Please reply to this issue if you have requests for additional models.
Our focus is on audio models right now but more than happy to consider visual models too.
We don't plan on supporting language models in the near term. Thinking more about how to best support SLMs for developers.
Model selection criteria:
- Size: We focus on models under 1B parameters. Larger models (like LLM-based architectures) don't run well on CoreML due to architecture constraints and practical performance limitations. For models above 1B or with transformer architectures , we'd recommend MLX instead.
- Adoption: We only consider models with over 1,000 monthly downloads on HuggingFace.
Vote with 👍🏻 👎🏻
Join 
for the latest updates
Please reply to this issue if you have requests for additional models.
Our focus is on audio models right now but more than happy to consider visual models too.
We don't plan on supporting language models in the near term. Thinking more about how to best support SLMs for developers.
Model selection criteria:
Vote with 👍🏻 👎🏻
Join
for the latest updates