From 6e5f075b1d9f189dd4e5400a0a228c670aa4696e Mon Sep 17 00:00:00 2001
From: hnluo <haoneng.lhn@alibaba-inc.com>
Date: 星期四, 09 二月 2023 14:15:18 +0800
Subject: [PATCH] Merge pull request #80 from alibaba-damo-academy/dev
---
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/finetune.py | 35 +++++
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/finetune.py | 35 +++++
funasr/bin/build_trainer.py | 3
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/infer.py | 13 +
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py | 2
egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py | 2
25 files changed, 394 insertions(+), 9 deletions(-)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/finetune.py
new file mode 100644
index 0000000..68d7ba8
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-offline/infer.py
new file mode 100644
index 0000000..d23c7f4
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/finetune.py
new file mode 100644
index 0000000..462f266
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-de-16k-common-vocab3690-tensorflow1-online/infer.py
new file mode 100644
index 0000000..d7840c2
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py
index 6998f0c..397b7ff 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/finetune.py
@@ -30,6 +30,6 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py
index 201f794..0cfe939 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-en-16k-common-vocab1080-tensorflow1-offline/infer.py
@@ -6,7 +6,7 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py
index 79ef4b6..3846ff6 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/finetune.py
@@ -30,6 +30,6 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py
index cfd9e9d..6c416b2 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-es-16k-common-vocab3445-tensorflow1-offline/infer.py
@@ -6,7 +6,7 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/finetune.py
new file mode 100644
index 0000000..1aef9c6
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-offline/infer.py
new file mode 100644
index 0000000..85ddeee
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/finetune.py
new file mode 100644
index 0000000..3bdf1cc
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/infer.py
new file mode 100644
index 0000000..960c393
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/finetune.py
new file mode 100644
index 0000000..4746cc2
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-offline/infer.py
new file mode 100644
index 0000000..e541f27
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/finetune.py
new file mode 100644
index 0000000..75901db
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-fr-16k-common-vocab3472-tensorflow1-online/infer.py
new file mode 100644
index 0000000..f871665
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py
index 249aa28..fd9c442 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/finetune.py
@@ -30,6 +30,6 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py
index 5fbfdbb..7aba7ee 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ko-16k-common-vocab6400-tensorflow1-offline/infer.py
@@ -6,7 +6,7 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py
index e8d61d9..432266d 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/finetune.py
@@ -30,6 +30,6 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py
index 1051b1f..8b96ffb 100644
--- a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py
+++ b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-ru-16k-common-vocab1664-tensorflow1-offline/infer.py
@@ -6,7 +6,7 @@
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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/finetune.py
new file mode 100644
index 0000000..3a90ed2
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-offline/infer.py
new file mode 100644
index 0000000..b7fcd59
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/finetune.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/finetune.py
new file mode 100644
index 0000000..5be2585
--- /dev/null
+++ b/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)
diff --git a/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/infer.py b/egs_modelscope/asr/uniasr/speech_UniASR_asr_2pass-vi-16k-common-vocab1001-pytorch-online/infer.py
new file mode 100644
index 0000000..869082b
--- /dev/null
+++ b/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)
diff --git a/funasr/bin/build_trainer.py b/funasr/bin/build_trainer.py
index 5ef736a..bb1d7a7 100644
--- a/funasr/bin/build_trainer.py
+++ b/funasr/bin/build_trainer.py
@@ -49,7 +49,8 @@
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
--
Gitblit v1.9.1