tracel-ai/burn · error
all_int: unsupported dtype {:?}
Error message
all_int: unsupported dtype {:?} What it means
Dtype-dispatch exhaustiveness panic: `all_int` supports the full integer dtype set and panics for any non-integer dtype, meaning the Flex backend received an int all-reduction request for a tensor of the wrong kind.
Source
Thrown at crates/burn-flex/src/ops/comparison.rs:913
}
/// Check if any element along a dimension is non-zero (int tensors).
pub fn any_int_dim(tensor: FlexTensor, dim: usize, out_dtype: BoolDType) -> FlexTensor {
reduce_bool_dim_int(&tensor, dim, false, |a, b| a || b, out_dtype)
}
/// Check if all elements are non-zero (int tensors).
pub fn all_int(tensor: FlexTensor, out_dtype: BoolDType) -> FlexTensor {
let all = match tensor.dtype() {
DType::I64 => iter_elements::<i64>(&tensor).all(|x| x != 0),
DType::I32 => iter_elements::<i32>(&tensor).all(|x| x != 0),
DType::I16 => iter_elements::<i16>(&tensor).all(|x| x != 0),
DType::I8 => iter_elements::<i8>(&tensor).all(|x| x != 0),
DType::U64 => iter_elements::<u64>(&tensor).all(|x| x != 0),
DType::U32 => iter_elements::<u32>(&tensor).all(|x| x != 0),
DType::U16 => iter_elements::<u16>(&tensor).all(|x| x != 0),
DType::U8 => iter_elements::<u8>(&tensor).all(|x| x != 0),
_ => panic!("all_int: unsupported dtype {:?}", tensor.dtype()),
};
bool_scalar(all, out_dtype)
}
/// Check if all elements along a dimension are non-zero (int tensors).
pub fn all_int_dim(tensor: FlexTensor, dim: usize, out_dtype: BoolDType) -> FlexTensor {
reduce_bool_dim_int(&tensor, dim, true, |a, b| a && b, out_dtype)
}
/// Check if any bool element is true.
pub fn any_bool(tensor: FlexTensor, out_dtype: BoolDType) -> FlexTensor {
let tensor = tensor.to_contiguous();
let data: &[u8] = tensor.bytes();
bool_scalar(data.iter().any(|&x| x != 0), out_dtype)
}
/// Check if any bool element along a dimension is true.
pub fn any_bool_dim(tensor: FlexTensor, dim: usize, out_dtype: BoolDType) -> FlexTensor {View on GitHub (pinned to d16f7ba2ed)
Solutions
- Route float tensors to all_float instead
- Cast to an integer dtype first if appropriate
- Add a dtype check/assert before calling all_int
Example fix
// before let all = all_int(f64_tensor, BoolDType::Native); // after let all = all_float(f64_tensor, BoolDType::Native);
Defensive patterns
Strategy: type-guard
Validate before calling
if !matches!(t.dtype(), DType::I64|DType::I32|DType::I16|DType::I8|DType::U64|DType::U32|DType::U16|DType::U8) { /* route to all_float or cast */ } Type guard
fn is_int_dtype(d: &DType) -> bool {
matches!(d, DType::I64|DType::I32|DType::I16|DType::I8|DType::U64|DType::U32|DType::U16|DType::U8)
} Prevention
- Dispatch all_int/all_float by dtype family
- Check for upstream ops that silently promote tensors to float
- Cover int/float paths in tests for shared reduction utilities
When it happens
Trigger: Calling the public all_int(tensor, out_dtype) with a float or Bool tensor.
Common situations: Symmetric mistake to any_int: float tensor routed into the int reduction; dtype drifted upstream to F32 after a division/normalization.
Related errors
- any_float: unsupported dtype {:?}
- all_float: unsupported dtype {:?}
- any_int: unsupported dtype {:?}
- reduce_bool_dim: unsupported dtype {:?}
- reduce_bool_dim_int: unsupported dtype {:?}
AI-assisted analysis of tracel-ai/burn@d16f7ba2ed (2026-09-05).
Data as JSON: /api/errors/1ca5d9e0a27caf77.
Report an issue: GitHub.