Spaces:
Running
on
Zero
Running
on
Zero
Update infer.py
Browse files
infer.py
CHANGED
@@ -1,89 +1,62 @@
|
|
1 |
from PIL import Image
|
2 |
-
import cv2
|
3 |
import torch
|
4 |
from RealESRGAN import RealESRGAN
|
5 |
import tempfile
|
6 |
import numpy as np
|
7 |
-
|
8 |
-
import
|
9 |
-
from pydub import AudioSegment
|
10 |
-
from moviepy.editor import VideoFileClip, AudioFileClip
|
11 |
|
12 |
-
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
13 |
-
|
14 |
-
def infer_image(img: Image.Image, size_modifier: int ) -> Image.Image:
|
15 |
-
if img is None:
|
16 |
-
raise Exception("Image not uploaded")
|
17 |
-
|
18 |
-
width, height = img.size
|
19 |
-
|
20 |
-
if width >= 5000 or height >= 5000:
|
21 |
-
raise Exception("The image is too large.")
|
22 |
|
23 |
-
|
24 |
-
model.load_weights(f'weights/RealESRGAN_x{size_modifier}.pth', download=False)
|
25 |
|
26 |
-
result = model.predict(img.convert('RGB'))
|
27 |
-
print(f"Image size ({device}): {size_modifier} ... OK")
|
28 |
-
return result
|
29 |
|
30 |
def infer_video(video_filepath: str, size_modifier: int) -> str:
|
31 |
model = RealESRGAN(device, scale=size_modifier)
|
32 |
model.load_weights(f'weights/RealESRGAN_x{size_modifier}.pth', download=False)
|
33 |
-
|
34 |
-
# Extract audio from the original video file
|
35 |
-
audio = AudioSegment.from_file(video_filepath, format="mp4")
|
36 |
-
audio_array = np.array(audio.get_array_of_samples())
|
37 |
-
|
38 |
-
# Create a VideoCapture object for the video file
|
39 |
-
cap = cv2.VideoCapture(video_filepath)
|
40 |
|
41 |
-
|
|
|
42 |
tmpfile = tempfile.NamedTemporaryFile(suffix='.mp4', delete=False)
|
43 |
vid_output = tmpfile.name
|
44 |
tmpfile.close()
|
45 |
|
46 |
-
#
|
47 |
-
|
|
|
|
|
|
|
48 |
vid_output,
|
49 |
-
fourcc=
|
50 |
-
fps=cap.get(
|
51 |
-
frameSize=(int(cap.get(
|
52 |
)
|
53 |
|
54 |
-
|
55 |
-
|
56 |
-
for
|
57 |
-
# Read the next frame
|
58 |
ret, frame = cap.read()
|
59 |
if not ret:
|
60 |
break
|
61 |
|
62 |
-
|
63 |
-
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
|
64 |
frame = Image.fromarray(frame)
|
65 |
-
upscaled_frame = model.predict(frame.convert('RGB'))
|
66 |
|
67 |
-
|
|
|
68 |
upscaled_frame = np.array(upscaled_frame)
|
69 |
-
upscaled_frame =
|
70 |
|
71 |
-
# Write the upscaled frame to the output video
|
72 |
vid_writer.write(upscaled_frame)
|
73 |
|
74 |
-
# Release the VideoCapture and VideoWriter objects
|
75 |
-
cap.release()
|
76 |
vid_writer.release()
|
77 |
|
78 |
-
#
|
79 |
-
|
80 |
|
81 |
-
#
|
82 |
-
|
83 |
-
output_clip = output_clip.set_audio(audio_clip)
|
84 |
|
85 |
-
|
86 |
-
output_clip.write_videofile(f'output_{video_filepath}')
|
87 |
|
88 |
-
return
|
89 |
-
|
|
|
1 |
from PIL import Image
|
2 |
+
import cv2 as cv
|
3 |
import torch
|
4 |
from RealESRGAN import RealESRGAN
|
5 |
import tempfile
|
6 |
import numpy as np
|
7 |
+
import tqdm
|
8 |
+
import ffmpeg
|
|
|
|
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
|
11 |
+
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
|
|
12 |
|
|
|
|
|
|
|
13 |
|
14 |
def infer_video(video_filepath: str, size_modifier: int) -> str:
|
15 |
model = RealESRGAN(device, scale=size_modifier)
|
16 |
model.load_weights(f'weights/RealESRGAN_x{size_modifier}.pth', download=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
|
18 |
+
cap = cv.VideoCapture(video_filepath)
|
19 |
+
|
20 |
tmpfile = tempfile.NamedTemporaryFile(suffix='.mp4', delete=False)
|
21 |
vid_output = tmpfile.name
|
22 |
tmpfile.close()
|
23 |
|
24 |
+
# Extract audio from the input video
|
25 |
+
audio_file = video_filepath.replace(".mp4", ".wav")
|
26 |
+
ffmpeg.input(video_filepath).output(audio_file, format='wav', ac=1).run(overwrite_output=True)
|
27 |
+
|
28 |
+
vid_writer = cv.VideoWriter(
|
29 |
vid_output,
|
30 |
+
fourcc=cv.VideoWriter.fourcc(*'mp4v'),
|
31 |
+
fps=cap.get(cv.CAP_PROP_FPS),
|
32 |
+
frameSize=(int(cap.get(cv.CAP_PROP_FRAME_WIDTH)) * size_modifier, int(cap.get(cv.CAP_PROP_FRAME_HEIGHT)) * size_modifier)
|
33 |
)
|
34 |
|
35 |
+
n_frames = int(cap.get(cv.CAP_PROP_FRAME_COUNT))
|
36 |
+
|
37 |
+
for _ in tqdm.tqdm(range(n_frames)):
|
|
|
38 |
ret, frame = cap.read()
|
39 |
if not ret:
|
40 |
break
|
41 |
|
42 |
+
frame = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
|
|
|
43 |
frame = Image.fromarray(frame)
|
|
|
44 |
|
45 |
+
upscaled_frame = model.predict(frame.convert('RGB'))
|
46 |
+
|
47 |
upscaled_frame = np.array(upscaled_frame)
|
48 |
+
upscaled_frame = cv.cvtColor(upscaled_frame, cv.COLOR_RGB2BGR)
|
49 |
|
|
|
50 |
vid_writer.write(upscaled_frame)
|
51 |
|
|
|
|
|
52 |
vid_writer.release()
|
53 |
|
54 |
+
# Re-encode the video with the modified audio
|
55 |
+
ffmpeg.input(vid_output).output(video_filepath.replace(".mp4", "_upscaled.mp4"), vcodec='libx264', acodec='aac', audio_bitrate='320k').run(overwrite_output=True)
|
56 |
|
57 |
+
# Replace the original audio with the upscaled audio
|
58 |
+
ffmpeg.input(audio_file).output(video_filepath.replace(".mp4", "_upscaled.mp4"), acodec='aac', audio_bitrate='320k').run(overwrite_output=True)
|
|
|
59 |
|
60 |
+
print(f"Video file : {video_filepath}")
|
|
|
61 |
|
62 |
+
return vid_output.replace(".mp4", "_upscaled.mp4")
|
|