mirror of
https://github.com/wassname/DeepTime.git
synced 2026-07-28 11:14:59 +08:00
482 KiB
482 KiB
In [2]:
import os
from os.path import join
import math
import logging
from typing import Callable, Optional, Union, Dict, Tuple
from matplotlib import pyplot as plt
from pathlib import Path
import matplotlib.colors as mcolors
import gin
from fire import Fire
import numpy as np
import torch
from torch.utils.data import DataLoader
from torch import optim
from torch import nn
from experiments.base import Experiment
from data.datasets import ForecastDataset
from models import get_model
from utils.checkpoint import Checkpoint
from utils.ops import default_device, to_tensor
from utils.losses import get_loss_fn
from utils.metrics import calc_metrics
from experiments.forecast import get_data
gin.enter_interactive_mode()In [ ]:
In [3]:
gin.clear_config()
gin.parse_config(open("storage/experiments/Exchange/96M2/repeat=0/config.gin"))
train_set, train_loader = get_data(flag='train', batch_size=16)
model = get_model("deeptime",
dim_size=train_set.data_x.shape[1],
datetime_feats=train_set.timestamps.shape[-1]).to(default_device())
model.load_state_dict(torch.load('storage/experiments/Exchange/96M2/repeat=0/model.pth'))
model = model.eval()
b = train_set[1]
b = [bb[None, :] for bb in b]
x, y, x_time, y_time = map(to_tensor, b)
with torch.no_grad():
forecast = model(x, x_time, y_time)
plt.title('mlp inr2')
import matplotlib.colors as mcolors
colors = list(mcolors.BASE_COLORS.keys())
l = x.shape[1]
forecast2 = forecast[0].detach().cpu().numpy()
x2 = x[0].cpu()
y2 = y[0].cpu()
l2 = y.shape[1]
i_past = list(range(l))
i_future = list(range(l, l+l2))
for i in range(x.shape[-1]):
plt.plot(i_past, x2[:, i], c=colors[i])
for i in range(x.shape[-1]):
plt.plot(i_future, y2[:, i], c=colors[i])
for i in range(x.shape[-1]):
plt.plot(i_future, forecast2[:, i], c=colors[i], linestyle='--')
In [4]:
def plot(model_name="deeptime", save_path=Path("storage/experiments/Exchange/96M/repeat=0"), i=200, title=None, plot=True):
gin.clear_config()
gin.parse_config(open(save_path/"config.gin"))
train_set, train_loader = get_data(flag='train', batch_size=2)
model = get_model(model_name,
dim_size=train_set.data_x.shape[1],
datetime_feats=train_set.timestamps.shape[-1]).to(default_device())
model.load_state_dict(torch.load(save_path/'model.pth'))
model = model.eval()
b = train_set[i]
b = [bb[None, :] for bb in b]
x, y, x_time, y_time = map(to_tensor, b)
with torch.no_grad():
forecast = model(x, x_time, y_time)
if title is None:
title = str(save_path).split('/')[-3:]
title = "-".join(title)
colors = list(mcolors.BASE_COLORS.keys())
l = x.shape[1]
forecast2 = forecast[0].detach().cpu().numpy()
x2 = x[0].cpu()
y2 = y[0].cpu()
l2 = y.shape[1]
i_past = list(range(l))
i_future = list(range(l, l+l2))
if plot:
plt.title(title)
for i in range(x.shape[-1]):
plt.plot(i_past, x2[:, i], c=colors[i])
for i in range(x.shape[-1]):
plt.plot(i_future, y2[:, i], c=colors[i])
for i in range(x.shape[-1]):
plt.plot(i_future, forecast2[:, i], c=colors[i], linestyle='--')
return x2, y2, forecast2, i_past, i_future
In [54]:
def plot_multi(model_names=["deeptime"], save_paths=[Path("storage/experiments/Exchange/96M/repeat=0")], i=200, title=None, plot=True):
for j in range(len(model_names)):
model_name = model_names[j]
save_path = save_paths[j]
gin.clear_config()
gin.parse_config(open(save_path/"config.gin"))
train_set, train_loader = get_data(flag='train', batch_size=2)
model = get_model(model_name,
dim_size=train_set.data_x.shape[1],
datetime_feats=train_set.timestamps.shape[-1]).to(default_device())
model.load_state_dict(torch.load(save_path/'model.pth'))
model = model.eval()
b = train_set[i]
b = [bb[None, :] for bb in b]
x, y, x_time, y_time = map(to_tensor, b)
with torch.no_grad():
forecast = model(x, x_time, y_time)
# if title is None:
# title = str(save_path).split('/')[-3:]
# title = "-".join(title)
colors = list(mcolors.BASE_COLORS.keys())
l = x.shape[1]
forecast2 = forecast[0].detach().cpu().numpy()
x2 = x[0].cpu()
y2 = y[0].cpu()
l2 = y.shape[1]
i_past = list(range(l))
i_future = list(range(l, l+l2))
linestyles = [ '--', '-.', '-', ':']
if plot:
ls = linestyles[j]
print(ls)
if j==0:
for k in range(x.shape[-1]):
plt.plot(i_past, x2[:, k], c=colors[k], label=f"var {k}")
for k in range(x.shape[-1]):
plt.plot(i_future, y2[:, k], c=colors[k])
mtitle = str(save_path).split('/')[-2:-1]
mtitle = "-".join(mtitle)
for k in range(x.shape[-1]):
plt.plot(i_future, forecast2[:, k], c=colors[k], linestyle=ls, label=f"{mtitle}" if k==0 else None)
plt.legend()
plt.title(title)
return x2, y2, forecast2, i_past, i_future
In [62]:
plot_multi(model_names=["deeptime", "deeptime2"],
save_paths=[
Path("storage/experiments/Exchange/96M/repeat=0"),
Path("storage/experiments/Exchange/96Mplus/repeat=0"),
],
i=100
)
1Out [62]:
-- receptive field [690 378 242]=[138 18 2]*[[ 1 1 1] [ 1 2 4] [ 1 4 16] [ 1 6 36] [ 1 8 64]] -.
1
In [64]:
plot_multi(model_names=["deeptime", "deeptime2"],
save_paths=[
Path("storage/experiments/Exchange/96M2/repeat=0"),
Path("storage/experiments/Exchange/96Mplus2/repeat=0"),
])
1Out [64]:
-- receptive field [690 378 242]=[138 18 2]*[[ 1 1 1] [ 1 2 4] [ 1 4 16] [ 1 6 36] [ 1 8 64]] -.
1
In [67]:
plot("deeptime", Path("storage/experiments/Exchange/96S/repeat=0"));
In [57]:
plot("deeptime", Path("storage/experiments/Exchange/96M2/repeat=0"))Out [57]:
(tensor([[ 0.5292, 1.7510, 0.7053, ..., -0.8215, -0.8000, -0.0716],
[ 0.4468, 1.6824, 0.6341, ..., -0.8088, -0.7295, -0.0506],
[ 0.5728, 1.7182, 0.6890, ..., -0.8497, -0.7756, 0.0018],
...,
[ 0.5457, 0.0489, 0.8019, ..., -1.4872, -0.9826, -0.3539],
[ 0.5340, 0.0770, 0.7975, ..., -1.4627, -0.9620, -0.3570],
[ 0.5195, 0.0113, 0.8092, ..., -1.4745, -0.9472, -0.3507]]),
tensor([[ 5.4374e-01, 1.3666e-01, 8.0406e-01, -8.8760e-01, 1.9074e+00,
-1.4300e+00, -9.5311e-01, -3.3919e-01],
[ 5.5538e-01, 2.2618e-01, 8.1282e-01, -8.4952e-01, 1.9074e+00,
-1.4273e+00, -8.8462e-01, -3.0561e-01],
[ 5.7478e-01, 2.4110e-01, 8.2086e-01, -8.3329e-01, 1.9123e+00,
-1.3801e+00, -8.4853e-01, -3.2345e-01],
[ 5.8739e-01, 2.4588e-01, 8.2966e-01, -8.3030e-01, 1.9123e+00,
-1.3864e+00, -8.5757e-01, -3.2240e-01],
[ 5.7769e-01, 1.6949e-01, 8.3553e-01, -8.8549e-01, 1.9123e+00,
-1.4164e+00, -8.3038e-01, -3.2870e-01],
[ 5.8739e-01, 1.3964e-01, 8.3480e-01, -9.1691e-01, 1.9057e+00,
-1.4373e+00, -8.2736e-01, -3.5388e-01],
[ 5.9127e-01, 1.5934e-01, 8.3406e-01, -8.9600e-01, 1.9057e+00,
-1.4200e+00, -8.3341e-01, -3.4759e-01],
[ 5.7769e-01, 1.3069e-01, 8.4067e-01, -9.1274e-01, 1.9057e+00,
-1.4191e+00, -8.2736e-01, -3.2555e-01],
[ 5.7381e-01, 9.4885e-02, 8.5245e-01, -9.3146e-01, 1.9057e+00,
-1.4236e+00, -8.2736e-01, -3.3395e-01],
[ 5.3889e-01, 9.7869e-02, 8.5984e-01, -9.5007e-01, 1.9057e+00,
-1.4164e+00, -8.2736e-01, -3.3290e-01],
[ 5.3017e-01, -5.7301e-02, 8.5320e-01, -1.0214e+00, 1.8953e+00,
-1.4545e+00, -8.5457e-01, -3.3185e-01],
[ 5.4374e-01, -2.4828e-01, 8.5320e-01, -1.0635e+00, 1.8835e+00,
-1.5045e+00, -9.4125e-01, -3.6018e-01],
[ 5.6799e-01, -2.0232e-01, 8.6945e-01, -1.0535e+00, 1.8636e+00,
-1.4645e+00, -9.3532e-01, -3.1296e-01],
[ 5.9224e-01, 2.9236e-02, 8.5689e-01, -9.5007e-01, 1.8699e+00,
-1.4327e+00, -8.4853e-01, -3.0981e-01],
[ 6.2230e-01, 1.5158e-01, 8.5910e-01, -8.7495e-01, 1.8734e+00,
-1.4146e+00, -8.3643e-01, -3.0771e-01],
[ 6.0678e-01, -6.0420e-04, 8.5541e-01, -9.8697e-01, 1.8734e+00,
-1.4473e+00, -8.3643e-01, -3.2135e-01],
[ 5.9709e-01, 4.7140e-02, 8.5172e-01, -9.6858e-01, 1.8734e+00,
-1.4427e+00, -8.3825e-01, -3.2870e-01],
[ 5.9321e-01, 5.0124e-02, 8.7242e-01, -9.4884e-01, 1.8734e+00,
-1.4427e+00, -8.3643e-01, -3.3395e-01],
[ 5.9709e-01, 1.0980e-01, 8.7909e-01, -9.2648e-01, 1.8734e+00,
-1.4227e+00, -8.3341e-01, -3.2240e-01],
[ 5.9418e-01, 5.3108e-02, 8.8058e-01, -9.6036e-01, 1.8734e+00,
-1.4427e+00, -8.1219e-01, -3.1296e-01],
[ 6.0388e-01, 5.0124e-02, 8.6871e-01, -9.7268e-01, 1.8734e+00,
-1.4354e+00, -8.1219e-01, -3.1926e-01],
[ 5.3501e-01, 1.1577e-01, 8.7984e-01, -9.6448e-01, 1.8734e+00,
-1.3946e+00, -8.0609e-01, -3.4129e-01],
[ 6.0000e-01, 1.3069e-01, 8.7242e-01, -9.5337e-01, 1.8734e+00,
-1.3592e+00, -8.0609e-01, -3.1821e-01],
[ 6.0775e-01, 1.2174e-01, 8.7761e-01, -9.6776e-01, 1.8734e+00,
-1.3873e+00, -8.0913e-01, -2.6259e-01]]),
array([[ 0.52890754, 0.04601173, 0.818106 , -0.9959081 , 1.9366491 ,
-1.5069696 , -0.9808905 , -0.36548147],
[ 0.528438 , 0.03981169, 0.8184106 , -0.9991272 , 1.937193 ,
-1.5084361 , -0.9828327 , -0.36484447],
[ 0.5307723 , 0.04257247, 0.8204612 , -1.0004759 , 1.9423875 ,
-1.5103797 , -0.9852797 , -0.36599845],
[ 0.5303809 , 0.04465088, 0.8191587 , -0.9979849 , 1.9372461 ,
-1.505586 , -0.9807858 , -0.36503443],
[ 0.5305283 , 0.04264332, 0.8198152 , -1.0008417 , 1.9394246 ,
-1.5105733 , -0.9837851 , -0.3655746 ],
[ 0.5307227 , 0.04202083, 0.8191189 , -0.99933463, 1.9369458 ,
-1.506729 , -0.98293924, -0.36503708],
[ 0.5283215 , 0.04141821, 0.81586105, -0.9971103 , 1.9277518 ,
-1.5040027 , -0.978884 , -0.36278751],
[ 0.5252694 , 0.04255188, 0.8136384 , -0.9928628 , 1.9213691 ,
-1.4981959 , -0.9740302 , -0.36309323],
[ 0.5275271 , 0.04352546, 0.814939 , -0.9949809 , 1.9263753 ,
-1.5013223 , -0.9757192 , -0.3638668 ],
[ 0.5276257 , 0.04390374, 0.8143943 , -0.9922678 , 1.9258269 ,
-1.5007619 , -0.97613275, -0.36400792],
[ 0.52339524, 0.04113823, 0.81086344, -0.9905842 , 1.9132272 ,
-1.4942402 , -0.9708606 , -0.36080262],
[ 0.5248644 , 0.042987 , 0.81305283, -0.9912248 , 1.9210671 ,
-1.4969778 , -0.9736485 , -0.3629068 ],
[ 0.52518415, 0.04741649, 0.81146616, -0.98756695, 1.9176267 ,
-1.4939798 , -0.97169137, -0.36096278],
[ 0.5220071 , 0.04414371, 0.81153566, -0.98984647, 1.9137857 ,
-1.4954246 , -0.97107446, -0.36085606],
[ 0.52268267, 0.0460408 , 0.8109568 , -0.9874458 , 1.9127601 ,
-1.4930701 , -0.9692728 , -0.3606753 ],
[ 0.5223365 , 0.04568218, 0.8103212 , -0.98693126, 1.9110385 ,
-1.4933776 , -0.96929896, -0.3607128 ],
[ 0.52153313, 0.04434089, 0.8080093 , -0.9862203 , 1.9069732 ,
-1.4901737 , -0.96673113, -0.3586038 ],
[ 0.5211224 , 0.04329763, 0.80827016, -0.9870948 , 1.9067513 ,
-1.4907483 , -0.96722066, -0.35898498],
[ 0.51958317, 0.0411768 , 0.80564517, -0.9852966 , 1.90139 ,
-1.4851948 , -0.96316063, -0.35826832],
[ 0.52038646, 0.04568981, 0.80709517, -0.9835636 , 1.9032961 ,
-1.4881977 , -0.9650438 , -0.35861585],
[ 0.5194442 , 0.04407027, 0.80520827, -0.98266006, 1.8993053 ,
-1.4852973 , -0.9634341 , -0.35824707],
[ 0.51898205, 0.04415236, 0.8051576 , -0.9825941 , 1.897731 ,
-1.4861436 , -0.9634547 , -0.3574056 ],
[ 0.5183343 , 0.04260498, 0.80339235, -0.9823522 , 1.893536 ,
-1.4827589 , -0.9612396 , -0.3563275 ],
[ 0.5182621 , 0.03967791, 0.8052563 , -0.9854379 , 1.8969235 ,
-1.4861064 , -0.9637438 , -0.3569042 ]], dtype=float32),
[0,
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,
188,
189,
190,
191],
[192,
193,
194,
195,
196,
197,
198,
199,
200,
201,
202,
203,
204,
205,
206,
207,
208,
209,
210,
211,
212,
213,
214,
215])
In [42]:
plot("deeptime2", Path("storage/experiments/Exchange/96Mplus2/repeat=0"))receptive field [690 378 242]=[138 18 2]*[[ 1 1 1] [ 1 2 4] [ 1 4 16] [ 1 6 36] [ 1 8 64]]
In [ ]:
In [ ]:
In [ ]: