rxnn 0.2.64__py3-none-any.whl → 0.2.66__py3-none-any.whl

rxnn/rxt/models.py

@@ -268,6 +268,7 @@ class RxTAlphaMemoryAttention(nn.Module, PyTorchModelHubMixin, license="apache-2
             residual_per_slot_gate: bool = False,
             residual_init_gate: float = 0.0,
             use_dynamic_residual_gate: bool = False,
+            use_tanh_residual_gate: bool = False,
             debug_mode: bool = False,
             debug_interval: int = 10,
             **kwargs,
@@ -300,7 +301,7 @@ class RxTAlphaMemoryAttention(nn.Module, PyTorchModelHubMixin, license="apache-2
             stm, attention_layers, memory_norm_layers,
             use_gated_residual=use_gated_residual, per_slot_gate=residual_per_slot_gate,
             init_gate=residual_init_gate, use_dynamic_gate=use_dynamic_residual_gate,
-            debug_mode=debug_mode, debug_interval=debug_interval,
+            use_tanh_gate=use_tanh_residual_gate, debug_mode=debug_mode, debug_interval=debug_interval,
         )
 
     def freeze(self):

rxnn/training/utils.py

@@ -151,5 +151,9 @@ def get_gradient_norms(model: nn.Module):
         param_norm = p.grad.data.norm(2)
         total_norm += param_norm.item() ** 2
     total_norm = total_norm ** 0.5
-    mean_norm = total_norm / len(grad_params)
+    params_len = len(grad_params)
+    if params_len != 0:
+        mean_norm = total_norm / params_len
+    else:
+        mean_norm = 0.0
     return total_norm, mean_norm

{rxnn-0.2.64.dist-info → rxnn-0.2.66.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: rxnn
-Version: 0.2.64
+Version: 0.2.66
 Summary: RxNN: Reactive Neural Networks Platform
 License: Apache-2.0
 Keywords: deep-learning,ai,machine-learning

{rxnn-0.2.64.dist-info → rxnn-0.2.66.dist-info}/RECORD

@@ -9,7 +9,7 @@ rxnn/memory/attention.py,sha256=lSniKrf_skiM1V1zbfmV84PbKoQ-t_fVcKfwNKW3_OY,3844
 rxnn/memory/norm.py,sha256=cVjjhCLqR5K6-321SP_ObG17y-ddlcTJeCTXvW4vpk0,6675
 rxnn/memory/stm.py,sha256=jv57gsH9XW19sLbxpRDqsp1yfsii_4Ef4Ncr_ztk-i4,3937
 rxnn/rxt/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-rxnn/rxt/models.py,sha256=JrZQ78F4HGGklAy6mML4fbqdsMOcGSDRZpjhX55VXb8,15486
+rxnn/rxt/models.py,sha256=M_0nEfSgr5Wyv-Ku4TCLpIs5VndUccjtIR0wU0DSVRo,15574
 rxnn/training/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 rxnn/training/base.py,sha256=CqaArEZYOdH64nmKfx28U3GI46TzO4oNkjf_hrF23Cw,11835
 rxnn/training/bml.py,sha256=hw6gLpLkGvqLzxIvBg4MvCc5r8cHpEm2RDyh7nH6CtE,16914
@@ -22,7 +22,7 @@ rxnn/training/reward.py,sha256=uiSsBXmjMw2yv-1Bssy3RTlpU6zP8ape3490Sl-aT0M,16144
 rxnn/training/rl.py,sha256=hWtExxY-_pAmTOGYxyCNounUbaGWvLDVltC4sRC7MN4,7175
 rxnn/training/scheduler.py,sha256=LcjU35mEwz2U5x3U6tLfeeYlBqMxbFSxYzJYuXkWbSY,1408
 rxnn/training/tokenizer.py,sha256=umaLByMBx_NMrQElA45HLm9gkuzyKWDTFaKVd-CjXl0,8344
-rxnn/training/utils.py,sha256=QMNkJPQBY04DX9WN7GHnI2EZTBbAzWkjt2W-798oUII,6129
+rxnn/training/utils.py,sha256=ngDCm654NL3UsPy190Er4XPc9HI-OyEV6tDLMgEEvQc,6219
 rxnn/transformers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 rxnn/transformers/attention.py,sha256=KRnKT6XUqAXElxV9y72mSpdTeiMgCKCCLqqxCFNTHmA,16372
 rxnn/transformers/ff.py,sha256=WDjO-H9XWInoWnUnxiseIH6Kx5GlHP0zGJygwhcb1gc,2589
@@ -33,7 +33,7 @@ rxnn/transformers/moe.py,sha256=j6jEx6Ip0zttlUZKKn82azxo95lkLZs-H2GLSMD88hY,5859
 rxnn/transformers/positional.py,sha256=1PjcJybUzeQlIKJI4tahAGZcYgCRCL0otxs7mpsNuzM,4410
 rxnn/transformers/sampler.py,sha256=t6iiQTdLQ0TakUWnnhKkb5DKF2F_9-thXHBydDF3fxg,17389
 rxnn/utils.py,sha256=ihb6OTyDtPiocB_lOvnq7eOkjjpCkgs8wxvXUBNQ7mM,996
-rxnn-0.2.64.dist-info/LICENSE,sha256=C8coDFIUYuOcke4JLPwTqahQUCyXyGq6WOaigOkx8tY,11275
-rxnn-0.2.64.dist-info/METADATA,sha256=QaDWd-8W0vs3povCgRAUXXSmTNN8gkEJ1dY6mA7n9kQ,25997
-rxnn-0.2.64.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-rxnn-0.2.64.dist-info/RECORD,,
+rxnn-0.2.66.dist-info/LICENSE,sha256=C8coDFIUYuOcke4JLPwTqahQUCyXyGq6WOaigOkx8tY,11275
+rxnn-0.2.66.dist-info/METADATA,sha256=-vE86fZe4EISwHsFwRSx6K05iBGghFo20l0xS3OwpGs,25997
+rxnn-0.2.66.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+rxnn-0.2.66.dist-info/RECORD,,