AI 导读
🧵 4. Nori 的 fit 调用将 X_train 和 y_train 作为上下文存储，而非训练新权重，然后 predict 用 X_test 读取该上下文，在一次前向传播中为每个新行返回 1 个值。 他们平台上有扎实的文档 https://t.co/vn3nCybi7v https://t.co/d0NbHr5fzJ
正文
🧵 4. Nori’s fit call stores X_train and y_train as context instead of training new weights, then predict reads that context with X_test and returns 1 value per new row in a single forward pass.
Solid documentation on their platform
https://t.co/vn3nCybi7v https://t.co/d0NbHr5fzJ
来源：@rohanpaul_ai · x.com