Spaces:
Sleeping
Sleeping
AshDavid12
commited on
Commit
·
47058ca
1
Parent(s):
6af664b
trying ivrit model
Browse files- infer.py +80 -19
- requirements.txt +0 -2
infer.py
CHANGED
|
@@ -1,33 +1,94 @@
|
|
| 1 |
import faster_whisper
|
| 2 |
import requests
|
| 3 |
-
|
| 4 |
-
import
|
| 5 |
|
| 6 |
# Load the faster-whisper model that supports Hebrew
|
| 7 |
model = faster_whisper.WhisperModel("ivrit-ai/faster-whisper-v2-d4")
|
| 8 |
|
| 9 |
-
# URL of the
|
| 10 |
-
audio_url = "https://github.com/
|
| 11 |
|
| 12 |
-
# Download the
|
| 13 |
response = requests.get(audio_url)
|
| 14 |
if response.status_code != 200:
|
| 15 |
raise Exception("Failed to download audio file")
|
| 16 |
|
| 17 |
-
#
|
| 18 |
-
|
|
|
|
|
|
|
| 19 |
|
| 20 |
-
#
|
| 21 |
-
|
| 22 |
-
wav_audio = io.BytesIO()
|
| 23 |
-
audio.export(wav_audio, format="wav")
|
| 24 |
-
wav_audio.seek(0) # Reset the pointer to the beginning of the buffer
|
| 25 |
|
| 26 |
-
#
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
segments, info = model.transcribe(temp_wav_file, language="he")
|
| 30 |
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import faster_whisper
|
| 2 |
import requests
|
| 3 |
+
import tempfile
|
| 4 |
+
import os
|
| 5 |
|
| 6 |
# Load the faster-whisper model that supports Hebrew
|
| 7 |
model = faster_whisper.WhisperModel("ivrit-ai/faster-whisper-v2-d4")
|
| 8 |
|
| 9 |
+
# URL of the audio file (replace this with the actual URL of your audio)
|
| 10 |
+
audio_url = "https://github.com/AshDavid12/runpod-serverless-forked/blob/main/me-hebrew.wav"
|
| 11 |
|
| 12 |
+
# Download the audio file from the URL
|
| 13 |
response = requests.get(audio_url)
|
| 14 |
if response.status_code != 200:
|
| 15 |
raise Exception("Failed to download audio file")
|
| 16 |
|
| 17 |
+
# Create a temporary file to store the audio
|
| 18 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_audio_file:
|
| 19 |
+
tmp_audio_file.write(response.content)
|
| 20 |
+
tmp_audio_file_path = tmp_audio_file.name
|
| 21 |
|
| 22 |
+
# Perform the transcription
|
| 23 |
+
segments, info = model.transcribe(tmp_audio_file_path, language="he")
|
|
|
|
|
|
|
|
|
|
| 24 |
|
| 25 |
+
# Print transcription results
|
| 26 |
+
for segment in segments:
|
| 27 |
+
print(f"[{segment.start:.2f}s - {segment.end:.2f}s] {segment.text}")
|
|
|
|
| 28 |
|
| 29 |
+
# Clean up the temporary file
|
| 30 |
+
os.remove(tmp_audio_file_path)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
# import torch
|
| 51 |
+
# from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
| 52 |
+
# import requests
|
| 53 |
+
# import soundfile as sf
|
| 54 |
+
# import io
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
# # Load the Whisper model and processor from Hugging Face Model Hub
|
| 58 |
+
# model_name = "openai/whisper-base"
|
| 59 |
+
# processor = WhisperProcessor.from_pretrained(model_name)
|
| 60 |
+
# model = WhisperForConditionalGeneration.from_pretrained(model_name)
|
| 61 |
+
#
|
| 62 |
+
# # Use GPU if available, otherwise use CPU
|
| 63 |
+
# device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 64 |
+
# model.to(device)
|
| 65 |
+
#
|
| 66 |
+
# # URL of the audio file
|
| 67 |
+
# audio_url = "https://www.signalogic.com/melp/EngSamples/Orig/male.wav"
|
| 68 |
+
#
|
| 69 |
+
# # Download the audio file
|
| 70 |
+
# response = requests.get(audio_url)
|
| 71 |
+
# audio_data = io.BytesIO(response.content)
|
| 72 |
+
#
|
| 73 |
+
# # Read the audio using soundfile
|
| 74 |
+
# audio_input, _ = sf.read(audio_data)
|
| 75 |
+
#
|
| 76 |
+
# # Preprocess the audio for Whisper
|
| 77 |
+
# inputs = processor(audio_input, return_tensors="pt", sampling_rate=16000)
|
| 78 |
+
# attention_mask = inputs['input_features'].ne(processor.tokenizer.pad_token_id).long()
|
| 79 |
+
#
|
| 80 |
+
# # Move inputs and attention mask to the correct device
|
| 81 |
+
# inputs = {key: value.to(device) for key, value in inputs.items()}
|
| 82 |
+
# attention_mask = attention_mask.to(device)
|
| 83 |
+
#
|
| 84 |
+
# # Generate the transcription with attention mask
|
| 85 |
+
# with torch.no_grad():
|
| 86 |
+
# predicted_ids = model.generate(
|
| 87 |
+
# inputs["input_features"],
|
| 88 |
+
# attention_mask=attention_mask # Pass attention mask explicitly
|
| 89 |
+
# )
|
| 90 |
+
# # Decode the transcription
|
| 91 |
+
# transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
|
| 92 |
+
#
|
| 93 |
+
# # Print the transcription result
|
| 94 |
+
# print("Transcription:", transcription)
|
requirements.txt
CHANGED
|
@@ -4,6 +4,4 @@ requests
|
|
| 4 |
transformers
|
| 5 |
soundfile
|
| 6 |
faster-whisper
|
| 7 |
-
pydub
|
| 8 |
-
ffmpeg
|
| 9 |
|
|
|
|
| 4 |
transformers
|
| 5 |
soundfile
|
| 6 |
faster-whisper
|
|
|
|
|
|
|
| 7 |
|