Oh, thanks for this link. Looks great and it may suit me well. I need to settle in and learn Python but I am experiencing tremendously severe dysregulation at the moment, and my normal quick deep learning is simply not happening.
Multiple people have offered but decided against it for a few reasons.
- Proper reviews actually feel like they would take me as much time as doing it myself.
- One benefit of doing it all myself is that all the content has a familiar style.
- The downside of contributions is that a lot of the stuff that I see on YT just simply doesn't fit the style that I intend to have on calmcode. So before accepting contributions it also feels like I would have to vet the person who makes the contrib.
A lot of the aforementioned is more complex now as well due to the fact that folks can pay for the platform. It was a 100% free platform before, and right now it's a 99% free platform and some people pay a stipend to keep the site running. If contributions come in, I would also need to figure out a way to keep the incentives aligned, which also complicates things.
I've had a collaborator in the past and a bunch of things worked out there. But he's gone off to do other things, all of which is fair enough.
Hi, there are no free trials yet, but we do offer 14 days refund no questions asked. You can test it out and if it does not fit your needs, request for a refund :)
I'm always a bit worried if an extension gets permission to do anything it wants with all my files, including deleting them. Is there a way to restrict it and allow it to modify only the files it created?
The problem is to scale that properly. If you have millions of documents, that won't scale that well. You are not going to prompt the LLM millions of times, aren't you?
Embedding models usually have fewer parameters than the LLMs, and once we index the documents, their retrieval is also pretty fast. Using LLM as a judge makes sense, but only on a limited scale.
Why is that an issue? Training the tokenizer seems much more straightforward than training the model as it is based on the statistics of the input data. I guess it may take a while for massive datasets, but is calculating the frequencies impossible to be done on a bigger scale?
Are there any specific reasons for using BPE, not Unigram, in LLMs? I've been trying to understand the impact of the tokenization algorithm, and Unigram was reported to be a better alternative (e.g., Byte Pair Encoding is Suboptimal for Language Model Pretraining: https://arxiv.org/abs/2004.03720). I understand that the unigram training process should eliminate under-trained tokens if trained on the same data as the LLM itself.
I'm unsure if there is any comparison of LanceDB and Qdrant available out there, but there shouldn't be any issues with Python 3.12 and qdrant-client compatibility. Windows is also not a problem, as the typical local setup is usually based on Docker. Are there any specific features you are interested in?
If you will be the only app user, then the Python SDK's local mode might be suitable. However, in the long run, when you decide to publish the app, you rather have to switch to an on-premise or cloud environment. Using Qdrant from the very beginning might be a good idea, as the interfaces are kept the same, and the switch is seamless.
There is a whole movement around enshittification, and I see potential in this kind of app, even though it still seems to be a niche.
reply