I am thrilled to share that my first paper "Simple Model Adaptation for Sparse Retrievers" in collaboration with @zongyuxuan3@bpiwowar and @LaureSoulier has been accepted in the "short papers track"! Special thanks to @basilevancooten too for the supervision #ecir2024
Cool project I've been working on in the last few weeks: Integrating ChatGPT and @sinequa so that ChatGPT can access the context, documents and data of your company.
Cc @gdb@karpathy
https://t.co/INq4y2uZ1K
@jobergum@binarymax@vespaengine@rodrigfnogueira@lintool this is amazing work, thank you so much for all of your blog posts that have been very insightful - may I ask what you train your models with ? I'm struggling with the fact that MS MARCO is under a highly restrictive licence and therefor cannot use it for commercial use.
@psuraj28@narges_tabari@huggingface Thanks ! Speaking of, are there any techniques proven to work to keep seq2seq Transformer models from straight out inventing stuff when being applied on summarization/generative QA tasks ?
💥 In this report, we’ll walk through implementing residual neural networks, the underlying architecture and the theory behind skip connections.
📕Read:https://t.co/SFamazLl2X
✍️Code:https://t.co/N2SxUQo9WF
#MachineLearning#DeepLearning
We will host a roadmap discussion for ONNX this Thursday (9/3) at 10 am PST. The discussion will focus on data pre-processing and integrating ONNX in a ML pipeline. Please join us at https://t.co/JCUZWlhA9W and submit your ideas at https://t.co/VBuWiLylz8.
@paulg@lisperati To me, this is similar to minimizing economic inequalities rather than minimizing poverty.
Try to minimize bad code / bad design / redundant features, yes.
I don't see why there should be an upper bound on good stuff.