x-transformers 1.42.27__py3-none-any.whl → 1.42.28__py3-none-any.whl

x_transformers/x_transformers.py

@@ -1077,7 +1077,7 @@ class Attention(Module):
         logit_softclamp_value = 50.,
         neutreno_value_residual = False, # Nguyen et al. https://arxiv.org/abs/2312.00751
         neutreno_alpha = 0.4,
-        learned_value_residual_mix = True,
+        learned_value_residual_mix = False,
         laser = False, # https://arxiv.org/abs/2411.03493v1
         laser_softclamp_value = 15.,
         onnxable = False,
@@ -1584,7 +1584,7 @@ class AttentionLayers(Module):
         unet_skips = False,
         reinject_input = False,              # seen first in DEQ paper https://arxiv.org/abs/1909.01377, but later used in a number of papers trying to achieve depthwise generalization https://arxiv.org/abs/2410.03020v1
         add_value_residual = False,          # resformer from Zhou et al - https://arxiv.org/abs/2410.17897v1
-        learned_value_residual_mix = False,  # seeing big improvements when the value residual mix value is learned per token - credit goes to @faresobeid for taking the first step with learned scalar mix, then @Blinkdl for taking it a step further with data dependent. here we will use per token learned
+        learned_value_residual_mix = True,   # seeing big improvements when the value residual mix value is learned per token - credit goes to @faresobeid for taking the first step with learned scalar mix, then @Blinkdl for taking it a step further with data dependent. here we will use per token learned
         rel_pos_kwargs: dict = dict(),
         **kwargs
     ):

{x_transformers-1.42.27.dist-info → x_transformers-1.42.28.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: x-transformers
-Version: 1.42.27
+Version: 1.42.28
 Summary: X-Transformers - Pytorch
 Home-page: https://github.com/lucidrains/x-transformers
 Author: Phil Wang

{x_transformers-1.42.27.dist-info → x_transformers-1.42.28.dist-info}/RECORD

@@ -6,11 +6,11 @@ x_transformers/dpo.py,sha256=xt4OuOWhU8pN3OKN2LZAaC2NC8iiEnchqqcrPWVqf0o,3521
 x_transformers/multi_input.py,sha256=tCh-fTJDj2ib4SMGtsa-AM8MxKzJAQSwqAXOu3HU2mg,9252
 x_transformers/neo_mlp.py,sha256=XCNnnop9WLarcxap1kGuYc1x8GHvwkZiDRnXOxSl3Po,3452
 x_transformers/nonautoregressive_wrapper.py,sha256=2NU58hYMgn-4Jzg3mie-mXb0XH_dCN7fjlzd3K1rLUY,10510
-x_transformers/x_transformers.py,sha256=mLAqXQuZynqueJDkTEBs-kE9Uk8mSq_DF8UG9oY65Ns,96695
+x_transformers/x_transformers.py,sha256=X4HegsAtCnaL3MAxu07RkZ5WBMgtdbi0W-2c9bXQxew,96696
 x_transformers/xl_autoregressive_wrapper.py,sha256=CvZMJ6A6PA-Y_bQAhnORwjJBSl6Vjq2IdW5KTdk8NI8,4195
 x_transformers/xval.py,sha256=7S00kCuab4tWQa-vf-z-XfzADjVj48MoFIr7VSIvttg,8575
-x_transformers-1.42.27.dist-info/LICENSE,sha256=As9u198X-U-vph5noInuUfqsAG2zX_oXPHDmdjwlPPY,1066
-x_transformers-1.42.27.dist-info/METADATA,sha256=g6KI8a3WyHUyq9w5Tq3aQatgr89gpc5IMZ5c1zAGlHU,739
-x_transformers-1.42.27.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
-x_transformers-1.42.27.dist-info/top_level.txt,sha256=hO6KGpFuGucRNEtRfme4A_rGcM53AKwGP7RVlRIxS5Q,15
-x_transformers-1.42.27.dist-info/RECORD,,
+x_transformers-1.42.28.dist-info/LICENSE,sha256=As9u198X-U-vph5noInuUfqsAG2zX_oXPHDmdjwlPPY,1066
+x_transformers-1.42.28.dist-info/METADATA,sha256=txhDZvzsfiBEPBUg3Ipszv2cWu9sXyd7hhDz4BGsbfc,739
+x_transformers-1.42.28.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
+x_transformers-1.42.28.dist-info/top_level.txt,sha256=hO6KGpFuGucRNEtRfme4A_rGcM53AKwGP7RVlRIxS5Q,15
+x_transformers-1.42.28.dist-info/RECORD,,