Upload . with huggingface_hub

.summary/0/events.out.tfevents.1678297845.qgallouedec-MS-7C84

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2836d6f83bdd97fb224a50435a46023ef41f74e3c711cce678b3e2d48e325b7d
+size 67266

README.md

@@ -15,7 +15,7 @@ model-index:
       type: hand-insert-v2
     metrics:
     - type: mean_reward
-      value: 2267.66 +/- 2222.12
+      value: 12.85 +/- 5.03
       name: mean_reward
       verified: false
 ---

checkpoint_p0/best_000000224_114688_reward_147.069.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33e5942b92ad2a181416b6b51e681802b3057bcd32c79d3dd30a12706d398ff3
+size 98239

checkpoint_p0/checkpoint_000000456_233472.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20fed33ec75095ab20b7ac281215c7a1dca17ab295b893b99dddbdbcb4d26673
+size 98567

checkpoint_p0/checkpoint_000000696_356352.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d5ccf9a64aedfb8fc1d7e46638d7ce2e804d143f6e3c4e5162e37185f881793
+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": "resume",
+  "restart_behavior": "restart",
   "device": "cpu",
   "seed": null,
   "num_policies": 1,
@@ -24,7 +24,7 @@
   "recurrence": 1,
   "shuffle_minibatches": false,
   "gamma": 0.99,
-  "reward_scale": 0.01,
+  "reward_scale": 0.1,
   "reward_clip": 1000.0,
   "value_bootstrap": true,
   "normalize_returns": true,
@@ -120,17 +120,15 @@
   "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 --train_for_env_steps 100000000 --reward_scale 0.01",
+  "command_line": "--env hand-insert-v2 --experiment hand-insert-v2 --with_wandb True --wandb_user qgallouedec --wandb_project sample_facotry_metaworld",
   "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": "44aa76c42f03edbe61877f2b20879ac5319e5842",
+  "git_hash": "2bb8d8c9cd813ffeafaad42038b1e71364092d36",
   "git_repo_name": "https://github.com/huggingface/gia",
-  "wandb_unique_id": "hand-insert-v2_20230308_153436_728632"
+  "wandb_unique_id": "hand-insert-v2_20230308_185042_269468"
 }

git.diff

@@ -0,0 +1,61 @@
+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:b92a870963b0ae5689f23bcdc7107c9f09bf93f252cada23fae29eb71648fa93
-size 2182118
+oid sha256:8f119d516d12805f1f08703d15334d766c9c035ea58e09cd30be4bf7cfe43e2f
+size 2668517