V1培訓照片的解析度為1024p，使用H200 SXM訓練。

此LORA可以用於圖片編輯、文生圖。

V1基於”Flux.2 Klein 9B-base”訓練，可在base和蒸餾版本上使用。

在Flux.2 Klein 9B-base上使用時，建議使用” Turbo lora”，4步驟即可生成圖片。

生成全身視角解析度需大於1536，這樣臉部才會清晰。

動漫轉寫實時，不要將這個lora用於具有巨大頭部的動漫人物，最好用於頭和身體比例正常的動漫人物。

The V1 training images have a resolution of 1024p and were trained using H200 SXM.

This LoRA can be used for image editing and text-to-image generation.

V1 is trained based on “Flux.2 Klein 9B-base” and can be used with both the base and distilled versions.

When using it with Flux.2 Klein 9B-base, it is recommended to use the “Turbo lora,” as images can be generated in just 4 steps.

For full-body shots, the resolution should be greater than 1536 to ensure facial clarity.

When converting anime to realistic styles, do not use this LoRA on anime characters with oversized heads; it works best on characters with normal head-to-body proportions.

觸發詞(Trigger words):

a pornmaster zimage woman,

--

無須觸發詞也能改變人物的臉型。

You can change a character's face without needing a trigger word.

--

動漫轉寫實(Anime to Realism):

The photo was processed into a high-quality image with a realistic style. a realistic cosplay photo. a pornmaster zimage woman,