fix some model

model.py

@@ -45,10 +45,7 @@ class MusicAudioClassifier(pl.LightningModule):
     def _process_audio_batch(self, x: torch.Tensor) -> torch.Tensor:
         B, S = x.shape[:2]  # [B, S, C, M, T] or [B, S, C, T] for wav, [B, S, 1?, embsize] for emb
         x = x.view(B*S, *x.shape[2:])  # [B*S, C, M, T] 
-        if self.is_emb == False:
-            _, embeddings = self.emb_model(x)  # [B*S, emb_dim]
-        else:
-            embeddings = x
+        embeddings=x
         if embeddings.dim() == 3:
             pooled_features = embeddings.mean(dim=1) # transformer
         else:

requirements.txt

@@ -10,4 +10,5 @@ scipy>=1.10.0
 soundfile>=0.12.0
 datasets>=2.0.0
 accelerate>=0.20.0
-spaces
+spaces
+nnAudio