Vayuvahana Technologies Private Limited VajraV1 is a state-of-the-art (SOTA) real time object detection model inspired by the YOLO model architectures. VajraV1 is a family of fast, lightweight models that can be used for a variety of tasks like object detection and tracking, instance segmentation, oriented object detection, pose detection, and image classification.

Enterprise License

To request for an Enterprise License please get in touch via Email

Performance on VisDrone Dataset

Model	size (pixels)	mAPtest-dev 50-95	mAPval 50-95	Speed RTX 4090 TensorRT10 Latency (ms)	params (M)	FLOPs (B)
VajraV1-nano-det	640	25.5	20.8	1.4	3.32	8.0
VajraV1-small-det	640	27.3	24.3	1.4	12.36	27.7
VajraV1-medium-det	640	29.7	27.7	1.8	21.09	74.8
VajraV1-large-det	640	30.0	28.0	2.4	25.70	92.8
VajraV1-xlarge-det	640	30.4	29.7	2.9	57.75	207.8

Results on COCO dataset to be published soon!

Documentation

Install

Git clone the VayuAI SDK including all requirements in a Python>=3.8 environment.

git clone https://github.com/NamanMakkar/VayuAI.git
cd VayuAI
pip install .

Usage

CLI

Vajra can be used in the Command Line Interface with a vajra or vayuvahan or vayuai command:

vajra predict model=vajra-v1-nano-det img_size=640 source="path/to/source.jpg"

Python

Vajra can also be used directly in a Python environment, and accepts the same arguments as in the CLI example above:

from vajra import Vajra, VajraDEYO
model = Vajra("vajra-v1-nano-det")
model_vajra_deyo = VajraDEYO("vajra-deyo-v1-nano-det")

train_results = model.train(
    data="coco8.yaml",
    epochs=100,
    img_size=640,
    device="cpu"
)

metrics = model.val()
results = model("path/to/img.jpg")
results[0].show()

path = model.export(format="onnx")

Model Architectures

• VajraV1-det
• VajraV1-cls
• VajraV1-pose
• VajraV1-seg
• VajraV1-obb
• VajraV1-world
• VajraV1-DEYO-det
• VajraV1-DEYO-seg (Coming Soon!)
• VajraV1-DEYO-pose (Coming Soon!)
• SAM
• EfficientNetV1
• EfficientNetV2
• VajraEffNetV1
• VajraEffNetV2
• ConvNeXtV1
• ConvNeXtV2
• ResNet
• ResNeSt
• ResNeXt (Coming Soon!)
• ResNetV2 (Coming Soon!)
• EdgeNeXt
• ME-NeSt
• VajraME-NeSt
• MixConvNeXt
• ViT (Coming Soon!)
• Swin (Coming Soon!)
• SwinV2 (Coming Soon!)

Tasks Supported

• detect
• small_obj_detect
• classify
• multilabel_classify
• pose
• obb
• segment
• world
• panoptic (Coming Soon!)

Model Architecture Details

To be published

Acknowledgements

• https://github.com/ultralytics/ultralytics
• https://github.com/ultralytics/yolov5
• https://github.com/ouyanghaodong/DEYOv1.5
• https://github.com/WongKinYiu/yolov9
• https://github.com/meituan/YOLOv6
• https://github.com/huggingface/pytorch-image-models
• https://github.com/pytorch/vision

License

Vayuvahana Technologies Private Limited offers two licensing options:

• AGPL-3.0 License: This is an OSI-approved open-source license for researchers for the purpose of promoting collaboration. See the LICENSE file for details.

• Enterprise License: This license is designed for commercial use and enables integration of VayuAI software and AI models into commercial goods and services, bypassing the open-source requirements of AGPL-3.0. If your product requires embedding the software for commercial purposes or require access to more capable enterprise AI models in the future, reach out via Email.