From fc628cb792beb5c39d095de5bda8209f00a6795e Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 14:51:28 +0800 Subject: [PATCH 01/10] Add phylstm hydra --- docs/zh/examples/phylstm.md | 28 ++-- examples/phylstm/conf/phylstm2.yaml | 39 ++++++ examples/phylstm/conf/phylstm3.yaml | 39 ++++++ examples/phylstm/phylstm2.py | 184 ++++++++++++++++++++++--- examples/phylstm/phylstm3.py | 203 +++++++++++++++++++++++----- 5 files changed, 425 insertions(+), 68 deletions(-) create mode 100644 examples/phylstm/conf/phylstm2.yaml create mode 100644 examples/phylstm/conf/phylstm3.yaml diff --git a/docs/zh/examples/phylstm.md b/docs/zh/examples/phylstm.md index fefb384bb..4e2703365 100644 --- a/docs/zh/examples/phylstm.md +++ b/docs/zh/examples/phylstm.md @@ -29,9 +29,9 @@ $$ 在 PhyLSTM 问题中，建立 LSTM 网络 Deep LSTM network，用 PaddleScience 代码表示如下 -``` py linenums="105" +``` py linenums="101" --8<-- -examples/phylstm/phylstm2.py:105:106 +examples/phylstm/phylstm2.py:101:106 --8<-- ``` @@ -47,9 +47,9 @@ wget -P ./ https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_ 本案例涉及读取数据构建，如下所示 -``` py linenums="38" +``` py linenums="36" --8<-- -examples/phylstm/phylstm2.py:38:105 +examples/phylstm/phylstm2.py:36:99 --8<-- ``` @@ -57,9 +57,9 @@ examples/phylstm/phylstm2.py:38:105 设置训练数据集和损失计算函数，返回字段，代码如下所示： -``` py linenums="118" +``` py linenums="116" --8<-- -examples/phylstm/phylstm2.py:118:136 +examples/phylstm/phylstm2.py:116:135 --8<-- ``` @@ -67,9 +67,9 @@ examples/phylstm/phylstm2.py:118:136 设置评估数据集和损失计算函数，返回字段，代码如下所示： -``` py linenums="139" +``` py linenums="137" --8<-- -examples/phylstm/phylstm2.py:139:158 +examples/phylstm/phylstm2.py:137:157 --8<-- ``` @@ -77,9 +77,9 @@ examples/phylstm/phylstm2.py:139:158 接下来我们需要指定训练轮数，此处我们按实验经验，使用 100 轮训练轮数。 -``` py linenums="36" +``` py linenums="167" --8<-- -examples/phylstm/phylstm2.py:36:36 +examples/phylstm/phylstm2.py:167:167 --8<-- ``` @@ -87,9 +87,9 @@ examples/phylstm/phylstm2.py:36:36 训练过程会调用优化器来更新模型参数，此处选择 `Adam` 优化器并设定 `learning_rate` 为 1e-3。 -``` py linenums="163" +``` py linenums="160" --8<-- -examples/phylstm/phylstm2.py:163:163 +examples/phylstm/phylstm2.py:160:160 --8<-- ``` @@ -97,9 +97,9 @@ examples/phylstm/phylstm2.py:163:163 完成上述设置之后，只需要将上述实例化的对象按顺序传递给 `ppsci.solver.Solver`。 -``` py linenums="164" +``` py linenums="161" --8<-- -examples/phylstm/phylstm2.py:164:175 +examples/phylstm/phylstm2.py:161:175 --8<-- ``` diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml new file mode 100644 index 000000000..dfd207732 --- /dev/null +++ b/examples/phylstm/conf/phylstm2.yaml @@ -0,0 +1,39 @@ +hydra: + run: + # dynamic output directory according to running time and override name + dir: outputs_PhyLSTM2/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname} + job: + name: ${mode} # name of logfile + chdir: false # keep current working direcotry unchaned + sweep: + # output directory for multirun + dir: ${hydra.run.dir} + subdir: ./ + +# general settings +mode: train # running mode: train/eval +seed: 42 +output_dir: ${hydra:run.dir} +log_freq: 20 +data_file: data_boucwen.mat + +# model settings +MODEL: + phylstm2_net: + input_size: 1 + hidden_size: 100 + model_type: 2 + +# training settings +TRAIN: + epochs: 100 + iters_per_epoch: 1 + save_freq: 50 + pretrained_model_path: null + checkpoint_path: null + learning_rate: 0.001 + +# evaluation settings +EVAL: + pretrained_model_path: null + eval_with_no_grad: true \ No newline at end of file diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml new file mode 100644 index 000000000..b88b63b47 --- /dev/null +++ b/examples/phylstm/conf/phylstm3.yaml @@ -0,0 +1,39 @@ +hydra: + run: + # dynamic output directory according to running time and override name + dir: outputs_PhyLSTM3/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname} + job: + name: ${mode} # name of logfile + chdir: false # keep current working direcotry unchaned + sweep: + # output directory for multirun + dir: ${hydra.run.dir} + subdir: ./ + +# general settings +mode: train # running mode: train/eval +seed: 42 +output_dir: ${hydra:run.dir} +log_freq: 20 +data_file: data_boucwen.mat + +# model settings +MODEL: + phylstm3_net: + input_size: 1 + hidden_size: 100 + model_type: 3 + +# training settings +TRAIN: + epochs: 200 + iters_per_epoch: 1 + save_freq: 50 + pretrained_model_path: null + checkpoint_path: null + learning_rate: 0.001 + +# evaluation settings +EVAL: + pretrained_model_path: null + eval_with_no_grad: true \ No newline at end of file diff --git a/examples/phylstm/phylstm2.py b/examples/phylstm/phylstm2.py index 0dcefa957..3e0f9373f 100755 --- a/examples/phylstm/phylstm2.py +++ b/examples/phylstm/phylstm2.py @@ -16,26 +16,25 @@ Reference: https://github.com/zhry10/PhyLSTM.git """ +from os import path as osp + import functions +import hydra import numpy as np import scipy.io +from omegaconf import DictConfig import ppsci -from ppsci.utils import config from ppsci.utils import logger -if __name__ == "__main__": - args = config.parse_args() + +def train(cfg: DictConfig): # set random seed for reproducibility - ppsci.utils.misc.set_random_seed(42) - # set output directory - OUTPUT_DIR = "./output_PhyLSTM2" if not args.output_dir else args.output_dir + ppsci.utils.misc.set_random_seed(cfg.seed) # initialize logger - logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info") - # set training hyper-parameters - EPOCHS = 100 if not args.epochs else args.epochs + logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") - mat = scipy.io.loadmat("data_boucwen.mat") + mat = scipy.io.loadmat(cfg.data_file) ag_data = mat["input_tf"] # ag, ad, av u_data = mat["target_X_tf"] ut_data = mat["target_Xd_tf"] @@ -90,9 +89,6 @@ eta_tt_star = u_tt_all[0:10] ag_c_star = ag_all[0:50] lift_star = -ag_c_star - eta_c_star = u_all[0:50] - eta_t_c_star = u_t_all[0:50] - eta_tt_c_star = u_tt_all[0:50] eta = eta_star ag = ag_star @@ -103,7 +99,12 @@ g = -eta_tt - ag phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) - model = ppsci.arch.DeepPhyLSTM(1, eta.shape[2], 100, 2) + model = ppsci.arch.DeepPhyLSTM( + cfg.MODEL.phylstm2_net.input_size, + eta.shape[2], + cfg.MODEL.phylstm2_net.hidden_size, + cfg.MODEL.phylstm2_net.model_type, + ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -113,7 +114,7 @@ label_dict_train, input_dict_val, label_dict_val, - ) = dataset_obj.get(EPOCHS) + ) = dataset_obj.get(cfg.TRAIN.epochs) sup_constraint_pde = ppsci.constraint.SupervisedConstraint( { @@ -159,22 +160,161 @@ validator_pde = {sup_validator_pde.name: sup_validator_pde} # initialize solver - ITERS_PER_EPOCH = 1 - optimizer = ppsci.optimizer.Adam(1e-3)(model) + optimizer = ppsci.optimizer.Adam(cfg.TRAIN.learning_rate)(model) solver = ppsci.solver.Solver( model, constraint_pde, - OUTPUT_DIR, + cfg.output_dir, optimizer, None, - EPOCHS, - ITERS_PER_EPOCH, - save_freq=50, + cfg.TRAIN.epochs, + cfg.TRAIN.iters_per_epoch, + save_freq=cfg.TRAIN.save_freq, + log_freq=cfg.log_freq, + seed=cfg.seed, validator=validator_pde, - eval_with_no_grad=True, + checkpoint_path=cfg.TRAIN.checkpoint_path, + eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) # train model solver.train() # evaluate after finished training solver.eval() + + +def evaluate(cfg: DictConfig): + # set random seed for reproducibility + ppsci.utils.misc.set_random_seed(cfg.seed) + # initialize logger + logger.init_logger("ppsci", osp.join(cfg.output_dir, "eval.log"), "info") + + mat = scipy.io.loadmat(cfg.data_file) + ag_data = mat["input_tf"] # ag, ad, av + u_data = mat["target_X_tf"] + ut_data = mat["target_Xd_tf"] + utt_data = mat["target_Xdd_tf"] + ag_data = ag_data.reshape([ag_data.shape[0], ag_data.shape[1], 1]) + u_data = u_data.reshape([u_data.shape[0], u_data.shape[1], 1]) + ut_data = ut_data.reshape([ut_data.shape[0], ut_data.shape[1], 1]) + utt_data = utt_data.reshape([utt_data.shape[0], utt_data.shape[1], 1]) + + t = mat["time"] + dt = t[0, 1] - t[0, 0] + + ag_all = ag_data + u_all = u_data + u_t_all = ut_data + u_tt_all = utt_data + + # finite difference + N = u_data.shape[1] + phi1 = np.concatenate( + [ + np.array([-3 / 2, 2, -1 / 2]), + np.zeros([N - 3]), + ] + ) + temp1 = np.concatenate([-1 / 2 * np.identity(N - 2), np.zeros([N - 2, 2])], axis=1) + temp2 = np.concatenate([np.zeros([N - 2, 2]), 1 / 2 * np.identity(N - 2)], axis=1) + phi2 = temp1 + temp2 + phi3 = np.concatenate( + [ + np.zeros([N - 3]), + np.array([1 / 2, -2, 3 / 2]), + ] + ) + phi_t0 = ( + 1 + / dt + * np.concatenate( + [ + np.reshape(phi1, [1, phi1.shape[0]]), + phi2, + np.reshape(phi3, [1, phi3.shape[0]]), + ], + axis=0, + ) + ) + phi_t0 = np.reshape(phi_t0, [1, N, N]) + + ag_star = ag_all[0:10] + eta_star = u_all[0:10] + eta_t_star = u_t_all[0:10] + eta_tt_star = u_tt_all[0:10] + ag_c_star = ag_all[0:50] + lift_star = -ag_c_star + + eta = eta_star + ag = ag_star + lift = lift_star + eta_t = eta_t_star + eta_tt = eta_tt_star + ag_c = ag_c_star + g = -eta_tt - ag + phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) + + model = ppsci.arch.DeepPhyLSTM( + cfg.MODEL.phylstm2_net.input_size, + eta.shape[2], + cfg.MODEL.phylstm2_net.hidden_size, + cfg.MODEL.phylstm2_net.model_type, + ) + model.register_input_transform(functions.transform_in) + model.register_output_transform(functions.transform_out) + + dataset_obj = functions.Dataset(eta, eta_t, g, ag, ag_c, lift, phi_t) + ( + _, + _, + input_dict_val, + label_dict_val, + ) = dataset_obj.get(1) + + sup_validator_pde = ppsci.validate.SupervisedValidator( + { + "dataset": { + "name": "NamedArrayDataset", + "input": input_dict_val, + "label": label_dict_val, + }, + }, + ppsci.loss.FunctionalLoss(functions.train_loss_func2), + { + "eta_pred": lambda out: out["eta_pred"], + "eta_dot_pred": lambda out: out["eta_dot_pred"], + "g_pred": lambda out: out["g_pred"], + "eta_t_pred_c": lambda out: out["eta_t_pred_c"], + "eta_dot_pred_c": lambda out: out["eta_dot_pred_c"], + "lift_pred_c": lambda out: out["lift_pred_c"], + }, + metric={"metric": ppsci.metric.FunctionalMetric(functions.metric_expr)}, + name="sup_valid", + ) + validator_pde = {sup_validator_pde.name: sup_validator_pde} + + # initialize solver + solver = ppsci.solver.Solver( + model, + output_dir=cfg.output_dir, + seed=cfg.seed, + validator=validator_pde, + checkpoint_path=cfg.TRAIN.checkpoint_path, + eval_with_no_grad=cfg.EVAL.eval_with_no_grad, + ) + # evaluate after finished training + solver.eval() + + +@hydra.main(version_base=None, config_path="./conf", config_name="phylstm2.yaml") +def main(cfg: DictConfig): + if cfg.mode == "train": + train(cfg) + elif cfg.mode == "eval": + evaluate(cfg) + else: + raise ValueError(f"cfg.mode should in ['train', 'eval'], but got '{cfg.mode}'") + + +if __name__ == "__main__": + main() diff --git a/examples/phylstm/phylstm3.py b/examples/phylstm/phylstm3.py index 718249f05..46d1996ab 100755 --- a/examples/phylstm/phylstm3.py +++ b/examples/phylstm/phylstm3.py @@ -16,27 +16,25 @@ Reference: https://github.com/zhry10/PhyLSTM.git """ +from os import path as osp + import functions +import hydra import numpy as np import scipy.io +from omegaconf import DictConfig import ppsci -from ppsci.utils import config from ppsci.utils import logger -if __name__ == "__main__": - args = config.parse_args() + +def train(cfg: DictConfig): # set random seed for reproducibility - ppsci.utils.misc.set_random_seed(42) - # set output directory - OUTPUT_DIR = "./output_PhyLSTM3" if not args.output_dir else args.output_dir + ppsci.utils.misc.set_random_seed(cfg.seed) # initialize logger - logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info") - # set training hyper-parameters - EPOCHS = 200 if not args.epochs else args.epochs - - mat = scipy.io.loadmat("data_boucwen.mat") + logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") + mat = scipy.io.loadmat(cfg.data_file) t = mat["time"] dt = 0.02 n1 = int(dt / 0.005) @@ -96,9 +94,6 @@ eta_tt_star = utt_data ag_c_star = np.concatenate([ag_data, ag_pred[0:53]]) lift_star = -ag_c_star - eta_c_star = np.concatenate([u_data, u_pred[0:53]]) - eta_t_c_star = np.concatenate([ut_data, ut_pred[0:53]]) - eta_tt_c_star = np.concatenate([utt_data, utt_pred[0:53]]) eta = eta_star ag = ag_star @@ -108,18 +103,14 @@ g = -eta_tt - ag ag_c = ag_c_star - # Training Data - eta_train = eta - ag_train = ag - lift_train = lift - eta_t_train = eta_t - eta_tt_train = eta_tt - g_train = g - ag_c_train = ag_c - phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) - model = ppsci.arch.DeepPhyLSTM(1, eta.shape[2], 100, 3) + model = ppsci.arch.DeepPhyLSTM( + cfg.MODEL.phylstm3_net.input_size, + eta.shape[2], + cfg.MODEL.phylstm3_net.hidden_size, + cfg.MODEL.phylstm3_net.model_type, + ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -129,7 +120,7 @@ label_dict_train, input_dict_val, label_dict_val, - ) = dataset_obj.get(EPOCHS) + ) = dataset_obj.get(cfg.TRAIN.epochs) sup_constraint_pde = ppsci.constraint.SupervisedConstraint( { @@ -179,22 +170,170 @@ validator_pde = {sup_validator_pde.name: sup_validator_pde} # initialize solver - ITERS_PER_EPOCH = 1 - optimizer = ppsci.optimizer.Adam(1e-3)(model) + optimizer = ppsci.optimizer.Adam(cfg.TRAIN.learning_rate)(model) solver = ppsci.solver.Solver( model, constraint_pde, - OUTPUT_DIR, + cfg.output_dir, optimizer, None, - EPOCHS, - ITERS_PER_EPOCH, - save_freq=50, + cfg.TRAIN.epochs, + cfg.TRAIN.iters_per_epoch, + save_freq=cfg.TRAIN.save_freq, + log_freq=cfg.log_freq, + seed=cfg.seed, validator=validator_pde, - eval_with_no_grad=True, + checkpoint_path=cfg.TRAIN.checkpoint_path, + eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) # train model solver.train() # evaluate after finished training solver.eval() + + +def evaluate(cfg: DictConfig): + # set random seed for reproducibility + ppsci.utils.misc.set_random_seed(cfg.seed) + # initialize logger + logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") + + mat = scipy.io.loadmat(cfg.data_file) + t = mat["time"] + dt = 0.02 + n1 = int(dt / 0.005) + t = t[::n1] + + ag_data = mat["input_tf"][:, ::n1] # ag, ad, av + u_data = mat["target_X_tf"][:, ::n1] + ut_data = mat["target_Xd_tf"][:, ::n1] + utt_data = mat["target_Xdd_tf"][:, ::n1] + ag_data = ag_data.reshape([ag_data.shape[0], ag_data.shape[1], 1]) + u_data = u_data.reshape([u_data.shape[0], u_data.shape[1], 1]) + ut_data = ut_data.reshape([ut_data.shape[0], ut_data.shape[1], 1]) + utt_data = utt_data.reshape([utt_data.shape[0], utt_data.shape[1], 1]) + + ag_pred = mat["input_pred_tf"][:, ::n1] # ag, ad, av + u_pred = mat["target_pred_X_tf"][:, ::n1] + ut_pred = mat["target_pred_Xd_tf"][:, ::n1] + utt_pred = mat["target_pred_Xdd_tf"][:, ::n1] + ag_pred = ag_pred.reshape([ag_pred.shape[0], ag_pred.shape[1], 1]) + u_pred = u_pred.reshape([u_pred.shape[0], u_pred.shape[1], 1]) + ut_pred = ut_pred.reshape([ut_pred.shape[0], ut_pred.shape[1], 1]) + utt_pred = utt_pred.reshape([utt_pred.shape[0], utt_pred.shape[1], 1]) + + N = u_data.shape[1] + phi1 = np.concatenate( + [ + np.array([-3 / 2, 2, -1 / 2]), + np.zeros([N - 3]), + ] + ) + temp1 = np.concatenate([-1 / 2 * np.identity(N - 2), np.zeros([N - 2, 2])], axis=1) + temp2 = np.concatenate([np.zeros([N - 2, 2]), 1 / 2 * np.identity(N - 2)], axis=1) + phi2 = temp1 + temp2 + phi3 = np.concatenate( + [ + np.zeros([N - 3]), + np.array([1 / 2, -2, 3 / 2]), + ] + ) + phi_t0 = ( + 1 + / dt + * np.concatenate( + [ + np.reshape(phi1, [1, phi1.shape[0]]), + phi2, + np.reshape(phi3, [1, phi3.shape[0]]), + ], + axis=0, + ) + ) + phi_t0 = np.reshape(phi_t0, [1, N, N]) + + ag_star = ag_data + eta_star = u_data + eta_t_star = ut_data + eta_tt_star = utt_data + ag_c_star = np.concatenate([ag_data, ag_pred[0:53]]) + lift_star = -ag_c_star + + eta = eta_star + ag = ag_star + lift = lift_star + eta_t = eta_t_star + eta_tt = eta_tt_star + g = -eta_tt - ag + ag_c = ag_c_star + + phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) + + model = ppsci.arch.DeepPhyLSTM( + cfg.MODEL.phylstm3_net.input_size, + eta.shape[2], + cfg.MODEL.phylstm3_net.hidden_size, + cfg.MODEL.phylstm3_net.model_type, + ) + model.register_input_transform(functions.transform_in) + model.register_output_transform(functions.transform_out) + + dataset_obj = functions.Dataset(eta, eta_t, g, ag, ag_c, lift, phi_t) + ( + _, + _, + input_dict_val, + label_dict_val, + ) = dataset_obj.get(1) + + sup_validator_pde = ppsci.validate.SupervisedValidator( + { + "dataset": { + "name": "NamedArrayDataset", + "input": input_dict_val, + "label": label_dict_val, + }, + }, + ppsci.loss.FunctionalLoss(functions.train_loss_func3), + { + "eta_pred": lambda out: out["eta_pred"], + "eta_dot_pred": lambda out: out["eta_dot_pred"], + "g_pred": lambda out: out["g_pred"], + "eta_t_pred_c": lambda out: out["eta_t_pred_c"], + "eta_dot_pred_c": lambda out: out["eta_dot_pred_c"], + "lift_pred_c": lambda out: out["lift_pred_c"], + "g_t_pred_c": lambda out: out["g_t_pred_c"], + "g_dot_pred_c": lambda out: out["g_dot_pred_c"], + }, + metric={"metric": ppsci.metric.FunctionalMetric(functions.metric_expr)}, + name="sup_valid", + ) + validator_pde = {sup_validator_pde.name: sup_validator_pde} + + # initialize solver + solver = ppsci.solver.Solver( + model, + output_dir=cfg.output_dir, + seed=cfg.seed, + validator=validator_pde, + checkpoint_path=cfg.TRAIN.checkpoint_path, + eval_with_no_grad=cfg.EVAL.eval_with_no_grad, + ) + + # evaluate after finished training + solver.eval() + + +@hydra.main(version_base=None, config_path="./conf", config_name="phylstm3.yaml") +def main(cfg: DictConfig): + if cfg.mode == "train": + train(cfg) + elif cfg.mode == "eval": + evaluate(cfg) + else: + raise ValueError(f"cfg.mode should in ['train', 'eval'], but got '{cfg.mode}'") + + +if __name__ == "__main__": + main() From bb5697e9f36d4cfb2d3e56e77893b2747f7e874b Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 14:55:13 +0800 Subject: [PATCH 02/10] Fix --- docs/zh/examples/phylstm.md | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/docs/zh/examples/phylstm.md b/docs/zh/examples/phylstm.md index 4e2703365..c4158a3b0 100644 --- a/docs/zh/examples/phylstm.md +++ b/docs/zh/examples/phylstm.md @@ -29,9 +29,9 @@ $$ 在 PhyLSTM 问题中，建立 LSTM 网络 Deep LSTM network，用 PaddleScience 代码表示如下 -``` py linenums="101" +``` py linenums="102" --8<-- -examples/phylstm/phylstm2.py:101:106 +examples/phylstm/phylstm2.py:102:107 --8<-- ``` @@ -47,9 +47,9 @@ wget -P ./ https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_ 本案例涉及读取数据构建，如下所示 -``` py linenums="36" +``` py linenums="37" --8<-- -examples/phylstm/phylstm2.py:36:99 +examples/phylstm/phylstm2.py:37:100 --8<-- ``` @@ -57,9 +57,9 @@ examples/phylstm/phylstm2.py:36:99 设置训练数据集和损失计算函数，返回字段，代码如下所示： -``` py linenums="116" +``` py linenums="119" --8<-- -examples/phylstm/phylstm2.py:116:135 +examples/phylstm/phylstm2.py:119:137 --8<-- ``` @@ -67,9 +67,9 @@ examples/phylstm/phylstm2.py:116:135 设置评估数据集和损失计算函数，返回字段，代码如下所示： -``` py linenums="137" +``` py linenums="140" --8<-- -examples/phylstm/phylstm2.py:137:157 +examples/phylstm/phylstm2.py:140:159 --8<-- ``` @@ -77,9 +77,9 @@ examples/phylstm/phylstm2.py:137:157 接下来我们需要指定训练轮数，此处我们按实验经验，使用 100 轮训练轮数。 -``` py linenums="167" +``` py linenums="170" --8<-- -examples/phylstm/phylstm2.py:167:167 +examples/phylstm/phylstm2.py:170:170 --8<-- ``` @@ -87,9 +87,9 @@ examples/phylstm/phylstm2.py:167:167 训练过程会调用优化器来更新模型参数，此处选择 `Adam` 优化器并设定 `learning_rate` 为 1e-3。 -``` py linenums="160" +``` py linenums="163" --8<-- -examples/phylstm/phylstm2.py:160:160 +examples/phylstm/phylstm2.py:163:163 --8<-- ``` @@ -97,17 +97,17 @@ examples/phylstm/phylstm2.py:160:160 完成上述设置之后，只需要将上述实例化的对象按顺序传递给 `ppsci.solver.Solver`。 -``` py linenums="161" +``` py linenums="164" --8<-- -examples/phylstm/phylstm2.py:161:175 +examples/phylstm/phylstm2.py:164:178 --8<-- ``` 最后启动训练、评估即可： -``` py linenums="177" +``` py linenums="180" --8<-- -examples/phylstm/phylstm2.py:177:180 +examples/phylstm/phylstm2.py:180:183 --8<-- ``` From 15bfc2ff06c442cb79dbb1295a483e0f00a53ab1 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 15:43:03 +0800 Subject: [PATCH 03/10] Fix --- examples/phylstm/phylstm2.py | 2 +- examples/phylstm/phylstm3.py | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/examples/phylstm/phylstm2.py b/examples/phylstm/phylstm2.py index 3e0f9373f..89f0c01b0 100755 --- a/examples/phylstm/phylstm2.py +++ b/examples/phylstm/phylstm2.py @@ -299,7 +299,7 @@ def evaluate(cfg: DictConfig): output_dir=cfg.output_dir, seed=cfg.seed, validator=validator_pde, - checkpoint_path=cfg.TRAIN.checkpoint_path, + checkpoint_path=cfg.EVAL.pretrained_model_path, eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) # evaluate after finished training diff --git a/examples/phylstm/phylstm3.py b/examples/phylstm/phylstm3.py index 46d1996ab..bc693125b 100755 --- a/examples/phylstm/phylstm3.py +++ b/examples/phylstm/phylstm3.py @@ -317,7 +317,7 @@ def evaluate(cfg: DictConfig): output_dir=cfg.output_dir, seed=cfg.seed, validator=validator_pde, - checkpoint_path=cfg.TRAIN.checkpoint_path, + checkpoint_path=cfg.EVAL.pretrained_model_path, eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) From f7d68973ae01e2e60f25dc612e22a3fc654cc516 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 18:37:46 +0800 Subject: [PATCH 04/10] Fix --- docs/zh/examples/phylstm.md | 24 ++++++++++++++++++++++-- examples/phylstm/conf/phylstm2.yaml | 8 +++++--- examples/phylstm/conf/phylstm3.yaml | 8 +++++--- examples/phylstm/phylstm2.py | 4 ++-- examples/phylstm/phylstm3.py | 4 ++-- 5 files changed, 36 insertions(+), 12 deletions(-) diff --git a/docs/zh/examples/phylstm.md b/docs/zh/examples/phylstm.md index c4158a3b0..cc85e338e 100644 --- a/docs/zh/examples/phylstm.md +++ b/docs/zh/examples/phylstm.md @@ -1,5 +1,25 @@ # PhyLSTM +=== "模型训练命令" + + ``` sh + # linux + wget https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat + # windows + # curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat --output data_boucwen.mat + python phylstm2.py + ``` + +=== "模型评估命令" + + ``` sh + # linux + wget https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat + # windows + # curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/PhyLSTM/data_boucwen.mat --output data_boucwen.mat + python phylstm2.py mode=eval EVAL.pretrained_model_path=https://paddle-org.bj.bcebos.com/paddlescience/models/phylstm/phylstm2_pretrained.pdparams + ``` + ## 1. 背景简介 我们引入了一种创新的物理知识LSTM框架，用于对缺乏数据的非线性结构系统进行元建模。基本概念是将可用但尚不完整的物理知识（如物理定律、科学原理）整合到深度长短时记忆（LSTM）网络中，该网络在可行的解决方案空间内限制和促进学习。物理约束嵌入在损失函数中，以强制执行模型训练，即使在可用训练数据集非常有限的情况下，也能准确地捕捉潜在的系统非线性。特别是对于动态结构，考虑运动方程的物理定律、状态依赖性和滞后本构关系来构建物理损失。嵌入式物理可以缓解过拟合问题，减少对大型训练数据集的需求，并提高训练模型的鲁棒性，使其具有外推能力，从而进行更可靠的预测。因此，物理知识指导的深度学习范式优于传统的非物理指导的数据驱动神经网络。 @@ -77,9 +97,9 @@ examples/phylstm/phylstm2.py:140:159 接下来我们需要指定训练轮数，此处我们按实验经验，使用 100 轮训练轮数。 -``` py linenums="170" +``` py linenums="31" --8<-- -examples/phylstm/phylstm2.py:170:170 +examples/phylstm/conf/phylstm2.yaml:31:31 --8<-- ``` diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml index dfd207732..94c326653 100644 --- a/examples/phylstm/conf/phylstm2.yaml +++ b/examples/phylstm/conf/phylstm2.yaml @@ -15,7 +15,9 @@ mode: train # running mode: train/eval seed: 42 output_dir: ${hydra:run.dir} log_freq: 20 -data_file: data_boucwen.mat + +# set data file path +DATA_FILE_PATH: data_boucwen.mat # model settings MODEL: @@ -29,11 +31,11 @@ TRAIN: epochs: 100 iters_per_epoch: 1 save_freq: 50 + learning_rate: 0.001 pretrained_model_path: null checkpoint_path: null - learning_rate: 0.001 # evaluation settings EVAL: pretrained_model_path: null - eval_with_no_grad: true \ No newline at end of file + eval_with_no_grad: true diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml index b88b63b47..22115c6ae 100644 --- a/examples/phylstm/conf/phylstm3.yaml +++ b/examples/phylstm/conf/phylstm3.yaml @@ -15,7 +15,9 @@ mode: train # running mode: train/eval seed: 42 output_dir: ${hydra:run.dir} log_freq: 20 -data_file: data_boucwen.mat + +# set data file path +DATA_FILE_PATH: data_boucwen.mat # model settings MODEL: @@ -29,11 +31,11 @@ TRAIN: epochs: 200 iters_per_epoch: 1 save_freq: 50 + learning_rate: 0.001 pretrained_model_path: null checkpoint_path: null - learning_rate: 0.001 # evaluation settings EVAL: pretrained_model_path: null - eval_with_no_grad: true \ No newline at end of file + eval_with_no_grad: true diff --git a/examples/phylstm/phylstm2.py b/examples/phylstm/phylstm2.py index 89f0c01b0..130d7942a 100755 --- a/examples/phylstm/phylstm2.py +++ b/examples/phylstm/phylstm2.py @@ -34,7 +34,7 @@ def train(cfg: DictConfig): # initialize logger logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") - mat = scipy.io.loadmat(cfg.data_file) + mat = scipy.io.loadmat(cfg.DATA_FILE_PATH) ag_data = mat["input_tf"] # ag, ad, av u_data = mat["target_X_tf"] ut_data = mat["target_Xd_tf"] @@ -189,7 +189,7 @@ def evaluate(cfg: DictConfig): # initialize logger logger.init_logger("ppsci", osp.join(cfg.output_dir, "eval.log"), "info") - mat = scipy.io.loadmat(cfg.data_file) + mat = scipy.io.loadmat(cfg.DATA_FILE_PATH) ag_data = mat["input_tf"] # ag, ad, av u_data = mat["target_X_tf"] ut_data = mat["target_Xd_tf"] diff --git a/examples/phylstm/phylstm3.py b/examples/phylstm/phylstm3.py index bc693125b..af5e4d9cb 100755 --- a/examples/phylstm/phylstm3.py +++ b/examples/phylstm/phylstm3.py @@ -34,7 +34,7 @@ def train(cfg: DictConfig): # initialize logger logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") - mat = scipy.io.loadmat(cfg.data_file) + mat = scipy.io.loadmat(cfg.DATA_FILE_PATH) t = mat["time"] dt = 0.02 n1 = int(dt / 0.005) @@ -199,7 +199,7 @@ def evaluate(cfg: DictConfig): # initialize logger logger.init_logger("ppsci", osp.join(cfg.output_dir, "train.log"), "info") - mat = scipy.io.loadmat(cfg.data_file) + mat = scipy.io.loadmat(cfg.DATA_FILE_PATH) t = mat["time"] dt = 0.02 n1 = int(dt / 0.005) From a13430e828c97dec0871e3f15dfb787f0272e4cb Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 18:42:31 +0800 Subject: [PATCH 05/10] Fix --- examples/phylstm/conf/phylstm2.yaml | 1 + examples/phylstm/conf/phylstm3.yaml | 1 + 2 files changed, 2 insertions(+) diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml index 94c326653..e6bc06f2d 100644 --- a/examples/phylstm/conf/phylstm2.yaml +++ b/examples/phylstm/conf/phylstm2.yaml @@ -39,3 +39,4 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true + \ No newline at end of file diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml index 22115c6ae..e6dba65c8 100644 --- a/examples/phylstm/conf/phylstm3.yaml +++ b/examples/phylstm/conf/phylstm3.yaml @@ -39,3 +39,4 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true + \ No newline at end of file From 409f04566efb3ff2aedc3b0f314d80272357b6e7 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 19:24:33 +0800 Subject: [PATCH 06/10] Fix --- examples/phylstm/conf/phylstm2.yaml | 2 +- examples/phylstm/conf/phylstm3.yaml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml index e6bc06f2d..908101ac6 100644 --- a/examples/phylstm/conf/phylstm2.yaml +++ b/examples/phylstm/conf/phylstm2.yaml @@ -39,4 +39,4 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true - \ No newline at end of file + diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml index e6dba65c8..895389c2f 100644 --- a/examples/phylstm/conf/phylstm3.yaml +++ b/examples/phylstm/conf/phylstm3.yaml @@ -39,4 +39,4 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true - \ No newline at end of file + From 8239de2b12d5182a7ccff9e25ddcd6012e8ec5c9 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 19:34:53 +0800 Subject: [PATCH 07/10] Fix --- examples/phylstm/conf/phylstm2.yaml | 1 - examples/phylstm/conf/phylstm3.yaml | 1 - 2 files changed, 2 deletions(-) diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml index 908101ac6..94c326653 100644 --- a/examples/phylstm/conf/phylstm2.yaml +++ b/examples/phylstm/conf/phylstm2.yaml @@ -39,4 +39,3 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true - diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml index 895389c2f..22115c6ae 100644 --- a/examples/phylstm/conf/phylstm3.yaml +++ b/examples/phylstm/conf/phylstm3.yaml @@ -39,4 +39,3 @@ TRAIN: EVAL: pretrained_model_path: null eval_with_no_grad: true - From 820a462817e64724042f5dd2381b2f222ca70cd0 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 20:58:23 +0800 Subject: [PATCH 08/10] Fix --- docs/zh/examples/phylstm.md | 4 ++-- examples/phylstm/phylstm2.py | 14 +++++++------- examples/phylstm/phylstm3.py | 14 +++++++------- 3 files changed, 16 insertions(+), 16 deletions(-) diff --git a/docs/zh/examples/phylstm.md b/docs/zh/examples/phylstm.md index cc85e338e..cc8b44029 100644 --- a/docs/zh/examples/phylstm.md +++ b/docs/zh/examples/phylstm.md @@ -97,9 +97,9 @@ examples/phylstm/phylstm2.py:140:159 接下来我们需要指定训练轮数，此处我们按实验经验，使用 100 轮训练轮数。 -``` py linenums="31" +``` py linenums="39" --8<-- -examples/phylstm/conf/phylstm2.yaml:31:31 +examples/phylstm/conf/phylstm2.yaml:39:39 --8<-- ``` diff --git a/examples/phylstm/phylstm2.py b/examples/phylstm/phylstm2.py index 130d7942a..759cb9a5e 100755 --- a/examples/phylstm/phylstm2.py +++ b/examples/phylstm/phylstm2.py @@ -100,10 +100,10 @@ def train(cfg: DictConfig): phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) model = ppsci.arch.DeepPhyLSTM( - cfg.MODEL.phylstm2_net.input_size, + cfg.MODEL.input_size, eta.shape[2], - cfg.MODEL.phylstm2_net.hidden_size, - cfg.MODEL.phylstm2_net.model_type, + cfg.MODEL.hidden_size, + cfg.MODEL.model_type, ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -255,10 +255,10 @@ def evaluate(cfg: DictConfig): phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) model = ppsci.arch.DeepPhyLSTM( - cfg.MODEL.phylstm2_net.input_size, + cfg.MODEL.input_size, eta.shape[2], - cfg.MODEL.phylstm2_net.hidden_size, - cfg.MODEL.phylstm2_net.model_type, + cfg.MODEL.hidden_size, + cfg.MODEL.model_type, ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -299,7 +299,7 @@ def evaluate(cfg: DictConfig): output_dir=cfg.output_dir, seed=cfg.seed, validator=validator_pde, - checkpoint_path=cfg.EVAL.pretrained_model_path, + pretrained_model_path=cfg.EVAL.pretrained_model_path, eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) # evaluate after finished training diff --git a/examples/phylstm/phylstm3.py b/examples/phylstm/phylstm3.py index af5e4d9cb..4119f7a3d 100755 --- a/examples/phylstm/phylstm3.py +++ b/examples/phylstm/phylstm3.py @@ -106,10 +106,10 @@ def train(cfg: DictConfig): phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) model = ppsci.arch.DeepPhyLSTM( - cfg.MODEL.phylstm3_net.input_size, + cfg.MODEL.input_size, eta.shape[2], - cfg.MODEL.phylstm3_net.hidden_size, - cfg.MODEL.phylstm3_net.model_type, + cfg.MODEL.hidden_size, + cfg.MODEL.model_type, ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -271,10 +271,10 @@ def evaluate(cfg: DictConfig): phi_t = np.repeat(phi_t0, ag_c_star.shape[0], axis=0) model = ppsci.arch.DeepPhyLSTM( - cfg.MODEL.phylstm3_net.input_size, + cfg.MODEL.input_size, eta.shape[2], - cfg.MODEL.phylstm3_net.hidden_size, - cfg.MODEL.phylstm3_net.model_type, + cfg.MODEL.hidden_size, + cfg.MODEL.model_type, ) model.register_input_transform(functions.transform_in) model.register_output_transform(functions.transform_out) @@ -317,7 +317,7 @@ def evaluate(cfg: DictConfig): output_dir=cfg.output_dir, seed=cfg.seed, validator=validator_pde, - checkpoint_path=cfg.EVAL.pretrained_model_path, + pretrained_model_path=cfg.EVAL.pretrained_model_path, eval_with_no_grad=cfg.EVAL.eval_with_no_grad, ) From 744235fde88649313051b4ffcf729c66a05d8798 Mon Sep 17 00:00:00 2001 From: co63oc Date: Thu, 19 Oct 2023 21:00:14 +0800 Subject: [PATCH 09/10] Fix --- examples/phylstm/conf/phylstm2.yaml | 16 ++++++++++++---- examples/phylstm/conf/phylstm3.yaml | 16 ++++++++++++---- 2 files changed, 24 insertions(+), 8 deletions(-) diff --git a/examples/phylstm/conf/phylstm2.yaml b/examples/phylstm/conf/phylstm2.yaml index 94c326653..3fc184d71 100644 --- a/examples/phylstm/conf/phylstm2.yaml +++ b/examples/phylstm/conf/phylstm2.yaml @@ -5,6 +5,15 @@ hydra: job: name: ${mode} # name of logfile chdir: false # keep current working direcotry unchaned + config: + override_dirname: + exclude_keys: + - TRAIN.checkpoint_path + - TRAIN.pretrained_model_path + - EVAL.pretrained_model_path + - mode + - output_dir + - log_freq sweep: # output directory for multirun dir: ${hydra.run.dir} @@ -21,10 +30,9 @@ DATA_FILE_PATH: data_boucwen.mat # model settings MODEL: - phylstm2_net: - input_size: 1 - hidden_size: 100 - model_type: 2 + input_size: 1 + hidden_size: 100 + model_type: 2 # training settings TRAIN: diff --git a/examples/phylstm/conf/phylstm3.yaml b/examples/phylstm/conf/phylstm3.yaml index 22115c6ae..0be68339b 100644 --- a/examples/phylstm/conf/phylstm3.yaml +++ b/examples/phylstm/conf/phylstm3.yaml @@ -5,6 +5,15 @@ hydra: job: name: ${mode} # name of logfile chdir: false # keep current working direcotry unchaned + config: + override_dirname: + exclude_keys: + - TRAIN.checkpoint_path + - TRAIN.pretrained_model_path + - EVAL.pretrained_model_path + - mode + - output_dir + - log_freq sweep: # output directory for multirun dir: ${hydra.run.dir} @@ -21,10 +30,9 @@ DATA_FILE_PATH: data_boucwen.mat # model settings MODEL: - phylstm3_net: - input_size: 1 - hidden_size: 100 - model_type: 3 + input_size: 1 + hidden_size: 100 + model_type: 3 # training settings TRAIN: From 51a255da50a1fe18f7c35090454ec6ec99e55ab1 Mon Sep 17 00:00:00 2001 From: co63oc Date: Fri, 20 Oct 2023 10:19:38 +0800 Subject: [PATCH 10/10] ci_trigger