FreedomIntelligence

/

EchoX-3B

like 7

Follow

FreedomAI 971

This PR enhances the model card by:

• Adding pipeline_tag: audio-to-audio to ensure discoverability for models processing audio input to audio output. You can find models using this pipeline tag at https://huggingface.co/models?pipeline_tag=audio-to-audio.
• Specifying library_name: transformers, which indicates compatibility with the Hugging Face Transformers library, enabling the automated display of a "how to use" widget. The config.json file confirms transformers_version: 4.48.1 and architectures: ["ACLlamaForCausalLM"].
• Including a direct link to the project page: https://freedomintelligence.github.io/EchoX/.
• Adding a concrete Python sample usage code snippet from the official GitHub repository to demonstrate how to run inference.
• Correcting the training data duration in the "Model Description" from 10k hours to 6k hours for consistency with other sections and the paper's details.

Please review and merge if these improvements align with the repository's goals.