Update README.md

README.md

@@ -8,7 +8,7 @@ license: apache-2.0
 
 ## News
 
-- [07/28] We released [**model checkpoint**](https://huggingface.co/JoshuaW1997/FUTGA) and **training/inference code** based on [**SALMONN-7B**](https://huggingface.co/tsinghua-ee/SALMONN) backbone!
+- [07/28] We released [**FUTGA-7B**](https://huggingface.co/JoshuaW1997/FUTGA) and **training/inference code** based on [**SALMONN-7B**](https://huggingface.co/tsinghua-ee/SALMONN) backbone!
 
 ## Overview
 FUTGA is an audio LLM with fine-grained music understanding, learning from generative augmentation with temporal compositions. By leveraging existing music caption datasets and large language models (LLMs), we synthesize detailed music captions with structural descriptions and time boundaries for full-length songs. This synthetic dataset enables FUTGA to identify temporal changes at key transition points, their musical functions, and generate dense captions for full-length songs.
@@ -27,3 +27,7 @@ We build FUTGA based on [**SALMONN**](https://huggingface.co/tsinghua-ee/SALMONN
 3. [vicuna 7B v1.5](https://huggingface.co/lmsys/vicuna-7b-v1.5/tree/main) to ```vicuna_path```,
 4. [FUTGA-7b](https://huggingface.co/JoshuaW1997/FUTGA/blob/main/salomnn_7b.bin) to ```ckpt_path```.
 
+
+## Datasets
+We generate dense captions for full-length songs in [MusicCaps](https://huggingface.co/JoshuaW1997/FUTGA/tree/main/Data-MusicCaps) and [SongDescriber](https://huggingface.co/JoshuaW1997/FUTGA/tree/main/Data-SongDescriber), 
+where ''raw captions'' are directly generated from FUTGA-7B and ''seg_captions_features'' contain automatically segmented captions with structures and textual-audio features for each segment.