speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge pull request #976 from alibaba-damo-academy/dev_lhn

Browse Source 
Dev lhn
This commit is contained in:
hnluo 2023-09-21 16:30:43 +08:00 committed by GitHub
commit bf4b3ef9cb
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 0 additions and 105 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

59
funasr/models/decoder/sanm_decoder.py
View File

@ -1035,65 +1035,6 @@ class ParaformerSANMDecoder(BaseTransformerDecoder):
)
return logp.squeeze(0), state
#def forward_chunk(
# self,
# memory: torch.Tensor,
# tgt: torch.Tensor,
# cache: dict = None,
#) -> Tuple[torch.Tensor, torch.Tensor]:
# """Forward decoder.
# Args:
# hs_pad: encoded memory, float32 (batch, maxlen_in, feat)
# hlens: (batch)
# ys_in_pad:
# input token ids, int64 (batch, maxlen_out)
# if input_layer == "embed"
# input tensor (batch, maxlen_out, #mels) in the other cases
# ys_in_lens: (batch)
# Returns:
# (tuple): tuple containing:
# x: decoded token score before softmax (batch, maxlen_out, token)
# if use_output_layer is True,
# olens: (batch, )
# """
# x = tgt
# if cache["decode_fsmn"] is None:
# cache_layer_num = len(self.decoders)
# if self.decoders2 is not None:
# cache_layer_num += len(self.decoders2)
# new_cache = [None] * cache_layer_num
# else:
# new_cache = cache["decode_fsmn"]
# for i in range(self.att_layer_num):
# decoder = self.decoders[i]
# x, tgt_mask, memory, memory_mask, c_ret = decoder.forward_chunk(
# x, None, memory, None, cache=new_cache[i]
# )
# new_cache[i] = c_ret
# if self.num_blocks - self.att_layer_num > 1:
# for i in range(self.num_blocks - self.att_layer_num):
# j = i + self.att_layer_num
# decoder = self.decoders2[i]
# x, tgt_mask, memory, memory_mask, c_ret = decoder.forward_chunk(
# x, None, memory, None, cache=new_cache[j]
# )
# new_cache[j] = c_ret
# for decoder in self.decoders3:
# x, tgt_mask, memory, memory_mask, _ = decoder.forward_chunk(
# x, None, memory, None, cache=None
# )
# if self.normalize_before:
# x = self.after_norm(x)
# if self.output_layer is not None:
# x = self.output_layer(x)
# cache["decode_fsmn"] = new_cache
# return x
def forward_chunk(
self,
memory: torch.Tensor,

46
funasr/models/encoder/sanm_encoder.py
View File

@ -873,52 +873,6 @@ class SANMEncoderChunkOpt(AbsEncoder):
cache["feats"] = overlap_feats[:, -(cache["chunk_size"][0] + cache["chunk_size"][2]):, :]
return overlap_feats
#def forward_chunk(self,
# xs_pad: torch.Tensor,
# ilens: torch.Tensor,
# cache: dict = None,
# ctc: CTC = None,
# ):
# xs_pad *= self.output_size() ** 0.5
# if self.embed is None:
# xs_pad = xs_pad
# else:
# xs_pad = self.embed(xs_pad, cache)
# if cache["tail_chunk"]:
# xs_pad = to_device(cache["feats"], device=xs_pad.device)
# else:
# xs_pad = self._add_overlap_chunk(xs_pad, cache)
# encoder_outs = self.encoders0(xs_pad, None, None, None, None)
# xs_pad, masks = encoder_outs[0], encoder_outs[1]
# intermediate_outs = []
# if len(self.interctc_layer_idx) == 0:
# encoder_outs = self.encoders(xs_pad, None, None, None, None)
# xs_pad, masks = encoder_outs[0], encoder_outs[1]
# else:
# for layer_idx, encoder_layer in enumerate(self.encoders):
# encoder_outs = encoder_layer(xs_pad, None, None, None, None)
# xs_pad, masks = encoder_outs[0], encoder_outs[1]
# if layer_idx + 1 in self.interctc_layer_idx:
# encoder_out = xs_pad
# # intermediate outputs are also normalized
# if self.normalize_before:
# encoder_out = self.after_norm(encoder_out)
# intermediate_outs.append((layer_idx + 1, encoder_out))
# if self.interctc_use_conditioning:
# ctc_out = ctc.softmax(encoder_out)
# xs_pad = xs_pad + self.conditioning_layer(ctc_out)
# if self.normalize_before:
# xs_pad = self.after_norm(xs_pad)
# if len(intermediate_outs) > 0:
# return (xs_pad, intermediate_outs), None, None
# return xs_pad, ilens, None
def forward_chunk(self,
xs_pad: torch.Tensor,
ilens: torch.Tensor,