Seg-Zero exhibits emergent test-time reasoning ability. It generates a reasoning chain before producing the final segmentation mask. Seg-Zero is trained exclusively using reinforcement learning, ...
Introducing MMR1-Math-v0, a Large Multimodal Model specialized in mathematical tasks. Remarkably, MMR1-Math-v0 achieves state-of-the-art performance among open-source 7B multimodal models, competing ...
近期阿里通义实验室在 Hugging Face 和 ModelScope 上开源了 Qwen2.5-VL 的 Base 和 Instruct 模型，包含 3B、7B 和 72B 在内的 3 个模型尺寸。其中，Qwen2.5-VL-7B ...
有消息称，阿里云未来还将发布基于Qwen2.5-Max的推理模型，其复杂任务处理能力及推理能力还将大幅提升。 全球开源生态领跑者 前三大开源模型中 2席为中国公司 自2023年开源以来，阿里千问 ...
The TikTok logo is displayed on a smartphone with owner ByteDance's name in the background. Investors are keen on ByteDance's AI potential. ByteDance cofounder Zhang Yiming has become China’s ...
“Manus didn’t just slap an API on a model. They built an autonomous system that can execute deep research, deep thinking, and multi-step tasks in a way that no other AI have.” by Supreeth Koundinya ...
QwQ-32B是阿里Qwen团队最新发布的推理模型，基于Qwen2.5-32B+强化学习构建。 据官方公示的基准评测结果，在测试数学能力的 AIME24 评测集上，以及评估 ...
IT之家3 月 6 日消息，AMD 今日宣布，为 Radeon RX 9070 系列开源 Linux 驱动程序，还宣布开源 Instella —— 一个完全开源的 3B 参数语言模型。 AMD Instella 代表“完全开源的尖端 30 亿参数语言模型（LMs ...
仅需 12 台 H800 上 6 小时即可训练完成，从没有长思维链的 Qwen2.5-32B-Instruct 出发，仅使用 7 万条数学数据训练，得到 Light-R1-32B，在 AIME24 测试基准中 ...
IT之家3 月 6 日消息，研究表明，强化学习可以显著提高模型的推理能力，例如 DeepSeek-R1 通过整合冷启动数据和多阶段训练，实现了最先进的性能，使其能够进行深度思考和复杂推理。 阿里云通 ...
Alibaba said its new model is accessible via its chatbot service, Qwen Chat, for which users can choose various Qwen models including Qwen2.5-Max, the most powerful language model in the Qwen series.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback