https://github.com/interactivetech/yolov8-determinedai
Proof of Concept on how to integrate YoloV8 with DeterminedAI
https://github.com/interactivetech/yolov8-determinedai
Last synced: 9 months ago
JSON representation
Proof of Concept on how to integrate YoloV8 with DeterminedAI
- Host: GitHub
- URL: https://github.com/interactivetech/yolov8-determinedai
- Owner: interactivetech
- License: mit
- Created: 2023-01-13T19:04:09.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2024-04-12T21:06:25.000Z (about 2 years ago)
- Last Synced: 2025-09-24T01:56:04.057Z (9 months ago)
- Language: Jupyter Notebook
- Size: 6.33 MB
- Stars: 2
- Watchers: 1
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# YoloV8 with DeterminedAI (WIP, Experimental)
Proof of Concept on how to integrate YoloV8(https://github.com/ultralytics/ultralytics) with DeterminedAI
Author: Andrew Mendez
# Install Dependencies
`bash install-dep.sh`
# Current Limitations
Currently the code only supports single node multi-gpu training. More work is needed to support multiple node.
If you want to train with multiple gpus, only edit the: `device` config setting. You should edit where you specify all the cuda devices you want to use. Examples are follows:
* Train on single GPU: `device: 0`
* Train on 4 GPUs: `device: 0,1,2,3`
* Train on 8 GPUs: `device: 0,1,2,3,4,5,6,7`
# Spin up Determined Cluster (No GPU cluster)
`det deploy local cluster-up --no-gpu --master-port 8081`
# Spin up Determined Cluster (GPU cluster)
`det deploy local cluster-up --master-port 8081`
# Run Training if using M1 Mac (Arm64)
`bash run_exp_arm64.sh`
# Run Training if using Linux/Mac (Intel-based, Amd64)
`bash run_exp.sh`
# Steps to run Vanilla Training without Determined
`python train.py`
# Questions
Any Questions, please reach out to me at andrew.mendez@hpe.com