Update README.md

README.md

@@ -111,7 +111,7 @@ CLaMP 2 is a music information retrieval model compatible with 101 languages, de
 
 ### Links
 - [CLaMP 2 Code](https://github.com/sanderwood/clamp2)
-- [CLaMP 2 Paper](https://arxiv.org/)
+- [CLaMP 2 Paper](https://arxiv.org/pdf/2410.13267)
 - [CLaMP 2 Model Weights](https://huggingface.co/sander-wood/clamp2/blob/main/weights_clamp2_h_size_768_lr_5e-05_batch_128_scale_1_t_length_128_t_model_FacebookAI_xlm-roberta-base_t_dropout_True_m3_True.pth)
 - [M3 Model Weights](https://huggingface.co/sander-wood/clamp2/blob/main/weights_m3_p_size_64_p_length_512_t_layers_3_p_layers_12_h_size_768_lr_0.0001_batch_16_mask_0.45.pth)
 
@@ -172,6 +172,7 @@ conda activate clamp2
     ]
   }
   ```
+   The filepaths field contains relative paths starting from the shortest common root directory (e.g., abc/ or mtf/). This ensures that only the minimal shared part of the path is included, and each file is represented with a concise relative path from this root.
 
    **Output Example**: The output will be a JSON file containing the structured summary in both English and a selected non-English language. Here’s an example of the expected output:
 
@@ -197,6 +198,19 @@ conda activate clamp2
 }
 ```
 
+After generating the individual JSON files:
+
+1. Merge all JSON files into a single JSONL file.
+ 
+2. Place the merged JSONL file and the shortest common root directories (e.g., abc/ and/or mtf/) in the same folder, structured like this:
+
+```
+/your-target-folder/
+├── abc/
+├── mtf/
+├── merged_output.jsonl
+```
+
 ### Training and Feature Extraction
 2. **Training Models**: If you want to train CLaMP 2 or M3 models, check the scripts in the `code/` folder.
    - Modify the `config.py` files to set your training hyperparameters and paths.
@@ -217,12 +231,11 @@ Benchmark datasets related to the experiments conducted with CLaMP 2 and M3, inc
 If you use CLaMP 2 or M3 in your research, please cite the following paper:
 
 ```bibtex
-@inproceedings{clamp2,
-  title={CLaMP 2: Multimodal Music Information Retrieval Across 101 Languages Using Large Language Models},
-  author={Author Name and Coauthor Name},
-  booktitle={Proceedings of the Conference on Music Information Retrieval},
-  year={2024},
-  publisher={Publisher Name},
-  address={Conference Location},
-  url={https://placeholder.url}
-}
+@misc{wu2024clamp2multimodalmusic,
+      title={CLaMP 2: Multimodal Music Information Retrieval Across 101 Languages Using Large Language Models}, 
+      author={Shangda Wu and Yashan Wang and Ruibin Yuan and Zhancheng Guo and Xu Tan and Ge Zhang and Monan Zhou and Jing Chen and Xuefeng Mu and Yuejie Gao and Yuanliang Dong and Jiafeng Liu and Xiaobing Li and Feng Yu and Maosong Sun},
+      year={2024},
+      eprint={2410.13267},
+      archivePrefix={arXiv},
+      primaryClass={cs.SD},
+      url={https://arxiv.org/abs/2410.13267},