python/FunASR-XL.git

			@@ -10,7 +10,7 @@
			from torch import Tensor
			from torch import nn
			from torch.cuda.amp import autocast
			from funasr.metrics.compute_acc import compute_accuracy
			from funasr.metrics.compute_acc import compute_accuracy, th_accuracy
			from funasr.losses.label_smoothing_loss import LabelSmoothingLoss
			from funasr.train_utils.device_funcs import force_gatherable
			from . import whisper_lib as whisper
			@@ -662,9 +662,11 @@
			else:
			encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)

			loss_att, acc_att, cer_att, wer_att = self._calc_att_loss(
			encoder_out, encoder_out_lens, text, text_lengths, target_mask=target_mask
			)
			with autocast(False):
			loss_att, acc_att, cer_att, wer_att = self._calc_att_loss(
			encoder_out, encoder_out_lens, text, text_lengths, target_mask=target_mask
			)

			loss = loss_att
			stats = {}
			stats["acc"] = acc_att
			@@ -1390,275 +1392,3 @@

			from funasr.models.paraformer.search import Hypothesis
			from funasr.utils import postprocess_utils


			@tables.register("model_classes", "SenseVoiceSANMCTC")
			class SenseVoiceSANMCTC(nn.Module):
			"""CTC-attention hybrid Encoder-Decoder model"""

			def __init__(
			self,
			specaug: str = None,
			specaug_conf: dict = None,
			normalize: str = None,
			normalize_conf: dict = None,
			encoder: str = None,
			encoder_conf: dict = None,
			ctc_conf: dict = None,
			input_size: int = 80,
			vocab_size: int = -1,
			ignore_id: int = -1,
			blank_id: int = 0,
			sos: int = 1,
			eos: int = 2,
			length_normalized_loss: bool = False,
			**kwargs,
			):

			super().__init__()

			if specaug is not None:
			specaug_class = tables.specaug_classes.get(specaug)
			specaug = specaug_class(**specaug_conf)
			if normalize is not None:
			normalize_class = tables.normalize_classes.get(normalize)
			normalize = normalize_class(**normalize_conf)
			encoder_class = tables.encoder_classes.get(encoder)
			encoder = encoder_class(input_size=input_size, **encoder_conf)
			encoder_output_size = encoder.output_size()

			if ctc_conf is None:
			ctc_conf = {}
			ctc = CTC(odim=vocab_size, encoder_output_size=encoder_output_size, **ctc_conf)

			self.blank_id = blank_id
			self.sos = sos if sos is not None else vocab_size - 1
			self.eos = eos if eos is not None else vocab_size - 1
			self.vocab_size = vocab_size
			self.ignore_id = ignore_id
			self.specaug = specaug
			self.normalize = normalize
			self.encoder = encoder
			self.error_calculator = None

			self.ctc = ctc

			self.length_normalized_loss = length_normalized_loss
			self.encoder_output_size = encoder_output_size

			self.lid_dict = {"zh": 3, "en": 4, "yue": 7, "ja": 11, "ko": 12, "nospeech": 13}
			self.textnorm_dict = {"withtextnorm": 14, "wotextnorm": 15}
			self.embed = torch.nn.Embedding(8 + len(self.lid_dict) + len(self.textnorm_dict), 560)

			def forward(
			self,
			speech: torch.Tensor,
			speech_lengths: torch.Tensor,
			text: torch.Tensor,
			text_lengths: torch.Tensor,
			**kwargs,
			):
			"""Encoder + Decoder + Calc loss
			Args:
			speech: (Batch, Length, ...)
			speech_lengths: (Batch, )
			text: (Batch, Length)
			text_lengths: (Batch,)
			"""
			# import pdb;
			# pdb.set_trace()
			if len(text_lengths.size()) > 1:
			text_lengths = text_lengths[:, 0]
			if len(speech_lengths.size()) > 1:
			speech_lengths = speech_lengths[:, 0]

			batch_size = speech.shape[0]

			# 1. Encoder
			encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)

			loss_ctc, cer_ctc = None, None
			stats = dict()

			loss_ctc, cer_ctc = self._calc_ctc_loss(encoder_out, encoder_out_lens, text, text_lengths)

			loss = loss_ctc

			# Collect total loss stats
			stats["loss"] = torch.clone(loss.detach())

			# force_gatherable: to-device and to-tensor if scalar for DataParallel
			if self.length_normalized_loss:
			batch_size = int((text_lengths + 1).sum())
			loss, stats, weight = force_gatherable((loss, stats, batch_size), loss.device)
			return loss, stats, weight

			def encode(
			self,
			speech: torch.Tensor,
			speech_lengths: torch.Tensor,
			**kwargs,
			):
			"""Frontend + Encoder. Note that this method is used by asr_inference.py
			Args:
			speech: (Batch, Length, ...)
			speech_lengths: (Batch, )
			ind: int
			"""

			# Data augmentation
			if self.specaug is not None and self.training:
			speech, speech_lengths = self.specaug(speech, speech_lengths)

			# Normalization for feature: e.g. Global-CMVN, Utterance-CMVN
			if self.normalize is not None:
			speech, speech_lengths = self.normalize(speech, speech_lengths)

			# Forward encoder
			# feats: (Batch, Length, Dim)
			# -> encoder_out: (Batch, Length2, Dim2)
			encoder_out, encoder_out_lens = self.encoder(speech, speech_lengths)

			return encoder_out, encoder_out_lens

			def _calc_ctc_loss(
			self,
			encoder_out: torch.Tensor,
			encoder_out_lens: torch.Tensor,
			ys_pad: torch.Tensor,
			ys_pad_lens: torch.Tensor,
			):
			# Calc CTC loss
			loss_ctc = self.ctc(encoder_out, encoder_out_lens, ys_pad, ys_pad_lens)

			# Calc CER using CTC
			cer_ctc = None
			if not self.training and self.error_calculator is not None:
			ys_hat = self.ctc.argmax(encoder_out).data
			cer_ctc = self.error_calculator(ys_hat.cpu(), ys_pad.cpu(), is_ctc=True)
			return loss_ctc, cer_ctc

			def inference(
			self,
			data_in,
			data_lengths=None,
			key: list = None,
			tokenizer=None,
			frontend=None,
			**kwargs,
			):

			if kwargs.get("batch_size", 1) > 1:
			raise NotImplementedError("batch decoding is not implemented")

			meta_data = {}
			if (
			isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank"
			): # fbank
			speech, speech_lengths = data_in, data_lengths
			if len(speech.shape) < 3:
			speech = speech[None, :, :]
			if speech_lengths is None:
			speech_lengths = speech.shape[1]
			else:
			# extract fbank feats
			time1 = time.perf_counter()
			audio_sample_list = load_audio_text_image_video(
			data_in,
			fs=frontend.fs,
			audio_fs=kwargs.get("fs", 16000),
			data_type=kwargs.get("data_type", "sound"),
			tokenizer=tokenizer,
			)
			time2 = time.perf_counter()
			meta_data["load_data"] = f"{time2 - time1:0.3f}"
			speech, speech_lengths = extract_fbank(
			audio_sample_list, data_type=kwargs.get("data_type", "sound"), frontend=frontend
			)
			time3 = time.perf_counter()
			meta_data["extract_feat"] = f"{time3 - time2:0.3f}"
			meta_data["batch_data_time"] = (
			speech_lengths.sum().item() * frontend.frame_shift * frontend.lfr_n / 1000
			)

			speech = speech.to(device=kwargs["device"])
			speech_lengths = speech_lengths.to(device=kwargs["device"])

			language = kwargs.get("language", None)
			if language is not None:
			language_query = self.embed(
			torch.LongTensor(
			[[self.lid_dict[language] if language in self.lid_dict else 0]]
			).to(speech.device)
			).repeat(speech.size(0), 1, 1)
			else:
			language_query = self.embed(torch.LongTensor([[0]]).to(speech.device)).repeat(
			speech.size(0), 1, 1
			)
			textnorm = kwargs.get("text_norm", "wotextnorm")
			textnorm_query = self.embed(
			torch.LongTensor([[self.textnorm_dict[textnorm]]]).to(speech.device)
			).repeat(speech.size(0), 1, 1)
			speech = torch.cat((textnorm_query, speech), dim=1)
			speech_lengths += 1

			event_emo_query = self.embed(torch.LongTensor([[1, 2]]).to(speech.device)).repeat(
			speech.size(0), 1, 1
			)
			input_query = torch.cat((language_query, event_emo_query), dim=1)
			speech = torch.cat((input_query, speech), dim=1)
			speech_lengths += 3

			# Encoder
			encoder_out, encoder_out_lens = self.encode(speech, speech_lengths)
			if isinstance(encoder_out, tuple):
			encoder_out = encoder_out[0]

			# c. Passed the encoder result and the beam search
			ctc_logits = self.ctc.log_softmax(encoder_out)

			results = []
			b, n, d = encoder_out.size()
			if isinstance(key[0], (list, tuple)):
			key = key[0]
			if len(key) < b:
			key = key * b
			for i in range(b):
			x = ctc_logits[i, : encoder_out_lens[i], :]
			yseq = x.argmax(dim=-1)
			yseq = torch.unique_consecutive(yseq, dim=-1)
			yseq = torch.tensor([self.sos] + yseq.tolist() + [self.eos], device=yseq.device)
			nbest_hyps = [Hypothesis(yseq=yseq)]

			for nbest_idx, hyp in enumerate(nbest_hyps):
			ibest_writer = None
			if kwargs.get("output_dir") is not None:
			if not hasattr(self, "writer"):
			self.writer = DatadirWriter(kwargs.get("output_dir"))
			ibest_writer = self.writer[f"{nbest_idx + 1}best_recog"]

			# remove sos/eos and get results
			last_pos = -1
			if isinstance(hyp.yseq, list):
			token_int = hyp.yseq[1:last_pos]
			else:
			token_int = hyp.yseq[1:last_pos].tolist()

			# remove blank symbol id, which is assumed to be 0
			token_int = list(
			filter(
			lambda x: x != self.eos and x != self.sos and x != self.blank_id, token_int
			)
			)

			# Change integer-ids to tokens
			text = tokenizer.decode(token_int)

			result_i = {"key": key[i], "text": text}
			results.append(result_i)

			if ibest_writer is not None:
			ibest_writer["token"][key[i]] = " ".join(token)
			ibest_writer["text"][key[i]] = text_postprocessed

			return results, meta_data