2024-01-26 04:26:20 +00:00
|
|
|
import torch
|
2024-01-26 07:38:46 +00:00
|
|
|
import numpy as np
|
2024-01-26 19:12:09 +00:00
|
|
|
import os
|
|
|
|
import time
|
|
|
|
import random
|
|
|
|
import string
|
2024-01-28 23:31:49 +00:00
|
|
|
import cv2
|
2024-01-26 07:38:46 +00:00
|
|
|
|
2024-01-26 04:22:04 +00:00
|
|
|
|
2024-01-28 16:27:43 +00:00
|
|
|
def compute_cond_mark(cond_or_uncond, sigmas):
|
|
|
|
cond_or_uncond_size = int(sigmas.shape[0])
|
|
|
|
|
|
|
|
cond_mark = []
|
|
|
|
for cx in cond_or_uncond:
|
|
|
|
cond_mark += [cx] * cond_or_uncond_size
|
|
|
|
|
|
|
|
cond_mark = torch.Tensor(cond_mark).to(sigmas)
|
|
|
|
return cond_mark
|
|
|
|
|
|
|
|
|
2024-01-26 19:12:09 +00:00
|
|
|
def generate_random_filename(extension=".txt"):
|
|
|
|
timestamp = time.strftime("%Y%m%d-%H%M%S")
|
|
|
|
random_string = ''.join(random.choices(string.ascii_lowercase + string.digits, k=5))
|
|
|
|
filename = f"{timestamp}-{random_string}{extension}"
|
|
|
|
return filename
|
|
|
|
|
|
|
|
|
2024-01-26 07:38:46 +00:00
|
|
|
@torch.no_grad()
|
|
|
|
@torch.inference_mode()
|
|
|
|
def pytorch_to_numpy(x):
|
|
|
|
return [np.clip(255. * y.cpu().numpy(), 0, 255).astype(np.uint8) for y in x]
|
|
|
|
|
|
|
|
|
|
|
|
@torch.no_grad()
|
|
|
|
@torch.inference_mode()
|
|
|
|
def numpy_to_pytorch(x):
|
|
|
|
y = x.astype(np.float32) / 255.0
|
|
|
|
y = y[None]
|
|
|
|
y = np.ascontiguousarray(y.copy())
|
|
|
|
y = torch.from_numpy(y).float()
|
|
|
|
return y
|
2024-01-26 19:12:09 +00:00
|
|
|
|
|
|
|
|
|
|
|
def write_images_to_mp4(frame_list: list, filename=None, fps=6):
|
|
|
|
from modules.paths_internal import default_output_dir
|
|
|
|
|
|
|
|
video_folder = os.path.join(default_output_dir, 'svd')
|
|
|
|
os.makedirs(video_folder, exist_ok=True)
|
|
|
|
|
|
|
|
if filename is None:
|
|
|
|
filename = generate_random_filename('.mp4')
|
|
|
|
|
|
|
|
full_path = os.path.join(video_folder, filename)
|
|
|
|
|
|
|
|
try:
|
|
|
|
import av
|
|
|
|
except ImportError:
|
|
|
|
from launch import run_pip
|
|
|
|
run_pip(
|
|
|
|
"install imageio[pyav]",
|
|
|
|
"imageio[pyav]",
|
|
|
|
)
|
|
|
|
import av
|
|
|
|
|
|
|
|
options = {
|
|
|
|
"crf": str(23)
|
|
|
|
}
|
|
|
|
|
|
|
|
output = av.open(full_path, "w")
|
|
|
|
|
|
|
|
stream = output.add_stream('libx264', fps, options=options)
|
|
|
|
stream.width = frame_list[0].shape[1]
|
|
|
|
stream.height = frame_list[0].shape[0]
|
|
|
|
for img in frame_list:
|
|
|
|
frame = av.VideoFrame.from_ndarray(img)
|
|
|
|
packet = stream.encode(frame)
|
|
|
|
output.mux(packet)
|
|
|
|
packet = stream.encode(None)
|
|
|
|
output.mux(packet)
|
|
|
|
output.close()
|
|
|
|
|
|
|
|
return full_path
|
2024-01-28 23:31:49 +00:00
|
|
|
|
|
|
|
|
|
|
|
def pad64(x):
|
|
|
|
return int(np.ceil(float(x) / 64.0) * 64 - x)
|
|
|
|
|
|
|
|
|
|
|
|
def safer_memory(x):
|
|
|
|
# Fix many MAC/AMD problems
|
|
|
|
return np.ascontiguousarray(x.copy()).copy()
|
|
|
|
|
|
|
|
|
|
|
|
def resize_image_with_pad(img, resolution):
|
|
|
|
H_raw, W_raw, _ = img.shape
|
|
|
|
k = float(resolution) / float(min(H_raw, W_raw))
|
|
|
|
interpolation = cv2.INTER_CUBIC if k > 1 else cv2.INTER_AREA
|
|
|
|
H_target = int(np.round(float(H_raw) * k))
|
|
|
|
W_target = int(np.round(float(W_raw) * k))
|
|
|
|
img = cv2.resize(img, (W_target, H_target), interpolation=interpolation)
|
|
|
|
H_pad, W_pad = pad64(H_target), pad64(W_target)
|
|
|
|
img_padded = np.pad(img, [[0, H_pad], [0, W_pad], [0, 0]], mode='edge')
|
|
|
|
|
|
|
|
def remove_pad(x):
|
|
|
|
return safer_memory(x[:H_target, :W_target])
|
|
|
|
|
|
|
|
return safer_memory(img_padded), remove_pad
|