Questions tagged [tensorflow]

For questions related to Google's open-source library for machine learning and machine intelligence. However, note that programming questions are off-topic here.

201 questions with no upvoted or accepted answers
Filter by
Sorted by
Tagged with
6
votes
0answers
167 views

How to design 4D Deep Recurrent Neural Networks using Tensorflow?

I want to design a simple model that predicts the movement of coordinates with RNNs. In a typical three-dimensional LSTM model, one feature is encoded as one hot encoding, and the ...
5
votes
0answers
44 views

It is possible to use deep learning to give approximate solutions to NP-hard graph theory problems?

It is possible to use deep learning to give approximate solutions to NP-hard graph theory problems? If we take, for example, the travelling salesman problem (or the dominating set problem). Let's say ...
5
votes
1answer
62 views

Validation accuracy higher than training accurarcy

I implemented the unet in tensorflow for the segmentation of MRI images of the thigh. I noticed I always get a higher validation accuracy by a small gap, independently of the initial split. One ...
4
votes
0answers
225 views

YOLO v3 complete architecture

I am attempting to implement YOLO v3 in Tensorflow-Keras from scratch, with the aim of training my own model on a custom dataset. By that, I mean without using pretrained weights. I have gone through ...
4
votes
0answers
153 views

Game AI - Modify image classification model for analog output

I'm developing a Game AI which tries to master racing simulation. I already trained a CNN (alexnet) on ingame footage of me playing the game and the pressed keys as the target. I had two main issues ...
3
votes
0answers
27 views

If I want to predict two unrelated values given the same sequence of data points, should I have a model with two outputs or two models?

I want to predict two separate y-values (not really logically connected) based on an input sequence of data (values x). Using LSTM cells. Should I train two models separately or should I just increase ...
3
votes
0answers
44 views

Why shouldn't batch normalisation layers be learnable during fine-tuning?

I have been reading this TensorFlow tutorial on transfer learning, where they unfroze the whole model and then they say: When you unfreeze a model that contains ...
3
votes
0answers
99 views

Doing backpropagation in an Tensorflow.js Neural Network

I have a neural network (which I am making from scratch). In order to make the neural network "learn" I need to conduct back-propagation. Using the code at the below how would I conduct back-...
3
votes
0answers
81 views

Understanding the TensorFlow implementation of the policy gradient method

I was trying to understand the implementation of a basic policy gradient (REINFORCE) method using TensorFlow. I think I got almost everything. The only thing that still bothers me is the loss function ...
3
votes
0answers
144 views

How can I increase the speed and performance of my implementation of an AI for Reversi?

I made an AI for Reversi, aka Othello (8×8), like Alpha Zero, using this book. This book is written in Japanese. The source code of the AI I implemented can be found in this Github repository. There ...
3
votes
0answers
207 views

How to use TPU for real-time low-latency inference?

I use Google's Cloud TPU hardware extensively using Tensorflow for training models and inference, however, when I run inference I do it in large batches. The TPU takes about 3 minutes to warm up ...
3
votes
0answers
604 views

Getting worse performance when training a pre-trained model with the existing class

I am training pre-trained SSD-InceptionV2-Coco to detect the "car", which is one of the classes in mscoco label. I train the model with ~50k sample from KITTI, 500k iteration with batch size 2. I ...
3
votes
1answer
122 views

If neurons are only defined for values between 0 and 1, how does ReLU differ from the identity?

I'm struggling to understand the underlying mechanics of CNNs so any help is appreciated. I have a network with a ReLU activation function which does perform signifigantly better than one with sigmoid....
2
votes
0answers
48 views

What parameters or hyper-parameters of my model for time-series should I change to improve the MAE?

The following time series exercise is about writing the best possible model, minimizing the MAE. Helper functions normalize_series, ...
2
votes
0answers
39 views

Hand Landmark Detector Not Converging

I'm currently trying to train a custom model with TensorFlow to detect 17 landmarks/keypoints on each of 2 hands shown in an image (fingertips, first knuckles, bottom knuckles, wrist, and palm), for ...
2
votes
0answers
35 views

Positional Encoding in Transformer on multi-variate time series data hurts performance

I set up a transformer model that embeds positional encodings in the encoder. The data is multi-variate time series-based data. As I just experiment with the positional encoding portion of the code I ...
2
votes
0answers
53 views

What causes high differences in neural network accuracy each run?

I trained a CNN using Keras in R to multi-dimensional image data for image classification of five classes. I realized that each run (I retrained the network on the same data for ten times), although I ...
2
votes
0answers
18 views

Neural Network Regression Experiment Going Wrong

I've been trying to get a simple regression experiment going with a neural network and I would like some help interpreting what is going wrong. My goal is to see what level of regression accuracy I ...
2
votes
0answers
26 views

What's new in LaBSE v2?

I can't find what's new in LaBSE v2 (https://tfhub.dev/google/LaBSE/2). What are the main highlights of v2 versus v1? And how did you find out?
2
votes
3answers
74 views

How can I model any structure for a neural network?

Hello I am currently doing research on the effect of altering a neural network's structure. Particularly I am investigating what affect would putting a random DAG (directed acyclic graph) in the ...
2
votes
0answers
36 views

CNN leaf segmentation throught classification of edges how to improve

I am trying to design a CNN that can do pixel wise segmentation of edges leaves in dense foliage agriculture images. Such as these: On the basis of this article https://arxiv.org/pdf/1904.03124.pdf, ...
2
votes
0answers
30 views

NN to find arbitrary transformation

Problem description I'm creating a clock with 4 seven-segment LED displays. In an effort to get more familiar with tensorflow, I figured I should try to drive this clock with use of a Neural Network. ...
2
votes
0answers
32 views

Single-value loss/training in a CNN with a tensor output

I am playing around with an idea of using using Q-learning with a DQN (Deep Q-Network), to determine the optimal position of a number of 'units' on a grid of allowed locations, according to some ...
2
votes
0answers
61 views

Why do I get higher average dice accuracy for less data

I am working on image segmentation of MRI thigh images with deep learning (Unet). I noticed that I get a higher average dice accuracy over my predicted masks if I have less samples in the test data ...
2
votes
0answers
24 views

Why does the loss stops reducing after a point in this Transformer Model?

Context I was making a Transformer Model to convert English Sentences to German Sentences. But the loss stops reducing after some time. Code ...
2
votes
0answers
30 views

Loss function decays linearly in segmentation MRI fascia

I am working on a segmentation of MRI images of the thigh. I am trying to segment the fascia, there is a slight imbalance between the background and the mask. I have about 1400 images from 30 patients ...
2
votes
0answers
448 views

Why is DDPG not learning and it does not converge?

I have used a different setting, but DDPG is not learning and it does not converge. I have used these codes 1,2, and 3 and I used different optimizers, activation functions, and learning rate but ...
2
votes
0answers
38 views

Incorporating domain knowledge into recurrent network

I am currently trying to solve a classification task with a recurrent artificial neural network (RNN). Situation There are up to 350 inputs (X) mapped on one categorical output (y)(13 differnt ...
2
votes
0answers
33 views

How to use one-hot encoding for multiple columns (multi-class) with varying number of labels in each class?

I am a beginner in TensorFlow as well as in AI. I am basically from Pharma background and learning AI from scratch. I have data with 5038 input (Float64) and 826 output (Categorical - Multi Labels in ...
2
votes
1answer
286 views

Finding patterns in binary files using deep learning

I am a newbie in deep learning and wanted to know if the problem I have at hand is a suitable fit for deep learning algorithms. I have thousands of fragments each of about 1000 bytes size (i.e. ...
2
votes
1answer
274 views

Why isn't my implementation of DQN using TensorFlow on the FrozenWorld environment working?

I am trying to test DQN on FrozenWorld environment in gym using TensorFlow 2.x. The update rule is (off policy) $$Q(s,a) \leftarrow Q(s,a)+\alpha (r+\gamma~ max_{a'}Q(s',a')-Q(s,a))$$ I am using an ...
2
votes
0answers
78 views

How do I make my LSTM model more sensitive to changes in the sequence?

I have a many to one LSTM model for multiclass classification. For reference, this is the architecture of the model ...
2
votes
0answers
74 views

External GPU for Mac

I'd like to buy an eGPU for my MaxBook Pro to use for simple deep learning tasks. My setup is: ...
2
votes
0answers
154 views

What is the equivalent PyTorch version of tensorflow lite

Update Checked the PyTorch Mobile which is designed to Android and iOS. Although according to the document, it says it can build for ARM CPUs, but there isn't any documentation mention about how to ...
2
votes
0answers
21 views

How to handle set-like size agnostic input format

Let's set up some hypothetical simplified scenario: Each instance $i$ of my imaginary dataset $D=\{i_{1}, \ldots, i_{MAX}\}$ has different number $k_{i}$ of $n$-dimensional vectors as input into my ...
2
votes
0answers
35 views

How can I train a Deep Learning model using degraded photos and their clean version to correct photos

I have 5000 degraded pictures ( pixelated, blurry, too much luminosity ... ) and their clean versions, and I would like to train a model so that it can predict how to correct future pictures. I've ...
2
votes
0answers
27 views

Can't figure out what's going wrong with my dataset construction for multivariate regression

TL;DR: I can't figure out why my neural network wont give me a sensible output. I assume it's something to do with how I'm presenting the input data to it but I have no idea how to fix it. Background:...
2
votes
0answers
27 views

How are batch statistics computed in Recurrent Batch Normalization?

I'm implementing recurrent BN per this paper in Keras, but looking at it and those citing it, a detail remains unclear to me: how are batch statistics computed? Authors omit explicit clarification, ...
2
votes
0answers
40 views

How should I make output layer of my neural network so that I can get outputs ranging from [-20,-1]

I am trying to make a neural network which takes in 0 and 1 as it's input and should give me output ranging from [-20,-1].I am using three layers with sigmoid as the activation function .How should I ...
2
votes
0answers
55 views

How to implement loss function of H-GAN model

I was trying to implement the loss function of H-GAN. Here is my code . But it seem somethings wrong, maybe is recognition loss on z (EQ 9). I used the EQ 5 on MISO to calculate it. Here is my code: ...
2
votes
0answers
314 views

How do I create a chatbot using tensorflow or pytorch using like the one defined in dialogflow?

How do I create a chatbot using TensorFlow or PyTorch using like the one defined in DialogFlow? What are the best datasets that I can use so to create my own personal assistant like google assistant? ...
2
votes
0answers
27 views

How to voxelize multiple frames at the time and append them together?

I'm trying to implement this approach for object detection and tracking. In this approach, the first step is voxelize each frame to construct a 3D tensor, the second step is to append multiple voxels ...
2
votes
1answer
68 views

Why feeding the correct output as input during training of seq2seq models?

So, I've read about seq2seq for time-series and it seemed really promising, but when I went to implement it, all the tutorial I've found use the correct output as input to the decoder phase during ...
2
votes
0answers
18 views

In addition to matrix algebra, can GPU's also handle the various Kernel functions for Neural Networks?

I've read a number of articles on how GPUs can speed up matrix algebra calculations, but I'm wondering how calculations are performed when one uses various kernel functions in a neural network. If ...
2
votes
0answers
46 views

RNN: Different test results on balanced and unbalanced data

I trained a recurrent neural network (if it matters - it contains three CuDNNLSTM cells and 3 Dense layers, Dropout = 0.2). The result of data preparation is one array of ~330.000 sequences. Each ...
2
votes
0answers
682 views

Can we combine multiple different neural networks in one?

I want to make a kind of robotic brain, i.e. a big neural network, which includes an NLP model (for understanding human voice), real-time object recognition system (so that it can identify particular ...
2
votes
0answers
435 views

How many episodes does it take for a vanilla one-step actor-critic agent to master the OpenAI BipedalWalker-v2 problem?

I'm trying to solve the OpenAI BipedalWalker-v2 by using a one-step actor-critic agent. I'm implementing the solution using python and tensorflow. I'm following this pseudo-code taken from the book ...
2
votes
0answers
30 views

Using two generative adversarial nets to classify articles - what is a good approach?

I'm trying to create a deep learning network to classify news article based on the text and associated image. The idea comes from a novel use of GANs to classify based on generated data. My approach ...
2
votes
0answers
143 views

How to measure times on TensorFlow Slim?

I would like to measure time for forward and backward times on TF-Slim (over all network and per-layer) like caffe does. However, it just logs the step/iteration time, and I have no idea of how to do ...
2
votes
0answers
138 views

Wide & Deep Learning Explanation

I was going through "Wide & Deep Learning" tensorflow tutorial & it's quite simply explained the process. But I missed few of the things. If someone can please explain them to me, it will be ...

1
2 3 4 5