experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update instruction page to include sample quantized FPNLite model.

Browse Source 
PiperOrigin-RevId: 248022612
This commit is contained in:
A. Unique TensorFlower 2019-05-13 15:25:40 -07:00 committed by TensorFlower Gardener
parent b1c6a1accb
commit 0c7c7596b4
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
tensorflow/lite/java/ovic/README.md
View File

@ -23,7 +23,7 @@ Note: all commands should be called from your tensorflow installation folder (un
* Download the [testdata package](https://storage.googleapis.com/download.tensorflow.org/data/ovic_2018_10_23.zip): * Download the [testdata package](https://storage.googleapis.com/download.tensorflow.org/data/ovic_2018_10_23.zip):
```sh ```sh
curl -L https://storage.googleapis.com/download.tensorflow.org/data/ovic_2018_10_23.zip -o /tmp/ovic.zip curl -L https://storage.googleapis.com/download.tensorflow.org/data/ovic_2019_04_30.zip -o /tmp/ovic.zip
``` ```
* Unzip the package into the testdata folder: * Unzip the package into the testdata folder:
@ -185,6 +185,7 @@ Note: the benchmarking results can be quite different depending on the backgroun
| -------------------- |:---------------------:| | -------------------- |:---------------------:|
| detect.lite | 331 | | detect.lite | 331 |
| quantized_detect.lite | 95 | | quantized_detect.lite | 95 |
| quantized_fpnlite.lite | 119 |
Since Pixel 2 has excellent support for 8-bit quantized models, we strongly recommend you to check out the [quantization training tutorial](https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/quantize). Since Pixel 2 has excellent support for 8-bit quantized models, we strongly recommend you to check out the [quantization training tutorial](https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/quantize).