Publish exact DistillationProject training split and provenance

README.md

@@ -0,0 +1,28 @@
+---
+language:
+- en
+task_categories:
+- text-generation
+tags:
+- math
+- sft
+- distillation-project
+source_datasets:
+- jasonrqh/Countdown-CoT-20k
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-10000.parquet
+  - split: test
+    path: data/test-500.parquet
+license: mit
+---
+
+# DistillationProject: Countdown 10K training set
+
+The exact two-phase Countdown curriculum shared by the nine Numina5K through Numina45K → Countdown10K models. The first 5,000 train rows reproduce the first Countdown phase in its actual loader order; the next 5,000 reproduce the continuation phase. The continuation used a fresh optimizer and schedule, so this dataset does not imply a single uninterrupted 10K-epoch optimizer schedule. The 500-puzzle test split is the unchanged separate Countdown benchmark and is disjoint from both train phases. Byte-identical copies of both original prepared 5K parquets are included under provenance. Training puzzle final expressions were checked with the original safe arithmetic verifier and input multiplicity rules; explanatory reasoning was not formally verified.
+
+Exact original question and response text is preserved. No resampling, solution rewriting or truncation is performed for publication. Source revisions, row counts, order and file SHA256 values are recorded in `publication.json` and `provenance/`. Prepared tables retain their original columns. Source validity labels do not establish independently verified mathematical correctness.
+
+Collection: https://huggingface.co/collections/zbeeb/disstillation-project-6ace05e5f40dbe715407516c

data/test-500.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45
+size 1002048

data/train-10000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39a5e9d2af63b42ad0f1937e4f0e69f5d6e23baab00f57231c8bafaf1b6fec6d
+size 13838358

provenance/countdown_new/manifest.json

@@ -0,0 +1,67 @@
+{
+  "at": "2026-10-14T13:51:39.630987+00:00",
+  "data_config_sha256": "57fcf454387826aa44cb173399a6f2372e5f464a88dc4a869947b1a95ad12574",
+  "dataset": {
+    "id": "jasonrqh/Countdown-CoT-20k",
+    "revision": "f5464c1062e8f90cf8103f9016bdde39d91e3bb4",
+    "split": "train",
+    "available_rows": 20480,
+    "files": [
+      "Countdown-CoT-20k.parquet"
+    ],
+    "local_source_files": {
+      "Countdown-CoT-20k.parquet": {
+        "path": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT/data/raw/Countdown-CoT-20k.parquet",
+        "sha256": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf"
+      }
+    }
+  },
+  "examples": 5000,
+  "test_examples": 500,
+  "previous_examples": 5000,
+  "cumulative_examples": 10000,
+  "files_sha256": {
+    "data/sft-5000.parquet": "a89fc5f71ebc1ab02da937b7331e3123967fdaacb5e290e86b42dc6da024e643",
+    "data/train.arrow": "18c22002d0ce02570cae7ce2d73ad76c8b95e981eed2f352f4683e74f4974597",
+    "data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "data/prompt-order.json": "a875bb0a59c856c9a18a5bb6ac52c9e1dec14c7168551e0f078b3e3bcaf92887"
+  },
+  "external_sources_sha256": {
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina5K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina5K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina5K/data/manifest.json": "bcf8476f5ff500d1e768582283c7d114efec6ccbaaf34f0100d2f728b6347a75",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K/data/manifest.json": "b58ce20bddbdbd8e53b259cd4497611662e2cfb79d681763c0cb41a03ee176a9",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina15K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina15K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina15K/data/manifest.json": "cf0335c5d7c7161706d7e4cb291f1dda88613d45e4293b278ff395adcb878789",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina20K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina20K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina20K/data/manifest.json": "0dcbc1d0f26c4d6e3dce187dc339f32fab7b5fa37e54618afbde46faa4cf5079",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina25K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina25K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina25K/data/manifest.json": "5e9503e1aba6888afcb7b4f99c3c8fe5a9b8cf4a37565f9abe7b882ee6efd93b",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina30K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina30K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina30K/data/manifest.json": "fd7e085c9b41bf6131bbf003fe6d806183a6c5db158cecc22d407e7b99e9f7a8",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina35K/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina35K/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina35K/data/manifest.json": "6fdc393db5a624768ac8bd2fe4222d8fbd1711a4c55de5c8231d61d9820ed0c9",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina40K-HP/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina40K-HP/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina40K-HP/data/manifest.json": "af801f597fa1cefc144375548259a7578cc478724cb4a3e0fb31c96f296c083b",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina45K-HP/data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina45K-HP/data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina45K-HP/data/manifest.json": "b8de7d17c1d9d2e49642e80038e2718c7700970acf88b814373213d6de3f0f64",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT/data/raw/Countdown-CoT-20k.parquet": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf"
+  },
+  "ordered_train_prompt_ids_sha256": "0744f5ee618b45003fa89bcf4793397ff83536aeb87815644272f5d23cfec753",
+  "original_train_prompt_ids_sha256": "087e6c4bd559d1b121e159b0416b2bf17ecb6715a5672de98da3b7f33343c1c5",
+  "test_prompt_ids_sha256": "02e34f8bee60377a7dc3af0ad2398803fdc82aeb91cafe657887f86733a793a7",
+  "source_question_response_text": "unchanged",
+  "selection": "seed42_hash_priority_then_shared_seed42_shuffle",
+  "truncated_examples": 0,
+  "overlap_previous5000": 0,
+  "overlap_test500": 0
+}

provenance/countdown_new/selection_config.json

@@ -0,0 +1,48 @@
+{
+  "dataset": {
+    "id": "jasonrqh/Countdown-CoT-20k",
+    "revision": "f5464c1062e8f90cf8103f9016bdde39d91e3bb4",
+    "split": "train",
+    "available_rows": 20480,
+    "files": [
+      "Countdown-CoT-20k.parquet"
+    ],
+    "local_source_files": {
+      "Countdown-CoT-20k.parquet": {
+        "path": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT/data/raw/Countdown-CoT-20k.parquet",
+        "sha256": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf"
+      }
+    }
+  },
+  "selection": {
+    "examples": 5000,
+    "test_examples": 500,
+    "seed": 42,
+    "maximum_prompt_tokens": 1024,
+    "overlength_policy": "reject",
+    "preserve_raw_text": true,
+    "deduplication": "target and sorted multiset of input numbers",
+    "test_split": "separate seeded puzzle split before training selection",
+    "previous_examples": 5000
+  },
+  "training": {
+    "maximum_sequence_tokens": 4096,
+    "system_prompt": "Solve the mathematical problem and explain your reasoning."
+  },
+  "original_projects": [
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina5K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina15K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina20K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina25K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina30K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina35K",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina40K-HP",
+    "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina45K-HP"
+  ],
+  "canonical_project": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K",
+  "expected_test_sha256": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+  "tokenizer_source": "/mnt/nfs/home/mohamadzbib/projects/DistillationProject/models/base",
+  "order": "seed42_hash_priority_then_shared_seed42_shuffle",
+  "disjoint_policy": "target_and_sorted_input_multiset"
+}

provenance/countdown_new/training_loader.py

@@ -0,0 +1,89 @@
+import os
+import random
+from collections import defaultdict
+
+from core import ROOT, Plan, completion_mask, config, digest, world_size
+
+
+class PreparedLoader:
+    def __init__(self):
+        import pyarrow as pa
+
+        self.cfg = config()
+        self.rank = int(os.environ["RANK"])
+        self.world = int(os.environ["WORLD_SIZE"])
+        expected = world_size(self.cfg)
+        if self.world != expected:
+            raise ValueError("Actual distributed world differs from the prepared topology")
+        training = self.cfg["training"]
+        self.plan = Plan(self.cfg["selection"]["examples"], training["snapshot_examples"], training["global_batch_size"], self.world)
+        self.mapping = pa.memory_map(str(ROOT / "data/train.arrow"), "r")
+        self.table = pa.ipc.open_file(self.mapping).read_all()
+        if len(self.table) != self.plan.examples:
+            raise ValueError("Prepared example count differs")
+        self.order = list(range(len(self.table)))
+        self.cursor = 0
+        self.num_samples, self.num_tokens = defaultdict(int), defaultdict(int)
+        self.input_tokens = 0
+        self.dataset = self
+        self.manifest_hash = digest(ROOT / "data/manifest.json")
+
+    @property
+    def step(self):
+        return self.plan.processed(self.cursor)
+
+    @property
+    def epoch(self):
+        return self.step / self.plan.examples
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        import torch
+
+        if self.cursor == self.plan.total_steps * self.plan.accumulation:
+            raise StopIteration
+        index = self.plan.index(self.cursor, self.rank)
+        dummy = index is None
+        row = self.table.slice(self.order[0 if dummy else index], 1).to_pylist()[0]
+        tokens, prefix = row["input_ids"], row["prompt_length"]
+        if not 0 < prefix < len(tokens) <= self.cfg["training"]["maximum_sequence_tokens"]:
+            raise ValueError("Prepared completion boundary or sequence length is invalid")
+        length = len(tokens) - 1
+        mask = torch.tensor(completion_mask(tokens, prefix, dummy), dtype=torch.bool)
+        if not dummy:
+            self.num_samples["countdown"] += 1
+            self.num_tokens["countdown"] += int(mask.sum())
+            self.input_tokens += length
+        self.cursor += 1
+        return {"input_ids": torch.tensor(tokens[:-1], dtype=torch.long).unsqueeze(0),
+                "target_ids": torch.tensor(tokens[1:], dtype=torch.long).unsqueeze(0),
+                "position_ids": torch.arange(length, dtype=torch.long).unsqueeze(0),
+                "loss_mask": mask.unsqueeze(0), "seq_lens": torch.tensor([length], dtype=torch.long)}
+
+    def state_dict(self):
+        import numpy as np
+        import torch
+
+        return {"cursor": self.cursor, "rank": self.rank, "world": self.world, "manifest_sha256": self.manifest_hash,
+                "cpu_rng": torch.get_rng_state(), "cuda_rng": torch.cuda.get_rng_state(),
+                "python_rng": random.getstate(), "numpy_rng": np.random.get_state(),
+                "num_samples": dict(self.num_samples), "num_tokens": dict(self.num_tokens), "input_tokens": self.input_tokens}
+
+    def load_state_dict(self, state):
+        import numpy as np
+        import torch
+
+        if (state["rank"], state["world"], state["manifest_sha256"]) != (self.rank, self.world, self.manifest_hash):
+            raise ValueError("Resume rank, topology or dataset contract differs; no fallback to rank zero")
+        if not 0 <= state["cursor"] <= self.plan.total_steps * self.plan.accumulation or state["cursor"] % self.plan.accumulation:
+            raise ValueError("Saved cursor is not at an optimizer boundary")
+        self.cursor = state["cursor"]
+        self.num_samples.update(state["num_samples"])
+        self.num_tokens.update(state["num_tokens"])
+        self.input_tokens = state["input_tokens"]
+        torch.set_rng_state(state["cpu_rng"])
+        torch.cuda.set_rng_state(state["cuda_rng"])
+        random.setstate(state["python_rng"])
+        np.random.set_state(state["numpy_rng"])

provenance/countdown_previous/manifest.json

@@ -0,0 +1,51 @@
+{
+  "at": "2026-10-13T16:03:09.259536+00:00",
+  "experiment_sha256": "8d7185002f088f425ca0b684b6cf02f745460daaf55fc578421eb6f6cc2854bc",
+  "files_sha256": {
+    "data/sft-5000.parquet": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+    "data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+    "data/train.arrow": "5a84490fccbf2a5fb03bcfbe1d4dfaac1f144f07ffaedda9b93712a1d2739442",
+    "data/raw/Countdown-CoT-20k.parquet": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf",
+    "models/base/README.md": "79a4bd77728fd3a3d26d6fc976f8f99f404b7991c3a1e78c8e983f686d4d72c3",
+    "models/base/export.json": "ac5abf7907f7e6e515b3021eceaa518185ceaf9cb736e2b558bbaf6b15c8d39a",
+    "models/base/model-00002-of-00002.safetensors": "244d89a2a1d713642c72cb92cb4810d3c684f5c76414f2ef779991193d89c87f",
+    "models/base/model.safetensors.index.json": "933e30035b9c3ef2b1bfc488c644f8124f0b54695e246ce7af74390de64217a0",
+    "models/base/config.json": "40581758f4416a4efb80079aebdb2aff7bd9abda5bbfc95aae383830a7d570e6",
+    "models/base/model-00001-of-00002.safetensors": "0b1c37560c85d4e48f06ef40984f1e39a37f78a089a642859fd7183fd5f8e162",
+    "models/base/tokenizer_config.json": "cbea7bca9904d56693d2226b00fd564e3be053609d1e50f8f1885510f0f6e790",
+    "models/base/chat_template.jinja": "44d5f08f3f72b837eaad09f13a54c1f9f4eb58d75240334548b7fd52a5437fa5",
+    "models/base/generation_config.json": "7553c0106ada1290d2d9501159b74d9381312eeb8dfa14414c0466d6420f2eaf",
+    "models/base/tokenizer.json": "3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8"
+  },
+  "examples": 5000,
+  "test_examples": 500,
+  "optimizer_steps": 80,
+  "snapshots": [
+    {
+      "optimizer_step": 16,
+      "examples_processed": 1000
+    },
+    {
+      "optimizer_step": 32,
+      "examples_processed": 2000
+    },
+    {
+      "optimizer_step": 48,
+      "examples_processed": 3000
+    },
+    {
+      "optimizer_step": 64,
+      "examples_processed": 4000
+    },
+    {
+      "optimizer_step": 80,
+      "examples_processed": 5000
+    }
+  ],
+  "raw_dataset_sha256": {
+    "data/raw/Countdown-CoT-20k.parquet": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf"
+  },
+  "selected_prompt_ids_sha256": "087e6c4bd559d1b121e159b0416b2bf17ecb6715a5672de98da3b7f33343c1c5",
+  "test_prompt_ids_sha256": "02e34f8bee60377a7dc3af0ad2398803fdc82aeb91cafe657887f86733a793a7",
+  "selection_method": "seeded hash priority over verified unique puzzles; reserve first 500 for test, then select complete train responses fitting context"
+}

provenance/countdown_previous/selection_config.json

@@ -0,0 +1,101 @@
+{
+  "model": {
+    "id": "local:Qwen2.5-1.5B-NuminaMath-SFT-10k",
+    "revision": "483643b2336275849a0849bc970bcd1c05503b93bd57eb76e8869aadd6de7885",
+    "local_source": "/mnt/nfs/home/mohamadzbib/projects/DistillationProject/models/snapshots/samples_010000",
+    "source_optimizer_step": 158,
+    "source_examples": 10000,
+    "source_contract": "eeb34a30b8666f32f58755d2ad3f1d969f79cbd0d90ed7d4ecbf13b100d0ac14",
+    "revision_type": "native_export_metadata_sha256"
+  },
+  "dataset": {
+    "id": "jasonrqh/Countdown-CoT-20k",
+    "revision": "f5464c1062e8f90cf8103f9016bdde39d91e3bb4",
+    "split": "train",
+    "available_rows": 20480,
+    "files": [
+      "Countdown-CoT-20k.parquet"
+    ],
+    "local_source_files": {
+      "Countdown-CoT-20k.parquet": {
+        "path": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT/data/raw/Countdown-CoT-20k.parquet",
+        "sha256": "98c4144be7b327ee835e62c91127f869ed72f342651150fce42b01e2042643cf"
+      }
+    }
+  },
+  "selection": {
+    "examples": 5000,
+    "test_examples": 500,
+    "seed": 42,
+    "maximum_prompt_tokens": 1024,
+    "overlength_policy": "reject",
+    "preserve_raw_text": true,
+    "deduplication": "target and sorted multiset of input numbers",
+    "test_split": "separate seeded puzzle split before training selection"
+  },
+  "training": {
+    "epochs": 1,
+    "learning_rate": 5e-06,
+    "global_batch_size": 64,
+    "micro_batch_size": 1,
+    "maximum_sequence_tokens": 4096,
+    "warmup_ratio": 0.03,
+    "scheduler": "cosine",
+    "weight_decay": 0.01,
+    "gradient_clip": 1.0,
+    "packing": false,
+    "snapshot_examples": 1000,
+    "keep_all_checkpoints": true,
+    "system_prompt": "Solve the mathematical problem and explain your reasoning.",
+    "loss": "completion_only_token_mean",
+    "finetuning": "full_parameters"
+  },
+  "topology": {
+    "nodes": 1,
+    "gpus_per_node": 8,
+    "minimum_gpu_memory_gib": 79,
+    "maximum_hp_gpus": 18,
+    "master_port": 29631,
+    "timeout_seconds": 600,
+    "node_gpu_counts": [
+      8
+    ],
+    "minimum_gpu_memory_gib_by_node": [
+      79
+    ],
+    "secondary_nodes": []
+  },
+  "runtime": {
+    "profile": "prime-rl",
+    "source_hashes": {
+      "prime_rl.trainer.sft.train": "ac5f90594634bf3c57d1770ba6a169ae45fc9f541e25222a27c2e1b0c818e298",
+      "prime_rl.trainer.ckpt": "c7488f3ded3b56bb0633b75b361054d9df135b723bf4d9dd0909672a39a9d7ee"
+    },
+    "expected_packages": {
+      "torch": "2.11.0+cu128",
+      "transformers": "5.6.2"
+    }
+  },
+  "benchmark_exclusions": [],
+  "launch_status": "additional separate 5000-example Countdown run authorized on 2026-10-13",
+  "scheduling": {
+    "account": "grad-students",
+    "partition": "low-priority",
+    "qos": "normal",
+    "nodes": [
+      "deep-chungus-9"
+    ],
+    "gpus_per_node": 8,
+    "exclusive": true,
+    "after_job": "2153058"
+  },
+  "comparison": {
+    "expected_data_sha256": {
+      "data/test-500.parquet": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45"
+    },
+    "test_prompt_ids_sha256": "02e34f8bee60377a7dc3af0ad2398803fdc82aeb91cafe657887f86733a793a7",
+    "prefix_examples": 2000,
+    "prefix_prompt_ids_sha256": "ed29c54a8c64201c9d639cbb6af94321b802213a2ad69832abadacb52b340c8b",
+    "same_size_reference_project": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina5K"
+  }
+}

provenance/countdown_previous/training_loader.py

@@ -0,0 +1,90 @@
+import os
+import random
+from collections import defaultdict
+
+from core import ROOT, Plan, completion_mask, config, digest, world_size
+
+
+class PreparedLoader:
+    def __init__(self):
+        import pyarrow as pa
+
+        self.cfg = config()
+        self.rank = int(os.environ["RANK"])
+        self.world = int(os.environ["WORLD_SIZE"])
+        expected = world_size(self.cfg)
+        if self.world != expected:
+            raise ValueError("Actual distributed world differs from the prepared topology")
+        training = self.cfg["training"]
+        self.plan = Plan(self.cfg["selection"]["examples"], training["snapshot_examples"], training["global_batch_size"], self.world)
+        self.mapping = pa.memory_map(str(ROOT / "data/train.arrow"), "r")
+        self.table = pa.ipc.open_file(self.mapping).read_all()
+        if len(self.table) != self.plan.examples:
+            raise ValueError("Prepared example count differs")
+        self.order = list(range(len(self.table)))
+        random.Random(self.cfg["selection"]["seed"]).shuffle(self.order)
+        self.cursor = 0
+        self.num_samples, self.num_tokens = defaultdict(int), defaultdict(int)
+        self.input_tokens = 0
+        self.dataset = self
+        self.manifest_hash = digest(ROOT / "data/manifest.json")
+
+    @property
+    def step(self):
+        return self.plan.processed(self.cursor)
+
+    @property
+    def epoch(self):
+        return self.step / self.plan.examples
+
+    def __iter__(self):
+        return self
+
+    def __next__(self):
+        import torch
+
+        if self.cursor == self.plan.total_steps * self.plan.accumulation:
+            raise StopIteration
+        index = self.plan.index(self.cursor, self.rank)
+        dummy = index is None
+        row = self.table.slice(self.order[0 if dummy else index], 1).to_pylist()[0]
+        tokens, prefix = row["input_ids"], row["prompt_length"]
+        if not 0 < prefix < len(tokens) <= self.cfg["training"]["maximum_sequence_tokens"]:
+            raise ValueError("Prepared completion boundary or sequence length is invalid")
+        length = len(tokens) - 1
+        mask = torch.tensor(completion_mask(tokens, prefix, dummy), dtype=torch.bool)
+        if not dummy:
+            self.num_samples["countdown"] += 1
+            self.num_tokens["countdown"] += int(mask.sum())
+            self.input_tokens += length
+        self.cursor += 1
+        return {"input_ids": torch.tensor(tokens[:-1], dtype=torch.long).unsqueeze(0),
+                "target_ids": torch.tensor(tokens[1:], dtype=torch.long).unsqueeze(0),
+                "position_ids": torch.arange(length, dtype=torch.long).unsqueeze(0),
+                "loss_mask": mask.unsqueeze(0), "seq_lens": torch.tensor([length], dtype=torch.long)}
+
+    def state_dict(self):
+        import numpy as np
+        import torch
+
+        return {"cursor": self.cursor, "rank": self.rank, "world": self.world, "manifest_sha256": self.manifest_hash,
+                "cpu_rng": torch.get_rng_state(), "cuda_rng": torch.cuda.get_rng_state(),
+                "python_rng": random.getstate(), "numpy_rng": np.random.get_state(),
+                "num_samples": dict(self.num_samples), "num_tokens": dict(self.num_tokens), "input_tokens": self.input_tokens}
+
+    def load_state_dict(self, state):
+        import numpy as np
+        import torch
+
+        if (state["rank"], state["world"], state["manifest_sha256"]) != (self.rank, self.world, self.manifest_hash):
+            raise ValueError("Resume rank, topology or dataset contract differs; no fallback to rank zero")
+        if not 0 <= state["cursor"] <= self.plan.total_steps * self.plan.accumulation or state["cursor"] % self.plan.accumulation:
+            raise ValueError("Saved cursor is not at an optimizer boundary")
+        self.cursor = state["cursor"]
+        self.num_samples.update(state["num_samples"])
+        self.num_tokens.update(state["num_tokens"])
+        self.input_tokens = state["input_tokens"]
+        torch.set_rng_state(state["cpu_rng"])
+        torch.cuda.set_rng_state(state["cuda_rng"])
+        random.setstate(state["python_rng"])
+        np.random.set_state(state["numpy_rng"])

provenance/phase1-prepared-5000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5
+size 7025077

provenance/phase2-prepared-5000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a89fc5f71ebc1ab02da937b7331e3123967fdaacb5e290e86b42dc6da024e643
+size 6833588

publication.json

@@ -0,0 +1,55 @@
+{
+  "publication_version": 1,
+  "raw_question_response_text": "unchanged",
+  "seed": 42,
+  "sources": {
+    "countdown_previous": {
+      "root": "/mnt/nfs/home/mohamadzbib/projects/CountdownSFT-5K-Numina10K",
+      "config": "configs/experiment.json",
+      "config_sha256": "8d7185002f088f425ca0b684b6cf02f745460daaf55fc578421eb6f6cc2854bc",
+      "manifest_sha256": "b58ce20bddbdbd8e53b259cd4497611662e2cfb79d681763c0cb41a03ee176a9",
+      "loader_sha256": "328ac1dc74339e30c6524912914c2fa13089f15888280596439bcddc9fa11add"
+    },
+    "countdown_new": {
+      "root": "/mnt/nfs/home/mohamadzbib/projects/DistillationProject/countdown-extension-20261014",
+      "config": "configs/data.json",
+      "config_sha256": "57fcf454387826aa44cb173399a6f2372e5f464a88dc4a869947b1a95ad12574",
+      "manifest_sha256": "35bd78d1fe50ae3f3f37cea14be9b42bb2d4712d01edc93d11d33a6a9daf2700",
+      "loader_sha256": "1ea26149570cb58eefb5d017c7bac0a72fd9920b021d00cd8103535f9522e5d6"
+    }
+  },
+  "details": {
+    "training_examples": 10000,
+    "phase1_examples": 5000,
+    "phase2_examples": 5000,
+    "test_examples": 500,
+    "phase_train_overlap": 0,
+    "train_test_overlap": 0,
+    "maximum_sequence_tokens": 4096,
+    "overlength_policy": "reject",
+    "training_order": "actual phase1 shuffled loader order, followed by phase2 identity loader order",
+    "continuation_optimizer": "fresh optimizer and cosine schedule for the second 5K phase"
+  },
+  "files": {
+    "data/train-10000.parquet": {
+      "sha256": "39a5e9d2af63b42ad0f1937e4f0e69f5d6e23baab00f57231c8bafaf1b6fec6d",
+      "bytes": 13838358,
+      "rows": 10000
+    },
+    "data/test-500.parquet": {
+      "sha256": "327c034439b7eabbb690ea9ca346cfa5f118d27685b59de68d21b98b46955f45",
+      "bytes": 1002048,
+      "rows": 500
+    },
+    "provenance/phase1-prepared-5000.parquet": {
+      "sha256": "ac3ca0e3f2d8cfd2b0d4f6eb219c7bac94502bce9be16ded8bb7e5284d7cabb5",
+      "bytes": 7025077,
+      "rows": 5000
+    },
+    "provenance/phase2-prepared-5000.parquet": {
+      "sha256": "a89fc5f71ebc1ab02da937b7331e3123967fdaacb5e290e86b42dc6da024e643",
+      "bytes": 6833588,
+      "rows": 5000
+    }
+  }
+}