-
Notifications
You must be signed in to change notification settings - Fork 502
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
docs(pt): examples for new dpa2 model (#4138)
small: 3 layers; w three-body; wo g2 attn; medium: 6 layers; w three-body; w g2 attn; large: 12 layers; w three-body; w g2 attn; <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit - **New Features** - Introduced comprehensive JSON configuration files for the DPA2 model, enhancing setup for molecular simulations. - Added detailed README documentation outlining model configurations and input files, aiding user selection based on precision and efficiency needs. - Added parameters for three-body interactions to improve model accuracy. - Configured learning rate settings and loss function preferences for better training dynamics. - **Bug Fixes** - Expanded test coverage by including multiple input file variations for the DPA2 example, ensuring more robust testing. - **Documentation** - Updated training example reference for clarity and included links to README for input variations. <!-- end of auto-generated comment: release notes by coderabbit.ai -->
- Loading branch information
Showing
6 changed files
with
260 additions
and
9 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,15 @@ | ||
| ## Inputs for DPA-2 model | ||
|
|
||
| This directory contains the input files for training the DPA-2 model (currently supporting PyTorch backend only). Depending on your precision/efficiency requirements, we provide three different levels of model complexity: | ||
|
|
||
| - `input_torch_small.json`: Our smallest DPA-2 model, optimized for speed. | ||
| - `input_torch_medium.json` (Recommended): Our well-performing DPA-2 model, balancing efficiency and precision. This is a good starting point for most users. | ||
| - `input_torch_large.json`: Our most complex model with the highest precision, suitable for very intricate data structures. | ||
|
|
||
| For detailed differences in their configurations, please refer to the table below: | ||
|
|
||
| | Input | Repformer layers | Three-body embedding in Repinit | Pair-wise attention in Repformer | Tuned sub-structures in [#4089](https://github.com/deepmodeling/deepmd-kit/pull/4089) | Description | | ||
| | ------------------------- | ---------------- | ------------------------------- | -------------------------------- | ------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- | | ||
| | `input_torch_small.json` | 3 | ✓ | ✗ | ✓ | Smallest DPA-2 model, optimized for speed. | | ||
| | `input_torch_medium.json` | 6 | ✓ | ✓ | ✓ | Recommended well-performing DPA-2 model, balancing efficiency and precision. | | ||
| | `input_torch_large.json` | 12 | ✓ | ✓ | ✓ | Most complex model with the highest precision. | |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,112 @@ | ||
| { | ||
| "_comment": "that's all", | ||
| "model": { | ||
| "type_map": [ | ||
| "O", | ||
| "H" | ||
| ], | ||
| "descriptor": { | ||
| "type": "dpa2", | ||
| "repinit": { | ||
| "tebd_dim": 8, | ||
| "rcut": 6.0, | ||
| "rcut_smth": 0.5, | ||
| "nsel": 120, | ||
| "neuron": [ | ||
| 25, | ||
| 50, | ||
| 100 | ||
| ], | ||
| "axis_neuron": 12, | ||
| "activation_function": "tanh", | ||
| "three_body_sel": 40, | ||
| "three_body_rcut": 4.0, | ||
| "three_body_rcut_smth": 3.5, | ||
| "use_three_body": true | ||
| }, | ||
| "repformer": { | ||
| "rcut": 4.0, | ||
| "rcut_smth": 3.5, | ||
| "nsel": 40, | ||
| "nlayers": 6, | ||
| "g1_dim": 128, | ||
| "g2_dim": 32, | ||
| "attn2_hidden": 32, | ||
| "attn2_nhead": 4, | ||
| "attn1_hidden": 128, | ||
| "attn1_nhead": 4, | ||
| "axis_neuron": 4, | ||
| "update_h2": false, | ||
| "update_g1_has_conv": true, | ||
| "update_g1_has_grrg": true, | ||
| "update_g1_has_drrd": true, | ||
| "update_g1_has_attn": false, | ||
| "update_g2_has_g1g1": false, | ||
| "update_g2_has_attn": true, | ||
| "update_style": "res_residual", | ||
| "update_residual": 0.01, | ||
| "update_residual_init": "norm", | ||
| "attn2_has_gate": true, | ||
| "use_sqrt_nnei": true, | ||
| "g1_out_conv": true, | ||
| "g1_out_mlp": true | ||
| }, | ||
| "add_tebd_to_repinit_out": false | ||
| }, | ||
| "fitting_net": { | ||
| "neuron": [ | ||
| 240, | ||
| 240, | ||
| 240 | ||
| ], | ||
| "resnet_dt": true, | ||
| "seed": 1, | ||
| "_comment": " that's all" | ||
| }, | ||
| "_comment": " that's all" | ||
| }, | ||
| "learning_rate": { | ||
| "type": "exp", | ||
| "decay_steps": 5000, | ||
| "start_lr": 0.001, | ||
| "stop_lr": 3.51e-08, | ||
| "_comment": "that's all" | ||
| }, | ||
| "loss": { | ||
| "type": "ener", | ||
| "start_pref_e": 0.02, | ||
| "limit_pref_e": 1, | ||
| "start_pref_f": 1000, | ||
| "limit_pref_f": 1, | ||
| "start_pref_v": 0, | ||
| "limit_pref_v": 0, | ||
| "_comment": " that's all" | ||
| }, | ||
| "training": { | ||
| "stat_file": "./dpa2.hdf5", | ||
| "training_data": { | ||
| "systems": [ | ||
| "../data/data_0", | ||
| "../data/data_1", | ||
| "../data/data_2" | ||
| ], | ||
| "batch_size": 1, | ||
| "_comment": "that's all" | ||
| }, | ||
| "validation_data": { | ||
| "systems": [ | ||
| "../data/data_3" | ||
| ], | ||
| "batch_size": 1, | ||
| "_comment": "that's all" | ||
| }, | ||
| "numb_steps": 1000000, | ||
| "warmup_steps": 0, | ||
| "gradient_max_norm": 5.0, | ||
| "seed": 10, | ||
| "disp_file": "lcurve.out", | ||
| "disp_freq": 100, | ||
| "save_freq": 2000, | ||
| "_comment": "that's all" | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,112 @@ | ||
| { | ||
| "_comment": "that's all", | ||
| "model": { | ||
| "type_map": [ | ||
| "O", | ||
| "H" | ||
| ], | ||
| "descriptor": { | ||
| "type": "dpa2", | ||
| "repinit": { | ||
| "tebd_dim": 8, | ||
| "rcut": 6.0, | ||
| "rcut_smth": 0.5, | ||
| "nsel": 120, | ||
| "neuron": [ | ||
| 25, | ||
| 50, | ||
| 100 | ||
| ], | ||
| "axis_neuron": 12, | ||
| "activation_function": "tanh", | ||
| "three_body_sel": 40, | ||
| "three_body_rcut": 4.0, | ||
| "three_body_rcut_smth": 3.5, | ||
| "use_three_body": true | ||
| }, | ||
| "repformer": { | ||
| "rcut": 4.0, | ||
| "rcut_smth": 3.5, | ||
| "nsel": 40, | ||
| "nlayers": 3, | ||
| "g1_dim": 128, | ||
| "g2_dim": 32, | ||
| "attn2_hidden": 32, | ||
| "attn2_nhead": 4, | ||
| "attn1_hidden": 128, | ||
| "attn1_nhead": 4, | ||
| "axis_neuron": 4, | ||
| "update_h2": false, | ||
| "update_g1_has_conv": true, | ||
| "update_g1_has_grrg": true, | ||
| "update_g1_has_drrd": true, | ||
| "update_g1_has_attn": false, | ||
| "update_g2_has_g1g1": false, | ||
| "update_g2_has_attn": false, | ||
| "update_style": "res_residual", | ||
| "update_residual": 0.01, | ||
| "update_residual_init": "norm", | ||
| "attn2_has_gate": true, | ||
| "use_sqrt_nnei": true, | ||
| "g1_out_conv": true, | ||
| "g1_out_mlp": true | ||
| }, | ||
| "add_tebd_to_repinit_out": false | ||
| }, | ||
| "fitting_net": { | ||
| "neuron": [ | ||
| 240, | ||
| 240, | ||
| 240 | ||
| ], | ||
| "resnet_dt": true, | ||
| "seed": 1, | ||
| "_comment": " that's all" | ||
| }, | ||
| "_comment": " that's all" | ||
| }, | ||
| "learning_rate": { | ||
| "type": "exp", | ||
| "decay_steps": 5000, | ||
| "start_lr": 0.001, | ||
| "stop_lr": 3.51e-08, | ||
| "_comment": "that's all" | ||
| }, | ||
| "loss": { | ||
| "type": "ener", | ||
| "start_pref_e": 0.02, | ||
| "limit_pref_e": 1, | ||
| "start_pref_f": 1000, | ||
| "limit_pref_f": 1, | ||
| "start_pref_v": 0, | ||
| "limit_pref_v": 0, | ||
| "_comment": " that's all" | ||
| }, | ||
| "training": { | ||
| "stat_file": "./dpa2.hdf5", | ||
| "training_data": { | ||
| "systems": [ | ||
| "../data/data_0", | ||
| "../data/data_1", | ||
| "../data/data_2" | ||
| ], | ||
| "batch_size": 1, | ||
| "_comment": "that's all" | ||
| }, | ||
| "validation_data": { | ||
| "systems": [ | ||
| "../data/data_3" | ||
| ], | ||
| "batch_size": 1, | ||
| "_comment": "that's all" | ||
| }, | ||
| "numb_steps": 1000000, | ||
| "warmup_steps": 0, | ||
| "gradient_max_norm": 5.0, | ||
| "seed": 10, | ||
| "disp_file": "lcurve.out", | ||
| "disp_freq": 100, | ||
| "save_freq": 2000, | ||
| "_comment": "that's all" | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters