In deep learning training, I notice some weird stuff. Let me try to explain. The overall number of training occurrences during laptop training differs from GPU training. Using TensorFlow Keras, I created a Convolutional Neural Network. My training instances total 31561, and my batch size is set at 32. Training on Laptop When the code is ..
I have a question regarding the use of UUID identifiers when doing machine learning. Let’s say I’m building a collaborative filtering based movie recommender using a version of the MovieLens dataset. However, instead of having the user IDs as integers, I’m using UUIDs. For training, I can process the data in at least a couple ..
I’m using Google Vision OCR for blood results. Users can take picture of their blood results and my job is to extract meaningful informations. For example, in the picture below, I might want to extract the number related to the TSH (here: 0.722). Example: But the pictures I get are never that straight, they are ..
I’m trying to convert my pytorch model to pytorch lightning, and I couldn’t find any example similar to my own online to confirm. This is my model: import torch import torch.nn as nn from resnet import ResNet, Bottleneck class KeyPointDetector(ResNet): def __init__(self, **kwargs): super().__init__(Bottleneck, [3, 4, 6, 3], **kwargs) self.conv1 = nn.Conv2d(1, 64, kernel_size=7, stride=2, ..
Yet I am working on project where I am using dlib 68 landmarks. But its not enough stable for further use. In some research I found some dataset to train facial landmarks. But I am new to dlib. I don’t know more about .dat file and dlib. But I found some python source code for ..
I am working on a classifier project based on udemy course(Deep Learning Computer Vision™ CNN, OpenCV, YOLO, SSD & GANs) The system i use has configuration below: Corei7 CPU, 32GB RAM, windows10, gtx1060 by 6GB Memory GPU, msvc 2019, tensorflow 2.4.0, cuda toolkit 11.0, cudnn 8.0 (the config is set based on tensorflow official website ..
I am working on APML (severe type of Blood cancer) Detection, a classification problem, I collected data from real environment which is imbalanced. I have just 31 samples for positive class and 574 samples for negative class. I build an SVM model, but not getting good results. I have also tried PCA, up sampling and ..
I need help on how to pullout my pertained weights and baises from my PyTorch cnn model. I would like to use them with other language so I prefer to have them on binary or text form. I tried but the only way I found is cvs which won’t work well with ,y c++ code. ..
I’m trying to split my data into training, validation, and test sets using Fast_ml for a machine learning purpose. Both my input and output data are read from .npy files through np.load. The input "P" is an array with the shape of (100000, 4, 4, 6, 1) and the target "Q" is a vector of ..
I’m running into an incompatible shapes error that I am not able to trace. I’m trying to use the code provided here: https://data-flair.training/blogs/face-mask-detection-with-python/ and I’m working on google colab. I’m at step 5 where I train the model with model.fit_generator() which is where the [10,2] vs [10,3] error happens, Using fit() gets the same error. ..