imitation learning maybe

from zuri

Uploaded by Rowan G. Last edited by Rowan G.

Training information

Algorithm
behavior_cloning
Seed
42
Progress unit
updates

The highest uploaded checkpoint may not be the final progress of the local run. Initialization-only runs do not indicate training.

Uploaded checkpoints

Related evaluations

Checkpoint evidence
[
  {
    "checkpoint_id": "0a7dedd8-3f08-4194-bd1c-c67bca473000",
    "id": "0a7dedd8-3f08-4194-bd1c-c67bca473000",
    "label": "imitation learning maybe \u00b7 1.5K updates",
    "progress": 1500,
    "progress_unit": "optimizer_updates",
    "run_id": "98f326db-874d-47d9-a5f8-7bc3326344f5",
    "run_name": "imitation learning maybe",
    "sha256": "8d086d8bd3f18e751e08974145dae825f0e18ed4b1edaed93515c7b4e474c46a",
    "source_checkpoint_id": "run:f4b9a13d-e19d-4eda-896c-6a2f7479861d:sha256:8d086d8bd3f18e751e08974145dae825f0e18ed4b1edaed93515c7b4e474c46a",
    "training_run_name": "imitation learning maybe"
  }
]
Training configuration
{
  "ACTIVATION": "relu",
  "ALGORITHM": "ippo_rnn",
  "ANNEAL_LR": false,
  "CHECKPOINT_INTERVAL_UPDATES": 1,
  "CHECKPOINT_KEEP_LAST": 2,
  "CLIP_EPS": 0.2,
  "ENT_COEF": 0.01,
  "ENV_KWARGS": {
    "agent_view_size": null,
    "layout": "oc2_sushi_1_3",
    "rules": {
      "max_steps": 600
    }
  },
  "ENV_NAME": "overcooked_v2",
  "FC_DIM_SIZE": 32,
  "GAE_LAMBDA": 0.95,
  "GAMMA": 0.99,
  "GRU_HIDDEN_DIM": 32,
  "IMITATION": {
    "action_weighting": "inverse_square_root_frequency",
    "batch_episodes": 1,
    "dataset_sha256": {
      "canonical.npz": "c43c80b0223257be89d7123782d58b7bdc0e0ec1cfc1c634f70717b6a0a0aa0d",
      "pose0-00.npz": "29a445cdc9b9865a0c6f677922c473d0c04fd6a2e25796eb1600633483ad90d5",
      "pose0-01.npz": "3d7688523b229a3bf5e63320d5e72a96e24373ac1b9cfb53f70827a0805eb000",
      "pose0-02.npz": "a51b178c57500348b240b00a036eb762410b34e66f805b8530c97eb872ef7775",
      "pose0-03.npz": "5eb5f16a44d2c93d8628d754b80ecf52d97f4a46d82bd5bb1e61f7517a57e7b1",
      "pose0-04.npz": "8eb53485c338cea2117cecf2d3c15eeeb84c57b3b596c4b3b682efb7596621bd",
      "pose0-05.npz": "feaa05ae6431a019b2054f5c64409665023ae199f32f67e2b2a2309dac00e288",
      "pose0-06.npz": "96ab142c125ee43eef1bf4b23116c88415b4b5fc0507c7b362a6ebcd41c711cd",
      "pose0-07.npz": "dff1d227b532aba679e7a5ac01bdc20c8db1fb960a35d1111313f9cbe7a29a3d",
      "pose0-08.npz": "8b7342fc264283b0d182f976ff21f93393e75ab8dec81252a49c7338dc7a9bd6",
      "pose0-09.npz": "d66594473a94d95719e039bfa26cdb1a8210d69053cfbea4344b41299e511fd7",
      "pose0-10.npz": "b6a4eb2e4cebde8b6c1514f047eeccdbf882974510885b6c6801dff581e37281",
      "pose0-11.npz": "ef220c5502876bba699e4790cb58c098967854c0f697c5a820e1e7dc2eac668e",
      "pose0-12.npz": "7ad758be05f9b10b9e34396243e6644c0ec9c9a0bb9cb733603b6774c7c32c62",
      "pose0-13.npz": "c0e4eb55383b6158755ea498b9d0ded83316fd9033d135ba339ffff2ab22b147",
      "pose0-14.npz": "54e6a24b3791828692cc3e21deb7ce029aaf87360360c43520cfc0459a5199f2",
      "pose0-15.npz": "ee279d78930fb239c1d06fae4ad3da4c87d2e3af6c16a69637724c9e1a03a40a",
      "pose0-16.npz": "9a167d108e53dac0b15bd45b42145a149f5e43cde3894c0e61eb0f3a2b85a0f0",
      "pose0-17.npz": "5f83224e0be6c9966809be9b4f973476994204c3b9baebb8d3f6e690b6bcb742",
      "pose0-18.npz": "5a3495023d2fd267779b5bf89142cbf6bcdd083ae383454ce139562415d44ed2",
      "pose0-19.npz": "d7a2a2aa56e89243cf3f50454402c460c911e640dca77659d7bbd3b77f9dd462",
      "pose1-00.npz": "017ceefe3ae333c397b8a075bd87491d821e3d9b3a0dccddb0edb0a031e195a2",
      "pose1-01.npz": "abd2b5edc7ff7edd902ea8cd762b32810206ad056d6afda6bc1063594454140b",
      "pose1-02.npz": "a648f0fb7f9321e870a6fed2d6bad7db1f5bc5f065f017c61269bb5c3c60c22c",
      "pose1-03.npz": "ed02b6d3755b4fb6617d5e4a24997c7b39257d06e3d1bed2e7c22f74951e01a5",
      "pose1-04.npz": "44ac4cedab0ace39fe6dd0795b27aa86fea36b7941d95b0190b5c69ae9a326f3",
      "pose1-05.npz": "41385e854f3140d6eb7b2e31d52641bd3087f2fdbeb21b920e6ecc32569bf0eb",
      "pose1-06.npz": "9548f912f6a0a10c7e6f2cb2538bf50522ed99bb99acc959b131c146aa97a9d7",
      "pose1-07.npz": "b3125552a4be2d29174f7515eda41c1e9a9fd8b0b2f0dc6b59b678e4572246a7",
      "pose1-08.npz": "07030c7a4b63c6a781ad8cd23690006beb6c12037608289f1a4589b72fcc9876",
      "pose1-09.npz": "66b0b11848429ac9c2cf6267068726ef7d2dfe4d0eddb658b72b3bc62d13c728",
      "pose1-10.npz": "795d4b265192b772ba0e5af4c48021e3b511be1fe8389d55df63466fc435da68",
      "pose1-11.npz": "c4b74179dca68c7d8a2ffd5fb4c877658610dea0beaff38c38b5d64568f5d592",
      "pose1-12.npz": "2965b3b192fe6fc94c21ae3d5fe6691ea5b5f2b5e49d952e384315c9dd09916a",
      "pose1-13.npz": "f37121d6fc10571c585d9bf9148c729a953eab5b73804e9c9911507f2d4a5896",
      "pose1-14.npz": "cbf0a812b900ae8d2c9fd5669861327fb8df324220fe2506cd980a8cd7f7afe2",
      "pose1-15.npz": "c5c9fb7efde9576e9728dd1877ab93a97f6bf79684fa36829f103309e24fdff9",
      "pose1-16.npz": "9245e0e26372ac9675d9e2b20af037cf8c59afb4fbe38b0b34785edaf49861b0",
      "pose1-17.npz": "9ee7e5017b057d7bb666d1d796018c1b8b0a35c1a11c376cba635ca1008ce90d",
      "pose1-18.npz": "89a169202962f2e85c7e5b81fce9b3e04f6a7963b0a06b76c06e08ebb8d43bb5",
      "pose1-19.npz": "6b34c9823f1c371acc05d2fcbcd89915295a0697587e4ba1c2b24387edcbf6b8",
      "seed200.npz": "9fd175e0c42d66c8104ec81e1a740279f443df29fbb8d871905edcabe49b69d9",
      "seed201.npz": "f9f4fc09cf44ca64efef92d7e141eb67f7807c789e9812a7b7411c07673a317e",
      "seed202.npz": "4c5f9f5dbc873bb0a36fefc7204aaeb6fd6e420547e8302b19e8b78b38283a75",
      "seed203.npz": "9eabdb1b1ef3f60096c7c5fae00567d1210c4ce7201d2765bc341bee2533e97a",
      "seed204.npz": "c3c0dedcdf56b6cd91601760316ef843a67359d76869d6a8dce09264d13bb8cf",
      "seed205.npz": "eb900587295baa5237bb2a39ecfa23ed24e62009a81da8427cf00d45c5595ea3",
      "seed206.npz": "cfcd79365c0bb13d8dd6923f39a4ba417646ede5dcc973b70f9acc07c5f5bdf7",
      "seed207.npz": "35987d2ca811a3cb1eabca06f06960a21e20634e60d5c6e034f0966d1111d5e4",
      "seed208.npz": "3b046f4025047d490f0bc93d551448c6e4599d2cef029f4d2761a0ceb8188553",
      "seed209.npz": "9690410325889c152ad647f4fbeead43eca0267638a2f42ed0255c1ba4bddebf",
      "seed210.npz": "d66e029a69c90121691e9032906c28be6116676dc3d445558b637450d697aeea",
      "seed211.npz": "09ec83de3d464202ebe8111079ba8f4283f220d6f31df65eb60c6a59bd3e5135",
      "seed212.npz": "081222b0675948ba9ce46af564e71b2154dac972ad8cc8eb19acac640a3ce5b8",
      "seed213.npz": "f010cacf7cc3934469366a80896e9c8ba231b098cc955661c36faee5bfeff13d",
      "seed214.npz": "53742c42462d8409680b886a8e1cd046a2e4dc5817ce885ba791f29ec10562bf",
      "seed215.npz": "499161936b1b42c402a90109fca3a3161ebe65343c3a46b9cfda414c90becbda"
    },
    "environment_source_sha256": "ffe3196e464d8b0a570fcdb9f3d4f8ff5d2946438f9230315954d1e8d2b926ad",
    "initial_weights_sha256": "9705019dbedeaa9d10283e35b1718c5b622461fd4771c3ca7d1438d21ddd40b8",
    "lr": 0.0001,
    "model_source_sha256": "aac7932be5e7ce08375e30c3b20d78c44d0ec3d9c25a9b0be1b39c7537dde7b1",
    "observations": "full_map",
    "padded_steps": 640,
    "supervision": "expert real-action traces; injected waits and padding masked",
    "updates": 5000
  },
  "LR": 0.0005,
  "LR_WARMUP": 0.0,
  "MAX_GRAD_NORM": 0.25,
  "NUM_ENVS": 2,
  "NUM_MINIBATCHES": 2,
  "NUM_SEEDS": 1,
  "NUM_STEPS": 8,
  "REW_SHAPING_HORIZON": 32,
  "RUNTIME": {
    "METAL_COMPATIBILITY": true
  },
  "SCALE_CLIP_EPS": false,
  "SEED": 42,
  "TOTAL_TIMESTEPS": 32,
  "TRAINING_METHOD": "behavior_cloning",
  "UPDATE_EPOCHS": 1,
  "VF_COEF": 0.5,
  "WANDB_MODE": "disabled"
}
Training code and dependency provenance
{
  "action_counts": [
    11805,
    5975,
    11036,
    5417,
    16144,
    9554,
    5042
  ],
  "backend": "METAL",
  "demonstration_lengths": [
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    600,
    592,
    592,
    592,
    592,
    592,
    592,
    592,
    592
  ],
  "git_commit": "a81406291b336e6a8632eff3d77311e8297fa010",
  "git_dirty": true,
  "jax": "0.4.38",
  "method": "behavior cloning; no reward optimization or expert at evaluation",
  "parameter_count": 190976,
  "upstream_revision": "976aeb152cb184a5095021968bba94da96eb6394"
}
Training device
{}

Permanently delete this record?

This cannot be undone. Related training runs and checkpoints are kept.