I tried to create a model with keras, which has the following code structure:
import tensorflow as tf
import keras
import os
from tensorflow.keras.models import save_model
from keras.activations import relu, elu, linear, sigmoid
from tensorflow.keras.models import Sequentialimport tensorflow as tf
import keras
…
from tensorflow.keras.models import save_model
from keras.activations import relu, elu, linear, sigmoid
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Activation
…
model = Sequential()
model.add(Dense(24, input_dim=X_train.shape[1]))
model.add(Dense(10, activation=‘relu’))
model.add(Dropout(0.1))
model.add(Dense(1, activation=‘sigmoid’))
model.compile(optimizer=Adam(0.001), loss=‘binary_crossentropy’, metrics=[‘accuracy’]) print(model.summary())
model.fit(X_train, y_train, epochs=80, batch_size=10)
Accuracy= model.evaluate(X_test, y_test)
print(Accuracy)
save_model(model,‘ANN-model.h5’)
The training dataset of this model is an excel file in xlsx format, and its output should be 0 or 1.
The training dataset is like this:
After studying TVM’s tutorial(Compile Keras Models — tvm 0.10.dev0 documentation), I added the following code to try to run the model using TVM:
which the oneperson1.xlsx is like this:
data = np.genfromtxt(‘oneperson1.xlsx’,delimiter=’,’)
data = data.astype(“float64”)
print(“input_1”, data.shape) # (8,) (1,8)
reconstructed_model = keras.models.load_model(“ANN-model.h5”)
shape_dict = {“input_1”: data.shape}
mod, params = relay.frontend.from_keras(reconstructed_model, shape_dict)
#compile the model
target = “llvm”
dev = tvm.cpu(0)
with tvm.transform.PassContext(opt_level=0): model = relay.build_module.create_executor(“graph”, mod, dev, target, params).evaluate()
#Execute on TVM
dtype = “float32”
tvm_out = model(tvm.nd.array(data.astype(dtype)))
print(tvm_out)
when I run this py script, I got this error:
How should I run a keras model for a binary classification problem using TVM? Or where can I find a more detailed TVM tutorial?