xingjian-bai/opal

选用指南

AI 根据 README 整理 · 2026/10/22 07:21:17

使用方式

通过 Python 库使用，将现有优化器包装在 opal.wrap 中，并传递目标句子和总步数。

查看 README 依据

The package is called `opal-optimizer`, because `opal` on PyPI is an unrelated web framework. Both install as `import opal`, so do not install both. ```python import torch, opal opt = torch.optim.AdamW(model.parameters(), lr=3e-4) # your optimizer, unchanged sched = torch.optim.lr_scheduler.CosineAnnealingLR(opt, T_max=steps) # your scheduler, unchanged opt = opal.wrap(opt, model, scheduler=sched, total_steps=steps, goal="ResNet-18 on CIFAR-10; judged by test accuracy after the last update",

部署要求

需要 Python 3.9 或更高版本，PyTorch 2.3 或更高版本，以及一个使用 torch.optim 优化器的训练脚本。对于代理运行，需要安装并登录 Claude Code。

查看 README 依据

The guide shows how to check the keep run against your plain run byte for byte, and how to use the Hugging Face Trainer: [Try it without an agent](docs/GUIDE.md#try-it-without-an-agent). ## Prerequisites - **A PyTorch training script** that uses `torch.optim` optimizers, on CPU, one GPU or several GPUs of one node (DDP). Python 3.9 or newer, PyTorch 2.3 or newer. - **For an agent run, Claude Code, logged in on the machine that runs rank 0.** Install [Claude Code](https://docs.claude.com/en/docs/claude-code), run `claude` once and log in with a Claude subscription

已知限制

不支持多节点、FSDP 和 DeepSpeed。代理可能使训练变差，且无法回滚训练。

查看 README 依据

Multi-node, FSDP and DeepSpeed are not supported. Accelerate works if you wrap after `accelerator.prepare`, without fp16 GradScaler or Accelerate's gradient accumulation; the HF Trainer works through `opal.hf.OpalTrainer`. Both are tested on CPU. Lightning should work through `opt.observe(loss)` but we have not tried it. The agent can make a run worse: it cannot restart or roll back training, it can only fix its programs at the next decision. ## Citation ```bibtex @misc{bai2026reactive, title = {Reactive Optimizers},