ZhennanQin commented on a change in pull request #13697: [MKLDNN] Enable signed
int8 support for convolution.
URL: https://github.com/apache/incubator-mxnet/pull/13697#discussion_r244627239
##########
File path: example/quantization/imagenet_gen_qsym_mkldnn.py
##########
@@ -140,8 +140,8 @@ def save_params(fname, arg_params, aux_params,
logger=None):
' thresholds. This mode is expected to produce
the best inference accuracy of all three'
' kinds of quantized models if the calibration
dataset is representative enough of the'
' inference dataset.')
- parser.add_argument('--quantized-dtype', type=str, default='uint8',
- choices=['int8', 'uint8'],
+ parser.add_argument('--quantized-dtype', type=str, default='auto',
Review comment:
Good idea. It's a trade off between int8 and uint8. One line suggestion is,
we should use uint8 as much as we could to get best performance and accuracy.
Int8 is better used for negitive inputs only. That's why we introduced `auto`
out_type.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
