BruceDai

Results 15 comments of BruceDai

@dontcallmedom @Honry I've updated this PR, please take another look, thanks.

I happen to meet this issue. @KaptenJansson @samdutton Any suggestion, thanks

Just checked #514 the duplicated issue likes this one. :(

- FCN8 (OpenVino) 1. Paper: [Fully Convolutional Networks for Semantic Segmentation](https://arxiv.org/abs/1605.06211) 2. Could convert from [fcn8s-heavy-pascal.caffemodel](http://dl.caffe.berkeleyvision.org/fcn8s-heavy-pascal.caffemodel) by [OpenVino Model Optimizer](https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_convert_model_Convert_Model_From_Caffe.html), converted OpenVino model file size: 537.9MB 3. This model is...

- [Mask R-CNN2GO](https://github.com/caffe2/models/tree/master/mask_rcnn_2go) Reference [Enabling full body AR with Mask R-CNN2Go](https://research.fb.com/blog/2018/01/enabling-full-body-ar-with-mask-r-cnn2go/)

Current I use workaround of reloading model for each inference to not block by this issue.

@mingmingtasd Did enable both SIMD and Threads flags?There's an issue #1381 Failed to run Wasm + SIMD + Threads on https://intel.github.io/webml-polyfill/examples. For testing on WebGL backend, I couldn't reproduce the...

@mingmingtasd I have update the github-page with tf.js 2.0.1. I check this issue on my side on Linux and Windows platform, it isn't reproduced. Would you please verify it on...

@mingmingtasd I think this isn't real root cause, we may investigate [the updated commits of NNAPI](https://android.googlesource.com/platform/frameworks/ml/) to find root cause.