Upload 11 files

README.md

@@ -1,10 +1,10 @@
 ---
-title: Mc Map Resolver
-emoji: 🐠
-colorFrom: purple
-colorTo: blue
+title: MC Map Captcha Resolver
+emoji: 📈
+colorFrom: blue
+colorTo: pink
 sdk: gradio
-sdk_version: 4.31.0
+sdk_version: 3.41.2
 app_file: app.py
 pinned: false
 license: mit

app.py

@@ -0,0 +1,103 @@
+import gradio as gr
+import io
+def solve_captcha(data_bytes: bytes = None, file_name: str = None):
+    import time
+    import os
+    import sys
+    import cv2
+    import numpy as np
+    import onnxruntime
+
+
+    real_tests = False
+    from config import characters, img_height, img_width, img_type, max_length, transpose_perm, OUTPUT_ONNX
+
+    sess = onnxruntime.InferenceSession(f"out.model.onnx")
+    name = sess.get_inputs()[0].name
+    def get_result(pred):
+        """CTC decoder of the output tensor
+        https://distill.pub/2017/ctc/
+        https://en.wikipedia.org/wiki/Connectionist_temporal_classification
+        :return string, float
+        """
+        accuracy = 1
+        last = None
+        ans = []
+        # pred - 3d tensor, we need 2d array - first element
+        for item in pred[0]:
+            # get index of element with max accuracy
+            char_ind = item.argmax()
+            # ignore duplicates and special characters
+            if char_ind != last and char_ind != 0 and char_ind != len(characters)+1:
+                # this element is a character - append it to answer
+                ans.append(characters[char_ind - 1])
+                # Get accuracy for current character and multiply global accuracy by it
+                accuracy *= item[char_ind]
+            last = char_ind
+
+        answ = "".join(ans)[:max_length]
+        return answ, accuracy
+
+
+    def decode_img(data_bytes: bytes):
+        # same actions, as for tensorflow
+        image = cv2.imdecode(np.asarray(bytearray(data_bytes), dtype=np.uint8), 1)
+        image: "np.ndarray" = image.astype(np.float32) / 255.
+        if image.shape != (img_height, img_width, 3):
+            image = cv2.resize(image, (img_width, img_height))
+        image = image.transpose(transpose_perm)
+        #  Creating tensor ( adding 4d dimension )
+        image = np.array([image])
+        return image
+        
+    def decode_img_array(nump_array):
+        # same actions, as for tensorflow
+        #image = cv2.imdecode(nump_array, 1)
+        image: "np.ndarray" = nump_array.astype(np.float32) / 255.
+        if image.shape != (img_height, img_width, 3):
+            image = cv2.resize(image, (img_width, img_height))
+        image = image.transpose(transpose_perm)
+        #  Creating tensor ( adding 4d dimension )
+        image = np.array([image])
+        return image
+
+    def solve(data_bytes: bytes=None, file_name=None):
+        if file_name:
+            with open(file_name, 'rb') as F:
+                data_bytes = F.read()
+        if data_bytes is None:
+            print('[CAPTCHA RESOLVER NN] ПУСТОТА ВМЕСТО БАЙТОВ!')
+            return None
+        if isinstance(data_bytes,np.ndarray):
+            print('Img is ndarray')
+            img = decode_img_array(data_bytes)
+        else:
+            print('Img is bytes')
+            img_byte_arr = io.BytesIO()
+            data_bytes.save(img_byte_arr, format='PNG')
+            img_byte_arr = img_byte_arr.getvalue()
+            img = decode_img(data_bytes)
+        #print(img)
+        pred_onx = sess.run(None, {name: img})[0]
+        ans = get_result(pred_onx)
+        return ans
+    result = solve(data_bytes,file_name)
+    print ('solved', result)
+    return {"result":result[0], "predict":result[1]}
+    
+def image_classifier(inp):
+    if inp is None:
+        return "Не загружена картинка"
+    print('INPUT GOT>>>', inp)
+    #gradio img inp - numpy ndarray
+    
+    
+    result = solve_captcha(inp)
+    return result
+#inputs=gr.Image(type="pil")
+examples_names = ["1089","3569","5989","18769","73295","89238"]
+examples = []
+for n in examples_names:
+    examples.append("examples/"+n+".png")
+demo = gr.Interface(fn=image_classifier, inputs=gr.Image(), outputs="textbox", examples=examples)
+demo.launch(show_api = True)

config.py

@@ -0,0 +1,30 @@
+img_type = "*.png" # "*.jpeg"  # "*.png"
+characters = ['0', '1', '2', '3', '4', '5', '6', '7', '8', '9']
+# Desired image dimensions
+img_width = 70#130
+img_height = 50#50
+# Maximum length of any captcha in the dataset
+max_length = 5
+
+MODEL_FNAME_TRANING = "output/output.traning2.model"
+MODEL_FNAME = "output/output2.model"
+OUTPUT_ONNX = "output/out.model.onnx"
+
+# Training config
+batch_size = 16 # 16 #48 мое норм 77%
+epochs = 100 # 400мое норм 77
+early_stopping_patience = 20 #было 10 #40 мое норм 77%
+train_ratio = 0.8
+
+# соотношение TRAIN IMAGES к TEST IMAGES, 0.9 означает 0.9 для трейн и 0.1 для тест.
+# Ставь 1.0 если для теста юзать только те что в папке тест
+
+transpose_perm = [1, 0, 2]
+# батч епохи патиенс
+
+# 1
+# 48 400 40 нормас 77% вроде 170 эпоха
+
+# 2
+#96 800 80
+# loss 0.2-0.1 на 200 эпохе; 300 эпоха стабильно 0.1  потом меньше 0.1

out.model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22fce976f92bc8abe8ee3b1f4fcd3a1032e17cf006bca2b88624a332ef4031dc
+size 1750703

requirements.txt

@@ -0,0 +1,4 @@
+onnx
+onnxruntime
+opencv-python
+numpy