machine learining

₹750-1250 INR / hour

종료됨

게시됨

7개월 전

₹750-1250 INR / hour

we have to do the follwing, for my dataset not the author's : We used an encoder with a structure similar to the classification network used in DGCNN (Wang et al., 2019). The input point cloud with dimension 3×512 passes through five EdgeConv layers with point-wise latent space dimensions (64, 64, 128, 256) and a max pooling layer (we do not use a batch normalization layer unlike the original DGCNN classification network), then we can obtain a 1024-dimensional feature vector. Other settings are the same (e.g., k = 20, leaky relu activation) Then this feature vector again passes through three fully-connected neural networks with dimension (512, 256, 2) with leaky relu activation function and linear output activation function; the latent space is two-dimensional. For the decoder model, we simply use a fully-connected neural network as the decoder. The two-dimensional vector on the latent space passes through three fully-connected neural networks with dimension (256, 512, 3×512) with relu activation function and linear output activation function; the output is a 3D point cloud with the number of points 512. Section 5.1.1: To train the networks, we use ADAM with a learning rate of 0.001 and batch size of 16; the total number of the epochs is 500. The mean value of MEDs of the dataset is 0.0339, and we use the bandwidth value k to 0.5. We use Chamfer distance as the reconstruction loss; code: [login to view URL]

Engineering

Machine Learning (ML)

Python

Pytorch

Deep Learning

프로젝트 ID: 37375961

프로젝트 정보

13 제안서

원격근무 프로젝트

활동 중 5개월 전

돈을 좀 벌 생각이십니까?

이메일 주소

프리랜서 입찰의 이점

예산 및 기간 설정

작업 결과에 대한 급여 수급

제안의 개요를 자세히 쓰세요

무료로 프로젝트에 신청하고 입찰할 수 있습니다

13 이 프로젝트에 프리랜서들의 평균 입찰은 ₹1,008 INR입니다./시간

@sajjadtaghvaeifr

Hi, I hope you are doing fine. I have almost 10 years of experience in machine learning algorithms. I can implement various types of artificial intelligence algorithms including yours with Matlab, Python and etc. I have PhD from Tohoku University and have several journal publications on the subjects. You can see portfolio for my previous projects. I read about your project and am interested in working with you. Please send me a message so that we can discuss more. Best regards.

₹1,000 INR 40일에

4.9

(123 건의 리뷰)

7.6

@rochaksharma7

Hello! My name is Rochak and I am a top 1% elite Python freelancer. I have been working with Python for 8 years and have a lot of experience in machine learning (ML), Python, artificial intelligence (AI), computer vision (CV), data science and deep learning. I understand that you need to do the following for your dataset: * Developing a machine learning system to classify point clouds with dimensions 3×512. * Using an encoder with a structure similar to the classification network used in DGCNN (Wang et al., 2019). * Using point-wise latent space dimensions 64, 64, 128, 256 and max pooling layer.

₹1,000 INR 40일에

5.0

(39 건의 리뷰)

5.0

@ChiSquareX

Chi SquareX provide services in the domain of Machine Learning, AI, Deep Learning, Computer Vision, Natural Language Processing, Algo Trading. Expertise: Training Machine learning, Deep Learning, Reinforcement Learning Models Frameworks are compatible with Pytorch, Tensorflow, Keras. Training ML models like Bayesian Networks, XGBoost, Random Forest, Decision Tree, KNN Classifier, GBM, GLM, SVM. Training DL models BERT and all its variants, DNN, CNN, LSTM, GRU, Variational Autoencoders using Bayesian Neural Networks. eXplainable AI (XAI) using LIME, SHAP, GRAD-CAM, DeepLIFT, DeepRED, CausalNex. Here are some of our previous works: Hair transplant startups require some image generation for their customers. Masters Thesis in Art Generation using GANs Head Gesture recognition using temporal signal data Hairstyle Classification Heartbeat recognition using Multilingual BERT for Document Ranking Algorithmic Trading on Indian Stock Market Yolov5 for Object Detection Kinetics 700 classification Aircraft Optimization Bio-BERT for Medical Dataset SciBERT for Explanation regeneration I have read your project details and I am confident that I can contribute towards it. Please message me so that we can talk further about your project.

₹1,000 INR 40일에

5.0

(24 건의 리뷰)

4.8

@usamae

For your dataset, we will implement an encoder-decoder network inspired by the structure of DGCNN (Wang et al., 2019), with a few modifications to suit your specific requirements. The input point cloud, which is of dimensions 3×512, will traverse through five EdgeConv layers with varying point-wise latent space dimensions (64, 64, 128, 256) and a max pooling layer. Unlike the original DGCNN classification network, we will not incorporate a batch normalization layer. This process will yield a 1024-dimensional feature vector. Following this, the feature vector will pass through three fully-connected neural networks with dimensions (512, 256, 2) and appropriate activation functions. The latent space will be two-dimensional. For the decoder model, we will use a fully-connected neural network. The two-dimensional vector in the latent space will navigate through three fully-connected neural networks with dimensions (256, 512, 3×512) to generate a 3D point cloud with 512 points. To train the networks, we will utilize the ADAM optimizer with a learning rate of 0.001, a batch size of 16, and a total of 500 epochs. The reconstruction loss will be measured using the Chamfer distance, and the bandwidth value 'k' will be set to 0.5.

₹1,000 INR 40일에

5.0

(1 건의 리뷰)

4.1

@mapsbox

Hello, I am very familiar with the requirements of your projects, and I can complete them quickly. Let's connect over chat to discuss this further. Regards,

₹1,000 INR 40일에

5.0

(6 건의 리뷰)

3.5

@EvanMartin419

Hi, Dear Client. I provide you with top expertise in training and testing deep learning models using Python. I have rich of experienced Deep learning for five years. I can make your AI project with full source code which includes data visualization, deep learning, classification neural network and predicting neural network models and etc. I am available now and can start right away. So let's connect and start the project. Best Regards.

₹1,000 INR 40일에

0.0

(0 건의 리뷰)

0.0

@sumeetkumar309

Hello there, looks interesting from the first look, I'll give a skeleton approach to start this with, have a look, and if you find it a good starting point, hit me up. Approach: 1. Data Preprocessing: Prepare and preprocess your dataset to match the required input format (3x512). Normalize and clean the data as needed for training stability. 2. Encoder Architecture: Implement an encoder network with the specified structure: Five EdgeConv layers with point-wise latent space dimensions (64, 64, 128, 256). Utilize max pooling without batch normalization. Apply Leaky ReLU activation after each layer. 3. Fully-Connected Layers (Latent Space): Design three fully-connected layers with dimensions (512, 256, 2). Apply Leaky ReLU activation functions for non-linearity. Use a linear activation function for the final layer to produce a 2-dimensional latent space representation. 4. Decoder Architecture: Three fully-connected layers with dimensions (256, 512, 3x512). Apply ReLU activation after the first two layers and linear activation function for the final layer. The output should be a 3D point cloud with 512 points. 5. Training: Use the Chamfer distance as the reconstruction loss function to minimize the difference between the input and output point clouds. 6. Evaluation: Assess the model's performance by calculating the Mean Euclidean Distance (MED) on the reconstructed point clouds.

₹1,100 INR 20일에