(Go: >> BACK << -|- >> HOME <<)
Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

EMO_UNK禁用和Merge VAD修复 #1940

Merged
merged 9 commits into from
Jul 22, 2024
Merged

EMO_UNK禁用和Merge VAD修复 #1940

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
48c41c6
添加富文本解码约束
May 8, 2024
ff1f1a7
special token
May 8, 2024
9c14826
bug fix
May 8, 2024
ee43244
fix
May 8, 2024
eb8ff89
增加unk score的参数
May 13, 2024
3f2d5bd
Merge branch 'main' into gcf
Jul 22, 2024
b2ae86b
emobaned
Jul 22, 2024
5ce52d6
kwargs2cfg
Jul 22, 2024
a1eedb9
merge_vad bug fix
Jul 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion funasr/auto/auto_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -337,7 +337,7 @@ def inference_with_vad(self, input, input_len=None, **cfg):
end_vad = time.time()

# FIX(gcf): concat the vad clips for sense vocie model for better aed
if kwargs.get("merge_vad", False):
if cfg.get("merge_vad", False):
for i in range(len(res)):
res[i]["value"] = merge_vad(
res[i]["value"], kwargs.get("merge_length_s", 15) * 1000
Expand Down
5 changes: 4 additions & 1 deletion funasr/models/sense_voice/model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -644,6 +644,7 @@ def __init__(
self.embed = torch.nn.Embedding(
7 + len(self.lid_dict) + len(self.textnorm_dict), input_size
)
self.emo_dict = {"unk": 25009, "happy": 25001, "sad": 25002, "angry": 25003, "neutral": 25004}

self.criterion_att = LabelSmoothingLoss(
size=self.vocab_size,
Expand Down Expand Up @@ -870,7 +871,9 @@ def inference(

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

if kwargs.get("ban_emo_unk", False):
ctc_logits[:, :, self.emo_dict["unk"]] = -float("inf")

results = []
b, n, d = encoder_out.size()
if isinstance(key[0], (list, tuple)):
Expand Down
20 changes: 12 additions & 8 deletions funasr/utils/vad_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,8 +32,10 @@ def slice_padding_audio_samples(speech, speech_lengths, vad_segments):
return speech_list, speech_lengths_list


def merge_vad(vad_result, max_length=15000):
def merge_vad(vad_result, max_length=15000, min_length=0):
new_result = []
if len(vad_result) <= 1:
return vad_result
time_step = [t[0] for t in vad_result] + [t[1] for t in vad_result]
time_step = sorted(list(set(time_step)))
if len(time_step) == 0:
Expand All @@ -43,13 +45,15 @@ def merge_vad(vad_result, max_length=15000):
time = time_step[i]
if time_step[i + 1] - bg < max_length:
continue
if time - bg < max_length * 1.5:
if time - bg > min_length:
new_result.append([bg, time])
else:
split_num = int(time - bg) // max_length + 1
spl_l = int(time - bg) // split_num
for j in range(split_num):
new_result.append([bg + j * spl_l, bg + (j + 1) * spl_l])
# if time - bg < max_length * 1.5:
# new_result.append([bg, time])
# else:
# split_num = int(time - bg) // max_length + 1
# spl_l = int(time - bg) // split_num
# for j in range(split_num):
# new_result.append([bg + j * spl_l, bg + (j + 1) * spl_l])
bg = time
new_result.append([bg, time_step[-1]])
return new_result
return new_result