NeuralNetworks 0.2.5__tar.gz → 0.2.6__tar.gz

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: NeuralNetworks
-Version: 0.2.5
+Version: 0.2.6
 Summary: Multi-Layer Perceptrons with multi-Fourier encoding, variable learning rate, visualization and PyTorch compilation
 Author-email: Alexandre Brun <alexandre51160@gmail.com>
 License: GPL-3.0-or-later

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks/Dependances/pytorch.py

@@ -6,7 +6,6 @@
 # (at your option) any later version.
 
 import os
-os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "expandable_segments:True"
 
 import platform
 
@@ -20,6 +19,8 @@ from torch.utils.data import TensorDataset, DataLoader
 from torchmetrics.image import PeakSignalNoiseRatio as PSNR
 from torchvision.transforms import ToTensor, Resize, Compose
 
+import visualtorch
+
 torch.cuda.empty_cache()
 def get_best_device():
 
@@ -51,6 +52,7 @@ def get_best_device():
 
     # =========== Unknown OS ===========
     return torch.device("cpu")
+    
 device = get_best_device()
 
 # --- Optimisations CUDA ---

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks/MLP/__init__.py

@@ -10,7 +10,9 @@ from .FourierFeatures import encode
 from .Layers import create_layers
 from .inference import infer
 
-class MLP (nn.Module):
+from ..shared import Module
+
+class MLP (Module):
     """
     Multi-Layer Perceptron avec encodage de Fourier optionnel.
 
@@ -49,18 +51,13 @@ class MLP (nn.Module):
         nb_fourier = 8,
         norm = "Relu",
         name = "Net"):
-        super ().__init__ ()
+        super ().__init__ (name)
 
         # --- Activation ---
         self.norm = norm_list.get (norm)
         if self.norm is None:
-            print (f"Warning: '{norm}' not recognized, falling back to 'Relu'")
+            print (f"Warning: '{norm}' not recognized, fRelung back to 'm is'")
             self.norm = norm_list.get ("Relu")
-
-        # --- Attributs ---
-        self.losses = []
-        self.learnings = []
-        self.name = name
         
         ## --- Encodage Fourier ou passthrough ---
         self.encodings, self.f = encode (
@@ -82,18 +79,8 @@ class MLP (nn.Module):
             self.norm
         )
 
-    def forward (self, x):
-        """
-        Effectue une passe avant du réseau.
-
-        Parameters
-        ----------
-        x : torch.Tensor
-            Entrées du réseau de shape `(N, input_size)`.
+    def _forward (self, x):
+        return infer (self, x)
 
-        Returns
-        -------
-        torch.Tensor
-            Sortie du MLP de shape `(N, output_size)`.
-        """
+    def train_forward (self, x):
         return infer (self, x)

neuralnetworks-0.2.6/src/NeuralNetworks/MLP/inference.py

@@ -0,0 +1,12 @@
+# NeuralNetworks - Multi-Layer Perceptrons avec encodage Fourier
+# Copyright (C) 2026 Alexandre Brun
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+from ..Dependances import torch, np, device
+
+def infer (net, x):
+    results_list = [net.model (encoding (x)) for encoding in net.encodings]
+    return net.f (torch.cat (results_list, dim = 1))

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks/Trainer/__init__.py

@@ -5,7 +5,7 @@
 # the Free Software Foundation, either version 3 of the License, or
 # (at your option) any later version.
 
-from ..Dependances import crit_list, optim_list
+from ..Dependances import crit_list, optim_list, np
 from .train import train_f
 from .sample_data import sample_data
 
@@ -15,7 +15,7 @@ class Trainer:
         *nets,
         inputs,
         outputs,
-        test_size = None,
+        train_size = None,
         optim = 'Adam',
         init_lr = 0.01,
         crit = 'MSE',
@@ -24,12 +24,19 @@ class Trainer:
         self.batch_size = batch_size
         self.nets = nets
         self.init_lr = init_lr
-
-        self.X_train, self.X_test, self.y_train, self.y_test = sample_data (
-            inputs,
-            outputs,
-            test_size
-        )
+        if train_size is None:
+
+            self.X_train = inputs
+            self.X_test  = inputs
+            self.y_train = outputs
+            self.y_test  = outputs
+        else:
+            
+            self.X_train, self.X_test, self.y_train, self.y_test = sample_data (
+                inputs,
+                outputs,
+                min(1,np.abs(1 - train_size))
+            )
 
         # --- Fonction de perte ---
         self.crit = crit_list.get(crit)

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks/Trainer/train.py

@@ -27,7 +27,8 @@ def train_f (Trainer, num_epochs = 1, activate_tqdm = True):
             pbar = tqdm (
                 range (num_epochs),
                 desc = f"train epoch",
-                disable = not (activate_tqdm)
+                disable = not (activate_tqdm),
+                mininterval=0.5
             )
 
             for epoch in pbar:
@@ -43,14 +44,8 @@ def train_f (Trainer, num_epochs = 1, activate_tqdm = True):
                     def closure ():
                         Trainer.optims [k].zero_grad (set_to_none = True)
                         with autocast (dev):
-                            loss = Trainer.crit (
-                                net.f (
-                                    torch.cat (
-            [net.model (encoding (Trainer.X_train [idx]))for encoding in net.encodings],
-            dim = 1
-                                    )
-                                ),
-                                Trainer.y_train[idx]
+                            loss = Trainer.crit (net.train_forward (Trainer.X_train [idx]),
+                                Trainer.y_train [idx]
                             )
                             scaler.scale (loss).backward ()
                             return loss
@@ -66,7 +61,7 @@ def train_f (Trainer, num_epochs = 1, activate_tqdm = True):
                 for param_group in Trainer.optims [k].param_groups:
                     param_group ['lr'] = net.learnings[-1]
                 
-                pbar.set_postfix(loss=f"{epoch_loss:.5f}",lr=f"{net.learnings[-1]:.5f}")
+                #pbar.set_postfix(loss=f"{epoch_loss:.5f}",lr=f"{net.learnings[-1]:.5f}")
                 
             net = net.to ('cpu')
             net.learnings.pop(-1)

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks/__init__.py

@@ -15,4 +15,4 @@ from .Trainer import Trainer
 
 from .UI import *
 
-__version__ = "0.2.5"
+__version__ = "0.2.6"

neuralnetworks-0.2.6/src/NeuralNetworks/shared/__init__.py

@@ -0,0 +1,8 @@
+# NeuralNetworks - Multi-Layer Perceptrons avec encodage Fourier
+# Copyright (C) 2026 Alexandre Brun
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+from .module import Module

neuralnetworks-0.2.6/src/NeuralNetworks/shared/module.py

@@ -0,0 +1,41 @@
+# NeuralNetworks - Multi-Layer Perceptrons avec encodage Fourier
+# Copyright (C) 2026 Alexandre Brun
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+from ..Dependances import norm_list, nn, torch, device
+
+class Module (nn.Module):
+
+    def __init__ (self, name = "Net"):
+        super ().__init__ ()
+
+        # --- Attributs ---
+        self.losses = []
+        self.learnings = []
+        self.name = name
+
+    def _forward (self,x):
+        raise Exception ("_forward n'est pas défini dans la classe")
+
+    def train_forward (self,x):
+        raise Exception ("train_forward n'est pas défini dans la classe")
+            
+    def forward (self, x):
+
+        with torch.no_grad ():
+            x = x.unsqueeze (0) if x.dim () == 1 else x
+            self = self.to (device)
+            x = x.to (device)
+
+            output = self._forward(x).cpu ().numpy ().flatten ()
+            
+            x = x.to ('cpu')
+            self = self.to ('cpu')
+
+            return output
+
+            
+    

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: NeuralNetworks
-Version: 0.2.5
+Version: 0.2.6
 Summary: Multi-Layer Perceptrons with multi-Fourier encoding, variable learning rate, visualization and PyTorch compilation
 Author-email: Alexandre Brun <alexandre51160@gmail.com>
 License: GPL-3.0-or-later

{neuralnetworks-0.2.5 → neuralnetworks-0.2.6}/src/NeuralNetworks.egg-info/SOURCES.txt

@@ -20,4 +20,6 @@ src/NeuralNetworks/Trainer/sample_data.py
 src/NeuralNetworks/Trainer/train.py
 src/NeuralNetworks/UI/Learnings.py
 src/NeuralNetworks/UI/Losses.py
-src/NeuralNetworks/UI/__init__.py
+src/NeuralNetworks/UI/__init__.py
+src/NeuralNetworks/shared/__init__.py
+src/NeuralNetworks/shared/module.py

neuralnetworks-0.2.5/src/NeuralNetworks/MLP/inference.py

@@ -1,26 +0,0 @@
-# NeuralNetworks - Multi-Layer Perceptrons avec encodage Fourier
-# Copyright (C) 2026 Alexandre Brun
-# This program is free software: you can redistribute it and/or modify
-# it under the terms of the GNU General Public License as published by
-# the Free Software Foundation, either version 3 of the License, or
-# (at your option) any later version.
-
-from ..Dependances import torch, np, device
-
-def infer (net, x):
-    with torch.no_grad ():
-        x = x.unsqueeze (0) if x.dim () == 1 else x
-
-        net = net.to (device)
-        x = x.to (device)
-        results_list = [net.model (encoding (x)) for encoding in net.encodings]
-        x = x.to ('cpu')
-
-        output = np.array (
-            net.f (
-                torch.cat (results_list, dim = 1)
-            ).cpu ().numpy ().flatten ()
-        )
-        net = net.to ('cpu')
-
-    return output