Skip to content
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

Jump to bottom

fix some bugs #55

Merged
merged 1 commit into from
Jan 20, 2022
Merged

fix some bugs #55

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
28 changes: 10 additions & 18 deletions Source/JackFramework/Core/build_training_graph.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -92,14 +92,18 @@ def __pass_data2device(self, data: list) -> list:

return data

def __calculation_process(self, model_item: object, input_data: list,
model_id: int) -> tuple:
output_data, _, _ = self.__init_calculation_result()
def __calculation_process(self, model_item: object, input_data: list, label_data: list,
model_id: int, is_training: bool = True) -> tuple:
output_data, loss, acc = self.__init_calculation_result()

# get ouput
output_data = self.__jf_model.inference(model_item, input_data, model_id)
# loss and acc
if is_training:
loss = self.__jf_model.loss(output_data, label_data, model_id)
acc = self.__jf_model.accuary(output_data, label_data, model_id)

return output_data
return output_data, loss, acc

def __variable2tensor(self, data: list) -> None:
res = []
Expand Down Expand Up @@ -174,15 +178,9 @@ def train_model(self, input_data: list, label_data: list) -> list:
input_data = self.__pass_data2device(input_data)
label_data = self.__pass_data2device(label_data)

output_data_list = []
for i, model_item in enumerate(self.__model):
self.__opt[i].zero_grad()
output_data = self.__calculation_process(model_item, input_data, i)
output_data_list.append(output_data)

for i, output_data in enumerate(output_data_list):
loss = self.__jf_model.loss(output_data, label_data, i)
acc = self.__jf_model.accuary(output_data, label_data, i)
_, loss, acc = self.__calculation_process(model_item, input_data, label_data, i)

loss[self.OPT_LOSS_ID].backward()
self.__opt[i].step()
Expand All @@ -202,14 +200,8 @@ def val_model(self, input_data: list, label_data: list) -> list:
label_data = self.__pass_data2device(label_data)

with torch.no_grad():
output_data_list = []
for i, model_item in enumerate(self.__model):
output_data = self.__calculation_process(model_item, input_data, i)
output_data_list.append(output_data)

for i, output_data in enumerate(output_data_list):
loss = self.__jf_model.loss(output_data, label_data, i)
acc = self.__jf_model.accuary(output_data, label_data, i)
_, loss, acc = self.__calculation_process(model_item, input_data, label_data, i)
tower_loss_iteration.append(self.__variable2tensor(loss))
tower_acc_iteration.append(self.__variable2tensor(acc))

Expand Down