Vadim Alperovich comments

Results 7 comments of


                                            Vadim Alperovich

Update pip release

Hi, I have the same problem with importing of the subdirectories. Could you tell if there is any solution?

Training on labeled only can achieve much better results than what is reported

Hi! I've obtained similar results to @IssamLaradji! @crux82 did you consider that increasing training epochs number for BERT model can be better in terms of quality metric than applying whole...

How model_seqlen affects quantization quality

@BlackSamorez Thanks for the answer! I am trying to quantize finetuned version of mixtral and I had no such long samples (8192) in the training set. Then should I decrease...

How model_seqlen affects quantization quality

@Godofnothing Thanks, good point!

Gan BERT for Multilabel Intent Classification

@rbhatia46 Hi! I've implemented this approach for multi-label classification, please see https://github.com/VirtualRoyalty/gan-token-classification

[Performance]: guided generation is very slow in offline mode

Hi, is this issue still relevant?

Implementation of Unit Test File Generation

Hi :) Any updates on this request?