0.16.0

bartzbeielstein · bartzbeielstein · commit 936084262d92 · 2024-10-23T18:35:39.000+02:00
condNet
diff --git a/notebooks/00_spotPython_tests.ipynb b/notebooks/00_spotPython_tests.ipynb
diff --git a/pyproject.toml b/pyproject.toml
@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "spotpython"
-version = "0.15.40"
+version = "0.16.0"
 authors = [
   { name="T. Bartz-Beielstein", email="tbb@bartzundbartz.de" }
 ]
diff --git a/src/spotpython/hyperdict/light_hyper_dict.json b/src/spotpython/hyperdict/light_hyper_dict.json
@@ -1,4 +1,115 @@
 {
+    "NNCondNetRegressor": {
+        "l1": {
+            "type": "int",
+            "default": 3,
+            "transform": "transform_power_2_int",
+            "lower": 3,
+            "upper": 8
+        },
+        "epochs": {
+            "type": "int",
+            "default": 4,
+            "transform": "transform_power_2_int",
+            "lower": 4,
+            "upper": 9
+        },
+        "batch_size": {
+            "type": "int",
+            "default": 4,
+            "transform": "transform_power_2_int",
+            "lower": 1,
+            "upper": 4
+        },
+        "act_fn": {
+            "levels": [
+                "Sigmoid",
+                "Tanh",
+                "ReLU",
+                "LeakyReLU",
+                "ELU",
+                "Swish"
+            ],
+            "type": "factor",
+            "default": "ReLU",
+            "transform": "None",
+            "class_name": "spotpython.torch.activation",
+            "core_model_parameter_type": "instance()",
+            "lower": 0,
+            "upper": 5
+        },
+        "optimizer": {
+            "levels": [
+                "Adadelta",
+                "Adagrad",
+                "Adam",
+                "AdamW",
+                "SparseAdam",
+                "Adamax",
+                "ASGD",
+                "NAdam",
+                "RAdam",
+                "RMSprop",
+                "Rprop",
+                "SGD"
+            ],
+            "type": "factor",
+            "default": "SGD",
+            "transform": "None",
+            "class_name": "torch.optim",
+            "core_model_parameter_type": "str",
+            "lower": 0,
+            "upper": 11
+        },
+        "dropout_prob": {
+            "type": "float",
+            "default": 0.01,
+            "transform": "None",
+            "lower": 0.0,
+            "upper": 0.25
+        },
+        "lr_mult": {
+            "type": "float",
+            "default": 1.0,
+            "transform": "None",
+            "lower": 0.1,
+            "upper": 10.0
+        },
+        "patience": {
+            "type": "int",
+            "default": 2,
+            "transform": "transform_power_2_int",
+            "lower": 2,
+            "upper": 6
+        },
+        "batch_norm": {
+            "levels": [
+                0,
+                1
+            ],
+            "type": "factor",
+            "default": 0,
+            "transform": "None",
+            "core_model_parameter_type": "bool",
+            "lower": 0,
+            "upper": 1
+        },
+        "initialization": {
+            "levels": [
+                "Default",
+                "kaiming_uniform",
+                "kaiming_normal",
+                "xavier_uniform",
+                "xavier_normal"
+            ],
+            "type": "factor",
+            "default": "Default",
+            "transform": "None",
+            "core_model_parameter_type": "str",
+            "lower": 0,
+            "upper": 4
+        }
+    },
     "TransformerLightRegression": {
         "d_mult": {
             "type": "int",
diff --git a/src/spotpython/light/classification/netlightbasemapk.py b/src/spotpython/light/classification/netlightbasemapk.py
@@ -73,6 +73,8 @@ def __init__(
         patience: int,
         _L_in: int,
         _L_out: int,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the NetLightBase object.
diff --git a/src/spotpython/light/cvmodel.py b/src/spotpython/light/cvmodel.py
@@ -26,6 +26,7 @@ def cv_model(config: dict, fun_control: dict) -> float:
         >>> fun_control = {
         ...     "_L_in": 10,
         ...     "_L_out": 1,
+        ...     "_L_cond": 0,
         ...     "enable_progress_bar": True,
         ...     "core_model": MyModel,
         ...     "num_workers": 4,
@@ -38,6 +39,7 @@ def cv_model(config: dict, fun_control: dict) -> float:
     """
     _L_in = fun_control["_L_in"]
     _L_out = fun_control["_L_out"]
+    _L_cond = fun_control["_L_cond"]
     _torchmetric = fun_control["_torchmetric"]
     if fun_control["enable_progress_bar"] is None:
         enable_progress_bar = False
@@ -52,7 +54,9 @@ def cv_model(config: dict, fun_control: dict) -> float:
     for k in range(num_folds):
         print("k:", k)
 
-        model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _torchmetric=_torchmetric)
+        model = fun_control["core_model"](
+            **config, _L_in=_L_in, _L_out=_L_out, _L_cond=_L_cond, _torchmetric=_torchmetric
+        )
 
         dm = LightCrossValidationDataModule(
             k=k,
diff --git a/src/spotpython/light/litmodel.py b/src/spotpython/light/litmodel.py
@@ -47,6 +47,8 @@ def __init__(
         learning_rate: float = 2e-4,
         _L_in: int = 28 * 28,
         _L_out: int = 10,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the LitModel object.
diff --git a/src/spotpython/light/loadmodel.py b/src/spotpython/light/loadmodel.py
@@ -51,6 +51,7 @@ def load_light_from_checkpoint(config: dict, fun_control: dict, postfix: str = "
         default_root_dir,
         _L_in=fun_control["_L_in"],
         _L_out=fun_control["_L_out"],
+        _L_cond=fun_control["_L_cond"],
         _torchmetric=fun_control["_torchmetric"],
     )
     # disable randomness, dropout, etc...
diff --git a/src/spotpython/light/predictmodel.py b/src/spotpython/light/predictmodel.py
@@ -53,6 +53,7 @@ def predict_model(config: dict, fun_control: dict) -> Tuple[float, float]:
     """
     _L_in = fun_control["_L_in"]
     _L_out = fun_control["_L_out"]
+    _L_cond = fun_control["_L_cond"]
     _torchmetric = fun_control["_torchmetric"]
     if fun_control["enable_progress_bar"] is None:
         enable_progress_bar = False
@@ -74,7 +75,7 @@ def predict_model(config: dict, fun_control: dict) -> Tuple[float, float]:
     # TODO: Check if this is necessary:
     # dm.setup(stage="train")
     # Init model from datamodule's attributes
-    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _torchmetric=_torchmetric)
+    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _L_cond=_L_cond, _torchmetric=_torchmetric)
 
     trainer = L.Trainer(
         # Where to save models
diff --git a/src/spotpython/light/regression/netlightregression.py b/src/spotpython/light/regression/netlightregression.py
@@ -107,6 +107,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the NetLightRegression object.
diff --git a/src/spotpython/light/regression/nn_condnet_regressor.py b/src/spotpython/light/regression/nn_condnet_regressor.py
@@ -146,6 +146,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the NNLinearRegressor object.
@@ -212,12 +214,12 @@ def __init__(
         hidden_sizes = self._get_hidden_sizes()
 
         # Conditional Layer
-        self.cond_layer = ConditionalLayer(self._L_in, self._L_cond, self._L_in)
-        
+        self.cond_layer = ConditionalLayer(self._L_in, self._L_cond, self.hparams.l1)
+
         if batch_norm:
             # Add batch normalization layers
             layers = []
-            layer_sizes = [self._L_in] + hidden_sizes
+            layer_sizes = [self.hparams.l1] + hidden_sizes
             for i in range(len(layer_sizes) - 1):
                 current_layer_size = layer_sizes[i]
                 next_layer_size = layer_sizes[i + 1]
@@ -230,7 +232,7 @@ def __init__(
             layers += [nn.Linear(layer_sizes[-1], self._L_out)]
         else:
             layers = []
-            layer_sizes = [self._L_in] + hidden_sizes
+            layer_sizes = [self.hparams.l1] + hidden_sizes
             for i in range(len(layer_sizes) - 1):
                 current_layer_size = layer_sizes[i]
                 next_layer_size = layer_sizes[i + 1]
@@ -244,8 +246,6 @@ def __init__(
         # Wrap the layers into a sequential container
         self.layers = nn.Sequential(*layers)
 
-
-
         # Initialization (Xavier, Kaiming, or Default)
         self.apply(self._init_weights)
 
diff --git a/src/spotpython/light/regression/nn_linear_regressor.py b/src/spotpython/light/regression/nn_linear_regressor.py
@@ -126,6 +126,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the NNLinearRegressor object.
diff --git a/src/spotpython/light/regression/nn_resnet_regressor.py b/src/spotpython/light/regression/nn_resnet_regressor.py
@@ -53,6 +53,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         super().__init__()
         self._L_in = _L_in
diff --git a/src/spotpython/light/regression/nn_transformer_regressor.py b/src/spotpython/light/regression/nn_transformer_regressor.py
@@ -24,6 +24,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         super().__init__()
 
diff --git a/src/spotpython/light/regression/rnnlightregression.py b/src/spotpython/light/regression/rnnlightregression.py
@@ -109,6 +109,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the NetLightRegression object.
diff --git a/src/spotpython/light/regression/transformerlightregression.py b/src/spotpython/light/regression/transformerlightregression.py
@@ -114,6 +114,8 @@ def __init__(
         _L_in: int,
         _L_out: int,
         _torchmetric: str,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the TransformerLightRegression object.
diff --git a/src/spotpython/light/testmodel.py b/src/spotpython/light/testmodel.py
@@ -54,6 +54,7 @@ def test_model(config: dict, fun_control: dict) -> Tuple[float, float]:
     """
     _L_in = fun_control["_L_in"]
     _L_out = fun_control["_L_out"]
+    _L_cond = fun_control["_L_cond"]
     _torchmetric = fun_control["_torchmetric"]
     if fun_control["enable_progress_bar"] is None:
         enable_progress_bar = False
@@ -75,7 +76,7 @@ def test_model(config: dict, fun_control: dict) -> Tuple[float, float]:
     # TODO: Check if this is necessary:
     # dm.setup()
     # Init model from datamodule's attributes
-    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _torchmetric=_torchmetric)
+    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _L_cond=_L_cond, _torchmetric=_torchmetric)
 
     trainer = L.Trainer(
         # Where to save models
diff --git a/src/spotpython/light/trainmodel.py b/src/spotpython/light/trainmodel.py
@@ -72,6 +72,7 @@ def train_model(config: dict, fun_control: dict, timestamp: bool = True) -> floa
     """
     _L_in = fun_control["_L_in"]
     _L_out = fun_control["_L_out"]
+    _L_cond = fun_control["_L_cond"]
     _torchmetric = fun_control["_torchmetric"]
     if fun_control["enable_progress_bar"] is None:
         enable_progress_bar = False
@@ -86,7 +87,7 @@ def train_model(config: dict, fun_control: dict, timestamp: bool = True) -> floa
         # so that the model can be loaded from a checkpoint,
         # the config id is generated here without a timestamp.
         config_id = generate_config_id(config, timestamp=False) + "_TRAIN"
-    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _torchmetric=_torchmetric)
+    model = fun_control["core_model"](**config, _L_in=_L_in, _L_out=_L_out, _L_cond=_L_cond, _torchmetric=_torchmetric)
 
     dm = LightDataModule(
         dataset=fun_control["data_set"],
diff --git a/src/spotpython/light/transformer/transformerlightpredictor.py b/src/spotpython/light/transformer/transformerlightpredictor.py
@@ -37,6 +37,8 @@ def __init__(
         max_iters: int,
         input_dropout: float,
         dropout: float,
+        *args,
+        **kwargs,
     ):
         """
         Initializes the TransformerLightRegression object.
diff --git a/src/spotpython/utils/init.py b/src/spotpython/utils/init.py
@@ -20,6 +20,7 @@
 def fun_control_init(
     _L_in=None,
     _L_out=None,
+    _L_cond=None,
     _torchmetric=None,
     PREFIX="00",
     TENSORBOARD_CLEAN=False,
@@ -101,6 +102,8 @@ def fun_control_init(
             The number of input features.
         _L_out (int):
             The number of output features.
+        _L_cond (int):
+            The number of conditional features.
         _torchmetric (str):
             The metric to be used by the Lighting Trainer.
             For example "mean_squared_error",
@@ -297,6 +300,7 @@ def fun_control_init(
                 'TENSORBOARD_PATH': 'runs/',
                 '_L_in': 64,
                 '_L_out': 11,
+                '_L_cond': None,
                 'accelerator': "auto",
                 'core_model': None,
                 'core_model_name': None,
@@ -363,6 +367,7 @@ def fun_control_init(
         "TENSORBOARD_CLEAN": TENSORBOARD_CLEAN,
         "_L_in": _L_in,
         "_L_out": _L_out,
+        "_L_cond": _L_cond,
         "_torchmetric": _torchmetric,
         "accelerator": accelerator,
         "converters": converters,
diff --git a/test/test_fun_control_init.py b/test/test_fun_control_init.py
@@ -6,6 +6,7 @@ def test_fun_control_init():
     fun_control = fun_control_init(_L_in=64, _L_out=11, num_workers=0, device=None)
     assert fun_control["_L_in"] == 64
     assert fun_control["_L_out"] == 11
+    assert fun_control["_L_cond"] is None
     assert fun_control["num_workers"] == 0
     assert fun_control["device"] is None
     assert fun_control["task"] is None

Original file line number	Diff line number	Diff line change
`@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"`
`7`	`7`
`8`	`8`	`[project]`
`9`	`9`	`name = "spotpython"`
`10`		`-version = "0.15.40"`
	`10`	`+version = "0.16.0"`
`11`	`11`	`authors = [`
`12`	`12`	`{ name="T. Bartz-Beielstein", email="tbb@bartzundbartz.de" }`
`13`	`13`	`]`
Original file line number	Diff line number	Diff line change
`@@ -51,6 +51,7 @@ def load_light_from_checkpoint(config: dict, fun_control: dict, postfix: str = "`
`51`	`51`	`default_root_dir,`
`52`	`52`	`_L_in=fun_control["_L_in"],`
`53`	`53`	`_L_out=fun_control["_L_out"],`
	`54`	`+ _L_cond=fun_control["_L_cond"],`
`54`	`55`	`_torchmetric=fun_control["_torchmetric"],`
`55`	`56`	`)`
`56`	`57`	`# disable randomness, dropout, etc...`