Unity-Technologies/ml-agents · error · NotImplementedException

Only float data types are currently supported

Error message

Only float data types are currently supported

What it means

FillTensorWithRandomNormal only supports float tensors; it throws NotImplementedException if the TensorProxy's DataType is anything else. Random-normal filling (used to generate random inputs, e.g. for testing) has no implementation for other dtypes.

Source

Thrown at com.unity.ml-agents/Runtime/Inference/TensorProxy.cs:189

        }

        /// <summary>
        /// Fill a pre-allocated Tensor with random numbers
        /// </summary>
        /// <param name="tensorProxy">The pre-allocated Tensor to fill</param>
        /// <param name="randomNormal">RandomNormal object used to populate tensor</param>
        /// <exception cref="NotImplementedException">
        /// Throws when trying to fill a Tensor of type other than float
        /// </exception>
        /// <exception cref="ArgumentNullException">
        /// Throws when the Tensor is not allocated
        /// </exception>
        public static void FillTensorWithRandomNormal(
            TensorProxy tensorProxy, RandomNormal randomNormal)
        {
            if (tensorProxy.DataType != typeof(float))
            {
                throw new NotImplementedException("Only float data types are currently supported");
            }

            if (tensorProxy.data == null)
            {
                throw new ArgumentNullException();
            }

            tensorProxy.data.CompleteAllPendingOperations();

            for (var i = 0; i < tensorProxy.data.Length(); i++)
            {
                ((Tensor<float>)tensorProxy.data)[i] = (float)randomNormal.NextDouble();
            }
        }
    }
}

View on GitHub (pinned to 3ecb446f75)

Solutions

1. Ensure the TensorProxy.DataType is typeof(float) before calling FillTensorWithRandomNormal
2. Write a custom fill routine for non-float dtypes
3. Convert the tensor data to float before filling

Example fix

// before
var proxy = new TensorProxy { DataType = typeof(int), shape = ... };
TensorUtils.FillTensorWithRandomNormal(proxy, randomNormal);
// after
var proxy = new TensorProxy { DataType = typeof(float), data = new TensorFloat(...) };
TensorUtils.FillTensorWithRandomNormal(proxy, randomNormal);

if (tensorProxy.DataType != typeof(float))
    throw new InvalidOperationException("FillTensorWithRandomNormal requires a float tensor");

bool IsFloatTensor(TensorProxy t) => t.DataType == typeof(float);

try { TensorProxy.FillTensorWithRandomNormal(proxy, randomNormal); }
catch (NotImplementedException)
{ /* fall back to a custom dtype-specific fill routine */ }

When it happens

Trigger: Calling TensorProxy.FillTensorWithRandomNormal with a tensorProxy whose DataType is not typeof(float) (e.g. int, double).

Common situations: Custom inference/utility code building non-float tensors and reusing the random-fill helper; test-generation code for quantized or integer-input models.

Related errors

• Unknown tensorProxy expected as output : {tensor.name}
• Unknown tensorProxy expected as input : {tensor.name}
• Index out of bounds, expected a number between 0 and {Length
• Enumerator not started.
• Enumerator has reached the end already.

AI-assisted analysis of Unity-Technologies/ml-agents@3ecb446f75 (2026-09-02). Data as JSON: /api/errors/7d64b06968ec9b04. Report an issue: GitHub.