tracel-ai/burn · error

{other:?} doesn't support native packing

Error message

{other:?} doesn't support native packing

What it means

On the `QuantStore::PackedNative` arm of `dtype_to_elem_type`, values other than E2M1 that still don't support native packing trigger `{other:?} doesn't support native packing`. This documents that PackedNative storage is only valid for a subset of QuantValues; attempting to use it with an unsupported sub-byte or exotic value panics with the value's debug name embedded.

Source

Thrown at crates/burn-backend/src/cubecl.rs:106

            BoolStore::U32 => ElemType::UInt(UIntKind::U32),
        },
        DType::QFloat(scheme) => match scheme.store {
            QuantStore::Native => match scheme.value {
                QuantValue::Q8F | QuantValue::Q8S => ElemType::Int(IntKind::I8),
                QuantValue::E4M3 => ElemType::Float(FloatKind::E4M3),
                QuantValue::E5M2 => ElemType::Float(FloatKind::E5M2),
                QuantValue::Q4F
                | QuantValue::Q4S
                | QuantValue::Q2F
                | QuantValue::Q2S
                | QuantValue::E2M1 => {
                    panic!("Can't store native sub-byte values")
                }
            },
            QuantStore::PackedU32(_) => ElemType::UInt(UIntKind::U32),
            QuantStore::PackedNative(_) => match scheme.value {
                QuantValue::E2M1 => panic!("Can't store native sub-byte values"),
                other => panic!("{other:?} doesn't support native packing"),
            },
        },
    }
}

/// Convert a burn [`DType`] into the corresponding cubecl [`ElemType`].
///
/// Handles sub-byte packed quantization configurations that cannot be expressed
/// as a plain [`ElemType`] by emitting a direct `ElemType`.
pub fn dtype_to_storage_type(dtype: DType) -> ElemType {
    match dtype {
        DType::QFloat(QuantScheme {
            store: QuantStore::PackedNative(_),
            value: QuantValue::E2M1,
            ..
        }) => ElemType::Float(FloatKind::E2M1x2),
        _ => dtype_to_elem_type(dtype),
    }

View on GitHub (pinned to d16f7ba2ed)

Solutions

1. Read the embedded `{other:?}` value in the panic to identify the unsupported QuantValue.
2. Switch to `QuantStore::PackedU32`, which supports the full sub-byte value set.
3. Re-quantize the model with a supported store/value combination if loading legacy checkpoints.
4. Validate the scheme (store mode vs value width) at config-load time before it reaches kernel dispatch.

Example fix

// before
let scheme = QuantScheme::new(QuantStore::PackedNative(f), QuantValue::Q2S, ...);

// after
let scheme = QuantScheme::new(QuantStore::PackedU32(f), QuantValue::Q2S, ...);

// only allow PackedNative for values known to support it
fn supports_packed_native(v: &burn::tensor::QuantValue) -> bool {
    matches!(v, burn::tensor::QuantValue::E4M3 | burn::tensor::QuantValue::E5M2)
}

fn safe_scheme(store: &burn::tensor::QuantStore, value: &burn::tensor::QuantValue) -> bool {
    !matches!(store, burn::tensor::QuantStore::PackedNative(_))
        || supports_packed_native(value)
}

When it happens

Trigger: `dtype_to_elem_type` called (via `dtype_to_storage_type`, `fuse_base`, `create_key`, `generate_reduce_autotune_key`, `compute_input_grad`, `with_bounds`) with `QuantStore::PackedNative(_)` and a QuantValue that is neither E2M1 nor accepted by the native packing path (e.g. Q4S/Q2S variants not enabled for native packing).

Common situations: Deserializing a quantization scheme from an old checkpoint format whose PackedNative + value combination is no longer supported; hand-building schemes with mismatched store/value pairs; version upgrades changing which values PackedNative supports.

Related errors

• Can't store native sub-byte values
• Quantization scheme is not valid for dtype {other:?}
• q_matmul inputs are on different backends
• float_storage_as_f32: unsupported dtype {:?}
• Expected quantized dtype, got {:?}

AI-assisted analysis of tracel-ai/burn@d16f7ba2ed (2026-09-05). Data as JSON: /api/errors/a33620adc8ecea03. Report an issue: GitHub.