From c804e3cb916be4cace8ec81a06b57349c79dba9f Mon Sep 17 00:00:00 2001 From: DeliberateEnsemble Date: Wed, 14 Oct 2026 15:50:37 -0400 Subject: [PATCH 1/2] docs: fix KTO data-shape row to use completion column The choosing-a-trainer table documented KTO rows as {prompt, response, label: bool}, but the KTO loader requires completion and label columns. Updated in English and Turkish. Fixes #885 --- docs/usermanuals/en/concepts/choosing-trainer.md | 2 +- docs/usermanuals/tr/concepts/choosing-trainer.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/usermanuals/en/concepts/choosing-trainer.md b/docs/usermanuals/en/concepts/choosing-trainer.md index 16d8b971..801b7a03 100644 --- a/docs/usermanuals/en/concepts/choosing-trainer.md +++ b/docs/usermanuals/en/concepts/choosing-trainer.md @@ -48,7 +48,7 @@ flowchart TD |---|---| | `{prompt, completion}` rows only | [SFT](#/training/sft) | | `{prompt, chosen, rejected}` rows | [DPO](#/training/dpo) (or [SimPO](#/training/simpo) if VRAM-tight) | -| `{prompt, response, label: bool}` rows | [KTO](#/training/kto) | +| `{prompt, completion, label: bool}` rows | [KTO](#/training/kto) | | Mix of SFT and preference rows | [SFT](#/training/sft) → [DPO](#/training/dpo) sequence, or [ORPO](#/training/orpo) for one-stage | | Programmatic reward (math grader, code tests) | [GRPO](#/training/grpo) | diff --git a/docs/usermanuals/tr/concepts/choosing-trainer.md b/docs/usermanuals/tr/concepts/choosing-trainer.md index e56f9a86..b3f0db67 100644 --- a/docs/usermanuals/tr/concepts/choosing-trainer.md +++ b/docs/usermanuals/tr/concepts/choosing-trainer.md @@ -48,7 +48,7 @@ flowchart TD |---|---| | Sadece `{prompt, completion}` satırları | [SFT](#/training/sft) | | `{prompt, chosen, rejected}` satırları | [DPO](#/training/dpo) (VRAM dar ise [SimPO](#/training/simpo)) | -| `{prompt, response, label: bool}` satırları | [KTO](#/training/kto) | +| `{prompt, completion, label: bool}` satırları | [KTO](#/training/kto) | | SFT ve tercih satırlarının karışımı | [SFT](#/training/sft) → [DPO](#/training/dpo) sıralı, veya tek aşama için [ORPO](#/training/orpo) | | Programatik reward (math grader, kod testi) | [GRPO](#/training/grpo) | From e93450894e4d3721b0dbea8394948409005a1048 Mon Sep 17 00:00:00 2001 From: Cemil ILIK Date: Sat, 17 Oct 2026 20:05:58 +0300 Subject: [PATCH 2/2] docs(changelog): record the KTO data-shape fix Adds the [Unreleased] > Fixed entry for the "Choosing a Trainer" KTO row. The SFT row of the same table is a separate defect tracked in #403 and is left unchanged. Resolves: F-W1-DOCS-TRAIN-02 Refs #885 Co-Authored-By: Claude Sonnet 5.5 --- CHANGELOG.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/CHANGELOG.md b/CHANGELOG.md index 7c510f6d..a3f8eaa7 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -54,6 +54,10 @@ All notable changes to ForgeLM are documented here. article and gender slips in the touched German, French and Spanish strings, and use one German term for the external stack. The new validation and interval keys keep their English fallback; native review of the deferred locales remains open (refs #948). +- Name the right columns for KTO in the EN/TR "Choosing a Trainer" user manual: its data-shape table + said `{prompt, response, label: bool}`, but the loader requires `{prompt, completion, label}` and + rejects the old shape with a `KeyError` (F-W1-DOCS-TRAIN-02, #885). The table's SFT row is a separate + defect, tracked in #403, and is unchanged here. - Say what Annex IV §9 asks for: a detailed description of the post-market monitoring system, its plan included. The wizard hint in every locale and the EN/TR configuration and compliance pages said it "includes the detailed plan"; German, French and Spanish now follow the Regulation's wording, and