Update resnet50_imagenet.py

llvm/projects/keras/src/resnet50_imagenet.py

@@ -29,7 +29,7 @@ data_format = 'channels_first'
 
 
 IMAGENET_DIR = '/home/nz11/ILSVRC2012/'
-OUTPUT_DIR = 'data/resnet50_imagenet_tune/'
+OUTPUT_DIR = 'data/resnet50_imagenet_tune_regenerate/'
 WEIGHTS_PATH = 'data/resnet50_imagenet/weights.h5'
 
 NUM_CLASSES = 200
@@ -217,6 +217,9 @@ y_true = np.array(y_true)
 X_tune = np.array(X_tune)
 y_tune = np.array(y_tune)
 
+print ('tune size', len(X_tune))
+print ('test size', len(X_test))
+
 
 
 
@@ -267,19 +270,20 @@ model.compile(optimizer=keras.optimizers.Adam(lr=0.00001), loss='categorical_cro
 if os.path.exists(WEIGHTS_PATH):
     model.load_weights(WEIGHTS_PATH)
 else:
-    model.fit_generator(generate(), steps_per_epoch=1000, validation_data=(X_test, to_categorical(y_true, num_classes=1000)), epochs=6)
-    model.save_weights(OUTPUT_DIR + 'weights.h5')
+    pass
+#     model.fit_generator(generate(), steps_per_epoch=1000, validation_data=(X_test, to_categorical(y_true, num_classes=1000)), epochs=6)
+#     model.save_weights(OUTPUT_DIR + 'weights.h5')
 
-translate_to_approxhpvm(model, OUTPUT_DIR, X_tune, y_tune, 1000)
+translate_to_approxhpvm(model, OUTPUT_DIR, X_tune, y_tune, 1000, dump_weights=False)
 
-# dumpCalibrationData2(OUTPUT_DIR + 'test_input_10K.bin', X_test, OUTPUT_DIR + 'test_labels_10K.bin', y_true)
-dumpCalibrationData2(OUTPUT_DIR + 'tune_input.bin', X_tune, OUTPUT_DIR + 'tune_labels.bin', y_tune)
-dumpCalibrationData2(OUTPUT_DIR + 'test_input.bin', X_test, OUTPUT_DIR + 'test_labels.bin', y_true)
+# # dumpCalibrationData2(OUTPUT_DIR + 'test_input_10K.bin', X_test, OUTPUT_DIR + 'test_labels_10K.bin', y_true)
+# dumpCalibrationData2(OUTPUT_DIR + 'tune_input.bin', X_tune, OUTPUT_DIR + 'tune_labels.bin', y_tune)
+# dumpCalibrationData2(OUTPUT_DIR + 'test_input.bin', X_test, OUTPUT_DIR + 'test_labels.bin', y_true)
 
 
 pred = np.argmax(model.predict(X_test), axis=1)
 print ('val accuracy', np.sum(pred == y_true.ravel()) / len(X_test))
     
-pred = np.argmax(model.predict(X_tune), axis=1)
-print ('val accuracy', np.sum(pred == y_tune.ravel()) / len(X_tune))
+# pred = np.argmax(model.predict(X_tune), axis=1)
+# print ('val accuracy', np.sum(pred == y_tune.ravel()) / len(X_tune))
     
\ No newline at end of file