Step-by-Step Guide for E-Commerce Startups to Create 3D Product Catalogs Using E2E Cloud

February 16, 2024


The recent integration of artificial intelligence with computer graphics has resulted in significant breakthroughs in the realm of digital content creation. Two of the most notable innovations in this field are Neuralangelo and NeRF (Neural Radiance Fields). These cutting-edge technologies have revolutionized our approach to image synthesis and the capture of 3D scenes, reshaping our understanding of these processes.


Neuralangelo, named after the legendary Renaissance artist Michelangelo, represents a blend of artistic insight with computational capability. Located at the intersection of deep learning and art, this system uses generative adversarial networks (GANs) as well as other neural network architectures to create visually stunning realistic images, paintings, or even sculptures. Through machine learning’s tremendous capabilities, artists and designers can now explore new frontiers in creativity – catalyzing a blurring of distinctions between human imagination and digital representation.


Neural Radiance Fields (NeRF) are a type of fully connected neural network that can generate new perspectives of complex 3D scenes from a subset of 2D images. They are trained to reproduce the appearance of a scene as seen in the input views by using a rendering loss function. To render a complete scene, NeRF interpolates between these input images, which represent different views of the scene. This makes NeRF a powerful tool for image creation in artificial intelligence.

NeRF networks use volume rendering to produce new views, and they are trained to map a 5D input (consisting of viewing direction and spatial location) to a 4D output (color and opacity). However, NeRF is a computationally intensive algorithm, and rendering complex scenes can take several hours or even days. Despite this, recent developments in algorithms have significantly improved its efficiency.

Synthetic views are generated by querying 5D coordinates along the paths of camera rays. The resulting colors and densities are then projected into an image using conventional volume rendering methods. The primary requirement for optimizing our representation is a collection of images accompanied by their known camera poses, as volume rendering inherently allows for differentiation. By effectively optimizing neural radiance fields, we demonstrate the ability to render new, photorealistic views of scenes with intricate geometry and appearance. This approach surpasses previous achievements in neural rendering and view synthesis in terms of results.

Our Problem Statement 

A three-dimensional product catalog is a sophisticated way for customers to interact with products in an online store. It presents each item in three dimensions so they can view it from different perspectives. A 3D product catalog, in contrast to traditional catalogs (with its static images and simple videos), immerses customers in a more dynamic and engaging shopping experience.

The use of 3D models – digital representations of real objects made with computer graphics techniques – is the primary characteristic of a 3D product catalog. These models are extremely realistic in their portrayal of the form, feel, and look of products in a virtual setting. When it comes to product presentation, 3D models provide more flexibility and versatility than just using traditional techniques like photography.

In this blog post, we’ll convert 2D product images into 3D by using NeRF on E2E’s Cloud GPU. 

E2E Networks: Leveraging Its Cloud GPU

Running the Neural Radiance Fields (NeRF) model, or any other computationally intensive deep learning model, on local computers can be challenging, often necessitating the use of cloud-based GPU resources.

The necessity for high-powered GPUs in operating NeRF models stems from the model's architecture and training process, which involve extensive computational demands. A dedicated, high-powered GPU is essential to efficiently handle these requirements.

A typical GPU architecture is shown in the figure below. However, instead of buying advanced GPUs, developers can get access to the same capabilities through a cloud GPU platform.

E2E Networks is a leading hyperscaler from India that focuses on advanced Cloud GPU infrastructure. E2E provides accelerated cloud computing solutions, including cutting-edge Cloud GPUs like A100/H100 and the AI Supercomputer HGX 8xH100 GPUs. We offer a range of advanced cloud GPUs at extremely competitive rates. To learn about the products provided by E2E Networks, visit here. As for the best GPU for Stable Diffusion model implementation, it largely depends on your specific requirements and budget. I used a GPU dedicated compute with A100–80 GB.

The best cloud GPU architectures allow you to access the capabilities offered by the GPU stack, which includes GPU clusters, faster bandwidth, and memory efficiency.

To proceed with E2E Networks, add your SSH key by going to Settings.

Then create a node by going to Compute.

Launch Visual Studio Code and download the Remote Explorer and Remote SSH extensions. Launch a fresh terminal. To gain access to your local system, just enter the code below:

ssh root@

SSH will be used to log you in remotely on your local computer. Let's begin putting the code into practice now.

Implementation with Nerf Model: Generating 3D Model Product Videos for E-Commerce

Let’s download a dataset from Kaggle using the Opendatasets library. It will require your Kaggle Username and API key, which you can access through your Kaggle account by going to Settings.

%pip install opendatasets

import opendatasets as od"")

This command installs the latest version of PyTorch, Torchvision, and Matplotlib.

The torch is used because it is an open-source deep-learning framework that provides tensor computation and GPU acceleration.

%pip install torch torchvision matplotlib

In our VS Code, the Python environment does not have the libraries that we want to use installed. So we’ll start installing all the important libraries.

# Imports
import os
import torch
import torch.nn as nn
import torch.optim as optim
from import DataLoader, Dataset
from torchvision import transforms
from PIL import Image
import numpy as np
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt

# Define a Neural Radiance Field (NeRF) model
class ComplexNeRF(nn.Module):
    def __init__(self, in_features=6, hidden_features=256, out_features=3):
        super(ComplexNeRF, self).__init__()
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.fc2 = nn.Linear(hidden_features, hidden_features)
        self.fc3 = nn.Linear(hidden_features, hidden_features)
        self.fc4 = nn.Linear(hidden_features, out_features)
    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = torch.relu(self.fc2(x))
        x = torch.relu(self.fc3(x))
        x = self.fc4(x)
        return x

The below-described procedures are followed in this implementation, which yields a dictionary with the image, RGB values, and 3D points for every sample.

# Loading Synthetic Dataset (Customize based on your dataset)
class CustomDataset(Dataset):
    def __init__(self, data_folder, transform=None):
        self.data_folder = data_folder
        self.transform = transform
        self.image_list = os.listdir(data_folder)

    def __len__(self):
        return len(self.image_list)

    def __getitem__(self, idx):
        img_path = os.path.join(self.data_folder, self.image_list[idx])
        image ='RGB')

        # Generate random 3D points for each pixel in the image
        height, width = image.size
        points_3d = np.column_stack((np.random.rand(height, width, 1), np.random.rand(height, width, 1), np.random.rand(height, width, 1)))

        # Generate random RGB values for each pixel
        rgb_values = np.random.randint(0, 256, size=(height, width, 3))

        sample = {'points_3d': points_3d, 'rgb_values': rgb_values, 'image': image}

        if self.transform:
            sample = self.transform(sample)
        return sample

This is the sample we received as output.

After completing the data processing, we need to develop a 360-degree video transformation feature for this e-commerce product.

The essential actions needed to carry out the Rescale transformation include emphasis on returning the transformed sample and resizing the image while maintaining the aspect ratio.

# Data processing Transformations
class Rescale(object):
    def __init__(self, output_size):
        assert isinstance(output_size, (int, tuple))
        self.output_size = output_size

    def __call__(self, sample):
        image, points_3d, rgb_values = sample['image'], sample['points_3d'], sample['rgb_values']

        h, w = image.size[:2]
        if isinstance(self.output_size, int):
            if h > w:
                new_h, new_w = self.output_size, int(self.output_size * w / h)
                new_h, new_w = int(self.output_size * h / w), self.output_size
            new_h, new_w = self.output_size
        new_h, new_w = int(new_h), int(new_w)
        img = transforms.Resize((new_h, new_w))(image)
        return {'image': img, 'points_3d': points_3d, 'rgb_values': rgb_values}

This hint gives instructions on how to use the ToTensor transformation to create PyTorch tensors from the image, RGB values, and 3D points.

class ToTensor(object):
    def __call__(self, sample):
        image, points_3d, rgb_values = sample['image'], sample['points_3d'], sample['rgb_values']
        img = transforms.ToTensor()(image)

        return {'image': img, 'points_3d': torch.tensor(points_3d, dtype=torch.float32),
                'rgb_values': torch.tensor(rgb_values, dtype=torch.float32)}

By minimizing the MSE loss between the predicted and ground truth 3D points, this function trains the model. It optimizes using the Adam optimizer.

# Training function
def train_complex_nerf(model, train_loader, num_epochs=10, lr=0.001):
    criterion = nn.MSELoss()
    optimizer = optim.Adam(model.parameters(), lr=lr)
    for epoch in range(num_epochs):
        for batch in train_loader:
            inputs, targets = batch['image'], batch['points_3d']
            outputs = model(inputs)
            loss = criterion(outputs, targets)

        print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}')

This feature indicates how well the model uses the input images to reconstruct the 3D scene.

# Visualization function
def visualize_3d_reconstruction(model, test_loader):
    with torch.no_grad():
        for batch in test_loader:
            inputs, targets = batch['image'], batch['points_3d']
            outputs = model(inputs)

            # Visualizing the 3D reconstruction
            fig = plt.figure()
            ax = fig.add_subplot(111, projection='3d')
            ax.scatter(targets[:, 0], targets[:, 1], targets[:, 2], c='r', marker='o', label='Ground Truth')
            ax.scatter(outputs[:, 0], outputs[:, 1], outputs[:, 2], c='b', marker='s', label='Reconstruction')

It shows how to load and prepare datasets, train NeRF, and view the results of the 3D reconstruction.

# Data processing, applying transformations
data_transform = transforms.Compose([Rescale(256), ToTensor()])
# Load train dataset
synthetic_dataset = CustomDataset(data_folder='/NERF/fashion-images/data/Apparel/Boys/Images/images_with_product_ids', transform=data_transform)
train_loader = DataLoader (synthetic_dataset, batch_size=64, shuffle=True)
# Load test dataset
test_dataset = CustomDataset(data_folder='/NERF/fashion-images/data/Apparel/Girls/Images/images_with_product_ids', transform=data_transform)
test_loader = DataLoader (test_dataset, batch_size=1, shuffle=False)
# Create and train a NeRF model
complex_nerf_model = ComplexNeRF()
train_complex_nerf(complex_nerf_model, train_loader)
# Visualize 3D reconstruction on the test dataset
visualize_3d_reconstruction(complex_nerf_model, test_loader)

Voila! The following are the 3D videos as sample outputs. 

Product 1 - Rotating 3D video of a pair of trousers.

Product 2 - Rotating 3D video of a t-shirt.

This process can be used by any e-commerce firm to convert still images to engaging 3D videos.


In conclusion, the Stable Diffusion model's fine-tuning for e-commerce image generation was greatly improved by integrating E2E Networks' A100–80 GB GPU dedicated compute. The computational power of the A100 GPU effectively handled complex model operations, leading to faster training and seamless processing.

The versatility of the A100 allowed for quick experimentation and effective model customization through fine-tuning unique datasets. The A100 GPU guaranteed responsiveness for real-time image generation, cutting down on training times and improving user experience. 

In summary, the synergistic environment that was created by the partnership between E2E Networks’ A100 GPU and Stable Diffusion model’s fine-tuning was marked by accessibility, computational efficiency, and accelerated model training, making the process of creating 3D content for e-commerce both efficient and pleasurable.

Latest Blogs
This is a decorative image for: A Complete Guide To Customer Acquisition For Startups
October 18, 2022

A Complete Guide To Customer Acquisition For Startups

Any business is enlivened by its customers. Therefore, a strategy to constantly bring in new clients is an ongoing requirement. In this regard, having a proper customer acquisition strategy can be of great importance.

So, if you are just starting your business, or planning to expand it, read on to learn more about this concept.

The problem with customer acquisition

As an organization, when working in a diverse and competitive market like India, you need to have a well-defined customer acquisition strategy to attain success. However, this is where most startups struggle. Now, you may have a great product or service, but if you are not in the right place targeting the right demographic, you are not likely to get the results you want.

To resolve this, typically, companies invest, but if that is not channelized properly, it will be futile.

So, the best way out of this dilemma is to have a clear customer acquisition strategy in place.

How can you create the ideal customer acquisition strategy for your business?

  • Define what your goals are

You need to define your goals so that you can meet the revenue expectations you have for the current fiscal year. You need to find a value for the metrics –

  • MRR – Monthly recurring revenue, which tells you all the income that can be generated from all your income channels.
  • CLV – Customer lifetime value tells you how much a customer is willing to spend on your business during your mutual relationship duration.  
  • CAC – Customer acquisition costs, which tells how much your organization needs to spend to acquire customers constantly.
  • Churn rate – It tells you the rate at which customers stop doing business.

All these metrics tell you how well you will be able to grow your business and revenue.

  • Identify your ideal customers

You need to understand who your current customers are and who your target customers are. Once you are aware of your customer base, you can focus your energies in that direction and get the maximum sale of your products or services. You can also understand what your customers require through various analytics and markers and address them to leverage your products/services towards them.

  • Choose your channels for customer acquisition

How will you acquire customers who will eventually tell at what scale and at what rate you need to expand your business? You could market and sell your products on social media channels like Instagram, Facebook and YouTube, or invest in paid marketing like Google Ads. You need to develop a unique strategy for each of these channels. 

  • Communicate with your customers

If you know exactly what your customers have in mind, then you will be able to develop your customer strategy with a clear perspective in mind. You can do it through surveys or customer opinion forms, email contact forms, blog posts and social media posts. After that, you just need to measure the analytics, clearly understand the insights, and improve your strategy accordingly.

Combining these strategies with your long-term business plan will bring results. However, there will be challenges on the way, where you need to adapt as per the requirements to make the most of it. At the same time, introducing new technologies like AI and ML can also solve such issues easily. To learn more about the use of AI and ML and how they are transforming businesses, keep referring to the blog section of E2E Networks.

Reference Links

This is a decorative image for: Constructing 3D objects through Deep Learning
October 18, 2022

Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

3D reconstruction is one of the most complex issues of deep learning systems. There have been multiple types of research in this field, and almost everything has been tried on it — computer vision, computer graphics and machine learning, but to no avail. However, that has resulted in CNN or convolutional neural networks foraying into this field, which has yielded some success.

The Main Objective of the 3D Object Reconstruction

Developing this deep learning technology aims to infer the shape of 3D objects from 2D images. So, to conduct the experiment, you need the following:

  • Highly calibrated cameras that take a photograph of the image from various angles.
  • Large training datasets can predict the geometry of the object whose 3D image reconstruction needs to be done. These datasets can be collected from a database of images, or they can be collected and sampled from a video.

By using the apparatus and datasets, you will be able to proceed with the 3D reconstruction from 2D datasets.

State-of-the-art Technology Used by the Datasets for the Reconstruction of 3D Objects

The technology used for this purpose needs to stick to the following parameters:

  • Input

Training with the help of one or multiple RGB images, where the segmentation of the 3D ground truth needs to be done. It could be one image, multiple images or even a video stream.

The testing will also be done on the same parameters, which will also help to create a uniform, cluttered background, or both.

  • Output

The volumetric output will be done in both high and low resolution, and the surface output will be generated through parameterisation, template deformation and point cloud. Moreover, the direct and intermediate outputs will be calculated this way.

  • Network architecture used

The architecture used in training is 3D-VAE-GAN, which has an encoder and a decoder, with TL-Net and conditional GAN. At the same time, the testing architecture is 3D-VAE, which has an encoder and a decoder.

  • Training used

The degree of supervision used in 2D vs 3D supervision, weak supervision along with loss functions have to be included in this system. The training procedure is adversarial training with joint 2D and 3D embeddings. Also, the network architecture is extremely important for the speed and processing quality of the output images.

  • Practical applications and use cases

Volumetric representations and surface representations can do the reconstruction. Powerful computer systems need to be used for reconstruction.

Given below are some of the places where 3D Object Reconstruction Deep Learning Systems are used:

  • 3D reconstruction technology can be used in the Police Department for drawing the faces of criminals whose images have been procured from a crime site where their faces are not completely revealed.
  • It can be used for re-modelling ruins at ancient architectural sites. The rubble or the debris stubs of structures can be used to recreate the entire building structure and get an idea of how it looked in the past.
  • They can be used in plastic surgery where the organs, face, limbs or any other portion of the body has been damaged and needs to be rebuilt.
  • It can be used in airport security, where concealed shapes can be used for guessing whether a person is armed or is carrying explosives or not.
  • It can also help in completing DNA sequences.

So, if you are planning to implement this technology, then you can rent the required infrastructure from E2E Networks and avoid investing in it. And if you plan to learn more about such topics, then keep a tab on the blog section of the website

Reference Links

This is a decorative image for: Comprehensive Guide to Deep Q-Learning for Data Science Enthusiasts
October 18, 2022

A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

For all data science enthusiasts who would love to dig deep, we have composed a write-up about Q-Learning specifically for you all. Deep Q-Learning and Reinforcement learning (RL) are extremely popular these days. These two data science methodologies use Python libraries like TensorFlow 2 and openAI’s Gym environment.

So, read on to know more.

What is Deep Q-Learning?

Deep Q-Learning utilizes the principles of Q-learning, but instead of using the Q-table, it uses the neural network. The algorithm of deep Q-Learning uses the states as input and the optimal Q-value of every action possible as the output. The agent gathers and stores all the previous experiences in the memory of the trained tuple in the following order:

State> Next state> Action> Reward

The neural network training stability increases using a random batch of previous data by using the experience replay. Experience replay also means the previous experiences stocking, and the target network uses it for training and calculation of the Q-network and the predicted Q-Value. This neural network uses openAI Gym, which is provided by taxi-v3 environments.

Now, any understanding of Deep Q-Learning   is incomplete without talking about Reinforcement Learning.

What is Reinforcement Learning?

Reinforcement is a subsection of ML. This part of ML is related to the action in which an environmental agent participates in a reward-based system and uses Reinforcement Learning to maximize the rewards. Reinforcement Learning is a different technique from unsupervised learning or supervised learning because it does not require a supervised input/output pair. The number of corrections is also less, so it is a highly efficient technique.

Now, the understanding of reinforcement learning is incomplete without knowing about Markov Decision Process (MDP). MDP is involved with each state that has been presented in the results of the environment, derived from the state previously there. The information which composes both states is gathered and transferred to the decision process. The task of the chosen agent is to maximize the awards. The MDP optimizes the actions and helps construct the optimal policy.

For developing the MDP, you need to follow the Q-Learning Algorithm, which is an extremely important part of data science and machine learning.

What is Q-Learning Algorithm?

The process of Q-Learning is important for understanding the data from scratch. It involves defining the parameters, choosing the actions from the current state and also choosing the actions from the previous state and then developing a Q-table for maximizing the results or output rewards.

The 4 steps that are involved in Q-Learning:

  1. Initializing parameters – The RL (reinforcement learning) model learns the set of actions that the agent requires in the state, environment and time.
  2. Identifying current state – The model stores the prior records for optimal action definition for maximizing the results. For acting in the present state, the state needs to be identified and perform an action combination for it.
  3. Choosing the optimal action set and gaining the relevant experience – A Q-table is generated from the data with a set of specific states and actions, and the weight of this data is calculated for updating the Q-Table to the following step.
  4. Updating Q-table rewards and next state determination – After the relevant experience is gained and agents start getting environmental records. The reward amplitude helps to present the subsequent step.  

In case the Q-table size is huge, then the generation of the model is a time-consuming process. This situation requires Deep Q-learning.

Hopefully, this write-up has provided an outline of Deep Q-Learning and its related concepts. If you wish to learn more about such topics, then keep a tab on the blog section of the E2E Networks website.

Reference Links

This is a decorative image for: GAUDI: A Neural Architect for Immersive 3D Scene Generation
October 13, 2022

GAUDI: A Neural Architect for Immersive 3D Scene Generation

The evolution of artificial intelligence in the past decade has been staggering, and now the focus is shifting towards AI and ML systems to understand and generate 3D spaces. As a result, there has been extensive research on manipulating 3D generative models. In this regard, Apple’s AI and ML scientists have developed GAUDI, a method specifically for this job.

An introduction to GAUDI

The GAUDI 3D immersive technique founders named it after the famous architect Antoni Gaudi. This AI model takes the help of a camera pose decoder, which enables it to guess the possible camera angles of a scene. Hence, the decoder then makes it possible to predict the 3D canvas from almost every angle.

What does GAUDI do?

GAUDI can perform multiple functions –

  • The extensions of these generative models have a tremendous effect on ML and computer vision. Pragmatically, such models are highly useful. They are applied in model-based reinforcement learning and planning world models, SLAM is s, or 3D content creation.
  • Generative modelling for 3D objects has been used for generating scenes using graf, pigan, and gsn, which incorporate a GAN (Generative Adversarial Network). The generator codes radiance fields exclusively. Using the 3D space in the scene along with the camera pose generates the 3D image from that point. This point has a density scalar and RGB value for that specific point in 3D space. This can be done from a 2D camera view. It does this by imposing 3D datasets on those 2D shots. It isolates various objects and scenes and combines them to render a new scene altogether.
  • GAUDI also removes GANs pathologies like mode collapse and improved GAN.
  • GAUDI also uses this to train data on a canonical coordinate system. You can compare it by looking at the trajectory of the scenes.

How is GAUDI applied to the content?

The steps of application for GAUDI have been given below:

  • Each trajectory is created, which consists of a sequence of posed images (These images are from a 3D scene) encoded into a latent representation. This representation which has a radiance field or what we refer to as the 3D scene and the camera path is created in a disentangled way. The results are interpreted as free parameters. The problem is optimized by and formulation of a reconstruction objective.
  • This simple training process is then scaled to trajectories, thousands of them creating a large number of views. The model samples the radiance fields totally from the previous distribution that the model has learned.
  • The scenes are thus synthesized by interpolation within the hidden space.
  • The scaling of 3D scenes generates many scenes that contain thousands of images. During training, there is no issue related to canonical orientation or mode collapse.
  • A novel de-noising optimization technique is used to find hidden representations that collaborate in modelling the camera poses and the radiance field to create multiple datasets with state-of-the-art performance in generating 3D scenes by building a setup that uses images and text.

To conclude, GAUDI has more capabilities and can also be used for sampling various images and video datasets. Furthermore, this will make a foray into AR (augmented reality) and VR (virtual reality). With GAUDI in hand, the sky is only the limit in the field of media creation. So, if you enjoy reading about the latest development in the field of AI and ML, then keep a tab on the blog section of the E2E Networks website.

Reference Links

Build on the most powerful infrastructure cloud

A vector illustration of a tech city using latest cloud technologies & infrastructure