added some slot value corrections & different num of epochs in training scripts

data/prompt-learning/100-dpd/train.soloist.json

@@ -14879,7 +14879,7 @@
       "taxi"
     ],
     "belief_states": [
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -14892,7 +14892,7 @@
       "taxi"
     ],
     "belief_states": [
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },
@@ -14909,7 +14909,7 @@
     ],
     "belief_states": [
       "destination = loch fyne",
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },
@@ -14928,7 +14928,7 @@
     ],
     "belief_states": [
       "destination = loch fyne",
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },

data/prompt-learning/250-dpd/train.soloist.json

@@ -16037,7 +16037,7 @@
     ],
     "belief_states": [
       "destination = pizza hut fen ditton",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -16056,7 +16056,7 @@
     "belief_states": [
       "leave = 11:45",
       "destination = pizza hut fen ditton",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -16077,7 +16077,7 @@
     "belief_states": [
       "leave = 11:45",
       "destination = pizza hut fen ditton",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -32656,7 +32656,7 @@
       "taxi"
     ],
     "belief_states": [
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -32669,7 +32669,7 @@
       "taxi"
     ],
     "belief_states": [
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },
@@ -32686,7 +32686,7 @@
     ],
     "belief_states": [
       "destination = loch fyne",
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },
@@ -32705,7 +32705,7 @@
     ],
     "belief_states": [
       "destination = loch fyne",
-      "departure = saint catharines college",
+      "departure = saint catherines college",
       "arrive = 01:15"
     ]
   },
@@ -38563,7 +38563,7 @@
       "taxi"
     ],
     "belief_states": [
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -38577,7 +38577,7 @@
     ],
     "belief_states": [
       "destination = rosas bed and breakfast",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -38594,7 +38594,7 @@
     "belief_states": [
       "leave = 13:00",
       "destination = rosas bed and breakfast",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {
@@ -38613,7 +38613,7 @@
     "belief_states": [
       "leave = 13:00",
       "destination = rosas bed and breakfast",
-      "departure = saint catharines college"
+      "departure = saint catherines college"
     ]
   },
   {

prompt-learning/prompt_utils.py

@@ -12,7 +12,7 @@ PROMPT_TEMPLATES = {
         "generate": "belief states: value = $value, slot ="
     },
     "inverse-prompt": {
-        "training": "belief states: $slot = $value",
+        "training": "belief states: slot = $slot, value = $value",
     },
     "prompt-ensemble": {
         "training": {

prompt-learning/train_prompting.sh

@@ -46,15 +46,22 @@ datetime_now=$(date +"%Y%m%dT%H%M%S")
 experiment_folder="${data_split}"/experiment-${datetime_now}
 SAVE_DIR="${SAVED_MODELS_PROMPT}"/"${experiment_folder}"
 
-echo "Trained Models (checkpoints/epochs) are saved in ${SAVE_DIR}"
+echo "Trained Models (epochs) will be saved in ${SAVE_DIR}"
+
+# different number of epoch for different training sets
+if [ "$data_split" = "5-dpd" ] || [ "$data_split" = "10-dpd" ]; then
+  epochs=5
+else
+  epochs=8
+fi
 
 python prompt_train.py \
 --save_model_dir="${SAVE_DIR}" \
 --pretrained_model_path="${PRE_TRAINED_SOLOIST}" \
 --train_data_file="${TRAIN_DATA_FILE}"  \
 --validation_file=../data/prompt-learning/valid/valid.soloist.json \
---num_epochs 10 \
+--num_epochs $epochs \
 --learning_rate 5e-5 \
+--with_prompt_ensemble \
 --with_inverse_prompt \
---inverse_prompt_weight 0.1 \
---with_prompt_ensemble
+--inverse_prompt_weight 0.1

utils/create_dataset.py

@@ -39,7 +39,9 @@ CORRECTIONS = {
     "christs college": "christ college",
     "museums": "museum",
     "alexander": "alexander bed and breakfast",
-    "ian hong house": "lan hong house"
+    "ian hong house": "lan hong house",
+    "saint catharines college": "saint catherines college",
+    "gandhi": "the gandhi"
 }
 
 

utils/regexner.rules

@@ -43,7 +43,7 @@ stevenage train station	PLACE
 parkside police station	PLACE
 birmingham new street	PLACE
 jesus green outdoor pool	PLACE
-sheep's green and lammas land park fen causeway	PLACE
+sheeps green and lammas land park fen causeway	PLACE
 broxbourne	PLACE
 broxbourne train station	PLACE
 duxford	PLACE
@@ -57,7 +57,6 @@ huntingdon marriott hotel	PLACE
 glastonbury	PLACE
 city hall	PLACE
 hughes hall	PLACE
-city centre north	PLACE
 city centre north b and b	PLACE
 cafe jello gallery	PLACE
 the junction	PLACE
@@ -68,11 +67,11 @@ hobsons house	PLACE
 funky fun house	PLACE
 avalon	PLACE
 regency gallery	PLACE
+ruskin gallery	PLACE
 churchills college	PLACE
 christs college	PLACE
 holy trinity church	PLACE
 cineworld cinema	PLACE
-hotel	PLACE_TYPE
 guesthouse	PLACE_TYPE
 architecture	PLACE_TYPE
 boat	PLACE_TYPE
@@ -118,10 +117,13 @@ frankie and bennys	PLACE
 kymmoy	PLACE
 dojo noodle bar	PLACE
 the bedouin	PLACE
+mumford theatre	PLACE
 restaurant alimentum	PLACE
 gourmet burger kitchen	PLACE
 la margherita	PLACE
 golden house	PLACE
+kirkwood house	PLACE
+warkworth house	PLACE
 chiquito	PLACE
 darrys cookhouse and wine shop	PLACE
 scudamores punt	PLACE
@@ -149,6 +151,7 @@ rajmahal	PLACE
 ali baba	PLACE
 limehouse	PLACE
 the grafton hotel	PLACE
+lensfield hotel	PLACE
 barbakan	PLACE
 sesame restaurant and bar	PLACE
 golden wok	PLACE
@@ -179,7 +182,6 @@ european	PLACE
 saffron brasserie	PLACE
 gardenia	PLACE
 de luca cucina and bar	PLACE
-two two	PLACE
 ashley hotel	PLACE
 the hotpot	PLACE
 michaelhouse cafe	PLACE
@@ -224,6 +226,7 @@ sitar	PLACE
 alex	PLACE
 cambridge chop house	PLACE
 cambridge arts theatre	PLACE
+cambridge and county folk museum	PLACE
 the missing sock	PLACE
 primavera	PLACE
 the meze bar	PLACE
@@ -340,6 +343,7 @@ mexican	FOOD_TYPE
 middle eastern	FOOD_TYPE
 modern american	FOOD_TYPE
 modern english	FOOD_TYPE
+modern european	FOOD_TYPE
 moroccan	FOOD_TYPE
 north african	FOOD_TYPE
 north american	FOOD_TYPE