[YOLOv8-Seg] Validation

[dbogunowicz]

The validators obey the following hierarchy of inheritance (to reduce the bloating of the codebase):

DeepSparseSegmentationValidator has parents:
- SegmentationValidator (Ultralytics)
- DeepsparseValidator (that handles the main inference method `__call__`)

DeepSparseDetectionValidator has parents:
- DetectionValidator (Ultralytics)
- DeepsparseValidator (that handles the main inference method `__call__`)

adding classification task potentially in the future will be trivial.

Testing plan:

• Test annotation for YOLO (sanity check)
• Test annotation for YOLOv8 Det (sanity check)
• Test annotation for YOLOv8 Seg (sanity check)
• Test validation for YOLOv8 Det
• Test validation for YOLOv8 Seg

Validation on yolov8n.onnx reproduces the results reported by the authors:

vs

/home/ubuntu/damian/deepsparse_venv/bin/python /home/ubuntu/damian/deepsparse/src/deepsparse/yolov8/validation.py --dataset_yaml coco.yaml --model-path yolov8n-seg.onnx --subtask segmentation 
DeepSparse, Copyright 2021-present / Neuralmagic, Inc. version: 1.4.0.20230211 COMMUNITY | (bacdaef4) (release) (optimized) (system=avx2, binary=avx2)
Ultralytics YOLOv8.0.30 🚀 Python-3.8.10 torch-1.13.1+cu117 CUDA:0 (NVIDIA RTX A4000, 16117MiB)
val: Scanning /home/ubuntu/damian/deepsparse/datasets/coco/labels/val2017.cache... 4952 images, 48 backgrounds, 0 corrupt: 100%|██████████| 5000/5000 [00:00<?, ?it/s]
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100%|██████████| 313/313 [13:58<00:00,  2.68s/it]
                   all       5000      36335      0.621      0.479      0.516      **0.363**      0.617      0.456      0.489      **0.305**

...

[bfineran]

LGTM - sanity check looks good

[KSGulin]

LGTM!

Wanted to note for the casual observer that all commits in this PR were authored by @dbogunowicz and any attributions to me are due to an internal server environment hiccup