Initialize results dataset

README.md

@@ -0,0 +1,39 @@
+# French Medical NLP Leaderboard - Results Dataset
+
+This dataset contains evaluation results for the French Medical NLP Leaderboard.
+
+## Results Format
+
+Each results file should be a JSON file with the following structure:
+
+```json
+{
+  "config": {
+    "model_dtype": "float16",
+    "model_name": "model_name",
+    "model_sha": "revision"
+  },
+  "results": {
+    "emea_ner": {
+      "f1": 0.85,
+      "precision": 0.83,
+      "recall": 0.87
+    },
+    "medline_ner": {
+      "f1": 0.82,
+      "precision": 0.80,
+      "recall": 0.84
+    }
+  }
+}
+```
+
+## Tasks
+- `emea_ner`: French medical NER on EMEA texts
+- `medline_ner`: French medical NER on MEDLINE abstracts
+
+## Metrics
+All metrics use seqeval with IOB2 scheme:
+- `f1`: Micro F1 score
+- `precision`: Micro precision
+- `recall`: Micro recall

sample_results.json

@@ -0,0 +1,19 @@
+{
+  "config": {
+    "model_dtype": "float16",
+    "model_name": "almanach/camembert-bio-base",
+    "model_sha": "main"
+  },
+  "results": {
+    "emea_ner": {
+      "f1": 0.006,
+      "precision": 0.0035,
+      "recall": 0.0198
+    },
+    "medline_ner": {
+      "f1": 0.0089,
+      "precision": 0.0051,
+      "recall": 0.0234
+    }
+  }
+}