Add OrbFlow QM9 and MD checkpoints

README.md

@@ -0,0 +1,80 @@
+---
+license: mit
+tags:
+  - chemistry
+  - electron-density
+  - flow-matching
+  - equivariant-neural-networks
+  - qm9
+datasets:
+  - divelab/OrbFlow-data
+---
+
+# OrbFlow checkpoints
+
+Trained checkpoints for **OrbFlow**, which predicts atomic-orbital expansion coefficients (and the resulting
+electron density) with equivariant flow matching.
+
+- Code: https://github.com/TODO/OrbFlow
+- Data: [divelab/OrbFlow-data](https://huggingface.co/datasets/divelab/OrbFlow-data)
+- Paper: TODO
+
+## Contents
+
+Each folder is one training run with its **best-validation checkpoint**, `config.yaml` and `metadata.json`.
+Optimizer state has been removed, so these files are for evaluation and inference; they cannot resume training.
+
+| Folder | Model | Seeds | Checkpoint size |
+|--------|-------|-------|-----------------|
+| `qm9_qhf_eqv3_gt_2ph_final`, `…_s42`, `…_s123` | QM9: EquiformerV3, 8 layers, ℓmax = 6; 75k endpoint → 450k integrated-density training | 0, 42, 123 | 1.2 GB |
+| `md_<mol>_k3l3_slim_10k100k`, `…_seed42`, `…_seed505` | MD: slim EquiformerV3, 3 layers, ℓmax = 3; 10k → 110k | 0, 42, 505 | 155 MB |
+
+MD molecules: benzene, ethane, ethanol, malonaldehyde, phenol, resorcinol.
+The seed-505 phenol and resorcinol runs stopped at ~60k of their 110k training steps (job time limit); their best-validation checkpoints are from step ~60k.
+
+## Usage
+
+Download into the root of the OrbFlow code repository, so the folder names match what the evaluation scripts expect:
+
+```bash
+git clone https://github.com/TODO/OrbFlow.git && cd OrbFlow
+# set up the environment and .env as described in the code README, then:
+hf download divelab/OrbFlow --local-dir .
+
+# QM9 test set (Euler K = 2 and 3; full test split and last 1600 molecules)
+CKPT_PATH=qm9_qhf_eqv3_gt_2ph_final sbatch vision/test_qhf_eqv3_gt_2ph_final_full.slurm
+
+# MD test sets (6 molecules); for other seeds use EXP_SUFFIX=md_%s_k3l3_slim_10k100k_seed42
+sbatch vision/test_flow_md_paper_2phase_beta13_all6.slurm
+```
+
+To download a single run: `hf download divelab/OrbFlow --include "qm9_qhf_eqv3_gt_2ph_final/*" --local-dir .`
+
+## Results
+
+Test nMAPE (%) with Euler K = 2, mean ± std over training seeds, computed with the evaluation scripts above.
+
+**QM9** (10,000 test molecules, seeds 0 / 42 / 123)
+
+| Split | nMAPE (%) |
+|-------|-----------|
+| Full test set | 0.1518 ± 0.0008 |
+| Last 1,600 molecules | 0.1646 ± 0.0008 |
+
+**MD** (seeds 0 / 42 / 505)
+
+| Molecule | nMAPE (%) |
+|----------|-----------|
+| benzene | 0.4210 ± 0.0017 |
+| ethane | 0.8706 ± 0.0084 |
+| ethanol | 1.0784 ± 0.0229 |
+| malonaldehyde | 1.1771 ± 0.0127 |
+| phenol | 0.5753 ± 0.0083 |
+| resorcinol | 0.6626 ± 0.0136 |
+| **average (6 molecules)** | **0.7975 ± 0.0067** |
+
+## Citation
+
+TODO
+
+OrbFlow builds on [SCDP](https://github.com/kyonofx/scdp) (Fu et al., NeurIPS 2024).

md_benzene_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,191 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/benzene
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.8172615989393406
+  flow_prior_mu: 0.008670088241267512
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: all
+      log_freq: 100
+core:
+  expname: md_benzene_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_benzene_k3l3_slim_10k100k/epoch=1730-step=108990.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c97567e39c5591b60498012578d1c3c2418bf082d774c8dc60e51318cf7fa72c
+size 154821792

md_benzene_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.049426741898059845, "target_var": 0.09023545682430267, "avg_num_neighbors": 11, "unique_atom_types": [0, 1, 6]}

md_benzene_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/benzene
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.8172615989393406
+  flow_prior_mu: 0.008670088241267512
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_benzene_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_benzene_k3l3_slim_10k100k_seed42/epoch=1741-step=109683.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e91a9ef42a774f8dff79251293f099fb4cfbc5e329ae564d75bb6a927b6f0163
+size 154821984

md_benzene_k3l3_slim_10k100k_seed42/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.049426741898059845, "target_var": 0.09023545682430267, "avg_num_neighbors": 11, "unique_atom_types": [0, 1, 6]}

md_benzene_k3l3_slim_10k100k_seed505/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/benzene
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/benzene/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.8172615989393406
+  flow_prior_mu: 0.008670088241267512
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 505
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 2
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_benzene_k3l3_slim_10k100k_seed505
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_benzene_k3l3_slim_10k100k_seed505/epoch=872-step=109872.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ee9b9966635023e8fad9ce3e75cb9ab9c4e1ce1d5007b20c7dd95a2b62901b7
+size 154820962

md_benzene_k3l3_slim_10k100k_seed505/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.049426741898059845, "target_var": 0.09023545682430267, "avg_num_neighbors": 11, "unique_atom_types": [0, 1, 6]}

md_ethane_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,191 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethane
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.6557021047032014
+  flow_prior_mu: 0.008435532788597242
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: all
+      log_freq: 100
+core:
+  expname: md_ethane_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethane_k3l3_slim_10k100k/epoch=440-step=55000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51067711ace62d1214d47c14504b1da7a21e9a3cf91f531895d8419de6712090
+size 154820260

md_ethane_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.02306615188717842, "target_var": 0.039020102471113205, "avg_num_neighbors": 7, "unique_atom_types": [0, 1, 6]}

md_ethane_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethane
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.6557021047032014
+  flow_prior_mu: 0.008435532788597242
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_ethane_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethane_k3l3_slim_10k100k_seed42/epoch=829-step=103625.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d91ce76b904463ced039346bb8f9e07e7bee59bb74e42d539860791c9669507
+size 154820962

md_ethane_k3l3_slim_10k100k_seed42/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.02306615188717842, "target_var": 0.039020102471113205, "avg_num_neighbors": 7, "unique_atom_types": [0, 1, 6]}

md_ethane_k3l3_slim_10k100k_seed505/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethane
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethane/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.6557021047032014
+  flow_prior_mu: 0.008435532788597242
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 505
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 2
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_ethane_k3l3_slim_10k100k_seed505
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethane_k3l3_slim_10k100k_seed505/epoch=429-step=107250.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf78c4065d59b7ec8904098082fecc46f38ce35da8755530c115851b5bfc1f2d
+size 154820962

md_ethane_k3l3_slim_10k100k_seed505/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.02306615188717842, "target_var": 0.039020102471113205, "avg_num_neighbors": 7, "unique_atom_types": [0, 1, 6]}

md_ethanol_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethanol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.0378894564782692
+  flow_prior_mu: 0.005635817718369937
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_ethanol_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethanol_k3l3_slim_10k100k/epoch=1748-step=110000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c083b8317177438ba0cef691baa499b28124ca88869f11002a2edcb0566ac9b
+size 154837998

md_ethanol_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.03295100852847099, "target_var": 0.09332586824893951, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_ethanol_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethanol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.0378894564782692
+  flow_prior_mu: 0.005635817718369937
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_ethanol_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethanol_k3l3_slim_10k100k_seed42/epoch=1748-step=110000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:423b04f267c2a83d8c67e32108598b61d74ce5e9b6af7f49ed406cae85dd2473
+size 154837998

md_ethanol_k3l3_slim_10k100k_seed42/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.03295100852847099, "target_var": 0.09332586824893951, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_ethanol_k3l3_slim_10k100k_seed505/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/ethanol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/ethanol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 3.0378894564782692
+  flow_prior_mu: 0.005635817718369937
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 505
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 2
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_ethanol_k3l3_slim_10k100k_seed505
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_ethanol_k3l3_slim_10k100k_seed505/epoch=434-step=54684.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b60101e69ed3c4360d32402406e3296228345b62b06ee2d39d7d84e75ee717c9
+size 154836454

md_ethanol_k3l3_slim_10k100k_seed505/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.03295100852847099, "target_var": 0.09332586824893951, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_malonaldehyde_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,191 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/malonaldehyde
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 2.7930136827969125
+  flow_prior_mu: 0.017925494043203455
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: all
+      log_freq: 100
+core:
+  expname: md_malonaldehyde_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_malonaldehyde_k3l3_slim_10k100k/epoch=148-step=18500.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c43acd17309b1623fcd3186b9f2b47cb5a52909d43c82647c19df27185cbe1fa
+size 154836262

md_malonaldehyde_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.04613230377435684, "target_var": 0.15922808647155762, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_malonaldehyde_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/malonaldehyde
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 2.7930136827969125
+  flow_prior_mu: 0.017925494043203455
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_malonaldehyde_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_malonaldehyde_k3l3_slim_10k100k_seed42/epoch=320-step=40000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8823014baa4602a894e42f5ebc8d98fa388f821b63dfbdf0166c77e1557a1579
+size 154836454

md_malonaldehyde_k3l3_slim_10k100k_seed42/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.04613230377435684, "target_var": 0.15922808647155762, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_malonaldehyde_k3l3_slim_10k100k_seed505/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/malonaldehyde
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/malonaldehyde/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 2.7930136827969125
+  flow_prior_mu: 0.017925494043203455
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 505
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 2
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_malonaldehyde_k3l3_slim_10k100k_seed505
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_malonaldehyde_k3l3_slim_10k100k_seed505/epoch=108-step=27000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e279e4d237d3ff3d23fa315a615286b10456fbecee8dc67033004f94719dcac
+size 154836454

md_malonaldehyde_k3l3_slim_10k100k_seed505/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.04613230377435684, "target_var": 0.15922808647155762, "avg_num_neighbors": 7, "unique_atom_types": [8, 1, 6, 0]}

md_phenol_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/phenol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.9846056192284764
+  flow_prior_mu: 0.005037187932829464
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_phenol_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_phenol_k3l3_slim_10k100k/epoch=1748-step=110000.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37fa2defb2da0615e627725c16261cde54b7fe837aadca4a3f7201d59c56b5c7
+size 154837934

md_phenol_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.059311117976903915, "target_var": 0.1463594138622284, "avg_num_neighbors": 12, "unique_atom_types": [8, 1, 6, 0]}

md_phenol_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/phenol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.9846056192284764
+  flow_prior_mu: 0.005037187932829464
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_phenol_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_phenol_k3l3_slim_10k100k_seed42/epoch=712-step=44856-v1.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f06f639fa3ccfcd032f5e245a8fe10aae8219220256d3131ebe14956c40426a
+size 154838514

md_phenol_k3l3_slim_10k100k_seed42/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.059311117976903915, "target_var": 0.1463594138622284, "avg_num_neighbors": 12, "unique_atom_types": [8, 1, 6, 0]}

md_phenol_k3l3_slim_10k100k_seed505/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/phenol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/phenol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 1.9846056192284764
+  flow_prior_mu: 0.005037187932829464
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 505
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 2
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_phenol_k3l3_slim_10k100k_seed505
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_phenol_k3l3_slim_10k100k_seed505/epoch=477-step=60102.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf856c66cb543c94a6f5b37b8105c6bae357487ebc8d3f911e8ed097577d8883
+size 154836454

md_phenol_k3l3_slim_10k100k_seed505/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.059311117976903915, "target_var": 0.1463594138622284, "avg_num_neighbors": 12, "unique_atom_types": [8, 1, 6, 0]}

md_resorcinol_k3l3_slim_10k100k/config.yaml

@@ -0,0 +1,191 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/resorcinol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/resorcinol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/resorcinol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 2.0182929296544465
+  flow_prior_mu: 0.004724867169487745
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 0
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_resorcinol_k3l3_slim_10k100k
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'

md_resorcinol_k3l3_slim_10k100k/epoch=698-step=43974.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b8c7b823cf4c1390b16482e2d5cf64cf8ffd5d4cc0edfa37fa190c56884f68d
+size 154836262

md_resorcinol_k3l3_slim_10k100k/metadata.json

@@ -0,0 +1 @@
+{"target_mean": 0.06919773668050766, "target_var": 0.20266202092170715, "avg_num_neighbors": 13, "unique_atom_types": [8, 1, 6, 0]}

md_resorcinol_k3l3_slim_10k100k_seed42/config.yaml

@@ -0,0 +1,196 @@
+data:
+  _target_: scdp.data.datamodule.ProbeDataModule
+  dataset:
+    _target_: scdp.data.dataset.LmdbDataset
+    path: ${oc.env:DATAPATH}/MD/scdp_lmdb/resorcinol
+    gt_coeffs_path: ${oc.env:DATAPATH}/MD/scdp_lmdb_gt_ridge_beta1.3/resorcinol
+    sad_coeffs_path: null
+  split_file: ${oc.env:DATAPATH}/MD/scdp_lmdb/resorcinol/datasplits.json
+  num_workers:
+    train: 8
+    val: 8
+    test: 8
+  batch_size:
+    train: 4
+    val: 4
+    test: 4
+  n_probe:
+    train: 125000
+    val: 125000
+    test: 125000
+model:
+  model:
+    _target_: scdp.model.equiformer_v3.EquiformerV3
+    enable_flow_coeff_entangle: true
+    flow_bridge_time_dim: ${..flow_time_dim}
+    max_num_elements: 100
+    num_layers: 3
+    lmax_list:
+    - 3
+    mmax_list:
+    - 2
+    cutoff: 6.0
+    sphere_channels: 96
+    hidden_channels: 192
+    edge_channels: 96
+    num_radial_basis: 300
+    distance_resolution: 0.02
+    basis_width_scalar: 2.0
+    attn_hidden_channels: 64
+    num_heads: 8
+    attn_alpha_channels: 32
+    attn_value_channels: 16
+    attn_grid_resolution_list:
+    - 20
+    - 8
+    ffn_grid_resolution_list:
+    - 20
+    - 20
+    norm_type: merge_layer_norm
+    use_atom_edge_embedding: true
+    use_envelope: true
+    use_attn_renorm: true
+    use_add_merge: false
+    use_grid_mlp: true
+    attn_activation: sep-merge_gates2_swiglu
+    ffn_activation: sep-merge_gates2_swiglu
+    alpha_drop: 0.0
+    attn_mask_rate: 0.0
+    attn_weights_drop: 0.0
+    value_drop: 0.0
+    drop_path_rate: 0.0
+    proj_drop: 0.0
+    ffn_drop: 0.0
+    gradient_checkpointing_block_list: null
+  _target_: scdp.model.module_flow.ChgFlowMatchingModule
+  criterion: mae
+  vnode_elem: 8
+  pbc: false
+  dft_basis_set: def2-qzvppd
+  dft_wt_aug: true
+  concat_basis: false
+  beta: 1.3
+  lmax_restriction: true
+  lmax_relax: 0
+  uncontracted: true
+  orb_cutoff: 5.0
+  expo_trainable: false
+  magnitude_weighting: false
+  density_synthesis_mode: linear
+  density_eval_mode: realspace
+  hybrid_fft_zeta_max: 5.0
+  hybrid_fft_phase_sign: -1.0
+  hybrid_fft_chunk_G: 65536
+  flow_velocity_mlp_hidden: 512
+  flow_time_dim: 64
+  flow_head: escn
+  flow_e3nn_hidden_multiplier: 2
+  bridge_mode: cfm
+  flow_loss_mode: endpoint
+  flow_endpoint_loss: integrated_density
+  flow_combined_integrated_weight: 1.0
+  flow_target_mode: gt
+  flow_use_t_scale: false
+  flow_t_scale_max: 0.9
+  flow_prior_mode: irrep_gaussian
+  flow_prior_sigma: 2.0182929296544465
+  flow_prior_mu: 0.004724867169487745
+  flow_prior_nelectron_normalize: false
+  flow_nelectron_mc_max_probes: 0
+  flow_val_integrate_steps: 2
+  flow_val_integrate_sweep_steps:
+  - 1
+  - 2
+  - 3
+  flow_integrate_train_steps: 2
+  flow_integrate_min_t: 0.01
+  flow_integrate_schedule: uniform
+  flow_integrate_t_hi: 0.99
+  flow_integrate_power: 4.0
+  flow_val_max_n_probe_per_pass: 400000
+  flow_val_log_readout: false
+  sb_sigma: 0.1
+  sb_noise_mode: isotropic
+  sb_per_l_sigma_mode: inv_sqrt_2lp1
+  sb_per_l_sigma_scales: null
+train:
+  seed: 42
+  deterministic: true
+  run_post_fit_test: false
+  trainer:
+    fast_dev_run: false
+    accelerator: gpu
+    strategy: ddp_find_unused_parameters_true
+    devices: 4
+    precision: 32
+    max_epochs: 100000
+    max_steps: 110000
+    num_sanity_val_steps: 2
+    gradient_clip_val: 0.5
+    val_check_interval: 1.0
+    deterministic: ${train.deterministic}
+  optim:
+    _target_: torch.optim.Adam
+    lr: 0.001
+    betas:
+    - 0.9
+    - 0.999
+    eps: 1.0e-08
+    weight_decay: 0.0
+  lr_scheduler:
+    _target_: scdp.model.utils.PowerDecayScheduler
+    warmup_steps: 1000
+    alpha: 0.96
+    beta: 4433
+  lr_schedule_freq: 1
+  ema:
+    decay: 0.995
+  restore:
+    ckpt_or_run_path: null
+    mode: null
+    weights_only: false
+  monitor:
+    metric: nmape/val_flow
+    mode: min
+  callbacks:
+  - _target_: lightning.pytorch.callbacks.EarlyStopping
+    patience: 10000
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.ModelCheckpoint
+    dirpath: ${core.storage_dir}
+    save_top_k: 2
+    save_last: true
+    verbose: false
+    monitor: ${train.monitor.metric}
+    mode: ${train.monitor.mode}
+  - _target_: lightning.pytorch.callbacks.LearningRateMonitor
+    logging_interval: step
+    log_momentum: false
+  - _target_: lightning.pytorch.callbacks.progress.tqdm_progress.TQDMProgressBar
+    refresh_rate: 20
+  logging:
+    upload:
+      run_files: false
+      source: true
+    wandb:
+      name: ${core.expname}
+      project: ${core.project_name}
+      entity: null
+      log_model: ${..upload.run_files}
+      mode: online
+      tags: ${core.tags}
+      resume: allow
+    wandb_watch:
+      log: null
+      log_freq: 100
+core:
+  expname: md_resorcinol_k3l3_slim_10k100k_seed42
+  project_name: orbflow
+  storage_dir: ${core.expname}
+  version: 0.0.1
+  tags: null
+conventions:
+  x_key: x
+  y_key: 'y'