venzen

Perhaps persistence across sessions can be implemented by using nvim.sqlite as is done in nvim-neoclip.lua https://github.com/AckslD/nvim-neoclip.lua

@titipata Thank you for your response. I started implementing word similarity checking before I saw your reply. Can use `difftool.SequenceMatcher()` with a text file of Thai words and find correct...

Unexpected behavior from `deepcut`: I am passing a custom dictionary that contains both the words 'หรือ' and 'อิริยาบถ'. Each word is a separate entry on its own line and without...

The issue was that the custom dictionary contained duplicate words (words also present in the deepcut dictionary). When I made a new blank custom dictionary deepcut works as expected.

@nexthybrid you can pass --device="mps" on the command line or specify `device="mps"` in the _finetune_shakespeare.py_ script

@codetiger When you say "tune the encoder/decoder a little bit", can you give a brief explanation or example?

Your train.py script looks OK. What is the output of the prepare.py script with your new data? To run train.py, do you use cpu or GPU? I ask because you...

Not sure that this is going to solve your issue, but I'd recommend you try the process of finetuning again with the following practice: 1. Execute commands from the project...

Tokenizing as you propose - read a line, tokenize it. write it to file, repeat - could work. However, the final step of creating a _train.bin_ and _val.bin_ files of...

@aartivnkt unfortunately I cannot advise about GPU - don't have one and know naathing :) What I can say is that the prepare.py script does not use GPU, so you...