google-research
jina
google-research | jina | |
---|---|---|
98 | 126 | |
32,915 | 20,121 | |
1.1% | 1.3% | |
9.6 | 9.1 | |
4 days ago | 23 days ago | |
Jupyter Notebook | Python | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
google-research
-
Show HN: Next-token prediction in JavaScript – build fast LLMs from scratch
People on here will be happy to say that I do a similar thing, however my sequence length is dynamic because I also use a 2nd data structure - I'll use pretentious academic speak: I use a simple bigram LM (2-gram) for single next-word likeliness and separately a trie that models all words and phrases (so, n-gram). Not sure how many total nodes because sentence lengths vary in training data, but there are about 200,000 entry points (keys) so probably about 2-10 million total nodes in the default setup.
"Constructing 7-gram LM": They likely started with bigrams (what I use) which only tells you the next word based on 1 word given, and thought to increase accuracy by modeling out more words in a sequence, and eventually let the user (developer) pass in any amount they want to model (https://github.com/google-research/google-research/blob/5c87...). I thought of this too at first, but I actually got more accuracy (and speed) out of just keeping them as bigrams and making a totally separate structure that models out an n-gram of all phrases (e.g. could be a 24-token long sequence or 100+ tokens etc. I model it all) and if that phrase is found, then I just get the bigram assumption of the last token of the phrase. This works better when the training data is more diverse (for a very generic model), but theirs would probably outperform mine on accuracy when the training data has a lot of nearly identical sentences that only change wildly toward the end - I don't find this pattern in typical data though, maybe for certain coding and other tasks there are those patterns though. But because it's not dynamic and they make you provide that number, even a low number (any phrase longer than 2 words) - theirs will always have to do more lookup work than with simple bigrams and they're also limited by that fixed number as far as accuracy. I wonder how scalable that is - if I need to train on occasional ~100-word long sentences but also (and mostly) just ~3-word long sentences, I guess I set this to 100 and have a mostly "undefined" trie.
I also thought of the name "LMJS", theirs is "jslm" :) but I went with simply "next-token-prediction" because that's what it ultimately does as a library. I don't know what theirs is really designed for other than proving a concept. Most of their code files are actually comments and hypothetical scenarios.
I recently added a browser example showing simple autocomplete using my library: https://github.com/bennyschmidt/next-token-prediction/tree/m... (video)
And next I'm implementing 8-dimensional embeddings that are converted to normalized vectors between 0-1 to see if doing math on them does anything useful beyond similarity, right now they look like this:
[nextFrequency, prevalence, specificity, length, firstLetter, lastLetter, firstVowel, lastVowel]
- Google Research website is down
-
Jpegli: A New JPEG Coding Library
The change was literally just made: https://github.com/google-research/google-research/commit/4a...
It appears this was in response to Hacker News comments.
- Multi-bitrate JPEG compression perceptual evaluation dataset 2023
-
Vector Databases: A Technical Primer [pdf]
There are options such as Google's ScaNN that may let you go farther before needing to consider specialized databases.
https://github.com/google-research/google-research/blob/mast...
-
Labs.Google
I feel it was unnecesary to create this because https://research.google/ already exists? It just seems like they want to take another URL with a "pure" domain name instead of psubdirectories, etc parts.
-
Smerf: Streamable Memory Efficient Radiance Fields
https://github.com/google-research/google-research/blob/mast...
-
Shisa 7B: a new JA/EN bilingual model based on Mistral 7B
You could also try some dedicated translation models like https://huggingface.co/facebook/nllb-moe-54b (or https://github.com/google-research/google-research/tree/master/madlad_400 for something smaller) and see how they do.
-
Translate to and from 400+ languages locally with MADLAD-400
Google released T5X checkpoints for MADLAD-400 a couple of months ago, but nobody could figure out how to run them. Turns out the vocabulary was wrong, but they uploaded the correct one last week.
- Mastering ROUGE Matrix: Your Guide to Large Language Model Evaluation for Summarization with Examples
jina
- Jina.ai: Self-host Multimodal models
- FLaNK Stack Weekly for 30 Oct 2023
-
Cross data type search that wasn’t supported well using Elasticsearch
Jina mainly because of their use of neural networks and AI.
- Recommend a Lightweight Launcher with Nested Folders
-
I plan to build my own AI powered search engine for my portfolio. Do you know ones that are open-source?
Jina - It’s an open-source project where you can build search engines. Well maybe not no code but it claims that you only need a few lines of code for creating projects. The project supports semantic, text, image, audio, and video search. What I’m also interested in is with their neural search and generative AI. I’m also interested in the amount of github repo that they have. I have this on my radar since this is also something I was interested in.
-
How can we match images in our database?
Do you guys have any ideas how we can match images on our database? We’re working on a project that about matching images on our database. We were trying to use SIFT and some other similar methods, but for some reason, nothing doesn’t seem to be working that well. Does anyone have any suggestions for the most effective way to do this? Maybe some open-source solutions like HuggingFace or Jina AI? We just want to make sure our image matching is correct and that part’s been a bit of a struggle on our part.
-
Can AI 3D model search engines be a thing this year?
The tech lets you find 3D models without sifting through tons of text - An information retrieval framework does the heavy lifting and compares models to each other, no descriptions or keywords needed.
-
Any MLOps platform you use?
Jina AI -They offer a neural search solution that can help build smarter, more efficient search engines. They also have a list of cool github repos that you can check out. Similar to Vertex AI, they have image classification tools, NLPs, fine tuners etc.
-
This week(s) in DocArray
Well, it's not exactly a new feature, but we've been working on early support for DocArray v2 in Jina.
-
Multi-model serving options
Jina let’s you serve all of your models through the same Gateway while deploying them as individual microservices. You can also tie your models together in a pipeline if needed. Also some nice ML focussed features such as dynamic batching.
What are some alternatives?
qdrant - Qdrant - High-performance, massive-scale Vector Database for the next generation of AI. Also available in the cloud https://cloud.qdrant.io/
Weaviate - Weaviate is an open-source vector database that stores both objects and vectors, allowing for the combination of vector search with structured filtering with the fault tolerance and scalability of a cloud-native database.
fast-soft-sort - Fast Differentiable Sorting and Ranking
haystack - :mag: LLM orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited for building RAG, question answering, semantic search or conversational agent chatbots.
faiss - A library for efficient similarity search and clustering of dense vectors.
dalle-flow - 🌊 A Human-in-the-Loop workflow for creating HD images from text
ml-agents - The Unity Machine Learning Agents Toolkit (ML-Agents) is an open-source project that enables games and simulations to serve as environments for training intelligent agents using deep reinforcement learning and imitation learning.
whoogle-search - A self-hosted, ad-free, privacy-respecting metasearch engine
Milvus - A cloud-native vector database, storage for next generation AI applications
es-clip-image-search - Sample implementation of natural language image search with OpenAI's CLIP and Elasticsearch or Opensearch.
struct2depth - Models and examples built with TensorFlow
growthbook - Open Source Feature Flagging and A/B Testing Platform