RefalMachine/RuadaptQwen2.5-14B-Instruct
TEXT GENERATIONConcurrency Cost:1Model Size:14.8BQuant:FP8Ctx Length:32kPublished:Feb 3, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

RefalMachine/RuadaptQwen2.5-14B-Instruct is a 14.8 billion parameter instruction-tuned language model, adapted from Qwen2.5-14B for enhanced Russian language performance. It features a replaced tokenizer and continued pretraining on a Russian corpus, followed by Learned Embedding Propagation (LEP). This model is optimized for generating Russian texts, achieving up to 60% faster generation speed compared to the original Qwen-2.5-14B-Instruct for Russian content.

Loading preview...