Shamane Siri

Thanks for the detailed answer. Yes, this does answer my question. But let's say I have my custom data loader, which gets data from a spark parquet. So is it...

Yes it is wrong. The paper says we should get the normalize distribution over attentions at each outputs from the fusion layer .

@SaeedNajafi Saee Well if we take the conditional probability of generating the next word on given content . It's not only the previous word that keep an account of ....

@SaeedNajafi Can you please answer on this also . It will be a great help. [Issue #27](https://github.com/Marsan-Ma/tf_chatbot_seq2seq_antilm/issues/27)

Any luck with this? Can we extract a representation for the sentence "the cow goes moo

Hi is there any updates on BART pertaining script?

hi I am the person who revamped the RAG code base with the latest PL. Please follow @aRyBernAlTEglOTRO 's instructions. The current version is only working with the RAY. https://github.com/shamanez/transformers/blob/main/examples/research_projects/rag/lightning_base.py#L396...

Hi other than that your code worked really fine. Nice mate:). BTW did you try GAIL which is a recent inverse reinforcement learning algorithm?

As @geert-heyman mentioned you can go through those scrips . Simply these are the following steps 1. Create a vocabulary using both languages . 2. Index them (word 2 Index)...

This is nice. I took a different (dumb :D) approach previously. 1. TIES merge two models and then slice only the embedding layer 2. Then use the pass-through method Actually,...