Upload clm_backbone_5lang.meta.json with huggingface_hub

clm_backbone_5lang.meta.json

@@ -0,0 +1,52 @@
+{
+  "dataset": "allenai/c4",
+  "configs": {
+    "ko": "ko",
+    "en": "en",
+    "zh": "zh",
+    "ru": "ru",
+    "ja": "ja"
+  },
+  "license": "odc-by",
+  "gated": false,
+  "real_fraction": 1.0,
+  "is_sample": true,
+  "note": "BOUNDED streaming SAMPLE (mC4), not the full 60B-tok pretrain corpus (= step ③ H100). Balanced 5-way, round-robin interleaved.",
+  "preferred_blocked": "uonlp/CulturaX is GATED; this token has no data access (403 DatasetNotFoundError on stream) — fell back to allenai/c4.",
+  "vocab": 256,
+  "encoding": "byte-utf8",
+  "langs": [
+    "ko",
+    "en",
+    "zh",
+    "ru",
+    "ja"
+  ],
+  "per_lang_docs_target": 4000,
+  "per_lang_docs_fetched": {
+    "ko": 4000,
+    "en": 4000,
+    "zh": 4000,
+    "ru": 4000,
+    "ja": 4000
+  },
+  "per_lang_docs_balanced": 4000,
+  "per_lang_lines_written": {
+    "ko": 4000,
+    "en": 4000,
+    "zh": 4000,
+    "ru": 4000,
+    "ja": 4000
+  },
+  "per_lang_bytes_written": {
+    "ko": 16679968,
+    "en": 7739055,
+    "zh": 9317034,
+    "ru": 14529169,
+    "ja": 19468200
+  },
+  "total_lines": 20000,
+  "total_bytes": 67733426,
+  "max_line_bytes": 8192,
+  "corpus_sha256": "7cc86c9afd5347920fb40be48d8f22b17ca2f72e04ea172f2496b13f42df3b04"
+}