lol...

vall_e/emb/process.py

@@ -84,18 +84,10 @@ def process_batched_jobs( jobs, speaker_id="", device=None, raise_exceptions=Tru
 	# sort to avoid egregious padding
 	jobs = sorted(jobs, key=lambda x: x[1].shape[-1], reverse=True)
 
-	buffer = []
 	batches = []
-
-	for job in jobs:
-		buffer.append(job)
-		if len(buffer) >= batch_size:
-			batches.append(buffer)
-			buffer = []
-	
-	if buffer:
-		batches.append(buffer)
-		buffer = []
+	while jobs:
+		batches.append(jobs[:batch_size])
+		jobs = jobs[batch_size:]
 
 	for batch in tqdm(batches, desc=f'Quantizing {speaker_id} (batch size: {batch_size})'):
 		wavs = []
@@ -283,11 +275,14 @@ def process(
 			if f'{group_name}/{speaker_id}' not in dataset:
 				dataset.append(f'{group_name}/{speaker_id}')
 
-			jobs = []
 
+			jobs = []
 			use_slices = slice == True or (slice == "auto" and len(metadata.keys()) == 1) or group_name in always_slice_groups
+			if min_utterances and len(metadata.keys()) < min_utterances:
+				continue
 
 			for filename in sorted(metadata.keys()):
+
 				inpath = Path(f'./{input_audio}/{group_name}/{speaker_id}/{filename}')
 
 				"""
@@ -335,9 +330,6 @@ def process(
 				else:
 					i = 0
 					presliced = not inpath.exists()
-					
-					if min_utterances and len(metadata[filename]["segments"]) < min_utterances:
-						continue
 
 					for segment in metadata[filename]["segments"]:
 						id = pad(i, 4)