> ## Documentation Index > Fetch the complete documentation index at: https://starforge.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # sf submit > 提交训练作业的完整参数参考 ```bash theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}} sf submit [选项] ``` 提交前自动校验 config 与超参；工作流讲解见[提交训练](/zh-Hans/guides/submit)。 ## 资源 | 选项 | 说明 | | --------------------------- | -------------------------------------------------------------------------------------------------------------------- | | `--profile [ROLE=]名称[:总卡数]` | 目标硬件与资源，一个参数说清：`h200`（注册表默认形状）、`h200:4`、`h200:16`（2 满节点）。可重复以按角色分池：`--profile train=h200:8 --profile rollout=h100:2` | 项目名来自仓库根的 `starforge.yaml`，提交时自动带上，不用再传 `--project`。 ## 方法与超参 | 选项 | 说明 | | --------------------- | ------------------------------ | | `-m, --method` | 方法标识；不传则读实验 `recipe.lock.json` | | `-s, --set KEY=VALUE` | 覆盖超参，可重复；本地按方法声明校验类型与区间 | | `--upgrade-recipe` | 提交前把实验锁升级到当前 catalog | | `--framework-version` | 配合 `--upgrade-recipe` 切框架版本 | ## 模型与数据 | 选项 | 说明 | | ---------------------------------------------- | -------------------------------------------------------- | | `--model` | 基座模型路径或 Hub id；**verl / TRL 必填** | | `--train-data` / `--validation-data` | 数据路径；verl / TRL 必填。引用平台数据集时写数据集内相对文件名 | | `--train-dataset` / `--validation-dataset` | 平台数据集引用 `/[@version]`；推荐写在 config，CLI 仅临时覆盖 | | `--init-from run//checkpoint[@step=N]` | 从上一阶段产物起训（流水线） | ## 框架与观测（custom 专属） | 选项 | 说明 | | --------------------- | ---------------------------------------------------- | | `--image` | 本次作业覆盖 OCI 镜像；一等框架可选、custom 必填。tag 可用，生产建议 digest | | `--observability-url` | 仅 external observability recipe；platform recipe 禁止设置 | ## 流程控制 | 选项 | 说明 | | --------------- | -------------------------------- | | `--then ACTION` | 训练成功后自动执行（`export` / `eval`），可重复 | | `--allow-dirty` | 允许工作区有未提交改动（默认拒绝，保证可追溯） | | `--no-validate` | 跳过提交前校验 | ## 示例 ```bash theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}} # NeMo-RL GRPO，8 卡 sf submit my-grpo --profile h200:8 # verl 0.9 + 平台数据集 + 临时超参 sf submit my-verl --profile h200:4 \ --model Qwen/Qwen3.5-9B \ --train-dataset alice/gsm8k-zh@v2 --train-data train.parquet \ -s actor_rollout_ref.actor.optim.lr=1e-6 # 异构多池 + 训完自动导出 sf submit my-grpo --profile train=h200:8 --profile rollout=h100:2 --then export # custom sf submit my-custom --profile h100:1 \ --image registry.example.com/train@sha256:abc… # 当前 catalog 默认是外部观测时还要加 --observability-url # 曲线怎么打进控制台：见文档「自定义镜像」 ``` ## `sf submit` Submit a training job ```bash theme={"theme":{"light":"github-light","dark":"github-dark-dimmed"}} sf submit [options] ``` | 参数 | 说明 | | ----- | -------------------------------- | | `EXP` | Experiment name or path (**必填**) | | 选项 | 说明 | | ---------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `--profile` | Target hardware and resources in one argument: h200 (the registry default shape), h200:4 (4 GPUs), h200:16 (2 full nodes). Repeat it to split pools per role (the heterogeneous extension point): --profile train=h200:8 --profile rollout=h100:2. Comma-separate roles that share a pool: --profile actor,rollout=h200:8 --profile teacher=h200:8 (可重复) | | `--method`, `-m` | Method identifier \/\ (e.g. nemo-rl/grpo, verl/grpo); read from the experiment directory's recipe.lock.json when omitted | | `--set`, `-s` | Override a hyperparameter, repeatable. Types and ranges are checked locally against the method declaration, so a typo fails immediately (可重复) | | `--init-from` | Start from the previous stage's artifact (the SFT -> DPO -> GRPO pipeline) | | `--model` | Base model path or Hub id; required for verl/TRL | | `--train-data` | Training data path; required for verl/TRL. With a platform dataset declared (in the config or via --train-dataset), write the file name relative to the dataset (e.g. train.parquet) and the job side resolves it into the cache directory | | `--validation-data` | Validation data path; required for verl/TRL, used like --train-data | | `--train-dataset` | Platform dataset reference: pulled into the shared cache when the job starts, with \\_DATA\_DIR injected. Prefer declaring it as data.train.dataset in the experiment config; this flag is only a temporary override | | `--validation-dataset` | Platform dataset reference for the validation set; matches data.validation.dataset in the config | | `--output-dataset` | Publish what this run produces as this dataset version. The job writes files into FORGE\_DATASET\_OUT\_DIR and the platform publishes them when the run succeeds -- the job is never given a credential for the registry, and the published dataset inherits the classification of the material the run read. Used by a preprocessing run over a corpus | | `--environment` | Agent RL environment reference. Delivered like a corpus and for a stronger reason: a private environment is non-exportable, so the platform resolves the version into the job runtime and mounts it read-only. Run `sf env ls` to see the environments you can reference | | `--image` | Job image (a tag is enough, e.g. registry/nemo-rl-ext:v0.7.0). Required for custom; for other frameworks it overrides the default image the console configured for them | | `--then` | Run automatically after training succeeds (export/eval), repeatable (可重复) | | `--observability-url` | For an external observability recipe only; a platform recipe must not set it | | `--framework-version` | Only together with --upgrade-recipe; on its own it does not rewrite the lock file | | `--upgrade-recipe` | Upgrade this experiment's lock to the current catalog before submitting, reusing sf recipe upgrade | | `--allow-dirty` | Allow uncommitted changes in the working tree (refused by default, to keep submissions traceable) | | `--no-validate` | Skip the pre-submission validation |