8 ms·
You probably shouldn't use OpenAI's embeddings
- nico 3y agoIs someone doing embeddings<>embeddings mapping? For example, mapping embeddings of Llama to GPT-3? That way you can see how similar the models “understand the world”.
- fzliu 3y agoI'm curious about this as well. There are potentially many different versions of embedding models used in production and correlating different versions together could be very important.
- sroussey 3y agoI’d be interested to see this as well. I guess you can make a test and see what happens. Ping me if you do!
- KRAKRISMOTT 3y agoIsn't this the whole point behind the CLIP architecture?
- nico 3y agoCould you explain a bit more what the CLIP architecture is? Any good links to a short video demoing it? Thank you!
- KRAKRISMOTT 3y agohttps://www.pinecone.io/learn/clip-image-search/ https://www.pinecone.io/learn/clip-image-search/
- mustacheemperor 3y agoCould anyone point me towards a relatively beginner-friendly guide to do something like >download all my tweets (about 20k) and build a semantic searcher on top ? How can utilize 3rd party embeddings with OpenAI's LLM API? Am I correct to understand from this article that this is possible?
- diego 3y agoThat's exactly what I did here. https://github.com/dbasch/semantic-search-tweets https://github.com/dbasch/semantic-search-tweets
- mustacheemperor 3y agoThank you! Comparing this and the link the other commenter posted, what handles the actual search querying? Does instructor-xl include an LLM in addition to the embeddings? The other commenter's repo uses Pinecone for the embeddings and OpenAI for the LLM. My apologies if I am completely mangling the vocabulary here - I have an, at best, rudimentary understanding of this stuff that I am trying to hack my education on. Edit: If you're at the SF meetup tomorrow, I'd happily buy you a beverage in return for this explanation :)
- eternalban 3y agoIt's in the repo: You first create embeddings. What is this? It's an n-dimensional vector space with your tweets 'embedded' in that space. Each word is an n-dimensional vector in this space. The vectorization is supposed to maintain 'semantic distance'. Basically, if two words are very close in meaning or related (by say frequently appearing next to each other in corpus) they should be 'close' in some of those n-dimensions as well. The result at the end is the '.bin' file, the 'semantic model' of your corpus. https://github.com/dbasch/semantic-search-tweets/blob/main/embed-tweets-instructor.py#L25 https://github.com/dbasch/semantic-search-tweets/blob/main/e... For semantic search, you run the same embedding algorithm against the query and take the resultant vectors and do similarity search via matrix ops, resulting in a set of results, with probabilities. These point back to the original source, here the tweets, and you just print the tweet(s) that you select from that result set (here the top 10). https://github.com/dbasch/semantic-search-tweets/blob/main/search-instructor.py https://github.com/dbasch/semantic-search-tweets/blob/main/s... Experts can chime in here but there are knobs such as 'batch size' and the functions you use to index. (cosine was used here.) So the various performance dimensions of the process should also be clear. There is a fixed cost of making the embeddings of your data. There is a per-op embedding of your query, and then running the similarity algorithm to find the result set.
- nomadiccoder 3y agoThe heat map of availability time, 98.58 (Jan), 99.07 (Feb), and 99.71 (Mar) trends upwards.
- reportgunner 3y agoI don't think 3 data points is enough to establish a trend.
- celestialcheese 3y agoVery interested in this - I've been using embeddings / semantic search doing information retrieval from PDFs, using ada-002, and have been impressed by the results in testing. The reasons the article listed, namely a) lock-in and b) cost, have given me pause with embedding our whole corpus of data. I'd much rather use an open model but don't have much experience in evaluating these embedding models and search performance - still very new to me. Like what you did with ada-002 vs Instruct XL, has there been any papers or prior work done evaluating the different embedding models?
- VHRanger 3y agoYou can find some comparisons and evaluation datasets/tasks here: https://www.sbert.net/docs/pretrained_models.html https://www.sbert.net/docs/pretrained_models.html Generally MiniLM is a good baseline. For faster models you want this library: https://github.com/oborchers/Fast_Sentence_Embeddings https://github.com/oborchers/Fast_Sentence_Embeddings For higher quality ones, just take the bigger/slower models in the SentenceTransformers library
- sroussey 3y agoIs there performance comparisons for Apple Silicon machines?
- VHRanger 3y agoPerformance in terms of model quality would be the same. The fast-se library uses C++ code and word embeddings being averaged to generate sentence embeddings, so would be similarly fast, or faster on apple silicon than x86. For the SentenceTransformer library models I'm not sure, but I think it would run off the CPU for a M1/M2 computer
- breckenedge 3y agoIt’s fine to use their embeddings for a proof of concept, but since you don’t own it, you probably shouldn’t rely on it because it could go away at any time.
- adeelk93 3y agoCouldn’t you make that argument against all SaaS?
- anshumankmr 3y agoWell some SaaS purists may believe that to be true about Microsoft Office 365. Hence we have, Microsoft Office 2021. (Although there is a lot more advantages to just having Office 2021 like the flat fee)
- breckenedge 3y agoYes and no. Sometimes you get contracts that require “source code escrow” so that companies can run your source if you ever go out of business.
- fzliu 3y agoI've done some quick-and-dirty testing with OpenAI's embedding API + Zilliz Cloud. The 1st gen embeddings leave something to be desired (https://medium.com/@nils_reimers/openai-gpt-3-text-embeddings-really-a-new-state-of-the-art-in-dense-text-embeddings-6571fe3ec9d9 https://medium.com/@nils_reimers/openai-gpt-3-text-embedding...), but the 2nd gen embeddings are actually fairly performant relative to many open source models with MLM loss. I'll have to dig out the notebook that I created for this, but I'll try to post it here once I find it.
- shirkey 3y agoPlease do and thanks in advance for any insights you can provide -- it would be great to understand any benchmarking improvement with ada-002 from your previous findings, and whether you tested the specific OpenAI text-search-*-{query,doc} models as a comparison for large document search.
- lee101 3y ago[dead]