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.
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
[test/tvm] Add tvm ssat test for image labeling @open sesame 11/16 11:00 #3886
base: main
Are you sure you want to change the base?
Conversation
📝 TAOS-CI Version: 1.5.20200925. Thank you for submitting PR #3886. Please a submit 1commit/1PR (one commit per one PR) policy to get comments quickly from reviewers. Your PR must pass all verificiation processes of cibot before starting a review process from reviewers. If you are new member to join this project, please read manuals in documentation folder and wiki page. In order to monitor a progress status of your PR in more detail, visit http://ci.nnstreamer.ai/. |
PATH_TO_IMAGE="../test_models/data/orange.png" | ||
PATH_TO_CLASS="class.out.log" | ||
|
||
gstTest "--gst-plugin-path=${PATH_TO_PLUGIN} filesrc location=${PATH_TO_IMAGE} ! pngdec ! videoscale ! imagefreeze ! videoconvert ! video/x-raw,format=RGB,framerate=0/1 ! tensor_converter ! tensor_filter framework=tvm model=${PATH_TO_MODEL} custom=device=CPU,num_input_tensors:1 ! tensor_decoder mode=image_labeling option1=${PATH_TO_LABEL} ! filesink location=${PATH_TO_CLASS}" 1 0 0 $PERFORMANCE |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
custom=device=CPU,...
==>
custom=device:CPU...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
updated
testResult $? 1 "Golden test comparison" 0 1 | ||
|
||
# Negative test: wrong input type float32 instead of uint8 | ||
gstTest "--gst-plugin-path=${PATH_TO_PLUGIN} filesrc location=${PATH_TO_IMAGE} ! pngdec ! videoscale ! imagefreeze ! videoconvert ! tensor_converter ! tensor_transform mode=arithmetic option=typecast:float32,div:127.5,add:-1.0 ! tensor_filter framework=tvm model=${PATH_TO_MODEL} custom=device=CPU,num_input_tensors:1 ! tensor_decoder mode=image_labeling option1=${PATH_TO_LABEL} ! filesink location=${PATH_TO_CLASS} " 1_n 0 1 $PERFORMANCE |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
custom=device=CPU,...
==>
custom=device:CPU...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
updated, thanks
If you can supply lighter models for testing, it'd be great. |
Looks like $(arch) is often not supported in a few Linux distro. |
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202208190947010.022424936294556-8d6f2826f3597edcaa391cd0d6ed438faae5da22/. |
unfortunately I have no smaller compiled model than this one right now |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202208191633460.31454110145569-2ad5767436a5feab65f700255a8ec9a79ef57191/. |
Unfortunately, there are failed unit tests (TVM tests. Both in x64 arch. Maybe the x64 binary has some problems?)
|
May check this: #3412 (comment) |
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202208260935430.74763894081116-2ad5767436a5feab65f700255a8ec9a79ef57191/. |
2ad5767
to
1a2488f
Compare
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202210110001490.90890288352966-1a2488fa92bf4464814cc89c1a0f8d9baa3fc6f6/. |
This commit adds a compiled tvm model and two gst test commands Sample model (mobilenet_v1_0.75_224_{arch}.so) is compiled from public tflite model as follow: - Get the model from: http://download.tensorflow.org/models/mobilenet_v1_2018_08_02/mobilenet_v1_0.75_224_quant.tgz - Follow the steps described in: https://tvm.apache.org/docs/how_to/compile_models/from_tflite.html Supported architectures are armv7l, aarch64 and x86_64 Signed-off-by: Aymen Sghaier <aymen.sghaier@nxp.com>
1a2488f
to
da5dd81
Compare
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202210111852150.17939901351929-da5dd81548bf3e930b254a527a5a3465f058b2da/. |
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202210121359540.48101806640625-da5dd81548bf3e930b254a527a5a3465f058b2da/. |
@ksghaier @myungjoo |
@anyj0527 PTAL if it's now ok. |
cibot: @ksghaier, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/3886-202211161128370.079869031906128-da5dd81548bf3e930b254a527a5a3465f058b2da/. |
Tizen tvm package need upgrade (0.7.0 -> 0.9+). I'll make a new issue for this. |
This commit adds a compiled tvm model and two gst test commands
Sample model (mobilenet_v1_0.75_224_{arch}.so) is compiled from public
tflite model as follow:
http://download.tensorflow.org/models/mobilenet_v1_2018_08_02/mobilenet_v1_0.75_224_quant.tgz
https://tvm.apache.org/docs/how_to/compile_models/from_tflite.html
Signed-off-by: Aymen Sghaier aymen.sghaier@nxp.com
[Template] PR Description
In general, github system automatically copies your commit message for your convenience.
Please remove unused part of the template after writing your own PR description with this template.