Reinforcement Learning
Transformers
HalfCheetah-v4
decision-transformer
deep-reinforcement-learning
custom-implementation
Instructions to use SriramSohan/Cheetah-v4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SriramSohan/Cheetah-v4 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("SriramSohan/Cheetah-v4", dtype="auto") - Notebooks
- Google Colab
- Kaggle
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.