Yeah. I doubt that there will be a follow-up paper though. I think the point was more like "transformers are the shiny new toy, but don't forget about also implementing simple baselines when you start a new project".
CODE
-
Gzip and kNN: Text Similarity Through Compression
By
–
In short, the idea behind this gzip + kNN approach is that a given text A is close to the size of two concatenated texts A + B when compressed using Gzip. Or, the compressed size of A+A is similar to the compressed size of just A (plus a constant value for the back reference)
-
HyperDreamBooth: Fast Text-to-Image Model Personalization with HyperNetworks
By
–
HyperDreamBooth: HyperNetworks for Fast Personalization of Text-to-Image Models paper page: https://
huggingface.co/papers/2307.06
949
… Personalization has emerged as a prominent aspect within the field of generative AI, enabling the synthesis of individuals in diverse contexts and styles, while -
H2O.ai Catalog: 100 Hydrogen Torch Use Cases Collection
By
–
3/ We’re very excited to share the http://
H2O.ai Catalog: Collection of Hydrogen Torch Use Cases. We’ve curated a collection of 100 diverse use cases, datasets, models, and outputs; all developed using Hydrogen Torch. -
PyTorch and Deep Learning Software Podcast with Weights and Biases
By
–
Just did a podcast covering a bunch of PyTorch / DL Software related topics, hosted by @l2k from @wandb
. It was a lot of fun.
Listen here https://
bit.ly/3NIGFBa -
Development Framework for Building Character Agents
By
–
We just did a webinar on creating character agents today, and this is a fantastic way to follow that up – a development framework for building your own!!!
-
KNN Classification Works Through Majority Vote Among Neighbors
By
–
Yes, I assume this should work because the compute the label based on majority vote among the nearest neighbors.
-
Stacking Chains as Tools in AI Agents
By
–
A great overview of using an agent to interact with custom data! Step 1: Upload data to Metal
Step 2: Create a retrieval QA chain
Step 3: Use that chain as a **tool** in your final agent This idea of stacking chains as tools (or agents as tools) is super powerful! -
Always-on System Instability and Deployment Architecture Challenges
By
–
Hey Marwan! That's definitely frustrating. Sorry, always-on has been unstable lately. There are fundamental issues with the design of that system that makes it hard to keep reliable, namely that dev is also prod. Instead we've rolled out Deployments which achieves significantly
-
Nearest-Neighbor Approach on Normalized Bag-of-Words Vectors
By
–
Thinking about this a bit more, I think nearest-neighbor on normalized bag-of-words vectors would probably also perform well here. Because the count vectors of two similar documents are similar just like the compressed + concatenated compressed docs are similar.