Internal Server Error on the blob converter API

seanmabli

I am trying to run the following code to transform by tensorflow model into a blob, but I am getting an internal server error.

Code:
`import requests

url = "https://blobconverter.luxonis.com/compile"
payload = {
'compile_type': 'model',
'model_type': 'caffe',
'intermediate_compiler_params': '--data_type=FP16 --mean_values [127.5,127.5,127.5] --scale_values [255,255,255]',
'compiler_params': '-ip U8 -VPU_MYRIAD_PLATFORM VPU_MYRIAD_2480 -VPU_NUMBER_OF_SHAVES 4 -VPU_NUMBER_OF_CMX_SLICES 4'
}
files = {
'model': open('/home/team4169/vision/content/fine_tuned_model/saved_model/saved_model.pb', 'rb'),
}
params = {
'version': '2020.1',
}

try:
response = requests.post(url, data=payload, files=files, params=params)
response.raise_for_status()
print(response.text)
except requests.exceptions.RequestException as e:
print("Request failed:", e)
Error:
Request failed: 500 Server Error: INTERNAL SERVER ERROR for url: https://blobconverter.luxonis.com/compile?version=2020.1
`

jakaskerl

Hi seanmabli
What model are you trying to convert?

Thanks,
Jaka

seanmabli

Hello @jakaskerl,

I tried to upload both a .zip and .pb, but the website wouldn't allow me to upload them. Here is a google drive link where you can download a .zip containing the saved_model.pb:

Thank you,
Sean

JanCuhel

Hi @seanmabli,

I apologize for the delay in our response. I see from the code you shared that you are setting the model type as a caffe, which is incorrect because you are trying to convert a Tensorflow model. However, even when I set the correct model type, I obtained the following error:
[ FRAMEWORK ERROR ] Cannot load input model: TensorFlow cannot read the model file: "/tmp/blobconverter/ac6cd1d00994439f80bf19aca9531512/saved_model/FP16/saved_model.pb" is incorrect TensorFlow model file. The file should contain one of the following TensorFlow graphs: 1. frozen graph in text or binary format 2. inference graph for freezing with checkpoint (--input_checkpoint) in text or binary format 3. meta graph Make sure that --input_model_is_text is provided for a model in text format. By default, a model is interpreted in binary format. Framework error details: Error parsing message. For more information please refer to Model Optimizer FAQ, question #43. (https://docs.openvinotoolkit.org/latest/openvino_docs_MO_DG_prepare_model_Model_Optimizer_FAQ.html?question=43#question-43)

It seems that the model format is incorrect. If you haven't done so, I'd recommend that you try to freeze your model. Here's a tutorial on how to do it. Also, I'd recommend using our web app or blobconverter cli package for more user-friendliness.

Here's an example of how to use the blobconverter cli package to convert a Tensorflow model:
import blobconverter

blob_path = blobconverter.from_tf(
frozen_pb="/path/to/deeplabv3_mnv2_pascal_train_aug.pb",
data_type="FP16",
shaves=5,
optimizer_params=[
"--reverse_input_channels",
"--input_shape=[1,513,513,3]",
"--input=1:mul_1",
"--output=ArgMax",
],
)

Best,
Jan