refactor: merge tensor module into value

backends/tract/tests/session.rs

@@ -4,7 +4,7 @@ use image::{ImageBuffer, Luma, Pixel, imageops::FilterType};
 use ort::{
 	inputs,
 	session::{Session, builder::GraphOptimizationLevel},
-	tensor::ArrayExtensions,
+	util::ArrayExt,
 	value::TensorRef
 };
 

examples/cudarc/cudarc.rs

@@ -7,8 +7,7 @@ use ort::{
 	ep,
 	memory::{AllocationDevice, AllocatorType, MemoryInfo, MemoryType},
 	session::Session,
-	tensor::Shape,
-	value::TensorRefMut
+	value::{Shape, TensorRefMut}
 };
 use show_image::{AsImageView, WindowOptions, event};
 use tracing_subscriber::{layer::SubscriberExt, util::SubscriberInitExt};

examples/custom-ops/custom-ops.rs

@@ -5,8 +5,7 @@ use ort::{
 		kernel::{Kernel, KernelAttributes, KernelContext}
 	},
 	session::Session,
-	tensor::TensorElementType,
-	value::Tensor
+	value::{Tensor, TensorElementType}
 };
 
 struct CustomOpOne;

src/editor/tests.rs

@@ -5,8 +5,7 @@ use crate::{
 	inputs,
 	memory::Allocator,
 	session::builder::SessionBuilder,
-	tensor::{Shape, SymbolicDimensions, TensorElementType},
-	value::{Tensor, ValueType}
+	value::{Shape, SymbolicDimensions, Tensor, TensorElementType, ValueType}
 };
 
 #[test]

src/lib.rs

@@ -36,7 +36,6 @@ pub mod logging;
 pub mod memory;
 pub mod operator;
 pub mod session;
-pub mod tensor;
 #[cfg(feature = "training")]
 #[cfg_attr(docsrs, doc(cfg(feature = "training")))]
 pub mod training;

src/operator/io.rs

@@ -1,4 +1,4 @@
-use crate::{memory::MemoryType, tensor::TensorElementType};
+use crate::{memory::MemoryType, value::TensorElementType};
 
 #[repr(i32)]
 #[derive(Debug, Default, Clone, Copy, PartialEq, Eq)]

src/operator/kernel.rs

@@ -12,9 +12,8 @@ use crate::{
 	logging::Logger,
 	memory::{Allocator, MemoryInfo, MemoryType},
 	ortsys,
-	tensor::Shape,
 	util::with_cstr,
-	value::{DowncastableTarget, DynValue, Outlet, Value, ValueRef, ValueRefMut, ValueType}
+	value::{DowncastableTarget, DynValue, Outlet, Shape, Value, ValueRef, ValueRefMut, ValueType}
 };
 
 pub trait Kernel {

src/operator/tests.rs

@@ -6,8 +6,7 @@ use crate::{
 		kernel::{Kernel, KernelAttributes, KernelContext}
 	},
 	session::Session,
-	tensor::TensorElementType,
-	value::Tensor
+	value::{Tensor, TensorElementType}
 };
 
 struct CustomOpOne;

src/session/mod.rs

@@ -194,7 +194,7 @@ impl Session {
 	///
 	/// ```
 	/// # use std::sync::Arc;
-	/// # use ort::{session::{run_options::RunOptions, Session}, tensor::TensorElementType, value::{Value, ValueType, TensorRef}};
+	/// # use ort::{session::{run_options::RunOptions, Session}, value::{Value, ValueType, TensorRef, TensorElementType}};
 	/// # fn main() -> ort::Result<()> {
 	/// let mut session = Session::builder()?.commit_from_file("tests/data/upsample.onnx")?;
 	/// let input = ndarray::Array4::<f32>::zeros((1, 64, 64, 3));
@@ -223,7 +223,7 @@ impl Session {
 	/// ```no_run
 	/// # // no_run because upsample.onnx is too simple of a model for the termination signal to be reliable enough
 	/// # use std::sync::Arc;
-	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef}, tensor::TensorElementType};
+	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef, TensorElementType}};
 	/// # fn main() -> ort::Result<()> {
 	/// # 	let mut session = Session::builder()?.commit_from_file("tests/data/upsample.onnx")?;
 	/// # 	let input = Value::from_array(ndarray::Array4::<f32>::zeros((1, 64, 64, 3)))?;
@@ -385,7 +385,7 @@ impl Session {
 	///
 	/// ```
 	/// # use std::sync::Arc;
-	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef}, tensor::TensorElementType};
+	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef, TensorElementType}};
 	/// # fn main() -> ort::Result<()> { tokio::runtime::Builder::new_current_thread().enable_all().build().unwrap().block_on(async {
 	/// let mut session = Session::builder()?.with_intra_threads(2)?.commit_from_file("tests/data/upsample.onnx")?;
 	/// let input = ndarray::Array4::<f32>::zeros((1, 64, 64, 3));
@@ -501,7 +501,7 @@ impl Session {
 	///
 	/// ```
 	/// # use std::sync::Arc;
-	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef}, tensor::TensorElementType};
+	/// # use ort::{session::{Session, run_options::RunOptions}, value::{Value, ValueType, TensorRef, TensorElementType}};
 	/// # fn main() -> ort::Result<()> { tokio::runtime::Builder::new_current_thread().enable_all().build().unwrap().block_on(async {
 	/// let mut session = Session::builder()?.with_intra_threads(2)?.commit_from_file("tests/data/upsample.onnx")?;
 	/// let input = ndarray::Array4::<f32>::zeros((1, 64, 64, 3));
@@ -614,7 +614,7 @@ impl Session {
 	///
 	/// ```
 	/// # use std::sync::Arc;
-	/// # use ort::{session::{run_options::RunOptions, Session, WorkloadType}, tensor::TensorElementType, value::{Value, ValueType, TensorRef}};
+	/// # use ort::{session::{run_options::RunOptions, Session, WorkloadType}, value::{Value, ValueType, TensorRef, TensorElementType}};
 	/// # fn main() -> ort::Result<()> {
 	/// let mut session = Session::builder()?.commit_from_file("tests/data/upsample.onnx")?;
 	/// session.set_workload_type(WorkloadType::Efficient)?;

src/test_util.rs

@@ -5,7 +5,7 @@ pub mod mnist {
 
 	use crate::{
 		Result,
-		tensor::ArrayExtensions,
+		util::ArrayExt,
 		value::{TensorValueTypeMarker, Value}
 	};
 

src/training/trainer.rs

@@ -15,9 +15,8 @@ use crate::{
 	memory::Allocator,
 	ortsys,
 	session::{RunOptions, SessionInputValue, SessionInputs, SessionOutputs, builder::SessionBuilder},
-	tensor::IntoTensorElementType,
 	util::{char_p_to_string, with_cstr_ptr_array},
-	value::{Tensor, Value}
+	value::{IntoTensorElementType, Tensor, Value}
 };
 
 #[derive(Debug)]

src/util/mod.rs

@@ -24,6 +24,11 @@ pub(crate) use self::{
 	once_lock::OnceLock,
 	stack::*
 };
+#[cfg(feature = "ndarray")]
+mod ndarray;
+#[cfg(feature = "ndarray")]
+#[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
+pub use self::ndarray::ArrayExt;
 
 /// Preloads the dynamic library at the given `path`.
 ///

src/util/ndarray.rs

@@ -1,10 +1,8 @@
-//! Helper traits to extend [`ndarray`] functionality.
-
 /// Trait extending [`ndarray::ArrayBase`](https://docs.rs/ndarray/latest/ndarray/struct.ArrayBase.html)
 /// with useful tensor operations.
-#[cfg(feature = "std")]
-#[cfg_attr(docsrs, doc(cfg(feature = "std")))]
-pub trait ArrayExtensions<S, T, D> {
+#[cfg(feature = "ndarray")]
+#[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
+pub trait ArrayExt<S, T, D> {
 	/// Calculate the [softmax](https://en.wikipedia.org/wiki/Softmax_function) of the tensor along a given axis.
 	fn softmax(&self, axis: ndarray::Axis) -> ndarray::Array<T, D>
 	where
@@ -14,9 +12,9 @@ pub trait ArrayExtensions<S, T, D> {
 		T: ndarray::NdFloat + core::ops::SubAssign + core::ops::DivAssign;
 }
 
-#[cfg(feature = "std")]
-#[cfg_attr(docsrs, doc(cfg(feature = "std")))]
-impl<S, T, D> ArrayExtensions<S, T, D> for ndarray::ArrayBase<S, D>
+#[cfg(feature = "ndarray")]
+#[cfg_attr(docsrs, doc(cfg(feature = "ndarray")))]
+impl<S, T, D> ArrayExt<S, T, D> for ndarray::ArrayBase<S, D>
 where
 	D: ndarray::RemoveAxis,
 	S: ndarray::RawData + ndarray::Data + ndarray::RawData<Elem = T>,

src/value/impl_map.rs

@@ -13,14 +13,13 @@ use std::collections::HashMap;
 
 use super::{
 	DowncastableTarget, DynValue, Value, ValueInner, ValueRef, ValueRefMut, ValueType, ValueTypeMarker,
-	impl_tensor::{DynTensor, Tensor}
+	impl_tensor::{DynTensor, IntoTensorElementType, PrimitiveTensorElementType, Tensor, TensorElementType}
 };
 use crate::{
 	AsPointer, ErrorCode,
 	error::{Error, Result},
 	memory::Allocator,
-	ortsys,
-	tensor::{IntoTensorElementType, PrimitiveTensorElementType, TensorElementType}
+	ortsys
 };
 
 pub trait MapValueTypeMarker: ValueTypeMarker {

src/value/impl_sequence.rs

@@ -109,8 +109,9 @@ impl<T: ValueTypeMarker + DowncastableTarget + Debug + Sized + 'static> Value<Se
 	/// This `Value<T>` must be either a [`Tensor`] or [`Map`].
 	///
 	/// ```
-	/// # use ort::value::{Sequence, Tensor};
+	/// # use ort::{memory:: Allocator, value::{Sequence, Tensor}};
 	/// # fn main() -> ort::Result<()> {
+	/// # let allocator = Allocator::default();
 	/// let tensor1 = Tensor::<f32>::new(&allocator, [1_usize, 128, 128, 3])?;
 	/// let tensor2 = Tensor::<f32>::new(&allocator, [1_usize, 224, 224, 3])?;
 	/// let value = Sequence::new([tensor1, tensor2])?;

src/value/impl_tensor/create.rs

@@ -11,13 +11,12 @@ use core::{
 #[cfg(feature = "ndarray")]
 use ndarray::{ArcArray, Array, ArrayView, ArrayViewMut, CowArray, Dimension};
 
-use super::{DynTensor, Tensor, TensorRef, TensorRefMut};
+use super::{DynTensor, PrimitiveTensorElementType, Shape, SymbolicDimensions, Tensor, TensorElementType, TensorRef, TensorRefMut, Utf8Data};
 use crate::{
 	AsPointer,
 	error::{Error, ErrorCode, Result},
 	memory::{Allocator, MemoryInfo},
 	ortsys,
-	tensor::{PrimitiveTensorElementType, Shape, SymbolicDimensions, TensorElementType, Utf8Data},
 	value::{Value, ValueInner, ValueType}
 };
 

src/value/impl_tensor/extract.rs

@@ -11,12 +11,11 @@ use core::{
 	slice
 };
 
-use super::{DynTensor, Tensor, TensorValueTypeMarker};
+use super::{DynTensor, PrimitiveTensorElementType, Shape, Tensor, TensorElementType, TensorValueTypeMarker};
 use crate::{
 	AsPointer,
 	error::{Error, ErrorCode, Result},
 	ortsys,
-	tensor::{PrimitiveTensorElementType, Shape, TensorElementType},
 	value::{Value, ValueType}
 };
 

src/value/impl_tensor/mod.rs

@@ -1,7 +1,9 @@
 #[cfg(not(target_arch = "wasm32"))] // `ort-web` does not support synchronous `Run`, which `.clone()` depends on.
 mod copy;
 mod create;
+mod element;
 mod extract;
+mod shape;
 
 use alloc::sync::Arc;
 use core::{
@@ -11,14 +13,17 @@ use core::{
 	ptr::{self}
 };
 
-pub use self::create::{OwnedTensorArrayData, TensorArrayData, TensorArrayDataMut, TensorArrayDataParts, ToShape};
+pub use self::{
+	create::{OwnedTensorArrayData, TensorArrayData, TensorArrayDataMut, TensorArrayDataParts, ToShape},
+	element::{IntoTensorElementType, PrimitiveTensorElementType, TensorElementType, Utf8Data},
+	shape::{Shape, SymbolicDimensions}
+};
 use super::{DowncastableTarget, DynValue, Value, ValueInner, ValueRef, ValueRefMut, ValueType, ValueTypeMarker};
 use crate::{
 	AsPointer,
 	error::Result,
 	memory::{Allocator, MemoryInfo},
-	ortsys,
-	tensor::{IntoTensorElementType, Shape, SymbolicDimensions, TensorElementType}
+	ortsys
 };
 
 pub trait TensorValueTypeMarker: ValueTypeMarker {
@@ -86,7 +91,7 @@ impl DynTensor {
 	/// This can be used to create a tensor with data on a certain device. For example, to create a tensor with pinned
 	/// (CPU) memory for use with CUDA:
 	/// ```no_run
-	/// # use ort::{memory::{Allocator, MemoryInfo, MemoryType, AllocationDevice, AllocatorType}, session::Session, tensor::TensorElementType, value::DynTensor};
+	/// # use ort::{memory::{Allocator, MemoryInfo, MemoryType, AllocationDevice, AllocatorType}, session::Session, value::{DynTensor, TensorElementType}};
 	/// # fn main() -> ort::Result<()> {
 	/// # let session = Session::builder()?.commit_from_file("tests/data/upsample.onnx")?;
 	/// let allocator = Allocator::new(
@@ -345,10 +350,9 @@ mod tests {
 	#[cfg(feature = "ndarray")]
 	use ndarray::{ArcArray1, Array1, CowArray};
 
-	use super::Tensor;
+	use super::{Shape, SymbolicDimensions, Tensor, TensorElementType};
 	use crate::{
 		memory::Allocator,
-		tensor::{Shape, SymbolicDimensions, TensorElementType},
 		value::{TensorRef, ValueType}
 	};
 

src/value/impl_tensor/shape.rs

@@ -1,9 +1,5 @@
 //! Traits and types related to [`Tensor`](crate::value::Tensor)s.
 
-#[cfg(feature = "ndarray")]
-mod ndarray;
-mod types;
-
 use alloc::{string::String, vec::Vec};
 use core::{
 	fmt,
@@ -12,10 +8,6 @@ use core::{
 
 use smallvec::{SmallVec, smallvec};
 
-#[cfg(all(feature = "ndarray", feature = "std"))]
-pub use self::ndarray::ArrayExtensions;
-pub use self::types::{IntoTensorElementType, PrimitiveTensorElementType, TensorElementType, Utf8Data};
-
 #[derive(Default, Clone, PartialEq, Eq)]
 pub struct Shape {
 	inner: SmallVec<[i64; 4]>

src/value/mod.rs

@@ -37,8 +37,9 @@ pub use self::{
 		DynSequence, DynSequenceRef, DynSequenceRefMut, DynSequenceValueType, Sequence, SequenceRef, SequenceRefMut, SequenceValueType, SequenceValueTypeMarker
 	},
 	impl_tensor::{
-		DefiniteTensorValueTypeMarker, DynTensor, DynTensorRef, DynTensorRefMut, DynTensorValueType, OwnedTensorArrayData, Tensor, TensorArrayData,
-		TensorArrayDataMut, TensorArrayDataParts, TensorRef, TensorRefMut, TensorValueType, TensorValueTypeMarker, ToShape
+		DefiniteTensorValueTypeMarker, DynTensor, DynTensorRef, DynTensorRefMut, DynTensorValueType, IntoTensorElementType, OwnedTensorArrayData,
+		PrimitiveTensorElementType, Shape, SymbolicDimensions, Tensor, TensorArrayData, TensorArrayDataMut, TensorArrayDataParts, TensorElementType, TensorRef,
+		TensorRefMut, TensorValueType, TensorValueTypeMarker, ToShape, Utf8Data
 	},
 	r#type::{Outlet, ValueType}
 };
@@ -481,7 +482,6 @@ impl<Type: ValueTypeMarker + ?Sized> AsPointer for Value<Type> {
 #[cfg(test)]
 mod tests {
 	use super::{DynTensorValueType, Map, Sequence, Tensor, TensorRef, TensorRefMut, TensorValueType};
-	use crate::memory::Allocator;
 
 	#[test]
 	fn test_casting_tensor() -> crate::Result<()> {

src/value/type.rs

@@ -12,39 +12,45 @@ use smallvec::{SmallVec, smallvec};
 
 use crate::{
 	Result, ortsys,
-	tensor::{Shape, SymbolicDimensions, TensorElementType},
-	util::{self, run_on_drop, with_cstr, with_cstr_ptr_array}
+	util::{self, run_on_drop, with_cstr, with_cstr_ptr_array},
+	value::{Shape, SymbolicDimensions, TensorElementType}
 };
 
 /// The type of a [`Value`][super::Value], or a session input/output.
 ///
 /// ```
 /// # use std::sync::Arc;
-/// # use ort::{session::Session, tensor::{Shape, SymbolicDimensions}, value::{ValueType, Tensor}, tensor::TensorElementType};
+/// # use ort::{session::Session, value::{ValueType, Tensor, Shape, SymbolicDimensions, TensorElementType}};
 /// # fn main() -> ort::Result<()> {
 /// # 	let session = Session::builder()?.commit_from_file("tests/data/upsample.onnx")?;
 /// // `ValueType`s can be obtained from session inputs/outputs:
 /// let input = &session.inputs()[0];
-/// assert_eq!(input.dtype(), &ValueType::Tensor {
-/// 	ty: TensorElementType::Float32,
-/// 	// Our model's input has 3 dynamic dimensions, represented by -1
-/// 	shape: Shape::new([-1, -1, -1, 3]),
-/// 	// Dynamic dimensions may also have names.
-/// 	dimension_symbols: SymbolicDimensions::new([
-/// 		"unk__31".to_string(),
-/// 		"unk__32".to_string(),
-/// 		"unk__33".to_string(),
-/// 		String::default()
-/// 	])
-/// });
+/// assert_eq!(
+/// 	input.dtype(),
+/// 	&ValueType::Tensor {
+/// 		ty: TensorElementType::Float32,
+/// 		// Our model's input has 3 dynamic dimensions, represented by -1
+/// 		shape: Shape::new([-1, -1, -1, 3]),
+/// 		// Dynamic dimensions may also have names.
+/// 		dimension_symbols: SymbolicDimensions::new([
+/// 			"unk__31".to_string(),
+/// 			"unk__32".to_string(),
+/// 			"unk__33".to_string(),
+/// 			String::default()
+/// 		])
+/// 	}
+/// );
 ///
 /// // ...or by `Value`s created in Rust or output by a session.
 /// let value = Tensor::from_array(([5usize], vec![1_i64, 2, 3, 4, 5].into_boxed_slice()))?;
-/// assert_eq!(value.dtype(), &ValueType::Tensor {
-/// 	ty: TensorElementType::Int64,
-/// 	shape: Shape::new([5]),
-/// 	dimension_symbols: SymbolicDimensions::new([String::default()])
-/// });
+/// assert_eq!(
+/// 	value.dtype(),
+/// 	&ValueType::Tensor {
+/// 		ty: TensorElementType::Int64,
+/// 		shape: Shape::new([5]),
+/// 		dimension_symbols: SymbolicDimensions::new([String::default()])
+/// 	}
+/// );
 /// # 	Ok(())
 /// # }
 /// ```
@@ -215,7 +221,7 @@ impl ValueType {
 	/// Returns the element type of this value type if it is a tensor, or `None` if it is a sequence or map.
 	///
 	/// ```
-	/// # use ort::{tensor::TensorElementType, value::Tensor};
+	/// # use ort::value::{Tensor, TensorElementType};
 	/// # fn main() -> ort::Result<()> {
 	/// let value = Tensor::from_array(([5usize], vec![1_i64, 2, 3, 4, 5].into_boxed_slice()))?;
 	/// assert_eq!(value.dtype().tensor_type(), Some(TensorElementType::Int64));
@@ -374,7 +380,7 @@ mod tests {
 	use super::ValueType;
 	use crate::{
 		ortsys,
-		tensor::{Shape, SymbolicDimensions, TensorElementType}
+		value::{Shape, SymbolicDimensions, TensorElementType}
 	};
 
 	#[test]

tests/leak-check/main.rs

@@ -8,8 +8,7 @@ use ort::{
 		kernel::{Kernel, KernelAttributes, KernelContext}
 	},
 	session::{RunOptions, Session},
-	tensor::TensorElementType,
-	value::Tensor
+	value::{Tensor, TensorElementType}
 };
 
 struct CustomOpOne;

tests/mnist.rs

@@ -4,7 +4,7 @@ use image::{ImageBuffer, Luma, Pixel, imageops::FilterType};
 use ort::{
 	inputs,
 	session::{Session, builder::GraphOptimizationLevel},
-	tensor::ArrayExtensions,
+	util::ArrayExt,
 	value::TensorRef
 };
 

tests/squeezenet.rs

@@ -9,7 +9,7 @@ use ndarray::s;
 use ort::{
 	Error, inputs,
 	session::{Session, builder::GraphOptimizationLevel},
-	tensor::ArrayExtensions,
+	util::ArrayExt,
 	value::TensorRef
 };