LLMs1 min read
IBM Research has released the Granite Time Series PatchTST-FM-r2 model, a state-of-the-art model for time series data. This release includes a commercial-friendly license, enabling broader adoption and integration.
By OpenSmartRoute editorial · written through the router by writer-small
From Hugging Face blog - “IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license”
The Granite Time Series PatchTST-FM-r2 model is a new offering from IBM Research. It is designed for time series forecasting and analysis tasks. The model is available under a commercial-friendly license. This license facilitates its use in commercial applications and deployments. The model's architecture utilizes PatchTST-FM, a specific approach for time series modeling.
Source: https://huggingface.co/blog/ibm-research/ibm-releases-sota-granite-time-series
Keep reading
LLMs1 min read
Researchers introduced a sentence-trained LLM approach for cascaded simultaneous translation, utilizing prefix-aware fine-tuning to achieve flicker-free performance. This method, based on Qwen3-8B, demonstrates superior quality-latency tradeoffs compared to traditional wait-$k$ and target-suffix-deletion mechanisms.
LLMs1 min read
A new dataset, TamilEOT, and two audio-only detectors are released for semantic end-of-turn detection in Tamil telephone speech. The models achieve accuracy improvements of 13.41% on a held-out test set, with inference times under 150ms.
LLMs1 min read
Research found that document-level machine translation evaluation, presenting full documents to annotators, yields statistically equivalent scores and rankings compared to segment-level evaluations. This suggests current document-level systems and associated metrics may not be accurately measuring intended aspects of translation quality.
