evolutionary-policy-optimization 0.0.16__py3-none-any.whl → 0.0.18__py3-none-any.whl

evolutionary_policy_optimization/epo.py

@@ -15,6 +15,8 @@ from assoc_scan import AssocScan
 
 from adam_atan2_pytorch import AdoptAtan2
 
+from hl_gauss_pytorch import HLGaussLayer
+
 # helpers
 
 def exists(v):
@@ -204,6 +206,8 @@ class Actor(Module):
         assert len(dim_hiddens) >= 2
         dim_first, *_, dim_last = dim_hiddens
 
+        self.dim_latent = dim_latent
+
         self.init_layer = nn.Sequential(
             nn.Linear(dim_state, dim_first),
             nn.SiLU()
@@ -240,6 +244,8 @@ class Critic(Module):
         assert len(dim_hiddens) >= 2
         dim_first, *_, dim_last = dim_hiddens
 
+        self.dim_latent = dim_latent
+
         self.init_layer = nn.Sequential(
             nn.Linear(dim_state, dim_first),
             nn.SiLU()
@@ -488,6 +494,8 @@ class Agent(Module):
 
         self.latent_gene_pool = latent_gene_pool
 
+        assert actor.dim_latent == critic.dim_latent == latent_gene_pool.dim_latent
+
         # optimizers
 
         self.actor_optim = optim_klass(actor.parameters(), lr = actor_lr, **actor_optim_kwargs)

{evolutionary_policy_optimization-0.0.16.dist-info → evolutionary_policy_optimization-0.0.18.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: evolutionary-policy-optimization
-Version: 0.0.16
+Version: 0.0.18
 Summary: EPO - Pytorch
 Project-URL: Homepage, https://pypi.org/project/evolutionary-policy-optimization/
 Project-URL: Repository, https://github.com/lucidrains/evolutionary-policy-optimization
@@ -33,20 +33,17 @@ Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3.8
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Requires-Python: >=3.8
+Requires-Python: >=3.9
 Requires-Dist: adam-atan2-pytorch
 Requires-Dist: assoc-scan
 Requires-Dist: einops>=0.8.0
+Requires-Dist: hl-gauss-pytorch>=0.1.19
 Requires-Dist: torch>=2.2
 Requires-Dist: tqdm
 Provides-Extra: examples
 Requires-Dist: numpy; extra == 'examples'
 Requires-Dist: pufferlib>=2.0.6; extra == 'examples'
 Requires-Dist: tqdm; extra == 'examples'
-Provides-Extra: examples-gym
-Requires-Dist: box2d-py; extra == 'examples-gym'
-Requires-Dist: gymnasium[box2d]>=1.0.0; extra == 'examples-gym'
-Requires-Dist: tqdm; extra == 'examples-gym'
 Provides-Extra: test
 Requires-Dist: pytest; extra == 'test'
 Description-Content-Type: text/markdown
@@ -110,4 +107,15 @@ latent_pool.genetic_algorithm_step(fitness) # update latent genes with genetic a
 }
 ```
 
+```bibtex
+@article{Farebrother2024StopRT,
+    title   = {Stop Regressing: Training Value Functions via Classification for Scalable Deep RL},
+    author  = {Jesse Farebrother and Jordi Orbay and Quan Ho Vuong and Adrien Ali Taiga and Yevgen Chebotar and Ted Xiao and Alex Irpan and Sergey Levine and Pablo Samuel Castro and Aleksandra Faust and Aviral Kumar and Rishabh Agarwal},
+    journal = {ArXiv},
+    year   = {2024},
+    volume = {abs/2403.03950},
+    url    = {https://api.semanticscholar.org/CorpusID:268253088}
+}
+```
+
 *Evolution is cleverer than you are.* - Leslie Orgel

evolutionary_policy_optimization-0.0.18.dist-info/RECORD

@@ -0,0 +1,7 @@
+evolutionary_policy_optimization/__init__.py,sha256=Qavcia0n13jjaWIS_LPW7QrxSLT_BBeKujCjF9kQjbA,133
+evolutionary_policy_optimization/epo.py,sha256=3pvYPwAEZdrxwwV95Ea1qG4CQjLnyaxAr40opk07LDw,17747
+evolutionary_policy_optimization/experimental.py,sha256=ktBKxRF27Qsj7WIgBpYlWXqMVxO9zOx2oD1JuDYRAwM,548
+evolutionary_policy_optimization-0.0.18.dist-info/METADATA,sha256=BIyCXw2IbMs-x2hDbFs9NR5s2dYEbfbeK_LadUeUc8Q,4860
+evolutionary_policy_optimization-0.0.18.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+evolutionary_policy_optimization-0.0.18.dist-info/licenses/LICENSE,sha256=1yCiA9b5nhslTavxPjsQAO-wpOnwJR9-l8LTVi7GJuk,1066
+evolutionary_policy_optimization-0.0.18.dist-info/RECORD,,

evolutionary_policy_optimization-0.0.16.dist-info/RECORD

@@ -1,7 +0,0 @@
-evolutionary_policy_optimization/__init__.py,sha256=Qavcia0n13jjaWIS_LPW7QrxSLT_BBeKujCjF9kQjbA,133
-evolutionary_policy_optimization/epo.py,sha256=f_e-TkJRFF1VHG3psJDgLGNIzlEvDSjX0nOsbLaOBrw,17543
-evolutionary_policy_optimization/experimental.py,sha256=ktBKxRF27Qsj7WIgBpYlWXqMVxO9zOx2oD1JuDYRAwM,548
-evolutionary_policy_optimization-0.0.16.dist-info/METADATA,sha256=Go0IYw3PbRLeHIu_Volhn31YX8WKLiI_YdllA3mj-bE,4480
-evolutionary_policy_optimization-0.0.16.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-evolutionary_policy_optimization-0.0.16.dist-info/licenses/LICENSE,sha256=1yCiA9b5nhslTavxPjsQAO-wpOnwJR9-l8LTVi7GJuk,1066
-evolutionary_policy_optimization-0.0.16.dist-info/RECORD,,