Kevin Ta
Kevin Ta
Thanks for reporting. Have you noticed this issue with other optimizers from that package? Custom optimizers may behave unexpected, but we will look into this issue
Following up, @ldv1. The suggestion works on my end. IPEX 2.1 should be released soon. Please check again by end of the week. Alternatively, it should work with torch+ipex 2.0.1...
> @zhuhaozhe seems that installing IPEX 2.1 does not help Did you try setting: `model, optimizer = ipex.optimize(model, optimizer=optimizer, inplace=True)` ?
Not sure if this is related to hardware or that specific node. I was unable to reproduce this on a Intel(R) Xeon(R) Platinum 8358 system. Could you try targeting a...
> @kta-intel Could you try on IDC to see if this can be reproduced? I was able to reproduce the issue on an 6128, 8358, and 8480+ instance on the...
If I recall, the error started during backprop. I just tested on `Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz` instance without issues, though. It looks like oneAPI IDC was recently...
Hi, yes there is. Please follow the instructions here: https://intel.github.io/intel-extension-for-pytorch/index.html#installation?platform=cpu&version=v2.2.0%2Bcpu (under **_source_**) We provide a `compile_bundle.sh` script to help make the compilation smoother and highly recommend making use of it
@wswsmao Hi, were you successful with building from source?
Thanks, let me reproduce and get back to you
really sorry for the delay. I was able to reproduce your issue. not sure what is causing the accumulation, but you can empty the xpu cache between iterations to avoid...