state-machine

State-machine's group workspace

Group: xland-multi-task-test-v2

1-3

of 3

Tags

Notes

Author

state-machine

State

Finished

Start time

May 9th, 2024 8:28:00 PM

Runtime

14m 54s

Tracked hours

14m 48s

Run path

state-machine/xminigrid/gh7r0hw4

Linux-5.15.0-1044-nvidia-x86_64-with-glibc2.31

Python version

3.10.13

Git repository

git clone https://github.com/Howuhh/xland-minigrid-datasets.git

Git state

git checkout -b "ppo-rnn-multi-task-medium-XLand-MiniGrid-R1-9x9-trivial-1m-0" 5bb436c2e0ac10f4801e35efd959b847a51f97fc

Command

/home/jovyan/nikulin/xland-minigrid-datasets/src/xminigrid_datasets/collection/train.py --config_path=configs/trivial_base.yaml --group=xland-multi-task-test-v2 --name=ppo-rnn-multi-task-medium --dataset_path=/home/jovyan/nikulin/data --dataset_num_histories=32

System Hardware

CPU count	48
Logical CPU count	96
GPU count	1
GPU type	Tesla V100-SXM3-32GB

W&B CLI Version

0.17.0

Group

xland-multi-task-test-v2

Config parameters are your model's inputs. Learn more

▶
Config parameters:{} 39 keys
- action_emb_dim:
  16
- benchmark_id:
  "trivial-1m"
- checkpoint_path:
  null
- clip_eps:
  0.2
- context_dropout:
  0
- context_emb_dim:
  16
- context_hidden_dim:
  256
- dataset_name:
  "trivial-1m-0.hdf5"
- dataset_num_histories:
  32
- dataset_path:
  "/home/jovyan/nikulin/data"
- decay_lr:
  true
- ent_coef:
  0.001
- env_id:
  "XLand-MiniGrid-R1-9x9"
- eval_episodes:
  1,024
- eval_seed:
  42
- gae_lambda:
  0.999
- gamma:
  0.999
- group:
  "xland-multi-task-test-v2"
- head_hidden_dim:
  256
- lr:
  0.0005
- max_grad_norm:
  0.5
- name:
  "ppo-rnn-multi-task-medium-XLand-MiniGrid-R1-9x9-trivial-1m-0"
- num_envs:
  8,192
- num_minibatches:
  8
- num_steps:
  512
- obs_emb_dim:
  16
- pretrain_multitask:
  false
- pretrained_checkpoint_path:
  null
- project:
  "xminigrid"
- rnn_hidden_dim:
  256
- rnn_num_layers:
  1
- ruleset_id:
  0
- seed:
  0
- total_timesteps:
  500,000,000
- update_epochs:
  1
- upload_to_hf_repo:
  null
- use_bf16:
  false
- vf_coef:
  0.5
- wandb_logging:
  true

Summary metrics are your model's outputs. Learn more

▶
Summary metrics:{} 10 keys
- actor_loss:
  -0.00000774083309806883
- entropy:
  0.22124701738357544
- eval/lengths:
  30.25
- eval/returns:
  0.8878653049468994
- lr:
  0.00000420165088144131
- steps_per_second:
  603,988.7488375251
- total_loss:
  0.0033585582859814167
- training_time:
  827.8299901485443
- transitions:
  494,927,872
- value_loss:
  0.007175092585384846

This run consumed these artifacts as inputs. Learn more

job

job-https___github.com_Howuhh_xland-minigrid-datasets.git_src_xminigrid_datasets_collection_train.py:v0

This run produced these artifacts as outputs. Total: 1. Learn more

wandb-history

run-gh7r0hw4-history:v0