#16X / BuilderExperimental未读
Three days ago I left autoresearch tuning nanochat for ~2 days on depth=12 model. It found ~20 changes that improved the validation loss. I tested these changes yesterday and all o
来源:@karpathy / x ·
暂无摘要,建议先打开原文快速判断。
推荐理由:推荐理由待生成,可根据标题、标签和来源先判断优先级。
- missing_llm_fields