import itertools import os import zlib from typing import BinaryIO, Iterable, List, NamedTuple, Optional, Tuple, Union import ctranslate2 import numpy as np import tokenizers from faster_whisper.audio import decode_audio from faster_whisper.feature_extractor import FeatureExtractor from faster_whisper.tokenizer import Tokenizer from faster_whisper.utils import download_model class Word(NamedTuple): start: float end: float word: str probability: float class Segment(NamedTuple): start: float end: float text: str words: Optional[List[Word]] class AudioInfo(NamedTuple): language: str language_probability: float duration: float class TranscriptionOptions(NamedTuple): beam_size: int best_of: int patience: float length_penalty: float log_prob_threshold: Optional[float] no_speech_threshold: Optional[float] compression_ratio_threshold: Optional[float] condition_on_previous_text: bool temperatures: List[float] initial_prompt: Optional[str] prefix: Optional[str] suppress_blank: bool suppress_tokens: Optional[List[int]] without_timestamps: bool max_initial_timestamp: float word_timestamps: bool prepend_punctuations: str append_punctuations: str class WhisperModel: def __init__( self, model_size_or_path: str, device: str = "auto", device_index: Union[int, List[int]] = 0, compute_type: str = "default", cpu_threads: int = 0, num_workers: int = 1, ): """Initializes the Whisper model. Args: model_size_or_path: Size of the model to use (e.g. "large-v2", "small", "tiny.en", etc.) or a path to a converted model directory. When a size is configured, the converted model is downloaded from the Hugging Face Hub. device: Device to use for computation ("cpu", "cuda", "auto"). device_index: Device ID to use. The model can also be loaded on multiple GPUs by passing a list of IDs (e.g. [0, 1, 2, 3]). In that case, multiple transcriptions can run in parallel when transcribe() is called from multiple Python threads (see also num_workers). compute_type: Type to use for computation. See https://opennmt.net/CTranslate2/quantization.html. cpu_threads: Number of threads to use when running on CPU (4 by default). A non zero value overrides the OMP_NUM_THREADS environment variable. num_workers: When transcribe() is called from multiple Python threads, having multiple workers enables true parallelism when running the model (concurrent calls to self.model.generate() will run in parallel). This can improve the global throughput at the cost of increased memory usage. """ if os.path.isdir(model_size_or_path): model_path = model_size_or_path else: model_path = download_model(model_size_or_path) self.model = ctranslate2.models.Whisper( model_path, device=device, device_index=device_index, compute_type=compute_type, intra_threads=cpu_threads, inter_threads=num_workers, ) tokenizer_file = os.path.join(model_path, "tokenizer.json") if os.path.isfile(tokenizer_file): self.hf_tokenizer = tokenizers.Tokenizer.from_file(tokenizer_file) else: self.hf_tokenizer = tokenizers.Tokenizer.from_pretrained( "openai/whisper-tiny" + ("" if self.model.is_multilingual else ".en") ) self.feature_extractor = FeatureExtractor() self.num_samples_per_token = self.feature_extractor.hop_length * 2 self.frames_per_second = ( self.feature_extractor.sampling_rate // self.feature_extractor.hop_length ) self.tokens_per_second = ( self.feature_extractor.sampling_rate // self.num_samples_per_token ) self.input_stride = 2 self.time_precision = 0.02 self.max_length = 448 def transcribe( self, audio: Union[str, BinaryIO, np.ndarray], language: Optional[str] = None, task: str = "transcribe", beam_size: int = 5, best_of: int = 5, patience: float = 1, length_penalty: float = 1, temperature: Union[float, List[float], Tuple[float, ...]] = [ 0.0, 0.2, 0.4, 0.6, 0.8, 1.0, ], compression_ratio_threshold: Optional[float] = 2.4, log_prob_threshold: Optional[float] = -1.0, no_speech_threshold: Optional[float] = 0.6, condition_on_previous_text: bool = True, initial_prompt: Optional[str] = None, prefix: Optional[str] = None, suppress_blank: bool = True, suppress_tokens: Optional[List[int]] = [-1], without_timestamps: bool = False, max_initial_timestamp: float = 1.0, word_timestamps: bool = False, prepend_punctuations: str = "\"'“¿([{-", append_punctuations: str = "\"'.。,，!！?？:：”)]}、", ) -> Tuple[Iterable[Segment], AudioInfo]: """Transcribes an input file. Arguments: audio: Path to the input file (or a file-like object), or the audio waveform. language: The language spoken in the audio. It should be a language code such as "en" or "fr". If not set, the language will be detected in the first 30 seconds of audio. task: Task to execute (transcribe or translate). beam_size: Beam size to use for decoding. best_of: Number of candidates when sampling with non-zero temperature. patience: Beam search patience factor. length_penalty: Exponential length penalty constant. temperature: Temperature for sampling. It can be a tuple of temperatures, which will be successively used upon failures according to either `compression_ratio_threshold` or `logprob_threshold`. compression_ratio_threshold: If the gzip compression ratio is above this value, treat as failed. log_prob_threshold: If the average log probability over sampled tokens is below this value, treat as failed. no_speech_threshold: If the no_speech probability is higher than this value AND the average log probability over sampled tokens is below `logprob_threshold`, consider the segment as silent. condition_on_previous_text: If True, the previous output of the model is provided as a prompt for the next window; disabling may make the text inconsistent across windows, but the model becomes less prone to getting stuck in a failure loop, such as repetition looping or timestamps going out of sync. initial_prompt: Optional text to provide as a prompt for the first window. prefix: Optional text to provide as a prefix for the first window. suppress_blank: Suppress blank outputs at the beginning of the sampling. suppress_tokens: List of token IDs to suppress. -1 will suppress a default set of symbols as defined in the model config.json file. without_timestamps: Only sample text tokens. max_initial_timestamp: The initial timestamp cannot be later than this. word_timestamps: Extract word-level timestamps using the cross-attention pattern and dynamic time warping, and include the timestamps for each word in each segment. prepend_punctuations: If word_timestamps is True, merge these punctuation symbols with the next word append_punctuations: If word_timestamps is True, merge these punctuation symbols with the previous word Returns: A tuple with: - a generator over transcribed segments - an instance of AudioInfo """ if not isinstance(audio, np.ndarray): audio = decode_audio( audio, sampling_rate=self.feature_extractor.sampling_rate ) duration = audio.shape[0] / self.feature_extractor.sampling_rate features = self.feature_extractor(audio) whisper_encoder = WhisperEncoder(self.model) if language is None: if not self.model.is_multilingual: language = "en" language_probability = 1 else: segment = features[:, : self.feature_extractor.nb_max_frames] encoder_output = whisper_encoder(0, segment) results = self.model.detect_language(encoder_output) language_token, language_probability = results[0][0] language = language_token[2:-2] else: language_probability = 1 tokenizer = Tokenizer( self.hf_tokenizer, self.model.is_multilingual, task=task, language=language, ) options = TranscriptionOptions( beam_size=beam_size, best_of=best_of, patience=patience, length_penalty=length_penalty, log_prob_threshold=log_prob_threshold, no_speech_threshold=no_speech_threshold, compression_ratio_threshold=compression_ratio_threshold, condition_on_previous_text=condition_on_previous_text, temperatures=( temperature if isinstance(temperature, (list, tuple)) else [temperature] ), initial_prompt=initial_prompt, prefix=prefix, suppress_blank=suppress_blank, suppress_tokens=suppress_tokens, without_timestamps=without_timestamps, max_initial_timestamp=max_initial_timestamp, word_timestamps=word_timestamps, prepend_punctuations=prepend_punctuations, append_punctuations=append_punctuations, ) segments = self.generate_segments(features, whisper_encoder, tokenizer, options) audio_info = AudioInfo( language=language, language_probability=language_probability, duration=duration, ) return segments, audio_info def generate_segments( self, features: np.ndarray, whisper_encoder: "WhisperEncoder", tokenizer: Tokenizer, options: TranscriptionOptions, ) -> Iterable[Segment]: content_frames = features.shape[-1] - self.feature_extractor.nb_max_frames seek = 0 all_tokens = [] prompt_reset_since = 0 if options.initial_prompt is not None: initial_prompt = " " + options.initial_prompt.strip() initial_prompt_tokens = tokenizer.encode(initial_prompt) all_tokens.extend(initial_prompt_tokens) while seek < content_frames: time_offset = seek * self.feature_extractor.time_per_frame segment = features[:, seek : seek + self.feature_extractor.nb_max_frames] segment_size = min( self.feature_extractor.nb_max_frames, content_frames - seek ) segment_duration = segment_size * self.feature_extractor.time_per_frame previous_tokens = all_tokens[prompt_reset_since:] prompt = self.get_prompt( tokenizer, previous_tokens, without_timestamps=options.without_timestamps, prefix=options.prefix, ) encoder_output = whisper_encoder(seek, segment) result, avg_log_prob, temperature = self.generate_with_fallback( encoder_output, prompt, tokenizer, options ) if options.no_speech_threshold is not None: # no voice activity check should_skip = result.no_speech_prob > options.no_speech_threshold if ( options.log_prob_threshold is not None and avg_log_prob > options.log_prob_threshold ): # don't skip if the logprob is high enough, despite the no_speech_prob should_skip = False if should_skip: # fast-forward to the next segment boundary seek += segment_size continue tokens = result.sequences_ids[0] previous_seek = seek current_segments = [] single_timestamp_ending = ( len(tokens) >= 2 and tokens[-2] < tokenizer.timestamp_begin and tokens[-1] >= tokenizer.timestamp_begin ) consecutive_timestamps = [ i for i in range(len(tokens)) if i > 0 and tokens[i] >= tokenizer.timestamp_begin and tokens[i - 1] >= tokenizer.timestamp_begin ] if len(consecutive_timestamps) > 0: slices = list(consecutive_timestamps) if single_timestamp_ending: slices.append(len(tokens)) last_slice = 0 for current_slice in slices: sliced_tokens = tokens[last_slice:current_slice] start_timestamp_position = ( sliced_tokens[0] - tokenizer.timestamp_begin ) end_timestamp_position = ( sliced_tokens[-1] - tokenizer.timestamp_begin ) start_time = ( time_offset + start_timestamp_position * self.time_precision ) end_time = ( time_offset + end_timestamp_position * self.time_precision ) current_segments.append( dict( seek=seek, start=start_time, end=end_time, tokens=sliced_tokens, ) ) last_slice = current_slice if single_timestamp_ending: # single timestamp at the end means no speech after the last timestamp. seek += segment_size else: # otherwise, ignore the unfinished segment and seek to the last timestamp last_timestamp_position = ( tokens[last_slice - 1] - tokenizer.timestamp_begin ) seek += last_timestamp_position * self.input_stride else: duration = segment_duration timestamps = [ token for token in tokens if token >= tokenizer.timestamp_begin ] if len(timestamps) > 0 and timestamps[-1] != tokenizer.timestamp_begin: last_timestamp_position = timestamps[-1] - tokenizer.timestamp_begin duration = last_timestamp_position * self.time_precision current_segments.append( dict( seek=seek, start=time_offset, end=time_offset + duration, tokens=tokens, ) ) seek += segment_size if not options.condition_on_previous_text or temperature > 0.5: prompt_reset_since = len(all_tokens) if options.word_timestamps: self.add_word_timestamps( current_segments, tokenizer, encoder_output, segment_size, options.prepend_punctuations, options.append_punctuations, ) word_end_timestamps = [ w["end"] for s in current_segments for w in s["words"] ] if not single_timestamp_ending and len(word_end_timestamps) > 0: seek_shift = round( (word_end_timestamps[-1] - time_offset) * self.frames_per_second ) if seek_shift > 0: seek = previous_seek + seek_shift for segment in current_segments: tokens = segment["tokens"] text = tokenizer.decode(tokens) if segment["start"] == segment["end"] or not text.strip(): continue all_tokens.extend(tokens) yield Segment( start=segment["start"], end=segment["end"], text=text, words=( [Word(**word) for word in segment["words"]] if options.word_timestamps else None ), ) def generate_with_fallback( self, encoder_output: ctranslate2.StorageView, prompt: List[int], tokenizer: Tokenizer, options: TranscriptionOptions, ) -> Tuple[ctranslate2.models.WhisperGenerationResult, float, float]: result = None avg_log_prob = None final_temperature = None max_initial_timestamp_index = int( round(options.max_initial_timestamp / self.time_precision) ) for temperature in options.temperatures: if temperature > 0: kwargs = { "beam_size": 1, "num_hypotheses": options.best_of, "sampling_topk": 0, "sampling_temperature": temperature, } else: kwargs = { "beam_size": options.beam_size, "patience": options.patience, } final_temperature = temperature result = self.model.generate( encoder_output, [prompt], length_penalty=options.length_penalty, max_length=self.max_length, return_scores=True, return_no_speech_prob=True, suppress_blank=options.suppress_blank, suppress_tokens=options.suppress_tokens, max_initial_timestamp_index=max_initial_timestamp_index, **kwargs, )[0] tokens = result.sequences_ids[0] # Recover the average log prob from the returned score. seq_len = len(tokens) cum_log_prob = result.scores[0] * (seq_len**options.length_penalty) avg_log_prob = cum_log_prob / (seq_len + 1) text = tokenizer.decode(tokens).strip() compression_ratio = get_compression_ratio(text) needs_fallback = False if ( options.compression_ratio_threshold is not None and compression_ratio > options.compression_ratio_threshold ): needs_fallback = True # too repetitive if ( options.log_prob_threshold is not None and avg_log_prob < options.log_prob_threshold ): needs_fallback = True # average log probability is too low if not needs_fallback: break return result, avg_log_prob, final_temperature def get_prompt( self, tokenizer: Tokenizer, previous_tokens: List[int], without_timestamps: bool = False, prefix: Optional[str] = None, ) -> List[int]: prompt = [] if previous_tokens: prompt.append(tokenizer.sot_prev) prompt.extend(previous_tokens[-(self.max_length // 2 - 1) :]) prompt.extend(tokenizer.sot_sequence) if without_timestamps: prompt.append(tokenizer.no_timestamps) if prefix: prefix_tokens = tokenizer.encode(" " + prefix.strip()) if len(prefix_tokens) >= self.max_length // 2: prefix_tokens = prefix_tokens[: self.max_length // 2 - 1] prompt.extend(prefix_tokens) return prompt def add_word_timestamps( self, segments: List[dict], tokenizer: Tokenizer, encoder_output: ctranslate2.StorageView, num_frames: int, prepend_punctuations: str, append_punctuations: str, ): if len(segments) == 0: return text_tokens_per_segment = [ [token for token in segment["tokens"] if token < tokenizer.eot] for segment in segments ] text_tokens = list(itertools.chain.from_iterable(text_tokens_per_segment)) alignment = self.find_alignment( tokenizer, text_tokens, encoder_output, num_frames ) merge_punctuations(alignment, prepend_punctuations, append_punctuations) time_offset = ( segments[0]["seek"] * self.feature_extractor.hop_length / self.feature_extractor.sampling_rate ) word_index = 0 for segment, text_tokens in zip(segments, text_tokens_per_segment): saved_tokens = 0 words = [] while word_index < len(alignment) and saved_tokens < len(text_tokens): timing = alignment[word_index] if timing["word"]: words.append( dict( word=timing["word"], start=round(time_offset + timing["start"], 2), end=round(time_offset + timing["end"], 2), probability=timing["probability"], ) ) saved_tokens += len(timing["tokens"]) word_index += 1 if len(words) > 0: # adjust the segment-level timestamps based on the word-level timestamps segment["start"] = words[0]["start"] segment["end"] = words[-1]["end"] segment["words"] = words def find_alignment( self, tokenizer: Tokenizer, text_tokens: List[int], encoder_output: ctranslate2.StorageView, num_frames: int, median_filter_width: int = 7, ) -> List[dict]: if len(text_tokens) == 0: return [] result = self.model.align( encoder_output, tokenizer.sot_sequence, [text_tokens], num_frames, median_filter_width=median_filter_width, )[0] text_token_probs = result.text_token_probs alignments = result.alignments text_indices = np.array([pair[0] for pair in alignments]) time_indices = np.array([pair[1] for pair in alignments]) words, word_tokens = tokenizer.split_to_word_tokens( text_tokens + [tokenizer.eot] ) word_boundaries = np.pad(np.cumsum([len(t) for t in word_tokens[:-1]]), (1, 0)) jumps = np.pad(np.diff(text_indices), (1, 0), constant_values=1).astype(bool) jump_times = time_indices[jumps] / self.tokens_per_second start_times = jump_times[word_boundaries[:-1]] end_times = jump_times[word_boundaries[1:]] word_probabilities = [ np.mean(text_token_probs[i:j]) for i, j in zip(word_boundaries[:-1], word_boundaries[1:]) ] # hack: ensure the first and second word is not longer than twice the median word duration. # a better segmentation algorithm based on VAD should be able to replace this. word_durations = end_times - start_times word_durations = word_durations[word_durations.nonzero()] if len(word_durations) > 0: median_duration = np.median(word_durations) max_duration = median_duration * 2 if len(word_durations) >= 2 and word_durations[1] > max_duration: boundary = max(end_times[2] / 2, end_times[2] - max_duration) end_times[0] = start_times[1] = boundary if ( len(word_durations) >= 1 and end_times[0] - start_times[0] > max_duration ): start_times[0] = max(0, end_times[0] - max_duration) return [ dict( word=word, tokens=tokens, start=start, end=end, probability=probability ) for word, tokens, start, end, probability in zip( words, word_tokens, start_times, end_times, word_probabilities ) ] class WhisperEncoder: """Helper class to cache and reuse the encoder output.""" def __init__(self, model: ctranslate2.models.Whisper): self.model = model # When the model is running on multiple GPUs, the encoder output should be moved # to the CPU since we don't know which GPU will handle the next job. self.cache_on_cpu = len(model.device_index) > 1 self.last_seek = -1 self.last_output = None def __call__(self, seek: int, features: np.ndarray) -> ctranslate2.StorageView: if self.last_seek == seek: return self.last_output features = np.expand_dims(features, 0) features = get_ctranslate2_storage(features) output = self.model.encode(features, to_cpu=self.cache_on_cpu) if self.last_output is not None: del self.last_output self.last_seek = seek self.last_output = output return output def get_ctranslate2_storage(segment: np.ndarray) -> ctranslate2.StorageView: segment = np.ascontiguousarray(segment) segment = ctranslate2.StorageView.from_array(segment) return segment def get_compression_ratio(text: str) -> float: text_bytes = text.encode("utf-8") return len(text_bytes) / len(zlib.compress(text_bytes)) def merge_punctuations(alignment: List[dict], prepended: str, appended: str): # merge prepended punctuations i = len(alignment) - 2 j = len(alignment) - 1 while i >= 0: previous = alignment[i] following = alignment[j] if previous["word"].startswith(" ") and previous["word"].strip() in prepended: # prepend it to the following word following["word"] = previous["word"] + following["word"] following["tokens"] = previous["tokens"] + following["tokens"] previous["word"] = "" previous["tokens"] = [] else: j = i i -= 1 # merge appended punctuations i = 0 j = 1 while j < len(alignment): previous = alignment[i] following = alignment[j] if not previous["word"].endswith(" ") and following["word"] in appended: # append it to the previous word previous["word"] = previous["word"] + following["word"] previous["tokens"] = previous["tokens"] + following["tokens"] following["word"] = "" following["tokens"] = [] else: i = j j += 1