Commit 84760887 authored by Bharath Ramsundar's avatar Bharath Ramsundar
Browse files

Fixed some more bugs with multitasking output sizes

parent 58bf2bf9

deepchem/hyperparameters/tests/test_hyperparam_opt.py

+3 −2
Original line number Diff line number Diff line
@@ -104,7 +104,8 @@ class TestHyperparamOptAPI(TestAPI): 
        "batch_size": [32],

        "data_shape": [train_dataset.get_data_shape()],

    }

    classification_metric = Metric(metrics.matthews_corrcoef, np.mean)

    classification_metric = Metric(metrics.matthews_corrcoef, np.mean,

                                   mode="classification")

    def model_builder(tasks, task_types, model_params, task_model_dir,

                      verbosity=None):

      return SklearnModel(tasks, task_types, model_params, task_model_dir,
@@ -138,7 +139,7 @@ class TestHyperparamOptAPI(TestAPI): 
        splittype, compound_featurizers, 

        complex_featurizers, input_transformers,

        output_transformers, input_file, tasks)

    metric = Metric(metrics.matthews_corrcoef, np.mean)

    metric = Metric(metrics.matthews_corrcoef, np.mean, mode="classification")

    params_dict= {"nb_hidden": [5, 10],

                  "activation": ["relu"],

                  "dropout": [.5],

deepchem/models/__init__.py

+4 −6
Original line number Diff line number Diff line
@@ -138,10 +138,6 @@ class Model(object): 
    y_preds = []

    batch_size = self.model_params["batch_size"]

    for (X_batch, y_batch, w_batch, ids_batch) in dataset.iterbatches(batch_size):

      print("y_batch.shape")

      print(y_batch.shape)

      print("self.predict_on_batch(X_batch).shape")

      print(self.predict_on_batch(X_batch).shape)

      y_pred_batch = np.reshape(self.predict_on_batch(X_batch), y_batch.shape)

      y_pred_batch = undo_transforms(y_pred_batch, transformers)

      y_preds.append(y_pred_batch)
@@ -165,13 +161,15 @@ class Model(object): 
    n_tasks = len(self.tasks)

    for (X_batch, y_batch, w_batch, ids_batch) in dataset.iterbatches(batch_size):

      y_pred_batch = self.predict_proba_on_batch(X_batch)

      print("y_pred_batch.shape")

      print(y_pred_batch.shape)

      if n_classes is None:

        n_classes = y_pred_batch.shape[-1]

      batch_size = len(y_batch)

      print("y_pred_batch.shape")

      print(y_pred_batch.shape)

      print("batch_size, n_tasks, n_classes")

      print(batch_size, n_tasks, n_classes)

      print("(batch_size, n_tasks, n_classes)")

      print((batch_size, n_tasks, n_classes))

      y_pred_batch = np.squeeze(

          np.reshape(y_pred_batch, (batch_size, n_tasks, n_classes)))

      y_pred_batch = undo_transforms(y_pred_batch, transformers)

deepchem/models/keras_models/fcnet.py

+4 −4
Original line number Diff line number Diff line
@@ -139,14 +139,14 @@ class MultiTaskDNN(KerasModel): 
    """

    data = self.get_data_dict(X)

    y_pred_dict = self.raw_model.predict_on_batch(data)

    nb_samples = np.shape(X)[0]

    nb_tasks = len(self.tasks)

    y_pred = np.zeros((nb_samples, n_classes*nb_tasks))

    n_samples = np.shape(X)[0]

    n_tasks = len(self.tasks)

    y_pred = np.zeros((n_samples, n_tasks, n_classes))

    for ind, task in enumerate(self.tasks):

      task_type = self.task_types[task]

      taskname = "task%d" % ind

      y_pred_task = np.squeeze(y_pred_dict[taskname])

      y_pred[:, ind:ind+n_classes] = y_pred_task

      y_pred[:, ind] = y_pred_task

    y_pred = np.squeeze(y_pred)

    return y_pred

deepchem/models/multitask.py

+5 −6
Original line number Diff line number Diff line
@@ -62,8 +62,8 @@ class SingletaskToMultitask(Model): 
    Concatenates results from all singletask models.

    """

    N_tasks = len(self.tasks)

    N_samples = X.shape[0]

    y_pred = np.zeros((N_samples, N_tasks))

    n_samples = X.shape[0]

    y_pred = np.zeros((n_samples, N_tasks))

    for ind, task in enumerate(self.tasks):

      y_pred[:, ind] = self.models[task].predict_on_batch(X)[:, 0]

    return y_pred
@@ -73,11 +73,10 @@ class SingletaskToMultitask(Model): 
    Concatenates results from all singletask models.

    """

    n_tasks = len(self.tasks)

    N_samples = X.shape[0]

    y_pred = np.zeros((N_samples, n_classes*n_tasks))

    n_samples = X.shape[0]

    y_pred = np.zeros((n_samples, n_tasks, n_classes))

    for ind, task in enumerate(self.tasks):

      y_pred[:, ind*n_classes:(ind+1)*n_classes] = \

          self.models[task].predict_proba_on_batch(X)

      y_pred[:, ind] = self.models[task].predict_proba_on_batch(X)

    return y_pred



  def save(self):

CRA Git | Maintained and supported by SUSTech CRA and CCSE