Sachin Prasad

icon location San Jose, CA, USA

Results 312 comments of Sachin Prasad

Enable NNAPI Crash

@Whrothus , Apologies for the delayed response, I'm looking into your issue. I will get back to you if I need any details from you. Thanks

Enable NNAPI Crash

Could you try the option to `setUseNnapiCpu` to `true` in the `NnApiDelegate.Options` object and let us know if it is still crashing.

Dtype of reverse-mode gradient is incorrect when input dtype is a mixture of float32 and complex128

`tf.keras.losses.MSE` is doing the casting and it is converting to complex128 in both the cases, but in `grad_backward` when you calculate the gradient with respect to x, which is float32,...

Shifted/reordered YOLO output layers when converting with TfLite 2.8 or higher

Could you please try specifying the output signature in your keras saved model as mentioned in the issues which you have mentioned.

assertion failed: [Trying to access a placeholder that is not supposed to be executed. This means you are executing a graph generated from the cross-replica context in an in-replica context.]

Hi, Is the model which you are trying to load is trained on different Tensorflow version, if so could you please mention the version. Thanks!

Please update the Simplified Chinese documents, or at least set an indicator to tell readers the documents are out of date

I have created the pull request by making code changes, once the changes reflect in the master, feel free to add few more language specific changes here https://gitlocalize.com/tensorflow/docs-l10n

Please update the Simplified Chinese documents, or at least set an indicator to tell readers the documents are out of date

@code4lala , The linked PR has been merged, could you please close the issue. Thanks!

Please update the Simplified Chinese documents, or at least set an indicator to tell readers the documents are out of date

It will take few days to take effect, but there will be one more update with the translation with the description as well apart from the code changes submitted in...

Please update the Simplified Chinese documents, or at least set an indicator to tell readers the documents are out of date

Document is now updated with the latest code and information inline with the english version. Could you please verify here https://www.tensorflow.org/lite/performance/gpu_advanced?hl=zh-cn#android_cc and close the issue. Thanks!

gradient computation throws InvalidArgumentError in `tf.gather_nd`

@VictoriaGriffith , As per the above comment, please find the working [Gist](https://colab.sandbox.google.com/gist/sachinprasadhs/213513c68d4786382a9a90280e94d37d/56628.ipynb) with modified indices input.