nearswap_Merged_Qwen2-0.5B_0.4/mergekit_config.yml

base_model: /kaggle/working/Qwen2-0.5B
dtype: float32
merge_method: nearswap
models:
- model: /kaggle/working/debias_Qwen2-0.5B
parameters:
  t: 0.4
tokenizer:
  source: /kaggle/working/Qwen2-0.5B
初始化项目，由ModelHub XC社区提供模型 Model: trinhkhng/nearswap_Merged_Qwen2-0.5B_0.4 Source: Original Platform 2026-08-12 14:47:16 +08:00			`base_model: /kaggle/working/Qwen2-0.5B`
			`dtype: float32`
			`merge_method: nearswap`
			`models:`
			`- model: /kaggle/working/debias_Qwen2-0.5B`
			`parameters:`
			`t: 0.4`
			`tokenizer:`
			`source: /kaggle/working/Qwen2-0.5B`