# SANA LRM Checklist V2 ## 1. Scope and Dependency Rules - [x] Core parity scope is trainer architecture and runtime contracts across flux, lrm_15, lrm_xl, and lrm_sana. - [x] `information_related_to_sana` is treated as model-specific external adapter code. - [x] `information_related_to_flux` is treated as model-specific external adapter code. - [x] External model adapter modules are excluded from core trainer parity checks. ## 2. Codebase Parity Against Existing Variants - [x] SANA has expected variant-specific replacements for config/task/model/dataset/criterion modules. - [x] SANA keeps shared trainer scaffolding structure aligned with other variants. - [x] Launcher flow preserves profile-based behavior and runtime overrides. - [x] Optimizer/lr-scheduler defaults are set for runnable quick smoke behavior. - [x] Quick profile precision default updated to BF16 for stable H200 smoke runs. ## 3. Runtime Verification Evidence (centralized in log.log) - [x] Comparison inventory and focused diffs recorded in log output. - [x] Hydra composition succeeds (`Hydra compose bg exit code: 0`). - [x] Deterministic quick smoke run reaches final evaluation end marker (`EVAL END (final@gstep=1)`). - [x] Deterministic quick smoke run saves checkpoint (`checkpoint-gstep1`). - [x] All verification command outputs are consolidated under `lrm_sana/log.log`. ## 4. Remaining Validation Items - [ ] Initialization smoke checks for all four SANA model profiles. - [ ] Optional distributed_sharded path validation on nodes with CUDA toolchain available for DeepSpeed ops. - [ ] Extended multi-step sanity run (beyond max_steps=1). ## 5. Canonical Artifacts - Main aggregated run/comparison log: `lrm_sana/log.log` - Previous checklist: `lrm_sana/docs/checklist.md` - This checklist revision: `lrm_sana/docs/checklistv2.md`