RJ Honicky
@honicky
Followers
101
Following
43
Media
8
Statuses
75
Telecom, AI, technology for developing regions and percussion. "Minds that seek revenge destroy states, while those that seek reconciliation build nations."
Mostly SF Bay Area
Joined April 2009
Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.
821
3K
10K
Working with a large set of AnnData files on S3 is tough if you don't have the metadata indexed, so I created a tool to help. It uses partial downloads to dramatically speed up extracting the metadata without downloading the whole file.
1
1
4
I will talk about labeling 100M cells using a customized version of GenePT and scGPT. Register here:
linkedin.com
#GenerativeAI is transforming the way we analyze data — but what happens when it meets the complexity and scale of #MultiOmics? Join us for a new webinar of our #OmicsConnect series to hear from Dr....
0
0
0
Learn about what I'm doing with GenAI x Multiomics at this talk on May 29th
1
0
0
I promised to explain how you can build your very own, custom GenePT embeddings. Here's another Lab Note:. Enjoy!.
learning-exhaust.hashnode.dev
Better Gene Embeddings Through Prompt Engineering (Or, at Least, We Tried)
0
0
0
. and then on how we can push the limit on real world applications using novel approaches. It focus on the data and model side of things, so don't be intimidated if you're not familiar with genetics and biology data. Much more to come, enjoy!.
0
0
0
This is a quick "learning exhaust" post that talks about a little set of experiments I ran to prep for a larger collaborative series of posts about various exciting things that we can accomplish on real problems using existing technologies. .
learning-exhaust.hashnode.dev
In the spirit of “Learning in Public,” and “learning exhaust,” I’m going to start adding “Lab Notes” blog posts that chronicle little discoveries or failures that I have along the way to a larger...
1
0
1
I've been exploring AI models for Single Cell and Spatial Transcriptomics problems with my good friend Eugene Bolotin, Ph.D., and there is so much exciting work to do, great problems to tackle, and way less mainstream attention because it requires some basic Bio background.
1
0
0
Data Products are Different.
learning-exhaust.hashnode.dev
Why you have to manage data products differently than software and hardware.
0
0
1
Data products often scale cost like hardware, iterate like software, and scale performance like. data products. @swyx @eugeneyan
1
0
1
Oh man, writing up the Cerebras paper on Weight Streaming was a rabbit hole filled with parallel algorithms. @swyx @picocreator @eugeneyan @SarahChieng @latentspacepod.
1
0
1
I've been trying to post a bit more frequently and in smaller bites, so I'm going to try to pick one interesting thing I've learned from papers I read, and then write a quick post about them. Here's my first one:.
learning-exhaust.hashnode.dev
Modern training techniques for embedding models mean that we should probably include a prompt or fine tune to a specific type of task.
0
0
0
RT @latentspacepod: we are going thru OpenAI’s sCM paper on the LS Paper Club today with @honicky!!. (search our youtube or discord to join….
0
1
0
RT @latentspacepod: thanks to @honicky and @vibhuuuus for another great Paper Club on ReFT: Representation Fine Tuning by @aryaman2020 et a….
0
1
0
How am I supposed to accomplish anything when ChatGPT is screwing up whitespace?
1
0
1
I got a lot of positive feedback when I mentioned this at the Latent Space Paper Club this week, so I thought I'd write it up. Is it the model or the data that's low rank? . @swyx @latentspacepod @vibhuuuus.
learning-exhaust.hashnode.dev
Thoughts on LoRA Quantization Loss Recovery
0
1
4
Some ideas and experiments on quantization: Oooh, I'm so excited!.
learning-exhaust.hashnode.dev
or post-post-training-quantization-training :)
0
0
0