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
auto frontend
游雁870fa4a37a19
diff --git a/funasr/datasets/audio_datasets/samplers.py b/funasr/datasets/audio_datasets/samplers.pyindex 7195addc1..27fbdac25 100644--- a/funasr/datasets/audio_datasets/samplers.py+++ b/funasr/datasets/audio_datasets/samplers.py@@ -417,7 +417,7 @@ class CustomDistributedBufferDynamicBatchSampler(DistributedSampler):rank_batches[i % self.num_replicas].append(batch)# Assign all batches for the current rank directly- final_batches = rank_batches[self.rank][self.start_step :]+ final_batches = rank_batches[self.rank] # [self.start_step :]self.batch_num = len(final_batches)logging.info(