Description #

Qwen2.5-Omni multimodal audio understanding model – key features of audio content parsing and natural language description generation

Key features #

• Intelligent content understanding and description generation
• Automatically generates detailed and accurate natural language descriptions based on audio signals, and supports custom prompts via the prompt parameter
• Supports audio formats such as mp3, aac, and m4a
• Efficient model loading and inference optimization
• Supports multiple computation precision options (bfloat16, float16, float32) to meet different performance requirements
• Integrated FlashAttention2 acceleration technology, significantly improves inference efficiency
• Supports automatic or manual device allocation, and is fully compatible with single-card and multi-card environments
• Resource usage
• It is recommended to use a GPU with 48GB or more VRAM

Scenario optimization #

• Widely applicable to audio content analysis, intelligent audio description, multimedia content understanding, and many other application scenarios
• Built-in batch processing mechanism efficiently supports large-scale parallel inference of audio data
• Flexible configuration of audio data return options to meet diverse business requirements

Operator parameters #

Input #

Output #

The processed array, where each element is the text result of the audio understanding. For audio files that fail to process, an empty string is returned.

Parameters #

If a parameter does not have a default value, it is required

Examples #

The following code demonstrates how to use daft to run the operator for audio content understanding and generate descriptions as specified.

from __future__ import annotations

import os

import daft
from daft import col
from daft.las.functions.multimodal.qwen_omni_audio_understanding import QwenOmniAudioUnderstanding
from daft.las.functions.udf import las_udf

if __name__ == "__main__":

    if os.getenv("DAFT_RUNNER", "native") == "ray":
        import logging

        import ray

        def configure_logging():
            logging.basicConfig(
                level=logging.INFO,
                format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
                datefmt="%Y-%m-%d %H:%M:%S.%s".format(),
            )
            logging.getLogger("tracing.span").setLevel(logging.WARNING)
            logging.getLogger("daft_io.stats").setLevel(logging.WARNING)
            logging.getLogger("DaftStatisticsManager").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaScheduler").setLevel(logging.WARNING)
            logging.getLogger("DaftFlotillaDispatcher").setLevel(logging.WARNING)

        ray.init(dashboard_host="0.0.0.0", runtime_env={"worker_process_setup_hook": configure_logging})
        daft.set_runner_ray()
    daft.set_execution_config(actor_udf_ready_timeout=600)
    daft.set_execution_config(min_cpu_per_task=0)

    tos_dir_url = os.getenv("TOS_DIR_URL", "las-cn-beijing-public-online.tos-cn-beijing.volces.com")
    samples = {
        "audio_path": [
            f"https://{tos_dir_url}/public/shared_audio_dataset/.mp3"
        ],
        "prompt": [""],

    }

    model_path = os.getenv("MODEL_PATH", "/opt/las/models")
    model_name = "Qwen/Qwen2.5-Omni-7B"
    dtype = "bfloat16"
    use_flash_attention_2 = True
    default_prompt = None
    max_caption_length = 256
    batch_size = 1
    rank = 0
    num_gpus = int(os.getenv("NUM_GPUS", 1))
    num_gpus = 1

    ds = daft.from_pydict(samples)
    ds = ds.with_column(
        "caption",
        las_udf(
            QwenOmniAudioUnderstanding,
            construct_args={
                "model_path": model_path,
                "model_name": model_name,
                "dtype": dtype,
                "use_flash_attention_2": use_flash_attention_2,
                "prompt": default_prompt,
                "max_caption_length": max_caption_length,
                "batch_size": batch_size,
                "rank": rank,
            },
            num_gpus=num_gpus,
            batch_size=1,
            concurrency=1,
        )(col("audio_path"), col("prompt")),
    )
    print(ds.to_pandas()["caption"][0])

    ds.show()

    # ╭────────────────────────────────┬───────────────────────────────────────────────┬─────────────────────────────────────────────────────────────╮
    # │ audio_path                     ┆ prompt                                        ┆ caption                                                     │
    # │ ---                            ┆ ---                                           ┆ ---                                                         │
    # │ Utf8                           ┆ Utf8                                          ┆ Utf8                                                        │
    # ╞════════════════════════════════╪═══════════════════════════════════════════════╪═════════════════════════════════════════════════════════════╡
    # │ tos://las-cn-beijing-public-o… ┆ Please convert this audio directly into text without any explanation. …        ┆ I have saved the person and retrieved what I needed. I, Old Sun, do not seek any fame or fortune, only wish to return to the Mountain of Flowers and Fruit…           │
    # ╰────────────────────────────────┴───────────────────────────────────────────────┴─────────────────────────────────────────────────────────────╯