Upload folder using huggingface_hub

README.md

@@ -0,0 +1,31 @@
+---
+tags:
+- HalfCheetah-v4
+- reinforcement-learning
+- decision-transformer
+- deep-reinforcement-learning
+- custom-implementation
+library_name: transformers
+---
+
+# Decision Transformer for HalfCheetah-v4
+
+This is a trained Decision Transformer model for the HalfCheetah-v4 environment.
+
+## Model Details
+- Environment: HalfCheetah-v4
+- Model: Decision Transformer
+- Training framework: PyTorch
+- Final Training Loss: 0.07436713774998983
+
+## Hyperparameters
+{
+"max_ep_len": 1000,
+"state_dim": 17,
+"act_dim": 3,
+"target return": 12.0,
+"num_of_epochs": 120,
+"batch_size" : 64,
+"learning_rate": 1e-4
+}
+The model demonstrates the running behavior learned through Decision Transformer training.

hyperparameters.json

@@ -0,0 +1 @@
+{"env_id": "HalfCheetah-v4", "max_ep_len": 1000, "state_dim": 17, "act_dim": 6, "target_return": 12.0, "state_mean": [-0.044892117381095886, 0.032326120883226395, 0.060348208993673325, -0.17081618309020996, -0.1947702318429947, -0.057516805827617645, 0.09701419621706009, 0.032391782850027084, 11.047338485717773, -0.07997213304042816, -0.32363244891166687, 0.36296889185905457, 0.4232352375984192, 0.4083653688430786, 1.1085010766983032, -0.48743751645088196, -0.0737508088350296], "state_std": [0.04003537446260452, 0.41147083044052124, 0.5421720743179321, 0.4154345691204071, 0.23797930777072906, 0.6205318570137024, 0.30105698108673096, 0.2174210399389267, 2.211426258087158, 0.5726979970932007, 1.7259377241134644, 11.845534324645996, 12.067534446716309, 7.052667140960693, 13.506407737731934, 7.197610378265381, 5.027524948120117], "training_args": {"num_train_epochs": 120, "per_device_train_batch_size": 64, "learning_rate": 0.0001, "weight_decay": 0.0001, "warmup_ratio": 0.1, "max_grad_norm": 0.25}}

model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04b47884d5400489b8f094a0a5a8ca8a24e6c58922bb5a9a27b75d261c85fb54
+size 5044795

results.json

@@ -0,0 +1 @@
+{"env_id": "HalfCheetah-v4", "eval_datetime": "2025-01-21T17:49:02.231290", "training_loss": 0.07436713774998983, "metrics": {"train_runtime": 1433.1089, "train_samples_per_second": 83.734, "train_steps_per_second": 1.34, "total_flos": 147340224000000.0, "train_loss": 0.07436713774998983, "epoch": 120.0}}