case int:
				field.ReflectValueOf(ctx, value).SetFloat(float64(data))
			case int8:
				field.ReflectValueOf(ctx, value).SetFloat(float64(data))
			case int16:
				field.ReflectValueOf(ctx, value).SetFloat(float64(data))
			case int32:
				field.ReflectValueOf(ctx, value).SetFloat(float64(data))
			case uint:
				field.ReflectValueOf(ctx, value).SetFloat(float64(data))
			case uint8:

		}

		fA, okAf := a.bytesToFloat()
		fB, okBf := b.bytesToFloat()
		if okAf && okBf {
			a.setFloat(fA)
			b.setFloat(fB)
			return nil
		}

		// Check if they int and float combination.
		if okAi && okBf {
			a.setInt(iA)
			b.setFloat(fA)
			return nil
		}
		if okBi && okAf {
			a.setFloat(fA)
			b.setInt(iB)
			return nil
		}

		// Not numeric types at this point.

		// inferring untyped values are numbers.
		if !argVal.isNumeric() {
			if i, ok := argVal.bytesToInt(); ok {
				argVal.setInt(i)
			} else if f, ok := argVal.bytesToFloat(); ok {
				argVal.setFloat(f)
			} else {
				return errNonNumericArg(funcName)
			}
		}
	}

	// Mark that we have seen one non-null value.
	isFirstRow := false
	if !e.aggregate.seen {
		e.aggregate.seen = true

        issued that starts with Layer "layer-name" is casting an input tensor
        from dtype float64 to the layer's dtype of float32. To fix, either set
        the default dtype to float64 with
        `tf.keras.backend.set_floatx('float64')`, or pass `dtype='float64'` to
        each of the Layer constructors. See `tf.keras.layers.Layer` for more
        information.
    *   Some `tf.assert_*` methods now raise assertions at operation creation