Add a trackers tune CLI subcommand that runs hyperparameter optimization for any registered tracker using Optuna.

[SkalskiP]

Stage 1: Search Space Definitions

• Added search_space: ClassVar[dict[str, dict]] to SORTTracker and
  ByteTrackTracker
• Added validation in BaseTracker.__init_subclass__ ensuring every search_space key
  matches an __init__ parameter
• search_space is optional (trackers without it still work)

Stage 2: Tuner Class + Objective Function

• Tuner class that wraps Optuna and an internal objective function
• Update pyproject.toml with optuna as optional dependency under trackers[tune]
• Run evaluation with evaluate_mot_sequences from trackers/eval/evaluate.py
• Each trial calls tracker reset() method
• Tuner.run() creates Optuna study, runs trials, returns best params dict

Stage 3: CLI Subcommand

• Wire Tuner into trackers tune subcommand following trackers eval pattern
• CLI arguments: --tracker, --gt-dir, --detections-dir, --objective, --n-trials, --output

Target Usage

# Basic usage
trackers tune \
  --tracker bytetrack \
  --gt-dir path/to/gt/ \
  --detections-dir path/to/detections/ \
  --n-trials 100

# With custom objective
trackers tune \
  --tracker sort \
  --gt-dir path/to/gt/ \
  --detections-dir path/to/detections/ \
  --objective HOTA \
  --n-trials 200

[SkalskiP]

@omkar-334 I updated this branch to make it up to date before #301 merging.

[omkar-334]

hey @SkalskiP can you update this branch? I think my tuning experiment will get similar numbers because of #323 and #323 being merged. Did you get a chance to review my PR? #301