Fix typo in examples/whisper, Fix examples/whisper/run_faster_whisper.py

examples/whisper/run.py

@@ -373,14 +373,14 @@ def decode_dataset(
     tensorrt_llm.logger.set_level(args.log_level)
     model = WhisperTRTLLM(args.engine_dir, args.tokenizer_name, args.debug,
                           args.assets_dir)
-    normallizer = EnglishTextNormalizer()
+    normalizer = EnglishTextNormalizer()
     if args.enable_warmup:
         results, total_duration = decode_dataset(
             model,
             "hf-internal-testing/librispeech_asr_dummy",
             batch_size=args.batch_size,
             num_beams=args.num_beams,
-            normalizer=normallizer,
+            normalizer=normalizer,
             mel_filters_dir=args.assets_dir)
     start_time = time.time()
     if args.input_file:
@@ -398,7 +398,7 @@ def decode_dataset(
             dtype=args.dtype,
             batch_size=args.batch_size,
             num_beams=args.num_beams,
-            normalizer=normallizer,
+            normalizer=normalizer,
             mel_filters_dir=args.assets_dir)
     elapsed = time.time() - start_time
     results = sorted(results)

examples/whisper/run_faster_whisper.py

@@ -15,8 +15,8 @@
 import argparse
 import re
 import time
+from pathlib import Path
 
-import torch
 from datasets import load_dataset
 from torch.utils.data import DataLoader
 from whisper.normalizers import EnglishTextNormalizer
@@ -32,12 +32,14 @@
 def parse_arguments():
     parser = argparse.ArgumentParser()
     parser.add_argument('--input_file', type=str, default=None)
+    parser.add_argument('--results_dir', type=str, default='tmp')
     parser.add_argument(
         '--name',
         type=str,
         default="librispeech_dummy_faster_whisper_large_v3_warmup")
     parser.add_argument('--batch_size', type=int, default=1)
     parser.add_argument('--num_beams', type=int, default=1)
+    parser.add_argument('--enable_warmup', action='store_true')
 
     return parser.parse_args()
 
@@ -47,22 +49,20 @@ def decode_wav_file(
         model,
         text_prefix="<|startoftranscript|><|en|><|transcribe|><|notimestamps|>",
         num_beams=1,
-        normalizer=None):
-    mel, total_duration = log_mel_spectrogram(input_file_path,
-                                              device='cuda',
-                                              return_duration=True)
-    mel = mel.type(torch.float16)
-    mel = mel.unsqueeze(0)
-    predictions = model.process_batch(mel, text_prefix, num_beams)
-    prediction = predictions[0]
+        normalizer=None,
+        sample_rate=16000):
+    segments, info = model.transcribe(input_file_path,
+                                      beam_size=num_beams,
+                                      language="en")
+    prediction = " ".join([segment.text for segment in segments])
 
     # remove all special tokens in the prediction
     prediction = re.sub(r'<\|.*?\|>', '', prediction)
     if normalizer:
         prediction = normalizer(prediction)
     print(f"prediction: {prediction}")
     results = [(0, [""], prediction.split())]
-    return results, total_duration
+    return results, info.duration
 
 
 def collate_wrapper(batch):
@@ -116,16 +116,16 @@ def decode_dataset(
 
 if __name__ == '__main__':
     args = parse_arguments()
-    normallizer = EnglishTextNormalizer()
+    normalizer = EnglishTextNormalizer()
     model_size_or_path = "large-v3"
     model = WhisperModel(model_size_or_path,
                          device="cuda",
                          compute_type="float16")
-    # warmup
-    results, total_duration = decode_dataset(model,
-                                             batch_size=args.batch_size,
-                                             num_beams=args.num_beams,
-                                             normalizer=normallizer)
+    if args.enable_warmup:
+        results, total_duration = decode_dataset(model,
+                                                 batch_size=args.batch_size,
+                                                 num_beams=args.num_beams,
+                                                 normalizer=normalizer)
     start_time = time.time()
     if args.input_file:
         results, total_duration = decode_wav_file(args.input_file,
@@ -135,9 +135,11 @@ def decode_dataset(
         results, total_duration = decode_dataset(model,
                                                  batch_size=args.batch_size,
                                                  num_beams=args.num_beams,
-                                                 normalizer=normallizer)
+                                                 normalizer=normalizer)
     elapsed = time.time() - start_time
     results = sorted(results)
+
+    Path(args.results_dir).mkdir(parents=True, exist_ok=True)
     store_transcripts(filename=f"tmp/recogs-{args.name}.txt", texts=results)
 
     with open(f"tmp/errs-{args.name}.txt", "w") as f: