evolutionary-policy-optimization 0.0.56__py3-none-any.whl → 0.0.58__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (5) hide show
  1. evolutionary_policy_optimization/epo.py +5 -2
  2. {evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/METADATA +1 -1
  3. {evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/RECORD +5 -5
  4. {evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/WHEEL +0 -0
  5. {evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/licenses/LICENSE +0 -0
evolutionary_policy_optimization/epo.py CHANGED
@@ -1,4 +1,5 @@
1
1
  from __future__ import annotations
2
+ from typing import Callable
2
3
 
3
4
  from pathlib import Path
4
5
  from math import ceil
@@ -375,7 +376,7 @@ class LatentGenePool(Module):
375
376
  default_should_run_ga_gamma = 1.5,
376
377
  migrate_every = 100, # how many steps before a migration between islands
377
378
  apply_genetic_algorithm_every = 2, # how many steps before crossover + mutation happens for genes
378
- init_latent_fn: Callable = None
379
+ init_latent_fn: Callable | None = None
379
380
  ):
380
381
  super().__init__()
381
382
 
@@ -855,10 +856,12 @@ class Agent(Module):
855
856
  dones
856
857
  ) = memories
857
858
 
859
+ masks = 1. - dones.float()
860
+
858
861
  advantages = self.calc_gae(
859
862
  rewards[:-1],
860
863
  values,
861
- dones[:-1],
864
+ masks[:-1],
862
865
  )
863
866
 
864
867
  valid_episode = episode_ids >= 0
{evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: evolutionary-policy-optimization
3
- Version: 0.0.56
3
+ Version: 0.0.58
4
4
  Summary: EPO - Pytorch
5
5
  Project-URL: Homepage, https://pypi.org/project/evolutionary-policy-optimization/
6
6
  Project-URL: Repository, https://github.com/lucidrains/evolutionary-policy-optimization
{evolutionary_policy_optimization-0.0.56.dist-info → evolutionary_policy_optimization-0.0.58.dist-info}/RECORD RENAMED
@@ -1,9 +1,9 @@
1
1
  evolutionary_policy_optimization/__init__.py,sha256=0q0aBuFgWi06MLMD8FiHzBYQ3_W4LYWrwmCtF3u5H2A,201
2
2
  evolutionary_policy_optimization/distributed.py,sha256=7KgZdeS_wxBHo_du9XZFB1Cu318J-Bp66Xdr6Log_20,2423
3
- evolutionary_policy_optimization/epo.py,sha256=N7xmO3CRXeaJAy-2rysZg-DBvkZCZB2ySJT7Iq__r6w,35217
3
+ evolutionary_policy_optimization/epo.py,sha256=2PlLBrhX5_kkV4NOhxC5VB_plR-_jw3krrmMSjWAiCY,35288
4
4
  evolutionary_policy_optimization/experimental.py,sha256=-IgqjJ_Wk_CMB1y9YYWpoYqTG9GZHAS6kbRdTluVevg,1563
5
5
  evolutionary_policy_optimization/mock_env.py,sha256=202KJ5g57wQvOzhGYzgHfBa7Y2do5uuDvl5kFg5o73g,934
6
- evolutionary_policy_optimization-0.0.56.dist-info/METADATA,sha256=o2-1eCh8MuQVd0SH0GiUBBIAcqdK7cceuiu093cuEA4,6213
7
- evolutionary_policy_optimization-0.0.56.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
8
- evolutionary_policy_optimization-0.0.56.dist-info/licenses/LICENSE,sha256=1yCiA9b5nhslTavxPjsQAO-wpOnwJR9-l8LTVi7GJuk,1066
9
- evolutionary_policy_optimization-0.0.56.dist-info/RECORD,,
6
+ evolutionary_policy_optimization-0.0.58.dist-info/METADATA,sha256=ni9pyD5CS2QbBmwEUhCYt_RAxz3YnnOv3IY3ywNagSs,6213
7
+ evolutionary_policy_optimization-0.0.58.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
8
+ evolutionary_policy_optimization-0.0.58.dist-info/licenses/LICENSE,sha256=1yCiA9b5nhslTavxPjsQAO-wpOnwJR9-l8LTVi7GJuk,1066
9
+ evolutionary_policy_optimization-0.0.58.dist-info/RECORD,,