{ "schema_version": 1, "builder_plan": { "scenario": "You already need task-specific training and want to reduce its memory or compute cost.", "change": "Change only the training technique on a small controlled run after checking prerequisites.", "baseline": "Your current training recipe with the same model, data split, and evaluation.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Held-out quality, peak memory, training time, and full compute cost to the target quality.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2610.29110", "title": "GRAM: Group-wise Rank-Aware Model Merging via Subspace Alignment", "summary": "GRAM replaces single uniform SVD-based weight merging of fine-tuned task models with principal-angle-based task grouping followed by a standard SVD merge within each coalition. The cost is holding K merged models instead of one (K=2 doubles inference memory and serving footprint) plus an offline merge-time computation over fine-tuned weights only. It can fail when tasks do not form geometrically coherent groups, when per-group rank budgets are too tight, and it degrades to standard merging if K must be one due to deployment constraints.", "rationale": "The pipeline is cheap, training-free, and feasible on a single 24 GB GPU for LoRA-scale models, but the gain depends on having multiple same-base fine-tuned models whose geometry supports grouping, so workload-specific validation is required before production adoption.", "technique": "Group-wise rank-aware model merging (GRAM)", "family": "treino_eficiente", "recommendation": "testar", "published": "2026-10-21", "as_of": "2026-10-22", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.29110", "independent_implementations": 1, "total_implementations": 2, "stars": 126, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "GRAM replaces single uniform SVD-based weight merging of fine-tuned task models with principal-angle-based task grouping followed by a standard SVD merge within each coalition. The cost is holding K merged models instead of one (K=2 doubles inference memory and serving footprint) plus an offline merge-time computation over fine-tuned weights only. It can fail when tasks do not form geometrically coherent groups, when per-group rank budgets are too tight, and it degrades to standard merging if K must be one due to deployment constraints.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.29110", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "With only one additional merged model (K=2), GRAM significantly mitigates subspace interference and consistently outperforms prior merging methods on GLUE with Flan-T5, 8-task CLIP vision, and Qwen2-7B-Instruct medical instruction tuning; no multiplicative factor is reported.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.29110", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }