File size: 6,749 Bytes
dd58921 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 | # Basic Usage
## Two Ways to Run
After the dependencies have been installed, there are **two ways** to run
the CleanRL script under the uv virtual environments.
1. Using `uv run`:
```bash
uv run python cleanrl/ppo.py \
--seed 1 \
--env-id CartPole-v0 \
--total-timesteps 50000
```
<script id="asciicast-443649" src="https://asciinema.org/a/443649.js" async></script>
2. Using `uv venv`:
1. We first activate the virtual environment by using
`uv venv`
2. Then, run any desired CleanRL script
Attention: Each step must be executed separately!
```bash
uv venv
```
```bash
python cleanrl/ppo.py \
--seed 1 \
--env-id CartPole-v0 \
--total-timesteps 50000
```
<script id="asciicast-JL1FR00I2JNklAhMd2dwEAQuz" src="https://asciinema.org/a/JL1FR00I2JNklAhMd2dwEAQuz.js" async></script>
!!! note
We recommend `uv venv` workflow for development. When the shell is activated, you should
be seeing a prefix like `(cleanrl-iXg02GqF-py3.9)` in your shell's prompt, which is the name
of the poetry's virtual environment.
**We will assume to run other commands (e.g. `tensorboard`) in the documentation within the poetry's shell.**
!!! warning
If you are using NVIDIA ampere GPUs (e.g., 3060 TI), you might meet the following error
```bash
NVIDIA GeForce RTX 3060 Ti with CUDA capability sm_86 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70.
If you want to use the NVIDIA GeForce RTX 3060 Ti GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))
Traceback (most recent call last):
File "ppo_atari_envpool.py", line 240, in <module>
action, logprob, _, value = agent.get_action_and_value(next_obs)
File "ppo_atari_envpool.py", line 156, in get_action_and_value
hidden = self.network(x / 255.0)
RuntimeError: CUDA error: no kernel image is available for execution on the device
CUDA kernel errors might be asynchronously reported at some other API call,so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
```
This is because the `torch` wheel on PyPi is built with cuda 10.2. You would need to manually install the cuda 11.3 wheel like this:
```bash
uv pip install torch==1.12.1 --upgrade --extra-index-url https://download.pytorch.org/whl/cu113
```
Then, you can run the script again.
## Visualize Training Metrics
By default, the CleanRL scripts record all the training metrics via Tensorboard
into the `runs` folder. So, after running the training script above, feel free to run
```bash
tensorboard --logdir runs
```
## Visualize the Agent's Gameplay Videos
CleanRL helps record the agent's gameplay videos with a `--capture_video` flag,
which will save the videos in the `videos/{$run_name}` folder.
```bash linenums="1" hl_lines="5"
python cleanrl/ppo.py \
--seed 1 \
--env-id CartPole-v0 \
--total-timesteps 50000 \
--capture_video
```
## Get Documentation
You can directly obtained the documentation by using the `--help` flag.
```bash
python cleanrl/ppo.py --help
usage: ppo.py [-h] [--exp-name EXP_NAME] [--env-id ENV_ID]
[--learning-rate LEARNING_RATE] [--seed SEED]
[--total-timesteps TOTAL_TIMESTEPS]
[--torch-deterministic [TORCH_DETERMINISTIC]] [--cuda [CUDA]]
[--track [TRACK]] [--wandb-project-name WANDB_PROJECT_NAME]
[--wandb-entity WANDB_ENTITY] [--capture_video [CAPTURE_VIDEO]]
[--num-envs NUM_ENVS] [--num-steps NUM_STEPS]
[--anneal-lr [ANNEAL_LR]] [--gae [GAE]] [--gamma GAMMA]
[--gae-lambda GAE_LAMBDA] [--num-minibatches NUM_MINIBATCHES]
[--update-epochs UPDATE_EPOCHS] [--norm-adv [NORM_ADV]]
[--clip-coef CLIP_COEF] [--clip-vloss [CLIP_VLOSS]]
[--ent-coef ENT_COEF] [--vf-coef VF_COEF]
[--max-grad-norm MAX_GRAD_NORM] [--target-kl TARGET_KL]
optional arguments:
-h, --help show this help message and exit
--exp-name EXP_NAME the name of this experiment
--env-id ENV_ID the id of the environment
--learning-rate LEARNING_RATE
the learning rate of the optimizer
--seed SEED seed of the experiment
--total-timesteps TOTAL_TIMESTEPS
total timesteps of the experiments
--torch-deterministic [TORCH_DETERMINISTIC]
if toggled, `torch.backends.cudnn.deterministic=False`
--cuda [CUDA] if toggled, cuda will be enabled by default
--track [TRACK] if toggled, this experiment will be tracked with Weights
and Biases
--wandb-project-name WANDB_PROJECT_NAME
the wandb's project name
--wandb-entity WANDB_ENTITY
the entity (team) of wandb's project
--capture_video [CAPTURE_VIDEO]
weather to capture videos of the agent performances (check
out `videos` folder)
--num-envs NUM_ENVS the number of parallel game environments
--num-steps NUM_STEPS
the number of steps to run in each environment per policy
rollout
--anneal-lr [ANNEAL_LR]
Toggle learning rate annealing for policy and value
networks
--gae [GAE] Use GAE for advantage computation
--gamma GAMMA the discount factor gamma
--gae-lambda GAE_LAMBDA
the lambda for the general advantage estimation
--num-minibatches NUM_MINIBATCHES
the number of mini-batches
--update-epochs UPDATE_EPOCHS
the K epochs to update the policy
--norm-adv [NORM_ADV]
Toggles advantages normalization
--clip-coef CLIP_COEF
the surrogate clipping coefficient
--clip-vloss [CLIP_VLOSS]
Toggles whether or not to use a clipped loss for the value
function, as per the paper.
--ent-coef ENT_COEF coefficient of the entropy
--vf-coef VF_COEF coefficient of the value function
--max-grad-norm MAX_GRAD_NORM
the maximum norm for the gradient clipping
--target-kl TARGET_KL
the target KL divergence threshold
```
|