📘-概述

CTCForcedAlignmentFilter is a filtering operator that filters data based on CTC forced-alignment scores from speech recognition results.

__init__

def __init__(
    self,
    model_path: str = "MahmoudAshraf/mms-300m-1130-forced-aligner",
    device: Union[str, List[str]] = "cuda",
    num_workers: int = 1,
    sampling_rate: int = 16000,
    language: str = "en",
    micro_batch_size: int = 16,
    chinese_to_pinyin: bool = False,
    retain_word_level_alignment: bool = True,
    romanize: bool = True,
    threshold: float = 0.8,
    threshold_mode: str = "min",
)

init Parameters

Parameter	Type	Default	Description
model_path	str	MahmoudAshraf/mms-300m-1130-forced-aligner	The model identifier or path for the forced aligner used during evaluation.
device	Union[str, List[str]]	cuda	The device on which the model runs. Options: cuda or cpu. You can also specify a list of devices, such as ["cuda:0", "cuda:1"], to initialize multiple models on multiple GPUs.
num_workers	int	1	Degree of operator parallelism. Initializes num_workers model instances and assigns them to the devices specified by device. If num_workers exceeds the number of devices, multiple models will be initialized per device for concurrent execution. For example, with device=["cuda:0", "cuda:1"] and num_workers=4, two models will run on cuda:0 and two on cuda:1.
sampling_rate	int	16000	Audio sampling rate, default 16000.
language	str	en	Audio language, default en.
micro_batch_size	int	16	For long audio, the model splits it into multiple chunks. micro_batch_size specifies the chunk batch size per inference, default 16.
chinese_to_pinyin	bool	False	Whether to convert Chinese characters to Pinyin, default False.
romanize	bool	True	Whether to romanize characters, default True.
threshold	float	0.8	Alignment score threshold, default 0.8.
threshold_mode	str	min	How to apply the threshold: min (filter by the minimum alignment score within a span window) or mean (filter by the average alignment score within a span window). Samples with scores ≥ threshold are kept. Default is min.

run

def run(
    self,
    storage: DataFlowStorage,
    input_audio_key: str = "audio",
    input_conversation_key: str = "conversation",
)

Parameters

Parameter	Type	Default	Description
storage	DataFlowStorage	Required	The data storage instance used to hold input and output data.
input_audio_key	str	audio	The key name for audio data in the input data, default is audio.
input_conversation_key	str	conversation	The key name for conversation data in the input data, default is conversation.
output_answer_key	str	forced_alignment_results	The key name of the alignment results field in the retained output data. Default is forced_alignment_results.

🧠 Example Usage

from dataflow.utils.storage import FileStorage
from dataflow.operators.core_audio import CTCForcedAlignmentFilter
from dataflow.wrapper import BatchWrapper

class testCTCForcedAlignmentFilter:
    def __init__(self):
        self.storage = FileStorage(
            # See audio_asr_pipeline step 2 to step 3 for an example
            first_entry_file_name="/path/to/your/cache/audio_asr_pipeline/audio_asr_pipeline_step2.jsonl",
            cache_path="./cache",
            file_name_prefix="forced_alignment_filter",
            cache_type="jsonl",
        )
        
        self.filter = CTCForcedAlignmentFilter(
            model_path="MahmoudAshraf/mms-300m-1130-forced-aligner",
            device=["cuda:0"],
            num_workers=1,
            sampling_rate=16000,
            language="en",  
            micro_batch_size=16,
            chinese_to_pinyin=False,
            romanize=True,
            threshold=0.8,
            threshold_mode="mean"
        )
    
    def forward(self):
        self.filter.run(
            storage=self.storage.step(),
            input_audio_key='audio',
            input_conversation_key='conversation',    
        )
        self.filter.close()

if __name__ == "__main__":
    pipline = testCTCForcedAlignFilter()
    pipline.forward()

🧾 Default Output Format

Field	Type	Description
forced_alignment_results	dict	Forced-alignment results where spans represent frame-level character alignment quality and word_timestamps represent word-level timestamp alignment.
error	Optional[str]	If an error occurs during alignment score computation, the error message is stored here. If no error occurs, error is null.

Samples that meet the alignment score threshold are retained. If all data are filtered out, the message All data has been filtered out! is printed.

Example Input：

{"audio":["..\/example_data\/audio_asr_pipeline\/test.wav"],"conversation":[{"from":"human","value":""}],"transcript":"and says how do i get to dublin and the answer that comes back is well i would not start from here sonny that is to say much of political philosophy develops theories that take no account of where we actually are and how the theories that people argue about in the journals and in the literature actually could be implemented in the world if it"}

Example Output： Retained:

{"audio":["..\/example_data\/audio_asr_pipeline\/test.wav"],"conversation":[{"from":"human","value":""}],"transcript":"and says how do i get to dublin and the answer that comes back is well i would not start from here sonny that is to say much of political philosophy develops theories that take no account of where we actually are and how the theories that people argue about in the journals and in the literature actually could be implemented in the world if it","forced_alignment_results":{"alignment":[{"start":0.063,"end":0.147,"text":"and","score":0.9554548212},{"start":0.273,"end":0.462,"text":"says","score":0.9719064832},{"start":0.609,"end":0.735,"text":"how","score":0.9212982873},{"start":0.798,"end":0.84,"text":"do","score":0.9939799858},{"start":1.029,"end":1.029,"text":"i","score":null},{"start":1.113,"end":1.26,"text":"get","score":0.9985639263},{"start":1.365,"end":1.428,"text":"to","score":0.9945560943},{"start":1.554,"end":1.974,"text":"dublin","score":0.9609149893},{"start":2.856,"end":2.94,"text":"and","score":0.9309501759},{"start":2.982,"end":3.045,"text":"the","score":0.7141059392},{"start":3.192,"end":3.465,"text":"answer","score":0.5938632981},{"start":3.507,"end":3.633,"text":"that","score":0.9633214426},{"start":3.717,"end":4.011,"text":"comes","score":0.9843271526},{"start":4.116,"end":4.389,"text":"back","score":0.9842618417},{"start":4.515,"end":4.662,"text":"is","score":0.9815290374},{"start":5.25,"end":5.376,"text":"well","score":0.047969851},{"start":5.502,"end":5.502,"text":"i","score":null},{"start":5.544,"end":5.67,"text":"would","score":0.8428627272},{"start":5.754,"end":5.817,"text":"not","score":0.123845133},{"start":5.88,"end":6.153,"text":"start","score":0.9789600127},{"start":6.216,"end":6.363,"text":"from","score":0.9000720539},{"start":6.468,"end":6.657,"text":"here","score":0.9283110266},{"start":6.783,"end":7.035,"text":"sonny","score":0.8839239278},{"start":9.807,"end":9.975,"text":"that","score":0.7547208776},{"start":10.038,"end":10.122,"text":"is","score":0.8797863669},{"start":10.185,"end":10.248,"text":"to","score":0.8244834454},{"start":10.353,"end":10.542,"text":"say","score":0.9471999446},{"start":11.025,"end":11.34,"text":"much","score":0.9940719048},{"start":11.634,"end":11.802,"text":"of","score":0.9950778359},{"start":11.991,"end":12.621,"text":"political","score":0.9989232361},{"start":12.81,"end":13.629,"text":"philosophy","score":0.9465096714},{"start":14.217,"end":14.805,"text":"develops","score":0.9432990222},{"start":15.057,"end":15.666,"text":"theories","score":0.9267864129},{"start":17.136,"end":17.304,"text":"that","score":0.8086037475},{"start":17.43,"end":17.682,"text":"take","score":0.9565847912},{"start":17.829,"end":17.913,"text":"no","score":0.956001711},{"start":18.081,"end":18.648,"text":"account","score":0.9546385136},{"start":19.425,"end":19.656,"text":"of","score":0.8420175488},{"start":21.42,"end":21.567,"text":"where","score":0.7551332315},{"start":21.63,"end":21.693,"text":"we","score":0.9166198867},{"start":21.903,"end":22.323,"text":"actually","score":0.9312994611},{"start":22.512,"end":22.701,"text":"are","score":0.9616599245},{"start":22.89,"end":22.974,"text":"and","score":0.4025359219},{"start":23.079,"end":23.31,"text":"how","score":0.9633893459},{"start":23.436,"end":23.499,"text":"the","score":0.7716538814},{"start":23.625,"end":24.045,"text":"theories","score":0.9761697651},{"start":24.15,"end":24.36,"text":"that","score":0.9068021914},{"start":24.486,"end":24.78,"text":"people","score":0.9219708612},{"start":24.948,"end":25.2,"text":"argue","score":0.9620480049},{"start":25.242,"end":25.515,"text":"about","score":0.9651158228},{"start":25.641,"end":25.704,"text":"in","score":0.9931364561},{"start":25.767,"end":25.83,"text":"the","score":0.8166649179},{"start":25.956,"end":26.439,"text":"journals","score":0.9695284503},{"start":26.544,"end":26.607,"text":"and","score":0.9435737354},{"start":26.67,"end":26.712,"text":"in","score":0.778872343},{"start":26.754,"end":26.796,"text":"the","score":0.8787819404},{"start":26.88,"end":27.384,"text":"literature","score":0.928246194},{"start":27.804,"end":28.077,"text":"actually","score":0.9179609355},{"start":28.119,"end":28.266,"text":"could","score":0.8717020111},{"start":28.329,"end":28.392,"text":"be","score":0.9910494216},{"start":28.602,"end":29.169,"text":"implemented","score":0.9847475907},{"start":29.232,"end":29.274,"text":"in","score":0.9814222521},{"start":29.337,"end":29.379,"text":"the","score":0.8807633297},{"start":29.442,"end":29.736,"text":"world","score":0.9051810523},{"start":30.156,"end":30.24,"text":"if","score":0.7553217096},{"start":30.45,"end":30.471,"text":"it","score":0.0156467184}],"error":null}}

If everything is filtered out, the following string will be printed:

All data has been filtered out!