Spaces:
Sleeping
Sleeping
File size: 2,878 Bytes
2bcca4d |
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 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
import gradio as gr
from pytube import YouTube
from moviepy.editor import VideoFileClip
import torch
from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
device = "cuda:0" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
model_id = "openai/whisper-tiny"
model = AutoModelForSpeechSeq2Seq.from_pretrained(
model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
)
model.to(device)
processor = AutoProcessor.from_pretrained(model_id)
pipe = pipeline(
"automatic-speech-recognition",
model=model,
tokenizer=processor.tokenizer,
feature_extractor=processor.feature_extractor,
max_new_tokens=128,
chunk_length_s=15,
batch_size=16,
return_timestamps=True,
torch_dtype=torch_dtype,
device=device,
)
def transcribe(url):
# get video and extract video
def get_video(yt_url):
try:
video = YouTube(yt_url)
video.streams.get_by_itag(22).download(filename='video.mp4')
print('Video succesfully downloaded from Youtube')
except Exception as e:
print(f'Failed to download Youtube video \nerror : {e}')
def audio_from_video(video_path):
try:
video = VideoFileClip(video_path)
audio = video.audio
audio.write_audiofile('audio.wav')
video.close()
audio.close()
except Exception as e:
print(f'Failed to extract audio from video \nerror : {e}')
url = url
video_path = './video.mp4'
get_video(url)
audio_from_video(video_path)
# transcribe audio
audio = 'audio.wav'
text_audio = pipe(audio)
chunks = text_audio['chunks']
chunks_count = len(chunks)
chunk_id = []
timestamps = []
texts = []
start_time = []
end_time = []
for i in range(0, chunks_count):
chunk_id.append(i)
texts.append(chunks[i]['text'])
start_time.append(chunks[i]['timestamp'][0])
end_time.append(chunks[i]['timestamp'][1])
chunk_length = []
for i in range(0, chunks_count-1):
chunk_length.append(round(end_time[i] - start_time[i], 3))
output = list(zip(chunk_id, chunk_length, texts, start_time, end_time))
sample_output_list = []
for sublist in output:
chunk_dict = {
"chunk_id": sublist[0],
"chunk_length": sublist[1],
"text": sublist[2],
"start_time": sublist[3],
"end_time": sublist[4]
}
sample_output_list.append(chunk_dict)
return sample_output_list
intf = gr.Interface(
fn=transcribe,
inputs = ["text"],
outputs = ["text"]
)
intf.launch() |