Terry Heo 9b87db80e0 tflite: Prevent from reallocation of persistent tensors
After ResetAllocationsAfter() is called, CalculateAllocations() could be called
again for nodes which have persistent temporary tensors. The logic should
prevent from reallocation of these tensors since they're not going to be
initialized again.

This issue could be reproduced easily with hybrid quantized models since some
hybrid kernels are using persistent temporary tensors.

This PR resolves GitHub issue #44520.

PiperOrigin-RevId: 345569003
Change-Id: I1b9777b33a664ebd0f09df8d3236c7ece0118b1a
2020-12-04 11:51:44 +09:00
..
2020-10-19 19:04:06 -07:00
2020-10-20 22:58:14 -07:00
2020-11-23 15:41:54 -08:00
2020-10-19 19:04:06 -07:00
2020-09-30 16:03:47 -07:00
2020-07-17 10:45:33 -07:00
2020-10-14 21:11:06 -07:00
2020-11-23 15:41:54 -08:00
2020-10-19 19:04:06 -07:00
2020-09-17 19:14:22 -07:00
2020-09-17 19:14:22 -07:00
2020-10-08 10:22:02 -07:00
2020-09-17 19:14:22 -07:00
2019-09-05 13:53:45 -07:00
2020-04-01 01:48:15 -07:00
2019-03-18 12:35:50 -07:00

TensorFlow Lite

TensorFlow Lite is TensorFlow's lightweight solution for mobile and embedded devices. It enables low-latency inference of on-device machine learning models with a small binary size and fast performance supporting hardware acceleration.

See the documentation: https://www.tensorflow.org/lite/ Documentation edits can be made here: tensorflow/lite/g3doc