Fix FlorenceTemplate for florence2

[yaqiangsun]

Remove description when using task_prompts_with_input mode to get correct task.

self.task_prompts_with_input = {
            '<CAPTION_TO_PHRASE_GROUNDING>': "Locate the phrases in the caption: {input}",
            '<REFERRING_EXPRESSION_SEGMENTATION>': 'Locate {input} in the image with mask',
            '<REGION_TO_SEGMENTATION>': 'What is the polygon mask of region {input}',
            '<OPEN_VOCABULARY_DETECTION>': 'Locate {input} in the image.',
            '<REGION_TO_CATEGORY>': 'What is the region {input}?',
            '<REGION_TO_DESCRIPTION>': 'What does the region {input} describe?',
            '<REGION_TO_OCR>': 'What text is in the region {input}?',
        }

When using the mode in task_prompts_with_input, such as <OPEN_VOCABULARY_DETECTION>, the input should be like '<OPEN_VOCABULARY_DETECTION>animals'. And in post_process_generation, it should be <OPEN_VOCABULARY_DETECTION>.

PR type

• Bug Fix
• New Feature
• Document Updates
• More Models or Datasets Support

PR information

Write the detail information belongs to this PR.

Experiment results

Paste your experiment result here(if needed).

[Jintao-Huang]

thanks!

[Jintao-Huang]

please run:

pip install pre-commit
pre-commit run --all-files

[yaqiangsun]

@Jintao-Huang I have run the it, and update the PR.

please run:

pip install pre-commit
pre-commit run --all-files