in reference of this [kaggle notebook](https://www.kaggle.com/devang/transfer-learning-with-keras-and-efficientnets) i've written this [notebook](https://drive.google.com/file/d/1VyaUGXgaCDIbWbt0-F0iHTIdokxWesJH/view?usp=sharing) to try transfer learning through multiple datasets. through multiple soft runs on the notebook, I came to found out that I can't get deterministic/reproducible model out of it. Weird thing is, with the same virtual-environment, I can get deterministic/reproducible from a plain [CNN-MNIST](https://drive.google.com/file/d/1GpTupofffdaEufpMSvDkqnTpjiqP3mwC/view?usp=sharing) code. Spent a week on researching and can't find the solution. Any guidance or suggestions are much appreciated. [colab](https://colab.research.google.com/drive/1cqqdmrsQrlA7DDehnoEOXxqIt9WiwIMr?usp=sharing) code for reproducing the errors (dataset download and extractions included within) my virtual [environment](https://drive.google.com/file/d/1lfM5y2vylh9HkuduargAon9arNQKB9kQ/view?usp=sharing) p.s. link duplication from [issue44414](https://github.com/tensorflow/tensorflow/issues/44414) as I think this issue should be mentioned here