基于 TensorFlow 2 的端到端语音识别框架，支持 Conformer、ContextNet、DeepSpeech2 等模型，可做字符或子词级多语言识别与流式推理。
用 5 秒语音样本克隆音色，实时合成任意文本的语音，基于 PyTorch/TensorFlow 的说话人编码与声码器流水线。
收集大量预训练模型，并已转成 TensorFlow、PyTorch、ONNX、OpenVINO、TFLite、CoreML 等格式，省去自己跨框架转换的麻烦。
把 ONNX 模型一键转成 TFLite、TensorFlow、PyTorch 原生代码或 TorchScript，省去跨框架部署时手写转换脚本的麻烦。
用 TensorFlow/Keras 串联从文本预处理到 BERT、GPT、LLM 下游任务的实战教程，覆盖 NER、LoRA、DPO 等，适合系统学习 NLP 建模流程。
…sult ([ a0480fb ](https://github.com/TensorSpeech/TensorFlowASR/commit/a0480fbc563ec4518269568d2993bad66cd768b8)) - Device scope ([ 9379765 ](https://github.com/TensorSpe
… requirements.txt. Relax absl-py requirements for tensorflow model optimization. Mark compatibility with numpy v2.
…ws tests. by @fatcat-z in https://github.com/onnx/tensorflow-onnx/pull/2294 fix: make onnx node with attr() supports empty strings by @mys007 in https://github.com/onnx/t
…ws tests. by @fatcat-z in https://github.com/onnx/tensorflow-onnx/pull/2294 fix: make onnx node with attr() supports empty strings by @mys007 in https://github.com/onnx/t
Full Changelog : https://github.com/TensorSpeech/TensorFlowASR/compare/v2.0.1...v2.1.0
Full Changelog : https://github.com/TensorSpeech/TensorFlowASR/compare/v2.0.0...v2.0.1