Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

Dev gzf (#1424)

zhifu gao8ab670afc3b7
diff --git a/README.md b/README.md
index d436d5ebf..507306f37 100644
--- a/README.md
+++ b/README.md
@@ -27,6 +27,7 @@
<a name="whats-new"></a>
## What's new:
+- 2024/03/05:Added support for the Whisper-large-v3 model, a multitasking model that can perform multilingual speech recognition, speech translation, and language identification. It can be downloaded from the[modelscope](https://www.modelscope.cn/models/iic/Whisper-large-v3/summary), and [openai](https://github.com/alibaba-damo-academy/FunASR/tree/main/examples/industrial_data_pretraining/whisper).
- 2024/03/03: Offline File Transcription Service 4.4, Offline File Transcription Service of English 1.5,Real-time Transcription Service 1.9 released,Docker image supports ARM64 platform;([docs](runtime/readme.md))
- 2024/01/30:funasr-1.0 has been released ([docs](https://github.com/alibaba-damo-academy/FunASR/discussions/1319))
- 2024/01/30:emotion recognition models are new supported. [model link](https://www.modelscope.cn/models/iic/emotion2vec_base_finetuned/summary), modified from [repo](https://github.com/ddlBoJack/emotion2vec).
@@ -67,20 +68,21 @@ pip3 install -U modelscope
## Model Zoo
FunASR has open-sourced a large number of pre-trained models on industrial data. You are free to use, copy, modify, and share FunASR models under the [Model License Agreement](./MODEL_LICENSE). Below are some representative models, for more models please refer to the [Model Zoo]().
-(Note: ⭐ represents the ModelScope model zoo link, 🤗 represents the Huggingface model zoo link)
-
-
-| Model Name | Task Details | Training Data | Parameters |
-|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:--------------------------------------------------:|:--------------------------------:|:----------:|
-| paraformer-zh <br> ([⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) [🤗](https://huggingface.co/funasr/paraformer-tp) ) | speech recognition, with timestamps, non-streaming | 60000 hours, Mandarin | 220M |
-| <nobr>paraformer-zh-streaming <br> ( [⭐](https://modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/summary) [🤗](https://huggingface.co/funasr/paraformer-zh-streaming) )</nobr> | speech recognition, streaming | 60000 hours, Mandarin | 220M |
-| paraformer-en <br> ( [⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-en-16k-common-vocab10020/summary) [🤗](https://huggingface.co/funasr/paraformer-en) ) | speech recognition, with timestamps, non-streaming | 50000 hours, English | 220M |
-| conformer-en <br> ( [⭐](https://modelscope.cn/models/damo/speech_conformer_asr-en-16k-vocab4199-pytorch/summary) [🤗](https://huggingface.co/funasr/conformer-en) ) | speech recognition, non-streaming | 50000 hours, English | 220M |
-| ct-punc <br> ( [⭐](https://modelscope.cn/models/damo/punc_ct-transformer_cn-en-common-vocab471067-large/summary) [🤗](https://huggingface.co/funasr/ct-punc) ) | punctuation restoration | 100M, Mandarin and English | 1.1G |
-| fsmn-vad <br> ( [⭐](https://modelscope.cn/models/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/summary) [🤗](https://huggingface.co/funasr/fsmn-vad) ) | voice activity detection | 5000 hours, Mandarin and English | 0.4M |
-| fa-zh <br> ( [⭐](https://modelscope.cn/models/damo/speech_timestamp_prediction-v1-16k-offline/summary) [🤗](https://huggingface.co/funasr/fa-zh) ) | timestamp prediction | 5000 hours, Mandarin | 38M |
-| cam++ <br> ( [⭐](https://modelscope.cn/models/iic/speech_campplus_sv_zh-cn_16k-common/summary) [🤗](https://huggingface.co/funasr/campplus) ) | speaker verification/diarization | 5000 hours | 7.2M |
-| whisper-large-v2 <br> ([⭐](https://www.modelscope.cn/models/iic/speech_whisper-large_asr_multilingual/summary) [🤗]() ) | speech recognition, with timestamps, non-streaming | multilingual | 1G |
+(Note: ⭐ represents the ModelScope model zoo, 🤗 represents the Huggingface model zoo, 🍀 represents the OpenAI model zoo)
+
+
+| Model Name | Task Details | Training Data | Parameters |
+|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:-----------------------------------------------------:|:--------------------------------:|:----------:|
+| paraformer-zh <br> ([⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) [🤗](https://huggingface.co/funasr/paraformer-tp) ) | speech recognition, with timestamps, non-streaming | 60000 hours, Mandarin | 220M |
+| <nobr>paraformer-zh-streaming <br> ( [⭐](https://modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/summary) [🤗](https://huggingface.co/funasr/paraformer-zh-streaming) )</nobr> | speech recognition, streaming | 60000 hours, Mandarin | 220M |
+| paraformer-en <br> ( [⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-en-16k-common-vocab10020/summary) [🤗](https://huggingface.co/funasr/paraformer-en) ) | speech recognition, without timestamps, non-streaming | 50000 hours, English | 220M |
+| conformer-en <br> ( [⭐](https://modelscope.cn/models/damo/speech_conformer_asr-en-16k-vocab4199-pytorch/summary) [🤗](https://huggingface.co/funasr/conformer-en) ) | speech recognition, non-streaming | 50000 hours, English | 220M |
+| ct-punc <br> ( [⭐](https://modelscope.cn/models/damo/punc_ct-transformer_cn-en-common-vocab471067-large/summary) [🤗](https://huggingface.co/funasr/ct-punc) ) | punctuation restoration | 100M, Mandarin and English | 1.1G |
+| fsmn-vad <br> ( [⭐](https://modelscope.cn/models/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/summary) [🤗](https://huggingface.co/funasr/fsmn-vad) ) | voice activity detection | 5000 hours, Mandarin and English | 0.4M |
+| fa-zh <br> ( [⭐](https://modelscope.cn/models/damo/speech_timestamp_prediction-v1-16k-offline/summary) [🤗](https://huggingface.co/funasr/fa-zh) ) | timestamp prediction | 5000 hours, Mandarin | 38M |
+| cam++ <br> ( [⭐](https://modelscope.cn/models/iic/speech_campplus_sv_zh-cn_16k-common/summary) [🤗](https://huggingface.co/funasr/campplus) ) | speaker verification/diarization | 5000 hours | 7.2M |
+| Whisper-large-v2 <br> ([⭐](https://www.modelscope.cn/models/iic/speech_whisper-large_asr_multilingual/summary) [🍀](https://github.com/openai/whisper) ) | speech recognition, with timestamps, non-streaming | multilingual | 1.5G |
+| Whisper-large-v3 <br> ([⭐](https://www.modelscope.cn/models/iic/Whisper-large-v3/summary) [🍀](https://github.com/openai/whisper) ) | speech recognition, with timestamps, non-streaming | multilingual | 1.5G |
diff --git a/README_zh.md b/README_zh.md
index 8a34c82be..a6e1ac752 100644
--- a/README_zh.md
+++ b/README_zh.md
@@ -29,6 +29,7 @@ FunASR希望在语音识别的学术研究和工业应用之间架起一座桥
<a name="最新动态"></a>
## 最新动态
+- 2024/03/05:新增加Whisper-large-v3模型支持,多语言语音识别/翻译/语种识别,支持从[modelscope](https://www.modelscope.cn/models/iic/Whisper-large-v3/summary)仓库下载,也支持从[openai](https://github.com/alibaba-damo-academy/FunASR/tree/main/examples/industrial_data_pretraining/whisper)仓库下载模型。
- 2024/03/03: 中文离线文件转写服务 4.4、英文离线文件转写服务 1.5、中文实时语音听写服务 1.9 发布,docker镜像支持arm64平台;详细信息参阅([部署文档](runtime/readme_cn.md))
- 2024/01/30:funasr-1.0发布,更新说明[文档](https://github.com/alibaba-damo-academy/FunASR/discussions/1319)
- 2024/01/30:新增加情感识别 [模型链接](https://www.modelscope.cn/models/iic/emotion2vec_base_finetuned/summary),原始模型 [repo](https://github.com/ddlBoJack/emotion2vec).
@@ -69,20 +70,21 @@ pip3 install -U modelscope
FunASR开源了大量在工业数据上预训练模型,您可以在[模型许可协议](./MODEL_LICENSE)下自由使用、复制、修改和分享FunASR模型,下面列举代表性的模型,更多模型请参考[模型仓库]()。
-(注:⭐ 表示ModelScope模型仓库链接,🤗 表示Huggingface模型仓库链接)
-
-
-| 模型名字 | 任务详情 | 训练数据 | 参数量 |
-|:------------------------------------------------------------------------------------------------------------------------------------------------------------:|:---------------:|:------------:|:----:|
-| paraformer-zh <br> ([⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) [🤗](https://huggingface.co/funasr/paraformer-tp) ) | 语音识别,带时间戳输出,非实时 | 60000小时,中文 | 220M |
-| paraformer-zh-streaming <br> ( [⭐](https://modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/summary) [🤗](https://huggingface.co/funasr/paraformer-zh-streaming) ) | 语音识别,实时 | 60000小时,中文 | 220M |
-| paraformer-en <br> ( [⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-en-16k-common-vocab10020/summary) [🤗](https://huggingface.co/funasr/paraformer-en) ) | 语音识别,非实时 | 50000小时,英文 | 220M |
-| conformer-en <br> ( [⭐](https://modelscope.cn/models/damo/speech_conformer_asr-en-16k-vocab4199-pytorch/summary) [🤗](https://huggingface.co/funasr/conformer-en) ) | 语音识别,非实时 | 50000小时,英文 | 220M |
-| ct-punc <br> ( [⭐](https://modelscope.cn/models/damo/punc_ct-transformer_cn-en-common-vocab471067-large/summary) [🤗](https://huggingface.co/funasr/ct-punc) ) | 标点恢复 | 100M,中文与英文 | 1.1G |
-| fsmn-vad <br> ( [⭐](https://modelscope.cn/models/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/summary) [🤗](https://huggingface.co/funasr/fsmn-vad) ) | 语音端点检测,实时 | 5000小时,中文与英文 | 0.4M |
-| fa-zh <br> ( [⭐](https://modelscope.cn/models/damo/speech_timestamp_prediction-v1-16k-offline/summary) [🤗](https://huggingface.co/funasr/fa-zh) ) | 字级别时间戳预测 | 50000小时,中文 | 38M |
-| cam++ <br> ( [⭐](https://modelscope.cn/models/iic/speech_campplus_sv_zh-cn_16k-common/summary) [🤗](https://huggingface.co/funasr/campplus) ) | 说话人确认/分割 | 5000小时 | 7.2M |
-| whisper-large-v2 <br> ([⭐](https://www.modelscope.cn/models/iic/speech_whisper-large_asr_multilingual/summary) [🤗]() ) | 语音识别,带时间戳输出,非实时 | 多语言 | 1G |
+(注:⭐ 表示ModelScope模型仓库,🤗 表示Huggingface模型仓库,🍀表示OpenAI模型仓库)
+
+
+| 模型名字 | 任务详情 | 训练数据 | 参数量 |
+|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:---------------:|:------------:|:----:|
+| paraformer-zh <br> ([⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/summary) [🤗](https://huggingface.co/funasr/paraformer-tp) ) | 语音识别,带时间戳输出,非实时 | 60000小时,中文 | 220M |
+| paraformer-zh-streaming <br> ( [⭐](https://modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/summary) [🤗](https://huggingface.co/funasr/paraformer-zh-streaming) ) | 语音识别,实时 | 60000小时,中文 | 220M |
+| paraformer-en <br> ( [⭐](https://www.modelscope.cn/models/damo/speech_paraformer-large-vad-punc_asr_nat-en-16k-common-vocab10020/summary) [🤗](https://huggingface.co/funasr/paraformer-en) ) | 语音识别,非实时 | 50000小时,英文 | 220M |
+| conformer-en <br> ( [⭐](https://modelscope.cn/models/damo/speech_conformer_asr-en-16k-vocab4199-pytorch/summary) [🤗](https://huggingface.co/funasr/conformer-en) ) | 语音识别,非实时 | 50000小时,英文 | 220M |
+| ct-punc <br> ( [⭐](https://modelscope.cn/models/damo/punc_ct-transformer_cn-en-common-vocab471067-large/summary) [🤗](https://huggingface.co/funasr/ct-punc) ) | 标点恢复 | 100M,中文与英文 | 1.1G |
+| fsmn-vad <br> ( [⭐](https://modelscope.cn/models/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/summary) [🤗](https://huggingface.co/funasr/fsmn-vad) ) | 语音端点检测,实时 | 5000小时,中文与英文 | 0.4M |
+| fa-zh <br> ( [⭐](https://modelscope.cn/models/damo/speech_timestamp_prediction-v1-16k-offline/summary) [🤗](https://huggingface.co/funasr/fa-zh) ) | 字级别时间戳预测 | 50000小时,中文 | 38M |
+| cam++ <br> ( [⭐](https://modelscope.cn/models/iic/speech_campplus_sv_zh-cn_16k-common/summary) [🤗](https://huggingface.co/funasr/campplus) ) | 说话人确认/分割 | 5000小时 | 7.2M |
+| Whisper-large-v2 <br> ([⭐](https://www.modelscope.cn/models/iic/speech_whisper-large_asr_multilingual/summary) [🍀](https://github.com/openai/whisper) ) | 语音识别,带时间戳输出,非实时 | 多语言 | 1G |
+| Whisper-large-v3 <br> ([⭐](https://www.modelscope.cn/models/iic/Whisper-large-v3/summary) [🍀](https://github.com/openai/whisper) ) | 语音识别,带时间戳输出,非实时 | 多语言 | 1G |
<a name="快速开始"></a>
diff --git a/examples/industrial_data_pretraining/whisper/demo.py b/examples/industrial_data_pretraining/whisper/demo.py
index f010ea263..db8d92c7f 100644
--- a/examples/industrial_data_pretraining/whisper/demo.py
+++ b/examples/industrial_data_pretraining/whisper/demo.py
@@ -5,7 +5,7 @@
from funasr import AutoModel
-model = AutoModel(model="iic/speech_whisper-large_asr_multilingual",
+model = AutoModel(model="iic/Whisper-large-v3",
model_revision="v2.0.4",
)
diff --git a/examples/industrial_data_pretraining/whisper/demo_from_openai.py b/examples/industrial_data_pretraining/whisper/demo_from_openai.py
index 0b88a9598..046e9c677 100644
--- a/examples/industrial_data_pretraining/whisper/demo_from_openai.py
+++ b/examples/industrial_data_pretraining/whisper/demo_from_openai.py
@@ -7,8 +7,8 @@ from funasr import AutoModel
# model = AutoModel(model="Whisper-small", hub="openai")
# model = AutoModel(model="Whisper-medium", hub="openai")
-model = AutoModel(model="Whisper-large-v2", hub="openai")
-# model = AutoModel(model="Whisper-large-v3", hub="openai")
+# model = AutoModel(model="Whisper-large-v2", hub="openai")
+model = AutoModel(model="Whisper-large-v3", hub="openai")
res = model.generate(
language=None,
diff --git a/examples/industrial_data_pretraining/whisper/infer_from_local.sh b/examples/industrial_data_pretraining/whisper/infer_from_local.sh
index 14ccbb663..885dfc6c2 100644
--- a/examples/industrial_data_pretraining/whisper/infer_from_local.sh
+++ b/examples/industrial_data_pretraining/whisper/infer_from_local.sh
@@ -13,13 +13,19 @@ workspace=`pwd`
# download model
local_path_root=${workspace}/modelscope_models
mkdir -p ${local_path_root}
-local_path=${local_path_root}/speech_whisper-large_asr_multilingual
-git clone https://www.modelscope.cn/iic/speech_whisper-large_asr_multilingual.git ${local_path}
+#Whisper-large-v2
+#local_path=${local_path_root}/speech_whisper-large_asr_multilingual
+#git clone https://www.modelscope.cn/iic/speech_whisper-large_asr_multilingual.git ${local_path}
+#init_param="${local_path}/large-v2.pt"
+#Whisper-large-v3
+local_path=${local_path_root}/Whisper-large-v3
+git clone https://www.modelscope.cn/iic/Whisper-large-v3.git ${local_path}
+init_param="${local_path}/large-v3.pt"
device="cuda:0" # "cuda:0" for gpu0, "cuda:1" for gpu1, "cpu"
config="config.yaml"
-init_param="${local_path}/large-v2.pt"
+
python -m funasr.bin.inference \
--config-path "${local_path}" \
diff --git a/funasr/download/name_maps_from_hub.py b/funasr/download/name_maps_from_hub.py
index e1bc2955f..07cf6a095 100644
--- a/funasr/download/name_maps_from_hub.py
+++ b/funasr/download/name_maps_from_hub.py
@@ -8,7 +8,8 @@ name_maps_ms = {
"ct-punc-c": "damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch",
"fa-zh": "damo/speech_timestamp_prediction-v1-16k-offline",
"cam++": "damo/speech_campplus_sv_zh-cn_16k-common",
- "whisper-large-v2": "iic/speech_whisper-large_asr_multilingual",
+ "Whisper-large-v2": "iic/speech_whisper-large_asr_multilingual",
+ "Whisper-large-v3": "iic/Whisper-large-v3",
}
name_maps_hf = {
diff --git a/funasr/models/whisper/model.py b/funasr/models/whisper/model.py
index 1eac2ff4d..73d70d7a2 100644
--- a/funasr/models/whisper/model.py
+++ b/funasr/models/whisper/model.py
@@ -24,7 +24,7 @@ from funasr.register import tables
@tables.register("model_classes", "Whisper-large-v1")
@tables.register("model_classes", "Whisper-large-v2")
@tables.register("model_classes", "Whisper-large-v3")
[email protected]("model_classes", "Whisper-WhisperWarp")
[email protected]("model_classes", "WhisperWarp")
class WhisperWarp(nn.Module):
def __init__(self, *args, **kwargs):
super().__init__()
@@ -35,8 +35,8 @@ class WhisperWarp(nn.Module):
model_or_path = model_or_path.replace("Whisper-", "")
model = whisper.load_model(model_or_path)
else:
- whisper_dims = kwargs.get("whisper_dims", {})
- dims = whisper.model.ModelDimensions(**whisper_dims)
+ dims = kwargs.get("dims", {})
+ dims = whisper.model.ModelDimensions(**dims)
model = whisper.model.Whisper(dims=dims)
self.model = model
@@ -55,6 +55,13 @@ class WhisperWarp(nn.Module):
if kwargs.get("batch_size", 1) > 1:
raise NotImplementedError("batch decoding is not implemented")
+ if frontend is None and not hasattr(self, "frontend"):
+ frontend_class = tables.frontend_classes.get("WhisperFrontend")
+ frontend = frontend_class(n_mels=self.model.dims.n_mels, do_pad_trim=kwargs.get("do_pad_trim", True))
+ self.frontend = frontend
+ else:
+ frontend = frontend if frontend is not None else self.frontend
+
meta_data = {}
if isinstance(data_in, torch.Tensor) and kwargs.get("data_type", "sound") == "fbank": # fbank
speech, speech_lengths = data_in, data_lengths
@@ -65,7 +72,7 @@ class WhisperWarp(nn.Module):
else:
# extract fbank feats
time1 = time.perf_counter()
- audio_sample_list = load_audio_text_image_video(data_in, fs=frontend.fs, audio_fs=kwargs.get("fs", 16000),
+ audio_sample_list = load_audio_text_image_video(data_in, fs=frontend.fs if hasattr(frontend, "fs") else 16000, audio_fs=kwargs.get("fs", 16000),
data_type=kwargs.get("data_type", "sound"),
tokenizer=tokenizer)
time2 = time.perf_counter()