Description #

AudioSpeakerVerificationEres2net audio speaker verification processor

Key features #

• Speaker verification: Determines whether two audio segments are from the same speaker and outputs a similarity score.
• Multi-source audio support: Supports various audio input types including local files, URLs, object storage, base64, and binary.
• Batch processing: Efficiently verifies and handles exceptions for batches of audio pairs.

Technical features #

• Loads the ERes2Net speaker verification model based on the ModelScope framework.
• Supports multi-GPU inference, with automatic device selection or manual specification.
• 16kHz sample rate audio preprocessing, compatible with multiple audio formats.

Typical application scenarios #

• ✅ Voice identity authentication – Speaker consistency verification
• ✅ Intelligent customer service – Multi-channel voice normalization and verification
• ✅ Voice data cleansing – Automated speaker deduplication

Recommendations #

• It is recommended to standardize the input audio sample rate to 16kHz to improve verification accuracy.
• In multi-GPU environments, the GPU device can be flexibly specified using the rank parameter.

Operator parameters #

Input #

Output #

The speaker similarity score (float) for each audio pair; abnormal samples are returned as None.

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 to verify whether the speakers in two audio files are the same person.

from __future__ import annotations

import logging
import os

import ray

import daft
from daft import col
from daft.las.functions.audio.audio_speaker_verification_eres2net import AudioSpeakerVerificationEres2net
from daft.las.functions.udf import las_udf

if __name__ == "__main__":
    if os.getenv("DAFT_RUNNER", "ray") == "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)

        import ray

        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 = {
        "speaker_a": [f"https://{tos_dir_url}/public/shared_audio_dataset/.wav"],
        "speaker_b": [f"https://{tos_dir_url}/public/shared_audio_dataset/.wav"],
    }

    model_path = os.getenv("MODEL_PATH", "/opt/las/models")
    model_name = "iic/speech_eres2net_sv_zh-cn_16k-common"
    audio_src_type = "audio_url"
    rank = 0

    df = daft.from_pydict(samples)
    df = df.with_column(
        "speaker_verification_result",
        las_udf(
            AudioSpeakerVerificationEres2net,
            construct_args={
                "audio_src_type": audio_src_type,
                "model_path": model_path,
                "model_name": model_name,
                "rank": rank,
            },
            num_gpus=1,
            batch_size=1,
            concurrency=1,
        )(col("speaker_a"), col("speaker_b")),
    )

    df.show()

    # ╭────────────────────────────────┬────────────────────────────────┬─────────────────────────────╮
    # │ speaker_a                      ┆ speaker_b                      ┆ speaker_verification_result │
    # │ ---                            ┆ ---                            ┆ ---                         │
    # │ Utf8                           ┆ Utf8                           ┆ Float32                     │
    # ╞════════════════════════════════╪════════════════════════════════╪═════════════════════════════╡
    # │ tos://las-cn-beijing-public-o… ┆ tos://las-cn-beijing-public-o… ┆ 1                           │
    # ╰────────────────────────────────┴────────────────────────────────┴─────────────────────────────╯