Fixed log_param error (#1362)

* fix_qrun_error

* add_description

qlib/workflow/recorder.py

@@ -21,6 +21,8 @@ from ..log import TimeInspector, get_module_logger
 from mlflow.store.artifact.azure_blob_artifact_repo import AzureBlobArtifactRepository
 
 logger = get_module_logger("workflow", logging.INFO)
+# mlflow limits the length of log_param to 500, but this caused errors when using qrun, so we extended the mlflow limit.
+mlflow.utils.validation.MAX_PARAM_VAL_LENGTH = 1000
 
 
 class Recorder:

setup.py

@@ -62,6 +62,8 @@ REQUIRED = [
     "matplotlib>=3.3",
     "tables>=3.6.1",
     "pyyaml>=5.3.1",
+    # To ensure stable operation of the experiment manager, we have limited the version of mlflow,
+    # and we need to verify whether version 2.0 of mlflow can serve qlib properly.
     "mlflow>=1.12.1, <=1.30.0",
     "tqdm",
     "loguru",