Spaces:
Runtime error
Runtime error
replace request with aiohttp
Browse files- README.md +1 -1
- app.py +9 -10
- requirements.txt +3 -2
README.md
CHANGED
@@ -4,7 +4,7 @@ emoji: ✍️🎥📄
|
|
4 |
colorFrom: red
|
5 |
colorTo: gray
|
6 |
sdk: gradio
|
7 |
-
sdk_version: 3.0.
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
|
|
4 |
colorFrom: red
|
5 |
colorTo: gray
|
6 |
sdk: gradio
|
7 |
+
sdk_version: 3.0.9
|
8 |
app_file: app.py
|
9 |
pinned: false
|
10 |
---
|
app.py
CHANGED
@@ -5,6 +5,8 @@ import ffmpeg
|
|
5 |
import os
|
6 |
from pathlib import Path
|
7 |
import time
|
|
|
|
|
8 |
|
9 |
|
10 |
# Set true if you're using huggingface inference API API https://huggingface.co/inference-api
|
@@ -15,7 +17,6 @@ MODEL = "facebook/wav2vec2-base-960h"
|
|
15 |
# MODEL = "patrickvonplaten/wav2vec2-large-960h-lv60-self-4-gram"
|
16 |
if API_BACKEND:
|
17 |
from dotenv import load_dotenv
|
18 |
-
import requests
|
19 |
import base64
|
20 |
import asyncio
|
21 |
load_dotenv(Path(".env"))
|
@@ -51,8 +52,8 @@ for file in samples_data:
|
|
51 |
SAMPLES.append(sample)
|
52 |
VIDEOS = list(map(lambda x: [x['video']], SAMPLES))
|
53 |
|
54 |
-
total_inferences_since_reboot =
|
55 |
-
total_cuts_since_reboot =
|
56 |
|
57 |
|
58 |
async def speech_to_text(video_file_path):
|
@@ -81,7 +82,7 @@ async def speech_to_text(video_file_path):
|
|
81 |
for tries in range(4):
|
82 |
print(f'Transcribing from API attempt {tries}')
|
83 |
try:
|
84 |
-
inference_reponse = query_api(audio_memory)
|
85 |
transcription = inference_reponse["text"].lower()
|
86 |
timestamps = [[chunk["text"].lower(), chunk["timestamp"][0], chunk["timestamp"][1]]
|
87 |
for chunk in inference_reponse['chunks']]
|
@@ -185,7 +186,7 @@ def cut_timestamps_to_video(video_in, transcription, text_in, timestamps):
|
|
185 |
return (tokens, output_video)
|
186 |
|
187 |
|
188 |
-
def query_api(audio_bytes: bytes):
|
189 |
"""
|
190 |
Query for Huggingface Inference API for Automatic Speech Recognition task
|
191 |
"""
|
@@ -199,11 +200,9 @@ def query_api(audio_bytes: bytes):
|
|
199 |
"options": {"use_gpu": False}
|
200 |
}).encode("utf-8")
|
201 |
|
202 |
-
|
203 |
-
|
204 |
-
|
205 |
-
return json_reponse
|
206 |
-
|
207 |
|
208 |
# ---- Gradio Layout -----
|
209 |
video_in = gr.Video(label="Video file")
|
|
|
5 |
import os
|
6 |
from pathlib import Path
|
7 |
import time
|
8 |
+
import aiohttp
|
9 |
+
import asyncio
|
10 |
|
11 |
|
12 |
# Set true if you're using huggingface inference API API https://huggingface.co/inference-api
|
|
|
17 |
# MODEL = "patrickvonplaten/wav2vec2-large-960h-lv60-self-4-gram"
|
18 |
if API_BACKEND:
|
19 |
from dotenv import load_dotenv
|
|
|
20 |
import base64
|
21 |
import asyncio
|
22 |
load_dotenv(Path(".env"))
|
|
|
52 |
SAMPLES.append(sample)
|
53 |
VIDEOS = list(map(lambda x: [x['video']], SAMPLES))
|
54 |
|
55 |
+
total_inferences_since_reboot = 415
|
56 |
+
total_cuts_since_reboot = 1539
|
57 |
|
58 |
|
59 |
async def speech_to_text(video_file_path):
|
|
|
82 |
for tries in range(4):
|
83 |
print(f'Transcribing from API attempt {tries}')
|
84 |
try:
|
85 |
+
inference_reponse = await query_api(audio_memory)
|
86 |
transcription = inference_reponse["text"].lower()
|
87 |
timestamps = [[chunk["text"].lower(), chunk["timestamp"][0], chunk["timestamp"][1]]
|
88 |
for chunk in inference_reponse['chunks']]
|
|
|
186 |
return (tokens, output_video)
|
187 |
|
188 |
|
189 |
+
async def query_api(audio_bytes: bytes):
|
190 |
"""
|
191 |
Query for Huggingface Inference API for Automatic Speech Recognition task
|
192 |
"""
|
|
|
200 |
"options": {"use_gpu": False}
|
201 |
}).encode("utf-8")
|
202 |
|
203 |
+
async with aiohttp.ClientSession() as session:
|
204 |
+
async with session.post(API_URL, headers=headers, data=payload) as response:
|
205 |
+
return await response.json()
|
|
|
|
|
206 |
|
207 |
# ---- Gradio Layout -----
|
208 |
video_in = gr.Video(label="Video file")
|
requirements.txt
CHANGED
@@ -1,7 +1,8 @@
|
|
1 |
torch
|
2 |
transformers
|
3 |
-
gradio==3.0.
|
4 |
datasets
|
5 |
librosa
|
6 |
ffmpeg-python
|
7 |
-
python-dotenv
|
|
|
|
1 |
torch
|
2 |
transformers
|
3 |
+
gradio==3.0.9
|
4 |
datasets
|
5 |
librosa
|
6 |
ffmpeg-python
|
7 |
+
python-dotenv
|
8 |
+
aiohttp
|