xiguiw

@msaroufim What's your expected result?

@guytamir This error was caused during build ITREX packge from source code. Please note oneAPI 2024.0 is required. It seemed your environment is oneAPI 2024.1. I'll change the example to...

Hi @oyazdanb, Welcome to neural-compressor~ Yes, there is some function to resume qunatization from history.snapshot. I'll check the function and feedback to you ASAP.

@oyazdanb the recover is borken for some models (not for all). Development team is working to fix it. During the time, I show you the way to recover from history.snapshot,...

Fix borken recover. PR: https://github.com/intel/neural-compressor/pull/1788

Close as issue fixed.

@RongLei-intel From your log, ("/v1/chat/completions") is OK. While "/v1/askdoc/xxx' failed. 'askdoc' is retrieval api, please check the your neuralchat.yaml, make sure retrieval is enable and retrieval is in the task...

Hi @ZzEeKkAa, Thank you for tring the feature. We can reproduce this on linux Yes, you right. There is some unimplemented interface. `jaxlib.xla_extension.XlaRuntimeError: UNIMPLEMENTED: PJRT C API does not support...

@yiqianglee @jzhoulon Do you have idea this "jaxlib.xla_extension.XlaRuntimeError: UNIMPLEMENTED: PJRT C API does not support AcquireExternalReference"? Is this a problem of "PJRT C API" Interface design of NOT cover/support or...

@ZzEeKkAa Hope this helps for you. This works on recent Intel extension for Tensorflow 2.15 + oneAPI 2024.1. This is the software version I used to verify this problem. intel-extension-for-tensorflow...