Alibaba

38 summarised stories about Alibaba, each linking back to the original source. Browse all topics →

Sunday, 26 January 2025

Qwen2.5-1M: Deploy Your Own Qwen with Context Length up to 1M Tokens

Qwen 1 year ago

Alibaba released open-source Qwen2.5-1M models, including 7B and 14B parameter versions that support context lengths up to 1 million tokens. The inference framework achieves 3.2x to 6.7x faster processing speeds on 1M-token sequences compared to baseline approaches, with the 14B model matching GPT-4o-mini performance on short texts while supporting eight times longer context. Developers can now deploy these models locally using the open-sourced vLLM-based framework, which requires 120GB VRAM for the 7B model and 320GB for the 14B model.

AI Model Releases Alibaba Open Source

Qwen2.5 VL! Qwen2.5 VL! Qwen2.5 VL!

Qwen 1 year ago

Alibaba released Qwen2.5-VL, a vision-language model available in 3B, 7B, and 72B parameter sizes that can recognize objects, analyze documents and charts, process videos over 1 hour long, and function as a visual agent. The 7B model outperforms GPT-4o-mini on multiple tasks while the 72B model achieves competitive performance with much larger models like Llama-3-405B-Instruct. The model improves efficiency through a redesigned visual encoder with window attention and adds capabilities like structured output generation for financial documents and second-level event localization in videos.

AI Agents AI Model Releases Alibaba Vision-Language Models