Error in operator global_pool: [19:16:46] g:\deeplearn\mxnet\src\operator\./pooling-inl.h:196: Check failed: param_.kernel[0] <= dshape[2] + 2 * param_.pad[0] kernel size (7) exceeds input (2 padded to 2)

[paulfryer]

Any ideas what might cause that error? Intuition about what sort of things I should focus on? I'm getting this error when calling model.Predict

Environment info

Operating System: Windows
Compiler: Visual Studio
Package used (Python/R/Scala/Julia): C#

Error Message:

Error in operator global_pool: [19:16:46] g:\deeplearn\mxnet\src\operator./pooling-inl.h:196: Check failed: param_.kernel[0] <= dshape[2] + 2 * param_.pad[0] kernel size (7) exceeds input (2 padded to 2)

at mxnet.csharp.Util.CallCheck(Int32 ret) in F:\Projects\mxnet.csharp\mxnet.csharp\Util.cs:line 81
at mxnet.csharp.Symbol.InferShape(Dictionary2 argShapes, List1 inShape, List1 outShape, List1 auxShape) in F:\Projects\mxnet.csharp\mxnet.csharp\Symbol.cs:line 415
at mxnet.csharp.Symbol.SimpleBind(Context context, Dictionary2 inputShapes, OpReqType gradReq, Dictionary2 typeDict, Dictionary2 group2Ctx) in F:\Projects\mxnet.csharp\mxnet.csharp\Symbol.cs:line 575 at mxnet.csharp.FeedForward.InitPredictor(Dictionary2 inputShapes) in F:\Projects\mxnet.csharp\mxnet.csharp\FeedForward_Predict.cs:line 107
at mxnet.csharp.FeedForward.Predict(IDataIter inputX, Nullable`1 numBatch, Boolean returnData, Boolean reset) in F:\Projects\mxnet.csharp\mxnet.csharp\FeedForward_Predict.cs:line 23
at test.console.Program.TrainTest(String path) in F:\Projects\mxnet.csharp\test.console\Program.cs:line 163
at test.console.Program.Main(String[] args) in F:\Projects\mxnet.csharp\test.console\Program.cs:line 82

What have you tried to solve it?

1. Tried changing batch size

Minimum reproducible example

        int batch_size = 1;
        uint size ;
        IntPtr array = new IntPtr();
        IntPtr strARray ;
        NativeMethods.MXListAllOpNames(out size, out strARray);
        var pnet = Symbol.Load("Inception-BN-symbol.json");
        Context ctx = new Context(DeviceType.KCpu, 0);
        Speedometer speed = new Speedometer(batch_size, 50);
        DoCheckpoint doCheckpoint = new DoCheckpoint(path);

        CustomMetric customMetric = new CustomMetric((l, p) => Accuracy(l, p, batch_size), "Accuracy");

        Optimizer optimizer = new CcSgd(momentum: 0.9f, learningRate: 0.001f, wd: 0.00001f, rescaleGrad: 1.0f / batch_size);

        FeedForward model = null;
        try
        {
            var modelload = FeedForward.Load(path, ctx: ctx,
                numEpoch: 1,
                optimizer: optimizer,
                initializer: new Xavier(factorType: FactorType.In, magnitude: 2.34f));

             model = new FeedForward(pnet, new List<Context> { ctx },
                numEpoch: 1,
                optimizer: optimizer,
                initializer: new Xavier(factorType: FactorType.In, magnitude: 2.34f),
                argParams: modelload.ArgParams,
                auxParams: modelload.AuxParams
            );

        }
        catch (Exception)
        {
            // ignored
        }

        if (model == null)
        {
            model = new FeedForward(pnet, new List<Context> { ctx },
                numEpoch: 1,
                optimizer: optimizer,
                initializer: new Xavier(factorType: FactorType.In, magnitude: 2.34f)
            );
        }
        ReadData rdpredict = new ReadData("images\\", batch_size, true); //, true);
        var testOut = model.Predict(rdpredict, batch_size);

[szha]

This issue is closed due to lack of activity in the last 90 days. Feel free to ping me to reopen if this is still an active issue. Thanks!