曹潇

I found that if "args.dataset_type == 'llff' ", the code will select test images from the data. I'm wondering how could you make sure the depth points are the same...

Noised Image

1

Hi, I'm using text_to_image_lora, and I found the variavle "X_T" is removed which prevent us feeding the noised image to diffusion. Do you think it is better to add this...

if self.basis_type != BASIS_TYPE_MLP and imrend_fn_name in _C.__dict__ and not torch.is_grad_enabled() and not return_raylen: AttributeError: 'NoneType' object has no attribute '__dict__' The _C is always the nonetype

Error: `Image undistortion ============================================================================== I0619 00:23:26.301478 2228095 undistortion.cc:214] Undistorting image [1/20] E0619 00:23:26.376965 2228097 undistortion.cc:963] Check failed: distorted_camera.width == distorted_bitmap.Width() (4032 vs. 3972) E0619 00:23:26.390570 2228159 undistortion.cc:963] Check failed: distorted_camera.width...

Too slow fo webui

1

Why doing operation based on webui is so slow? I run it on L40s server which should be pretty fast for gaussian splatting task. Rotate the angle now takes minutes

Why two same negative prompt

1

Hi, I'm wondering the purpose of using two same negative prompt embedding. `prompt_embeds = torch.cat([prompt_embeds, negative_prompt_embeds, negative_prompt_embeds])` And you set ` noise_pred = ( noise_pred_uncond + self.guidance_scale * (noise_pred_text -...

Thanks for your interesting work. May I ask: in your work, you use dataset update strategy. Did you keep the old unedited images and sample from them randomly or just...

Missing part and confusing place

7

Hi, In training script, you use InterGen (train.py, def build_models: ...) But in eval script, you use InterClip (evaluator.py, def build_models: .... ) which contains "motion_encoder". This cause the mismatched...

How to render

1

Hi, I notice that there is only training script. Will you provide render script?