Merge pull request #80 from alibaba-damo-academy/dev

add speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline &…

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_de.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_de.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py

@@ -30,6 +30,6 @@ if __name__ == '__main__':
     params["dataset_type"] = "small"
     params["max_epoch"] = 50
     params["lr"] = 0.00005
-    params["model"] = "damo/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-online"
+    params["model"] = "damo/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline"
     params["model_revision"] = None
     modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py

@@ -6,7 +6,7 @@ if __name__ == "__main__":
     output_dir = "./results"
     inference_pipline = pipeline(
         task=Tasks.auto_speech_recognition,
-        model="damo/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-online",
+        model="damo/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline",
         output_dir=output_dir,
     )
     rec_result = inference_pipline(audio_in=audio_in)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py

@@ -30,6 +30,6 @@ if __name__ == '__main__':
     params["dataset_type"] = "small"
     params["max_epoch"] = 50
     params["lr"] = 0.00005
-    params["model"] = "damo/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-online"
+    params["model"] = "damo/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline"
     params["model_revision"] = None
     modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py

@@ -6,7 +6,7 @@ if __name__ == "__main__":
     output_dir = "./results"
     inference_pipline = pipeline(
         task=Tasks.auto_speech_recognition,
-        model="damo/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-online",
+        model="damo/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline",
         output_dir=output_dir,
     )
     rec_result = inference_pipline(audio_in=audio_in)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_fa.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_fa.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_fr.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_fr.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py

@@ -30,6 +30,6 @@ if __name__ == '__main__':
     params["dataset_type"] = "small"
     params["max_epoch"] = 50
     params["lr"] = 0.00005
-    params["model"] = "damo/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-online"
+    params["model"] = "damo/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline"
     params["model_revision"] = None
     modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py

@@ -6,7 +6,7 @@ if __name__ == "__main__":
     output_dir = "./results"
     inference_pipline = pipeline(
         task=Tasks.auto_speech_recognition,
-        model="damo/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-online",
+        model="damo/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline",
         output_dir=output_dir,
     )
     rec_result = inference_pipline(audio_in=audio_in)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py

@@ -30,6 +30,6 @@ if __name__ == '__main__':
     params["dataset_type"] = "small"
     params["max_epoch"] = 50
     params["lr"] = 0.00005
-    params["model"] = "damo/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-online"
+    params["model"] = "damo/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline"
     params["model_revision"] = None
     modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py

@@ -6,7 +6,7 @@ if __name__ == "__main__":
     output_dir = "./results"
     inference_pipline = pipeline(
         task=Tasks.auto_speech_recognition,
-        model="damo/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-online",
+        model="damo/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline",
         output_dir=output_dir,
     )
     rec_result = inference_pipline(audio_in=audio_in)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_vi.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/finetune.py

@@ -0,0 +1,35 @@
+import os
+from modelscope.metainfo import Trainers
+from modelscope.trainers import build_trainer
+from funasr.datasets.ms_dataset import MsDataset
+
+
+def modelscope_finetune(params):
+    if not os.path.exists(params["output_dir"]):
+        os.makedirs(params["output_dir"], exist_ok=True)
+    # dataset split ["train", "validation"]
+    ds_dict = MsDataset.load(params["data_dir"])
+    kwargs = dict(
+        model=params["model"],
+        model_revision=params["model_revision"],
+        data_dir=ds_dict,
+        dataset_type=params["dataset_type"],
+        work_dir=params["output_dir"],
+        batch_bins=params["batch_bins"],
+        max_epoch=params["max_epoch"],
+        lr=params["lr"])
+    trainer = build_trainer(Trainers.speech_asr_trainer, default_args=kwargs)
+    trainer.train()
+
+
+if __name__ == '__main__':
+    params = {}
+    params["output_dir"] = "./checkpoint"
+    params["data_dir"] = "./data"
+    params["batch_bins"] = 2000
+    params["dataset_type"] = "small"
+    params["max_epoch"] = 50
+    params["lr"] = 0.00005
+    params["model"] = "damo/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online"
+    params["model_revision"] = None
+    modelscope_finetune(params)

egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/infer.py

@@ -0,0 +1,13 @@
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+if __name__ == "__main__":
+    audio_in = "https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_vi.wav"
+    output_dir = "./results"
+    inference_pipline = pipeline(
+        task=Tasks.auto_speech_recognition,
+        model="damo/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online",
+        output_dir=output_dir,
+    )
+    rec_result = inference_pipline(audio_in=audio_in)
+    print(rec_result)

funasr/bin/build_trainer.py

@@ -49,7 +49,8 @@ def build_trainer(modelscope_dict,
                   scheduler_conf=None,
                   specaug=None,
                   specaug_conf=None,
-                  param_dict=None):
+                  param_dict=None,
+                  **kwargs):
     mode = modelscope_dict['mode']
     args, ASRTask = parse_args(mode=mode)
     # ddp related