I came across a peculiar difference when exporting a t5 model after fine-tuning.
When the code runs in docker using the original tensorflow image with Python 3.6.9 and exports to a folder shared between the docker and host (using ‘volumes’), it throws an error about mixture not found. Updating operative_config.gin for the mixture used in fine-tuning allows to export albeit with some other non-critical tensorflow errors.
When the code runs on the host with Python 3.8.5, there are no export issues.
The only differences I noticed is that 1) docker changes the shared folder user:group ownership to root:root where 2) folder permission settings for group lack ‘w’ (i.e. write) and 3) tensorflow image in docker uses Python 3.6.9.
Has anyone encountered and solved?