Skip to content
This repository was archived by the owner on Nov 17, 2023. It is now read-only.

[Quantization] ops: 'quantize', 'dequantize'#5843

Merged
piiswrong merged 11 commits intoapache:masterfrom
ZihengJiang:quantization
Apr 26, 2017
Merged

[Quantization] ops: 'quantize', 'dequantize'#5843
piiswrong merged 11 commits intoapache:masterfrom
ZihengJiang:quantization

Conversation

@ZihengJiang
Copy link
Copy Markdown
Contributor

Make it public for other guys to contribute and use.

@mli @thatindiandude @arank

DMLC_REGISTER_PARAMETER(DequantizeParam);

NNVM_REGISTER_OP(_contrib_dequantize)
.MXNET_DESCRIBE("Dequantize")
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

add some document describing the math and reference papers.

struct DequantizeParam : public dmlc::Parameter<DequantizeParam> {
float min_range;
float max_range;
int out_type;
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

arguments looks verbose. min, max, dtype looks better

@piiswrong piiswrong merged commit 5efd91a into apache:master Apr 26, 2017
Guneet-Dhillon pushed a commit to Guneet-Dhillon/mxnet that referenced this pull request Sep 13, 2017
* [Quantization] ops: 'quantize', 'dequantize'

* [Quantization] Add description

* Fix macro

* Fix macro

* [Quantization] Refine interface

* [Quantization] Fix lint

* Fix GPU version
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants