You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+9-8Lines changed: 9 additions & 8 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -22,22 +22,23 @@ Once deployed, the guidance features a full 3D reconstruction back-end system wi
22
22
4.**Gaussian Splat Training**: Optimization of 3D Gaussian primitives to represent the scene using AI/ML
23
23
5.**Export & Delivery**: Generation of the final 3D asset in standard formats for easy viewing and notification via email
24
24
25
-
By deploying this guidance, users gain access to a flexible infrastructure that handles the entire 3D reconstruction process programatically, from media upload to final 3D model delivery, while being highly modular through its componentized pipeline-based approach. This guidance addresses the significant challenges organizations face when trying to create photorealistic 3D content - traditionally a time-consuming, expensive, and technically complex process requiring specialized skills and equipment.
25
+
By deploying this guidance, users gain access to a flexible infrastructure that handles the entire 3D reconstruction process programmatically, from media upload to final 3D model delivery, while being highly modular through its componentized pipeline-based approach. This guidance addresses the significant challenges organizations face when trying to create photorealistic 3D content - traditionally a time-consuming, expensive, and technically complex process requiring specialized skills and equipment.
26
26
27
27
## Architecture
28
28
29
29
This guidance will:
30
30
31
31
- create the infrastructure required to create a gaussian splat from a video or set of images
32
32
- create the mechanism to run the code and perform 3D reconstruction
33
-
- enable a user to create a 3D gaussian splat from the backend (no UI) using open source tools and AWS by uploading a video (.mp4 or .mov) or images (.png or .jpg) and metadata (.json) into S3
33
+
- enable a user to create a 3D gaussian splat using open source tools and AWS by uploading a video (.mp4 or .mov) or images (.png or .jpg) and metadata (.json) into S3
34
+
- provide a 3D viewer for viewing the photo-realistic effects and performant nature of gaussian splats
<i>Figure 1: 3D Reconstruction Toolbox for Gaussian Splats on AWS Reference Architecture</i>
41
+
<i>Figure 1: 3D Reconstruction Toolbox for Gaussian Splats on AWS Reference Architecture</i>
41
42
</div>
42
43
43
44
### Architecture Steps
@@ -73,7 +74,7 @@ Amazon CloudWatch is used to monitor the training logs, surfacing errors to the
73
74
74
75
### Custom GS Pipeline Container
75
76
76
-
In this project, there is only one Docker container that contains all of the 3D reconstruction tools for Gaussian Splatting. This container has a `Dockerfile`, `main.py`, and helper script files and open source libraries under the `backend/container` directory. The main script processes each request from the SageMaker Training Job invoke message and saves the result to S3 upon successful completion. The list of open source libraries that make this project possible include:
77
+
In this project, there is only one Docker container that contains all of the 3D reconstruction tools for Gaussian Splatting. This container has a `Dockerfile`, `main.py`, and helper script files and open source libraries under the `source/container` directory. The main script processes each request from the SageMaker Training Job invoke message and saves the result to S3 upon successful completion. The list of open source libraries that make this project possible include:
@@ -91,30 +92,30 @@ In this project, there is only one Docker container that contains all of the 3D
91
92
92
93
- Git
93
94
- Docker
94
-
- Terraform (if chosing not to deploy infrastructure using CDK)
95
+
- Terraform (if choosing not to deploy infrastructure using CDK)
95
96
96
97
### AWS account requirements
97
98
98
99
An active AWS Account with IAM user or role with elevated permissions to deploy resources is required to deploy this guidance, along with either a local computer with appropriate AWS credentials to deploy the CDK or Terraform solution, or utilize an AWS EC2 workstation to build and deploy the CDK or Terraform solution. Instructions for doing this will be in the [Implementation Guide](https://implementationguides.kits.eventoutfitters.aws.dev/open-3drt-0403/compute/open-source-3d-reconstruction-toolbox-for-gaussian-splats-on-aws.html)
99
100
100
101
Resources included in this deployment:
101
102
102
-
- EC2 (if chosing not to deploy infrastructure from your local computer)
103
+
- EC2 (if choosing not to deploy infrastructure from your local computer)
103
104
- IAM roles with permissions
104
105
- CloudFormation
105
106
- ECR Image
106
107
- S3 Buckets
107
108
- DynamoDB Table
108
109
- Lambda Functions
109
110
- SageMaker Training Jobs
110
-
-Stepfunctions State Machine
111
+
-Step Functions State Machine
111
112
- CDK (bootstrap instructions will be included in the [Implementation Guide](https://implementationguides.kits.eventoutfitters.aws.dev/open-3drt-0403/compute/open-source-3d-reconstruction-toolbox-for-gaussian-splats-on-aws.html))
112
113
113
114
### Service limits
114
115
115
116
-[Service quotas](https://docs.aws.amazon.com/servicequotas/latest/userguide/intro.html) - increases can be requested via the AWS Management Console, AWS CLI, or AWS SDKs (see [Accessing Service Quotas](https://docs.aws.amazon.com/servicequotas/latest/userguide/intro.html#access))
116
117
117
-
- (Optional) SageMaker Training Jobs uses a Docker container to run the training. This deployment guide has a `Deploy backend container` section that walks through building a custom container image for SageMaker. You can optionally build and test this container locally (not running on SageMaker) on a GPU-enabled EC2 instance. If you plan to do this, increase the EC2 quota named "Running On-Demand G and VT instances" and/or "Running On-Demand P instances", depending on the instance family you plan to use, to a desired maximum number of vCPUs for running instances of the target family. Note, this is vCPUs NOT number of instances like the SageMaker Batch Transform quota.
118
+
- (Optional) SageMaker Training Jobs uses a Docker container to run the training. This deployment guide walks through building a custom container image for SageMaker. You can optionally build and test this container locally (not running on SageMaker) on a GPU-enabled EC2 instance. If you plan to do this, increase the EC2 quota named "Running On-Demand G and VT instances" and/or "Running On-Demand P instances", depending on the instance family you plan to use, to a desired maximum number of vCPUs for running instances of the target family. Note, this is vCPUs NOT number of instances like the SageMaker Training Jobs quota.
118
119
119
120
- Install and configure the AWS CLI (if not using the recommended EC2 deployment below)
0 commit comments