Upload manifest for synth_fr_500k_strat_v1

manifests/cpt_preparation_synth_fr_500k_strat_v1.json

@@ -0,0 +1,46 @@
+{
+  "dataset_id": "PleIAs/SYNTH",
+  "generated_at": "2026-03-12T10:53:32.448525+00:00",
+  "language": "fr",
+  "output_suffix": "synth_fr_500k_strat_v1",
+  "outputs": {
+    "eval": "data/processed/eval_cpt_fr_synth_fr_500k_strat_v1.jsonl",
+    "train": "data/processed/train_cpt_fr_synth_fr_500k_strat_v1.jsonl"
+  },
+  "seed": 42,
+  "selected_counts": {
+    "eval_examples": 10000,
+    "memorization_examples": 448866,
+    "non_memorization_examples": 51134,
+    "selected_per_exercise_counts": {
+      "constrained writing": 22553,
+      "creative writing": 22281,
+      "editing": 6300,
+      "memorization": 448866
+    },
+    "selected_total_words": 183869938,
+    "total_examples": 500000,
+    "train_examples": 490000
+  },
+  "selection": {
+    "eval_ratio": 0.02,
+    "min_text_chars": null,
+    "strategy": "all_non_memorization_plus_exact_hash_sample_of_memorization",
+    "target_examples": 500000
+  },
+  "source_counts": {
+    "fr_memorization_rows": 2406528,
+    "fr_non_memorization_rows": 51134,
+    "fr_per_exercise_counts": {
+      "constrained writing": 22553,
+      "creative writing": 22281,
+      "editing": 6300,
+      "memorization": 2406528
+    },
+    "fr_rejected_rows_after_serialization": 0,
+    "fr_total_rows": 2457662,
+    "fr_total_words": 761321441
+  },
+  "split": "train",
+  "task": "cpt"
+}