Extend inference SDK with client for (almost all) core models

[PawelPeczek-Roboflow]

Description

New client methods making it possible to call core models (apart from SAM)

New API:

from inference_sdk import InferenceHTTPClient

LOCALHOST_CLIENT = InferenceHTTPClient(
    api_url="http://127.0.0.1:9001",
    api_key=API_KEYS["MY_WORKSPACE_API_KEY"],
)

LOCALHOST_CLIENT.ocr_image("https://jeroen.github.io/images/testocr.png")
LOCALHOST_CLIENT.ocr_image(["https://jeroen.github.io/images/testocr.png"]*2)

LOCALHOST_CLIENT.detect_gazes(
    "https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg"
)
LOCALHOST_CLIENT.detect_gazes(
    ["https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg"] * 3
)

LOCALHOST_CLIENT.get_clip_image_embeddings(
    inference_input="https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg",
)
LOCALHOST_CLIENT.get_clip_image_embeddings(
    inference_input=["https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg"]*3,
)

LOCALHOST_CLIENT.get_clip_text_embeddings(
    text="Quick brawn fox",
)
LOCALHOST_CLIENT.get_clip_text_embeddings(
    text=["Quick brawn fox"]*5,
)

LOCALHOST_CLIENT.clip_compare(
    subject="Quick brawn fox",
    prompt=["Quick brown fox", "Quick brawn dog"],
    subject_type="text",
)

LOCALHOST_CLIENT.clip_compare(
    subject="https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg",
    prompt=["https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg"]*2,
    prompt_type="image",
)


LOCALHOST_CLIENT.prompt_cogvlm(
    visual_prompt="https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg",
    text_prompt="What is the topic of that picture",
)
LOCALHOST_CLIENT.prompt_cogvlm(
    visual_prompt="https://t3.ftcdn.net/jpg/02/43/12/34/360_F_243123463_zTooub557xEWABDLk0jJklDyLSGl2jrr.jpg",
    text_prompt="So - what is your final judgement about the content of the picture?",
    chat_history=[("I think the image shows a XXX", "You are wrong - the image shows YYY")]
)

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

How has this change been tested, please provide a testcase or example of how you tested the change?

• tested locally
• new tests in CI
• CI still green

Any specific deployment considerations

For example, documentation changes, usability, usage/costs, secrets, etc.

Docs

• Docs updated? What were the changes: