---
title: Qwen3-0.6B
canonical_url: "https://www.modelscope.cn/models/mlx-community/Qwen3-0.6B"
md_url: "https://www.modelscope.cn/models/mlx-community/Qwen3-0.6B.md"
repository: mlx-community/Qwen3-0.6B
last_updated: 2025-07-17
license: mit
base_model:
  - Qwen/Qwen3-0.6B
base_model_relation: finetune
parameters: 655360
tensor_type:
  - F32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 156
stars: 0
tags:
  - math
---

# Qwen3-0.6B

> Qwen3-0.6B - mlx-community 在 ModelScope 开源的模型。"batchsize": 1, "finetunetype": "lora", "gradcheckpoint": true, "iters": 1000, "learningrate": 2e-05, "loraparameters": { "rank": 8, "dropout": 0.0, "scale": 20.0 "testloss": 1.017, "Test ppl": 2.764

mlx-community/Qwen3-0.6B 是 ModelScope 魔搭社区上的 655360 参数机器学习模型，采用 mit 许可，基于 Qwen/Qwen3-0.6B 构建。

- **Repository**: mlx-community/Qwen3-0.6B
- **License**: mit
- **Parameters**: 655360
- **Base model**: Qwen/Qwen3-0.6B
- **Tags**: math
- **Downloads**: 156
- **Stars**: 0
- **Last updated**: 2025-07-17

Source: https://www.modelscope.cn/models/mlx-community/Qwen3-0.6B

---

"batch_size": 1,
    "fine_tune_type": "lora",
    "grad_checkpoint": true,
    "iters": 1000,
    "learning_rate": 2e-05,
    "lora_parameters": {
        "rank": 8,
        "dropout": 0.0,
        "scale": 20.0
    "test_loss": 1.017, 
    "Test ppl": 2.764