feat: add image/audio finetune

Author: gujieye

packages/cli/tests/e2e/dataset.e2e.test.ts

@@ -216,7 +216,7 @@ describe.skipIf(!isDashScopeE2EReady())("e2e: dataset (offline)", () => {
   });
 
   test("dataset upload --schema image --no-validate --dry-run 采用 1GB 媒体上限", async () => {
-    // image/video schemas raise the upload cap to 1 GiB (vs 300 MB for text).
+    // image schema raises the upload cap to 1 GiB (vs 300 MB for text).
     // --no-validate keeps this offline (the jsonl fixture is not a real zip).
     const file = join(__dirname, ".dataset-valid.jsonl");
     const { stdout, stderr, exitCode } = await runCli([
@@ -238,28 +238,59 @@ describe.skipIf(!isDashScopeE2EReady())("e2e: dataset (offline)", () => {
     expect(data.max_bytes).toBe(1024 * 1024 * 1024);
   });
 
-  test.each(["tts", "image", "video"])(
-    "dataset upload --dry-run 接受媒体 schema %s",
-    async (schema) => {
-      const file = join(__dirname, ".dataset-valid.jsonl");
-      const { stdout, stderr, exitCode } = await runCli([
-        "dataset",
-        "upload",
-        "--file",
-        file,
-        "--schema",
-        schema,
-        "--no-validate",
-        "--dry-run",
-        "--output",
-        "json",
-      ]);
-      expect(exitCode, stderr).toBe(0);
-      const data = parseStdoutJson<{ action: string; schema: string }>(stdout);
-      expect(data.action).toBe("dataset.upload");
-      expect(data.schema).toBe(schema);
-    },
-  );
+  test.each(["tts", "image"])("dataset upload --dry-run 接受媒体 schema %s", async (schema) => {
+    const file = join(__dirname, ".dataset-valid.jsonl");
+    const { stdout, stderr, exitCode } = await runCli([
+      "dataset",
+      "upload",
+      "--file",
+      file,
+      "--schema",
+      schema,
+      "--no-validate",
+      "--dry-run",
+      "--output",
+      "json",
+    ]);
+    expect(exitCode, stderr).toBe(0);
+    const data = parseStdoutJson<{ action: string; schema: string }>(stdout);
+    expect(data.action).toBe("dataset.upload");
+    expect(data.schema).toBe(schema);
+  });
+
+  test("dataset validate --schema video 拒绝（视频生成入口已隐藏）", async () => {
+    const file = join(__dirname, ".dataset-valid.jsonl");
+    const { stdout, stderr, exitCode } = await runCli([
+      "dataset",
+      "validate",
+      "--file",
+      file,
+      "--schema",
+      "video",
+      "--output",
+      "json",
+    ]);
+    expect(exitCode, stdout + stderr).not.toBe(0);
+    expect(`${stdout}\n${stderr}`).toMatch(/--schema video is not supported/);
+  });
+
+  test("dataset upload --schema video 拒绝（视频生成入口已隐藏）", async () => {
+    const file = join(__dirname, ".dataset-valid.jsonl");
+    const { stdout, stderr, exitCode } = await runCli([
+      "dataset",
+      "upload",
+      "--file",
+      file,
+      "--schema",
+      "video",
+      "--no-validate",
+      "--dry-run",
+      "--output",
+      "json",
+    ]);
+    expect(exitCode, stdout + stderr).not.toBe(0);
+    expect(`${stdout}\n${stderr}`).toMatch(/--schema video is not supported/);
+  });
 });
 
 describe.skipIf(!isDashScopeE2EReady())("e2e: dataset (DashScope)", () => {

packages/cli/tests/e2e/deploy.e2e.test.ts

@@ -56,69 +56,6 @@ describe.skipIf(!isDashScopeE2EReady())("e2e: deploy (offline)", () => {
     expect(data.body.capacity).toBe(1);
   });
 
-  test("deploy create --dry-run 组装视频 LoRA 的 aigc_config", async () => {
-    const { stdout, stderr, exitCode } = await runCli([
-      "deploy",
-      "create",
-      "--model",
-      "wan2.5-i2v-preview-ft-xxx",
-      "--name",
-      "my-video-lora",
-      "--aigc-prompt",
-      "a cat surfing",
-      "--aigc-lora-prompt-default",
-      "trigger-word",
-      "--aigc-use-input-prompt",
-      "true",
-      "--dry-run",
-      "--output",
-      "json",
-    ]);
-    expect(exitCode, stderr).toBe(0);
-    const data = parseStdoutJson<{
-      action: string;
-      body: {
-        plan: string;
-        aigc_config?: {
-          use_input_prompt?: boolean;
-          prompt?: string;
-          lora_prompt_default?: string;
-        };
-      };
-    }>(stdout);
-    expect(data.action).toBe("deploy.create");
-    expect(data.body.plan).toBe("lora");
-    expect(data.body.aigc_config).toEqual({
-      use_input_prompt: true,
-      prompt: "a cat surfing",
-      lora_prompt_default: "trigger-word",
-    });
-  });
-
-  test("deploy create --aigc-* 仅对 plan=lora 有效（plan=ptu 时报错）", async () => {
-    const { stdout, stderr, exitCode } = await runCli([
-      "deploy",
-      "create",
-      "--model",
-      "wan2.5-i2v-preview-ft-xxx",
-      "--name",
-      "my-video-lora",
-      "--plan",
-      "ptu",
-      "--input-tpm",
-      "10000",
-      "--output-tpm",
-      "1000",
-      "--aigc-prompt",
-      "a cat surfing",
-      "--dry-run",
-      "--output",
-      "json",
-    ]);
-    expect(exitCode, stdout + stderr).not.toBe(0);
-    expect(`${stdout}\n${stderr}`).toMatch(/--aigc-\* flags are only valid for plan=lora/);
-  });
-
   test("deploy create --plan mu --deploy-spec --dry-run 透传 deploy_spec", async () => {
     const { stdout, stderr, exitCode } = await runCli([
       "deploy",

packages/commands/src/commands/dataset/upload.ts

@@ -18,7 +18,7 @@ const UPLOAD_FLAGS = {
   file: {
     type: "string",
     valueHint: "<path>",
-    description: "Local dataset file (.jsonl or .zip; ≤300MB text, ≤1GB image/video)",
+    description: "Local dataset file (.jsonl or .zip; ≤300MB text, ≤1GB image)",
     required: true,
   },
   purpose: {
@@ -30,7 +30,7 @@ const UPLOAD_FLAGS = {
     type: "string",
     valueHint: "<s>",
     description:
-      'Record schema: "chatml" (SFT), "dpo" (chosen/rejected), "cpt" (raw text), "tts" (audio), "image" (image generation), or "video" (video generation). Default auto-detects per record.',
+      'Record schema: "chatml" (SFT), "dpo" (chosen/rejected), "cpt" (raw text), "tts" (audio), or "image" (image generation). Default auto-detects per record.',
   },
   noValidate: {
     type: "switch",
@@ -46,7 +46,7 @@ export default defineCommand({
   description: "Upload a dataset file (.jsonl or .zip) to Bailian",
   auth: "apiKey",
   usageArgs:
-    "--file <path> [--purpose <name>] [--schema <chatml|dpo|cpt|tts|image|video>] [--no-validate] [--full-validate]",
+    "--file <path> [--purpose <name>] [--schema <chatml|dpo|cpt|tts|image>] [--no-validate] [--full-validate]",
   flags: UPLOAD_FLAGS,
   exampleArgs: [
     "--file train.jsonl",
@@ -58,27 +58,32 @@ export default defineCommand({
     "--file train.jsonl --no-validate",
   ],
   notes: [
-    "Supports .jsonl (text) and .zip (audio/image/video archives with a",
-    "data.jsonl manifest). Six record schemas are recognized: chatml =",
-    "{messages:[...]} (SFT); dpo = {messages:[...], chosen, rejected};",
-    'cpt = {text:"..."} (continual pre-training, raw text); tts =',
-    '{wav_fn:"train/xxx.wav", text:"..."} (audio fine-tuning); image =',
-    '{img_path:"..."} (image generation); video = {first_frame_path:"...",',
-    'video_path:"..."} (video generation). With no --schema, a record',
-    "carrying wav_fn is validated as TTS, img_path as image, video_path /",
-    "first_frame_path as video, chosen/rejected as DPO, text (no messages)",
-    "as CPT, otherwise ChatML. Upload cap: 300MB text, 1GB image/video.",
-    "Upload uses the OpenAI-compatible /compatible-mode/v1/files endpoint so",
-    "the purpose tag is persisted (the DashScope-native /api/v1/files drops it).",
+    "Supports .jsonl (text) and .zip (audio/image archives with a data.jsonl",
+    "manifest). Five record schemas are recognized: chatml = {messages:[...]}",
+    '(SFT); dpo = {messages:[...], chosen, rejected}; cpt = {text:"..."}',
+    '(continual pre-training, raw text); tts = {wav_fn:"train/xxx.wav",',
+    'text:"..."} (audio fine-tuning); image = {img_path:"..."} (image',
+    "generation). With no --schema, a record carrying wav_fn is validated as",
+    "TTS, img_path as image, chosen/rejected as DPO, text (no messages) as CPT,",
+    "otherwise ChatML. Upload cap: 300MB text, 1GB image. Upload uses the",
+    "OpenAI-compatible /compatible-mode/v1/files endpoint so the purpose tag is",
+    "persisted (the DashScope-native /api/v1/files drops it).",
   ],
   async run(ctx) {
     const { identity, settings, flags } = ctx;
     const filePath = flags.file;
     const purpose = flags.purpose || "fine-tune";
     const schema = parseDatasetSchemaFlag(flags.schema);
+    if (schema === "video") {
+      throw new BailianError(
+        `--schema video is not supported.`,
+        ExitCode.USAGE,
+        `Supported schemas: chatml, dpo, cpt, tts, image.`,
+      );
+    }
     const format = detectOutputFormat(settings.output);
-    // Image / video schemas allow larger ZIPs (1 GB vs 300 MB for text).
-    const isMediaSchema = schema === "image" || schema === "video";
+    // Image schema allows larger ZIPs (1 GB vs 300 MB for text).
+    const isMediaSchema = schema === "image";
 
     if (!flags.noValidate) {
       const maxBytes = isMediaSchema ? MAX_MEDIA_ZIP_BYTES : MAX_DATASET_BYTES;

packages/commands/src/commands/dataset/validate.ts

@@ -36,15 +36,15 @@ const VALIDATE_FLAGS = {
     type: "string",
     valueHint: "<s>",
     description:
-      'Record schema: "chatml" (SFT), "dpo" (chosen/rejected), "cpt" (raw text), "tts" (audio), "image" (image generation), or "video" (video generation). Default auto-detects per record.',
+      'Record schema: "chatml" (SFT), "dpo" (chosen/rejected), "cpt" (raw text), "tts" (audio), or "image" (image generation). Default auto-detects per record.',
   },
 } satisfies FlagsDef;
 
 export default defineCommand({
   description: "Locally validate a dataset file (.jsonl or .zip) without uploading",
   // 纯本地校验，不触网、不需 API key（与 `pipeline validate` 一致）。
   auth: "none",
-  usageArgs: "--file <path> [--full-validate] [--schema <chatml|dpo|cpt|tts|image|video>]",
+  usageArgs: "--file <path> [--full-validate] [--schema <chatml|dpo|cpt|tts|image>]",
   flags: VALIDATE_FLAGS,
   exampleArgs: [
     "--file train.jsonl",
@@ -60,19 +60,25 @@ export default defineCommand({
     "Schemas: chatml = {messages:[...]} (SFT); dpo = {messages:[...], chosen,",
     'rejected}; cpt = {text:"..."} (continual pre-training, raw text);',
     'tts = {wav_fn:"train/xxx.wav", text:"..."} (audio fine-tuning);',
-    'image = {img_path:"..."} (image generation); video =',
-    '{first_frame_path:"...", video_path:"..."} (video generation). With no',
-    "--schema, a record carrying wav_fn is validated as TTS, img_path as",
-    "image, video_path / first_frame_path as video, chosen/rejected as DPO,",
-    "text (no messages) as CPT, otherwise ChatML. Pass --schema to require a",
-    "specific shape on every record. ZIP archives (.zip) are validated",
-    "structurally (data.jsonl present, media references resolve) in addition",
-    "to per-record content checks. Use --full-validate to JSON.parse every line.",
+    'image = {img_path:"..."} (image generation). With no --schema, a record',
+    "carrying wav_fn is validated as TTS, img_path as image, chosen/rejected",
+    "as DPO, text (no messages) as CPT, otherwise ChatML. Pass --schema to",
+    "require a specific shape on every record. ZIP archives (.zip) are",
+    "validated structurally (data.jsonl present, media references resolve) in",
+    "addition to per-record content checks. Use --full-validate to JSON.parse",
+    "every line.",
   ],
   async run(ctx) {
     const { settings, flags } = ctx;
     const filePath = flags.file;
     const schema = parseDatasetSchemaFlag(flags.schema);
+    if (schema === "video") {
+      throw new BailianError(
+        `--schema video is not supported.`,
+        ExitCode.USAGE,
+        `Supported schemas: chatml, dpo, cpt, tts, image.`,
+      );
+    }
     const format = detectOutputFormat(settings.output);
 
     if (settings.dryRun) {

packages/commands/src/commands/deploy/constants.ts

@@ -0,0 +1,41 @@
+/**
+ * Deploy-domain constants — billing plans, billing methods and template
+ * charge types. Centralised here so no `deploy` command carries a magic
+ * string for these server-contract values.
+ */
+
+/** Billing plan (`--plan` value, matches the server's deployment plan). */
+export const DEPLOY_PLAN = {
+  /** Token-billed; the CLI default. */
+  LORA: "lora",
+  /** Token-billed, provisioned throughput. */
+  PTU: "ptu",
+  /** Model-unit-billed. */
+  MU: "mu",
+} as const;
+
+export type DeployPlan = (typeof DEPLOY_PLAN)[keyof typeof DEPLOY_PLAN];
+
+/** CLI default plan when `--plan` is omitted. */
+export const DEFAULT_DEPLOY_PLAN: DeployPlan = DEPLOY_PLAN.LORA;
+
+/** Billing method (`billing_method`, plan=mu only). */
+export const BILLING_METHOD = {
+  /** Post-paid (currently the only server-supported value). */
+  POST_PAY: "POST_PAY",
+  /** Pre-paid. */
+  PRE_PAY: "PRE_PAY",
+} as const;
+
+export type BillingMethod = (typeof BILLING_METHOD)[keyof typeof BILLING_METHOD];
+
+/** Default billing method for plan=mu when `--billing-method` is omitted. */
+export const DEFAULT_BILLING_METHOD: BillingMethod = BILLING_METHOD.POST_PAY;
+
+/** Template charge type (`charge_type`) returned by the deployable-models catalog. */
+export const CHARGE_TYPE = {
+  POST_PAID: "post_paid",
+  PRE_PAID: "pre_paid",
+} as const;
+
+export type ChargeType = (typeof CHARGE_TYPE)[keyof typeof CHARGE_TYPE];