Commit Graph

  • bfc8ff6098 Fix training budget (19k steps / ~20B tokens), add Muon optimizer, correct variance scaling math, move figures to §Training Config, tighten abstract master ViperEkura 2026-07-08 22:49:43 +0800
  • 4a143b056d Tighten abstract, add Muon optimizer mention, update ckpt_comparison figure ViperEkura 2026-07-08 22:27:33 +0800
  • 5726bf6c84 Add optimizer/initialization comparison figure (ckpt_comparison) ViperEkura 2026-07-08 21:59:56 +0800
  • ff13a45d95 Restructure data pipeline: replace IFD analysis with length bias appendix ViperEkura 2026-07-05 18:59:29 +0800
  • 2d52a00149 Update ifd_loss_ratio_density.png ViperEkura 2026-07-05 12:38:43 +0800
  • f2c2b159ba Add IFD loss ratio density analysis, weight distribution appendix, and rewrite abstract ViperEkura 2026-07-05 12:31:40 +0800
  • 459523f0e4 Fix IFD citation: correct author list and venue (NAACL 2024, not NeurIPS) ViperEkura 2026-07-05 00:24:43 +0800
  • 4a7bf7ac3f Add GQA/SwiGLU/RoPE formula blocks, Alpaca-GPT4 citation, and improve IFD-Loss Ratio interpretation ViperEkura 2026-07-05 00:17:09 +0800
  • 5c915514b3 Add correlation analysis interpretation to IFD bias section ViperEkura 2026-07-04 23:45:55 +0800
  • bc8d4066c1 Add IFD analysis: density/length-bias figures and appendix ViperEkura 2026-07-04 23:33:12 +0800
  • 9534316bac Fix duplicate paragraph and misleading table caption ViperEkura 2026-07-04 22:39:16 +0800
  • b40fcc86c7 Refine abstract, switch to Times font, remove date, fix overfull hboxes with microtype ViperEkura 2026-07-04 22:31:45 +0800
  • dd412b3c9c Add IFD analysis, appendix, and improve abstract ViperEkura 2026-07-04 22:25:18 +0800
  • 73cb02f0aa fix training hyperparams to match pretraining script: lr 1.5e-4, wd 0.1, warmup 0.02 ViperEkura 2026-06-28 16:57:30 +0800
  • adc284920a add Alembic SFT data cleaning section with MinHash algorithm and fix bibliography order ViperEkura 2026-06-28 16:52:11 +0800
  • 13089b002a fix training config: grad_accum 32, remove unused validation & activation ckpt ViperEkura 2026-06-28 15:35:15 +0800
  • d3fbe0ecc6 refactor: numerical stability analysis with residual scaling comparison ViperEkura 2026-06-28 14:36:12 +0800
  • 3db1096c78 fix: add caption package, fix overfull hbox, correct RoFormer author names ViperEkura 2026-06-25 16:53:26 +0800
  • 6f9278b35b Initial commit: BF16 weight lock-in paper for 1.2B Transformer training with AstrAI ViperEkura 2026-06-25 15:51:05 +0800