tangzj / able-nerf Goto Github PK
View Code? Open in Web Editor NEWThis is the official repo for ABLE-NeRF, an attention based/ transformer novel view synthesis work.
Home Page: https://arxiv.org/abs/2303.13817
This is the official repo for ABLE-NeRF, an attention based/ transformer novel view synthesis work.
Home Page: https://arxiv.org/abs/2303.13817
May I ask how to train the Shiny Blender dataset? I can't find '/data/Shiny Blender/all/transforms_val. json' when running the code
when i was running the code,in the validation step,it raise this error:
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 1008, in forward
output = self._run_ddp_forward(*inputs, **kwargs)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/parallel/distributed.py", line 969, in _run_ddp_forward
return module_to_run(*inputs[0], **kwargs[0])
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/pytorch_lightning/overrides/base.py", line 110, in forward
return self._forward_module.validation_step(*inputs, **kwargs)
File "/data/niuzy/Python/able-nerf/ablenerf_litsystem.py", line 87, in validation_step
coarse_rgb, fine_rgb = self.render_image(batch)
File "/data/niuzy/Python/able-nerf/ablenerf_litsystem.py", line 118, in render_image
output_coarse, output_fine, weights = self(batch_rays)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/Python/able-nerf/ablenerf_litsystem.py", line 28, in forward
output = self.model(batch_rays)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/Python/able-nerf/models/able_nerf.py", line 183, in forward
ray_tokens, attn_weights = self.coarse(ray_tokens, attn_masks)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/Python/able-nerf/models/transformers.py", line 85, in forward
tokens, weights = layer(tokens, attn_masks)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/Python/able-nerf/models/transformers.py", line 59, in forward
attn_out, weights = self.attn(x,x,x, attn_mask=attn_masks)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/data/niuzy/anaconda3/envs/monosdf/lib/python3.8/site-packages/torch/nn/modules/activation.py", line 1113, in forward
return torch._native_multi_head_attention(
RuntimeError: Mask shape should match input
Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.
Epoch 0: 100%|████████████████████████████▉| 31250/31258 [3:11:00<00:02, 2.73it/s, loss=0.00284, v_num=0, train/psnr_fine=30.40, train/psnr_coarse=28.60]
I use the nerf_sththetic dataset,using the lego scene,and the default yaml configuration,I check the shape into the attn module,but It seems the same as the train_step.
in the validation_step and the training step,both the shape of tokens,attn_masks are the same:
tokens, weights = layer(tokens, attn_masks) #tokens:[1024,97,192],attn_masks:[3072,97,97]
so is there anything wrong?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.