> Users want to learn how to use the example for their project ! Not just test standard result ! If users want to use these codes in their project, how to use our code ? For example, there are some audios that needs to be recognized, how to get the result . Please show it in README
> What are the requirements for audio format ?
> if there are some bins or exes builded by speechx(such as `recognizer_main`), how to find the enterpoint source code ? how to use these program ? What are the meanings of these parameters ? What are the limits of these parameters ? Please show it in README, not just write it in run.sh .
1. convert deepspeech2 model to ONNX, using Paddle2ONNX.
2. check paddleinference and onnxruntime output equal.
3. optimize onnx model
@ -31,7 +33,18 @@ onnxruntime 1.11.0
bash run.sh --stage 0 --stop_stage 5
```
For more details please see `run.sh`.
**For more details please see `run.sh`.**
> write more detail in REAME, what's meaning of different stage ?
## how to use in user's project
> Users want to learn how to use the example for their project ! Not just test standard result ! If users want to use these codes in their project, how to use our code ? For example, there are some audios that needs to be recognized, how to get the result . Please show it in README
> What are the requirements for audio format ?
> if there are some bins or exes builded by speechx(such as `recognizer_main`), how to find the enterpoint source code ? how to use these program ? What are the meanings of these parameters ? What are the limits of these parameters ? Please show it in README, not just write it in run.sh .
## Outputs
The optimized onnx model is `exp/model.opt.onnx`, quanted model is `$exp/model.optset11.quant.onnx`.
> Users want to learn how to use the example for their project ! Not just test standard result ! If users want to use these codes in their project, how to use our code ? For example, there are some audios that needs to be recognized, how to get the result . Please show it in README
> What are the requirements for audio format ?
> if there are some bins or exes builded by speechx(such as `recognizer_main`), how to find the enterpoint source code ? how to use these program ? What are the meanings of these parameters ? What are the limits of these parameters ? Please show it in README, not just write it in run.sh .