Technical Registry

Index of open-source models, datasets, and specialized architectures.

Computer Vision

GolfCart YOLO11s

A lightweight object detection model optimized for identifying golf carts in aerial drone imagery. Fine-tuned on a dataset of 624 images to balance edge-device inference speed with high precision.

mAP@50 0.93

Precision 0.978

Model Size 22 MB

Latency 14ms

from ultralytics import YOLO

# Load fine-tuned model
model = YOLO('best-3.pt')

# Run inference
results = model('test_images/golfcart_drone_01.jpg')
results[0].show() 
                    

🤗 View on HuggingFace

cart 0.94

NLP / Generation

HaikuGPT-17M

A decoder-only transformer trained from scratch on a curated dataset of haikus. Despite having only 17M parameters, it learns to adhere to strict poetic structure (5-7-5) and semantic coherence without prompting.

Parameters 17M

Ctx Window 64 Tkn

Arch GPT-2

Loss 1.84

from transformers import pipeline

pipe = pipeline("text-generation", model="rwitz/haiku-17m")
print(pipe("The summer sun", max_length=20))
                    

🤗 View on HuggingFace

user@dev:~$ ./generate_haiku.py

"Cherry blossoms fall,

Petals dance on spring's soft breeze,

Nature's gentle waltz."

user@dev:~$

OCR / Segmentation

Pinball Score YOLOv11s

A lightweight detection model trained on A100 GPUs to identify and segment digital scores on vintage pinball machines. Specifically engineered to handle extreme glare, reflections, and low-light arcade conditions.

mAP@50 0.99

Recall 0.989

Data 1.1k Img

Epochs 100

from ultralytics import YOLO

model = YOLO('pinball-best.pt')
results = model('arcade_clip_04.mp4', stream=True)

for r in results:
    r.save_crop('output/')
                    

🤗 View on HuggingFace

SCORE_REGION 0.99

42,050,100