update

嘉渊

2023-04-24 6427c834dfd97b1f05c6659cdc7ccf010bf82fe1

 funasr/bin/sv_inference.py

@@ -1,4 +1,7 @@
#!/usr/bin/env python3
# Copyright FunASR (https://github.com/alibaba-damo-academy/FunASR). All Rights Reserved.
#  MIT License  (https://opensource.org/licenses/MIT)

import argparse
import logging
import os
@@ -26,14 +29,14 @@
from funasr.utils.types import str2bool
from funasr.utils.types import str2triple_str
from funasr.utils.types import str_or_none

from funasr.utils.misc import statistic_model_parameters

class Speech2Xvector:
    """Speech2Xvector class

    Examples:
        >>> import soundfile
        >>> speech2xvector = Speech2Xvector("sv_config.yml", "sv.pth")
        >>> speech2xvector = Speech2Xvector("sv_config.yml", "sv.pb")
        >>> audio, rate = soundfile.read("speech.wav")
        >>> speech2xvector(audio)
        [(text, token, token_int, hypothesis object), ...]
@@ -59,6 +62,7 @@
            device=device
        )
        logging.info("sv_model: {}".format(sv_model))
        logging.info("model parameter number: {}".format(statistic_model_parameters(sv_model)))
        logging.info("sv_train_args: {}".format(sv_train_args))
        sv_model.to(dtype=getattr(torch, dtype)).eval()

@@ -156,21 +160,22 @@


def inference_modelscope(
        output_dir: Optional[str],
        batch_size: int,
        dtype: str,
        ngpu: int,
        seed: int,
        num_workers: int,
        log_level: Union[int, str],
        key_file: Optional[str],
        sv_train_config: Optional[str],
        sv_model_file: Optional[str],
        model_tag: Optional[str],
        output_dir: Optional[str] = None,
        batch_size: int = 1,
        dtype: str = "float32",
        ngpu: int = 1,
        seed: int = 0,
        num_workers: int = 0,
        log_level: Union[int, str] = "INFO",
        key_file: Optional[str] = None,
        sv_train_config: Optional[str] = "sv.yaml",
        sv_model_file: Optional[str] =  "sv.pb",
        model_tag: Optional[str] = None,
        allow_variable_data_keys: bool = True,
        streaming: bool = False,
        embedding_node: str = "resnet1_dense",
        sv_threshold: float = 0.9465,
        param_dict: Optional[dict] = None,
        **kwargs,
):
    assert check_argument_types()
@@ -183,6 +188,7 @@
        level=log_level,
        format="%(asctime)s (%(module)s:%(lineno)d) %(levelname)s: %(message)s",
    )
    logging.info("param_dict: {}".format(param_dict))

    if ngpu >= 1 and torch.cuda.is_available():
        device = "cuda"
@@ -212,7 +218,9 @@
            data_path_and_name_and_type: Sequence[Tuple[str, str, str]] = None,
            raw_inputs: Union[np.ndarray, torch.Tensor] = None,
            output_dir_v2: Optional[str] = None,
            param_dict: Optional[dict] = None,
    ):
        logging.info("param_dict: {}".format(param_dict))
        if data_path_and_name_and_type is None and raw_inputs is not None:
            if isinstance(raw_inputs, torch.Tensor):
                raw_inputs = raw_inputs.numpy()
@@ -233,11 +241,10 @@

        # 7 .Start for-loop
        output_path = output_dir_v2 if output_dir_v2 is not None else output_dir
        embd_fd, ref_emb_fd, score_fd = None, None, None
        embd_writer, ref_embd_writer, score_writer = None, None, None
        if output_path is not None:
            os.makedirs(output_path, exist_ok=True)
            embd_writer = WriteHelper("ark:{}/xvector.ark".format(output_path))
            # embd_fd = open(os.path.join(output_path, "xvector.ark"), "wb")
            embd_writer = WriteHelper("ark,scp:{}/xvector.ark,{}/xvector.scp".format(output_path, output_path))
        sv_result_list = []
        for keys, batch in loader:
            assert isinstance(batch, dict), type(batch)
@@ -249,6 +256,7 @@
            embedding, ref_embedding, score = speech2xvector(**batch)
            # Only supporting batch_size==1
            key = keys[0]
            normalized_score = 0.0
            if score is not None:
                score = score.item()
                normalized_score = max(score - sv_threshold, 0.0) / (1.0 - sv_threshold) * 100.0
@@ -257,23 +265,21 @@
                item = {"key": key, "value": embedding.squeeze(0).cpu().numpy()}
            sv_result_list.append(item)
            if output_path is not None:
                # kaldiio.save_mat(embd_fd, embedding[0].cpu().numpy(), key)
                embd_writer(key, embedding[0].cpu().numpy())
                if ref_embedding is not None:
                    if ref_emb_fd is None:
                        # ref_emb_fd = open(os.path.join(output_path, "ref_xvector.ark"), "wb")
                        ref_embd_writer = WriteHelper("ark:{}/ref_xvector.ark".format(output_path))
                        score_fd = open(os.path.join(output_path, "score.txt"), "w")
                    # kaldiio.save_mat(ref_emb_fd, ref_embedding[0].cpu().numpy(), key)
                    if ref_embd_writer is None:
                        ref_embd_writer = WriteHelper(
                            "ark,scp:{}/ref_xvector.ark,{}/ref_xvector.scp".format(output_path, output_path)
                        )
                        score_writer = open(os.path.join(output_path, "score.txt"), "w")
                    ref_embd_writer(key, ref_embedding[0].cpu().numpy())
                    score_fd.write("{:.6f}\n".format(score.item()))
                    score_writer.write("{} {:.6f}\n".format(key, normalized_score))

        if output_path is not None:
            # embd_fd.close()
            embd_writer.close()
            if ref_emb_fd is not None:
                # ref_emb_fd.close()
                ref_emb_fd.close()
                score_fd.close()
            if ref_embd_writer is not None:
                ref_embd_writer.close()
                score_writer.close()

        return sv_result_list