jw9730 / setvae Goto Github PK
View Code? Open in Web Editor NEW[CVPR'21] SetVAE: Learning Hierarchical Composition for Generative Modeling of Set-Structured Data, in PyTorch
License: MIT License
[CVPR'21] SetVAE: Learning Hierarchical Composition for Generative Modeling of Set-Structured Data, in PyTorch
License: MIT License
Initially, i received an error regarding "maximum recursion depth exceeded while calling a Python object". I then set recursion limit to the highest python allows:
import sys, threading sys.setrecursionlimit(10**7)
Following which the training exits without any informative error
[2022-08-26 06:40:38,321] [INFO] [launch.py:286:sigkill_handler] Killing subprocess 1122823 [2022-08-26 06:40:38,321] [ERROR] [launch.py:292:sigkill_handler] ['/nfshome/USERNAME/.conda/envs/setvae/bin/python', '-u', 'train.py', '--local_rank=0', '--cates', 'tooth', '--input_dim', '3', '--max_outputs', '2500', '--init_dim', '32', '--n_mixtures', '4', '--z_dim', '16', '--z_scales', '1', '1', '2', '4', '8', '16', '32', '--hidden_dim', '64', '--num_heads', '4', '--num_workers', '0', '--kl_warmup_epochs', '250', '--fixed_gmm', '--train_gmm', '--lr', '1e-3', '--beta', '1.0', '--epochs', '1000', '--dataset_type', 'shapenet15k', '--log_name', 'gen/shapenet15k/camera-ready', '--shapenet_data_dir', '/train/SetVae/ShapeNetCore.v2.PC15k', '--save_freq', '25', '--viz_freq', '1000', '--log_freq', '10', '--val_freq', '10000', '--scheduler', 'linear', '--slot_att', '--ln', '--eval', '--seed', '42', '--distributed', '--deepspeed_config', 'batch_size.json'] exits with return code = -11
Hi there,
Thank you for providing the codes.
I got this exception when running shapenet_airplane_test.sh
in the docker image:
RuntimeError: CUDA kernel failed : 209
Any help would be appreciated.
tdn = self.top_down((~x_mask).sum(-1), list(reversed(bup['features'])))
AttributeError: 'SetVAE' object has no attribute 'optimizer'
Hello, thank you for releasing the code for your SetVAE paper.
I had a question about this line:
Line 205 in 7fdf4c6
For the first decoder block, why is the residual from the hidden state ignored?
Thanks
Unexpected key(s) in state_dict: "encoder.0.att1.ln_x.weight", "encoder.0.att1.ln_x.bias", "encoder.0.att1.ln_y.weight", "encoder.0.att1.ln_y.bias", "encoder.0.att1.ln_o1.weight", "encoder.0.att1.ln_o1.bias", "encoder.0.att1.ln_o2.weight", "encoder.0.att1.ln_o2.bias", "encoder.0.att2.ln_x.weight", "encoder.0.att2.ln_x.bias", "encoder.0.att2.ln_y.weight", "encoder.0.att2.ln_y.bias", "encoder.0.att2.ln_o1.weight", "encoder.0.att2.ln_o1.bias", "encoder.0.att2.ln_o2.weight", "encoder.0.att2.ln_o2.bias", "encoder.1.att1.ln_x.weight", "encoder.1.att1.ln_x.bias", "encoder.1.att1.ln_y.weight", "encoder.1.att1.ln_y.bias", "encoder.1.att1.ln_o1.weight", "encoder.1.att1.ln_o1.bias", "encoder.1.att1.ln_o2.weight", "encoder.1.att1.ln_o2.bias", "encoder.1.att2.ln_x.weight", "encoder.1.att2.ln_x.bias", "encoder.1.att2.ln_y.weight", "encoder.1.att2.ln_y.bias", "encoder.1.att2.ln_o1.weight", "encoder.1.att2.ln_o1.bias", "encoder.1.att2.ln_o2.weight", "encoder.1.att2.ln_o2.bias", "encoder.2.att1.ln_x.weight", "encoder.2.att1.ln_x.bias", "encoder.2.att1.ln_y.weight", "encoder.2.att1.ln_y.bias", "encoder.2.att1.ln_o1.weight", "encoder.2.att1.ln_o1.bias", "encoder.2.att1.ln_o2.weight", "encoder.2.att1.ln_o2.bias", "encoder.2.att2.ln_x.weight", "encoder.2.att2.ln_x.bias", "encoder.2.att2.ln_y.weight", "encoder.2.att2.ln_y.bias", "encoder.2.att2.ln_o1.weight", "encoder.2.att2.ln_o1.bias", "encoder.2.att2.ln_o2.weight", "encoder.2.att2.ln_o2.bias", "encoder.3.att1.ln_x.weight", "encoder.3.att1.ln_x.bias", "encoder.3.att1.ln_y.weight", "encoder.3.att1.ln_y.bias", "encoder.3.att1.ln_o1.weight", "encoder.3.att1.ln_o1.bias", "encoder.3.att1.ln_o2.weight", "encoder.3.att1.ln_o2.bias", "encoder.3.att2.ln_x.weight", "encoder.3.att2.ln_x.bias", "encoder.3.att2.ln_y.weight", "encoder.3.att2.ln_y.bias", "encoder.3.att2.ln_o1.weight", "encoder.3.att2.ln_o1.bias", "encoder.3.att2.ln_o2.weight", "encoder.3.att2.ln_o2.bias", "encoder.4.att1.ln_x.weight", "encoder.4.att1.ln_x.bias", "encoder.4.att1.ln_y.weight", "encoder.4.att1.ln_y.bias", "encoder.4.att1.ln_o1.weight", "encoder.4.att1.ln_o1.bias", "encoder.4.att1.ln_o2.weight", "encoder.4.att1.ln_o2.bias", "encoder.4.att2.ln_x.weight", "encoder.4.att2.ln_x.bias", "encoder.4.att2.ln_y.weight", "encoder.4.att2.ln_y.bias", "encoder.4.att2.ln_o1.weight", "encoder.4.att2.ln_o1.bias", "encoder.4.att2.ln_o2.weight", "encoder.4.att2.ln_o2.bias", "encoder.5.att1.ln_x.weight", "encoder.5.att1.ln_x.bias", "encoder.5.att1.ln_y.weight", "encoder.5.att1.ln_y.bias", "encoder.5.att1.ln_o1.weight", "encoder.5.att1.ln_o1.bias", "encoder.5.att1.ln_o2.weight", "encoder.5.att1.ln_o2.bias", "encoder.5.att2.ln_x.weight", "encoder.5.att2.ln_x.bias", "encoder.5.att2.ln_y.weight", "encoder.5.att2.ln_y.bias", "encoder.5.att2.ln_o1.weight", "encoder.5.att2.ln_o1.bias", "encoder.5.att2.ln_o2.weight", "encoder.5.att2.ln_o2.bias", "encoder.6.att1.ln_x.weight", "encoder.6.att1.ln_x.bias", "encoder.6.att1.ln_y.weight", "encoder.6.att1.ln_y.bias", "encoder.6.att1.ln_o1.weight", "encoder.6.att1.ln_o1.bias", "encoder.6.att1.ln_o2.weight", "encoder.6.att1.ln_o2.bias", "encoder.6.att2.ln_x.weight", "encoder.6.att2.ln_x.bias", "encoder.6.att2.ln_y.weight", "encoder.6.att2.ln_y.bias", "encoder.6.att2.ln_o1.weight", "encoder.6.att2.ln_o1.bias", "encoder.6.att2.ln_o2.weight", "encoder.6.att2.ln_o2.bias", "decoder.0.att1.ln_x.weight", "decoder.0.att1.ln_x.bias", "decoder.0.att1.ln_y.weight", "decoder.0.att1.ln_y.bias", "decoder.0.att1.ln_o1.weight", "decoder.0.att1.ln_o1.bias", "decoder.0.att1.ln_o2.weight", "decoder.0.att1.ln_o2.bias", "decoder.0.att2.ln_x.weight", "decoder.0.att2.ln_x.bias", "decoder.0.att2.ln_y.weight", "decoder.0.att2.ln_y.bias", "decoder.0.att2.ln_o1.weight", "decoder.0.att2.ln_o1.bias", "decoder.0.att2.ln_o2.weight", "decoder.0.att2.ln_o2.bias", "decoder.1.att1.ln_x.weight", "decoder.1.att1.ln_x.bias", "decoder.1.att1.ln_y.weight", "decoder.1.att1.ln_y.bias", "decoder.1.att1.ln_o1.weight", "decoder.1.att1.ln_o1.bias", "decoder.1.att1.ln_o2.weight", "decoder.1.att1.ln_o2.bias", "decoder.1.att2.ln_x.weight", "decoder.1.att2.ln_x.bias", "decoder.1.att2.ln_y.weight", "decoder.1.att2.ln_y.bias", "decoder.1.att2.ln_o1.weight", "decoder.1.att2.ln_o1.bias", "decoder.1.att2.ln_o2.weight", "decoder.1.att2.ln_o2.bias", "decoder.2.att1.ln_x.weight", "decoder.2.att1.ln_x.bias", "decoder.2.att1.ln_y.weight", "decoder.2.att1.ln_y.bias", "decoder.2.att1.ln_o1.weight", "decoder.2.att1.ln_o1.bias", "decoder.2.att1.ln_o2.weight", "decoder.2.att1.ln_o2.bias", "decoder.2.att2.ln_x.weight", "decoder.2.att2.ln_x.bias", "decoder.2.att2.ln_y.weight", "decoder.2.att2.ln_y.bias", "decoder.2.att2.ln_o1.weight", "decoder.2.att2.ln_o1.bias", "decoder.2.att2.ln_o2.weight", "decoder.2.att2.ln_o2.bias", "decoder.3.att1.ln_x.weight", "decoder.3.att1.ln_x.bias", "decoder.3.att1.ln_y.weight", "decoder.3.att1.ln_y.bias", "decoder.3.att1.ln_o1.weight", "decoder.3.att1.ln_o1.bias", "decoder.3.att1.ln_o2.weight", "decoder.3.att1.ln_o2.bias", "decoder.3.att2.ln_x.weight", "decoder.3.att2.ln_x.bias", "decoder.3.att2.ln_y.weight", "decoder.3.att2.ln_y.bias", "decoder.3.att2.ln_o1.weight", "decoder.3.att2.ln_o1.bias", "decoder.3.att2.ln_o2.weight", "decoder.3.att2.ln_o2.bias", "decoder.4.att1.ln_x.weight", "decoder.4.att1.ln_x.bias", "decoder.4.att1.ln_y.weight", "decoder.4.att1.ln_y.bias", "decoder.4.att1.ln_o1.weight", "decoder.4.att1.ln_o1.bias", "decoder.4.att1.ln_o2.weight", "decoder.4.att1.ln_o2.bias", "decoder.4.att2.ln_x.weight", "decoder.4.att2.ln_x.bias", "decoder.4.att2.ln_y.weight", "decoder.4.att2.ln_y.bias", "decoder.4.att2.ln_o1.weight", "decoder.4.att2.ln_o1.bias", "decoder.4.att2.ln_o2.weight", "decoder.4.att2.ln_o2.bias", "decoder.5.att1.ln_x.weight", "decoder.5.att1.ln_x.bias", "decoder.5.att1.ln_y.weight", "decoder.5.att1.ln_y.bias", "decoder.5.att1.ln_o1.weight", "decoder.5.att1.ln_o1.bias", "decoder.5.att1.ln_o2.weight", "decoder.5.att1.ln_o2.bias", "decoder.5.att2.ln_x.weight", "decoder.5.att2.ln_x.bias", "decoder.5.att2.ln_y.weight", "decoder.5.att2.ln_y.bias", "decoder.5.att2.ln_o1.weight", "decoder.5.att2.ln_o1.bias", "decoder.5.att2.ln_o2.weight", "decoder.5.att2.ln_o2.bias", "decoder.6.att1.ln_x.weight", "decoder.6.att1.ln_x.bias", "decoder.6.att1.ln_y.weight", "decoder.6.att1.ln_y.bias", "decoder.6.att1.ln_o1.weight", "decoder.6.att1.ln_o1.bias", "decoder.6.att1.ln_o2.weight", "decoder.6.att1.ln_o2.bias", "decoder.6.att2.ln_x.weight", "decoder.6.att2.ln_x.bias", "decoder.6.att2.ln_y.weight", "decoder.6.att2.ln_y.bias", "decoder.6.att2.ln_o1.weight", "decoder.6.att2.ln_o1.bias", "decoder.6.att2.ln_o2.weight", "decoder.6.att2.ln_o2.bias".
Hi,
Is the attention visualization created by training a model for one class?
can you please introduce how do you conduct the exp of generation of a scene layout?
Hi,
Thanks for the awesome work. I'm trying to run the code, via non-docker method (because I only have non-sudo access to my server), and I encounter this error when running bash install.sh
. Here is the error message.
mv: cannot stat 'build/lib.linux-x86_64-3.6/StructuralLosses': No such file or directory
Makefile:75: recipe for target 'all' failed
make: *** [all] Error 1
What should I do?
I notice that without above installation, errors will be encountered in metrics/evaluation_metrics.py/metrics.StructuralLosses
. For now, I only want to run the code with dummy data to see the data flow (value and each variable shape) to help me understand the details of the paper. Thus, running the non-optimal one (no cuda version) is also OK for me. Any suggestion?
Many thanks!
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ 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. ๐๐๐
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 โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.