Event ·Multi-source

谷歌发布TimesFM-3时间序列预测模型

谷歌推出TimesFM-3模型，拥有3.3亿参数，在超1万亿时间点数据上预训练。该模型支持零样本多变量时间序列预测，采用双注意力机制和连续补丁掩码技术。TimesFM-3可联合预测多个相关时间序列，捕捉依赖关系，提升准确率。原生支持多目标预测、过去协变量和动态协变量三种特性。

Current read

谷歌发布3.3亿参数的TimesFM-3，能零样本预测多变量时间序列，无需微调即可捕捉复杂依赖关系。

3 sources73° AI heatLast updated 2026/9/1 07:46:40

Evidence

3 sources

Primary sourceIT之家

2026/9/1 07:46:40View source

Related source 1marktechpost

2026/8/31 21:20:16View source

Related source 2Patrick Loeber

2026/9/1 15:48:30View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story