v0.4 | slimshadys Workspace – Weights & Biases

Skip to main content

Slimshadys's group workspace

Group: v0.4

1-3

of 3

Tags

Notes

Author

State

Crashed

Start time

December 9th, 2023 4:41:38 PM

Runtime

3h 24m 14s

Tracked hours

3h 2m 11s

Run path

slimshadys/SoftKick/7f5gt8ux

OS

Windows-10-10.0.22631-SP0

Python version

3.9.18

Git repository

git clone https://github.com/SlimShadys/SoftKick

Git state

git checkout -b "rlgym-ppo-run" 0554fcc2371f17c236c2bedd274babffa7e7a5c3

Command

c:\Users\gianm\Documents\Università-Git\SoftKick\train_rlgym_ppo.py

System Hardware

CPU count	8
Logical CPU count	16
GPU count	1
GPU type	NVIDIA GeForce RTX 3060 Laptop GPU

W&B CLI Version

0.16.0

Group

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 15 keys
- ▶
  critic_layer_sizes:[] 4 items
  - 0:
    1,024
  - 1:
    512
  - 2:
    512
  - 3:
    512
- critic_lr:
  0.001
- exp_buffer_size:
  150,000
- gae_gamma:
  0.996
- gae_lambda:
  0.95
- min_inference_size:
  40
- n_proc:
  45
- ▶
  policy_layer_sizes:[] 4 items
  - 0:
    1,024
  - 1:
    512
  - 2:
    512
  - 3:
    512
- policy_lr:
  0.001
- ppo_batch_size:
  50,000
- ppo_clip_range:
  0.2
- ppo_ent_coef:
  0.0001
- ppo_epochs:
  1
- ppo_minibatch_size:
  50,000
- ts_per_iteration:
  50,000

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 24 keys
- ball_height:
  390.000732421875
- ball_speed:
  436.6310424804687
- boost_held:
  0.036328110843896866
- car_height:
  20.50667381286621
- car_speed:
  628.5851440429688
- Collected Steps per Second:
  8,214.629182959312
- Cumulative Model Updates:
  2,761
- Cumulative Timesteps:
  69,081,448
- demoed_rate:
  0
- Mean KL Divergence:
  0.004542958806268871
- on_ground:
  0.4649882018566131
- Overall Steps per Second:
  6,384.587346300871
- Policy Entropy:
  0.48903368413448334
- Policy Reward:
  -225.87839414567424
- Policy Update Magnitude:
  0.07963341474533081
- PPO Batch Consumption Time:
  0.1490398645401001
- SB3 Clip Fraction:
  0.04098999872803688
- Timestep Collection Time:
  6.0898670999995375
- Timestep Consumption Time:
  1.7455649000003175
- Timesteps Collected:
  50,026
- Total Iteration Time:
  7.835431999999855
- touch_rate:
  0.007596049923449755
- Value Function Loss:
  0.04895556904375553
- Value Function Update Magnitude:
  0.07235431671142578

This run consumed these artifacts as inputs. Learn more

job-https___github.com_SlimShadys_SoftKick_train_rlgym_ppo.py:v7

Loading...

This run produced these artifacts as outputs. Learn more

run-7f5gt8ux-history:v0

Loading...