50 agent.encoder.feedforward.num_layers=2 agent.encoder.feedforward.feature_dim=50 agent.actor.num_layers=4 agent.multitask.task_encoder_cfg.model_cfg.pretrained_embedding_cfg.should_use=False

System Hardware

CPU count	2
Logical CPU count	4
GPU count	1
GPU type	Tesla T4

W&B CLI Version

0.15.3

Group

soft_modularization

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 21 keys
- actor/hidden_dim:
  326
- actor/num_layer:
  4
- agent:
  "state_sac"
- agent/encoder_feature_dim:
  50
- agent/num_filters:
  0
- agent/num_layers:
  0
- batch_size:
  1,280
- encoder:
  "feedforward"
- eval_freq:
  10,000
- lr/actor:
  0.0003
- lr/alpha:
  0.0003
- lr/critic:
  0.0003
- lr/decoder:
  0.0003
- lr/encoder:
  0.0003
- num_eval_episodes:
  1
- num_tasks:
  10
- num_train_steps:
  200,000
- pal/dim:
  50
- pal/residual:
  "none"
- pal/shared:
  true
- seed:
  10

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 32 keys
- Env 0/episode_reward_env_index_:
  411.7856918986021
- Env 0/success_env_index_:
  1
- Env 1/episode_reward_env_index_:
  126.91067610377964
- Env 1/success_env_index_:
  1
- Env 2/episode_reward_env_index_:
  8.112238940590132
- Env 2/success_env_index_:
  0
- Env 3/episode_reward_env_index_:
  -191.23665786157903
- Env 3/success_env_index_:
  0
- Env 4/episode_reward_env_index_:
  259.56861313477833
- Env 4/success_env_index_:
  0
- Env 5/episode_reward_env_index_:
  22.904445002928306
- Env 5/success_env_index_:
  0
- Env 6/episode_reward_env_index_:
  286.3579414263302
- Env 6/success_env_index_:
  1
- Env 7/episode_reward_env_index_:
  -10.432677001603928
- Env 7/success_env_index_:
  0
- Env 8/episode_reward_env_index_:
  393.7943212755215
- Env 8/success_env_index_:
  0
- Env 9/episode_reward_env_index_:
  104.59052700672146
- Env 9/success_env_index_:
  0
- Eval/episode:
  1,200
- Eval/episode_reward:
  174.5698589739223
- Eval/success:
  0.3
- Train/actor_entropy:
  2.676268221537272
- Train/actor_loss:
  -88.8094644165039
- Train/actor_target_entropy:
  -4
- Train/alpha_loss:
  0.0007444420487930378
- Train/batch_reward:
  1.0066827114423116
- Train/critic_loss:
  38.3063538424174
- Train/duration:
  4.970698118209839
- Train/episode:
  1,333
- Train/success:
  0.3

This run produced these artifacts as outputs. Total: 1. Learn more

wandb-history

run-kghkm5mn-history:v0