Update app.py

app.py

@@ -45,14 +45,14 @@ class AudioMathRockModel(nn.Module):
         self.amplitude_to_db = torchaudio.transforms.AmplitudeToDB(stype='power', top_db=80.0)
         
         self.cnn_extractor = nn.Sequential(
-            nn.KeepConv = nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),
+            nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),
             nn.BatchNorm2d(32), nn.ReLU(), nn.MaxPool2d(2, 2),
-            nn.KeepConv2d = nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),
+            nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),
             nn.BatchNorm2d(64), nn.ReLU(), nn.MaxPool2d(2, 2),
-            nn.KeepConv3d = nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1),
+            nn.Conv2d(64, 128, kernel_size=3, stride=1, padding=1),
             nn.BatchNorm2d(128), nn.ReLU(), nn.AdaptiveAvgPool2d((1, 1)),
             nn.Flatten(),
-            nn.KeepLinear = nn.Linear(128, 512)
+            nn.Linear(128, 512)
         )
         
         self.wlm_proj = nn.Linear(768, 512)
@@ -72,7 +72,7 @@ class AudioMathRockModel(nn.Module):
         )
         self.vibe_head = nn.Linear(512, len(VIBE_LABELS))
         self.int_head  = nn.Linear(512, len(INTENSITY_LABELS))
-        self.tmp_head  = nn.Linear(512, len(TEMPO_CLASSES))
+        self.tmp_head  = nn.Linear(512, len(TEMPO_LABELS))
 
     def forward(self, wavlm_values, clap_values):
         wavlm_feats = self.wavlm(wavlm_values).last_hidden_state.mean(dim=1)