fix(ia): Update T2I latency score to include num_inference_steps

[eliteprox]

What does this pull request do? Explain your changes. (required)

Updates the formula used to calculate latency score for T2I request so that num_inference_steps is included in the measurement.

Specific updates (required)

• Update ai_process.go to use the formula: took.Seconds() / float64(outPixels) / (float64(numImages) * float64(numInferenceSteps))

How did you test each of these updates (required)

Tested with high-end range values for num_inference_steps and num_images_per_prompt

Results:

25 inference
0.00000006567261756897  - 1 image
0.00000006524371048584 - 25 images

100 inference
0.00000004808495975494 - 1 image
0.00000004837596923332 - 25 images

Does this pull request close any open issues?

LIV-362

Checklist:

• [ ] Read the contribution guide
• [ ] make runs successfully
• [ ] All tests in ./test.sh pass
• [ ] README and other documentation updated
• [ ] Pending changelog updated

[rickstaa]

Replaced by #3074 because for some reason I don't have enough permissions to push changes to this pull request.