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.

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

Share the SavedModel(s) for the hand_landmark.tflite, hand_landmark_3d.tflite and palm_detection.tflite models #29

Closed
tzolov opened this issue Aug 20, 2019 · 3 comments
Labels
legacy:hands Hand tracking/gestures/etc

Comments

@tzolov
Copy link

tzolov commented Aug 20, 2019

I will be very interested to test the palm-detection and hand-landmark TF models.

Unfortunately it is not possible to convert the .tflite format back into frozen graphs or savedmodels.

Will appreciate a lot if you could share the original TF SavedModel(s) used to build those tflite files (e.g. https://github.com/google/mediapipe/tree/master/mediapipe/models)

@mgyong
Copy link

mgyong commented Aug 20, 2019

@tzolov Unfortunately, we are unable to share the TF saved model files for any of our examples except object detection, which is provided in the MediaPipe model zoo. The current tflite format can be used as it is together with our MediaPipe inference pipeline examples

@lyleaf
Copy link

lyleaf commented Aug 21, 2019

How about TF.JS models. We would love to build a light weight webapp without having to go through the app building process.

@tzolov
Copy link
Author

tzolov commented Aug 21, 2019

Thanks for the quick response @mgyong! May i as if the inability to share are due to technical or license/legal reasons?
IMO there are other (not mediapipe based) interesting use cases that could benefit from the original, pre-trained TF models. I would like to try them with Java pipelines.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
legacy:hands Hand tracking/gestures/etc
Projects
None yet
Development

No branches or pull requests

3 participants