add training.md skeleton

README.md

-ATTENTION: don't forget to add group_by_length in configs.
-
 # Parler-TTS
 
 [[Paper we reproduce]](https://arxiv.org/abs/2402.01912)
@@ -65,7 +63,6 @@ Then, run:
 python helpers/gradio_demo/app.py
 ```
 
-
 ## Acknowledgements
 
 This library builds on top of a number of open-source giants, to whom we'd like to extend our warmest thanks for providing these tools!
@@ -91,6 +88,8 @@ Namely, we're looking at ways to improve both quality and speed:
 - Optimization:
     - Compilation and static cache
     - Support to FA2 and SDPA
+- Evaluation:
+    - Add more evaluation metrics
 
 ## Citation
 ```

training/TRAINING.md

@@ -2,9 +2,22 @@
 
 This sub-folder contains all the information to train or finetune you own Parler-TTS model.
 
+## Getting started
 
+### Requirements
 
+### Initializing models
 
+### Datasets
+
+## Training
+
+
+## Discussions and tips
+
+
+
+ATTENTION: don't forget to add group_by_length in configs.
 
 
 # Init model
@@ -16,9 +29,3 @@ text_model = "google-t5/t5-small"
 encodec_version = "facebook/encodec_24khz"
 text_model = "google/flan-t5-base"
 encodec_version = "ylacombe/dac_44khZ_8kbps"
-
-
-
-## TODOs
-- [ ] Add PEFT compatibility to do Lora fine-tuning.
-- [ ] Enrich dataset with accent classifier
\ No newline at end of file