Looking forward to mxnet supporting bfloat16 too [image] Accelerating AI performance on 3rd Gen Intel® Xeon® Scalable processors with... Developers can now use the latest Intel build of TensorFlow to speed up current FP32 models using bfloat16 on 3rd Gen Xeon Scalable processors. Learn m…

Bfloat16

penli9 June 18, 2020, 6:14pm 1

Looking forward to mxnet supporting bfloat16 too

Topic		Replies	Views
Inference performance with fp16 float16 GTX 1080 ti Performance python , gluon , performance	2	3293	February 21, 2019
How to use all available cores in mxnet Performance	1	551	September 28, 2018
Support TensorRT in MXNET Performance	2	985	July 24, 2019
Is there a flag to enable tensor cores? No speedup from FP16 Performance	4	691	March 20, 2020
Performance issue on Intel CPU Performance	0	63	February 29, 2024