SensenGao

> I have tried to train on the NeRF representation, but the generator only returns white rendered images. > >   > > Can someone please help me? Are...

> @SensenGao Did you check this? #95 I have already tried replacing it with "ZeroCool94/stable-diffusion-v1-5", but once it reaches `pipe = model_cls.from_pretrained(ckpt_or_pretrained, torch_dtype=torch_dtype, **kwargs)`, i.e., `StableDiffusionControlNetPipeline.from_pretrained`, it gets stuck. There...

> @SensenGao try with base_model = "benjamin-paine/stable-diffusion-v1-5" Thank you for your reply. I will have a try.

原版代码好像有部分错误 ` sp_text = reqf.text # 找位置 # idx = sp_text.find('formDigestValue') # if idx != -1: # print(sp_text[idx-100:idx+3000]) # else: # print('没有找到 _spPageContextInfo') def extract_field(key, src): # 支持带引号的键名 pt =...

> 没测过几个 大概2 3个就有效果了吧epoch 学习率2e-5  > […](#) 可以问一下1w个视频使用lora训练2-3个epoch，大概需要多久呢

> 我用的64卡 几个小时就可以了 分辨率480*720 > […](#) 感谢回复！您那边用的是14B模型吗，我想评估一下我使用8卡 1.3B大概要多久

> 14B模型的 > […](#) 非常感谢🙏