Edward Gaere

Hey Kevin, Thanks for the quick response! I am only writing 64 columns... so that it way below Excel's 16K limit. In terms of rows, the problem occurs at around...

When writing the same dataframe with Pandas/openpyxl, there are no problems with the generated file. So it appears Excel can support the volume of inline strings. Of course openpyxl takes...

Understood. I just realised that in the example I provided there are only 26 unique values of the strings (string.ascii_lowercase) across all 500'000 * 64 cells. ``` df = DataFrame(random.choice(list(ascii_lowercase),...

I think the issue can be closed. Shivanandroy's answer above works. > T5 models are text to text models & they expect `source_text` and `target_text` to be a string. I...

Thanks for the clarification and quick response! Clear now.

Great contribution :-) I tested it on a GPU (NVIDIA TITAN V), batched predictions provide up to 4x speedup.

My suggestion would be to limit the batch size to 6 or 12, so that the model and data fit on the GPU

I had the same problem. It seems a solution has been found on [StackOverflow](https://stackoverflow.com/questions/70454977/vite-production-build-errors-is-not-a-constructor-for-node-modules), which worked for me. In your **vite.config.js,** add the following to the configuration and then build...