Build Real Agentic Apps with CUGA: 24 Working Examples
CUGA introduces a lightweight harness enabling developers to build real agentic applications. It includes 24 working examples demonstrating practical implementations across various use cases.
Introducing the FFASR Leaderboard: Benchmarking ASR in the Real World
The FFASR Leaderboard was launched to evaluate speech recognition systems in real-world conditions. It provides a benchmark for assessing the performance of automatic speech recognition models across diverse environments and use cases.
NVIDIA NeMo AutoModel Speeds Up Transformer Fine-Tuning
NVIDIA's NeMo AutoModel enables faster fine-tuning of transformer models by automating model selection and optimization. It reduces development time and improves efficiency in training large language models on NVIDIA hardware.
Shipping huggingface_hub weekly with AI, open tools, and human oversight
Hugging Face is releasing huggingface_hub weekly, integrating AI models, open-source tools, and a human review process to ensure quality and safety. The update emphasizes transparency, community involvement, and responsible AI development through continuous human-in-the-loop validation.
PP-OCRv6 Released on Hugging Face with 50-Language Support
PP-OCRv6, a new optical character recognition model, is now available on Hugging Face. It supports 50 languages and scales from 1.5 million to 34.5 million parameters, offering improved accuracy and efficiency across diverse languages.
Can You Beat LoRA in Fine-Tuning?
A new study explores alternatives to LoRA, the most popular fine-tuning technique, assessing whether other methods can achieve better performance with less computational cost. The research finds that while some approaches show promise, none consistently outperform LoRA across diverse tasks and datasets.
Testing Cross-Origin Storage API in Transformers.js
Transformers.js has begun experimenting with the proposed Cross-Origin Storage API. The initiative aims to enable secure, cross-origin data sharing in web applications without requiring user interaction or explicit permissions.
We got local models to triage the OpenClaw repo for FREE
OpenClaw has launched a free initiative to use local AI models for triaging its repository. This allows community contributors to efficiently manage issues and pull requests without relying on external services. The effort aims to improve transparency and accessibility in open-source project maintenance.
MosaicLeaks: Can your research agent keep a secret?
MosaicLeaks has released a report questioning whether research agents can reliably maintain confidentiality. The report highlights concerns about data exposure and trust in AI-driven research tools. It calls for stronger privacy safeguards and transparency in how such agents handle sensitive information.