doc/v632/RBatchGenerator__PyTorch_8py.html

import torch

import ROOT


tree_name = "sig_tree"

file_name = str(ROOT.gROOT.GetTutorialDir()) + "/tmva/data/Higgs_data.root"


batch_size = 128

chunk_size = 5_000


target = "Type"


# Returns two generators that return training and validation batches

# as PyTorch tensors.

gen_train, gen_validation = ROOT.TMVA.Experimental.CreatePyTorchGenerators(

    tree_name,

    file_name,

    batch_size,

    chunk_size,

    target=target,

    validation_split=0.3,

)


# Get a list of the columns used for training

input_columns = gen_train.train_columns

num_features = len(input_columns)


def calc_accuracy(targets, pred):

    return torch.sum(targets == pred.round()) / pred.size(0)


# Initialize PyTorch model

model = torch.nn.Sequential(

    torch.nn.Linear(num_features, 300),

    torch.nn.Tanh(),

    torch.nn.Linear(300, 300),

    torch.nn.Tanh(),

    torch.nn.Linear(300, 300),

    torch.nn.Tanh(),

    torch.nn.Linear(300, 1),

    torch.nn.Sigmoid(),

)

loss_fn = torch.nn.MSELoss(reduction="mean")

optimizer = torch.optim.SGD(model.parameters(), lr=0.01, momentum=0.9)


# Loop through the training set and train model

for i, (x_train, y_train) in enumerate(gen_train):

    # Make prediction and calculate loss

    pred = model(x_train).view(-1)

    loss = loss_fn(pred, y_train)


    # improve model

    model.zero_grad()

    loss.backward()

    optimizer.step()


    # Calculate accuracy

    accuracy = calc_accuracy(y_train, pred)


    print(f"Training => accuracy: {accuracy}")


#################################################################

# Validation

#################################################################


# Evaluate the model on the validation set

for i, (x_train, y_train) in enumerate(gen_validation):

    # Make prediction and calculate accuracy

    pred = model(x_train).view(-1)

    accuracy = calc_accuracy(y_train, pred)


    print(f"Validation => accuracy: {accuracy}")

TRangeDynCast
ROOT::Detail::TRangeCast< T, true > TRangeDynCast
TRangeDynCast is an adapter class that allows the typed iteration through a TCollection.
Definition TCollection.h:358

len
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char Pixmap_t Pixmap_t PictureAttributes_t attr const char char ret_data h unsigned char height h Atom_t Int_t ULong_t ULong_t unsigned char prop_list Atom_t Atom_t Atom_t Time_t UChar_t len
Definition TGWin32VirtualXProxy.cxx:249

ROOT::Detail::TRangeCast
Definition TCollection.h:311