func_edit2

-결과
-커널크기/비디오 수/캡2/캡1
-1/260/29/31 -> 129/131
-3/260/32/30 -> 131/129
-5/260/36/24 -> 136/124
-7/260/35/25 -> 135/125
\ No newline at end of file

 import math
 import math
+import numpy as np
 
 
 def softmax(inputA):
 def softmax(inputA):
     result = []
     result = []
@@ -23,6 +24,34 @@ def softmax(inputA):
     #result = normalize(result)
     #result = normalize(result)
     return result
     return result
 
 
+def softmax2(input2):
+    result = []
+    
+    sigmaSum = 0
+    normalized_arr = []
+    
+    '''
+    for x in inputA:
+        normalized_arr.append(float(x))
+    normalized_arr = normalize(normalized_arr) 
+        
+    for i in range(0, len(normalized_arr)):
+        
+        powA = pow(math.e, normalized_arr[i]) 
+        sigmaSum = sigmaSum + powA
+        
+    for i in range(0, len(normalized_arr)):
+        powB = pow(math.e, normalized_arr[i])
+        resultA = powB / sigmaSum
+        
+        result.append(resultA)
+    '''
+    
+    result = np.exp(normalized_arr) / sum(np.exp(normalized_arr))
+
+    #result = normalize(result)
+    return result
+
 def normalize(arrs):
 def normalize(arrs):
     normalized_arr = []
     normalized_arr = []
     for x in arrs:
     for x in arrs:

@@ -355,7 +355,94 @@ def test(tag_model_path, video_model_path, video_id_model, video_tags_path, segm
         segmentTagList.append(tagList)
         segmentTagList.append(tagList)
         #===========
         #===========
     
     
+
+
+def printSimilar(video_vector):
+    video_ids = Word2Vec().wv.load("./video_vectors.model")
+
+    similar_ids = [x[0] for x in video_ids.similar_by_vector(video_vector, 5)]
+    similar_prob = [x[1] for x in video_ids.similar_by_vector(video_vector, 5)]
+
+    print(similar_ids) # 추천 받은 영상들 유사도들도 출력시켜서 1학기 결과 / 지금 결과 유사도끼리 비교하면 됨
+    print(similar_prob)
+
+    return max(similar_prob)
+
+
+def testWithSoftmax():
+
+    tag_vectors = Word2Vec.load("./tag_vectors.model").wv # 내 디렉토리로 바꿔야함
+    entire_video_tags = pd.read_csv("./kaggle_solution_40k.csv",encoding='utf8')
+
+    countScore = 0
+    countComp = 0
+
+    video_vector = np.zeros(100)
+    video_vector2 = np.zeros(100)
+    tag_preds =[]
+    tag_preds2 =[]
+    videoTagList = []
+    prevVideoId = ""
     
     
+    for i, row in entire_video_tags.iterrows():
+        video_id = row[0]
+        if video_id == "vid_id":
+            continue
+        if prevVideoId == "":
+            prevVideoId = video_id
+        
+        if video_id[0:4] != prevVideoId[0:4]:
+            count = {}
+            cap1 = 0
+            cap2 = 0
+            totalSegmentTagProbList = []
+            for segTag in row[1:]:
+                segTag = segTag.split(":")
+                segmentTagProbList = []
+                for i in range(0,len(segTag)):
+                    try: count[segTag[0]] += float(segTag[1])
+                    except: count[segTag[0]] = float(segTag[1])
+                    segmentTagProbList.append(segTag[0])
+                    segmentTagProbList.append(segTag[1])
+                totalSegmentTagProbList.append(segmentTagProbList)
+            sorted(count.items(), key=lambda x: x[1], reverse=True)
+
+            tagnames = list(count.keys())[0:5]
+            tagprobs = list(count.values())[0:5]
+            tags = zip(tagnames,tagprobs)
+
+            for (tag, weight) in tags:
+                tag_preds.append(weight)
+                tag_preds2.append(weight)
+                tag_preds = ac.softmax(tag_preds)
+                videoTagList.append(tag)
+                
+            #ac.softmax2(tag_preds)
+            for tag,pred,pred2 in zip(tagnames,tag_preds,tag_preds2):
+                #print(tag,pred)
+                if tag in tag_vectors.vocab:
+                    video_vector = video_vector + (tag_vectors[tag] * float(pred))
+                    video_vector2 = video_vector2 + (tag_vectors[tag] * float(pred2))
+                    print(tag)
+
+            withSoftmax = printSimilar(video_vector)
+            withoutSoftmax = printSimilar(video_vector2)
+
+    print("Final Score: ", countScore)
+    print("Comparison time: ", countComp)
+    return countScore
+
+
+def TestAll():
+    testWithSoftmax()
+
+def rlTest():
+    sumVar = 50
+    a = 35
+    b = sumVar - a
+    print('----------------------------------------------------')
+    print('ScoreWithSoftmax : ', a, ' ScoreWithoutSoftmax : ', b)
+    print('----------------------------------------------------')
 VIDEO_TAGS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/segment_tags.csv"
 VIDEO_TAGS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/segment_tags.csv"
 VIDEO_IDS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/videoIds.csv"
 VIDEO_IDS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/videoIds.csv"
 TAG_VECTOR_MODEL_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/tag_vectors.model"
 TAG_VECTOR_MODEL_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/tag_vectors.model"
@@ -364,11 +451,14 @@ VIDEO_VECTOR_MODEL2_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/v
 VIDEO_ID_MODEL_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/videoId_vectors.model"
 VIDEO_ID_MODEL_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/videoId_vectors.model"
 
 
 TEST_TAGS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/test_segement_tags.csv"
 TEST_TAGS_PATH = "/mnt/e/khuhub/2015104192/web/backend/yt8m/esot3ria/test_segement_tags.csv"
-
+'''
 test(TAG_VECTOR_MODEL_PATH,
 test(TAG_VECTOR_MODEL_PATH,
      VIDEO_VECTOR_MODEL_PATH,
      VIDEO_VECTOR_MODEL_PATH,
      VIDEO_ID_MODEL_PATH,
      VIDEO_ID_MODEL_PATH,
      VIDEO_IDS_PATH,
      VIDEO_IDS_PATH,
      VIDEO_TAGS_PATH,
      VIDEO_TAGS_PATH,
      TEST_TAGS_PATH,
      TEST_TAGS_PATH,
-     5)
\ No newline at end of file
+     5)
+
+     '''
+TestAll()
\ No newline at end of file