Open heamina opened 2 years ago
Errors will also be printed into ckpt/
You go into the "opt" folder and run the command replacing
with some name you want the training results bo be saved by. is for most of us with a single gpu a 0 and must be a path from the "opt" folder to where you saved the downloaded example folder of the lego scene or another scene. has to be replaced with a path to the config file matching the type of scene you want to train. If it is a synthetic dataset use configs/syn.json. After you run the command training will start. But the command will exit in the console. To follow the informations live use this command: tail -f ckpt/
/log It will print everything to console. I hope this helps. To get informations about other args that can be added to the command you can take a look in this file: https://github.com/sxyu/svox2/blob/master/opt/opt.py
Hope that helps.
You go into the "opt" folder and run the command replacing
with some name you want the training results bo be saved by. is for most of us with a single gpu a 0 and must be a path from the "opt" folder to where you saved the downloaded example folder of the lego scene or another scene. has to be replaced with a path to the config file matching the type of scene you want to train. If it is a synthetic dataset use configs/syn.json. After you run the command training will start. But the command will exit in the console. To follow the informations live use this command: tail -f ckpt/
/log It will print everything to console. I hope this helps. To get informations about other args that can be added to the command you can take a look in this file: https://github.com/sxyu/svox2/blob/master/opt/opt.py
Hope that helps.
It's very kind of you, think you for helping me. In fact, my head was in whilr because this trouble. I will refer you advice to copy with this, and hope we can interact afterwards.
You go into the "opt" folder and run the command replacing
with some name you want the training results bo be saved by. is for most of us with a single gpu a 0 and must be a path from the "opt" folder to where you saved the downloaded example folder of the lego scene or another scene. has to be replaced with a path to the config file matching the type of scene you want to train. If it is a synthetic dataset use configs/syn.json. After you run the command training will start. But the command will exit in the console. To follow the informations live use this command: tail -f ckpt/
/log It will print everything to console. I hope this helps. To get informations about other args that can be added to the command you can take a look in this file: https://github.com/sxyu/svox2/blob/master/opt/opt.py
Hope that helps.
I ran this instruction "python .\opt.py .\nerf_synthetic\ship\ -t ckpt/res", but there is a error "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' ", can you tell me how to do then? Thanks in advance.
hi, did you solve the problem?
I ran this instruction "python .\opt.py .\nerf_synthetic\ship\ -t ckpt/res", but there is a error "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' ", can you tell me how to do then? Thanks in advance.
I'm getting this exact same error when running the launch.sh. AttributeError: 'NoneType' object has no attribute 'dict'
Here's additional info from log trace, pointing to line 1159 of svox2.py
Selecting random rays
Eval step
^M 0%| | 0/6 [00:00<?, ?it/s]^M 0%| | 0/6 [00:00<?, ?it/s]
Traceback (most recent call last):
File "opt.py", line 471, in
Any thoughts why dict might be empty?
You go into the "opt" folder and run the command replacing with some name you want the training results bo be saved by. is for most of us with a single gpu a 0 and must be a path from the "opt" folder to where you saved the downloaded example folder of the lego scene or another scene.