gardnr

15 ms·

241.
by gardnr 3y ago
I worry more about the paychecks staying the same size: https://www.epi.org/productivity-pay-gap/
242.
by gardnr 3y ago
The author is discussing fine-tuning a base model. How long it takes really depends on the dataset, the method, and the hyperparameters. DPO, for example, can achieve some great results with a fraction of the steps of other methods. Just li
243.
by gardnr 3y ago
Jeff Atwood has an opinion: https://meta.discourse.org/t/options-to-disable-hijack-of-cm...
244.
by gardnr 3y ago
I just noticed this has been added to langchain: https://python.langchain.com/docs/modules/data_connection/do...
245.
by gardnr 3y ago
Arguing isn't the correct word for that. I did express incredulity when asking for clarification. The person I was replying to did minimize the effort involved in upgrades by enclosing it in parentheses. The incredulity is based on exp
246.
by gardnr 3y ago
I haven't found any frameworks that offer it. The best explanation of an implementation that can take a stream of unformatted text and map over it to determine when a topic changes is explained in this video: https://youtu.b
247.
by gardnr 3y ago
I have been tempted to go back to turn of the century tech: html tables for layout with inline css for styling and smarty template for separation of concerns all driven by php.
248.
by gardnr 3y ago
So, your team upgraded to React 18 without having to rewrite anything?
249.
by gardnr 3y ago
Which libraries have you had trouble using? The official preact/compat library works great for me https://preactjs.com/guide/v10/switching-to-preact/
250.
by gardnr 3y ago
Do you plan to offer content aware chunking?
251.
by gardnr 3y ago
It will depend on the jurisdiction. It will be fun when a model is trained in Japan on material that would be copyrighted in the USA, then produces an image that is not able to be copyrighted in the USA. > Japan's government recentl
252.
by gardnr 3y ago
definitely sounds like an ssm.
253.
by gardnr 3y ago
I read the paper "Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modeling" that was published last week and started building a tool for people to collectively generate synthetic training data. The tool still
254.
by gardnr 3y ago
I just set it as my default search engine for a day. It's not quite there for my use cases. Can we help improve the search results?
255.
by gardnr 3y ago
What time series data do you have that you'd like to run through it?
256.
by gardnr 3y ago
A popular move with ai / ml folks: Use GitHub to publish information about a thing that is not open. Then "It's on GitHub".
257.
by gardnr 3y ago
I wonder how many of the samples from this demo video are authentic: https://arstechnica.com/information-technology/2023/12/googl...
258.
by gardnr 3y ago
It’s got a low power mode where it can still do some processing: https://youtube.com/watch?v=hS-NiaGeeVA
259.
by gardnr 3y ago
2.5 GbE can operate over CAT 5e.
260.
by gardnr 3y ago
The pascal language requires things to be declared in a certain order. It's a bit awkward some of the time bit it enables the compiler to work in a single pass. This meant that running an application was extremely fast by any standards
261.
by gardnr 3y ago
I was going to ask how you integrated BM25 but then I found this: https://docs.llamaindex.ai/en/stable/examples/retrievers/bm2...
262.
by gardnr 3y ago
How is it better than the model from the team that made the dataset? https://huggingface.co/Open-Orca/Mistral-7B-SlimOrca
263.
by gardnr 3y ago
Link: https://pypi.org/project/Mako/
264.
by gardnr 3y ago
I can fit 13B Q4 K M models on a 12GB RTX 3060. It OOMs when the context window goes above 3k. I get 25 tok/s.
265.
by gardnr 3y ago
I read a bit of sarcasm in parent comment. It seemed they were "telling it how it is" not agreeing with the way it is.
266.
by gardnr 3y ago
Then there's something about Heart Rate Variability (HRV), the vagus nerve, parasympathetic nervous system, and gut biome all being connected. I can measure one of those relatively easily with a chest strap. It seems like the only thin
267.
by gardnr 3y ago
Yeah, the almonds, cinnamon, and vanilla in the protein powder really even out all the weird stuff.
268.
by gardnr 3y ago
Somewhere between 3-5 grams.
269.
by gardnr 3y ago
I have a high powered blender. I try to get the fiber and the probiotics at the same time. Fruits and vegetables have probiotics inside the flesh, not just on the skin. The science is less clear on supplements. [1] The night before I soak:
270.
by gardnr 3y ago
I tried to watch MotoGP the other day on "real TV". I turned it off because the ads were so bad.

More ›