Open vineethanandh opened 10 months ago
I'm going to look into it
@regisss - Did you get some time to check this behaviour
@regisss - Did you get some time to check this behaviour
Not yet. I don't think I'll have time to do it this week and before releasing Optimum Habana v1.8. I'll investigate this by next week and will release a patch if needed.
System Info
Information
Tasks
examples
folder (such as GLUE/SQuAD, ...)Reproduction
Steps to Reproduce Writing down the steps to reproduce to run SwinT in 1x
Expected behavior
The expected behaviour is that 1.12.0-463 having similar perf with optimum-habana 1.6.1 and optimum-habana 1.7.4
But what is observed is that perf is better in optimum-habana 1.6.1 and comparitively lesser in 1.7.4
This is applicable for SwinT, ViT, Bert-Large in 8x and 1x. Eg: values in SwinT is given below
OH - 1.7.4 values 362.524 362.566 360.719 358.089
OH - 1.6.1 values 389.045 390.971 389.587
Almost 7.5% drop