Graph-feature-preprocessed gradient boosting: artefacts + measured metrics

export_report.json

@@ -1,4 +1,5 @@
 {
+  "browser_runnable": true,
   "fp32_bytes": 4606582,
   "fp32_max_abs_error": 2.384185791015625e-07,
   "fp32_path": "onnx/model.onnx",
@@ -10,6 +11,9 @@
   "int8_path": "",
   "int8_skipped_reason": "tree ensemble: parameters live in node attributes, not weight tensors, so dynamic int8 quantisation has nothing to quantise",
   "n_fixtures": 512,
+  "onnx_domains": [
+    "ai.onnx.ml"
+  ],
   "opset": 0,
   "passed": true
 }