-
Notifications
You must be signed in to change notification settings - Fork 2
/
inference_2_monthly.py
187 lines (157 loc) · 9.35 KB
/
inference_2_monthly.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
"""The inference script for DDIM model.
"""
import argparse
import logging
import os
import pickle
import warnings
from collections import OrderedDict, defaultdict
import numpy as np
import torch
from torch.nn.functional import mse_loss, l1_loss
from torch.utils.data import DataLoader
from data.mydataset_patch import SR3_Dataset_all
import model
from configs import Config, get_current_datetime
from utils import dict2str, setup_logger, construct_and_save_wbd_plots, \
construct_mask, set_seeds,psnr
import xarray as xr
import glob
import matplotlib.pyplot as plt
import matplotlib
matplotlib.use('Agg')
warnings.filterwarnings("ignore")
def loop_prediction(start_year,end_year,ddim_steps):
max_normal = np.load("/home/data/downscaling/downscaling_1023/data/train_dataset/max_new_10.npy", mmap_mode='r+')[list_idx]
min_normal = np.load("/home/data/downscaling/downscaling_1023/data/train_dataset/min_new_10.npy", mmap_mode='r+')[list_idx]
for year in range(start_year,end_year):
all_data=[]
member_data=[]
idx=0
batch=12
data_paths = sorted(glob.glob("/home/data/downscaling/downscaling_1023/data/test_dataset/sl/*{0}_monthly*npy".format(year)))
information=xr.open_dataset("/home/data/downscaling/downscaling_1023/data/ERA_deal/different_grid/10km/ERA5_land_10km_East_china_{0}_monthly.nc".format(year))
val_logger.info(f"Dataset- Testing] is created=========year: "+str(year))
val_dataset = SR3_Dataset_all(land_01_path,mask_path,lr_paths=data_paths,var=variable_name)
val_loader = DataLoader(val_dataset, batch_size=batch,shuffle=False, num_workers=3)
idx=0
with torch.no_grad():
for val_data in val_loader:
if idx % 5==0:
print(idx*batch)
idx = idx+1
diffusion.feed_data(val_data)
diffusion.infer_patch(continuous=False,use_ddim=True,use_dpm_solver=False,ddim_steps=ddim_steps)#infer_patch这个是平均,v2是不带平均,两个需要实验看看
visuals = diffusion.get_current_visuals(only_rec=True)
pred_norm=visuals["SR"]
all_data.append(pred_norm)#
if need_member:
sr_candidates = diffusion.infer_generate_multiple_candidates(n=sample_size,use_ddim=True,use_dpm_solver=False,ddim_steps=ddim_steps)
# mem_candidate = sr_candidates* torch.from_numpy(std_hr).float() +torch.from_numpy(mean_hr).float() # [n,B, C, H, W]
member_data.append(sr_candidates)
if variable_name == "tp":
new_data=torch.clamp(torch.cat(all_data,dim=0),0,5).numpy()
new_data=np.exp(new_data[:,0,:,:])-1
new_data[new_data<0]=0
else:
new_data=torch.clamp(torch.cat(all_data,dim=0),0,1).numpy()
new_data=new_data[:,0,:,:]*(max_normal-min_normal)+min_normal
new_data=new_data*std_hr[list_idx]+mean_hr[list_idx]
if need_member:
if variable_name == "tp":
new_member_data=torch.clamp(torch.cat(member_data,dim=1),0,5).numpy()
new_member_data=np.exp(new_member_data[:,:,0,:,:])-1
new_member_data[new_member_data<0]=0
else:
new_member_data=torch.clamp(torch.cat(member_data,dim=1),0,1).numpy()
new_member_data=new_member_data[:,:,0,:,:]*(max_normal-min_normal)+min_normal
new_member_data=new_member_data*std_hr[list_idx]+mean_hr[list_idx]
dataset_new=xr.Dataset({
variable_name:(["time", "latitude", "longitude"],new_data[:,:,:])
},
coords={
"time":information.time,
"latitude":information.latitude,
"longitude":information.longitude}
)
dataset_new.to_netcdf(result_path+"/"+"single_results/predict_{0}_.nc".format(year))
# np.save(result_path+"/"+f"single_results/predict_{year}_{locs}_.npy",new_data)
print(new_member_data.shape)
if need_member:
dataset_new=xr.Dataset({
variable_name:(["member","time", "latitude", "longitude"],new_member_data[:,:,:,:]) ,#if tp need exp
# "v10":(["time", "latitude", "longitude"], new_data[:,1,:,:]),
# "t2m":(["time", "latitude", "longitude"], new_data[:,2,:,:]),
# "sp":(["time", "latitude", "longitude"], new_data[:,3,:,:]),
# "tp":(["time", "latitude", "longitude"], np.exp(new_data[:,4,:,:])-1),
},
coords={"member":np.arange(sample_size),
"time":information.time,
"latitude":information.latitude,
"longitude":information.longitude}
)
dataset_new.to_netcdf(result_path+"/"+"multi_member/predict_{0}_.nc".format(year))
# np.save(result_path+"/"+f"multi_member/predict_{year}_{locs}_.npy",new_member_data)
val_logger.info(f"{year} member data is finished.")
if __name__ == "__main__":
set_seeds()
parser = argparse.ArgumentParser()
parser.add_argument("-c", "--config", type=str, help="JSON file for configuration")
parser.add_argument("-p", "--phase", type=str, choices=["train", "val"],
help="Run either training or validation(inference).", default="train")
parser.add_argument("-gpu", "--gpu_ids", type=str, default=None)
parser.add_argument("-var", "--variable_name", type=str, default=None)
parser.add_argument("-member", "--member", type=str, default=None)
step_s=25
need_member=True
inference_version="v1"
args = parser.parse_args()
configs = Config(args)
variable_name=args.variable_name
sample_size=int(args.member)
variable_list={"u":0,"v":1,"sp":3,"t2m":2,"tp":4}
list_idx=variable_list[variable_name]
if variable_name == "tp":
mean_hr=0
std_hr=1
else:
mean_hr=np.load("/home/data/downscaling/downscaling_1023/data/train_dataset/mean&std/hr_mean.npy").transpose(2,0,1)
std_hr=np.load("/home/data/downscaling/downscaling_1023/data/train_dataset/mean&std/hr_std.npy").transpose(2,0,1)
torch.backends.cudnn.enabled = True
torch.backends.cudnn.benchmark = True
# index_list=got_index_list(patch_size=128)
test_root = f"{configs.experiments_root}/test_{sample_size}member_{step_s}_{get_current_datetime()}"
os.makedirs(test_root, exist_ok=True)
setup_logger("test", test_root, "test", screen=True)
val_logger = logging.getLogger("test")
val_logger.info(dict2str(configs.get_hyperparameters_as_dict()))
land_01_path="/home/data/downscaling/downscaling_1023/data/land10.npy"
mask_path="/home/data/downscaling/downscaling_1023/data/mask10.npy"
diffusion = model.create_model(in_channel=configs.in_channel, out_channel=configs.out_channel,
norm_groups=configs.norm_groups, inner_channel=configs.inner_channel,
channel_multiplier=configs.channel_multiplier, attn_res=configs.attn_res,
res_blocks=configs.res_blocks, dropout=configs.dropout,
diffusion_loss=configs.diffusion_loss, conditional=configs.conditional,
gpu_ids=configs.gpu_ids, distributed=configs.distributed,
init_method=configs.init_method, train_schedule=configs.train_schedule,
train_n_timestep=configs.train_n_timestep,
train_linear_start=configs.train_linear_start,
train_linear_end=configs.train_linear_end,
val_schedule=configs.val_schedule, val_n_timestep=configs.val_n_timestep,
val_linear_start=configs.val_linear_start, val_linear_end=configs.val_linear_end,
finetune_norm=configs.finetune_norm, optimizer=None, amsgrad=configs.amsgrad,
learning_rate=configs.lr, checkpoint=configs.checkpoint,
resume_state=configs.resume_state,phase=configs.phase, height=configs.height)
result_path = f"{test_root}/results"
os.makedirs(result_path+"/"+"single_results", exist_ok=True)
os.makedirs(result_path+"/"+"multi_member", exist_ok=True)
val_logger.info("Model initialization is finished.")
val_logger.info("Testing dataset is ready.")
current_step, current_epoch = diffusion.begin_step, diffusion.begin_epoch
val_logger.info(f"Testing the model at epoch: {current_epoch}, iter: {current_step}.")
diffusion.register_schedule(beta_schedule=configs.test_schedule,
timesteps=configs.test_n_timestep,
linear_start=configs.test_linear_start,
linear_end=configs.test_linear_end)
loop_prediction(2016,2022,step_s)
val_logger.info("End of testing.")