hyper connect the audio attention models

README.md

@@ -599,3 +599,14 @@ $ accelerate launch train.py
     url     = {https://api.semanticscholar.org/CorpusID:275405495}
 }
 ```
+
+```bibtex
+@article{Zhu2024HyperConnections,
+    title   = {Hyper-Connections},
+    author  = {Defa Zhu and Hongzhi Huang and Zihao Huang and Yutao Zeng and Yunyao Mao and Banggu Wu and Qiyang Min and Xun Zhou},
+    journal = {ArXiv},
+    year    = {2024},
+    volume  = {abs/2409.19606},
+    url     = {https://api.semanticscholar.org/CorpusID:272987528}
+}
+```

audiolm_pytorch/audiolm_pytorch.py

@@ -21,6 +21,8 @@
 
 from audiolm_pytorch.t5 import t5_encode_text, get_encoded_dim, DEFAULT_T5_NAME
 
+from hyper_connections import get_init_and_expand_reduce_stream_functions
+
 from torchaudio.functional import resample
 
 from audiolm_pytorch.soundstream import SoundStream
@@ -421,6 +423,7 @@ def __init__(
         rel_pos_bias = True,
         flash_attn = False,
         add_value_residual = True,
+        num_residual_streams = 4,
         **kwargs
     ):
         super().__init__()
@@ -438,11 +441,17 @@ def __init__(
 
         self.rel_pos_bias = RelativePositionBias(dim = dim // 2, heads = heads) if rel_pos_bias else None
 
+        # hyper connections
+
+        init_hyper_conn, self.expand_streams, self.reduce_streams = get_init_and_expand_reduce_stream_functions(num_residual_streams, disable = num_residual_streams == 1)
+
+        # layers
+
         for _ in range(depth):
             self.layers.append(nn.ModuleList([
-                Attention(dim = dim, heads = heads, dropout = attn_dropout, flash = flash_attn, causal = True, **kwargs),
-                Attention(dim = dim, heads = heads, dropout = attn_dropout, dim_context = dim_context, flash = flash_attn, num_null_kv = 1, norm_context = True, **kwargs) if cross_attend else None,
-                FeedForward(dim = dim, dropout = ff_dropout)
+                init_hyper_conn(dim = dim, branch = Attention(dim = dim, heads = heads, dropout = attn_dropout, flash = flash_attn, causal = True, **kwargs)),
+                init_hyper_conn(dim = dim, branch = Attention(dim = dim, heads = heads, dropout = attn_dropout, dim_context = dim_context, flash = flash_attn, num_null_kv = 1, norm_context = True, **kwargs)) if cross_attend else None,
+                init_hyper_conn(dim = dim, branch = FeedForward(dim = dim, dropout = ff_dropout))
             ]))
 
         self.norm = LayerNorm(dim)
@@ -510,6 +519,10 @@ def forward(
         self_attn_value_residual = None
         cross_attn_value_residual = None
 
+        # expand residual streams
+
+        x = self.expand_streams(x)
+
         # transformer layers
 
         for attn, cross_attn, ff in self.layers:
@@ -523,18 +536,21 @@ def forward(
 
             new_kv_cache.append(layer_kv_cache)
 
-            x = x + residual
-
             if exists(cross_attn):
                 assert exists(context)
 
-                cross_attend_out, values = cross_attn(x, context = context, mask = context_mask, return_values = True, value_residual = cross_attn_value_residual)
-                x = cross_attend_out + x
+                x, values = cross_attn(x, context = context, mask = context_mask, return_values = True, value_residual = cross_attn_value_residual)
 
                 if self.add_value_residual:
                     cross_attn_value_residual = default(cross_attn_value_residual, values)
 
-            x = ff(x) + x
+            x = ff(x)
+
+        # reduce residual streams
+
+        x = self.reduce_streams(x)
+
+        # final norm
 
         x = self.norm(x)
 

audiolm_pytorch/version.py

@@ -1 +1 @@
-__version__ = '2.3.1'
+__version__ = '2.4.0'

setup.py

@@ -27,6 +27,7 @@
     'fairseq',
     'wandb',
     'gateloop-transformer>=0.2.3',
+    'hyper-connections>=0.1.8',
     'joblib',
     'local-attention>=1.9.0',
     'pytorch-warmup',