RT @Cohere_Labs: Don't forget to tune in tomorrow, August 26th as @ihor_step discusses Zero-Shot Named Entity Recognition with GLiNeR. Lea….

RT @Cohere_Labs: Join us for a deep dive into Zero-Shot Named Entity Recognition with GLiNeR presented by @ihor_step on Tuesday, August 26t….

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

RT @knowledgator: 🚀 GLiNER x SmolLM: a new joint encoder-decoder architecture 🚀. We are excited to release a new kind of GLiNER model built….

RT @knowledgator: 🚀 Our largest study on zero-shot text classification is out!.📄 We surpass cross-encoders while b….

RT @ClementDelangue: Every tech company can and should train their own deepseek R1, Llama or GPT5, just like every tech company writes thei….

RT @gm8xx8: GLiClass-V3: A family of encoder-only models that match or exceed DeBERTa-v3-Large in zero-shot accuracy, while delivering up t….

RT @knowledgator: 🚀 Introducing GLiClass‑V3 – a leap forward in zero-shot classification!. Matches or beats cross-encoder accuracy, while b….

RT @antoine_chaffin: Seeing ModernBERT and Ettin models being useful is heart warming.

RT @GoogleDeepMind: An advanced version of Gemini with Deep Think has officially achieved gold medal-level performance at the International….

🤔 Which task do you think is more complex for ML models?. 🔹 Multi-label classification .🔹 Entity recognition / Object detection (e.g., in computer vision). Vote and share your thoughts! 👇.

RT @knowledgator: 🚀 Introducing GLiNER-X: Breaking Language Barriers in Multilingual Zero-Shot NER! 🌍. We’re excited to roll out our latest….

RT @knowledgator: 🚀 GLiNER just reached 2K stars on GitHub — and it’s now 3x faster!. We’re thrilled to see this project grow into a core t….

RT @knowledgator: 🎉 Thanks to @GoogleStartups Ukraine Support Fund, we are receiving funding + long-term mentorship from @Google to grow ou….

RT @ClementDelangue: Biased but in my opinion ZeroGPU is one of the most impressive piece of infra in AI that no one is talking about. Po….

RT @Dorialexander: Breaking: @pleiasfr releases a new generation of small reasoning models for RAG and source synthesis. Pleias-RAG-350M an….

I just finished my small experiments comparing different encoder models on retrieval tasks. The goal was to check whether MLM is better than RTD for these tasks. I compared Electra's small models, both generator and discriminator, that have the same size. Additionally, it was

It became popular to think that architecture is no longer a thing; we achieved some local minimum in architectural design, and data right now is the limiting factor. This work empirically shows that this is not the case, that these aspects are more interesting and hide deeper.

RT @gm8xx8: ModernBERT vs DeBERTaV3. To isolate architecture from pretraining effects, ModernBERT is trained on the same dataset as CamemBE….

RT @knowledgator: ⚡ Introducing truly Flash DeBERTa implementation ⚡. DeBERTa remains a top-performing model in Named Entity Recognition, T….

RT @sijun_tan: Hey @sama, we know you're planning to open-source your reasoning model—but we couldn’t wait. Introducing DeepCoder-14B-Prev….