Improve dataset card: Correct license, add paper and code links

This PR updates the Balalaika dataset card to improve its accuracy and discoverability.

Specifically, it:
- Corrects the `license` metadata tag from `cc-by-nc-sa-4.0` to `cc-by-nc-nd-4.0` to accurately reflect the dataset's licensing terms as stated in the dataset card content.
- Adds explicit links to the paper and the GitHub repository immediately after the abstract, making these crucial resources more prominent and easily accessible.
- Adds `russian` to the `tags` metadata for better categorization.

README.md

@@ -1,16 +1,21 @@
 ---
-license: cc-by-nc-sa-4.0
 language:
 - ru
+license: cc-by-nc-nd-4.0
 task_categories:
 - text-to-speech
 pretty_name: Balalaika
+tags:
+- russian
 ---
 
 # A Data-Centric Framework for Addressing Phonetic and Prosodic Challenges in Russian Speech Generative Models
 
 Russian speech synthesis presents distinctive challenges, including vowel reduction, consonant devoicing, variable stress patterns, homograph ambiguity, and unnatural intonation. This paper introduces Balalaika, a novel dataset comprising more than 2,000 hours of studio-quality Russian speech with comprehensive textual annotations, including punctuation and stress markings. Experimental results show that models trained on Balalaika significantly outperform those trained on existing datasets in both speech synthesis and enhancement tasks.
 
+Paper: [A Data-Centric Framework for Addressing Phonetic and Prosodic Challenges in Russian Speech Generative Models](https://huggingface.co/papers/2507.13563)
+Code: https://github.com/mtuciru/balalaika
+
 ---
 
 ## Quick Start 👟