Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
bug fix for punc and umap
shixian.shiae4dceecf03c
diff --git a/funasr/models/campplus/cluster_backend.py b/funasr/models/campplus/cluster_backend.pyindex 3bac0a02c..e33a14d6a 100644--- a/funasr/models/campplus/cluster_backend.py+++ b/funasr/models/campplus/cluster_backend.py@@ -119,6 +119,7 @@ class UmapHdbscan:self.metric = metricdef __call__(self, X):+ from umap.umap_ import UMAPumap_X = umap.UMAP(n_neighbors=self.n_neighbors,min_dist=0.0,@@ -156,6 +157,7 @@ class ClusterBackend(torch.nn.Module):if X.shape[0] < 20:return np.zeros(X.shape[0], dtype='int')if X.shape[0] < 2048 or k is not None:+ # unexpected corner caselabels = self.spectral_cluster(X, k)else:labels = self.umap_hdbscan_cluster(X)diff --git a/funasr/models/ct_transformer/model.py b/funasr/models/ct_transformer/model.pyindex 330d7e554..1e53aa3d4 100644--- a/funasr/models/ct_transformer/model.py+++ b/funasr/models/ct_transformer/model.py@@ -336,10 +336,11 @@ class CTTransformer(torch.nn.Module):elif new_mini_sentence[-1] != "。" and new_mini_sentence[-1] != "?" and len(new_mini_sentence[-1].encode())!=1:new_mini_sentence_out = new_mini_sentence + "。"new_mini_sentence_punc_out = new_mini_sentence_punc[:-1] + [self.sentence_end_id]+ if len(punctuations): punctuations[-1] = 2elif new_mini_sentence[-1] != "." and new_mini_sentence[-1] != "?" and len(new_mini_sentence[-1].encode())==1:new_mini_sentence_out = new_mini_sentence + "."new_mini_sentence_punc_out = new_mini_sentence_punc[:-1] + [self.sentence_end_id]-+ if len(punctuations): punctuations[-1] = 2# keep a punctuations array for punc segmentif punc_array is None:punc_array = punctuations@@ -347,6 +348,5 @@ class CTTransformer(torch.nn.Module):punc_array = torch.cat([punc_array, punctuations], dim=0)result_i = {"key": key[0], "text": new_mini_sentence_out, "punc_array": punc_array}results.append(result_i)-return results, meta_datadiff --git a/setup.py b/setup.pyindex f4439b01e..561dea2b6 100644--- a/setup.py+++ b/setup.py@@ -38,7 +38,7 @@ requirements = {# "protobuf","tqdm","hdbscan",- "umap",+ "umap_learn","jaconv","hydra-core>=1.3.2",],