Funasr1.0

examples/industrial_data_pretraining/uniasr/demo.py

@@ -0,0 +1,29 @@
+#!/usr/bin/env python3
+# -*- encoding: utf-8 -*-
+# Copyright FunASR (https://github.com/alibaba-damo-academy/FunASR). All Rights Reserved.
+#  MIT License  (https://opensource.org/licenses/MIT)
+
+from funasr import AutoModel
+
+model = AutoModel(model="/Users/zhifu/Downloads/modelscope_models/speech_UniASR_asr_2pass-zh-cn-16k-common-vocab8358-tensorflow1-online", model_revision="v2.0.4",
+                  # vad_model="damo/speech_fsmn_vad_zh-cn-16k-common-pytorch",
+                  # vad_model_revision="v2.0.4",
+                  # punc_model="damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",
+                  # punc_model_revision="v2.0.4",
+                  )
+
+res = model.generate(input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav")
+print(res)
+
+
+''' can not use currently
+from funasr import AutoFrontend
+
+frontend = AutoFrontend(model="damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch", model_revision="v2.0.4")
+
+fbanks = frontend(input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav", batch_size=2)
+
+for batch_idx, fbank_dict in enumerate(fbanks):
+    res = model.generate(**fbank_dict)
+    print(res)
+'''

examples/industrial_data_pretraining/uniasr/infer.sh

@@ -0,0 +1,11 @@
+
+model="damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch"
+model_revision="v2.0.4"
+
+python funasr/bin/inference.py \
++model=${model} \
++model_revision=${model_revision} \
++input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav" \
++output_dir="./outputs/debug" \
++device="cpu" \
+

funasr/download/download_from_hub.py

@@ -63,7 +63,7 @@ def add_file_root_path(model_or_path: str, file_path_metas: dict, cfg = {}):
             elif isinstance(v, dict):
                 if k not in cfg:
                     cfg[k] = {}
-                return add_file_root_path(model_or_path, v, cfg[k])
+                add_file_root_path(model_or_path, v, cfg[k])
 
     return cfg
 

funasr/models/transformer/model.py

@@ -348,7 +348,7 @@ def init_beam_search(self,
         scorers["ngram"] = ngram
 
         weights = dict(
-            decoder=1.0 - kwargs.get("decoding_ctc_weight"),
+            decoder=1.0 - kwargs.get("decoding_ctc_weight", 0.0),
             ctc=kwargs.get("decoding_ctc_weight", 0.0),
             lm=kwargs.get("lm_weight", 0.0),
             ngram=kwargs.get("ngram_weight", 0.0),