Upload config.json with huggingface_hub

config.json

@@ -3,10 +3,10 @@
   "algo": "APPO",
   "env": "doom_health_gathering_supreme",
   "experiment": "default_experiment",
-  "train_dir": "/home/raj/repos/HF-DeepRL/8-Proximal-Policy-Optimization/train_dir",
+  "train_dir": "/content/train_dir",
   "restart_behavior": "resume",
   "device": "gpu",
-  "seed": 42,
+  "seed": null,
   "num_policies": 1,
   "async_rl": true,
   "serial_mode": false,
@@ -15,9 +15,9 @@
   "worker_num_splits": 2,
   "policy_workers_per_policy": 1,
   "max_policy_lag": 1000,
-  "num_workers": 20,
-  "num_envs_per_worker": 12,
-  "batch_size": 2048,
+  "num_workers": 8,
+  "num_envs_per_worker": 4,
+  "batch_size": 1024,
   "num_batches_per_epoch": 1,
   "num_epochs": 1,
   "rollout": 32,
@@ -42,7 +42,7 @@
   "adam_eps": 1e-06,
   "adam_beta1": 0.9,
   "adam_beta2": 0.999,
-  "max_grad_norm": 0.0,
+  "max_grad_norm": 4.0,
   "learning_rate": 0.0001,
   "lr_schedule": "constant",
   "lr_schedule_kl_threshold": 0.008,
@@ -50,7 +50,7 @@
   "obs_scale": 255.0,
   "normalize_input": true,
   "normalize_input_keys": null,
-  "decorrelate_experience_max_seconds": 1,
+  "decorrelate_experience_max_seconds": 0,
   "decorrelate_envs_on_one_worker": true,
   "actor_worker_gpus": [],
   "set_workers_cpu_affinity": true,
@@ -61,10 +61,10 @@
   "flush_summaries_interval": 30,
   "stats_avg": 100,
   "summaries_use_frameskip": true,
-  "heartbeat_interval": 10,
-  "heartbeat_reporting_interval": 300,
-  "train_for_env_steps": 20,
-  "train_for_seconds": 1000000,
+  "heartbeat_interval": 20,
+  "heartbeat_reporting_interval": 600,
+  "train_for_env_steps": 4000000,
+  "train_for_seconds": 10000000000,
   "save_every_sec": 120,
   "keep_checkpoints": 2,
   "load_checkpoint_kind": "latest",
@@ -83,10 +83,10 @@
   ],
   "use_rnn": true,
   "rnn_size": 512,
-  "rnn_type": "lstm",
+  "rnn_type": "gru",
   "rnn_num_layers": 1,
   "decoder_mlp_layers": [],
-  "nonlinearity": "relu",
+  "nonlinearity": "elu",
   "policy_initialization": "orthogonal",
   "policy_init_gain": 1.0,
   "actor_critic_share_weights": true,
@@ -95,6 +95,7 @@
   "initial_stddev": 1.0,
   "use_env_info_cache": false,
   "env_gpu_actions": false,
+  "env_gpu_observations": true,
   "env_frameskip": 4,
   "env_framestack": 1,
   "pixel_format": "CHW",
@@ -127,11 +128,14 @@
   "wide_aspect_ratio": false,
   "eval_env_frameskip": 1,
   "fps": 35,
-  "cli_args": {},
-  "git_hash": "0401714b01ee832562a0930e3744117f1ba51e10",
-  "git_repo_name": "https://github.com/tenkara/HF-DeepRL.git",
-  "command_line": "--env=doom_health_gathering_supreme --train_for_env_steps=20",
-  "lr_adaptive_min": 1e-06,
-  "lr_adaptive_max": 0.01,
-  "env_gpu_observations": true
+  "command_line": "--env=doom_health_gathering_supreme --num_workers=8 --num_envs_per_worker=4 --train_for_env_steps=4000000",
+  "cli_args": {
+    "env": "doom_health_gathering_supreme",
+    "num_workers": 8,
+    "num_envs_per_worker": 4,
+    "train_for_env_steps": 4000000
+  },
+  "git_hash": "unknown",
+  "git_repo_name": "not a git repository",
+  "train_script": ".usr.local.lib.python3.10.dist-packages.ipykernel_launcher"
 }