metacontroller-pytorch 0.0.10__tar.gz → 0.0.12__tar.gz

{metacontroller_pytorch-0.0.10 → metacontroller_pytorch-0.0.12}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: metacontroller-pytorch
-Version: 0.0.10
+Version: 0.0.12
 Summary: Transformer Metacontroller
 Project-URL: Homepage, https://pypi.org/project/metacontroller/
 Project-URL: Repository, https://github.com/lucidrains/metacontroller
@@ -35,9 +35,10 @@ Classifier: Programming Language :: Python :: 3.9
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9
 Requires-Dist: assoc-scan>=0.0.3
-Requires-Dist: discrete-continuous-embed-readout>=0.1.11
+Requires-Dist: discrete-continuous-embed-readout>=0.1.12
 Requires-Dist: einops>=0.8.1
 Requires-Dist: einx>=0.3.0
+Requires-Dist: loguru
 Requires-Dist: torch>=2.5
 Requires-Dist: x-evolution>=0.1.23
 Requires-Dist: x-mlps-pytorch

{metacontroller_pytorch-0.0.10 → metacontroller_pytorch-0.0.12}/metacontroller/metacontroller.py

@@ -3,6 +3,7 @@ from contextlib import nullcontext
 
 from functools import partial
 from collections import namedtuple
+from loguru import logger
 
 import torch
 from torch import nn, cat, stack, tensor
@@ -130,7 +131,8 @@ class MetaController(Module):
         residual_stream,
         cache: MetaControllerOutput | None = None,
         discovery_phase = False,
-        hard_switch = False
+        hard_switch = False,
+        temperature = 1.
     ):
 
         # destruct prev cache
@@ -142,6 +144,8 @@ class MetaController(Module):
         next_action_proposer_hidden = None
 
         if discovery_phase:
+            logger.warning('meta controller cache being passed back in for discovery phase, which does not make sense given bidirectional encoder')
+
             temporal_compressed, _ = self.bidirectional_temporal_compressor(residual_stream)
             temporal_compressed = reduce(temporal_compressed, '... (two d) -> ... d', 'mean', two = 2)
 
@@ -157,7 +161,7 @@ class MetaController(Module):
 
         action_dist = readout(proposed_action_hidden)
 
-        sampled_action = readout.sample(action_dist)
+        sampled_action = readout.sample(action_dist, temperature = temperature)
 
         # switching unit timer
 
@@ -287,8 +291,9 @@ class Transformer(Module):
         discovery_phase = False,
         no_grad_transformer = None,
         no_grad_meta_controller = None,
+        meta_controller_temperature = 1.,
         return_latents = False,
-        return_cache = False
+        return_cache = False,
     ):
         meta_controller = default(meta_controller, self.meta_controller)
 
@@ -319,7 +324,7 @@ class Transformer(Module):
         with meta_controller_context():
 
             if exists(meta_controller):
-                modified_residual_stream, next_meta_hiddens = meta_controller(residual_stream, cache = meta_hiddens, discovery_phase = discovery_phase)
+                modified_residual_stream, next_meta_hiddens = meta_controller(residual_stream, cache = meta_hiddens, discovery_phase = discovery_phase, temperature = meta_controller_temperature)
             else:
                 modified_residual_stream, next_meta_hiddens = residual_stream, None
 

{metacontroller_pytorch-0.0.10 → metacontroller_pytorch-0.0.12}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "metacontroller-pytorch"
-version = "0.0.10"
+version = "0.0.12"
 description = "Transformer Metacontroller"
 authors = [
     { name = "Phil Wang", email = "lucidrains@gmail.com" }
@@ -25,9 +25,10 @@ classifiers=[
 
 dependencies = [
     "assoc-scan>=0.0.3",
+    "discrete-continuous-embed-readout>=0.1.12",
     "einx>=0.3.0",
     "einops>=0.8.1",
-    "discrete-continuous-embed-readout>=0.1.11",
+    "loguru",
     "torch>=2.5",
     "x-evolution>=0.1.23",
     "x-mlps-pytorch",