wandb-autoresearch

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

W&B autoresearch

W&B 自动研究

Use W&B Launch to execute real training experiments, read their results, and choose the next bounded trial. Do not create a local

wandb.init()

run as a substitute for compute that was supposed to run through Launch.

The Launch implementation lives in

skills/wandb-primary

skills/wandb-primary/scripts/launch_helpers.py

the
```
Launch
```
section in
```
skills/wandb-primary/SKILL.md
```

Install missing packages explicitly with

uv

; do not assume a fixed working directory or preinstalled environment.

使用W&B Launch执行真实训练实验，读取实验结果，并选择下一个有边界的试验。请勿创建本地

wandb.init()

运行来替代本应通过Launch执行的计算任务。

Launch的实现位于

skills/wandb-primary

目录下：

skills/wandb-primary/scripts/launch_helpers.py

```
skills/wandb-primary/SKILL.md
```
中的
```
Launch
```
章节

使用

uv

显式安装缺失的包；请勿假设存在固定工作目录或预安装环境。

Flow

流程

Read only the reference for the current stage:

Stage	Entry condition	Read
Understand	Always first; inspect launchable code, queue capacity, data location, and prior state	`references/understand.md`
Set up	Launchable code or usable queue capacity is missing	`references/setup.md`
Search	Launchable code and queue capacity both exist	`references/search.md`

The readiness gate is

launchable code + queue with usable capacity

. A dataset may be a versioned Artifact or may be supplied by the code/image; its storage form alone does not determine readiness.

仅阅读当前阶段的参考文档：

阶段	进入条件	参考文档
理解	始终为第一个阶段；检查可启动代码、队列容量、数据位置及先前状态	`references/understand.md`
搭建	缺少可启动代码或可用队列容量	`references/setup.md`
搜索	同时存在可启动代码与队列容量	`references/search.md`

就绪条件为

可启动代码 + 具备可用容量的队列

。数据集可以是带版本的Artifact，也可以由代码/镜像提供；仅其存储形式不足以判断是否就绪。

Experiment records

实验记录

Use each W&B object for one job:

Runs hold trial config, metrics, system metrics, status, and code/job lineage.
Job or code Artifacts make code relaunchable.
Dataset Artifacts provide versioned inputs when the workload needs them.
Launch queues and agents provide compute.
A small local markdown file holds the hypothesis, trial tally, decisions, queue, and next step.

Keep local state at a user-visible path in the current project or a path the user supplied. Sync it to a W&B Artifact only when cross-session resume is useful and the user authorizes the write:

bash

S=skills/wandb-autoresearch/scripts/autoresearch_state.py
uv run --with wandb python "$S" load ENTITY PROJECT --dest .
uv run --with wandb python "$S" save ENTITY PROJECT --path AUTORESEARCH_STATE.md

每个W&B对象仅用于一项任务：

Runs存储试验配置、指标、系统指标、状态以及代码/任务谱系。
Job或代码Artifact可让代码重新启动。
当工作负载需要时，Dataset Artifact提供带版本的输入。
Launch队列与Agent提供计算资源。
一个本地小型markdown文件存储假设、试验计数、决策、队列信息及下一步操作。

将本地状态存储在当前项目中用户可见的路径或用户指定的路径下。仅当跨会话恢复有用且用户授权写入时，才将其同步至W&B Artifact：

bash

S=skills/wandb-autoresearch/scripts/autoresearch_state.py
uv run --with wandb python "$S" load ENTITY PROJECT --dest .
uv run --with wandb python "$S" save ENTITY PROJECT --path AUTORESEARCH_STATE.md

Compute safety

计算安全

Confirm the queue, metric direction, search axis, and trial/compute budget before the first non-smoke launch. Do not exceed the agreed budget.
Smoke-test the real job at a deliberately small setting before full trials.
Change only config fields the program actually reads. Use a code edit and a new job Artifact for architecture or pipeline changes.
Run trials serially unless the user explicitly authorizes parallel compute. Inspect each completed result before selecting the next trial.
Report failures, hardware utilization, and uncertainty. Re-run a likely winner with another seed when the apparent gap could be noise.
Never fabricate a run, metric, queue result, or completion state.

在首次非冒烟测试启动前，确认队列、指标方向、搜索轴以及试验/计算预算。请勿超出约定的预算。
在完整试验前，故意使用小规模设置对真实任务进行冒烟测试。
仅修改程序实际读取的配置字段。若需更改架构或流水线，请编辑代码并创建新的job Artifact。
除非用户明确授权并行计算，否则请串行执行试验。在选择下一个试验前，检查每个已完成的结果。
报告失败情况、硬件利用率及不确定性。当明显的差距可能是噪声导致时，使用另一个种子重新运行潜在的最优试验。
切勿伪造运行记录、指标、队列结果或完成状态。