Tag
6 articles
Liquid AI introduces LFM2.5-DSpark draft models that accelerate decoding by up to 3.18x without altering model outputs, using speculative decoding techniques.
Liquid AI has released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, which maintain fast performance at 8K context on CPU hardware.
Learn how Antidoom, a new open-source AI method, helps prevent AI models from getting stuck in repetitive loops that waste time and reduce efficiency.
Liquid AI introduces LFM2.5-Embedding-350M and LFM2.5-ColBERT-350M, two new models for fast multilingual search on edge devices.
This article explains vision-language models and how Liquid AI's new LFM2.5-VL-450M model brings powerful AI capabilities to edge devices.
Liquid AI has released LocalCowork, an open-source desktop agent application powered by the LFM2-24B-A2B model, enabling privacy-first AI workflows to run entirely on-device.
