From 3a93a88a2c6065375cc6a1a63dc74106b2ede4ee Mon Sep 17 00:00:00 2001 From: genisplaja Date: Wed, 25 Oct 2023 12:30:11 +0200 Subject: [PATCH] add missing link --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index fc8e2ee..d64ab54 100644 --- a/README.md +++ b/README.md @@ -25,7 +25,7 @@ Additional arguments can be passed to use a different model (`--model-name`), mo To train your own model, you should first prepare the data. See [how we do process Saraga](./dataset/prepare_saraga.py) before the training process detailed in the paper. The key idea is to have the chunked and aligned audio samples of the dataset with a naming like: `_.wav`, where `` corresponds to `mixture` and `vocals`. -Then, run model training in [train.py](./train.py). Checkpoints will be stored every X training steps, X is defined by user in the (./config.py) file. +Then, run model training in [train.py](./train.py). Checkpoints will be stored every X training steps, X is defined by user in the [config.py](./config.py) file. To start to train from previous checkpoint, `--load-step` is available.