Spaces:
Running
Running
Michael Hu
commited on
Commit
·
b591083
1
Parent(s):
6f92dbc
refactor(stt): replace whisper with faster-whisper for improved performance
Browse filesSwitch from transformers-based whisper implementation to faster-whisper for better speed and memory efficiency. The new implementation removes torch dependency for device detection and uses optimized compute types based on available hardware.
- utils/stt.py +38 -46
utils/stt.py
CHANGED
@@ -11,10 +11,8 @@ from abc import ABC, abstractmethod
|
|
11 |
|
12 |
logger = logging.getLogger(__name__)
|
13 |
|
14 |
-
import
|
15 |
-
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor
|
16 |
from pydub import AudioSegment
|
17 |
-
import soundfile as sf
|
18 |
|
19 |
class ASRModel(ABC):
|
20 |
"""Base class for ASR models"""
|
@@ -43,64 +41,58 @@ class ASRModel(ABC):
|
|
43 |
|
44 |
|
45 |
class WhisperModel(ASRModel):
|
46 |
-
"""Whisper ASR model implementation"""
|
47 |
|
48 |
def __init__(self):
|
49 |
self.model = None
|
50 |
-
|
51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
|
53 |
def load_model(self):
|
54 |
-
"""Load Whisper model"""
|
55 |
-
logger.info("Loading Whisper model")
|
56 |
logger.info(f"Using device: {self.device}")
|
|
|
57 |
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
)
|
64 |
-
|
65 |
-
self.processor = AutoProcessor.from_pretrained("unsloth/whisper-large-v3")
|
66 |
-
logger.info("Whisper model loaded successfully")
|
67 |
|
68 |
def transcribe(self, audio_path):
|
69 |
-
"""Transcribe audio using Whisper"""
|
70 |
-
if self.model is None
|
71 |
self.load_model()
|
72 |
|
73 |
wav_path = self.preprocess_audio(audio_path)
|
74 |
|
75 |
-
#
|
76 |
-
logger.info("
|
77 |
-
|
78 |
-
|
79 |
-
|
|
|
|
|
|
|
80 |
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
stride_length_s=10
|
89 |
-
).to(self.device)
|
90 |
-
|
91 |
-
# Transcription
|
92 |
-
logger.info("Generating transcription")
|
93 |
-
with torch.no_grad():
|
94 |
-
# Add max_length parameter to allow for longer outputs
|
95 |
-
outputs = self.model.generate(
|
96 |
-
**inputs,
|
97 |
-
language="en",
|
98 |
-
task="transcribe",
|
99 |
-
max_length=448, # Explicitly set max output length
|
100 |
-
no_repeat_ngram_size=3 # Prevent repetition in output
|
101 |
-
)
|
102 |
|
103 |
-
result =
|
104 |
logger.info(f"Transcription completed successfully")
|
105 |
return result
|
106 |
|
|
|
11 |
|
12 |
logger = logging.getLogger(__name__)
|
13 |
|
14 |
+
from faster_whisper import WhisperModel as FasterWhisperModel
|
|
|
15 |
from pydub import AudioSegment
|
|
|
16 |
|
17 |
class ASRModel(ABC):
|
18 |
"""Base class for ASR models"""
|
|
|
41 |
|
42 |
|
43 |
class WhisperModel(ASRModel):
|
44 |
+
"""Faster Whisper ASR model implementation"""
|
45 |
|
46 |
def __init__(self):
|
47 |
self.model = None
|
48 |
+
# Check for CUDA availability without torch dependency
|
49 |
+
try:
|
50 |
+
import torch
|
51 |
+
self.device = "cuda" if torch.cuda.is_available() else "cpu"
|
52 |
+
except ImportError:
|
53 |
+
# Fallback to CPU if torch is not available
|
54 |
+
self.device = "cpu"
|
55 |
+
self.compute_type = "float16" if self.device == "cuda" else "int8"
|
56 |
|
57 |
def load_model(self):
|
58 |
+
"""Load Faster Whisper model"""
|
59 |
+
logger.info("Loading Faster Whisper model")
|
60 |
logger.info(f"Using device: {self.device}")
|
61 |
+
logger.info(f"Using compute type: {self.compute_type}")
|
62 |
|
63 |
+
# Use large-v3 model with appropriate compute type based on device
|
64 |
+
self.model = FasterWhisperModel(
|
65 |
+
"large-v3",
|
66 |
+
device=self.device,
|
67 |
+
compute_type=self.compute_type
|
68 |
+
)
|
69 |
+
logger.info("Faster Whisper model loaded successfully")
|
|
|
|
|
70 |
|
71 |
def transcribe(self, audio_path):
|
72 |
+
"""Transcribe audio using Faster Whisper"""
|
73 |
+
if self.model is None:
|
74 |
self.load_model()
|
75 |
|
76 |
wav_path = self.preprocess_audio(audio_path)
|
77 |
|
78 |
+
# Transcription with Faster Whisper
|
79 |
+
logger.info("Generating transcription with Faster Whisper")
|
80 |
+
segments, info = self.model.transcribe(
|
81 |
+
wav_path,
|
82 |
+
beam_size=5,
|
83 |
+
language="en",
|
84 |
+
task="transcribe"
|
85 |
+
)
|
86 |
|
87 |
+
logger.info(f"Detected language '{info.language}' with probability {info.language_probability}")
|
88 |
+
|
89 |
+
# Collect all segments into a single text
|
90 |
+
result_text = ""
|
91 |
+
for segment in segments:
|
92 |
+
result_text += segment.text + " "
|
93 |
+
logger.debug(f"[{segment.start:.2f}s -> {segment.end:.2f}s] {segment.text}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
94 |
|
95 |
+
result = result_text.strip()
|
96 |
logger.info(f"Transcription completed successfully")
|
97 |
return result
|
98 |
|