Skip to content

Conversation

@apivovarov
Copy link
Contributor

LOGISTIC operator is used in TFLite SSD Resnet50 and Mobilenet models.
Model graph: ssd_resnet_50_fpn_coco_nopp.tflite.pdf

### ssd_resnet_50_fpn_coco summary:
# op_id: op_name - count
 0: ADD - 58
 2: CONCATENATION - 2
 3: CONV_2D - 110
14: LOGISTIC - 1
17: MAX_POOL_2D - 4
18: MUL - 42
22: RESHAPE - 14
34: PAD - 1

This PR adds LOGISTIC operator support to relay tflite frontend.


assert isinstance(op, Operator)
input_tensors = self.get_input_tensors(op)
print("LOGISTIC")
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Remove it.

Copy link
Contributor Author

@apivovarov apivovarov Jun 7, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

removed

assert isinstance(op, Operator)
input_tensors = self.get_input_tensors(op)
print("LOGISTIC")
for i in input_tensors:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Remove it

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

removed

@tqchen tqchen added status: need review status: need update need update based on feedbacks labels Jun 7, 2019
Copy link
Member

@yongwww yongwww left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@yongwww
Copy link
Member

yongwww commented Jun 7, 2019

pls fix ci

@apivovarov
Copy link
Contributor Author

apivovarov commented Jun 10, 2019

The test which I added run successfully
Running relay TFLite frontend test.. ok
http://ci.tvm.ai:8080/blue/organizations/jenkins/tvm/detail/PR-3313/3/pipeline/239
I think tests/scripts/task_python_frontend.sh "Frontend GPU" run for 1 hour and 2 seconds and was killed automatically (duration limit is 1 hour). Last outputs from task_python_frontend.sh are:

test_forward.test_forward_resnet50 ... Sending interrupt signal to process
sh: line 1: 17972 Terminated

@apivovarov
Copy link
Contributor Author

apivovarov commented Jun 10, 2019

Opened issue "Too small time limit for frontend: GPU integration test" #3334

@kevinthesun kevinthesun merged commit 7c1c97d into apache:master Jun 11, 2019
wweic pushed a commit to wweic/tvm that referenced this pull request Jun 26, 2019
wweic pushed a commit to neo-ai/tvm that referenced this pull request Jun 27, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

status: need review status: need update need update based on feedbacks

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants