Commit message (Expand) | Author | Age | |
---|---|---|---|
* | Update fake quant op to support bitwidths in the range 2 to 16, from 2 to 8. | 2018-02-08 | |
* | Remove unnecessary code paths in FakeQuantWithMinMaxVarsPerChannel[Gradient] | 2017-06-12 | |
* | Add a new attribute narrow_range to FakeQuant* operations. It quantizes into... | 2017-05-31 | |
* | Remove unused using-declarations | 2017-05-26 | |
* | Delete debugging code in FakeQuant* operations. They are stable now, and ext... | 2017-05-24 | |
* | Add support for variable number of bits to FakeQuant* ops. | 2017-04-26 | |
* | Add FakeQuant ops and kernels for use with quantized training. | 2016-10-24 |