TF-Keras

IndexError: too many indices for array: array is 0-dimensional, but 3 were indexed

#1
by Minjung17 - opened

모델카드를 사용했는데, 에러가 뜹니다.

model = from_pretrained_keras("Derendering/InkSight-Small-p")
cf = model.signatures['serving_default']
image='/content/image2.png'
prompt = "DANKE" # "Recognize and derender." or "Derender the ink:

input_text = tf.constant([prompt], dtype=tf.string)
image_encoded = tf.reshape(tf.io.encode_jpeg(np.array(image)[:, :, :3]), (1, 1))
output = cf(**{'input_text': input_text, 'image/encoded': image_encoded})

IndexError: too many indices for array: array is 0-dimensional, but 3 were indexed 라고요.

InkSight at Google Research org

Hi Minjung17, thanks for reaching out. The error you're seeing here is because of the usage of InkSight. Please refer to our colab for example. A fix would be:

file_path = '/content/image2.png'
input_image = Image.open(file_path)
image, _, _, _, _ = scale_and_pad(input_image) # this function is defined in the colab as well

prompt = "Recognize and derender."
cf = model.signatures['serving_default']
input_text = tf.constant([prompt], dtype=tf.string)
image_encoded = tf.reshape(tf.io.encode_jpeg(np.array(image)[:, :, :3]), (1, 1))
output = cf(**{'input_text': input_text, 'image/encoded': image_encoded})

Please let me know if this fixes your issue :)

chengkunli changed discussion status to closed

Sign up or log in to comment