---
title: moshika-vis-mlx
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/moshika-vis-mlx"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/moshika-vis-mlx.md"
repository: AI-ModelScope/moshika-vis-mlx
last_updated: 2025-03-21
license: cc-by-4.0
base_model:
  - google/paligemma2-3b-pt-448
  - kyutai/moshika-mlx-bf16
base_model_relation: finetune
parameters: 13.5B
tensor_type:
  - U32
  - F32
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 544
stars: 0
---

# moshika-vis-mlx

> moshika-vis-mlx - AI-ModelScope 在 ModelScope 开源的模型。Please refer to the main model card

AI-ModelScope/moshika-vis-mlx 是 ModelScope 魔搭社区上的 13.5B 参数机器学习模型，采用 cc-by-4.0 许可，基于 google/paligemma2-3b-pt-448、kyutai/moshika-mlx-bf16 构建。

- **Repository**: AI-ModelScope/moshika-vis-mlx
- **License**: cc-by-4.0
- **Parameters**: 13.5B
- **Base model**: google/paligemma2-3b-pt-448, kyutai/moshika-mlx-bf16
- **Downloads**: 544
- **Stars**: 0
- **Last updated**: 2025-03-21

Source: https://www.modelscope.cn/models/AI-ModelScope/moshika-vis-mlx

---

Please refer to the [main model card](https://huggingface.co/kyutai/moshika-vis-pytorch-bf16)

This model page contains the Moshika (female voice) model weights for the `MLX` backend of the MoshiVis repo, 
in `bfloat16`,`Q8` and `Q4` formats. We provide the same model weights for other backends and quantization formats in the associated model collection.