webml-polyfill
webml-polyfill copied to clipboard
intel
Deprecated, the Web Neural Network Polyfill project has been moved to https://github.com/webmachinelearning/webnn-polyfill
Hi, 1. Built [chromium ](https://github.com/intel/webml-polyfill/wiki/How-to-build-chromium-on-Linux) for Webnn support. 2. Installed openvino 2020.2.120. System configuration : Operating system: Linux (Ubuntu 18.04) Processor Generation: 8 Openvino Version: 2020.2.120 Issue : I have...
@cuiyanx Could you please help to add cts test for int8Softmax?
Task ID | Task Title | Assignee | Notes / Status -- | -- | -- | -- [issue-1167](https://github.com/intel/webml-polyfill/issues/1167) | [INT8][DNNL]Implementation of AVERAGE POOL | lisha | done [issue-1212](https://github.com/intel/webml-polyfill/issues/1212) |...
@huningxin Could you please help to add this support?
Since the last layer of mobilenet is softmax and the previous layer is Int8Conv, I think we also need to add TENSOR_QUANT8_ASYMM_SIGNED in Softmax. @huningxin could you please help to...
Here’s code to get GPU detail information: ```js var canvas = document.createElement('canvas'); var gl = canvas.getContext('webgl'); var debugInfo = gl.getExtension('WEBGL_debug_renderer_info'); var renderer = gl.getParameter(debugInfo.UNMASKED_RENDERER_WEBGL); // console.log(renderer); // "ANGLE (Intel(R) UHD...
**Test Env:** Chromium Version: nightly build 79.0.3917.0(d3768f3) Platform: Windows GPU: NVIDIA GeForce RTX 2080(26.21.14.4587) **Expected Result:** The inference time should be stable on NVIDIA graphic. **Actual Result:** The inference time...
Create test case based on MobileNet Int8 model: 1. Int8ConvRelu op -> CONV_2D - Create input operand with TENSOR_QUANT8_ASYMM_SIGNED, set scale based on scales of bias and weights (scale_bias/scale_weights), zero...
