Inquiry About Teacher, Student Models, and supported language in ConGen-BGE_M3-model-phayathaibert.
I am exploring the ConGen-BGE_M3-model-phayathaibert repository and would like to know more about the teacher and student models used in this setup. Could you please provide more details?
Additionally, can I assume that this model supports all the languages the BGE-M3 model supports, or is there a specific focus on Thai?
Thank you for your work on this!
To answer the question about the teacher and student setup. We trained the model using the method from ConGen, replacing the teacher model with bge-m3 and the student model with phayathaibert. We varied the parameters according to this setup ConGen Training Guide. See training code examples here: https://github.com/mrpeerat/Thai-Sentence-Vector-Benchmark/blob/main/ConGen-Thai.ipynb
For language support, this model will focus on the Thai language.