(l-onnx-doc-If)=

# If


(l-onnx-op-if-25)=

## If - 25

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `25`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 25**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(float4e2m1))`, `optional(tensor(float8e4m3fn))`, `optional(tensor(float8e4m3fnuz))`, `optional(tensor(float8e5m2))`, `optional(tensor(float8e5m2fnuz))`, `optional(tensor(float8e8m0))`, `optional(tensor(int16))`, `optional(tensor(int2))`, `optional(tensor(int32))`, `optional(tensor(int4))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint2))`, `optional(tensor(uint32))`, `optional(tensor(uint4))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(float4e2m1))`, `seq(tensor(float8e4m3fn))`, `seq(tensor(float8e4m3fnuz))`, `seq(tensor(float8e5m2))`, `seq(tensor(float8e5m2fnuz))`, `seq(tensor(float8e8m0))`, `seq(tensor(int16))`, `seq(tensor(int2))`, `seq(tensor(int32))`, `seq(tensor(int4))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint2))`, `seq(tensor(uint32))`, `seq(tensor(uint4))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(float4e2m1)`, `tensor(float8e4m3fn)`, `tensor(float8e4m3fnuz)`, `tensor(float8e5m2)`, `tensor(float8e5m2fnuz)`, `tensor(float8e8m0)`, `tensor(int16)`, `tensor(int2)`, `tensor(int32)`, `tensor(int4)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint2)`, `tensor(uint32)`, `tensor(uint4)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv13.
* **B** in ( `tensor(bool)` ):

  Only bool

### Examples

#### _if

```python
import numpy as np
import onnx

# Given a bool scalar input cond.
# return constant tensor x if cond is True, otherwise return constant tensor y.

then_out = onnx.helper.make_tensor_value_info(
    "then_out", onnx.TensorProto.FLOAT, [5]
)
else_out = onnx.helper.make_tensor_value_info(
    "else_out", onnx.TensorProto.FLOAT, [5]
)

x = np.array([1, 2, 3, 4, 5]).astype(np.float32)
y = np.array([5, 4, 3, 2, 1]).astype(np.float32)

then_const_node = onnx.helper.make_node(
    "Constant",
    inputs=[],
    outputs=["then_out"],
    value=onnx.numpy_helper.from_array(x),
)

else_const_node = onnx.helper.make_node(
    "Constant",
    inputs=[],
    outputs=["else_out"],
    value=onnx.numpy_helper.from_array(y),
)

then_body = onnx.helper.make_graph(
    [then_const_node], "then_body", [], [then_out]
)

else_body = onnx.helper.make_graph(
    [else_const_node], "else_body", [], [else_out]
)

if_node = onnx.helper.make_node(
    "If",
    inputs=["cond"],
    outputs=["res"],
    then_branch=then_body,
    else_branch=else_body,
)

cond = np.array(1).astype(bool)
res = x if cond else y
expect(
    if_node,
    inputs=[cond],
    outputs=[res],
    name="test_if",
    opset_imports=[onnx.helper.make_opsetid("", 11)],
)
```

#### _if_seq

```python
import numpy as np
import onnx

# Given a bool scalar input cond.
# return constant sequence x if cond is True, otherwise return constant sequence y.

then_out = onnx.helper.make_tensor_sequence_value_info(
    "then_out", onnx.TensorProto.FLOAT, shape=[5]
)
else_out = onnx.helper.make_tensor_sequence_value_info(
    "else_out", onnx.TensorProto.FLOAT, shape=[5]
)

x = [np.array([1, 2, 3, 4, 5]).astype(np.float32)]
y = [np.array([5, 4, 3, 2, 1]).astype(np.float32)]

then_const_node = onnx.helper.make_node(
    "Constant",
    inputs=[],
    outputs=["x"],
    value=onnx.numpy_helper.from_array(x[0]),
)

then_seq_node = onnx.helper.make_node(
    "SequenceConstruct", inputs=["x"], outputs=["then_out"]
)

else_const_node = onnx.helper.make_node(
    "Constant",
    inputs=[],
    outputs=["y"],
    value=onnx.numpy_helper.from_array(y[0]),
)

else_seq_node = onnx.helper.make_node(
    "SequenceConstruct", inputs=["y"], outputs=["else_out"]
)

then_body = onnx.helper.make_graph(
    [then_const_node, then_seq_node], "then_body", [], [then_out]
)

else_body = onnx.helper.make_graph(
    [else_const_node, else_seq_node], "else_body", [], [else_out]
)

if_node = onnx.helper.make_node(
    "If",
    inputs=["cond"],
    outputs=["res"],
    then_branch=then_body,
    else_branch=else_body,
)

cond = np.array(1).astype(bool)
res = x if cond else y
expect(
    if_node,
    inputs=[cond],
    outputs=[res],
    name="test_if_seq",
    opset_imports=[onnx.helper.make_opsetid("", 13)],
)
```

#### _if_optional

```python
import numpy as np
import onnx

# Given a bool scalar input cond, return an empty optional sequence of
# tensor if True, return an optional sequence with value x
# (the input optional sequence) otherwise.

ten_in_tp = onnx.helper.make_tensor_type_proto(
    onnx.TensorProto.FLOAT, shape=[5]
)
seq_in_tp = onnx.helper.make_sequence_type_proto(ten_in_tp)

then_out_tensor_tp = onnx.helper.make_tensor_type_proto(
    onnx.TensorProto.FLOAT, shape=[5]
)
then_out_seq_tp = onnx.helper.make_sequence_type_proto(then_out_tensor_tp)
then_out_opt_tp = onnx.helper.make_optional_type_proto(then_out_seq_tp)
then_out = onnx.helper.make_value_info("optional_empty", then_out_opt_tp)

else_out_tensor_tp = onnx.helper.make_tensor_type_proto(
    onnx.TensorProto.FLOAT, shape=[5]
)
else_out_seq_tp = onnx.helper.make_sequence_type_proto(else_out_tensor_tp)
else_out_opt_tp = onnx.helper.make_optional_type_proto(else_out_seq_tp)
else_out = onnx.helper.make_value_info("else_opt", else_out_opt_tp)

x = [np.array([1, 2, 3, 4, 5]).astype(np.float32)]
cond = np.array(0).astype(bool)
res = compute_if_outputs(x, cond)

opt_empty_in = onnx.helper.make_node(
    "Optional", inputs=[], outputs=["optional_empty"], type=seq_in_tp
)

then_body = onnx.helper.make_graph([opt_empty_in], "then_body", [], [then_out])

else_const_node = onnx.helper.make_node(
    "Constant",
    inputs=[],
    outputs=["x"],
    value=onnx.numpy_helper.from_array(x[0]),
)

else_seq_node = onnx.helper.make_node(
    "SequenceConstruct", inputs=["x"], outputs=["else_seq"]
)

else_optional_seq_node = onnx.helper.make_node(
    "Optional", inputs=["else_seq"], outputs=["else_opt"]
)

else_body = onnx.helper.make_graph(
    [else_const_node, else_seq_node, else_optional_seq_node],
    "else_body",
    [],
    [else_out],
)

if_node = onnx.helper.make_node(
    "If",
    inputs=["cond"],
    outputs=["sequence"],
    then_branch=then_body,
    else_branch=else_body,
)

expect(
    if_node,
    inputs=[cond],
    outputs=[res],
    name="test_if_opt",
    output_type_protos=[else_out_opt_tp],
    opset_imports=[onnx.helper.make_opsetid("", 16)],
)
```

```{toctree}
text_diff_If_24_25
```

(l-onnx-op-if-24)=

## If - 24

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `24`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 24**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(float4e2m1))`, `optional(tensor(float8e4m3fn))`, `optional(tensor(float8e4m3fnuz))`, `optional(tensor(float8e5m2))`, `optional(tensor(float8e5m2fnuz))`, `optional(tensor(float8e8m0))`, `optional(tensor(int16))`, `optional(tensor(int32))`, `optional(tensor(int4))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint32))`, `optional(tensor(uint4))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(float4e2m1))`, `seq(tensor(float8e4m3fn))`, `seq(tensor(float8e4m3fnuz))`, `seq(tensor(float8e5m2))`, `seq(tensor(float8e5m2fnuz))`, `seq(tensor(float8e8m0))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int4))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint4))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(float4e2m1)`, `tensor(float8e4m3fn)`, `tensor(float8e4m3fnuz)`, `tensor(float8e5m2)`, `tensor(float8e5m2fnuz)`, `tensor(float8e8m0)`, `tensor(int16)`, `tensor(int32)`, `tensor(int4)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint4)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv11.
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_23_25
text_diff_If_23_24
```

(l-onnx-op-if-23)=

## If - 23

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `23`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 23**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(float4e2m1))`, `optional(tensor(float8e4m3fn))`, `optional(tensor(float8e4m3fnuz))`, `optional(tensor(float8e5m2))`, `optional(tensor(float8e5m2fnuz))`, `optional(tensor(int16))`, `optional(tensor(int32))`, `optional(tensor(int4))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint32))`, `optional(tensor(uint4))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(float4e2m1))`, `seq(tensor(float8e4m3fn))`, `seq(tensor(float8e4m3fnuz))`, `seq(tensor(float8e5m2))`, `seq(tensor(float8e5m2fnuz))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int4))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint4))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(float4e2m1)`, `tensor(float8e4m3fn)`, `tensor(float8e4m3fnuz)`, `tensor(float8e5m2)`, `tensor(float8e5m2fnuz)`, `tensor(int16)`, `tensor(int32)`, `tensor(int4)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint4)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv11.
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_21_25
text_diff_If_21_24
text_diff_If_21_23
```

(l-onnx-op-if-21)=

## If - 21

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `21`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 21**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(float8e4m3fn))`, `optional(tensor(float8e4m3fnuz))`, `optional(tensor(float8e5m2))`, `optional(tensor(float8e5m2fnuz))`, `optional(tensor(int16))`, `optional(tensor(int32))`, `optional(tensor(int4))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint32))`, `optional(tensor(uint4))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(float8e4m3fn))`, `seq(tensor(float8e4m3fnuz))`, `seq(tensor(float8e5m2))`, `seq(tensor(float8e5m2fnuz))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int4))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint4))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(float8e4m3fn)`, `tensor(float8e4m3fnuz)`, `tensor(float8e5m2)`, `tensor(float8e5m2fnuz)`, `tensor(int16)`, `tensor(int32)`, `tensor(int4)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint4)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv10.
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_19_25
text_diff_If_19_24
text_diff_If_19_23
text_diff_If_19_21
```

(l-onnx-op-if-19)=

## If - 19

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `19`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 19**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(float8e4m3fn))`, `optional(tensor(float8e4m3fnuz))`, `optional(tensor(float8e5m2))`, `optional(tensor(float8e5m2fnuz))`, `optional(tensor(int16))`, `optional(tensor(int32))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint32))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(float8e4m3fn))`, `seq(tensor(float8e4m3fnuz))`, `seq(tensor(float8e5m2))`, `seq(tensor(float8e5m2fnuz))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(float8e4m3fn)`, `tensor(float8e4m3fnuz)`, `tensor(float8e5m2)`, `tensor(float8e5m2fnuz)`, `tensor(int16)`, `tensor(int32)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv9.
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_16_25
text_diff_If_16_24
text_diff_If_16_23
text_diff_If_16_21
text_diff_If_16_19
```

(l-onnx-op-if-16)=

## If - 16

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `16`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 16**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `optional(seq(tensor(bfloat16)))`, `optional(seq(tensor(bool)))`, `optional(seq(tensor(complex128)))`, `optional(seq(tensor(complex64)))`, `optional(seq(tensor(double)))`, `optional(seq(tensor(float)))`, `optional(seq(tensor(float16)))`, `optional(seq(tensor(int16)))`, `optional(seq(tensor(int32)))`, `optional(seq(tensor(int64)))`, `optional(seq(tensor(int8)))`, `optional(seq(tensor(string)))`, `optional(seq(tensor(uint16)))`, `optional(seq(tensor(uint32)))`, `optional(seq(tensor(uint64)))`, `optional(seq(tensor(uint8)))`, `optional(tensor(bfloat16))`, `optional(tensor(bool))`, `optional(tensor(complex128))`, `optional(tensor(complex64))`, `optional(tensor(double))`, `optional(tensor(float))`, `optional(tensor(float16))`, `optional(tensor(int16))`, `optional(tensor(int32))`, `optional(tensor(int64))`, `optional(tensor(int8))`, `optional(tensor(string))`, `optional(tensor(uint16))`, `optional(tensor(uint32))`, `optional(tensor(uint64))`, `optional(tensor(uint8))`, `seq(tensor(bfloat16))`, `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bfloat16)`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(int16)`, `tensor(int32)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor, Sequence(Tensor), Optional(Tensor), and Optional(Sequence(Tensor)) types up to IRv4.
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_13_25
text_diff_If_13_24
text_diff_If_13_23
text_diff_If_13_21
text_diff_If_13_19
text_diff_If_13_16
```

(l-onnx-op-if-13)=

## If - 13

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `13`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 13**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `seq(tensor(bool))`, `seq(tensor(complex128))`, `seq(tensor(complex64))`, `seq(tensor(double))`, `seq(tensor(float))`, `seq(tensor(float16))`, `seq(tensor(int16))`, `seq(tensor(int32))`, `seq(tensor(int64))`, `seq(tensor(int8))`, `seq(tensor(string))`, `seq(tensor(uint16))`, `seq(tensor(uint32))`, `seq(tensor(uint64))`, `seq(tensor(uint8))`, `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(int16)`, `tensor(int32)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor and Sequence types
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_11_25
text_diff_If_11_24
text_diff_If_11_23
text_diff_If_11_21
text_diff_If_11_19
text_diff_If_11_16
text_diff_If_11_13
```

(l-onnx-op-if-11)=

## If - 11

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `11`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 11**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same data type. The `then_branch` and `else_branch` may produce tensors with the same element type and different shapes. If corresponding outputs from the then-branch and the else-branch have static shapes S1 and S2, then the shape of the corresponding output variable of the if-node (if present) must be compatible with both S1 and S2 as it represents the union of both possible shapes.For example, if in a model file, the first output of `then_branch` is typed float tensor with shape [2] and the first output of `else_branch` is another float tensor with shape [3], If's first output should have (a) no shape set, or (b) a shape of rank 1 with neither `dim_value` nor `dim_param` set, or (c) a shape of rank 1 with a unique `dim_param`. In contrast, the first output cannot have the shape [2] since [2] and [3] are not compatible.

### Type Constraints

* **V** in ( `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(int16)`, `tensor(int32)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor types
* **B** in ( `tensor(bool)` ):

  Only bool

```{toctree}
text_diff_If_1_25
text_diff_If_1_24
text_diff_If_1_23
text_diff_If_1_21
text_diff_If_1_19
text_diff_If_1_16
text_diff_If_1_13
text_diff_If_1_11
```

(l-onnx-op-if-1)=

## If - 1

### Version

- **name**: [If (GitHub)](https://github.com/onnx/onnx/blob/main/docs/Operators.md#If)
- **domain**: `main`
- **since_version**: `1`
- **function**: `False`
- **support_level**: `SupportType.COMMON`
- **shape inference**: `True`

This version of the operator has been available
**since version 1**.

### Summary

If conditional

### Attributes

* **else_branch - GRAPH** (required) :

  Graph to run if condition is false. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the then_branch.

* **then_branch - GRAPH** (required) :

  Graph to run if condition is true. Has N outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.

### Inputs

- **cond** (heterogeneous) - **B**:

  Condition for the if. The tensor must contain a single element.

### Outputs

Between 1 and 2147483647 outputs.

- **outputs** (variadic) - **V**:

  Values that are live-out to the enclosing scope. The return values in the `then_branch` and `else_branch` must be of the same shape and same data type.

### Type Constraints

* **V** in ( `tensor(bool)`, `tensor(complex128)`, `tensor(complex64)`, `tensor(double)`, `tensor(float)`, `tensor(float16)`, `tensor(int16)`, `tensor(int32)`, `tensor(int64)`, `tensor(int8)`, `tensor(string)`, `tensor(uint16)`, `tensor(uint32)`, `tensor(uint64)`, `tensor(uint8)` ):

  All Tensor types
* **B** in ( `tensor(bool)` ):

  Only bool