Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

language: en license: apache-2.0 model-index:

  • name: CartPole-v1__functional_dqn__0__1671479574

DQN model applied to the this discrete environments CartPole-v1

Model Description

The model was trained from the CleanRl library using the DQN algorithm

Intended Use & Limitation

The model is intended to be used for the following environments CartPole-v1 and understand the implication of Quantization on this type of model from a pretrained state## Training Procdure

Training Hyperparameters

The folloing hyperparameters were used during training:

  • exp_name: functional_dqn
  • seed: 0
  • torch_deterministic: True
  • cuda: False
  • track: True
  • wandb_project_name: cleanRL
  • wandb_entity: compress_rl
  • capture_video: False
  • env_id: CartPole-v1
  • total_timesteps: 500000
  • learning_rate: 0.00025
  • buffer_size: 10000
  • gamma: 0.99
  • target_network_frequency: 500
  • batch_size: 128
  • start_e: 1
  • end_e: 0.05
  • exploration_fraction: 0.5
  • learning_starts: 10000
  • train_frequency: 10
  • optimizer: Adan
  • wandb_project: cleanrl

Framework and version

Pytorch 1.12.1+cu102 gym 0.23.1 Weights and Biases 0.13.3 Hugging Face Hub 0.11.1

Downloads last month
0
Unable to determine this model's library. Check the docs .