feat: add sharpness metric

[begumcig]

Description

• Adds SharpnessMetric, which computes focus via the variance of the Laplacian (higher = sharper).

• CPU‑side numpy/OpenCV implementation.

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

Additional Notes

You can use the SharpnessMetric directly or via the EvaluationAgent:

from pruna.evaluation.metrics.metric_sharpness import SharpnessMetric
from pruna.evaluation.task import Task
from pruna.evaluation.evaluation_agent import EvaluationAgent
from pruna.data.pruna_datamodule import PrunaDataModule

# 1) Direct use 
metric = SharpnessMetric(kernel_size=5)

inputs, gts = next(iter(my_dataloader))          # your B×C×H×W tensors
outputs = my_model(inputs)

metric.update(inputs, gts, outputs)
print("Sharpness:", metric.compute().result)     # higher = sharper

# 2) Via Task / EvaluationAgent
dm = PrunaDataModule.from_string("LAION256")
dm.limit_datasets(10)

task = Task(
    request=["sharpness"],        # or [SharpnessMetric(kernel_size=5)]
    datamodule=dm,
)
agent = EvaluationAgent(task=task)

results = agent.evaluate(my_model)
print(results[0])                 # MetricResult(name='sharpness', value=…)

[sharpenb]

Very cool new metric! This is close to merge. Beyond my comments, a last thing would be to have/extend a tutorial with these introduced metric (it can be in another PR). This tutorial could take an image compute the metrics on it, and then perturb it it (e.g. by adding blur or other things mentioned in this paper). The goal would be to justify the relevance of these metrics :)