Add `ProteinMPNN` model and example #10289

xnuohz · 2025-05-25T17:31:06Z

Feature summary

Add a new dataset ProteinMPNNDataset includes 2 versions (small and large)
Add a new model ProteinMPNN
Add unit test
Add an example for loading the dataset and training the model

Issue

Usage

python examples/llm/protein_mpnn.py --gradient_norm 5 --mixed_precision True

Highlight

Small dataset training log

python examples/llm/protein_mpnn.py --gradient_norm 5 --mixed_precision True
Processing...
Processing split: 100%|██████████████████████████████████████████████████████████████████████████████████████████| 473127/473127 [00:11<00:00, 41755.99it/s]
Processing:   0%|                                                                                                                | 0/438572 [00:00<?, ?it/s]/home/ubuntu/Projects/pytorch_geometric/torch_geometric/datasets/protein_mpnn_dataset.py:211: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  meta = torch.load(f'{prefix}.pt')
/home/ubuntu/Projects/pytorch_geometric/torch_geometric/datasets/protein_mpnn_dataset.py:243: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  c: torch.load(f'{prefix}_{c}.pt')
Processing: 100%|███████████████████████████████████████████████████████████████████████████████████████████████| 438572/438572 [00:01<00:00, 256820.89it/s]
Done!
Processing...
Load split
Processing: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 17957/17957 [00:00<00:00, 40089.40it/s]
Done!
Processing...
Load split
Processing: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████| 16598/16598 [00:00<00:00, 467106.31it/s]
Done!
epoch: 000, step: 0, train: 27.305, valid: 19.288, train_acc: 0.068, valid_acc: 0.111
epoch: 001, step: 0, train: 18.214, valid: 19.615, train_acc: 0.103, valid_acc: 0.078
epoch: 002, step: 0, train: 16.328, valid: 19.160, train_acc: 0.131, valid_acc: 0.097
epoch: 003, step: 0, train: 15.034, valid: 19.335, train_acc: 0.159, valid_acc: 0.118
epoch: 004, step: 0, train: 13.922, valid: 18.412, train_acc: 0.186, valid_acc: 0.120
epoch: 005, step: 0, train: 12.825, valid: 17.270, train_acc: 0.216, valid_acc: 0.143
epoch: 006, step: 0, train: 12.273, valid: 19.554, train_acc: 0.231, valid_acc: 0.132
epoch: 007, step: 0, train: 11.274, valid: 17.401, train_acc: 0.257, valid_acc: 0.151
epoch: 008, step: 0, train: 10.498, valid: 17.596, train_acc: 0.283, valid_acc: 0.145
epoch: 009, step: 0, train: 9.125, valid: 17.946, train_acc: 0.334, valid_acc: 0.145
epoch: 010, step: 0, train: 8.843, valid: 18.615, train_acc: 0.352, valid_acc: 0.152
epoch: 011, step: 0, train: 7.689, valid: 19.883, train_acc: 0.408, valid_acc: 0.144
epoch: 012, step: 0, train: 7.028, valid: 16.705, train_acc: 0.441, valid_acc: 0.166
epoch: 013, step: 0, train: 6.340, valid: 18.358, train_acc: 0.476, valid_acc: 0.159
epoch: 014, step: 0, train: 5.573, valid: 21.194, train_acc: 0.521, valid_acc: 0.148
epoch: 015, step: 0, train: 5.223, valid: 18.807, train_acc: 0.537, valid_acc: 0.154
epoch: 016, step: 0, train: 4.525, valid: 23.286, train_acc: 0.575, valid_acc: 0.160
epoch: 017, step: 0, train: 4.288, valid: 18.867, train_acc: 0.590, valid_acc: 0.173
epoch: 018, step: 0, train: 3.842, valid: 25.381, train_acc: 0.615, valid_acc: 0.159
epoch: 019, step: 0, train: 3.798, valid: 16.560, train_acc: 0.625, valid_acc: 0.196
epoch: 020, step: 0, train: 3.464, valid: 15.796, train_acc: 0.652, valid_acc: 0.222
epoch: 021, step: 0, train: 3.095, valid: 10.163, train_acc: 0.687, valid_acc: 0.328
epoch: 022, step: 0, train: 2.602, valid: 8.160, train_acc: 0.739, valid_acc: 0.396
epoch: 023, step: 0, train: 2.444, valid: 12.667, train_acc: 0.764, valid_acc: 0.321
epoch: 024, step: 0, train: 2.230, valid: 7.600, train_acc: 0.790, valid_acc: 0.443
epoch: 025, step: 0, train: 2.139, valid: 7.181, train_acc: 0.802, valid_acc: 0.469
epoch: 026, step: 0, train: 2.050, valid: 6.300, train_acc: 0.813, valid_acc: 0.495
epoch: 027, step: 0, train: 2.008, valid: 6.220, train_acc: 0.819, valid_acc: 0.508
epoch: 028, step: 0, train: 1.963, valid: 6.453, train_acc: 0.825, valid_acc: 0.509
epoch: 029, step: 0, train: 1.956, valid: 5.357, train_acc: 0.826, valid_acc: 0.546
epoch: 030, step: 0, train: 1.883, valid: 6.460, train_acc: 0.835, valid_acc: 0.522
epoch: 031, step: 0, train: 1.849, valid: 5.659, train_acc: 0.841, valid_acc: 0.553
epoch: 032, step: 0, train: 1.807, valid: 5.535, train_acc: 0.846, valid_acc: 0.561
epoch: 033, step: 0, train: 1.798, valid: 5.756, train_acc: 0.846, valid_acc: 0.541
epoch: 034, step: 0, train: 1.779, valid: 5.609, train_acc: 0.850, valid_acc: 0.552
epoch: 035, step: 0, train: 1.737, valid: 5.565, train_acc: 0.856, valid_acc: 0.555
epoch: 036, step: 0, train: 1.717, valid: 6.836, train_acc: 0.860, valid_acc: 0.534
epoch: 037, step: 0, train: 1.697, valid: 5.379, train_acc: 0.863, valid_acc: 0.557
epoch: 038, step: 0, train: 1.683, valid: 6.461, train_acc: 0.865, valid_acc: 0.543
epoch: 039, step: 0, train: 1.680, valid: 5.368, train_acc: 0.867, valid_acc: 0.556
epoch: 040, step: 0, train: 1.655, valid: 5.264, train_acc: 0.870, valid_acc: 0.560
epoch: 041, step: 0, train: 1.639, valid: 6.009, train_acc: 0.874, valid_acc: 0.543
epoch: 042, step: 0, train: 1.623, valid: 5.207, train_acc: 0.876, valid_acc: 0.557
epoch: 043, step: 0, train: 1.609, valid: 5.855, train_acc: 0.880, valid_acc: 0.549
epoch: 044, step: 0, train: 1.595, valid: 6.061, train_acc: 0.881, valid_acc: 0.554
epoch: 045, step: 0, train: 1.576, valid: 6.337, train_acc: 0.887, valid_acc: 0.547
epoch: 046, step: 0, train: 1.571, valid: 6.588, train_acc: 0.888, valid_acc: 0.540
epoch: 047, step: 0, train: 1.553, valid: 5.722, train_acc: 0.890, valid_acc: 0.551
epoch: 048, step: 0, train: 1.529, valid: 5.834, train_acc: 0.895, valid_acc: 0.552
epoch: 049, step: 0, train: 1.523, valid: 5.628, train_acc: 0.897, valid_acc: 0.567
Average Epoch Time: 4.2097s
Median Epoch Time: 4.2084s
Total Program Runtime: 225.3954s
test: 5.006, test_acc: 0.569

codecov · 2025-05-25T17:56:41Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.64%. Comparing base (c211214) to head (9b81985).
⚠️ Report is 117 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #10289      +/-   ##
==========================================
- Coverage   86.11%   85.64%   -0.47%     
==========================================
  Files         496      499       +3     
  Lines       33655    34297     +642     
==========================================
+ Hits        28981    29373     +392     
- Misses       4674     4924     +250

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

for more information, see https://pre-commit.ci

xnuohz · 2025-05-31T03:26:04Z

@jdhenaos plz share your large dataset training log here, thanks:)

jdhenaos · 2025-05-31T11:40:01Z

Large dataset training log

python pytorch_geometric/examples/llm/protein_mpnn.py --gradient_norm 5 --mixed_precision True --size large > pmpnn.log
/home/juan/anaconda3/envs/pygfix/lib/python3.10/site-packages/torch/amp/grad_scaler.py:132: UserWarning: torch.cuda.amp.GradScaler is enabled, but CUDA is not available.  Disabling.
  warnings.warn(
/home/juan/Documents/projects/pytorch_geometric/examples/llm/protein_mpnn.py:164: UserWarning: WARNING: Available system RAM (62.24 GB) is less than the dataset size (64.1 GB).
Consider freeing memory or using a machine with more available RAM.
  warnings.warn(
Downloading https://files.ipd.uw.edu/pub/training_sets/pdb_2021aug02.tar.gz
Extracting data/ProteinMPNN/raw/pdb_2021aug02.tar.gz
Processing...
Processing split: 100%|███████████████████████████████████████████████████████████████████| 473062/473062 [00:14<00:00, 32212.67it/s]
Processing: 100%|████████████████████████████████████████████████████████████████████████| 438507/438507 [00:00<00:00, 489343.72it/s]
Done!
Processing...
Processing: 100%|███████████████████████████████████████████████████████████████████████████| 17957/17957 [00:01<00:00, 15913.02it/s]
Done!
Processing...
Processing: 100%|███████████████████████████████████████████████████████████████████████████| 16598/16598 [00:01<00:00, 11169.99it/s]
Done!
Load split
Load split
epoch: 000, step: 0, train: 26.981, valid: 19.407, train_acc: 0.079, valid_acc: 0.090
epoch: 001, step: 0, train: 17.997, valid: 18.147, train_acc: 0.107, valid_acc: 0.121
epoch: 002, step: 0, train: 16.041, valid: 16.945, train_acc: 0.129, valid_acc: 0.134
epoch: 003, step: 0, train: 13.906, valid: 16.862, train_acc: 0.168, valid_acc: 0.146
epoch: 004, step: 0, train: 11.576, valid: 18.258, train_acc: 0.229, valid_acc: 0.139
epoch: 005, step: 0, train: 8.895, valid: 19.604, train_acc: 0.322, valid_acc: 0.138
epoch: 006, step: 0, train: 6.411, valid: 26.347, train_acc: 0.440, valid_acc: 0.124
epoch: 007, step: 0, train: 4.690, valid: 28.591, train_acc: 0.548, valid_acc: 0.136
epoch: 008, step: 0, train: 3.681, valid: 31.862, train_acc: 0.630, valid_acc: 0.126
epoch: 009, step: 0, train: 2.990, valid: 35.331, train_acc: 0.697, valid_acc: 0.124
epoch: 010, step: 0, train: 2.543, valid: 37.162, train_acc: 0.745, valid_acc: 0.118
epoch: 011, step: 0, train: 2.229, valid: 38.115, train_acc: 0.785, valid_acc: 0.126
epoch: 012, step: 0, train: 2.027, valid: 39.859, train_acc: 0.815, valid_acc: 0.113
epoch: 013, step: 0, train: 1.846, valid: 38.634, train_acc: 0.843, valid_acc: 0.118
epoch: 014, step: 0, train: 1.742, valid: 40.913, train_acc: 0.860, valid_acc: 0.131
epoch: 015, step: 0, train: 1.644, valid: 38.049, train_acc: 0.878, valid_acc: 0.118
epoch: 016, step: 0, train: 1.570, valid: 39.292, train_acc: 0.895, valid_acc: 0.123
epoch: 017, step: 0, train: 1.511, valid: 39.023, train_acc: 0.906, valid_acc: 0.119
epoch: 018, step: 0, train: 1.469, valid: 35.645, train_acc: 0.914, valid_acc: 0.121
epoch: 019, step: 0, train: 1.407, valid: 36.872, train_acc: 0.929, valid_acc: 0.109
epoch: 020, step: 0, train: 1.407, valid: 35.574, train_acc: 0.927, valid_acc: 0.124
epoch: 021, step: 0, train: 1.366, valid: 36.755, train_acc: 0.938, valid_acc: 0.123
epoch: 022, step: 0, train: 1.324, valid: 39.198, train_acc: 0.947, valid_acc: 0.111
epoch: 023, step: 0, train: 1.336, valid: 37.151, train_acc: 0.943, valid_acc: 0.132
epoch: 024, step: 0, train: 1.307, valid: 33.625, train_acc: 0.949, valid_acc: 0.124
epoch: 025, step: 0, train: 1.269, valid: 34.207, train_acc: 0.960, valid_acc: 0.122
epoch: 026, step: 0, train: 1.255, valid: 31.263, train_acc: 0.963, valid_acc: 0.141
epoch: 027, step: 0, train: 1.260, valid: 32.474, train_acc: 0.960, valid_acc: 0.139
epoch: 028, step: 0, train: 1.246, valid: 30.826, train_acc: 0.964, valid_acc: 0.154
epoch: 029, step: 0, train: 1.230, valid: 31.579, train_acc: 0.967, valid_acc: 0.132
epoch: 030, step: 0, train: 1.235, valid: 29.492, train_acc: 0.965, valid_acc: 0.156
epoch: 031, step: 0, train: 1.227, valid: 29.500, train_acc: 0.967, valid_acc: 0.141
epoch: 032, step: 0, train: 1.214, valid: 30.314, train_acc: 0.971, valid_acc: 0.156
epoch: 033, step: 0, train: 1.190, valid: 32.563, train_acc: 0.976, valid_acc: 0.168
epoch: 034, step: 0, train: 1.198, valid: 29.133, train_acc: 0.973, valid_acc: 0.157
epoch: 035, step: 0, train: 1.188, valid: 32.391, train_acc: 0.976, valid_acc: 0.145
epoch: 036, step: 0, train: 1.199, valid: 28.737, train_acc: 0.973, valid_acc: 0.173
epoch: 037, step: 0, train: 1.204, valid: 24.798, train_acc: 0.970, valid_acc: 0.202
epoch: 038, step: 0, train: 1.197, valid: 23.723, train_acc: 0.974, valid_acc: 0.216
epoch: 039, step: 0, train: 1.175, valid: 22.979, train_acc: 0.979, valid_acc: 0.216
epoch: 040, step: 0, train: 1.189, valid: 23.653, train_acc: 0.974, valid_acc: 0.228
epoch: 041, step: 0, train: 1.179, valid: 23.006, train_acc: 0.976, valid_acc: 0.231
epoch: 042, step: 0, train: 1.172, valid: 21.775, train_acc: 0.979, valid_acc: 0.238
epoch: 043, step: 0, train: 1.181, valid: 23.155, train_acc: 0.977, valid_acc: 0.239
epoch: 044, step: 0, train: 1.181, valid: 20.401, train_acc: 0.978, valid_acc: 0.255
epoch: 045, step: 0, train: 1.160, valid: 20.502, train_acc: 0.983, valid_acc: 0.254
epoch: 046, step: 0, train: 1.164, valid: 20.080, train_acc: 0.981, valid_acc: 0.273
epoch: 047, step: 0, train: 1.163, valid: 21.548, train_acc: 0.981, valid_acc: 0.262
epoch: 048, step: 0, train: 1.168, valid: 23.360, train_acc: 0.979, valid_acc: 0.263
epoch: 049, step: 0, train: 1.166, valid: 19.664, train_acc: 0.981, valid_acc: 0.282
Average Epoch Time: 104.2671s
Median Epoch Time: 105.9803s
Total Program Runtime: 9090.4755s
test: 17.504, test_acc: 0.304

puririshi98

see comments

examples/llm/README.md

test/nn/models/test_protein_mpnn.py

for more information, see https://pre-commit.ci

.github/actions/setup/action.yml

test/datasets/test_protein_mpnn_dataset.py

puririshi98 · 2025-06-11T21:59:05Z

left comments on the existing threads, please address and I will be ready to merge once comments addressed and CI is passing

xnuohz · 2025-06-12T13:10:55Z

@puririshi98 new ci tests work fine on 25.05 pyg container
@jdhenaos check if your environment is installed correctly

jdhenaos · 2025-06-12T14:16:53Z

@xnuohz thanks for double checking!

I have reinstalled the Docker image and rerun the test. However, the probe for spmm persists. I am clueless about what I am doing wrong. On the other hand, protein_mpnn worked.

root@246f883b8ec4:/workspace# pytest test/datasets/test_protein_mpnn_dataset.py test/nn/models/test_protein_mpnn.py test/utils/test_spmm.py 
========================================= test session starts ==========================================
platform linux -- Python 3.12.3, pytest-8.1.1, pluggy-1.6.0 -- /usr/bin/python
cachedir: .pytest_cache
hypothesis profile 'default' -> database=DirectoryBasedExampleDatabase(PosixPath('/workspace/.hypothesis/examples'))
rootdir: /workspace
configfile: pyproject.toml
plugins: cov-6.1.1, anyio-4.9.0, hypothesis-6.130.8, flakefinder-1.1.0, xdist-3.6.1, shard-0.1.2, rerunfailures-15.1, xdoctest-1.0.2, typeguard-4.3.0
collected 44 items                                                                                     
Running 44 items in this shard: test/datasets/test_protein_mpnn_dataset.py::test_protein_mpnn_dataset, test/nn/models/test_protein_mpnn.py::test_protein_mpnn, test/utils/test_spmm.py::test_spmm_basic[sum-cpu], test/utils/test_spmm.py::test_spmm_basic[sum-cuda:0], test/utils/test_spmm.py::test_spmm_basic[mean-cpu], test/utils/test_spmm.py::test_spmm_basic[mean-cuda:0], test/utils/test_spmm.py::test_spmm_reduce[min-cpu], test/utils/test_spmm.py::test_spmm_reduce[min-cuda:0], test/utils/test_spmm.py::test_spmm_reduce[max-cpu], test/utils/test_spmm.py::test_spmm_reduce[max-cuda:0], test/utils/test_spmm.py::test_spmm_layout[sum-layout0-cpu], test/utils/test_spmm.py::test_spmm_layout[sum-layout0-cuda:0], test/utils/test_spmm.py::test_spmm_layout[sum-layout1-cpu], test/utils/test_spmm.py::test_spmm_layout[sum-layout1-cuda:0], test/utils/test_spmm.py::test_spmm_layout[sum-layout2-cpu], test/utils/test_spmm.py::test_spmm_layout[sum-layout2-cuda:0], test/utils/test_spmm.py::test_spmm_layout[mean-layout0-cpu], test/utils/test_spmm.py::test_spmm_layout[mean-layout0-cuda:0], test/utils/test_spmm.py::test_spmm_layout[mean-layout1-cpu], test/utils/test_spmm.py::test_spmm_layout[mean-layout1-cuda:0], test/utils/test_spmm.py::test_spmm_layout[mean-layout2-cpu], test/utils/test_spmm.py::test_spmm_layout[mean-layout2-cuda:0], test/utils/test_spmm.py::test_spmm_layout[min-layout0-cpu], test/utils/test_spmm.py::test_spmm_layout[min-layout0-cuda:0], test/utils/test_spmm.py::test_spmm_layout[min-layout1-cpu], test/utils/test_spmm.py::test_spmm_layout[min-layout1-cuda:0], test/utils/test_spmm.py::test_spmm_layout[min-layout2-cpu], test/utils/test_spmm.py::test_spmm_layout[min-layout2-cuda:0], test/utils/test_spmm.py::test_spmm_layout[max-layout0-cpu], test/utils/test_spmm.py::test_spmm_layout[max-layout0-cuda:0], test/utils/test_spmm.py::test_spmm_layout[max-layout1-cpu], test/utils/test_spmm.py::test_spmm_layout[max-layout1-cuda:0], test/utils/test_spmm.py::test_spmm_layout[max-layout2-cpu], test/utils/test_spmm.py::test_spmm_layout[max-layout2-cuda:0], test/utils/test_spmm.py::test_spmm_jit[sum], test/utils/test_spmm.py::test_spmm_jit[mean], test/utils/test_spmm.py::test_spmm_edge_index[sum-cpu], test/utils/test_spmm.py::test_spmm_edge_index[sum-cuda:0], test/utils/test_spmm.py::test_spmm_edge_index[mean-cpu], test/utils/test_spmm.py::test_spmm_edge_index[mean-cuda:0], test/utils/test_spmm.py::test_spmm_edge_index[min-cpu], test/utils/test_spmm.py::test_spmm_edge_index[min-cuda:0], test/utils/test_spmm.py::test_spmm_edge_index[max-cpu], test/utils/test_spmm.py::test_spmm_edge_index[max-cuda:0]

test/datasets/test_protein_mpnn_dataset.py::test_protein_mpnn_dataset Load split
Processing: 100%|███████████████████████████████████████████| 438572/438572 [00:02<00:00, 158569.97it/s]
PASSED
test/nn/models/test_protein_mpnn.py::test_protein_mpnn SKIPPED (Package torch_cluster not found)
test/utils/test_spmm.py::test_spmm_basic[sum-cpu] PASSED
test/utils/test_spmm.py::test_spmm_basic[sum-cuda:0] FAILED
test/utils/test_spmm.py::test_spmm_basic[mean-cpu] PASSED
test/utils/test_spmm.py::test_spmm_basic[mean-cuda:0] FAILED
test/utils/test_spmm.py::test_spmm_reduce[min-cpu] PASSED
test/utils/test_spmm.py::test_spmm_reduce[min-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_reduce[max-cpu] PASSED
test/utils/test_spmm.py::test_spmm_reduce[max-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout0-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout0-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout1-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout1-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout2-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[sum-layout2-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout0-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout0-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout1-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout1-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout2-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[mean-layout2-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout0-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout0-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout1-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout1-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout2-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[min-layout2-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout0-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout0-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout1-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout1-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout2-cpu] PASSED
test/utils/test_spmm.py::test_spmm_layout[max-layout2-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_jit[sum] PASSED
test/utils/test_spmm.py::test_spmm_jit[mean] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[sum-cpu] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[sum-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[mean-cpu] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[mean-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[min-cpu] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[min-cuda:0] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[max-cpu] PASSED
test/utils/test_spmm.py::test_spmm_edge_index[max-cuda:0] PASSED

=============================================== FAILURES ===============================================
_____________________________________ test_spmm_basic[sum-cuda:0] ______________________________________

device = device(type='cuda', index=0), reduce = 'sum'

    @withCUDA
    @pytest.mark.parametrize('reduce', ['sum', 'mean'])
    def test_spmm_basic(device, reduce):
        src = torch.randn(5, 4, device=device)
        other = torch.randn(4, 8, device=device)
    
        out1 = (src @ other) / (src.size(1) if reduce == 'mean' else 1)
        out2 = spmm(src.to_sparse_csr(), other, reduce=reduce)
        assert out1.size() == (5, 8)
>       assert torch.allclose(out1, out2, atol=1e-6)
E       AssertionError: assert False
E        +  where False = <built-in method allclose of type object at 0x72d3ee1842a0>(tensor([[ 0.2263,  0.5490,  0.4093,  0.6030,  0.7371, -0.4968, -0.4992,  0.1147],\n        [ 0.3482, -0.7796,  0.1877, -0.6512,  0.0834, -0.9297,  0.3539,  1.9510],\n        [ 0.7085,  0.5063,  1.1544,  0.1988,  1.3384, -1.5084, -0.8598,  0.9801],\n        [ 0.4680,  0.6143, -1.0471, -2.4450,  1.0119, -1.0227,  0.8287,  1.4525],\n        [-0.5595,  0.7819, -2.7147,  5.6426,  1.4819, -0.3322,  1.3007,  4.8081]],\n       device='cuda:0'), tensor([[ 0.2263,  0.5490,  0.4094,  0.6029,  0.7368, -0.4966, -0.4993,  0.1147],\n        [ 0.3482, -0.7798,  0.1878, -0.6511,  0.0837, -0.9296,  0.3540,  1.9513],\n        [ 0.7086,  0.5063,  1.1549,  0.1991,  1.3383, -1.5082, -0.8600,  0.9802],\n        [ 0.4679,  0.6143, -1.0472, -2.4445,  1.0120, -1.0224,  0.8288,  1.4523],\n        [-0.5594,  0.7830, -2.7147,  5.6439,  1.4836, -0.3322,  1.3001,  4.8078]],\n       device='cuda:0'), atol=1e-06)
E        +    where <built-in method allclose of type object at 0x72d3ee1842a0> = torch.allclose

test/utils/test_spmm.py:25: AssertionError
_____________________________________ test_spmm_basic[mean-cuda:0] _____________________________________

device = device(type='cuda', index=0), reduce = 'mean'

    @withCUDA
    @pytest.mark.parametrize('reduce', ['sum', 'mean'])
    def test_spmm_basic(device, reduce):
        src = torch.randn(5, 4, device=device)
        other = torch.randn(4, 8, device=device)
    
        out1 = (src @ other) / (src.size(1) if reduce == 'mean' else 1)
        out2 = spmm(src.to_sparse_csr(), other, reduce=reduce)
        assert out1.size() == (5, 8)
>       assert torch.allclose(out1, out2, atol=1e-6)
E       AssertionError: assert False
E        +  where False = <built-in method allclose of type object at 0x72d3ee1842a0>(tensor([[-0.1546, -0.0307, -0.3225, -0.0970,  0.2301,  0.5281,  0.0804,  0.1057],\n        [ 0.1258, -0.0055,  0.1565, -0.4203, -0.2013, -0.4148, -0.5455, -0.2861],\n        [-0.5532, -0.4884, -0.3228,  0.3831,  0.1442,  0.7102,  0.8012,  0.7611],\n        [-0.5898, -0.0525, -0.7840, -0.3497,  0.6824,  1.4893,  0.0843,  0.2738],\n        [ 0.1357, -0.2766, -0.1663,  0.0844, -0.1866, -0.0933,  0.2581,  0.1851]],\n       device='cuda:0'), tensor([[-0.1548, -0.0308, -0.3225, -0.0970,  0.2301,  0.5281,  0.0804,  0.1059],\n        [ 0.1259, -0.0053,  0.1566, -0.4202, -0.2012, -0.4148, -0.5457, -0.2861],\n        [-0.5532, -0.4883, -0.3228,  0.3830,  0.1441,  0.7099,  0.8012,  0.7609],\n        [-0.5900, -0.0525, -0.7839, -0.3498,  0.6823,  1.4891,  0.0842,  0.2739],\n        [ 0.1357, -0.2766, -0.1662,  0.0845, -0.1866, -0.0933,  0.2581,  0.1851]],\n       device='cuda:0'), atol=1e-06)
E        +    where <built-in method allclose of type object at 0x72d3ee1842a0> = torch.allclose

test/utils/test_spmm.py:25: AssertionError
======================================= short test summary info ========================================
FAILED test/utils/test_spmm.py::test_spmm_basic[sum-cuda:0] - AssertionError: assert False
FAILED test/utils/test_spmm.py::test_spmm_basic[mean-cuda:0] - AssertionError: assert False
=============================== 2 failed, 41 passed, 1 skipped in 3.78s ================================

puririshi98 · 2025-06-13T17:59:37Z

my advice would be to run the full unit tests inside the latest nvidia container everytime this branch needs an update from master branch, and when you get clean passing lmk ill merge once your local tests and the github website CI are both green

jdhenaos · 2025-06-16T17:01:47Z

@puririshi98 @xnuohz

I was running a new round of CI tests. However, I encountered an error in moleculeGPT that appears to be more related to the LLM model used than the code itself. This is the unique error I am getting. Am I ignoring something?:

root@77489883e7c5:/workspace# pytest test/datasets/test_molecule_gpt_dataset.py 
================================================================================= test session starts ==================================================================================
platform linux -- Python 3.12.3, pytest-8.1.1, pluggy-1.6.0 -- /usr/bin/python
cachedir: .pytest_cache
hypothesis profile 'default' -> database=DirectoryBasedExampleDatabase(PosixPath('/workspace/.hypothesis/examples'))
rootdir: /workspace
configfile: pyproject.toml
plugins: cov-6.1.1, anyio-4.9.0, hypothesis-6.130.8, flakefinder-1.1.0, xdist-3.6.1, shard-0.1.2, rerunfailures-15.1, xdoctest-1.0.2, typeguard-4.3.0
collected 1 item                                                                                                                                                                       
Running 1 items in this shard: test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset

test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset The length of target_CID_list: 8329
  0%|                                                                                                                                                          | 0/4215 [00:00<?, ?it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
 56%|██████████████████████████████████████████████████████████████████████████████▍                                                              | 2346/4215 [00:00<00:00, 8049.19it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
 75%|█████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                   | 3151/4215 [00:00<00:00, 7962.51it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
 94%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████         | 3948/4215 [00:00<00:00, 7136.67it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4215/4215 [00:00<00:00, 7114.91it/s]
block id: 1 with 0 valid SDF file
In total: 4215 molecules
Setting up 'TinyLlama/TinyLlama-1.1B-Chat-v0.1' with configuration: {'revision': 'main'}
FAILED

======================================================================================= FAILURES =======================================================================================
______________________________________________________________________________ test_molecule_gpt_dataset _______________________________________________________________________________
DeprecationWarning: Type google._upb._message.MessageMapContainer uses PyType_Spec with a metaclass that has custom tp_new. This is deprecated and will no longer be allowed in Python 3.14.

The above exception was the direct cause of the following exception:

    @onlyOnline
    @withPackage('transformers', 'sentencepiece', 'accelerate', 'rdkit')
    def test_molecule_gpt_dataset():
>       dataset = MoleculeGPTDataset(root='./data/MoleculeGPT')

test/datasets/test_molecule_gpt_dataset.py:8: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/usr/local/lib/python3.12/dist-packages/torch_geometric/datasets/molecule_gpt_dataset.py:220: in __init__
    super().__init__(root, transform, pre_transform, pre_filter,
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/in_memory_dataset.py:81: in __init__
    super().__init__(root, transform, pre_transform, pre_filter, log,
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/dataset.py:115: in __init__
    self._process()
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/dataset.py:262: in _process
    self.process()
/usr/local/lib/python3.12/dist-packages/torch_geometric/datasets/molecule_gpt_dataset.py:430: in process
    llm = LLM(
/usr/local/lib/python3.12/dist-packages/torch_geometric/nn/nlp/llm.py:87: in __init__
    self.tokenizer = AutoTokenizer.from_pretrained(
/usr/local/lib/python3.12/dist-packages/transformers/models/auto/tokenization_auto.py:1013: in from_pretrained
    return tokenizer_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
/usr/local/lib/python3.12/dist-packages/transformers/tokenization_utils_base.py:2025: in from_pretrained
    return cls._from_pretrained(
/usr/local/lib/python3.12/dist-packages/transformers/tokenization_utils_base.py:2278: in _from_pretrained
    tokenizer = cls(*init_inputs, **init_kwargs)
/usr/local/lib/python3.12/dist-packages/transformers/models/llama/tokenization_llama.py:171: in __init__
    self.sp_model = self.get_spm_processor(kwargs.pop("from_slow", False))
/usr/local/lib/python3.12/dist-packages/transformers/models/llama/tokenization_llama.py:203: in get_spm_processor
    model_pb2 = import_protobuf(f"The new behaviour of {self.__class__.__name__} (with `self.legacy = False`)")
/usr/local/lib/python3.12/dist-packages/transformers/convert_slow_tokenizer.py:37: in import_protobuf
    from sentencepiece import sentencepiece_model_pb2
/usr/local/lib/python3.12/dist-packages/sentencepiece/sentencepiece_model_pb2.py:5: in <module>
    from google.protobuf.internal import builder as _builder
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/builder.py:41: in <module>
    from google.protobuf.internal import python_message
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/python_message.py:59: in <module>
    from google.protobuf.internal import api_implementation
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/api_implementation.py:74: in <module>
    if _CanImport('google._upb._message'):
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/api_implementation.py:64: in _CanImport
    mod = importlib.import_module(mod_name)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

name = 'google._upb._message', package = None

    def import_module(name, package=None):
        """Import a module.
    
        The 'package' argument is required when performing a relative import. It
        specifies the package to use as the anchor point from which to resolve the
        relative import to an absolute import.
    
        """
        level = 0
        if name.startswith('.'):
            if not package:
                raise TypeError("the 'package' argument is required to perform a "
                                f"relative import for {name!r}")
            for character in name:
                if character != '.':
                    break
                level += 1
>       return _bootstrap._gcd_import(name[level:], package, level)
E       SystemError: <class 'DeprecationWarning'> returned a result with an exception set

/usr/lib/python3.12/importlib/__init__.py:90: SystemError
=============================================================================== short test summary info ================================================================================
FAILED test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset - SystemError: <class 'DeprecationWarning'> returned a result with an exception set
================================================================================== 1 failed in 2.27s ===================================================================================

puririshi98 · 2025-06-26T22:45:50Z

@puririshi98 @xnuohz

I was running a new round of CI tests. However, I encountered an error in moleculeGPT that appears to be more related to the LLM model used than the code itself. This is the unique error I am getting. Am I ignoring something?:

root@77489883e7c5:/workspace# pytest test/datasets/test_molecule_gpt_dataset.py 
================================================================================= test session starts ==================================================================================
platform linux -- Python 3.12.3, pytest-8.1.1, pluggy-1.6.0 -- /usr/bin/python
cachedir: .pytest_cache
hypothesis profile 'default' -> database=DirectoryBasedExampleDatabase(PosixPath('/workspace/.hypothesis/examples'))
rootdir: /workspace
configfile: pyproject.toml
plugins: cov-6.1.1, anyio-4.9.0, hypothesis-6.130.8, flakefinder-1.1.0, xdist-3.6.1, shard-0.1.2, rerunfailures-15.1, xdoctest-1.0.2, typeguard-4.3.0
collected 1 item                                                                                                                                                                       
Running 1 items in this shard: test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset

test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset The length of target_CID_list: 8329
  0%|                                                                                                                                                          | 0/4215 [00:00<?, ?it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
 56%|██████████████████████████████████████████████████████████████████████████████▍                                                              | 2346/4215 [00:00<00:00, 8049.19it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
 75%|█████████████████████████████████████████████████████████████████████████████████████████████████████████▍                                   | 3151/4215 [00:00<00:00, 7962.51it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
 94%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████         | 3948/4215 [00:00<00:00, 7136.67it/s][16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
[16:57:48] WARNING: not removing hydrogen atom without neighbors
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4215/4215 [00:00<00:00, 7114.91it/s]
block id: 1 with 0 valid SDF file
In total: 4215 molecules
Setting up 'TinyLlama/TinyLlama-1.1B-Chat-v0.1' with configuration: {'revision': 'main'}
FAILED

======================================================================================= FAILURES =======================================================================================
______________________________________________________________________________ test_molecule_gpt_dataset _______________________________________________________________________________
DeprecationWarning: Type google._upb._message.MessageMapContainer uses PyType_Spec with a metaclass that has custom tp_new. This is deprecated and will no longer be allowed in Python 3.14.

The above exception was the direct cause of the following exception:

    @onlyOnline
    @withPackage('transformers', 'sentencepiece', 'accelerate', 'rdkit')
    def test_molecule_gpt_dataset():
>       dataset = MoleculeGPTDataset(root='./data/MoleculeGPT')

test/datasets/test_molecule_gpt_dataset.py:8: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/usr/local/lib/python3.12/dist-packages/torch_geometric/datasets/molecule_gpt_dataset.py:220: in __init__
    super().__init__(root, transform, pre_transform, pre_filter,
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/in_memory_dataset.py:81: in __init__
    super().__init__(root, transform, pre_transform, pre_filter, log,
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/dataset.py:115: in __init__
    self._process()
/usr/local/lib/python3.12/dist-packages/torch_geometric/data/dataset.py:262: in _process
    self.process()
/usr/local/lib/python3.12/dist-packages/torch_geometric/datasets/molecule_gpt_dataset.py:430: in process
    llm = LLM(
/usr/local/lib/python3.12/dist-packages/torch_geometric/nn/nlp/llm.py:87: in __init__
    self.tokenizer = AutoTokenizer.from_pretrained(
/usr/local/lib/python3.12/dist-packages/transformers/models/auto/tokenization_auto.py:1013: in from_pretrained
    return tokenizer_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
/usr/local/lib/python3.12/dist-packages/transformers/tokenization_utils_base.py:2025: in from_pretrained
    return cls._from_pretrained(
/usr/local/lib/python3.12/dist-packages/transformers/tokenization_utils_base.py:2278: in _from_pretrained
    tokenizer = cls(*init_inputs, **init_kwargs)
/usr/local/lib/python3.12/dist-packages/transformers/models/llama/tokenization_llama.py:171: in __init__
    self.sp_model = self.get_spm_processor(kwargs.pop("from_slow", False))
/usr/local/lib/python3.12/dist-packages/transformers/models/llama/tokenization_llama.py:203: in get_spm_processor
    model_pb2 = import_protobuf(f"The new behaviour of {self.__class__.__name__} (with `self.legacy = False`)")
/usr/local/lib/python3.12/dist-packages/transformers/convert_slow_tokenizer.py:37: in import_protobuf
    from sentencepiece import sentencepiece_model_pb2
/usr/local/lib/python3.12/dist-packages/sentencepiece/sentencepiece_model_pb2.py:5: in <module>
    from google.protobuf.internal import builder as _builder
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/builder.py:41: in <module>
    from google.protobuf.internal import python_message
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/python_message.py:59: in <module>
    from google.protobuf.internal import api_implementation
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/api_implementation.py:74: in <module>
    if _CanImport('google._upb._message'):
/usr/local/lib/python3.12/dist-packages/google/protobuf/internal/api_implementation.py:64: in _CanImport
    mod = importlib.import_module(mod_name)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

name = 'google._upb._message', package = None

    def import_module(name, package=None):
        """Import a module.
    
        The 'package' argument is required when performing a relative import. It
        specifies the package to use as the anchor point from which to resolve the
        relative import to an absolute import.
    
        """
        level = 0
        if name.startswith('.'):
            if not package:
                raise TypeError("the 'package' argument is required to perform a "
                                f"relative import for {name!r}")
            for character in name:
                if character != '.':
                    break
                level += 1
>       return _bootstrap._gcd_import(name[level:], package, level)
E       SystemError: <class 'DeprecationWarning'> returned a result with an exception set

/usr/lib/python3.12/importlib/__init__.py:90: SystemError
=============================================================================== short test summary info ================================================================================
FAILED test/datasets/test_molecule_gpt_dataset.py::test_molecule_gpt_dataset - SystemError: <class 'DeprecationWarning'> returned a result with an exception set
================================================================================== 1 failed in 2.27s ===================================================================================

this seems unrelated to us, hopefully it goes away with the new 25.07 container coming end of july, or you can file an issue with transformers and see if they fix it

xnuohz · 2025-07-02T09:43:57Z

hi @puririshi98 CI becomes green, can you merge?

puririshi98

lgtm now

xnuohz added 3 commits May 24, 2025 21:32

add proteinmpnn dataset and test case

0bee2c9

add proteinmpnn model

8885c10

Merge branch 'master' into models/proteinmpnn

0d2aeae

xnuohz requested review from EdisonLeeeee and wsad1 as code owners May 25, 2025 17:31

xnuohz added 3 commits May 26, 2025 01:32

add changelog

c12e614

update

e5a7625

update

877b4b7

update

e56190b

This comment was marked as off-topic.

Sign in to view

xnuohz and others added 14 commits May 26, 2025 20:47

update

eca4965

update

f58a089

add noam opt

0655b00

add train/valid/test

36b7dac

add gradient clip and mixed precision

252b2b9

update

9da381f

move edge_index and edge_attr into model

d0b6e36

update

771f8f5

update

0a2b2fd

Merge branch 'master' into models/proteinmpnn

c212363

Add ProteinMPNN complete dataset (#3)

ea3af64

[pre-commit.ci] auto fixes from pre-commit.com hooks

7ba66d9

for more information, see https://pre-commit.ci

update

9dcb079

update

f0b12a0

Merge branch 'master' into models/proteinmpnn

ac70b4a

Merge branch 'master' into models/proteinmpnn

95bdbe6

puririshi98 requested changes Jun 2, 2025

View reviewed changes

examples/llm/README.md Outdated Show resolved Hide resolved

test/nn/models/test_protein_mpnn.py Show resolved Hide resolved

xnuohz and others added 5 commits June 3, 2025 22:27

Merge branch 'master' into models/proteinmpnn

afa8f97

update

7d136ab

update

26b319d

Update README.md

fc40345

[pre-commit.ci] auto fixes from pre-commit.com hooks

2db820f

for more information, see https://pre-commit.ci

puririshi98 requested changes Jun 3, 2025

View reviewed changes

.github/actions/setup/action.yml Outdated Show resolved Hide resolved

puririshi98 requested changes Jun 3, 2025

View reviewed changes

test/datasets/test_protein_mpnn_dataset.py Show resolved Hide resolved

xnuohz added 4 commits June 4, 2025 18:20

Merge branch 'master' into models/proteinmpnn

853081c

update

a4fb642

Merge branch 'master' into models/proteinmpnn

0a7fe94

update

a6d1cb8

Merge branch 'master' into models/proteinmpnn

c4431c0

Merge branch 'master' into models/proteinmpnn

42bb74d

update

6fca81b

Merge branch 'master' into models/proteinmpnn

73b99a6

Merge branch 'master' into models/proteinmpnn

9b81985

puririshi98 approved these changes Jul 2, 2025

View reviewed changes

puririshi98 merged commit 237d077 into pyg-team:master Jul 2, 2025
19 checks passed

xnuohz deleted the models/proteinmpnn branch July 2, 2025 17:20

akihironitta mentioned this pull request Sep 21, 2025

Fix full test failures in master #10465

Draft

Add ProteinMPNN model and example #10289

Add ProteinMPNN model and example #10289

Uh oh!

Conversation

xnuohz commented May 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Feature summary

Issue

Usage

Highlight

Small dataset training log

Uh oh!

codecov bot commented May 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

This comment was marked as off-topic.

Uh oh!

xnuohz commented May 31, 2025

Uh oh!

jdhenaos commented May 31, 2025

Large dataset training log

Uh oh!

puririshi98 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

puririshi98 commented Jun 11, 2025

Uh oh!

xnuohz commented Jun 12, 2025

Uh oh!

jdhenaos commented Jun 12, 2025

Uh oh!

puririshi98 commented Jun 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jdhenaos commented Jun 16, 2025

Uh oh!

puririshi98 commented Jun 26, 2025

Uh oh!

xnuohz commented Jul 2, 2025

Uh oh!

puririshi98 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Add `ProteinMPNN` model and example #10289

Add `ProteinMPNN` model and example #10289

xnuohz commented May 25, 2025 •

edited

Loading

codecov bot commented May 25, 2025 •

edited

Loading

puririshi98 commented Jun 13, 2025 •

edited

Loading