stats making and reading files

aug.py

@@ -0,0 +1,33 @@
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+from os import listdir as ls
+from sys import argv, exit
+from tensorflow.keras.utils import array_to_img as img
+from tensorflow.keras import layers as kl
+
+if len(argv) != 3:
+    exit(1)
+
+files = ls(argv[1])
+
+for v, fn in enumerate(files):
+    i = np.array(Image.open(f"{argv[1]}/{fn}").convert('RGB').resize((300,300))) / 255.0
+    print(f"Processing {v+1:03d}/{len(files)}: {fn}")
+
+    for x1 in range(3):
+        i1 = tf.image.random_brightness(i, 0.6)
+        for x2 in range(3):
+            i2 = tf.image.random_saturation(i1, 0.1, 2.0)
+            for x3 in range(3):
+                i3 = tf.image.random_contrast(i2, 0.2, 1.9)
+                img(i3).save(f"{argv[2]}/{fn.rsplit('.', 1)[0]}_{x1}_{x2}_{x3}_1.png")
+
+                i3 = tf.image.flip_left_right(i3)
+                img(i3).save(f"{argv[2]}/{fn.rsplit('.', 1)[0]}_{x1}_{x2}_{x3}_2.png")
+
+                i3 = tf.image.flip_up_down(i3)
+                img(i3).save(f"{argv[2]}/{fn.rsplit('.', 1)[0]}_{x1}_{x2}_{x3}_3.png")
+
+                i3 = tf.image.flip_left_right(i3)
+                img(i3).save(f"{argv[2]}/{fn.rsplit('.', 1)[0]}_{x1}_{x2}_{x3}_4.png")

ident-live.py

@@ -0,0 +1,45 @@
+#from PIL import Image
+#from sys import argv, exit
+import numpy as np
+import cv2 as cv
+
+from tensorflow.keras.utils import array_to_img as img
+
+from matplotlib import pyplot as plt
+
+#if len(argv) != 2:
+#    exit(1)
+
+#i = np.array(Image.open(argv[1]).convert('RGB').resize((300,300)))
+
+from m import *
+mod.load_weights("model3-val-acc-0.74.model.keras")
+
+c = cv.VideoCapture(0)
+
+while True:
+    ret, frame = c.read()
+
+    pi = np.array(img(cv.cvtColor(frame, cv.COLOR_BGR2RGB)).resize((300, 300)).convert("RGB"))
+    #plt.imshow(pi)
+    #plt.show()
+
+    #print(pi.astype(float))
+
+    r = mod.predict(np.array([pi.astype(float)]))
+
+    cv.putText(frame, f"Y: {r[0][1]:.9f}", (10, 40), cv.FONT_HERSHEY_SIMPLEX, 1, (200,255,200), 2)
+    cv.putText(frame, f"N: {r[0][0]:.9f}", (10, 70), cv.FONT_HERSHEY_SIMPLEX, 1, (200,200,255), 2)
+
+    cv.imshow('me', frame)
+
+    if cv.waitKey(1) == ord('q'):
+        break
+
+#print(frame)
+
+c.release()
+cv.destroyAllWindows()
+
+#r = mod.predict(np.array([i]))
+#print(r)

ident.py

@@ -0,0 +1,15 @@
+from m import *
+from PIL import Image
+from sys import argv, exit
+import numpy as np
+
+
+if len(argv) != 2:
+    exit(1)
+
+i = np.array(Image.open(argv[1]).convert('RGB').resize((300,300)))
+
+mod.load_weights("model3.model.keras")
+
+r = mod.predict(np.array([i]))
+print(r)

m.py

@@ -0,0 +1,160 @@
+import tensorflow as tf
+import numpy as np
+
+from tensorflow import keras as k
+from tensorflow.keras import layers as kl
+from tensorflow.keras import models as km
+from tensorflow.keras import optimizers as ko
+from tensorflow.keras import callbacks as kc
+
+CONV_SIZE = 96
+MPPT_SIZE = 256
+DRPT_RATE = 0.3
+TOUT_AMNT = 2
+
+def conv(i, n, s, t = (1, 1)):
+    c = kl.Conv2D(n, s, padding = 'same', strides = t)(i)
+    b = kl.BatchNormalization()(c)
+    return kl.Activation('relu')(b)
+
+def avg(i, s):
+    return kl.AveragePooling2D(s, padding = 'same', strides = (1, 1))(i)
+
+def max(i, s):
+    return kl.MaxPooling2D(s, padding = 'same', strides = (1, 1))(i)
+
+def generate_start(i):
+    s = kl.Rescaling(1./255.)(i)
+    r1 = conv(s, CONV_SIZE, (3, 3))
+    r2 = conv(r1, CONV_SIZE, (3, 3))
+    r3 = conv(r2, CONV_SIZE, (3, 3))
+
+    m1 = kl.MaxPooling2D((3, 3))(r3)
+
+    r4 = conv(m1, CONV_SIZE, (1, 1))
+    r5 = conv(r4, CONV_SIZE, (3, 3))
+
+    return kl.MaxPooling2D((3, 3))(r5)
+
+def generate_type_a(i):
+    a11 = conv(i, CONV_SIZE, (1, 1))
+
+    a21 = avg(i, (3, 3))
+    a22 = conv(a21, CONV_SIZE, (1, 1))
+
+    a31 = conv(i, CONV_SIZE, (1, 1))
+    a32 = conv(a31, CONV_SIZE, (3, 3))
+
+    a41 = conv(i, CONV_SIZE, (1, 1))
+    a42 = conv(a41, CONV_SIZE, (3, 3))
+    a43 = conv(a42, CONV_SIZE, (3, 3))
+
+    return kl.Concatenate()([a11, a22, a32, a43])
+
+def generate_ab_bridge(i):
+    ab11 = conv(i, CONV_SIZE, (1, 1))
+    ab12 = conv(ab11, CONV_SIZE, (3, 3))
+    ab13 = conv(ab12, CONV_SIZE, (3, 3))
+
+    ab21 = conv(i, CONV_SIZE, (3, 3))
+
+    ab31 = max(i, (3, 3))
+
+    return kl.Concatenate()([ab13, ab21, ab31])
+
+def generate_type_b(i):
+    b11 = conv(i, CONV_SIZE, (1, 1))
+    b12 = conv(b11, CONV_SIZE, (7, 1))
+    b13 = conv(b12, CONV_SIZE, (1, 7))
+    b14 = conv(b13, CONV_SIZE, (7, 1))
+    b15 = conv(b14, CONV_SIZE, (1, 7))
+
+    b21 = conv(i, CONV_SIZE, (1, 1))
+    b22 = conv(b21, CONV_SIZE, (1, 7))
+    b23 = conv(b22, CONV_SIZE, (7, 1))
+
+    b31 = conv(i, CONV_SIZE, (1, 1))
+
+    b41 = avg(i, (3, 3))
+    b42 = conv(b41, CONV_SIZE, (1, 1))
+
+    return kl.Concatenate()([b15, b23, b31, b42])
+
+def generate_bc_bridge(i):
+    bc11 = conv(i, CONV_SIZE, (1, 1))
+    bc12 = conv(bc11, CONV_SIZE, (7, 1))
+    bc13 = conv(bc12, CONV_SIZE, (1, 7))
+    bc14 = conv(bc13, CONV_SIZE, (3, 3))
+
+    bc21 = conv(i, CONV_SIZE, (1, 1))
+    bc22 = conv(bc21, CONV_SIZE, (1, 1))
+
+    bc31 = max(i, (3, 3))
+
+    return kl.Concatenate()([bc14, bc22, bc31])
+
+def generate_aux(i):
+    u1 = kl.AveragePooling2D((5, 5))(i)
+    u2 = conv(u1, CONV_SIZE, (1, 1))
+    u3 = kl.Flatten()(u2)
+    u4 = kl.Dropout(DRPT_RATE)(u3)
+    u5 = kl.Dense(MPPT_SIZE)(u4)
+    u6 = kl.Dense(TOUT_AMNT)(u5)
+
+    return kl.Activation('softmax')(u6)
+
+def generate_type_c(i):
+    c11 = conv(i, CONV_SIZE, (1, 1))
+    c12 = conv(c11, CONV_SIZE, (3, 3))
+    c13 = conv(c12, CONV_SIZE, (1, 3))
+    c14 = conv(c12, CONV_SIZE, (3, 1))
+
+    c21 = conv(i, CONV_SIZE, (1, 1))
+    c22 = conv(c21, CONV_SIZE, (1, 3))
+    c23 = conv(c21, CONV_SIZE, (3, 1))
+
+    c31 = max(i, (3, 3))
+    c32 = conv(c31, CONV_SIZE, (1, 1))
+
+    c41 = conv(i, CONV_SIZE, (1, 1))
+
+    return kl.Concatenate()([c14, c23, c32, c41])
+
+def generate_finish(i):
+    f1 = kl.AveragePooling2D((5, 5))(i)
+    f2 = conv(f1, CONV_SIZE, (1, 1))
+    f3 = kl.Flatten()(f2)
+    f4 = kl.Dropout(DRPT_RATE)(f3)
+    f5 = kl.Dense(MPPT_SIZE)(f4)
+    f6 = kl.Dense(TOUT_AMNT)(f5)
+
+    return kl.Activation('softmax')(f6)
+
+
+gi = kl.Input((300, 300, 3))
+
+ds = generate_start(gi)
+
+for _ in range(3):
+    ds = generate_type_a(ds)
+
+ds = generate_ab_bridge(ds)
+
+for _ in range(4):
+    ds = generate_type_b(ds)
+
+uo = generate_aux(ds)
+
+go = generate_bc_bridge(ds)
+
+for _ in range(2):
+    go = generate_type_c(go)
+
+go = generate_finish(go)
+
+mod = k.Model(inputs = gi, outputs = go)
+
+mod.summary()
+mod.compile(optimizer = ko.Lion(learning_rate = 0.0001),
+            loss = 'categorical_crossentropy',
+            metrics = ['accuracy'])

stats-cm.py

@@ -0,0 +1,47 @@
+from m import *
+
+from matplotlib import pyplot as plt
+
+def __prep_conf_matr(data):
+    output_matrix = np.zeros([2, 2])
+
+    for p, r in zip(*data):
+        print(p, r)
+        output_matrix[np.argmax(p)][np.argmax(r)] += 1
+
+    return output_matrix
+
+
+def __plot_conf_matr(data):
+    matr = __prep_conf_matr(data)
+
+    _, ax = plt.subplots()
+
+    ax.matshow(matr, cmap = plt.cm.Blues)
+
+    for i, x in enumerate(matr):
+        for j, y in enumerate(x):
+            ax.text(i,
+                    j,
+                    str(round(y)),
+                    va = "center",
+                    ha = "center")
+
+    plt.show()
+
+ds = tf.keras.preprocessing.image_dataset_from_directory(
+    '../dataset-orig-aug-1-mini-1/',
+    labels = 'inferred',
+    label_mode = 'categorical',
+    color_mode = 'rgb',
+    image_size = (300, 300),
+    batch_size = 32,
+    verbose = True
+)
+
+#ds_short = ds.take(1)
+
+p = mod.predict(ds)
+r = np.concatenate([y for x, y in ds])
+
+__plot_conf_matr([p, r])

stats-generic.py

@@ -0,0 +1,20 @@
+from m import *
+
+from matplotlib import pyplot as plt
+
+ds = tf.keras.preprocessing.image_dataset_from_directory(
+    '../dataset-orig-aug-1-mini-1/',
+    labels = 'inferred',
+    label_mode = 'categorical',
+    color_mode = 'rgb',
+    image_size = (300, 300),
+    batch_size = 32,
+    verbose = True
+)
+
+p = mod.predict(ds)
+r = np.concatenate([y for x, y in ds])
+
+with open('results.txt', 'w') as f:
+    for i in zip(p, r):
+        f.write(f"{i[0][0]} {i[0][1]} {i[1][0]} {i[1][1]}\n")

stats-other.py

@@ -0,0 +1,30 @@
+from m import *
+from sklearn.metrics import f1_score, precision_score, recall_score, accuracy_score
+
+
+ds = tf.keras.preprocessing.image_dataset_from_directory(
+    '../dataset-orig-aug-1-mini-1/',
+    labels = 'inferred',
+    label_mode = 'categorical',
+    color_mode = 'rgb',
+    image_size = (300, 300),
+    batch_size = 32,
+    verbose = True
+)
+
+p = []
+r = []
+
+with open('results.txt', 'r') as f:
+    for i in f.read().split('\n'):
+        if i == '':
+            continue
+        
+        res = tuple(map(float, i.split()))
+        p.append([res[0] > 0.5, res[1] > 0.5])
+        r.append([res[2], res[3]])
+
+print(f"Accuracy  : {accuracy_score(p, r)}")
+print(f"Precision : {precision_score(p, r, average = 'micro')}")
+print(f"Recall    : {recall_score(p, r, average = 'micro')}")
+print(f"F1 Score  : {f1_score(p, r, average = 'micro')}")

stats-plot-1.py

@@ -0,0 +1,13 @@
+from matplotlib import pyplot as plt
+
+ta = [0.5396, 0.6266, 0.7137, 0.6768, 0.7114, 0.7821, 0.7373, 0.6866, 0.5987, 0.6861, 0.7236, 0.7147]
+va = [0.4273, 0.4714, 0.4273, 0.4229, 0.4317, 0.4229, 0.5022, 0.4449, 0.5947, 0.7489, 0.5815, 0.7048]
+
+x = [i+1 for i in range(12)]
+
+plt.plot(x, ta)
+plt.plot(x, va)
+
+plt.legend(["train_acc", "valid_acc"])
+
+plt.show()

test.py

@@ -1,116 +0,0 @@
-import tensorflow as tf
-import numpy as np
-
-from tensorflow.keras import layers as kl
-from tensorflow.keras import models as km
-from tensorflow.keras import optimizers as ko
-
-def generate_inception_model(a = 5, b = 4, c = 2):
-    # start
-    i = kl.Input(shape = (300, 300, 3))
-
-    r1 = kl.Conv2D(1024, (3, 3), padding = 'same', activation = 'relu')(i)
-    r2 = kl.Conv2D(1024, (5, 5), padding = 'same', activation = 'relu')(r1)
-    r3 = kl.Conv2D(1024, (7, 7), padding = 'same', activation = 'relu')(r2)
-
-    m1 = kl.MaxPooling2D((3, 3))(r3)
-
-    r4 = kl.Conv2D(1024, (3, 3), padding = 'same', activation = 'relu')(m1)
-    r5 = kl.Conv2D(1024, (5, 5), padding = 'same', activation = 'relu')(r4)
-
-    ia0 = kl.MaxPooling2D((2, 2))(r5)
-
-    # a types
-    for k in range(a):
-        exec(f"ia{k}_1_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(ia{k})")
-
-        exec(f"ia{k}_3_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(ia{k})")
-        exec(f"ia{k}_3_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ia{k}_3_1)")
-
-        exec(f"ia{k}_5_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(ia{k})")
-        exec(f"ia{k}_5_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ia{k}_5_1)")
-        exec(f"ia{k}_5_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ia{k}_5_2)")
-
-        exec(f"ia{k}_7_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(ia{k})")
-        exec(f"ia{k}_7_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ia{k}_7_1)")
-        exec(f"ia{k}_7_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ia{k}_7_2)")
-        exec(f"ia{k}_7_4 = kl.Conv2D(1024, (7, 1), padding = 'same', activation = 'relu')(ia{k}_7_3)")
-
-        exec(f"ia{k+1} = kl.Concatenate()([ia{k}_1_1, ia{k}_3_2, ia{k}_5_3, ia{k}_7_4])")
-
-    # grid size reductor 1
-    iab_1 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(eval(f"ia{a}"))
-
-    iab_2 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(eval(f"ia{a}"))
-    iab_3 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(iab_2)
-    iab_4 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(iab_3)
-
-    iab_6 = kl.Concatenate()([iab_1, iab_4, iab_5])
-
-    # b types
-    for k in range(b):
-        exec(f"ib{k}_1_1 = kl.MaxPooling2D((2, 2), padding = 'same')(iab_6)")
-        exec(f"ib{k}_1_2 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(i)")
-
-        exec(f"ib{k}_3_1 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(iab_6)")
-
-        exec(f"ib{k}_5_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(iab_6)")
-        exec(f"ib{k}_5_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ib{k}_5_1)")
-        exec(f"ib{k}_5_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ib{k}_5_2)")
-
-        exec(f"ib{k}_7_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(iab_6)")
-        exec(f"ib{k}_7_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ib{k}_7_1)")
-        exec(f"ib{k}_7_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ib{k}_7_2)")
-        exec(f"ib{k}_7_4 = kl.Conv2D(1024, (7, 1), padding = 'same', activation = 'relu')(ib{k}_7_3)")
-
-        exec(f"ib{k+1} = kl.Concatenate()([ib{k}_1_2, ib{k}_3_2, ib{k}_5_3, ib{k}_7_4])")
-
-    # grid size reductor 2
-
-    # c types
-    for k in range(c):
-        exec(f"ic{k}_1_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(i{'b' if k else 'a'}{k if k else a})")
-
-        exec(f"ic{k}_3_1 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(i{'b' if k else 'a'}{k if k else a})")
-
-        exec(f"ic{k}_5_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(i{'b' if k else 'a'}{k if k else a})")
-        exec(f"ic{k}_5_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ib{k}_5_1)")
-        exec(f"ic{k}_5_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ib{k}_5_2)")
-
-        exec(f"ic{k}_7_1 = kl.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(i{'b' if k else 'a'}{k if k else a})")
-        exec(f"ic{k}_7_2 = kl.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(ib{k}_7_1)")
-        exec(f"ic{k}_7_3 = kl.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(ib{k}_7_2)")
-        exec(f"ic{k}_7_4 = kl.Conv2D(1024, (7, 1), padding = 'same', activation = 'relu')(ib{k}_7_3)")
-
-        exec(f"ib{k+1} = kl.Concatenate()([ib{k}_1_1, ib{k}_3_2, ib{k}_5_3, ib{k}_7_4])")
-
-    # mppt
-
-    o = eval(f"ic{c}")
-    return tf.keras.Model(inputs = i, outputs = o)
-
-
-'''
-a_i = l.Input(shape = (300, 300, 3))
-
-a_1_1 = l.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(a_i)
-
-a_3_1 = l.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(a_i)
-a_3_2 = l.Conv2D(1024, (3, 1), padding = 'same', activation = 'relu')(a_3_1)
-
-a_5_1 = l.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(a_i)
-a_5_2 = l.Conv2D(1024, (5, 1), padding = 'same', activation = 'relu')(a_5_1)
-
-a_7_1 = l.Conv2D(1024, (1, 1), padding = 'same', activation = 'relu')(a_i)
-a_7_2 = l.Conv2D(1024, (7, 1), padding = 'same', activation = 'relu')(a_7_1)
-
-a_o = l.Concatenate()([a_1_1, a_3_2, a_5_2, a_7_2])
-
-inception_type_a = [a_i, a_o]
-
-tf.keras.Model(*inception_type_a)
-'''
-
-mod = generate_inception_model()
-mod.summary()
-mod.compile(optimizer = ko.Lion(learning_rate = 0.001))

train_l.py

@@ -0,0 +1,15 @@
+import tensorflow as tf
+
+ds_train, ds_valid = tf.keras.preprocessing.image_dataset_from_directory(
+    '../dataset-orig',
+    labels = 'inferred',
+    label_mode = 'categorical',
+    color_mode = 'rgb',
+    image_size = (300, 300),
+    shuffle = False,
+    validation_split = 0.05,
+    subset = 'both',
+    verbose = True
+)
+
+from m import mod

train_l_v2.py

@@ -0,0 +1,25 @@
+import tensorflow as tf
+
+ds_train, ds_valid = tf.keras.preprocessing.image_dataset_from_directory(
+    '/mnt/tmpfs1/ds-mini-1',
+    labels = 'inferred',
+    label_mode = 'categorical',
+    color_mode = 'rgb',
+    batch_size = 16,
+    image_size = (300, 300),
+    shuffle = False,
+    validation_split = 0.05,
+    subset = 'both',
+    verbose = True
+)
+
+from m import *
+
+ckpt = kc.ModelCheckpoint("model3.model.keras",
+                          monitor = 'val_accuracy',
+                          save_best_only = True)
+
+h = mod.fit(ds_train,
+            epochs = 9,
+            validation_data = ds_valid,
+            callbacks = [ckpt])