tracel-ai/burn · error

float_select: unsupported dtype {:?}

Error message

float_select: unsupported dtype {:?}

What it means

Dtype-dispatch exhaustiveness panic: `float_select` supports only float dtypes and panics otherwise; reaching it means a non-float tensor was dispatched into the float select (indexing) op in the Flex backend.

Source

Thrown at crates/burn-flex/src/ops/float.rs:385

            DType::F32 => crate::ops::gather_scatter::gather_nd::<f32>(data, indices),
            DType::F64 => crate::ops::gather_scatter::gather_nd::<f64>(data, indices),
            DType::F16 => crate::ops::gather_scatter::gather_nd::<f16>(data, indices),
            DType::BF16 => crate::ops::gather_scatter::gather_nd::<bf16>(data, indices),
            _ => panic!("float_gather_nd: unsupported dtype {:?}", data.dtype()),
        }
    }

    fn float_select(
        tensor: FloatTensor<Flex>,
        dim: usize,
        indices: IntTensor<Flex>,
    ) -> FloatTensor<Flex> {
        match tensor.dtype() {
            DType::F32 => crate::ops::gather_scatter::select::<f32>(tensor, dim, indices),
            DType::F64 => crate::ops::gather_scatter::select::<f64>(tensor, dim, indices),
            DType::F16 => crate::ops::gather_scatter::select::<f16>(tensor, dim, indices),
            DType::BF16 => crate::ops::gather_scatter::select::<bf16>(tensor, dim, indices),
            _ => panic!("float_select: unsupported dtype {:?}", tensor.dtype()),
        }
    }

    fn float_select_assign(
        tensor: FloatTensor<Flex>,
        dim: usize,
        indices: IntTensor<Flex>,
        value: FloatTensor<Flex>,
        update: burn_backend::tensor::IndexingUpdateOp,
    ) -> FloatTensor<Flex> {
        match update {
            burn_backend::tensor::IndexingUpdateOp::Assign => match tensor.dtype() {
                DType::F32 => {
                    crate::ops::gather_scatter::select_assign::<f32>(tensor, dim, indices, value)
                }
                DType::F64 => {
                    crate::ops::gather_scatter::select_assign::<f64>(tensor, dim, indices, value)
                }

View on GitHub (pinned to d16f7ba2ed)

Solutions

1. Confirm tensor.dtype() is a float type before select
2. Use the corresponding int/bool select op for non-float tensors
3. Cast the tensor to a float dtype before selecting
4. Add a match arm for any newly added DType

Example fix

// before: labels is I64 -> panic
let picked = labels.select(dim, indices);
// after
let picked = labels.cast(FloatDType::F32).select(dim, indices);

fn ensure_float_for_select(dt: DType) -> Result<(), String> {
    match dt {
        DType::F32 | DType::F64 | DType::F16 | DType::BF16 => Ok(()),
        other => Err(format!("float_select requires a float dtype, got {:?}", other)),
    }
}

fn is_float_dtype(dt: DType) -> bool {
    matches!(dt, DType::F32 | DType::F64 | DType::F16 | DType::BF16)
}

When it happens

Trigger: Calling float_select (Tensor::select along a dim) where the tensor's dtype is Int or Bool rather than a float type.

Common situations: Selecting rows/slices from an int tensor (e.g. label tensor) via the float op; dtype changes upstream; new DType variant missing from the dispatch table.

Related errors

• Should be float, got int
• Should be float, got bool
• Should be float, got quantized
• Should be float, got autodiff
• float_into_int: unsupported source dtype {:?}

AI-assisted analysis of tracel-ai/burn@d16f7ba2ed (2026-09-05). Data as JSON: /api/errors/e4be166cef84343f. Report an issue: GitHub.