Temporary public mirror: foundational-curriculum-v1.1 + eval batteries (snapshot @7c43edd)

.gitattributes

@@ -58,3 +58,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+battery/eval_locate16k/test.jsonl filter=lfs diff=lfs merge=lfs -text
+battery/eval_twohop/test.jsonl filter=lfs diff=lfs merge=lfs -text
+curriculum/train.jsonl filter=lfs diff=lfs merge=lfs -text

DATACARD.md

@@ -0,0 +1,178 @@
+# Foundational Structural Curriculum — DATACARD
+
+**Curriculum version:** foundational-curriculum-v1.1  ·  **Eval version:** foundational-eval-v1
+**Base model:** meta-llama/Llama-3.2-1B  ·  **Bank seed:** 20260711
+
+## Overview
+A procedurally-generated table-reasoning curriculum spanning nine general-audience tasks (column/row extraction, subtable and transpose views, predicate selection, argmax/argmin/median lookup, and odd-one-out anomaly detection) over two sources: synthetic structural-causal-model (SCM) worlds and real MMTU table grids. Sealed batteries hold out chained multi-hop lookups, row-address generalization, a 16k-token stress tier, out-of-distribution table sizes, and a small curated set, plus a replay/forgetting-probe slice drawn from Dolly-15k.
+
+## Construction
+- World bank: 1000 worlds (params_hash=b66e43ee018a6605, generator_git_sha=fd0e79a7b00fbe5cacc451ffb391dfe82bb709ae)
+- Augmentation: {'frac': 0.15, 'k': 1, 'seed': 4046}
+- Replay: databricks-dolly-15k @ bdd27f4d94b9c1f951818a7da7fd7aeea5dbff1a (3000 replay + 500 probe records)
+
+## Per-split record counts
+| split | n records |
+|---|---:|
+| train | 37260 |
+| eval_chained | 1500 |
+| eval_curated10 | 1200 |
+| eval_locate16k | 300 |
+| eval_ood_size | 792 |
+| eval_rowaddr | 600 |
+| eval_twohop | 2400 |
+| replay | 3000 |
+| probe_forgetting | 500 |
+
+## Per-task x per-source x per-split counts
+| task | source | split | n |
+|---|---|---|---:|
+| argmax | mmtu | eval_ood_size | 82 |
+| argmax | mmtu | train | 584 |
+| argmax | scm | eval_curated10 | 100 |
+| argmax | scm | eval_ood_size | 50 |
+| argmax | scm | train | 864 |
+| argmin | mmtu | eval_ood_size | 82 |
+| argmin | mmtu | train | 582 |
+| argmin | scm | eval_curated10 | 100 |
+| argmin | scm | eval_ood_size | 50 |
+| argmin | scm | train | 872 |
+| chained_lookup | scm | eval_chained | 1500 |
+| extract_column | mmtu | eval_ood_size | 82 |
+| extract_column | mmtu | train | 2272 |
+| extract_column | scm | eval_curated10 | 100 |
+| extract_column | scm | eval_ood_size | 50 |
+| extract_column | scm | train | 3493 |
+| extract_row | mmtu | eval_ood_size | 82 |
+| extract_row | mmtu | train | 1735 |
+| extract_row | scm | eval_curated10 | 100 |
+| extract_row | scm | eval_ood_size | 50 |
+| extract_row | scm | train | 2911 |
+| index_lookup | mmtu | eval_locate16k | 300 |
+| median | mmtu | eval_ood_size | 82 |
+| median | mmtu | train | 913 |
+| median | scm | eval_curated10 | 100 |
+| median | scm | eval_ood_size | 50 |
+| median | scm | train | 1141 |
+| odd_one_out | scm | eval_curated10 | 100 |
+| odd_one_out | scm | train | 2321 |
+| pk_lookup | mmtu | eval_rowaddr | 300 |
+| pk_lookup | mmtu | eval_twohop | 300 |
+| pk_lookup | mmtu | train | 1149 |
+| pk_lookup | scm | eval_curated10 | 100 |
+| pk_lookup | scm | eval_rowaddr | 300 |
+| pk_lookup | scm | eval_twohop | 300 |
+| pk_lookup | scm | train | 1716 |
+| select_cmp_single | mmtu | eval_twohop | 300 |
+| select_cmp_single | mmtu | train | 696 |
+| select_cmp_single | scm | eval_curated10 | 100 |
+| select_cmp_single | scm | eval_twohop | 300 |
+| select_cmp_single | scm | train | 1135 |
+| select_eq_single | mmtu | eval_twohop | 300 |
+| select_eq_single | mmtu | train | 1151 |
+| select_eq_single | scm | eval_curated10 | 100 |
+| select_eq_single | scm | eval_twohop | 300 |
+| select_eq_single | scm | train | 1731 |
+| select_predicate | mmtu | eval_ood_size | 82 |
+| select_predicate | mmtu | eval_twohop | 300 |
+| select_predicate | mmtu | train | 2278 |
+| select_predicate | scm | eval_curated10 | 100 |
+| select_predicate | scm | eval_ood_size | 50 |
+| select_predicate | scm | eval_twohop | 300 |
+| select_predicate | scm | train | 2872 |
+| subtable | mmtu | train | 1720 |
+| subtable | scm | eval_curated10 | 100 |
+| subtable | scm | train | 2255 |
+| transpose | mmtu | train | 1157 |
+| transpose | scm | eval_curated10 | 100 |
+| transpose | scm | train | 1712 |
+
+## Two-hop filter-type ramp
+`eval_twohop` is a v1.1-additive sealed battery; the original foundational-eval-v1 five batteries (eval_chained, eval_rowaddr, eval_locate16k, eval_ood_size, eval_curated10) are byte-identical and unchanged/reusable.
+| filter_type | split | n |
+|---|---|---:|
+| pk_eq | eval_curated10 | 100 |
+| pk_eq | eval_rowaddr | 600 |
+| pk_eq | eval_twohop | 600 |
+| pk_eq | train | 2865 |
+| eq_single | eval_curated10 | 100 |
+| eq_single | eval_twohop | 600 |
+| eq_single | train | 2882 |
+| cmp_single | eval_curated10 | 100 |
+| cmp_single | eval_twohop | 600 |
+| cmp_single | train | 1831 |
+| eq_multi | eval_curated10 | 59 |
+| eq_multi | eval_ood_size | 82 |
+| eq_multi | eval_twohop | 383 |
+| eq_multi | train | 3331 |
+| in_multi | eval_curated10 | 41 |
+| in_multi | eval_ood_size | 50 |
+| in_multi | eval_twohop | 217 |
+| in_multi | train | 1819 |
+
+## Distributions
+| field | min | p50 | max | mean |
+|---|---:|---:|---:|---:|
+| n_cols | 4.0 | 17.0 | 51.0 | 18.801 |
+| n_rows | 4.0 | 12.0 | 50.0 | 13.269 |
+| chain_k | 1.0 | 3.0 | 5.0 | 3.0 |
+
+### Token length (from census.json, exact dlm assembly path)
+| split/file | n | drop_pct | over_cap | max_len | cdf@2048 | cdf@4096 |
+|---|---:|---:|---:|---:|---:|---:|
+| curriculum/train | 37260 | 0.0% | 0 | 4000 | 53.5% | 100.0% |
+| eval_chained | 1500 | 0.0% | 0 | 4000 | 46.67% | 100.0% |
+| eval_curated10 | 1200 | 0.0% | 0 | 1444 | 100.0% | 100.0% |
+| eval_locate16k | 300 | 100.0% | 300 | 48859 | 0.0% | 0.0% |
+| eval_ood_size | 792 | 0.0% | 0 | 3997 | 4.04% | 100.0% |
+| eval_rowaddr | 600 | 0.0% | 0 | 3997 | 28.5% | 100.0% |
+| eval_twohop | 2400 | 0.0% | 0 | 3998 | 47.46% | 100.0% |
+| probe_forgetting | 500 | 0.0% | 0 | 528 | 100.0% | 100.0% |
+| replay | 3000 | 0.0% | 0 | 569 | 100.0% | 100.0% |
+
+## Disjointness verification
+- `world_disjoint`: **True**
+- `table_disjoint`: **True**
+- `disjoint_by_hash`: **True**
+- train worlds: 850  ·  battery worlds: 150
+- train grids: 84  ·  battery grids: 15
+
+## Census / verify gates
+- Census (cap 4096): drop_pct=0.0% (max 5.0%) — **PASS**
+- verify_dataset: 37260 train + 6792 battery records checked, 0 errors
+
+## Limitations
+- Only 99 distinct MMTU grids underlie the real-source split; this thins the real-source half of every (qtype, source) pool relative to the procedural half.
+- Procedural (SCM) worlds are synthetic — structurally faithful to the generator's mechanisms but not drawn from real-world tables.
+- Scorer/build normalizer policy: dlm's `_normalize` is the binding screen space; the exp040 `norm_key` variant is stricter and is applied ONLY at build time (screens), not by the shipped scorer.
+- odd_one_out is procedural-SCM-only (no MMTU arm) — a source asymmetry specific to this task, documented and accepted.
+- eval_locate16k is a 16k-token stress tier and is EXEMPT from the 4096 drop gate by design (J-7); its long-context tokens are expected to exceed cap.
+
+## Consumption contract
+Downstream experiments (047+) pin the exact DVC rev + version below and re-run their own drop gate at their chosen train cap (R76) using the published length CDF. Hydrate via `dvc pull` at the pinned rev, or a byte-verified HF mirror of the same payload (R60).
+```json
+{
+  "curriculum_version": "foundational-curriculum-v1.1",
+  "eval_version": "foundational-eval-v1",
+  "dvc_rev": "fd0e79a7b00fbe5cacc451ffb391dfe82bb709ae",
+  "bank_seed": 20260711,
+  "params_hash": "b66e43ee018a6605",
+  "generator_git_sha": "fd0e79a7b00fbe5cacc451ffb391dfe82bb709ae",
+  "scorer_module": "src/dlm/grading/denotation.py:denotation_match + foundational/screens.py cell-aligned (delimiters ' | ' / ' ; ')",
+  "split_names": [
+    "train",
+    "val",
+    "eval_chained",
+    "eval_rowaddr",
+    "eval_locate16k",
+    "eval_ood_size",
+    "eval_curated10",
+    "probe_forgetting"
+  ],
+  "replay_source_rev": "bdd27f4d94b9c1f951818a7da7fd7aeea5dbff1a"
+}
+```
+
+## Provenance
+Built as exp046 "Foundational" (structural curriculum). Effort estimate (author, SPEC §F): generator 2.5-3.5d, verify_world 1-1.5d, curriculum builder 3-4d, test battery 2-3d, DVC pipeline 1-1.5d, datacard/verify_dataset/census 1.5-2d — total ~12-17 engineering days, no pod time (dataset construction is CPU-only; GPU budget belongs to the downstream training experiments).
+Record-count budget per (task, source): see `params.yaml` `mix:` (R41). Task-mix weights used for this build: `{"extract_column": {"scm": 3000, "mmtu": 2000}, "extract_row": {"scm": 2500, "mmtu": 1500}, "subtable": {"scm": 2000, "mmtu": 1500}, "transpose": {"scm": 1500, "mmtu": 1000}, "select_predicate": {"scm": 2500, "mmtu": 2000}, "pk_lookup": {"scm": 1500, "mmtu": 1000}, "select_eq_single": {"scm": 1500, "mmtu": 1000}, "select_cmp_single": {"scm": 1000, "mmtu": 600}, "argmax": {"scm": 750, "mmtu": 500}, "argmin": {"scm": 750, "mmtu": 500}, "median": {"scm": 1000, "mmtu": 800}, "odd_one_out": {"scm": 2000, "mmtu": 0}}`.

README.md

@@ -0,0 +1,102 @@
+---
+license: cc-by-4.0
+language:
+  - en
+tags:
+  - table-reasoning
+  - table-question-answering
+  - synthetic
+pretty_name: Foundational Structural Curriculum (v1.1)
+configs:
+  - config_name: curriculum
+    data_files: curriculum/train.jsonl
+    default: true
+  - config_name: eval_chained
+    data_files: battery/eval_chained/test.jsonl
+  - config_name: eval_curated10
+    data_files: battery/eval_curated10/test.jsonl
+  - config_name: eval_locate16k
+    data_files: battery/eval_locate16k/test.jsonl
+  - config_name: eval_ood_size
+    data_files: battery/eval_ood_size/test.jsonl
+  - config_name: eval_rowaddr
+    data_files: battery/eval_rowaddr/test.jsonl
+  - config_name: eval_twohop
+    data_files: battery/eval_twohop/test.jsonl
+---
+
+# Foundational Structural Curriculum — v1.1
+
+> **TEMPORARY MIRROR.** This repository is a convenience copy for browsing and
+> exploration, published so a reader can look at the data without S3/DVC
+> credentials. It is **not** the canonical dataset and is **not** guaranteed to
+> stay in sync. The source of truth is the DVC-tracked corpus in the
+> `graph-reasoning-llm` repo (pin by the git SHA of `dvc.lock`, never "latest").
+> The usual byte-for-byte verification against `dvc.lock` was **skipped** for
+> this upload. Treat these files as a snapshot that may be removed or replaced.
+
+A procedurally generated table-reasoning curriculum plus sealed evaluation
+batteries, built for training a Llama-3.2-1B model that reasons over in-context
+structured tables.
+
+## What's here
+
+| Config | Records | Role |
+|---|---:|---|
+| `curriculum` | 37,260 | Training records (the default config). |
+| `eval_chained` | 1,500 | Held-out chained multi-hop lookups. |
+| `eval_curated10` | 1,200 | Small curated set. |
+| `eval_locate16k` | 300 | 16k-token stress tier. |
+| `eval_ood_size` | 792 | Out-of-distribution table sizes. |
+| `eval_rowaddr` | 600 | Row-address generalization. |
+| `eval_twohop` | 2,400 | Two-hop lookups (by filter type). |
+
+Also included: `splits.json`, `curriculum/build_manifest.json`,
+`battery/battery_manifest.json` (per-battery `n` + sha256), and `DATACARD.md`
+(the full committed data card with per-task / per-source / per-split counts).
+
+## Record schema
+
+Each line is one JSON object. Core fields shared by every record:
+
+- `question` — the natural-language query.
+- `gold` — the reference answer (a ` | `-joined string for multi-value answers).
+- `columns` — list of column names.
+- `rows` — the table body as a list of rows.
+- `cells` — the table as `[row_idx, col_idx]`-addressable cells.
+- `qtype` — task type (`extract_column`, `argmax`, `chained_lookup`, ...).
+- `source` — `scm` (synthetic structural-causal-model world) or `mmtu` (real
+  MMTU table grid).
+- `n_rows`, `n_cols`, `target_col_idx`, `target_col_name`, `target_row_idx`,
+  `id`, `source_id`, `split`.
+
+Battery records add task-specific fields (`filter_type`, `hop_count`,
+`key_col_idx`, `key_col_name`, `key_value`, `n_matches`).
+
+## Load
+
+```python
+from datasets import load_dataset
+
+train = load_dataset("akumch/graph-reasoning-foundational-curriculum", "curriculum", split="train")
+rowaddr = load_dataset("akumch/graph-reasoning-foundational-curriculum", "eval_rowaddr", split="train")
+print(train[0]["question"], "->", train[0]["gold"])
+```
+
+Or just browse the **Data Studio** viewer tab above — no download needed.
+
+## Provenance
+
+- **Curriculum version:** `foundational-curriculum-v1.1`
+- **Eval version:** `foundational-eval-v1`
+- **Base model:** meta-llama/Llama-3.2-1B
+- **World bank:** 1,000 SCM worlds, bank seed `20260711`
+- **Sources:** synthetic SCM worlds + real MMTU table grids
+- Snapshot taken from `graph-reasoning-llm` at commit `7c43edd`
+  (`dvc.lock` last changed at `7882746`).
+
+## License
+
+Released under **CC-BY-4.0**. The synthetic SCM portion is original. The `mmtu`
+portion re-hosts tables drawn from the MMTU benchmark; if a different license is
+needed for that portion, open an issue on the dataset repo and it will be added.

battery/battery_manifest.json

@@ -0,0 +1,47 @@
+{
+ "eval_chained": {
+  "n": 1500,
+  "n_pilot": 100,
+  "sha256": "f2b8360cfa273f47b5807e247fa684385aa794b7a11ad405de7d1e728970bccd",
+  "k_row_corr": 0.00561176317462427,
+  "per_k": {
+   "1": 300,
+   "2": 300,
+   "3": 300,
+   "4": 300,
+   "5": 300
+  }
+ },
+ "eval_rowaddr": {
+  "n": 600,
+  "n_pilot": 100,
+  "sha256": "8f6a01176569945bf6dbb467f573f0c6667754195b340245378c647f96e4c8c4"
+ },
+ "eval_locate16k": {
+  "n": 300,
+  "n_pilot": 100,
+  "sha256": "f864e3a2eae6c9beb6b585ac2cd09509afe54d9f28b5b47f6f1fc900864e1113"
+ },
+ "eval_ood_size": {
+  "n": 792,
+  "n_pilot": 100,
+  "sha256": "8b72accc208ba42666c56092e7a046271804d821823d12f61350c82dc15c960b"
+ },
+ "eval_curated10": {
+  "n": 1200,
+  "n_pilot": 100,
+  "sha256": "4b962d63474f47bc824feeba4b432450137a521d5dbd2493e1a9bdb2aeed1c21"
+ },
+ "eval_twohop": {
+  "n": 2400,
+  "n_pilot": 100,
+  "sha256": "741cd770d9292d7d99ee3ad08dc18091b293fc1dfe6692abdaaa8dafb8a29ed9",
+  "per_filter_type": {
+   "pk_eq": 600,
+   "eq_single": 600,
+   "cmp_single": 600,
+   "eq_multi": 383,
+   "in_multi": 217
+  }
+ }
+}

battery/eval_locate16k/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f864e3a2eae6c9beb6b585ac2cd09509afe54d9f28b5b47f6f1fc900864e1113
+size 14265896

battery/eval_twohop/test.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:741cd770d9292d7d99ee3ad08dc18091b293fc1dfe6692abdaaa8dafb8a29ed9
+size 10549331

curriculum/build_manifest.json

@@ -0,0 +1,41 @@
+{
+ "n_base": 32400,
+ "n_aug": 4860,
+ "n_total": 37260,
+ "counts": {
+  "extract_column/scm": 3000,
+  "extract_column/mmtu": 2000,
+  "extract_row/scm": 2500,
+  "extract_row/mmtu": 1500,
+  "subtable/scm": 2000,
+  "subtable/mmtu": 1500,
+  "transpose/scm": 1500,
+  "transpose/mmtu": 1000,
+  "select_predicate/scm": 2500,
+  "select_predicate/mmtu": 2000,
+  "pk_lookup/scm": 1500,
+  "pk_lookup/mmtu": 1000,
+  "select_eq_single/scm": 1500,
+  "select_eq_single/mmtu": 1000,
+  "select_cmp_single/scm": 1000,
+  "select_cmp_single/mmtu": 600,
+  "argmax/scm": 750,
+  "argmax/mmtu": 500,
+  "argmin/scm": 750,
+  "argmin/mmtu": 500,
+  "median/scm": 1000,
+  "median/mmtu": 800,
+  "odd_one_out/scm": 2000,
+  "odd_one_out/mmtu": 0
+ },
+ "aug": {
+  "frac": 0.15,
+  "k": 1,
+  "seed": 4046
+ },
+ "n_train_worlds": 850,
+ "n_held_worlds": 150,
+ "n_train_grids": 84,
+ "n_held_grids": 15,
+ "train_sha256": "0e8c8546df370823a1f89ef7949ce6729ebff8ae50f75aad3f62be85393d923c"
+}

curriculum/train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e8c8546df370823a1f89ef7949ce6729ebff8ae50f75aad3f62be85393d923c
+size 155991703

splits.json

@@ -0,0 +1,1115 @@
+{
+ "bank_seed": 20260711,
+ "held_out_frac": 0.15,
+ "worlds": {
+  "train": [
+   0,
+   1,
+   3,
+   4,
+   5,
+   7,
+   8,
+   11,
+   12,
+   13,
+   14,
+   15,
+   16,
+   17,
+   18,
+   22,
+   23,
+   24,
+   25,
+   26,
+   28,
+   29,
+   30,
+   31,
+   32,
+   33,
+   34,
+   35,
+   36,
+   37,
+   38,
+   40,
+   41,
+   42,
+   43,
+   44,
+   45,
+   46,
+   47,
+   50,
+   51,
+   52,
+   53,
+   54,
+   55,
+   56,
+   57,
+   58,
+   59,
+   60,
+   61,
+   62,
+   63,
+   65,
+   66,
+   67,
+   68,
+   69,
+   71,
+   72,
+   73,
+   74,
+   77,
+   78,
+   79,
+   80,
+   83,
+   84,
+   85,
+   87,
+   88,
+   90,
+   91,
+   92,
+   93,
+   94,
+   95,
+   96,
+   97,
+   98,
+   99,
+   100,
+   101,
+   102,
+   103,
+   104,
+   105,
+   106,
+   108,
+   109,
+   110,
+   111,
+   113,
+   114,
+   116,
+   117,
+   119,
+   121,
+   122,
+   123,
+   124,
+   125,
+   126,
+   127,
+   128,
+   129,
+   130,
+   131,
+   132,
+   133,
+   134,
+   135,
+   136,
+   137,
+   138,
+   139,
+   140,
+   141,
+   142,
+   143,
+   144,
+   146,
+   147,
+   148,
+   149,
+   150,
+   151,
+   152,
+   153,
+   154,
+   155,
+   156,
+   157,
+   158,
+   159,
+   160,
+   161,
+   162,
+   163,
+   164,
+   165,
+   166,
+   167,
+   168,
+   169,
+   170,
+   171,
+   172,
+   173,
+   174,
+   175,
+   176,
+   177,
+   178,
+   180,
+   181,
+   182,
+   183,
+   184,
+   185,
+   186,
+   187,
+   188,
+   189,
+   190,
+   191,
+   192,
+   193,
+   194,
+   195,
+   196,
+   197,
+   198,
+   199,
+   200,
+   201,
+   203,
+   205,
+   206,
+   207,
+   211,
+   213,
+   214,
+   215,
+   217,
+   218,
+   219,
+   220,
+   221,
+   222,
+   225,
+   226,
+   227,
+   228,
+   229,
+   231,
+   232,
+   233,
+   234,
+   236,
+   237,
+   238,
+   239,
+   240,
+   242,
+   243,
+   244,
+   245,
+   246,
+   247,
+   248,
+   249,
+   250,
+   251,
+   252,
+   254,
+   255,
+   256,
+   257,
+   258,
+   259,
+   260,
+   261,
+   262,
+   264,
+   267,
+   268,
+   270,
+   271,
+   272,
+   273,
+   274,
+   275,
+   276,
+   277,
+   279,
+   280,
+   281,
+   282,
+   283,
+   284,
+   285,
+   286,
+   288,
+   289,
+   290,
+   291,
+   292,
+   293,
+   294,
+   295,
+   296,
+   297,
+   298,
+   301,
+   302,
+   303,
+   304,
+   305,
+   306,
+   308,
+   309,
+   310,
+   311,
+   312,
+   313,
+   315,
+   317,
+   318,
+   319,
+   320,
+   321,
+   322,
+   323,
+   325,
+   326,
+   327,
+   328,
+   329,
+   330,
+   331,
+   332,
+   333,
+   334,
+   335,
+   336,
+   337,
+   338,
+   341,
+   342,
+   343,
+   344,
+   345,
+   347,
+   348,
+   349,
+   350,
+   351,
+   352,
+   354,
+   355,
+   357,
+   359,
+   360,
+   361,
+   362,
+   363,
+   364,
+   365,
+   366,
+   367,
+   368,
+   369,
+   370,
+   371,
+   372,
+   373,
+   374,
+   375,
+   376,
+   377,
+   378,
+   379,
+   380,
+   381,
+   382,
+   384,
+   385,
+   386,
+   387,
+   388,
+   390,
+   391,
+   392,
+   393,
+   394,
+   396,
+   397,
+   399,
+   400,
+   401,
+   402,
+   403,
+   404,
+   405,
+   408,
+   409,
+   410,
+   411,
+   412,
+   413,
+   415,
+   416,
+   417,
+   419,
+   420,
+   421,
+   422,
+   423,
+   425,
+   426,
+   427,
+   428,
+   429,
+   430,
+   431,
+   432,
+   433,
+   435,
+   436,
+   439,
+   441,
+   445,
+   446,
+   447,
+   448,
+   449,
+   450,
+   452,
+   453,
+   454,
+   455,
+   456,
+   457,
+   458,
+   459,
+   460,
+   461,
+   462,
+   463,
+   464,
+   465,
+   466,
+   467,
+   468,
+   469,
+   470,
+   471,
+   472,
+   474,
+   475,
+   477,
+   478,
+   479,
+   480,
+   482,
+   483,
+   484,
+   485,
+   486,
+   489,
+   490,
+   491,
+   492,
+   493,
+   494,
+   495,
+   496,
+   497,
+   498,
+   499,
+   500,
+   501,
+   502,
+   503,
+   504,
+   505,
+   506,
+   507,
+   508,
+   509,
+   510,
+   512,
+   513,
+   514,
+   515,
+   516,
+   517,
+   518,
+   519,
+   520,
+   521,
+   522,
+   523,
+   525,
+   526,
+   527,
+   530,
+   531,
+   532,
+   533,
+   534,
+   535,
+   536,
+   537,
+   538,
+   539,
+   540,
+   541,
+   542,
+   543,
+   544,
+   546,
+   547,
+   549,
+   550,
+   551,
+   552,
+   553,
+   554,
+   555,
+   557,
+   559,
+   560,
+   561,
+   562,
+   563,
+   564,
+   565,
+   566,
+   567,
+   568,
+   570,
+   572,
+   573,
+   574,
+   575,
+   577,
+   578,
+   579,
+   580,
+   581,
+   582,
+   583,
+   584,
+   587,
+   588,
+   589,
+   590,
+   592,
+   593,
+   594,
+   595,
+   596,
+   597,
+   598,
+   599,
+   601,
+   602,
+   603,
+   604,
+   606,
+   607,
+   608,
+   609,
+   610,
+   611,
+   612,
+   613,
+   614,
+   615,
+   616,
+   617,
+   618,
+   619,
+   620,
+   621,
+   622,
+   623,
+   624,
+   625,
+   626,
+   627,
+   628,
+   629,
+   630,
+   631,
+   632,
+   633,
+   634,
+   635,
+   636,
+   637,
+   638,
+   639,
+   640,
+   641,
+   642,
+   644,
+   645,
+   646,
+   647,
+   648,
+   650,
+   652,
+   653,
+   654,
+   655,
+   656,
+   657,
+   658,
+   660,
+   661,
+   662,
+   663,
+   664,
+   665,
+   666,
+   668,
+   669,
+   670,
+   671,
+   672,
+   673,
+   674,
+   675,
+   676,
+   677,
+   678,
+   679,
+   680,
+   681,
+   683,
+   685,
+   686,
+   687,
+   688,
+   689,
+   690,
+   691,
+   692,
+   693,
+   696,
+   697,
+   700,
+   701,
+   702,
+   703,
+   704,
+   705,
+   706,
+   710,
+   711,
+   712,
+   714,
+   715,
+   716,
+   717,
+   718,
+   719,
+   720,
+   721,
+   722,
+   723,
+   725,
+   726,
+   728,
+   729,
+   730,
+   731,
+   734,
+   736,
+   737,
+   739,
+   740,
+   741,
+   742,
+   744,
+   745,
+   746,
+   747,
+   748,
+   749,
+   750,
+   751,
+   752,
+   753,
+   754,
+   755,
+   756,
+   758,
+   759,
+   760,
+   761,
+   763,
+   764,
+   765,
+   766,
+   767,
+   769,
+   770,
+   773,
+   774,
+   775,
+   776,
+   778,
+   779,
+   780,
+   781,
+   782,
+   784,
+   785,
+   786,
+   787,
+   789,
+   790,
+   791,
+   792,
+   794,
+   795,
+   796,
+   797,
+   798,
+   799,
+   800,
+   801,
+   803,
+   804,
+   805,
+   806,
+   807,
+   808,
+   809,
+   810,
+   811,
+   812,
+   813,
+   814,
+   815,
+   816,
+   817,
+   818,
+   819,
+   821,
+   822,
+   823,
+   825,
+   826,
+   827,
+   828,
+   830,
+   832,
+   833,
+   834,
+   835,
+   836,
+   837,
+   839,
+   840,
+   841,
+   842,
+   843,
+   844,
+   845,
+   846,
+   848,
+   849,
+   850,
+   851,
+   852,
+   854,
+   855,
+   856,
+   857,
+   860,
+   861,
+   862,
+   863,
+   864,
+   865,
+   866,
+   867,
+   868,
+   869,
+   870,
+   871,
+   872,
+   874,
+   875,
+   876,
+   877,
+   879,
+   880,
+   882,
+   883,
+   884,
+   885,
+   886,
+   887,
+   888,
+   889,
+   890,
+   891,
+   892,
+   893,
+   894,
+   895,
+   896,
+   897,
+   898,
+   899,
+   900,
+   903,
+   905,
+   906,
+   907,
+   908,
+   909,
+   910,
+   911,
+   912,
+   913,
+   914,
+   915,
+   916,
+   917,
+   919,
+   920,
+   922,
+   923,
+   924,
+   925,
+   926,
+   927,
+   928,
+   929,
+   930,
+   931,
+   932,
+   933,
+   934,
+   935,
+   936,
+   937,
+   938,
+   939,
+   941,
+   942,
+   943,
+   944,
+   945,
+   946,
+   947,
+   948,
+   949,
+   950,
+   953,
+   954,
+   955,
+   956,
+   957,
+   958,
+   959,
+   960,
+   961,
+   962,
+   963,
+   964,
+   965,
+   966,
+   967,
+   968,
+   970,
+   971,
+   972,
+   973,
+   974,
+   975,
+   976,
+   977,
+   978,
+   979,
+   980,
+   981,
+   982,
+   983,
+   984,
+   985,
+   986,
+   987,
+   988,
+   990,
+   991,
+   993,
+   994,
+   995,
+   996,
+   997,
+   998,
+   999
+  ],
+  "held": [
+   2,
+   6,
+   9,
+   10,
+   19,
+   20,
+   21,
+   27,
+   39,
+   48,
+   49,
+   64,
+   70,
+   75,
+   76,
+   81,
+   82,
+   86,
+   89,
+   107,
+   112,
+   115,
+   118,
+   120,
+   145,
+   179,
+   202,
+   204,
+   208,
+   209,
+   210,
+   212,
+   216,
+   223,
+   224,
+   230,
+   235,
+   241,
+   253,
+   263,
+   265,
+   266,
+   269,
+   278,
+   287,
+   299,
+   300,
+   307,
+   314,
+   316,
+   324,
+   339,
+   340,
+   346,
+   353,
+   356,
+   358,
+   383,
+   389,
+   395,
+   398,
+   406,
+   407,
+   414,
+   418,
+   424,
+   434,
+   437,
+   438,
+   440,
+   442,
+   443,
+   444,
+   451,
+   473,
+   476,
+   481,
+   487,
+   488,
+   511,
+   524,
+   528,
+   529,
+   545,
+   548,
+   556,
+   558,
+   569,
+   571,
+   576,
+   585,
+   586,
+   591,
+   600,
+   605,
+   643,
+   649,
+   651,
+   659,
+   667,
+   682,
+   684,
+   694,
+   695,
+   698,
+   699,
+   707,
+   708,
+   709,
+   713,
+   724,
+   727,
+   732,
+   733,
+   735,
+   738,
+   743,
+   757,
+   762,
+   768,
+   771,
+   772,
+   777,
+   783,
+   788,
+   793,
+   802,
+   820,
+   824,
+   829,
+   831,
+   838,
+   847,
+   853,
+   858,
+   859,
+   873,
+   878,
+   881,
+   901,
+   902,
+   904,
+   918,
+   921,
+   940,
+   951,
+   952,
+   969,
+   989,
+   992
+  ]
+ },
+ "grids": {
+  "train": [
+   "experiment5/27136",
+   "experiment5/27137",
+   "experiment5/27138",
+   "experiment5/27139",
+   "experiment5/27140",
+   "experiment5/27141",
+   "experiment5/27142",
+   "experiment5/27145",
+   "experiment5/27146",
+   "experiment5/27148",
+   "experiment5/27149",
+   "experiment5/27151",
+   "experiment5/27153",
+   "experiment5/27155",
+   "experiment5/27156",
+   "experiment5/27157",
+   "experiment5/27158",
+   "experiment5/27159",
+   "experiment5/27160",
+   "experiment5/27161",
+   "experiment5/27162",
+   "experiment5/27164",
+   "experiment5/27166",
+   "experiment5/27173",
+   "experiment5/27174",
+   "experiment5/27176",
+   "experiment5/27177",
+   "experiment5/27178",
+   "experiment5/27180",
+   "experiment5/27181",
+   "experiment5/27182",
+   "experiment5/27184",
+   "experiment5/27186",
+   "experiment5/27188",
+   "experiment5/27191",
+   "experiment5/27198",
+   "experiment5/27199",
+   "experiment5/27200",
+   "experiment5/27201",
+   "experiment5/27203",
+   "experiment5/27209",
+   "experiment5/27210",
+   "experiment5/27215",
+   "experiment5/27218",
+   "experiment5/27222",
+   "experiment5/27226",
+   "experiment5/27227",
+   "experiment5/27229",
+   "experiment5/27235",
+   "experiment5/27240",
+   "experiment5/27245",
+   "experiment5/27251",
+   "experiment5/27253",
+   "experiment5/27254",
+   "experiment5/27255",
+   "experiment5/27256",
+   "experiment5/27258",
+   "experiment5/27262",
+   "experiment5/27273",
+   "experiment5/27278",
+   "experiment5/27283",
+   "experiment5/27285",
+   "experiment5/27286",
+   "experiment5/27289",
+   "experiment5/27293",
+   "experiment5/27310",
+   "experiment5/27316",
+   "experiment5/27328",
+   "experiment5/27334",
+   "experiment5/27339",
+   "experiment5/27360",
+   "experiment5/27362",
+   "experiment5/27365",
+   "experiment5/27376",
+   "experiment5/27398",
+   "experiment5/27403",
+   "experiment5/27421",
+   "experiment5/27423",
+   "experiment5/27445",
+   "experiment5/27459",
+   "experiment5/27490",
+   "experiment5/27496",
+   "experiment5/27601",
+   "experiment5/27829"
+  ],
+  "held": [
+   "experiment5/27147",
+   "experiment5/27152",
+   "experiment5/27163",
+   "experiment5/27165",
+   "experiment5/27167",
+   "experiment5/27168",
+   "experiment5/27169",
+   "experiment5/27195",
+   "experiment5/27213",
+   "experiment5/27243",
+   "experiment5/27247",
+   "experiment5/27260",
+   "experiment5/27265",
+   "experiment5/27324",
+   "experiment5/27449"
+  ]
+ }
+}