Docs updates for HUB, YOLOv4, YOLOv7, NAS (#3174)

Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

docs/datasets/segment/coco.md

@@ -1,6 +1,7 @@
 ---
 comments: true
 description: Learn about the COCO-Seg dataset, designed for simple training of YOLO models on instance segmentation tasks.
+keywords: COCO-Seg, COCO, instance segmentation, segmentation annotations, computer vision, deep learning, data science, YOLO models, image size, open-source datasets
 ---
 
 # COCO-Seg Dataset

docs/datasets/segment/coco8-seg.md

@@ -1,6 +1,7 @@
 ---
 comments: true
 description: Test and debug segmentation models on small, versatile COCO8-Seg instance segmentation dataset, now available for use with YOLOv8 and Ultralytics HUB.
+keywords: Ultralytics, COCO8-Seg, instance segmentation dataset, segmentation models, new detection approaches, COCO train 2017 set
 ---
 
 # COCO8-Seg Dataset

docs/datasets/segment/index.md

@@ -1,6 +1,7 @@
 ---
 comments: true
 description: Learn about the Ultralytics YOLO dataset format for segmentation models. Use YAML to train Detection Models. Convert COCO to YOLO format using Python.
+keywords: instance segmentation datasets, yolov8 segmentations, yaml dataset format, auto annotation, convert label formats
 ---
 
 # Instance Segmentation Datasets Overview