Interpreting non ascii codepoint 256 - during training of model

[Shreyas Skandan]

Hi,

I looked through the message boards extensively and couldn't find any posts related to this error, which hopefully means is something that can be easily fixed.

I'm using the ImageNet example to train my own model using my own images. I have a limited data set right now (40 training & 10 validation images) but I was hoping to run through the entire pipeline before gathering more data.

I managed to get through the compilation of the training and validation databases successfully.

./create_imagenet_mean.sh -> success

./make_imagenet_mean.sh -> success

But when I try to train my model I get the following error,

X-MacBook-Pro:imagenet X$ ~/technologies/caffe/build/tools/caffe train --solver ~/technologies/caffe/models/bvlc_reference_caffenet/solver.prototxt

[libprotobuf ERROR google/protobuf/text_format.cc:274] Error parsing text-format caffe.SolverParameter: 1:6: Interpreting non ascii codepoint 226.

[libprotobuf ERROR google/protobuf/text_format.cc:274] Error parsing text-format caffe.SolverParameter: 1:6: Expected string.

F0423 12:11:36.859617 2004807680 upgrade_proto.cpp:1063] Check failed: ReadProtoFromTextFile(param_file, param) Failed to parse SolverParameter file: /Users/X/technologies/caffe/models/bvlc_reference_caffenet/solver.prototxt

*** Check failure stack trace: ***

    @        0x10fdff9da  google::LogMessage::Fail()

    @        0x10fdff0d5  google::LogMessage::SendToLog()

    @        0x10fdff63b  google::LogMessage::Flush()

    @        0x10fe02a17  google::LogMessageFatal::~LogMessageFatal()

    @        0x10fdffcc7  google::LogMessageFatal::~LogMessageFatal()

    @        0x10fc56f65  caffe::ReadSolverParamsFromTextFileOrDie()

    @        0x10fadcbf9  train()

    @        0x10fadf711  main

    @     0x7fff8b2335ad  start

    @                0x4  (unknown)

Abort trap: 6

Does anyone have any ideas what might be causing this error? I think it could be an incorrect entry in solver.prototxt but I can't figure out what it is, so here's my prototxt file,

net: “train_val.prototxt”

test_iter: 10

test_interval: 10

base_lr: 0.01

lr_policy: "step"

gamma: 0.1

stepsize: 100000

display: 20

max_iter: 450000

momentum: 0.9

weight_decay: 0.0005

snapshot: 10

snapshot_prefix: “caffenet_train"

solver_mode: CPU

Thanks in advance,

 
Any help is much appreciated. 

[Jan]

It seems that there is something wrong with the encoding of your text file. Try to open it in an text editor that can convert encodings, try to encode it with utf-8 and try again.

Jan