Hackernews posts about HuggingFace
- New benchmarks (i.e. Nvidia canary) in HuggingFace ASR leaderboard (huggingface.co)
- Show HN: TabPFN-2.5 – SOTA foundation model for tabular data (priorlabs.ai)
- Show HN: Chonky – a neural text semantic chunking goes multilingual (huggingface.co)
- Show HN: Project AELLA – Open LLMs for structuring 100M research papers (aella.inference.net)
- Show HN: We tested 9 AI models with 37K+ security tests (www.modelred.ai)
- Show HN: BesiegeField – LLM Agents Learn to Build Machines in a Physics Sandbox (besiegefield.github.io)
- Show HN: Free unlimited AI video animation with no daily limits or signups (animateforever.com)
- Show HN: Fantail SLMs for Coding Agents (www.autohand.ai)
- The Smol Training Playbook: The Secrets to Building World-Class LLMs (huggingface.co)
- Drax: Speech Recognition with Discrete Flow Matching (huggingface.co)
- Sentence Transformers is joining Hugging Face (huggingface.co)
- Show HN: The Legal Embedding Benchmark (MLEB) (huggingface.co)
- The Smol Training Playbook: The Secrets to Building World-Class LLMs (huggingface.co)
- Kimi-K2-Thinking: open weights LLM with frontier performance (huggingface.co)
- Unlocking On-Policy Distillation for Any Model Family (huggingface.co)
- Australian-made LLM beats OpenAI and Google at legal retrieval (huggingface.co)
- The 1B Token Challenge: Finding the Perfect Pre-Training Mix (huggingface.co)
- Weibo launch open source AI, VibeThinker-1.5B (huggingface.co)
- Granite 4.0 Nano: Just how small can you go? (huggingface.co)
- The Smol Training Playbook: The Secrets to Building World-Class LLMs (huggingface.co)
- Maya1: Open-source 3B Voice Model (huggingface.co)