Upload . with huggingface_hub

.summary/0/events.out.tfevents.1678286079.qgallouedec-MS-7C84

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:69b89313cd1a881e17a47ff9388cb59a886a2aced9126dae7076d4f3b970673d
-size 1954598
+oid sha256:99cecd26758e5938e18fda367db078079b7867f0ebe0b1848c08776398a920d4
+size 4964113

README.md

@@ -15,7 +15,7 @@ model-index:
       type: hand-insert-v2
     metrics:
     - type: mean_reward
-      value: 12.85 +/- 5.03
+      value: 4295.54 +/- 1298.36
       name: mean_reward
       verified: false
 ---

checkpoint_p0/best_000184960_94699520_reward_4719.047.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68278cb6570af4eff04f454f5e06000abf938a6eaab3e9ed79bdc89174b389c4
+size 98239

checkpoint_p0/checkpoint_000195208_99946496.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:944b1247942c821f9c02fe77c31227c45bcca5e81836d0023daf80070bf44779
+size 98567

checkpoint_p0/checkpoint_000195328_100007936.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdfda9847b96835f2b8ce3c1909e6cd9fbc0ad8b1e68fa5f931931d17f5340a9
+size 98567

config.json

@@ -4,7 +4,7 @@
   "env": "hand-insert-v2",
   "experiment": "hand-insert-v2",
   "train_dir": "/home/qgallouedec/gia/data/envs/metaworld/train_dir",
-  "restart_behavior": "restart",
+  "restart_behavior": "resume",
   "device": "cpu",
   "seed": null,
   "num_policies": 1,
@@ -24,7 +24,7 @@
   "recurrence": 1,
   "shuffle_minibatches": false,
   "gamma": 0.99,
-  "reward_scale": 0.1,
+  "reward_scale": 0.01,
   "reward_clip": 1000.0,
   "value_bootstrap": true,
   "normalize_returns": true,
@@ -120,15 +120,17 @@
   "pbt_target_objective": "true_objective",
   "pbt_perturb_min": 1.1,
   "pbt_perturb_max": 1.5,
-  "command_line": "--env hand-insert-v2 --experiment hand-insert-v2 --with_wandb True --wandb_user qgallouedec --wandb_project sample_facotry_metaworld",
+  "command_line": "--env hand-insert-v2 --experiment hand-insert-v2 --with_wandb True --wandb_user qgallouedec --wandb_project sample_facotry_metaworld --train_for_env_steps 100000000 --reward_scale 0.01",
   "cli_args": {
     "env": "hand-insert-v2",
     "experiment": "hand-insert-v2",
+    "reward_scale": 0.01,
+    "train_for_env_steps": 100000000,
     "with_wandb": true,
     "wandb_user": "qgallouedec",
     "wandb_project": "sample_facotry_metaworld"
   },
-  "git_hash": "2bb8d8c9cd813ffeafaad42038b1e71364092d36",
+  "git_hash": "44aa76c42f03edbe61877f2b20879ac5319e5842",
   "git_repo_name": "https://github.com/huggingface/gia",
-  "wandb_unique_id": "hand-insert-v2_20230308_185042_269468"
+  "wandb_unique_id": "hand-insert-v2_20230308_153436_728632"
 }

git.diff

@@ -1,61 +0,0 @@
-diff --git a/data/envs/metaworld/train_all.sh b/data/envs/metaworld/train_all.sh
-index dbf328a..c412d01 100755
---- a/data/envs/metaworld/train_all.sh
-+++ b/data/envs/metaworld/train_all.sh
-@@ -1,28 +1,28 @@
- #!/bin/bash
- 
- ENVS=(
--    assembly
--    basketball
--    bin-picking
--    box-close
--    button-press-topdown
--    button-press-topdown-wall
--    button-press
--    button-press-wall
--    coffee-button
--    coffee-pull
--    coffee-push
--    dial-turn
--    disassemble
--    door-close
--    door-lock
--    door-open
--    door-unlock
--    drawer-close
--    drawer-open
--    faucet-close
--    faucet-open
--    hammer
-+    # assembly
-+    # basketball
-+    # bin-picking
-+    # box-close
-+    # button-press-topdown
-+    # button-press-topdown-wall
-+    # button-press
-+    # button-press-wall
-+    # coffee-button
-+    # coffee-pull
-+    # coffee-push
-+    # dial-turn
-+    # disassemble
-+    # door-close
-+    # door-lock
-+    # door-open
-+    # door-unlock
-+    # drawer-close
-+    # drawer-open
-+    # faucet-close
-+    # faucet-open
-+    # hammer
-     hand-insert
-     handle-press-side
-     handle-press
-@@ -55,4 +55,5 @@ ENVS=(
- 
- for ENV in "${ENVS[@]}"; do
-     python train.py --env $ENV-v2 --experiment $ENV-v2 --with_wandb True --wandb_user qgallouedec --wandb_project sample_facotry_metaworld
-+    python enjoy.py --algo=APPO --env $ENV-v2 --experiment $ENV-v2 --train_dir=./train_dir --max_num_episodes=10 --push_to_hub --hf_repository=qgallouedec/sample-factory-$ENV-v2 --save_video --no_render --enjoy_script=enjoy --train_script=train --load_checkpoint_kind best
- done

replay.mp4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f119d516d12805f1f08703d15334d766c9c035ea58e09cd30be4bf7cfe43e2f
-size 2668517
+oid sha256:87a1d832b52692f42e95a6859cf3363a2989963f4721106fe5a968db30eb90de
+size 2016827