Update TRAINING.md

hkchengrex · Jan 14, 2025 · e8a0603 · e8a0603
1 parent 34bf089
commit e8a0603
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/docs/TRAINING.md b/docs/TRAINING.md
@@ -18,7 +18,9 @@ The current training script does not support `_v2` training.
 
 ## Prerequisites
 
-Install [av-benchmark](https://github.com/hkchengrex/av-benchmark). We use this library to automatically evaluate on the validation set during training, and on test set after training.
+Install [av-benchmark](https://github.com/hkchengrex/av-benchmark). We use this library to automatically evaluate on the validation set during training, and on the test set after training.
+Extract features for evaluation using [av-benchmark](https://github.com/hkchengrex/av-benchmark) for the validation and test set as a [validation cache](https://github.com/hkchengrex/MMAudio/blob/34bf089fdd2e457cd5ef33be96c0e1c8a0412476/config/data/base.yaml#L38) and a [test cache](https://github.com/hkchengrex/MMAudio/blob/34bf089fdd2e457cd5ef33be96c0e1c8a0412476/config/data/base.yaml#L31). You can also download the precomputed evaluation cache [here](https://huggingface.co/datasets/hkchengrex/MMAudio-precomputed-results/tree/main).
+
 You will also need ffmpeg for video frames extraction. Note that `torchaudio` imposes a maximum version limit (`ffmpeg<7`). You can install it as follows:
 
 ```bash