Spaces:
Runtime error
Runtime error
fix fire inference
Browse files- app.py +4 -5
- climategan/fire.py +6 -5
- requirements.txt +1 -0
app.py
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
# thank you @NimaBoscarino
|
| 3 |
|
| 4 |
import os
|
|
|
|
| 5 |
from textwrap import dedent
|
| 6 |
from urllib import parse
|
| 7 |
-
from requests import get
|
| 8 |
|
| 9 |
import googlemaps
|
| 10 |
import gradio as gr
|
|
@@ -20,8 +20,8 @@ from gradio.components import (
|
|
| 20 |
Row,
|
| 21 |
Textbox,
|
| 22 |
)
|
|
|
|
| 23 |
from skimage import io
|
| 24 |
-
from datetime import datetime
|
| 25 |
|
| 26 |
from climategan_wrapper import ClimateGAN
|
| 27 |
|
|
@@ -76,9 +76,9 @@ TEXTS = [
|
|
| 76 |
|
|
| 77 |
Read the original
|
| 78 |
<a
|
| 79 |
-
href='https://
|
| 80 |
target='_blank'>
|
| 81 |
-
ICLR
|
| 82 |
</a>
|
| 83 |
</p>
|
| 84 |
"""
|
|
@@ -217,7 +217,6 @@ def predict(cg: ClimateGAN, api_key):
|
|
| 217 |
|
| 218 |
|
| 219 |
if __name__ == "__main__":
|
| 220 |
-
|
| 221 |
ip = get("https://api.ipify.org").content.decode("utf8")
|
| 222 |
print("My public IP address is: {}".format(ip))
|
| 223 |
|
|
|
|
| 2 |
# thank you @NimaBoscarino
|
| 3 |
|
| 4 |
import os
|
| 5 |
+
from datetime import datetime
|
| 6 |
from textwrap import dedent
|
| 7 |
from urllib import parse
|
|
|
|
| 8 |
|
| 9 |
import googlemaps
|
| 10 |
import gradio as gr
|
|
|
|
| 20 |
Row,
|
| 21 |
Textbox,
|
| 22 |
)
|
| 23 |
+
from requests import get
|
| 24 |
from skimage import io
|
|
|
|
| 25 |
|
| 26 |
from climategan_wrapper import ClimateGAN
|
| 27 |
|
|
|
|
| 76 |
|
|
| 77 |
Read the original
|
| 78 |
<a
|
| 79 |
+
href='https://arxiv.org/abs/2110.02871'
|
| 80 |
target='_blank'>
|
| 81 |
+
ICLR 2022 ClimateGAN paper
|
| 82 |
</a>
|
| 83 |
</p>
|
| 84 |
"""
|
|
|
|
| 217 |
|
| 218 |
|
| 219 |
if __name__ == "__main__":
|
|
|
|
| 220 |
ip = get("https://api.ipify.org").content.decode("utf8")
|
| 221 |
print("My public IP address is: {}".format(ip))
|
| 222 |
|
climategan/fire.py
CHANGED
|
@@ -1,7 +1,8 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
import torch.nn.functional as F
|
| 3 |
import random
|
|
|
|
| 4 |
import kornia
|
|
|
|
|
|
|
| 5 |
from torchvision.transforms.functional import adjust_brightness, adjust_contrast
|
| 6 |
|
| 7 |
from climategan.tutils import normalize, retrieve_sky_mask
|
|
@@ -105,9 +106,9 @@ def add_fire(x, seg_preds, fire_opts):
|
|
| 105 |
kernel_size = (fire_opts.get("kernel_size", 301), fire_opts.get("kernel_size", 301))
|
| 106 |
sigma = (fire_opts.get("kernel_sigma", 150.5), fire_opts.get("kernel_sigma", 150.5))
|
| 107 |
border_type = "reflect"
|
| 108 |
-
kernel =
|
| 109 |
-
|
| 110 |
-
|
| 111 |
sky_mask = filter2d(sky_mask, kernel, border_type)
|
| 112 |
|
| 113 |
filter_ = torch.ones(wildfire_tens.shape, device=x.device)
|
|
|
|
|
|
|
|
|
|
| 1 |
import random
|
| 2 |
+
|
| 3 |
import kornia
|
| 4 |
+
import torch
|
| 5 |
+
import torch.nn.functional as F
|
| 6 |
from torchvision.transforms.functional import adjust_brightness, adjust_contrast
|
| 7 |
|
| 8 |
from climategan.tutils import normalize, retrieve_sky_mask
|
|
|
|
| 106 |
kernel_size = (fire_opts.get("kernel_size", 301), fire_opts.get("kernel_size", 301))
|
| 107 |
sigma = (fire_opts.get("kernel_sigma", 150.5), fire_opts.get("kernel_sigma", 150.5))
|
| 108 |
border_type = "reflect"
|
| 109 |
+
kernel = kornia.filters.kernels.get_gaussian_kernel2d(kernel_size, sigma)
|
| 110 |
+
if kernel.ndim == 2:
|
| 111 |
+
kernel = kernel.unsqueeze(0)
|
| 112 |
sky_mask = filter2d(sky_mask, kernel, border_type)
|
| 113 |
|
| 114 |
filter_ = torch.ones(wildfire_tens.shape, device=x.device)
|
requirements.txt
CHANGED
|
@@ -2,6 +2,7 @@ gradio==3.44.1
|
|
| 2 |
torch
|
| 3 |
torch-optimizer
|
| 4 |
torchvision
|
|
|
|
| 5 |
addict
|
| 6 |
aiohttp
|
| 7 |
aiosignal
|
|
|
|
| 2 |
torch
|
| 3 |
torch-optimizer
|
| 4 |
torchvision
|
| 5 |
+
accelerate
|
| 6 |
addict
|
| 7 |
aiohttp
|
| 8 |
aiosignal
|