Issue when trying to parse json

[mikegazzaruso]

Hi, parser gives me the following error:

std::ifstream jsonStream("adaptivo_ac_weights.json", std::ifstream::binary);
auto model = RTNeural::json_parser::parseJson(jsonStream); <----- breaks here

JUCE v7.0.5
libc++abi: terminating with uncaught exception of type nlohmann::json_v3_11_1::detail::parse_error: [json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - unexpected end of input; expected '[', '{', or a literal
terminating with uncaught exception of type nlohmann::json_v3_11_1::detail::parse_error: [json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - unexpected end of input; expected '[', '{', or a literal
(lldb) 

My model json was created with your script, it's a common dense / relu / softmax model created with tensorflow and keras.

It says error on line 1, I will past first lines:

{
    "in_shape": [
        null,
        40
    ],
    "layers": [
        {
            "type": "dense",
            "activation": "",
            "shape": [
                null,
                100
            ],
            "weights": [
                [
                    [
                        -0.04106162488460541,
                        -0.008081752806901932,
                        0.05740608274936676,
                        0.0283
 ---- continue ----

[mikegazzaruso]

Solved, ifstream gives no warning when file is not found, I erroneusly mispelled a character.

Using if(myFile.fail) I was able to properly catch that.

Sorry and thanks for your marvellous work.

[jatinchowdhury18]

Ah yes, that particular exception usually means that the JSON parser was passed an empty stream.

In practice, if you need to be able to dynamically load models at run-time, it's possible to wrap the model loading process in a try/catch block, to avoid a crash in case an invalid or non-existent file is passed in.

[mikegazzaruso]

Thanks for the comment! Yes you are right. I'm experimenting other problems but I opened a separate issue, thanks for your help