Jina Reader

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

Jina Reader AI

@qorprate one off

@qorprate one off

@ChiragCX Oh man we thought Skills were the dead ones

@ChiragCX Oh man we thought Skills were the dead ones

Our official CLI for agents https://t.co/XLhRvLRuDc https://t.co/wFtN8i9YcA

Our official CLI for agents https://t.co/XLhRvLRuDc https://t.co/wFtN8i9YcA

The trend toward smaller embeddings is a shift. On-device retrieval, browser-based search, and edge deployment all demand models that fit in constrained memory budgets. Learn more about Small & Nano b

The trend toward smaller embeddings is a shift. On-device retrieval, browser-based search, and edge deployment all demand models that fit in constrained memory budgets. Learn more about Small & Nano below: - blog post: https://t.co/M8RJp2pczh - 🤗 weights including GGUFs and MLX: https://t.co/IwpUK9SzAV - arXiv: https://t.co/AsTenf1XDt

v5-text uses decoder-only backbones with last-token pooling instead of mean pooling. Four lightweight LoRA adapters are injected at each transformer layer, handling retrieval, text-matching, classific

v5-text uses decoder-only backbones with last-token pooling instead of mean pooling. Four lightweight LoRA adapters are injected at each transformer layer, handling retrieval, text-matching, classification, and clustering independently. Users select the appropriate adapter at inference time. For retrieval, queries get a "Query:" prefix and documents get "Document:". Context length is 32K tokens, a 4x increase over v3.

MMTEB (131 multilingual tasks): v5-small (677M) hits 67.0, next best sub-1B is 64.3. +2.7pt gap. MTEB English (41 tasks): v5-small leads at 71.7. v5-nano (239M) scores 71.0 -- matching models 2x its s

MMTEB (131 multilingual tasks): v5-small (677M) hits 67.0, next best sub-1B is 64.3. +2.7pt gap. MTEB English (41 tasks): v5-small leads at 71.7. v5-nano (239M) scores 71.0 -- matching models 2x its size. Retrieval (5 benchmarks): v5-small at 63.28 matches v4 (3.8B) while being 5.6x smaller. The nano model at 239M params has no peer in its weight class.

jina-embeddings-v5-text is here! Our fifth generation of jina embeddings, pushing the quality-efficiency frontier for sub-1B multilingual embeddings. Two versions: small & nano, available today o

jina-embeddings-v5-text is here! Our fifth generation of jina embeddings, pushing the quality-efficiency frontier for sub-1B multilingual embeddings. Two versions: small & nano, available today on Elastic Inference Service, vLLM, GGUF and MLX. https://t.co/68GGuBRdy4

@tmztmobile It will be a lossy compression, like impressionist lossy

@tmztmobile It will be a lossy compression, like impressionist lossy

Check out the live demo https://t.co/W1EXpDFCAL and see it in action. Our read our repo and paper for more technical details on training and decoding.

Check out the live demo https://t.co/W1EXpDFCAL and see it in action. Our read our repo and paper for more technical details on training and decoding.

Text embeddings are widely assumed to be safe, irreversible representations. We show we can reconstruct the original text using conditional masked diffusion. Existing inversions (Vec2Text, ALGEN, Zer

Text embeddings are widely assumed to be safe, irreversible representations. We show we can reconstruct the original text using conditional masked diffusion. Existing inversions (Vec2Text, ALGEN, Zero2Text) generate tokens autoregressively and require iterative re-embedding through the target encoder. We take a different approach: embedding inversion as conditional masked diffusion. Starting from a fully masked sequence, a denoising model reveals tokens at all positions in parallel, conditioned on the target embedding via adaptive layer normalization (AdaLN-Zero). Each denoising step refines all positions simultaneously using global context, without ever re-embedding the current hypothesis.

Most don't know (1) how easy it is to invert embedding vectors back into sentences, (2) this is a perfect task text diffusion models. Here's a 78M parameter model and live demo that recovers 80% of to

Most don't know (1) how easy it is to invert embedding vectors back into sentences, (2) this is a perfect task text diffusion models. Here's a 78M parameter model and live demo that recovers 80% of tokens from Qwen3-Embedding and EmbeddingGemma vectors. Works even on multilingual input.

@Prince_Canuma @liquidai @deepseek_ai @Alibaba_Qwen @allen_ai @TencentHunyuan @PaddlePaddle 🔥

@Prince_Canuma @liquidai @deepseek_ai @Alibaba_Qwen @allen_ai @TencentHunyuan @PaddlePaddle 🔥

0.6B params. Top3 on MTEB reranking task. 10× smaller than generative listwise rerankers. Read more about this Best Paper at AAAI Frontier IR here: https://t.co/vVKleKBqI9 https://t.co/DFur26280e

0.6B params. Top3 on MTEB reranking task. 10× smaller than generative listwise rerankers. Read more about this Best Paper at AAAI Frontier IR here: https://t.co/vVKleKBqI9 https://t.co/DFur26280e

jina-reranker-v3 was the first listwise reranker to throw all documents into one context window (where traditional rerankers loop over ⟨q,d⟩ pairs) and let them fight it out via self-attention—what we

jina-reranker-v3 was the first listwise reranker to throw all documents into one context window (where traditional rerankers loop over ⟨q,d⟩ pairs) and let them fight it out via self-attention—what we call "last but not late" interaction. Bold or stupid? But not mediocre. Today it won Best Paper at AAAI Frontier IR Workshop.

Here's why it works: embeddings lie on a hypersphere, so d-1 angles can replace d Cartesian coordinates. In high dimensions, those angles concentrate around pi/2, causing IEEE 754 exponents to collaps

Here's why it works: embeddings lie on a hypersphere, so d-1 angles can replace d Cartesian coordinates. In high dimensions, those angles concentrate around pi/2, causing IEEE 754 exponents to collapse to a single value. This makes the byte stream highly compressible. https://t.co/pejBpRsnY0