参数名 | 类型 | 默认值 | 取值范围 | 描述 |
---|---|---|---|---|
model | string | "flux-dev" | 固定为"flux-dev" | 使用的模型名称,目前仅支持"flux-dev" |
prompt | string | - | - | 用于生成图像的文本提示,是必填项。提示内容会直接影响生成图像的内容和风格,需要尽可能详细和具体地描述您想要生成的图像 |
image | string | - | - | 用于图像到图像(image to image)模式的输入图像。输出图像的宽高比将与此图像相同。图像需要以URI的形式提供,支持http/https URL或者data URL(base64编码的图像数据) |
prompt_strength | number | 0.8 | 0~1 | 在图像到图像模式下,文本提示相对于输入图像的强度。取值为1时,输入图像中的信息将被完全替换,即生成的图像与输入图像无关。取值为0时,输入图像将完全保留,文本提示不起作用。建议取值在0.6~0.9之间 |
num_outputs | integer | 1 | 1~4 | 要生成的图像数量 |
aspect_ratio | string | "1:1" | "1:1", "16:9", "21:9", "2:3", "3:2", "4:5", "5:4", "9:16", "9:21" | 生成图像的宽高比,提供了一些常见的宽高比选项,请根据您的需求选择 |
num_inference_steps | integer | 50 | 1~50 | 进行采样的推理步数。推理步数越大,图像质量越高,但生成时间也越长。推荐设置在28~50之间 |
guidance | number | 3.5 | 0~10 | 控制生成图像与文本提示的相关性和图像质量/多样性之间的平衡。在flux-dev模型中有效,在flux-schnell模型中忽略。较高的值(如7)会使输出更紧密地匹配提示,但可能会降低图像的整体质量和多样性。较低的值(如2)允许更多的创作自由度,但生成的图像可能与提示的关联度降低。建议在3~5之间取值 |
seed | integer | - | 0~2^32-1 | 随机种子。如果设置了seed,则每次生成的图像都是确定的。如果不设置seed,则每次生成的图像都是随机的。设置seed可以让您得到可复现的生成结果 |
output_format | string | "webp" | "webp", "jpg", "png" | 输出图像的格式,提供了三种常见的web友好格式,如果没有特殊需求,建议使用默认的webp格式 |
output_quality | integer | 80 | 0~100 | 输出图像的质量,仅对jpg和webp格式有效。100表示最佳质量,0表示最低质量。如果要在图像质量和文件大小之间取得平衡,建议设置在80左右 |
disable_safety_checker | boolean | true | true, false | 是否禁用内容安全检查器。默认为true,即不进行内容安全检查,以尽量减少过滤器对用户使用的影响。 |
response_format | string | "url" | "url", "b64_json" | 返回的图像数据格式,默认以url的形式返回图像地址。如果设置为"b64_json",则会在json中以base64编码的形式直接返回图像数据,节省了额外的网络请求,但response的数据量会增大 |
curl --location --request POST 'https://cn2us02.opapi.win/api/v1/ai/draw/flux/dev' \
--header 'Authorization: Bearer <token>' \
--data-urlencode 'model=flux-dev' \
--data-urlencode 'prompt=black forest gateau cake spelling out the words \"OhMyGPT\", tasty, food photography, dynamic shot' \
--data-urlencode 'response_format=b64_json'
{ "statusCode": 200, "message": "Success", "data": { "outputs": [ { "url": "https://replicate.delivery/yhqm/xXbDDptSA7oAD9UrEp8u4nkjpIUi7lvllWNzt9btzEJQvi1E/out-0.webp", "b64_json": "data:image/webp;base64,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