Spaces:
Build error
Build error
File size: 1,431 Bytes
ec2398c e0bc8d0 ec2398c e0bc8d0 ec2398c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 |
import os
import whisper
import torch
import time
def transcribe_segments(speakers):
print(f"Whisper models {whisper.available_models()}")
if torch.cuda.is_available():
print("transcribe_segments Using CUDA")
device = "cuda"
else:
device = "cpu"
print("transcribe_segments Using CPU")
model = whisper.load_model("tiny.en", device=device)
#model = whisper.load_model("medium.en", device="cuda")
# model = whisper.load_model("turbo", device="cuda")
#model = whisper.load_model("large-v3-turbo", device="cuda")
transcripts = []
input_file = ""
print("Transcribing ALL segments")
total_start = time.time()
for speaker in speakers:
# {'speaker': speaker, 'start': round(turn.start, 1),
# 'end': round(turn.end, 1), 'clipFile':clipName}
input_file = speaker['clipFile']
print("TRANSCRIBING " + input_file)
start = time.time()
transcript = model.transcribe(input_file)
print("Elapsed " + str(time.time() - start))
segments = transcript["segments"]
outText = ""
for segment in segments:
outText += segment['text']
transcripts.append(speaker['speaker']+" : "+outText)
os.remove(input_file)
print("Total Elapsed " + str(time.time() - total_start))
currdir= input_file[0:input_file.index('\\')]
os.rmdir(currdir)
return transcripts
|