MiniMax H3 Image to Video Audio LoRA Trainer
整合
模型 id
minimax/h3/i2v/trainer
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://vit.doitong.com/api/v1/run/minimax/h3/i2v/trainer?wait=20" \
-H "Authorization: Bearer $DOITONG_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"training_data_url": "https://example.com/your-file.zip"
}'
# Still "queued" or "running"? Poll with the id from the response:
curl "https://vit.doitong.com/api/v1/requests/REQUEST_ID?wait=20" \
-H "Authorization: Bearer $DOITONG_API_KEY"
const API = "https://vit.doitong.com/api/v1";
const headers = {
"Authorization": "Bearer " + process.env.DOITONG_API_KEY,
"Content-Type": "application/json",
};
// Run the model; wait up to 20 s for the finished result in the same call
let res = await fetch(API + "/run/minimax/h3/i2v/trainer?wait=20", {
method: "POST",
headers,
body: JSON.stringify({
"training_data_url": "https://example.com/your-file.zip"
}),
});
let request = await res.json();
if (!res.ok) throw new Error(request.error.code + ": " + request.error.message);
// Not finished yet: long-poll until it is
while (request.status === "queued" || request.status === "running") {
res = await fetch(API + "/requests/" + request.id + "?wait=20", { headers });
request = await res.json();
}
if (request.status !== "succeeded") throw new Error(JSON.stringify(request.error));
console.log(request.outputs.map((o) => o.url), request.cost.credits + " API credits ($" + request.cost.usd + ")");
import os
import requests
API = "https://vit.doitong.com/api/v1"
headers = {"Authorization": "Bearer " + os.environ["DOITONG_API_KEY"]}
payload = {
"training_data_url": "https://example.com/your-file.zip",
}
# Run the model; wait up to 20 s for the finished result in the same call
r = requests.post(API + "/run/minimax/h3/i2v/trainer", params={"wait": 20}, json=payload, headers=headers)
request = r.json()
if not r.ok:
raise RuntimeError(request["error"]["code"] + ": " + request["error"]["message"])
# Not finished yet: long-poll until it is
while request["status"] in ("queued", "running"):
request = requests.get(API + "/requests/" + request["id"], params={"wait": 20}, headers=headers).json()
if request["status"] != "succeeded":
raise RuntimeError(request["error"])
print([o["url"] for o in request["outputs"]], request["cost"]["credits"], "API credits", "$" + request["cost"]["usd"])
# 1. Connect once (Claude Code; Cursor, VS Code, Claude Desktop: https://doitong.com/zh-TW/developers/docs/mcp)
claude mcp add --transport http doitong-models https://vit.doitong.com/api/v1/mcp \
--header "Authorization: Bearer $DOITONG_API_KEY"
# 2. Your agent calls the "run" tool with:
{
"model_id": "minimax/h3/i2v/trainer",
"input": {
"training_data_url": "https://example.com/your-file.zip"
}
}
透過 Doitong API 呼叫 MiniMax H3 Image to Video Audio LoRA Trainer 模型(minimax/h3/i2v/trainer)。代理程式文件:https://doitong.com/zh-TW/developers/models/minimax/h3/i2v/trainer.md
POST https://vit.doitong.com/api/v1/run/minimax/h3/i2v/trainer?wait=20
Authorization: Bearer $DOITONG_API_KEY
Content-Type: application/json
{
"training_data_url": "https://example.com/your-file.zip"
}
如果回應狀態為 "queued" 或 "running",請輪詢 GET https://vit.doitong.com/api/v1/requests/REQUEST_ID?wait=20,直到狀態變為 "succeeded";檔案位於 outputs[].url。使用相同主體傳送 POST https://vit.doitong.com/api/v1/quote/minimax/h3/i2v/trainer,可在執行前取得確切價格。
此請求: $20.00 20,000 API 點數 在此頁面執行
您的新 API 金鑰。 僅會在此時顯示,且已填入下方的程式碼與 API 金鑰欄位。請妥善保密:切勿將其放入在瀏覽器中執行的程式碼。
建立請求
控制項是根據模型結構描述生成的。變更設定時,程式碼和價格會同步更新。
POST
/run/minimax/h3/i2v/trainer?wait=20
報價 → $20.00
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://vit.doitong.com/api/v1/run/minimax/h3/i2v/trainer?wait=20" \
-H "Authorization: Bearer $DOITONG_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"training_data_url": "https://example.com/your-file.zip"
}'
# Still "queued" or "running"? Poll with the id from the response:
curl "https://vit.doitong.com/api/v1/requests/REQUEST_ID?wait=20" \
-H "Authorization: Bearer $DOITONG_API_KEY"
const API = "https://vit.doitong.com/api/v1";
const headers = {
"Authorization": "Bearer " + process.env.DOITONG_API_KEY,
"Content-Type": "application/json",
};
// Run the model; wait up to 20 s for the finished result in the same call
let res = await fetch(API + "/run/minimax/h3/i2v/trainer?wait=20", {
method: "POST",
headers,
body: JSON.stringify({
"training_data_url": "https://example.com/your-file.zip"
}),
});
let request = await res.json();
if (!res.ok) throw new Error(request.error.code + ": " + request.error.message);
// Not finished yet: long-poll until it is
while (request.status === "queued" || request.status === "running") {
res = await fetch(API + "/requests/" + request.id + "?wait=20", { headers });
request = await res.json();
}
if (request.status !== "succeeded") throw new Error(JSON.stringify(request.error));
console.log(request.outputs.map((o) => o.url), request.cost.credits + " API credits ($" + request.cost.usd + ")");
import os
import requests
API = "https://vit.doitong.com/api/v1"
headers = {"Authorization": "Bearer " + os.environ["DOITONG_API_KEY"]}
payload = {
"training_data_url": "https://example.com/your-file.zip",
}
# Run the model; wait up to 20 s for the finished result in the same call
r = requests.post(API + "/run/minimax/h3/i2v/trainer", params={"wait": 20}, json=payload, headers=headers)
request = r.json()
if not r.ok:
raise RuntimeError(request["error"]["code"] + ": " + request["error"]["message"])
# Not finished yet: long-poll until it is
while request["status"] in ("queued", "running"):
request = requests.get(API + "/requests/" + request["id"], params={"wait": 20}, headers=headers).json()
if request["status"] != "succeeded":
raise RuntimeError(request["error"])
print([o["url"] for o in request["outputs"]], request["cost"]["credits"], "API credits", "$" + request["cost"]["usd"])
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "run",
"arguments": {
"model_id": "minimax/h3/i2v/trainer",
"input": {
"training_data_url": "https://example.com/your-file.zip"
},
"wait_seconds": 20
}
}
}
關於 MiniMax H3 Image to Video Audio LoRA Trainer
從 `training_data_url` 中的影片片段 zip 檔訓練 LoRA,使用首幀條件讓靜態影像轉為動畫影片。資料集只能包含影片。設定包括 8 至 128 的 `rank`、最多 15000 的 `number_of_steps`、22 至 124 的 `number_of_frames`、8 至 60 的 `frame_rate` 與 `trigger_phrase`。輸出 LoRA 權重與設定檔。
參數
名稱類型
預設值允許值
說明
training_data_url必填
類型string
包含 .mp4、.mov、.avi 或 .mkv 格式訓練影片的 ZIP URL;建議至少 10 個片段,且集合越大越好。圖像不能作為訓練媒體:純圖像資料集,以及混合訓練圖像和影片的封存檔都會被拒絕,但部分端點允許圖像作為條件控制側載檔。片段可附帶可選的說明檔,每個文字檔與其影片共用基本檔名。
rank選填
類型enum
預設值32
8163264128
LoRA 適配階數;設定越高,記憶體消耗越大,但容量也越高。
number_of_steps選填
類型integer
預設值2000
1 – 15,000
訓練步驟總數。
learning_rate選填
類型number
預設值0.0002
0.000001 – 1
最佳化學習率,在提高時權衡訓練可能加速與過度擬合的風險。
number_of_frames選填
類型integer
預設值73
22 – 124
每個訓練樣本的影格數;除以 17 後餘數必須為 5。範例包括 22、39、56、73、90、107 和 124。
frame_rate選填
類型integer
預設值24
8 – 60
目標影片影格速率,以每秒影格數表示。
resolution選填
類型enum
預設值medium
lowmediumhigh
訓練解析度。提高解析度會增加記憶體需求。
aspect_ratio選填
類型enum
預設值16:9
21:916:94:31:13:49:16
從 MiniMax H3 API 提供的相同集合中選擇訓練長寬比。
trigger_phrase選填
類型string
預設值
LoRA 風格啟用片語,訓練時插入說明的開頭。
first_frame_conditioning_p選填
類型number
預設值0.5
0 – 1
透過提高訓練使用第一影格條件控制的頻率,改善圖像轉影片效能;此值設定該機率。
auto_scale_input選填
類型boolean
預設值false
是否透過自動影片縮放來匹配目標 fps 和影格數;true 會啟用此調整。
split_input_into_scenes選填
類型boolean
預設值true
超過時長門檻之影片的場景分割,為 true 時啟用。
split_input_duration_threshold選填
類型number
預設值30
1 – 60
以秒數表示的場景分割觸發門檻:只有長於該門檻的影片才會分割。
debug_dataset選填
類型boolean
預設值false
預處理檢查模式,用於完整訓練前檢查影片、圖像與說明。啟用後,訓練器會將處理過的訓練資料集放入可下載的封存檔,供手動檢查。
strict_dataset選填
類型boolean
預設值false
嚴格處理預處理問題;否則這些問題會被記錄並使用文件所述的備援方式解決。啟用後,若說明需要空白提示詞、缺少側載關鍵影格,或影片音訊無法讀取,請求會被拒絕,而不會分別以空白提示詞、片段影格或靜音替代。
必填
輸出
GET /requests/req_8k2m4q7x1z9c3v5b6n0p2r4t
{
"id": "req_8k2m4q7x1z9c3v5b6n0p2r4t",
"object": "request",
"model": "minimax/h3/i2v/trainer",
"status": "succeeded",
"created_at": "2026-10-20T12:00:00Z",
"started_at": "2026-10-20T12:00:02Z",
"completed_at": "2026-10-20T12:00:09Z",
"queue_position": null,
"cost": {
"credits": 20000,
"usd": "20.00",
"charged": 20000,
"refunded": 0
},
"input": {
"training_data_url": "https://example.com/your-file.zip"
},
"output": {
"lora_file": {
"url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
"content_type": "image/jpeg",
"file_name": "0-output.jpeg",
"file_size": 1843200
},
"config_file": {
"url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
"content_type": "image/jpeg",
"file_name": "0-output.jpeg",
"file_size": 1843200
},
"debug_dataset": {
"url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
"content_type": "image/jpeg",
"file_name": "0-output.jpeg",
"file_size": 1843200
}
},
"outputs": [
{
"type": "file",
"url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
"content_type": "image/jpeg",
"file_size": 1843200,
"stored": true
}
],
"error": null,
"metrics": {
"inference_seconds": 6.8
},
"urls": {
"self": "https://vit.doitong.com/api/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t",
"cancel": "https://vit.doitong.com/api/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t/cancel"
}
}
您的檔案,由您選擇每個 outputs[].url 都是儲存空間中的副本:可永久保留(按 GB-月計費),或免費保留 7 天;可透過帳戶設定或 ?storage= 設定。
所有模型使用相同結構圖片、影片和音訊皆包含 type、url、content_type、width、height 和 duration。output 是模型自己的結果物件。
輸出欄位
- lora_file
- object {url, content_type, file_name, file_size}`.safetensors` 格式訓練 LoRA 權重的 URL。
- config_file
- object {url, content_type, file_name, file_size}推論端點設定的組態。
- debug_dataset
- object {url, content_type, file_name, file_size} | null僅適用於啟用 `debug_dataset` 的請求;此可下載封存檔包含預處理後的訓練資料。
其他 MiniMax H3 變體
更多模型:訓練
給 AI 代理程式:此頁面的 Markdown 位於 https://doitong.com/zh-TW/developers/models/minimax/h3/i2v/trainer.md. 所有模型與指南: https://doitong.com/developers/llms.txt