Add metadata for repo root

README.md

@@ -0,0 +1,22 @@
+# YeMoKoo/flamedata
+
+Exact tokenized train/test splits used for the FLAME-MoE continual-learning experiments.
+
+Repository layout:
+- `wiki/train`
+- `wiki/test`
+- `code/train`
+- `code/test`
+
+Dataset entries:
+
+| Folder | Summary |
+| --- | --- |
+| `wiki/train` | Exact wiki train split: first 1800 steps worth of tokens. |
+| `wiki/test` | Exact wiki test split: remainder after the first 1800 steps. |
+| `code/train` | Exact code train split: first 1800 steps worth of tokens. |
+| `code/test` | Exact code test split trimmed to match the wiki test token count. |
+
+These are Megatron indexed-dataset artifacts, not Hugging Face `datasets` parquet exports.
+
+Source code: https://github.com/YeMoKoo/FLAME-MoE

dataset_manifest.json

@@ -0,0 +1,30 @@
+[
+  {
+    "key": "wiki_train",
+    "title": "Wiki Train Exact",
+    "source_dir": ".local/dataset/wikipedia-full/tokenized/EleutherAI/pythia-12b-step1800-train-exact",
+    "path_in_repo": "wiki/train",
+    "summary": "Exact wiki train split: first 1800 steps worth of tokens."
+  },
+  {
+    "key": "wiki_test",
+    "title": "Wiki Test Full Remainder Exact",
+    "source_dir": ".local/dataset/wikipedia-full/tokenized/EleutherAI/pythia-12b-step1800-test-fullremainder-exact",
+    "path_in_repo": "wiki/test",
+    "summary": "Exact wiki test split: remainder after the first 1800 steps."
+  },
+  {
+    "key": "code_train",
+    "title": "Code Train Exact",
+    "source_dir": ".local/dataset/python-code-full/tokenized/EleutherAI/pythia-12b-step1800-train-exact",
+    "path_in_repo": "code/train",
+    "summary": "Exact code train split: first 1800 steps worth of tokens."
+  },
+  {
+    "key": "code_test",
+    "title": "Code Test Match Wiki Exact",
+    "source_dir": ".local/dataset/python-code-full/tokenized/EleutherAI/pythia-12b-step1800-test-matchwiki-exact",
+    "path_in_repo": "code/test",
+    "summary": "Exact code test split trimmed to match the wiki test token count."
+  }
+]