Hi All,
We are excited to announce that we have published another end-to-end tutorial of How to Create a Cartoonizer with TensorFlow Lite!
Here’s the link to the tutorial: https://blog.tensorflow.org/2020/09/how-to-create-cartoonizer-with-tf-lite.html
Please retweet / like it on Twitter: https://twitter.com/TensorFlow/status/1303727371841875969
The tutorial covers the following:
Export model checkpoints as SavedModel with TF 1.x
Convert the SavedModel to a TFLite model in TF 2.x with different quantization recipes
Model inference in Python
Add metadata to the TFLite models
Benchmark the TFLite models
Deploy TFLite model to Android in just a few lines of code with ML Model Binding
The Android app uses TensorFlow Lite models converted from the state of art White-box CartoonGAN.
Unlike the Selfie2Anime project, this one does not have the biased data issues which is why we were able to publish the tflite models on TensorFlow Hub.
The models are smaller and faster, and cartoonizes not only the selfies but also other types of images.
It has been an incredible learning and collaboration experience.
Many thanks to the support of the TensorFlow Lite team, TF Hub team, and the ML Kit team; and the ML GDE community for making this happen.
Thanks to the TensorFlow team for your help publishing it on the official TF blog.
This is part of an ongoing effort writing a series of E2E TFLite tutorials with sample code, and we are looking forward to more upcoming ones!
Best,
Margaret / Sayak
guys! this is awesome!sharing right now! and i'll try to make some time this weekend to try it out!cheers!
--
You received this message because you are subscribed to the Google Groups "ML GDE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ml-gde+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ml-gde/CAA2dEiW43VTSDCCFq5d6A923ZfvNJei1mZbxmCQK95O0n_P0Dw%40mail.gmail.com.
--
You received this message because you are subscribed to the Google Groups "ML GDE Project - E2E Model Deployment" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ml-gde-e2e-deplo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ml-gde-e2e-deployment/CAA2dEiW43VTSDCCFq5d6A923ZfvNJei1mZbxmCQK95O0n_P0Dw%40mail.gmail.com.
Very cool post Margaret - looking forward to trying it out!--Christiaan
--
You received this message because you are subscribed to the Google Groups "ML on Mobile OS Working Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mlwg-mobile...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mlwg-mobile/CAA2dEiW43VTSDCCFq5d6A923ZfvNJei1mZbxmCQK95O0n_P0Dw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
To view this discussion on the web visit https://groups.google.com/d/msgid/mlwg-mobile/CAP7uL88HPQWXJQevWM2tYVPmcGBZps8rdBsiT-yhjtY46jdmnw%40mail.gmail.com.
This is awesome!On Thu, Sep 10, 2020 at 1:45 AM Jerry Wu <je...@ap-mic.com> wrote:Awesome!!!'Khanh LeViet' via ML GDE <ml-...@googlegroups.com>於 2020年9月10日 週四,07:41寫道:
--You received this message because you are subscribed to the Google Groups "ML GDE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ml-gde+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ml-gde/CAMDTK-7zQfoxmZu8iU18xZzGMc_O-2kigwrY%2BmzBOD-sAo1AAg%40mail.gmail.com.
=====================================================Jerry Wu Ph.D. CandidateChief Executive OfficerGoogle Developers Expert (Machine Learning)2F., No.73, Ln. 72, Guangfu S. Rd., Da’an Dist., Taipei, Taiwan (R.O.C.)WeChat: dsjerry2017APMIC OpenTalk(歐奔頭殼股份有限公司)=====================================================----
You received this message because you are subscribed to the Google Groups "ML GDE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ml-gde+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ml-gde/CAKCEWQp1N40NCN1BcoeM1mdZSci6xja7JHZ5ZQnu48Q3yVzp7g%40mail.gmail.com.