dnn test

recommend-model-produce/src/main/python/models/dnn/config.yaml

@@ -35,7 +35,8 @@ runner:
   #use inference save model
   # model_init_path: "output_model_dnn/2" # init model
   use_inference: True
-  save_inference_feed_varnames: ["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","dense_input"]
+  #save_inference_feed_varnames: ["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","dense_input"]
+  save_inference_feed_varnames: ["C1","C2","C3","dense_input"]
   save_inference_fetch_varnames: ["label", "SparseFeatFactors"]
 
   # distribute_config

recommend-model-produce/src/main/python/tools/static_trainer.py

@@ -260,7 +260,7 @@ def dataset_train(epoch_id, dataset, fetch_vars, exe, config):
     fetch_info = [
         "Epoch {} Var {}".format(epoch_id, var_name) for var_name in fetch_vars
     ]
-    logger.info("var_name: {}".format(var_name)) for var_name in fetch_vars
+    logger.info("var_name: {}".format(fetch_vars)
     fetch_vars = [var for _, var in fetch_vars.items()]
     print_interval = config.get("runner.print_interval")
     exe.train_from_dataset(