Timeout issue when quantizing BERT large squad model

[ppalania21]

I have been following these steps to quantize BERT model - https://github.com/intel/lpot/tree/master/examples/tensorflow/nlp/bert_large_squad.

While running the below command for quantization:
python tune_squad.py --config=./bert.yaml --input_model=./bert_fp32.pb --output_model=./int8.pb --tune

I see a timeout/crash with the following message:
[ERROR] Specified timeout or max trials is reached! Not found any quantized model which meet accuracy goal.

Error log - bert_quantization_error.txt
yaml config file I am using - bert.yaml.txt

[guomingz]

@ppalania21 will try tf1.15up2 on it and send the update later.

[ppalania21]

I was able to confirm that BERT Int8 model works fine on TF1.15up2.