|
@@ -176,6 +176,21 @@ python finetune.py &> log.txt &
|
|
|
- `max_epoch`: number of training epoch
|
|
- `max_epoch`: number of training epoch
|
|
|
- `lr`: learning rate
|
|
- `lr`: learning rate
|
|
|
|
|
|
|
|
|
|
+- Training data formats:
|
|
|
|
|
+```sh
|
|
|
|
|
+cat ./example_data/text
|
|
|
|
|
+BAC009S0002W0122 而 对 楼 市 成 交 抑 制 作 用 最 大 的 限 购
|
|
|
|
|
+BAC009S0002W0123 也 成 为 地 方 政 府 的 眼 中 钉
|
|
|
|
|
+english_example_1 hello world
|
|
|
|
|
+english_example_2 go swim 去 游 泳
|
|
|
|
|
+
|
|
|
|
|
+cat ./example_data/wav.scp
|
|
|
|
|
+BAC009S0002W0122 /mnt/data/wav/train/S0002/BAC009S0002W0122.wav
|
|
|
|
|
+BAC009S0002W0123 /mnt/data/wav/train/S0002/BAC009S0002W0123.wav
|
|
|
|
|
+english_example_1 /mnt/data/wav/train/S0002/english_example_1.wav
|
|
|
|
|
+english_example_2 /mnt/data/wav/train/S0002/english_example_2.wav
|
|
|
|
|
+```
|
|
|
|
|
+
|
|
|
- Then you can run the pipeline to finetune with:
|
|
- Then you can run the pipeline to finetune with:
|
|
|
```shell
|
|
```shell
|
|
|
python finetune.py
|
|
python finetune.py
|
|
@@ -210,4 +225,4 @@ CUDA_VISIBLE_DEVICES=1,2 python -m torch.distributed.launch --nproc_per_node 2 f
|
|
|
--njob 64 \
|
|
--njob 64 \
|
|
|
--checkpoint_dir "./checkpoint" \
|
|
--checkpoint_dir "./checkpoint" \
|
|
|
--checkpoint_name "valid.cer_ctc.ave.pb"
|
|
--checkpoint_name "valid.cer_ctc.ave.pb"
|
|
|
-```
|
|
|
|
|
|
|
+```
|