liuxin

training result

14

hi, I trained openfold and use the ckpt to test sequence, but it was not correct, lots of atoms overlapped together ![Uploading 微信图片_20220525170716.jpg…]()

CUBLAS_STATUS_INTERNAL_ERROR

3

I met this problem: ```python RuntimeError: CUDA error: CUBLAS_STATUS_INTERNAL_ERROR when calling `cublasSgemmStridedBatched( handle, opa, opb, m, n, k, &alpha, a, lda, stridea, b, ldb, strideb, &beta, c, ldc, stridec, num_batches)`...

nan problem while training

7

according to debug the code, I find there is a problem with the data entering the model, some values which prefixed with template is nan

train error about chain_data_cache

8

hi , I generated chain_data_cache.json，but when I run train_openfold.py, it will error in row 301 in openfold/data/data_module.py，because train_dataset is much larger than chain_data_cache.json，chain_data_cache.json cannot fully contain train_dataset

different dap

12

hi, I try different dap_size, such as dap_size=2, dap_size=4, but with the increase of dap, the decrease of gpu memory is not obvious, have you tried this？

use dap in openfold

1

hi, I see how to use dap is described in the readme, as follows, ```python from fastfold.distributed import init_dap torch.distributed.init_process_group(backend='nccl', init_method='env://') init_dap(args.dap_size) ``` I want to know is it possible...

data problem when use dap

9

I'm confused about dap, 1. Can the parameter dap_size only take 2? means row and column？ 2. Is the input data complete or do I need to divide the data...

scn.load() error

3

when I run python train_pre.py, it get error at scn.load(), can not get the pkl file，how can I get the pkl file before I run the program