AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
翻译准备中
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证PyTorch's API that trains in a browser tab on WebGPU — autograd, nn, optim, capture/replay, int8, ONNX out; held to real torch's values on 4,764 golden cases; measured on one page against TF.js, jax-js, Burn and ONNX Runtime Web (4–12× faster training, ahead of ORT at inference)
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
PyTorch's shape, in a browser tab — no install, no server, the data never leaves the
machine, and every value held to real PyTorch's. Three implementations of one
arithmetic — a numpy core (import borch as torch), a TypeScript runtime on WebGPU
(borch-ts), and a Python binding over that runtime for Pyodide (borch_webgpu) — held
to PyTorch's values, errors and printed form within the range a curriculum uses.查看完整 README ↗目前无法加载完整 README。请通过原文链接阅读。 以下为已保存的摘录。
PyTorch's shape, in a browser tab — no install, no server, the data never leaves the
machine, and every value held to real PyTorch's. Three implementations of one
arithmetic — a numpy core (import borch as torch), a TypeScript runtime on WebGPU
(borch-ts), and a Python binding over that runtime for Pyodide (borch_webgpu) — held
to PyTorch's values, errors and printed form within the range a curriculum uses.查看完整 README ↗