ferg-lab / diamondback Goto Github PK
View Code? Open in Web Editor NEWDiAMoNDBack: Diffusion-denoising Autoregressive Model for Non-Deterministic Backmapping of Cα Protein Traces
DiAMoNDBack: Diffusion-denoising Autoregressive Model for Non-Deterministic Backmapping of Cα Protein Traces
Hi,
Thank you for sharing this work!
I am trying to run the code on the provided test data as shared in the README.
When I "backmap the PDB test data using the PDB-trained model" with this command:
python run_eval.py --training_set PDB --data_type aa --pdb_dir ../data/PDB_test_pdbs/
It finishes all the steps successfully but errors when converting to mdtraj:
Traceback (most recent call last):
File "/DiAMoNDBack/scripts/run_eval.py", line 260, in <module>
iterative_inference(data, training_set, cond_type, n_cond_res, model_name, pdb_save_names=pb_save_names, dtime=dtime,
File "/DiAMoNDBack/scripts/run_eval.py", line 249, in iterative_inference
crd_to_mdtrj(all_seqs, all_crds_ref, all_crds, end_id_list,
File "/DiAMoNDBack/scripts/utils.py", line 792, in crd_to_mdtrj
sb_ref = scn.StructureBuilder(s, crd=c)
File "/home/md-mi/miniconda3/envs/diffback/lib/python3.10/site-packages/sidechainnet/structure/StructureBuilder.py", line 95, in __init__
raise ValueError(
ValueError: The length of the coordinate matrix must match the sequence length. You have provided coords. shape[0] = 3080.
Additionally, when I "backmap the DES test set using the DES-finetuned model" with this command:
python run_eval.py --training_set PDB_DES-FT --data_type aa --pdb_dir ../data/all_train_test_pdbs/DES_test
It gives this error:
trained models/train_PDB_DES-FT/32-1-2-4-8-6128-11.0-05-gammal.0--rescale20.0- 1
Milestone = 15
Traceback (most recent call last):
File "/DiAMoNDBack/scripts/run_eval.py", line 260, in <module>
iterative_inference(data, training_set, cond_type, n_cond_res, model_name, pdb_save_names=pdb_save_names, dtime=dtime,
File "/DiAMoNDBack/scripts/run_eval.py", line 181, in iterative_inference
for i in range (max(n_res_list)):
ValueError: max() arg is an empty sequence
May I ask how to fix this?
Thanks
Hello!
First of all, thank you very much for making this work available!
I'm trying to use the code but having some trouble making it work.
First, I tried to install it by following the instructions on my MacOS machine, but then I found out that the environment needed to be Linux.
Then, I tried to install it on a Linux machine, but then found out that it requires a specific CUDA:
Traceback (most recent call last):
File "run_eval.py", line 262, in <module>
gamma=gamma, train_lr=train_lr, adamw=adamw, rescale=rescale, round2=False, decode_termini=decode_termini)
File "run_eval.py", line 194, in iterative_inference
gen_xyzs = run_inference(trainer, data_name=test_iter_prefix, folder_name=test_save_dir, batch_size=5000)
File "/home/diego/bkp.bmp/DiAMoNDBack/scripts/utils.py", line 442, in run_inference
trainer.op_number, batch_size=n, samples = next(sample_dl).cuda()[:n, :]), batches))
File "/home/diego/bkp.bmp/DiAMoNDBack/scripts/utils.py", line 442, in <lambda>
trainer.op_number, batch_size=n, samples = next(sample_dl).cuda()[:n, :]), batches))
File "/home/diego/.local/conda/envs/diffback/lib/python3.7/site-packages/torch/autograd/grad_mode.py", line 26, in decorate_context
return func(*args, **kwargs)
File "../denoising_diffusion_pytorch/denoising_diffusion_pytorch_backmap_combined.py", line 702, in sample
return self.p_sample_loop((batch_size, 1, op_number), samples)
File "/home/diego/.local/conda/envs/diffback/lib/python3.7/site-packages/torch/autograd/grad_mode.py", line 26, in decorate_context
return func(*args, **kwargs)
File "../denoising_diffusion_pytorch/denoising_diffusion_pytorch_backmap_combined.py", line 655, in p_sample_loop
state = torch.randn(shape, device=device)
RuntimeError: CUDA error: no kernel image is available for execution on the device
I wonder if you could update the conda environment script, make available a CPU version, or at least write down a list of minimum requirements that are needed to try/test/reproduce the work.
Thank you!
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.