CtrlK
CommunityDocumentationLog inGet started
Tessl Logo

tessl/pypi-vllm

tessl install tessl/pypi-vllm@0.10.0

A high-throughput and memory-efficient inference and serving engine for LLMs

Agent Success

Agent success rate when using this tile

69%

Improvement

Agent success rate improvement when using this tile compared to baseline

1.33x

Baseline

Agent success rate without this tile

52%

task.mdevals/scenario-9/

Image Description Service

A Python service that uses a vision-language model to generate descriptions of images.

Capabilities

Image Description Generation

  • Given an image file path and a text prompt "Describe this image in detail", the service returns a text description of the image content @test
  • Given an image file path and no prompt (using a default prompt), the service returns a basic description of the image @test

Multi-Image Analysis

  • Given a list of two image file paths and a prompt asking to compare them, the service returns a comparison of the images @test

Error Handling

  • Given an invalid image file path, the service raises an appropriate error @test

Implementation

@generates

API

class ImageDescriber:
    """A service that generates descriptions of images using a vision-language model."""

    def __init__(self, model_name: str = "llava-hf/llava-1.5-7b-hf"):
        """
        Initialize the image describer with a vision-language model.

        Args:
            model_name: The name of the vision-language model to use
        """
        pass

    def describe_image(self, image_path: str, prompt: str = "Describe this image.") -> str:
        """
        Generate a text description of a single image.

        Args:
            image_path: Path to the image file
            prompt: The text prompt to accompany the image (default: "Describe this image.")

        Returns:
            A text description of the image

        Raises:
            FileNotFoundError: If the image file does not exist
        """
        pass

    def describe_images(self, image_paths: list[str], prompt: str) -> str:
        """
        Generate a text response analyzing multiple images.

        Args:
            image_paths: List of paths to image files
            prompt: The text prompt to accompany the images

        Returns:
            A text response analyzing the images

        Raises:
            FileNotFoundError: If any image file does not exist
        """
        pass

Dependencies { .dependencies }

vLLM { .dependency }

High-throughput inference engine with vision-language model support

@satisfied-by

Version

Workspace
tessl
Visibility
Public
Created
Last updated
Describes
pypipkg:pypi/vllm@0.10.x
tile.json