Update README.md
Browse files
README.md
CHANGED
|
@@ -20,60 +20,6 @@ Citation update is coming up...
|
|
| 20 |
}
|
| 21 |
```
|
| 22 |
***
|
| 23 |
-
#
|
| 24 |
-
*
|
| 25 |
-
*
|
| 26 |
-
* [numpy](https://github.com/numpy/numpy)
|
| 27 |
-
* [bgmol](https://github.com/noegroup/bgmol) (for Chignolin)
|
| 28 |
-
* [bgflow](https://github.com/noegroup/bgmol) (for Chignolin)
|
| 29 |
-
* [mdtraj=1.9.9](https://github.com/mdtraj/mdtraj) (for Chignolin)
|
| 30 |
-
|
| 31 |
-
## Training
|
| 32 |
-
* All hyper-parameters and training details are provided in config files (), and free feel to tune these parameters../configs/*.yml
|
| 33 |
-
|
| 34 |
-
```bash
|
| 35 |
-
python train.py ./configs/GMM10D_default.yml ./models/GMM10D
|
| 36 |
-
python train.py ./configs/GMM1000D_default.yml ./models/GMM1000D
|
| 37 |
-
python train.py ./configs/CGN_default.yml ./models/CGN
|
| 38 |
-
```
|
| 39 |
-
|
| 40 |
-
* The model checkpoints will be saved in the specified directory,e.g., ./models/GMM10D
|
| 41 |
-
|
| 42 |
-
## Metropolis Monte Carlo (MC) simulations
|
| 43 |
-
* The Metropolis MC simulations are provided in the following files:
|
| 44 |
-
|
| 45 |
-
```bash
|
| 46 |
-
python MC.py ./configs/CGN_default.yml ./models/CGN --method 0 --eps_type Rademacher
|
| 47 |
-
```
|
| 48 |
-
|
| 49 |
-
* The first argument specifies the configuration file, while the second argument indicates the model directory. The `--method` option determines the approach for running the Monte Carlo (MC) simulations:
|
| 50 |
-
- `0`: Flow Perturbation (FP)
|
| 51 |
-
- `-1`: Jacobian-based method
|
| 52 |
-
- `-2`: Stochastic Normalizing Flow (SNF)
|
| 53 |
-
- `1-n`: Hutchinson trace estimator
|
| 54 |
-
- `--eps_type` parameter: Defines the type of perturbation to use, such as Rademacher, Gaussian, etc.
|
| 55 |
-
|
| 56 |
-
## Sequential Monte Carlo (SMC)
|
| 57 |
-
* The Sequential Monte Carlo in the following files:
|
| 58 |
-
```bash
|
| 59 |
-
python SMC.py ./configs/GMM10D_default.yml ./models/GMM10D --method 0 --eps_type Rademacher
|
| 60 |
-
python SMC.py ./configs/GMM1000D_default.yml ./models/GMM1000D --method 0 --eps_type Rademacher
|
| 61 |
-
python SMC.py ./configs/CGN_default.yml ./models/CGN --method 0 --eps_type Rademacher
|
| 62 |
-
```
|
| 63 |
-
* The first argument specifies the configuration file, while the second argument indicates the model directory. The `--method` option determines the approach for running the Sequential Monte Carlo (SMC) simulations:
|
| 64 |
-
- `0`: Flow Perturbation (FP)
|
| 65 |
-
- `-1`: Jacobian-based method
|
| 66 |
-
- `-2`: Stochastic Normalizing Flow (SNF)
|
| 67 |
-
- `1-n`: Hutchinson trace estimator
|
| 68 |
-
- `--eps_type` parameter: Defines the type of perturbation to use, such as Rademacher, Gaussian, etc.
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
## Model and Data
|
| 72 |
-
* The result datasets presented in the paper can be found in the `Result` folder of this repository.
|
| 73 |
-
|
| 74 |
-
* We provide all pre-trained model checkpoints as well as other result datasets from the paper. These resources are hosted on Hugging Face:
|
| 75 |
-
* [Model and Result Folder](https://huggingface.co/XinPeng76/Flow_Perturbation)
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
## [License](#dependencies)
|
| 79 |
-
[MIT License](LICENSE)
|
|
|
|
| 20 |
}
|
| 21 |
```
|
| 22 |
***
|
| 23 |
+
# Project GitHub Repository
|
| 24 |
+
* The GitHub repository for this project is available at:
|
| 25 |
+
* https://github.com/XinPeng76/Flow_Perturbation
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|