attach_speaker_embedding_s2s.py

import os
from os.path import expanduser

import shutil
from soundfile import LibsndfileError
from datasets import load_dataset, DatasetDict, Audio

direction = os.getenv("DIRECTION", "enA-jaA")
sides = set(direction.split("-"))
dataset_id = os.getenv("DATASET_ID", 0)
num_proc = int(os.getenv("NUM_PROC", 1))
hf_org = os.getenv("HF_ORG", "asahi417")
hf_dataset = f"seamless-align-{direction}"
dataset = load_dataset(f"{hf_org}/{hf_dataset}", f"subset_{dataset_id}", split="train")
audio_loader = Audio()
se_model = os.getenv("SE_MODEL", "metavoice")
max_seq_length = 1000000
min_seq_length = 50000

if se_model == "metavoice":
    from speaker_embedding_metavoice import MetaVoiceSE
    speaker_embedder = MetaVoiceSE()
elif se_model == "pyannote":
    from speaker_embedding_pyannote import PyannoteSE
    speaker_embedder = PyannoteSE()
elif se_model == "w2vbert-600m":
    from speaker_embedding_hf import W2VBERTEmbedding
    speaker_embedder = W2VBERTEmbedding()
elif se_model == "xlsr-2b":
    from speaker_embedding_hf import XLSR2BEmbedding
    speaker_embedder = XLSR2BEmbedding()
elif se_model == "hubert-xl":
    from speaker_embedding_hf import HuBERTXLEmbedding
    speaker_embedder = HuBERTXLEmbedding()
else:
    raise ValueError(f"unknown speaker embedding: {se_model}")


def error_file(example):
    for side in sides:
        try:
            wav = audio_loader.decode_example(example[f"{side}.audio"])
            if len(wav["array"]) < min_seq_length or len(wav["array"]) > max_seq_length:
                return False
        except ValueError:
            return False
        except LibsndfileError:
            return False
    return True


print(f"Num examples: {len(dataset)}")
for s in sides:
    dataset = dataset.cast_column(f"{s}.audio", Audio(decode=False))
dataset = dataset.filter(error_file, num_proc=num_proc, desc="drop broken audio")
for s in sides:
    dataset = dataset.cast_column(f"{s}.audio", Audio())
print(f"Num examples (after filtering): {len(dataset)}")


def speaker_embedding(example):
    for side in sides:
        # print(len(example[f"{side}.audio"]["array"]))
        embedding = speaker_embedder.get_speaker_embedding(
            example[f"{side}.audio"]["array"], example[f"{side}.audio"]["sampling_rate"]
        )
        if embedding.ndim == 1:
            example[f"{side}.audio.speaker_embedding"] = embedding
        else:
            example[f"{side}.audio.speaker_embedding"] = embedding.mean(0)
            example[f"{side}.audio.speaker_embedding.full"] = embedding
    return example


dataset = dataset.map(
    function=speaker_embedding,
    remove_columns=[f"{s}.audio" for s in sides] + [f"{s}.url" for s in sides] + [f"{s}.duration_start" for s in sides] + [f"{s}.duration_end" for s in sides],
    num_proc=num_proc,
    desc="attach speaker embedding dataset"
)
DatasetDict({"train": dataset}).push_to_hub(f"{hf_org}/{hf_dataset}.speaker-embedding.{se_model}", config_name=f"subset_{dataset_id}")
cache_dir = f"{expanduser('~')}/.cache/huggingface/datasets/{hf_org}___{hf_dataset.replace('A', '_a')}/subset_{dataset_id}"
if os.path.exists(cache_dir):
    shutil.rmtree(cache_dir)