Subculture Science Research Group's Blog
  • Blog
  • Portfolio
  • About us
  • Tags
  • 日本語
Subculture Science Research Group's Blog

Apple Silicon


Fine-tuning a Japanese LLM Without a GPU: One Mac, One Pipeline

 Posted on Sun Nov 15 2026  |  5 min  |  989 words  |    • Other languages: 日本語

No CUDA GPU, no cloud quota, no separate evaluation scripts. How FT_LLM turns fine-tuning a Japanese LLM into one Mac workflow, and where the approach still falls short. [Read More]
MLX  LoRA  LLM  Apple Silicon  fine-tuning  problem solving 

A LoRA Training Pipeline for Japanese LLMs on MLX: LFM2.5-1.2B-JP

 Posted on Sun Nov 15 2026  |  6 min  |  1212 words  |    • Other languages: 日本語

How FT_LLM fine-tunes LFM2.5-1.2B-JP with LoRA on Apple Silicon: config resolution, two execution modes, a chunked loss for long sequences, early stopping, and character-level evaluation, with code excerpts and measured numbers. [Read More]
MLX  LoRA  LLM  Apple Silicon  fine-tuning  engineering 

 • © 2026  •  Subculture Science Research Group's Blog

Hugo v0.142.0 powered  •  Theme Beautiful Hugo adapted from Beautiful Jekyll

  • 日本語
  • English