Deep-Reinforcement-Learning-Algorithms-with-PyTorch

Deep-Reinforcement-Learning-Algorithms-with-PyTorch copied to clipboard

A question about critic-loss in discrete sac？

5

I applied the code of discrete sac to a custom discrete action environment. During the training process, I found that the loss of critic did not decrease but increased, and...

outshine-J

Bump numpy from 1.15.2 to 1.22.0

1

Bumps [numpy](https://github.com/numpy/numpy) from 1.15.2 to 1.22.0. Release notes Sourced from numpy's releases. v1.22.0 NumPy 1.22.0 Release Notes NumPy 1.22.0 is a big release featuring the work of 153 contributors spread...

dependabot[bot]

According to the paper, the target network should be updated several steps after local network update, but your code seem to be not like this. In your code, the local...

LLYYKK

four-rooms中，if __name__== '__main__': AGENTS = [A3C] #DIAYN] # DDQN] #SNN_HRL] #, DDQN] trainer = Trainer(config, AGENTS) trainer.run_games_for_agents() 调用A3C算法，报错：File "D:\Pycharm\test\Deep-Reinforcement-Learning-Algorithms-with-PyTorch-master\agents\actor_critic_agents\A3C.py", line 98, in __init__ self.exploration_worker_difference = self.config.hyperparameters["exploration_worker_difference"] KeyError: 'exploration_worker_difference'

HHY1123

torch版本问题

4

requirement.txt中的torch版本找不到请问如何处理?

noc-turne

Implement model saves

2

Hi. Great work on the library, it's working like a charm. Right now, only the DQN Agent implements the `locally_save_policy` that allows for saving the current model. Would it be...

LucCADORET

Updating setup info

2

Hello, I tried cloning this repo to test out your CartPole agent. I had to go through a few extra steps to have a successful setup. I thought I would...

simonalford42

terminate called after throwing an instance of 'at::Error' what(): CUDA error (3): initialization error (check_status at /pytorch/aten/src/ATen/cuda/detail/CUDAHooks.cpp:36)

3

**/home/account/anaconda3/envs/RL17/bin/python /home/account/Documents/Deep_RL_Implementations/results/Cart_Pole.py /home/account/anaconda3/envs/RL17/lib/python3.7/site-packages/gym/envs/registration.py:14: PkgResourcesDeprecationWarning: Parameters to load are deprecated. Call .resolve and .require separately. result = entry_point.load(False)** AGENT NAME: A3C 1.1: A3C TITLE CartPole layer info [20, 10, [2, 1]]...

JinQiangWang2021

In SAC.py Line 120 https://github.com/p-christ/Deep-Reinforcement-Learning-Algorithms-with-PyTorch/blob/b338c87bebb672e39304e47e0eed55aeb462b243/agents/actor_critic_agents/SAC.py#L120 However, the output of `produce_action_and_action_info(state)` is https://github.com/p-christ/Deep-Reinforcement-Learning-Algorithms-with-PyTorch/blob/b338c87bebb672e39304e47e0eed55aeb462b243/agents/actor_critic_agents/SAC.py#L135 So, even though SAC algorithm can work in practice, is it a mistake?

fokx

ConnectionResetError: [Errno 104] Connection reset by peer

2

AGENT NAME: A3C 1.1: A3C TITLE CartPole layer info [20, 10, [2, 1]] layer info [20, 10, [2, 1]] {'learning_rate': 0.005, 'linear_hidden_units': [20, 10], 'final_layer_activation': ['SOFTMAX', None], 'gradient_clipping_norm': 5.0, 'discount_rate':...

JinQiangWang2021