Category : neural-network

In Neural Networks and Deep Learning, there’s an object called network3 (which is a PY file, written for python 2.7 and theano 0.7). I modified it to run with python 3.9 and theano 1.0.3. However, when I run the following code (in google colab): import network3 from network3 import Network from network3 import ConvPoolLayer , ..

Read more

I’m creating a image recognition neural network with three layers with these dimensions: 400 features, 40 nodes, 40 nodes, 10 targets (images of digits 0 to 9), therefore these are my weights (theta): theta1 = np.random.uniform(low=0.00001, high=0.0001, size=(40,401)) theta2 = np.random.uniform(low=0.00001, high=0.0001, size=(40,41)) theta3 = np.random.uniform(low=0.00001, high=0.0001, size=(10,41)) I am following Andrew Ng’s approach. I ..

Read more

I’m training my model on around 4000 Chest X-Ray images using Transfer Learning (EfficientNet B1). The training is as follows: Epoch 1/10 99/99 [==============================] – 1437s 14s/step – loss: 0.6446 – accuracy: 0.8617 – val_loss: 0.3128 – val_accuracy: 0.9178 Epoch 2/10 99/99 [==============================] – 193s 2s/step – loss: 0.2179 – accuracy: 0.9356 – val_loss: 0.4521 ..

Read more

I’m trying to train the model. This is the epoch loop seed_val = 17 random.seed(seed_val) np.random.seed(seed_val) torch.manual_seed(seed_val) torch.cuda.manual_seed_all(seed_val) device = torch.device(‘cuda’ if torch.cuda.is_available() else ‘cpu’) # model.to(device) training_stats = [] for epoch_i in range(0, epochs): print("") print(‘======== Epoch {:} / {:} ========’.format(epoch_i + 1, epochs)) print(‘Training…’) # Reset the total loss for this epoch. total_train_loss ..

Read more

I’ve looked at some explanation. Here But I understand what is going wrong I think, but my error occurs not at the loss. For example the snippet where the error is occurring is the line outputs = model(**inputs). # INPUTS # Pulling out the inputs in the form of dictionary inputs = {‘input_ids’: batch[0], ‘attention_mask’: ..

Read more

I’m looking to train a neural network that I can provide two inputs such as the following. input2 = {"instrument": [‘drums’, ‘bass’, ‘guitar’, ‘vocals’], ‘effect’: [[‘reverb’, ‘compressor’, ‘distortion’], [‘flanger’], [], []]} input1 = "drums" And it produce an output like: "instrument_1" This data could change and get quite complex. for example if input1 remains the ..

Read more