Fix for llava models not generating text with test failures in 1.19

[tthakkal]

What does this PR do?

Seeing regression on 1.19 for llava model accuracy with fp8.
Using scale mode to fix the issue.
Updated readme and tests to use maxabs_quant_scale_format_const.json for llava models/

Fixes # (issue)

Current output:
11/26/2024 00:56:50 - INFO - main - result = [[{'generated_text': 'USER: \nWhat is shown in this image? ASSISTANT:'}]]
11/26/2024 00:56:50 - INFO - main - time = 160.23870641365647ms, Throughput (including tokenization) = -12.481378842618692 tokens/second

output after fix:
11/26/2024 00:52:23 - INFO - {}main{} - result = [[
{'generated_text': 'USER: \nWhat is shown in this image? ASSISTANT: The image shows a pier extending out over a body of water, with a mountain in the background.'}
]]
11/26/2024 00:52:23 - INFO - {}main{} - time = 160.74028802104294ms, Throughput (including tokenization) = 111.98188221265083 tokens/second

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.