语言:

• 英文 标签:
• 开放式视觉推理与神经网络优化工具包

MIT/ast微调语音指令识别模型v2

此模型为MIT/ast-finetuned-speech-commands-v2转换而成的OpenVINO版本，用于加速推理运算。

以下为调用该模型进行推理的示例代码：

from optimum.intel.openvino import OVModelForAudioClassification
from transformers import AutoFeatureExtractor, pipeline

# model_id可设置为本地目录路径或HuggingFace平台上的模型名称
model_id = "helenai/MIT-ast-finetuned-speech-commands-v2-ov"
feature_extractor = AutoFeatureExtractor.from_pretrained(model_id)
model = OVModelForAudioClassification.from_pretrained(model_id)
pipe = pipeline("audio-classification", model=model, feature_extractor=feature_extractor)
result = pipe("https://datasets-server.huggingface.co/assets/speech_commands/--/v0.01/test/38/audio/audio.mp3")
print(result)