https://nitter.cf/t.co/jlZmucuX7G
Bangalore, India
Joined September 2012
- Tweets232
- Following1.8K
- Followers438
- Likes752
Pinned Tweet
🧠Introducing 𝐛𝐚𝐛𝐲𝐚𝐠𝐢-𝐚𝐬-𝐚-𝐬𝐞𝐫𝐯𝐢𝐜𝐞.
🤝Integrate @yoheinakajima 's babyagi with your own applications - thanks to @langchain and @JinaAI_'s langchain-serve with one simple command - 𝐥𝐜-𝐬𝐞𝐫𝐯𝐞 𝐝𝐞𝐩𝐥𝐨𝐲 𝐛𝐚𝐛𝐲𝐚𝐠𝐢
github.com/jina-ai/langchain…
Deepankar Mahapatro retweeted
Today we’re announcing Open Responses: an open-source spec for building multi-provider, interoperable LLM interfaces built on top of the original OpenAI Responses API.
✅ Multi-provider by default
✅ Useful for real-world workflows
✅ Extensible without fragmentation
Build agentic systems without rewriting your stack for every model: openresponses.org
I wrote about the issues we faced trusting "OpenAI-compatible" claims across models and providers. @simonw highlighted the same problem in his AAIF post. Hope @linuxfoundation considers this - a real spec would open doors for languages beyond Python/TS.
deepankarm.github.io/posts/o…
Deepankar Mahapatro retweeted
🤨 This fine-tuning is too good to be true? Just tell us what you want your embeddings to excel at, e.g., car insurance claims, financial news, Spanish dialogs. Specify your wish in a prompt; and remember this is your only input to our API. In about 30 mins, we then deliver a ready-to-use, fine-tuned embedding model that can be loaded via SentenceTransformers. Behind the scenes, we take care of everything else: from generating useful synthetic data to managing the train-eval-test ML workflow, and finally, uploading the fine-tuned model to the @huggingface Hub. Yep, under this very minimal UI abstraction, so much happens!
Deepankar Mahapatro retweeted
We are releasing two new open-source reranker models: jina-reranker-v1-turbo-en and jina-reranker-v1-tiny-en, the latter has only 30M parameters and four layers! 🤯 These two new rerankers enjoy 5X faster inference speed than our last jina-reranker-v1-base model at only a very small cost on the quality. You can find these two models on @huggingface
Turbo (6-layer, 37.8M parameters): huggingface.co/jinaai/jina-r…
Tiny (4-layer, 33M parameters):
huggingface.co/jinaai/jina-r…
Deepankar Mahapatro retweeted
Feeding webpages to LLMs is crucial for grounding, but it's hard to do right. Scraping webpages is complex and unreliable, especially with dynamic pages. 🥁Introduce Jina Reader: simply prefix any URL with 𝗵𝘁𝘁𝗽𝘀://𝗿.𝗷𝗶𝗻𝗮.𝗮𝗶 and get an LLM-friendly input! Our Reader API acts as a proxy that processes any URL by performing browser rendering, content extraction, and cleaning, ultimately converting it into well-formatted text that is ready to use. It does all the nasty job so you don't have to.
Deepankar Mahapatro retweeted
Jina's new reranker can dramatically improve the quality of your RAG application!
Vector search by itself can help you search your data by meaning, but paper after paper shows using a post-retrieval reranker like @JinaAI_'s just-released jina-reranker-v1-base-en can give the quality of your results and the resulting query to the LLM a big boost.
Step through using their reranker step-by-step with LlamaIndex and @MistralAI:
jina.ai/news/precise-rag-wit…
Deepankar Mahapatro retweeted
🚀 Introducing Reranker API! Leveraging our latest research for top-notch performance, enjoy a boost of +20% in accuracy on vector search & RAG systems. Kickstart with 1 million tokens on us! jina.ai/reranker?nocache=1
Deepankar Mahapatro retweeted
Our open-source bilingual Spanish-English embedding model is now ready for download through @huggingface. Start using it right away to power your state-of-the-art AI applications.
huggingface.co/jinaai/jina-e…
Deepankar Mahapatro retweeted
Absolutely DELIGHTED to finally reveal details of my brand-new book series 'We Solve Murders', about Amy and Steve, a daughter-in-law/father-in-law detective duo. If you like The Thursday Murder Club, I think you'll love this. Out in September, and I can't wait for you to read it!
Deepankar Mahapatro retweeted
Gemini 1.5 Pro - A highly capable multimodal model with a 10M token context length
Today we are releasing the first demonstrations of the capabilities of the Gemini 1.5 series, with the Gemini 1.5 Pro model. One of the key differentiators of this model is its incredibly long context capabilities, supporting millions of tokens of multimodal input. The multimodal capabilities of the model means you can interact in sophisticated ways with entire books, very long document collections, codebases of hundreds of thousands of lines across hundreds of files, full movies, entire podcast series, and more.
Gemini 1.5 was built by an amazing team of people from @GoogleDeepMind, @GoogleResearch, and elsewhere at @Google. @OriolVinyals (my co-technical lead for the project) and I are incredibly proud of the whole team, and we’re so excited to be sharing this work and what long context and in-context learning can mean for you today!
There’s lots of material about this, some of which are linked to below.
Main blog post:
blog.google/technology/ai/go…
Technical report:
“Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context”
goo.gle/GeminiV1-5
Videos of interactions with the model that highlight its long context abilities:
Understanding the three.js codebase: youtube.com/watch?v=SSnsmqIj…
Analyzing a 45 minute Buster Keaton movie: youtube.com/watch?v=wa0MT8Ow…
Apollo 11 transcript interaction: youtube.com/watch?v=LHKL_210…
Starting today, we’re offering a limited preview of 1.5 Pro to developers and enterprise customers via AI Studio and Vertex AI. Read more about this on these blogs:
Google for Developers blog:
developers.googleblog.com/20…
Google Cloud blog:
cloud.google.com/blog/produc…
We’ll also introduce 1.5 Pro with a standard 128,000 token context window when the model is ready for a wider release. Coming soon, we plan to introduce pricing tiers that start at the standard 128,000 context window and scale up to 1 million tokens, as we improve the model.
Early testers can try the 1 million token context window at no cost during the testing period. We’re excited to see what developer’s creativity unlocks with a very long context window.
Let me walk you through the capabilities of the model and what I’m excited about!
Deepankar Mahapatro retweeted
Introducing Sora, our text-to-video model.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
openai.com/sora
Prompt: “Beautiful, snowy Tokyo city is bustling. The camera moves through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes.”
Deepankar Mahapatro retweeted
Jina AI's new bilingual Spanish-English embedding model brings the state-of-the-art in AI to half a billion Spanish speakers.
jina.ai/news/aqui-se-habla-e…
Deepankar Mahapatro retweeted
😇 New week, new embeddings for you! Announce 𝗷𝗶𝗻𝗮-𝗲𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝘀-𝘃𝟮-𝗯𝗮𝘀𝗲-𝗰𝗼𝗱𝗲, optimized for code search. This powerful model supports searches between English and 30 programming languages, all with an impressive 8K token length and SOTA performance!🚀
Deepankar Mahapatro retweeted
Channeling JFK's "Ich bin ein Berliner" vibes, we've just launched our German/English bilingual embedding model 🇩🇪🇺🇸
8k tokens, superior benchmark performance, 1/7th the size of similar models, setting a new standard in open-source language models
jina.ai/news/ich-bin-ein-ber…
Deepankar Mahapatro retweeted
大家好！
We're happy to announce the world's first bilingual Chinese-English embedding model with 8192 token-length.
Encode text in either language, and use either as query or target document. And with 8k tokens, encode long documents with ease!
jina.ai/news/8k-token-length…
Deepankar Mahapatro retweeted
Jina AI and @MongoDB have collaborated to bring you a tutorial on using Jina Embeddings v2 to bring state-of-the-art semantic search to MongoDB Atlas. Supercharge MongoDB's multi-cloud vector search solutions with Jina AI's industry-leading embeddings!
mongodb.com/developer/produc…
Deepankar Mahapatro retweeted
Announcing our $14.1M Seed round
extropic.ai/accelerate
