Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Remove unused protobuf header inclusions | A. Unique TensorFlower | 2017-06-06 |
| | | | | PiperOrigin-RevId: 158120864 | ||
* | Switched to Eigen min/max to speed up RequantizationRange op | Pete Warden | 2016-12-15 |
| | | | | Change: 142206704 | ||
* | Change quantize_graph to use two steps - RequantizationRange and Requantize - | A. Unique TensorFlower | 2016-10-21 |
in place of QuantizeDownAndShrinkRange. This will alow replacing RequantizationRange with constants. Also remove restriction on requested_output_max that was not present in QuantizeDownAndShrinkRange. Change: 136906788 |