Skip to content

Fix cpu inference #145

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 21 commits into from
Nov 14, 2022
Merged

Fix cpu inference #145

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
21 commits
Select commit Hold shift + click to select a range
8e68585
update docs
JX278 Nov 11, 2022
8c3ecc6
Update input.rst
JX278 Nov 11, 2022
89ed980
Update input.rst
JX278 Nov 11, 2022
bc4812d
Delete pytorch.rst
JX278 Nov 11, 2022
170c8f4
Delete cvode.rst
JX278 Nov 11, 2022
cadac15
Update install.rst
JX278 Nov 11, 2022
bc1ab5c
Update and rename input.rst to input.md
JX278 Nov 11, 2022
650b107
Update input.md
JX278 Nov 11, 2022
8da359d
Update input.md
JX278 Nov 11, 2022
70d0dba
Update and rename input.md to input.rst
JX278 Nov 11, 2022
7142e09
Merge branch 'deepmodeling:master' into master
JX278 Nov 11, 2022
0e29757
Update install.rst
zhixchen Nov 13, 2022
ce72f82
Merge branch 'deepmodeling:master' into master
JX278 Nov 13, 2022
3dbf425
Merge branch 'deepmodeling:master' into master
JX278 Nov 13, 2022
ea772ee
Merge branch 'deepmodeling:master' into master
JX278 Nov 13, 2022
e2ecb3c
Update install.rst
zhixchen Nov 13, 2022
43ca8cb
Add compile_success figure
JX278 Nov 13, 2022
e004f5f
Update install.rst
zhixchen Nov 13, 2022
1926cd6
Merge branch 'deepmodeling:master' into master
JX278 Nov 13, 2022
ee101c5
Merge branch 'deepmodeling:master' into master
JX278 Nov 14, 2022
f307c0d
fix cpu inference
JX278 Nov 14, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion examples/df0DFoam/zeroD_cubicReactor/H2/pytorchIntegrator/constant/CanteraTorchProperties
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ zeroDReactor

TorchSettings
{
torch off;
torch on;
GPU off;
log on;
torchModel1 "ESH2-sub1.pt";
Expand Down
34 changes: 27 additions & 7 deletions examples/df0DFoam/zeroD_cubicReactor/H2/pytorchIntegrator/inference.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,8 @@
import torch.profiler
import os



torch.set_printoptions(precision=10)


Expand Down Expand Up @@ -91,11 +93,12 @@ def forward(self, x):
setting0 = json2Parser(str("pytorchDNN/"+moduleName1+".json"))
setting1 = json2Parser(str("pytorchDNN/"+moduleName2+".json"))
setting2 = json2Parser(str("pytorchDNN/"+moduleName3+".json"))

#print(str("pytorchDNN/"+moduleName1+".json"))
lamda = setting0.power_transform
delta_t = setting0.delta_t
dim = setting0.dim
layers = setting0.layers


Xmu0 = torch.tensor(setting0.Xmu).unsqueeze(0).to(device)
Xstd0 = torch.tensor(setting0.Xstd).unsqueeze(0).to(device=device)
Expand All @@ -116,22 +119,33 @@ def forward(self, x):
model0 = Net()
model1 = Net()
model2 = Net()
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"))
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"))
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"))

if torch.cuda.is_available()==False:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"), map_location='cpu')
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"), map_location='cpu')
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"), map_location='cpu')
else:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"))
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"))
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"))

model0.load_state_dict(check_point0)
model1.load_state_dict(check_point1)
model2.load_state_dict(check_point2)
model0.to(device=device)
model1.to(device=device)
model2.to(device=device)

if len(device_ids) > 1:
model0 = torch.nn.DataParallel(model0, device_ids=device_ids)
model1 = torch.nn.DataParallel(model1, device_ids=device_ids)
model2 = torch.nn.DataParallel(model2, device_ids=device_ids)
except Exception as e:
print(e.args)

# print("check_point0")
# print(check_point0)
# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])

def inference(vec0, vec1, vec2):
'''
Expand Down Expand Up @@ -159,7 +173,8 @@ def inference(vec0, vec1, vec2):
input0_normalized = (input0_bct - Xmu0) / Xstd0
# input0_normalized[:, -1] = 0 #set Y_AR to 0
input0_normalized = input0_normalized.float()

#print("input0_normalized")
#print(input0_normalized[0])
rho1 = input1_[:, 0].unsqueeze(1)
input1_Y = input1_[:, 3:].clone()
input1_bct = input1_[:, 1:]
Expand All @@ -181,8 +196,13 @@ def inference(vec0, vec1, vec2):
output0_normalized = model0(input0_normalized)
output1_normalized = model1(input1_normalized)
output2_normalized = model2(input2_normalized)
#print("output0_normalized")
#print(output0_normalized[0])
# for name in model0.state_dict():
# print(name)


# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])
# post_processing
output0_bct = (output0_normalized * Ystd0 + Ymu0) * delta_t + input0_bct
output0_Y = (lamda * output0_bct[:, 2:] + 1)**(1 / lamda)
Expand Down
85 changes: 69 additions & 16 deletions examples/dfLowMachFoam/oneD_freelyPropagation/H2/pytorchIntegrator/inference.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,12 +10,10 @@
import torch.profiler
import os

torch.set_printoptions(precision=10)
print('position 0 in inference.py')
device = torch.device("cuda")
device_ids = range(torch.cuda.device_count())


torch.set_printoptions(precision=10)


class MyGELU(torch.nn.Module):
def __init__(self):
Expand Down Expand Up @@ -51,16 +49,56 @@ def forward(self, x):
x = self.fc(x)
return x
try:
#load variables from constant/CanteraTorchProperties
path_r = r"./constant/CanteraTorchProperties"
with open(path_r, "r") as f:
data = f.read()
i = data.index('torchModel1')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName1 = data[a+1:b]

i = data.index('torchModel2')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName2 = data[a+1:b]

i = data.index('torchModel3')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName3 = data[a+1:b]

i = data.index('GPU')
a = data.index(';', i)
b = data.rfind(' ',i+1,a)
switch_GPU = data[b+1:a]

#load OpenFOAM switch
switch_on = ["true", "True", "on", "yes", "y", "t", "any"]
switch_off = ["false", "False", "off", "no", "n", "f", "none"]
if switch_GPU in switch_on:
device = torch.device("cuda")
device_ids = range(torch.cuda.device_count())
elif switch_GPU in switch_off:
device = torch.device("cpu")
device_ids = [0]
else:
print("invalid setting!")
os._exit(0)



#glbal variable will only init once when called interperter
#load parameters from json
setting0 = json2Parser('pytorchDNN/settings1.json')
setting1 = json2Parser('pytorchDNN/settings2.json')
setting2 = json2Parser('pytorchDNN/settings3.json')

setting0 = json2Parser(str("pytorchDNN/"+moduleName1+".json"))
setting1 = json2Parser(str("pytorchDNN/"+moduleName2+".json"))
setting2 = json2Parser(str("pytorchDNN/"+moduleName3+".json"))
#print(str("pytorchDNN/"+moduleName1+".json"))
lamda = setting0.power_transform
delta_t = setting0.delta_t
dim = setting0.dim
layers = setting0.layers


Xmu0 = torch.tensor(setting0.Xmu).unsqueeze(0).to(device)
Xstd0 = torch.tensor(setting0.Xstd).unsqueeze(0).to(device=device)
Expand All @@ -76,29 +114,38 @@ def forward(self, x):
Xstd2 = torch.tensor(setting2.Xstd).unsqueeze(0).to(device=device)
Ymu2 = torch.tensor(setting2.Ymu).unsqueeze(0).to(device=device)
Ystd2 = torch.tensor(setting2.Ystd).unsqueeze(0).to(device=device)
print('position 1 in inference.py')

#load module
model0 = Net()
model1 = Net()
model2 = Net()
check_point0 = torch.load('pytorchDNN/ESH2-sub1.pt')
check_point1 = torch.load('pytorchDNN/ESH2-sub2.pt')
check_point2 = torch.load('pytorchDNN/ESH2-sub3.pt')

if torch.cuda.is_available()==False:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"), map_location='cpu')
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"), map_location='cpu')
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"), map_location='cpu')
else:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"))
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"))
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"))

model0.load_state_dict(check_point0)
model1.load_state_dict(check_point1)
model2.load_state_dict(check_point2)
model0.to(device=device)
model1.to(device=device)
model2.to(device=device)

if len(device_ids) > 1:
model0 = torch.nn.DataParallel(model0, device_ids=device_ids)
model1 = torch.nn.DataParallel(model1, device_ids=device_ids)
model2 = torch.nn.DataParallel(model2, device_ids=device_ids)
print('call init')
except Exception as e:
print(e.args)

# print("check_point0")
# print(check_point0)
# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])

def inference(vec0, vec1, vec2):
'''
Expand Down Expand Up @@ -126,7 +173,8 @@ def inference(vec0, vec1, vec2):
input0_normalized = (input0_bct - Xmu0) / Xstd0
# input0_normalized[:, -1] = 0 #set Y_AR to 0
input0_normalized = input0_normalized.float()

#print("input0_normalized")
#print(input0_normalized[0])
rho1 = input1_[:, 0].unsqueeze(1)
input1_Y = input1_[:, 3:].clone()
input1_bct = input1_[:, 1:]
Expand All @@ -148,8 +196,13 @@ def inference(vec0, vec1, vec2):
output0_normalized = model0(input0_normalized)
output1_normalized = model1(input1_normalized)
output2_normalized = model2(input2_normalized)
#print("output0_normalized")
#print(output0_normalized[0])
# for name in model0.state_dict():
# print(name)


# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])
# post_processing
output0_bct = (output0_normalized * Ystd0 + Ymu0) * delta_t + input0_bct
output0_Y = (lamda * output0_bct[:, 2:] + 1)**(1 / lamda)
Expand Down
85 changes: 69 additions & 16 deletions examples/dfLowMachFoam/threeD_reactingTGV/H2/pytorchIntegrator/inference.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,12 +10,10 @@
import torch.profiler
import os

torch.set_printoptions(precision=10)
print('position 0 in inference.py')
device = torch.device("cuda")
device_ids = range(torch.cuda.device_count())


torch.set_printoptions(precision=10)


class MyGELU(torch.nn.Module):
def __init__(self):
Expand Down Expand Up @@ -51,16 +49,56 @@ def forward(self, x):
x = self.fc(x)
return x
try:
#load variables from constant/CanteraTorchProperties
path_r = r"./constant/CanteraTorchProperties"
with open(path_r, "r") as f:
data = f.read()
i = data.index('torchModel1')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName1 = data[a+1:b]

i = data.index('torchModel2')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName2 = data[a+1:b]

i = data.index('torchModel3')
a = data.index('"',i)
b = data.index('.',a+1)
moduleName3 = data[a+1:b]

i = data.index('GPU')
a = data.index(';', i)
b = data.rfind(' ',i+1,a)
switch_GPU = data[b+1:a]

#load OpenFOAM switch
switch_on = ["true", "True", "on", "yes", "y", "t", "any"]
switch_off = ["false", "False", "off", "no", "n", "f", "none"]
if switch_GPU in switch_on:
device = torch.device("cuda")
device_ids = range(torch.cuda.device_count())
elif switch_GPU in switch_off:
device = torch.device("cpu")
device_ids = [0]
else:
print("invalid setting!")
os._exit(0)



#glbal variable will only init once when called interperter
#load parameters from json
setting0 = json2Parser('pytorchDNN/settings1.json')
setting1 = json2Parser('pytorchDNN/settings2.json')
setting2 = json2Parser('pytorchDNN/settings3.json')

setting0 = json2Parser(str("pytorchDNN/"+moduleName1+".json"))
setting1 = json2Parser(str("pytorchDNN/"+moduleName2+".json"))
setting2 = json2Parser(str("pytorchDNN/"+moduleName3+".json"))
#print(str("pytorchDNN/"+moduleName1+".json"))
lamda = setting0.power_transform
delta_t = setting0.delta_t
dim = setting0.dim
layers = setting0.layers


Xmu0 = torch.tensor(setting0.Xmu).unsqueeze(0).to(device)
Xstd0 = torch.tensor(setting0.Xstd).unsqueeze(0).to(device=device)
Expand All @@ -76,29 +114,38 @@ def forward(self, x):
Xstd2 = torch.tensor(setting2.Xstd).unsqueeze(0).to(device=device)
Ymu2 = torch.tensor(setting2.Ymu).unsqueeze(0).to(device=device)
Ystd2 = torch.tensor(setting2.Ystd).unsqueeze(0).to(device=device)
print('position 1 in inference.py')

#load module
model0 = Net()
model1 = Net()
model2 = Net()
check_point0 = torch.load('pytorchDNN/ESH2-sub1.pt')
check_point1 = torch.load('pytorchDNN/ESH2-sub2.pt')
check_point2 = torch.load('pytorchDNN/ESH2-sub3.pt')

if torch.cuda.is_available()==False:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"), map_location='cpu')
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"), map_location='cpu')
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"), map_location='cpu')
else:
check_point0 = torch.load(str("pytorchDNN/"+moduleName1+".pt"))
check_point1 = torch.load(str("pytorchDNN/"+moduleName2+".pt"))
check_point2 = torch.load(str("pytorchDNN/"+moduleName3+".pt"))

model0.load_state_dict(check_point0)
model1.load_state_dict(check_point1)
model2.load_state_dict(check_point2)
model0.to(device=device)
model1.to(device=device)
model2.to(device=device)

if len(device_ids) > 1:
model0 = torch.nn.DataParallel(model0, device_ids=device_ids)
model1 = torch.nn.DataParallel(model1, device_ids=device_ids)
model2 = torch.nn.DataParallel(model2, device_ids=device_ids)
print('call init')
except Exception as e:
print(e.args)

# print("check_point0")
# print(check_point0)
# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])

def inference(vec0, vec1, vec2):
'''
Expand Down Expand Up @@ -126,7 +173,8 @@ def inference(vec0, vec1, vec2):
input0_normalized = (input0_bct - Xmu0) / Xstd0
# input0_normalized[:, -1] = 0 #set Y_AR to 0
input0_normalized = input0_normalized.float()

#print("input0_normalized")
#print(input0_normalized[0])
rho1 = input1_[:, 0].unsqueeze(1)
input1_Y = input1_[:, 3:].clone()
input1_bct = input1_[:, 1:]
Expand All @@ -148,8 +196,13 @@ def inference(vec0, vec1, vec2):
output0_normalized = model0(input0_normalized)
output1_normalized = model1(input1_normalized)
output2_normalized = model2(input2_normalized)
#print("output0_normalized")
#print(output0_normalized[0])
# for name in model0.state_dict():
# print(name)


# print("fc.0.weight")
# print(model0.state_dict()['fc.0.weight'])
# post_processing
output0_bct = (output0_normalized * Ystd0 + Ymu0) * delta_t + input0_bct
output0_Y = (lamda * output0_bct[:, 2:] + 1)**(1 / lamda)
Expand Down
Loading