``` with torch.inference_mode(): seed = 3407+j prompt = data['caption'][0] idx = data['idx'][0] image = pipe( prompt, height=1024, width=1024, guidance_scale=3.5, num_inference_steps=30, max_sequence_length=512, generator=torch.Generator(device="cuda").manual_seed(seed) ).images[0] ``` 针对leaderboard上开源模型的分数有一个问题请教一下,在第一步推理生成图像的时候,开源模型全部都是基于flux_zh_multi_node_inference.py中上述推理参数生成的,还是每个开源模型各自采用了官方的原始推理参数?
Activity
CodeGoat24 commentedon Mar 14, 2026
感谢关注~
seed = 3407+j是所有开源模型都一样的。