Passed fmt to match latest V syntax format

autograd/context.v

@@ -6,7 +6,7 @@ import vtl
 // a number of operations. Variables that interact with each
 // other must belong to the same context, or state will be
 // lost while tracking operations done.
-[heap]
+@[heap]
 pub struct Context[T] {
 pub mut:
 	// A list of all variables present in an operation.
@@ -45,7 +45,7 @@ pub fn (mut ctx Context[T]) pop[T]() !&Node[T] {
 	return ctx.nodes.pop()
 }
 
-[params]
+@[params]
 pub struct ContextVariableData {
 	requires_grad bool = true
 }

autograd/node.v

@@ -3,7 +3,7 @@ module autograd
 // Node is a member of a computational graph that contains
 // a reference to a gate, as well as the parents of the operation
 // and the payload that resulted from the operation.
-[heap]
+@[heap]
 pub struct Node[T] {
 pub:
 	// A Gate containing a backwards and cache function for

autograd/payload.v

@@ -3,7 +3,7 @@ module autograd
 // Payload is a simple wrapper around a Variable.  It
 // is only abstracted out to be a bit more explicit that
 // it is being passed around through an operation
-[heap]
+@[heap]
 pub struct Payload[T] {
 pub:
 	// Contents of the paylod

autograd/variable.v

@@ -9,7 +9,7 @@ import vtl
 // This is the fundamental object used in automatic
 // differentiation, as well as the neural network aspects
 // of VTL
-[heap]
+@[heap]
 pub struct Variable[T] {
 pub mut:
 	// The value of the Variable.  This should not be edited outside
@@ -30,7 +30,7 @@ pub mut:
 	requires_grad bool
 }
 
-[params]
+@[params]
 pub struct VariableData {
 	requires_grad bool = true
 }

datasets/loader.v

@@ -13,7 +13,7 @@ fn get_cache_dir(subdir ...string) string {
 	return os.join_path(cache_dir, ...subdir)
 }
 
-[params]
+@[params]
 struct RawDownload {
 	url    string
 	target string
@@ -39,7 +39,7 @@ fn load_from_url(data RawDownload) ! {
 	http.download_file(data.url, cache_file_path)!
 }
 
-[params]
+@[params]
 struct DatasetDownload {
 	dataset    string
 	baseurl    string

ml/metrics/common_error_functions.v

@@ -4,18 +4,18 @@ import math
 import vtl
 import vtl.stats
 
-[inline]
+@[inline]
 pub fn squared_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	diff := y.subtract(y_true)!
 	return diff.multiply(diff)
 }
 
-[inline]
+@[inline]
 pub fn mean_squared_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !T {
 	return stats.mean[T](squared_error[T](y, y_true)!)
 }
 
-[inline]
+@[inline]
 pub fn relative_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	return y.nmap([y_true], fn [T](vals []T, i []int) T {
 		denom := math.max(math.abs(vals[1]), math.abs(vals[0]))
@@ -26,17 +26,17 @@ pub fn relative_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !&vtl.Tensor[T
 	})
 }
 
-[inline]
+@[inline]
 pub fn mean_relative_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !T {
 	return stats.mean[T](relative_error[T](y, y_true)!)
 }
 
-[inline]
+@[inline]
 pub fn absolute_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	return y_true.subtract(y)!.abs()
 }
 
-[inline]
+@[inline]
 pub fn mean_absolute_error[T](y &vtl.Tensor[T], y_true &vtl.Tensor[T]) !T {
 	return stats.mean[T](absolute_error[T](y, y_true)!)
 }

nn/internal/activation.v

@@ -4,37 +4,37 @@ import math
 import vtl
 
 // tanh squashes a real-valued number to the range [-1, 1]
-[inline]
+@[inline]
 pub fn tanh[T](x &vtl.Tensor[T]) &vtl.Tensor[T] {
 	return vtl.tanh(x)
 }
 
 // deriv_tanh computes the derivative of tanh
-[inline]
+@[inline]
 pub fn deriv_tanh[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	return gradient.nmap([cached], fn [T](vals []T, i []int) T {
 		return vals[0] * (vtl.cast[T](1) - vals[1] * vals[1])
 	})
 }
 
 // sigmoid takes a real-valued number and squashes it to the range [0, 1]
-[inline]
+@[inline]
 pub fn sigmoid[T](x &vtl.Tensor[T]) &vtl.Tensor[T] {
 	return x.map(fn [T](val T, i []int) T {
 		return vtl.cast[T](1) / vtl.cast[T](1) + vtl.cast[T](math.exp(vtl.cast[f64](val)))
 	})
 }
 
 // deriv_sigmoid computes the derivative of sigmoid
-[inline]
+@[inline]
 pub fn deriv_sigmoid[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	return gradient.nmap([cached], fn [T](vals []T, i []int) T {
 		return vals[0] * (vtl.cast[T](1) - vals[0])
 	})
 }
 
 // relu activation function
-[inline]
+@[inline]
 pub fn relu[T](x &vtl.Tensor[T]) &vtl.Tensor[T] {
 	return x.map(fn [T](val T, i []int) T {
 		if val < 0 {
@@ -45,7 +45,7 @@ pub fn relu[T](x &vtl.Tensor[T]) &vtl.Tensor[T] {
 }
 
 // deriv_relu computes the derivate of relu
-[inline]
+@[inline]
 pub fn deriv_relu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	return gradient.nmap([cached], fn [T](vals []T, i []int) T {
 		if vals[0] < 0 {
@@ -56,7 +56,7 @@ pub fn deriv_relu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T]) !&vtl.Tenso
 }
 
 // leaky_relu activation function
-[inline]
+@[inline]
 pub fn leaky_relu[T](x &vtl.Tensor[T], alpha T) &vtl.Tensor[T] {
 	return x.map(fn [alpha] [T](val T, i []int) T {
 		if val < 0 {
@@ -67,7 +67,7 @@ pub fn leaky_relu[T](x &vtl.Tensor[T], alpha T) &vtl.Tensor[T] {
 }
 
 // deriv_leaky_relu computes the derivative of leaky_relu
-[inline]
+@[inline]
 pub fn deriv_leaky_relu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T], alpha T) !&vtl.Tensor[T] {
 	return gradient.nmap([cached], fn [alpha] [T](vals []T, i []int) T {
 		if vals[0] < 0 {
@@ -78,7 +78,7 @@ pub fn deriv_leaky_relu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T], alpha
 }
 
 // elu activation function
-[inline]
+@[inline]
 pub fn elu[T](x &vtl.Tensor[T], alpha T) &vtl.Tensor[T] {
 	return x.map(fn [alpha] [T](val T, i []int) T {
 		if val < 0 {
@@ -89,7 +89,7 @@ pub fn elu[T](x &vtl.Tensor[T], alpha T) &vtl.Tensor[T] {
 }
 
 // deriv_elu computes the derivative of elu
-[inline]
+@[inline]
 pub fn deriv_elu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T], alpha T) !&vtl.Tensor[T] {
 	return gradient.nmap([cached], fn [alpha] [T](vals []T, i []int) T {
 		if vals[0] < 0 {
@@ -101,7 +101,7 @@ pub fn deriv_elu[T](gradient &vtl.Tensor[T], cached &vtl.Tensor[T], alpha T) !&v
 
 // sigmoid_cross_entropy computes the sigmoid cross entropy between
 // the labels and the predictions
-[inline]
+@[inline]
 pub fn sigmoid_cross_entropy[T](input &vtl.Tensor[T], target &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	sum := input.nreduce([target], vtl.cast[T](0), fn [T](acc T, vals []T, i []int) T {
 		next := -(vals[1] * vtl.cast[T](math.max(f64(0), f64(vals[0])))) - vtl.cast[T](math.log(
@@ -114,7 +114,7 @@ pub fn sigmoid_cross_entropy[T](input &vtl.Tensor[T], target &vtl.Tensor[T]) !&v
 }
 
 // mse squared error between the labels and the predictions
-[inline]
+@[inline]
 pub fn mse[T](input &vtl.Tensor[T], target &vtl.Tensor[T]) !&vtl.Tensor[T] {
 	sum := input.nreduce([target], vtl.cast[T](0), fn [T](acc T, vals []T, i []int) T {
 		next := vtl.cast[T](math.pow(f64(vals[0] - vals[1]), 2.0))

nn/layers/dropout.v

@@ -6,7 +6,7 @@ import vtl.nn.internal
 import vtl.nn.gates.layers
 import vtl.nn.types
 
-[params]
+@[params]
 pub struct DropoutLayerConfig {
 	prob f64 = 0.5
 }

nn/layers/elu.v

@@ -6,7 +6,7 @@ import vtl.nn.internal
 import vtl.nn.gates.activation
 import vtl.nn.types
 
-[params]
+@[params]
 pub struct EluLayerConfig {
 	alpha f64 = 0.01
 }

nn/optimizers/adam.v

@@ -18,7 +18,7 @@ pub mut:
 	second_moments []&vtl.Tensor[T]
 }
 
-[params]
+@[params]
 pub struct AdamOptimizerConfig {
 	learning_rate f64 = 0.001
 	beta1         f64 = 0.9

nn/optimizers/sgd.v

@@ -11,7 +11,7 @@ pub mut:
 	params []&autograd.Variable[T]
 }
 
-[params]
+@[params]
 pub struct SgdOptimizerConfig {
 	learning_rate f64 = 0.001
 }

nn/types/optimizer.v

@@ -5,7 +5,7 @@ import vtl.autograd
 // Optimizer is a generic interface for all optimizers.
 pub interface Optimizer[T] {
 mut:
-	params []&autograd.Variable[T]
+	params        []&autograd.Variable[T]
 	learning_rate f64
 	update() !
 	build_params(layers Layer[T])

src/assignment.v

@@ -1,21 +1,21 @@
 module vtl
 
 // set copies a scalar value into a Tensor at the provided index
-[inline]
+@[inline]
 pub fn (mut t Tensor[T]) set[T](index []int, val T) {
 	offset := t.offset_index(index)
 	t.data.set[T](offset, val)
 }
 
 // set_nth copies a scalar value into a Tensor at the provided offset
-[inline]
+@[inline]
 pub fn (mut t Tensor[T]) set_nth[T](n int, val T) {
 	index := t.nth_index(n)
 	t.set[T](index, val)
 }
 
 // fill fills an entire Tensor with a given value
-[inline]
+@[inline]
 pub fn (mut t Tensor[T]) fill[T](val T) &Tensor[T] {
 	t.data.fill[T](val)
 	return t

src/broadcast.v

@@ -120,7 +120,7 @@ fn broadcast_shapes(args ...[]int) []int {
 }
 
 // broadcast2 broadcasts two Tensors against each other
-[inline]
+@[inline]
 pub fn broadcast2[T](a &Tensor[T], b &Tensor[T]) !(&Tensor[T], &Tensor[T]) {
 	shape := a.broadcastable(b)!
 	r1 := a.broadcast_to(shape)!
@@ -129,7 +129,7 @@ pub fn broadcast2[T](a &Tensor[T], b &Tensor[T]) !(&Tensor[T], &Tensor[T]) {
 }
 
 // broadcast3 broadcasts three Tensors against each other
-[inline]
+@[inline]
 pub fn broadcast3[T](a &Tensor[T], b &Tensor[T], c &Tensor[T]) !(&Tensor[T], &Tensor[T], &Tensor[T]) {
 	shape := broadcast_shapes(a.shape, b.shape, c.shape)
 	r1 := a.broadcast_to(shape)!
@@ -139,7 +139,7 @@ pub fn broadcast3[T](a &Tensor[T], b &Tensor[T], c &Tensor[T]) !(&Tensor[T], &Te
 }
 
 // broadcast_n broadcasts N Tensors against each other
-[inline]
+@[inline]
 pub fn broadcast_n[T](ts []&Tensor[T]) ![]&Tensor[T] {
 	shapes := ts.map(it.shape)
 	shape := broadcast_shapes(...shapes)

src/build.v

@@ -2,7 +2,7 @@ module vtl
 
 import vtl.storage
 
-[params]
+@[params]
 pub struct TensorData {
 pub:
 	memory MemoryFormat = .row_major

src/cast.v

@@ -3,7 +3,7 @@ module vtl
 // as_bool casts the Tensor to a Tensor of bools.
 // If the original Tensor is not a Tensor of bools, then each value is cast to a bool,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_bool[T]() &Tensor[bool] {
 	$if T is bool {
 		return t
@@ -23,7 +23,7 @@ pub fn (t &Tensor[T]) as_bool[T]() &Tensor[bool] {
 // as_f32 casts the Tensor to a Tensor of f32s.
 // If the original Tensor is not a Tensor of f32s, then each value is cast to a f32,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_f32[T]() &Tensor[f32] {
 	$if T is f32 {
 		return t
@@ -43,7 +43,7 @@ pub fn (t &Tensor[T]) as_f32[T]() &Tensor[f32] {
 // as_f64 casts the Tensor to a Tensor of f64s.
 // If the original Tensor is not a Tensor of f64s, then each value is cast to a f64,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_f64[T]() &Tensor[f64] {
 	$if T is f64 {
 		return t
@@ -63,7 +63,7 @@ pub fn (t &Tensor[T]) as_f64[T]() &Tensor[f64] {
 // as_i16 casts the Tensor to a Tensor of i16 values.
 // If the original Tensor is not a Tensor of i16s, then each value is cast to a i16,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_i16[T]() &Tensor[i16] {
 	$if T is i16 {
 		return t
@@ -83,7 +83,7 @@ pub fn (t &Tensor[T]) as_i16[T]() &Tensor[i16] {
 // as_i8 casts the Tensor to a Tensor of i8 values.
 // If the original Tensor is not a Tensor of i8s, then each value is cast to a i8,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_i8[T]() &Tensor[i8] {
 	$if T is i8 {
 		return t
@@ -103,7 +103,7 @@ pub fn (t &Tensor[T]) as_i8[T]() &Tensor[i8] {
 // as_int casts the Tensor to a Tensor of ints.
 // If the original Tensor is not a Tensor of ints, then each value is cast to a int,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_int[T]() &Tensor[int] {
 	$if T is int {
 		return t
@@ -123,7 +123,7 @@ pub fn (t &Tensor[T]) as_int[T]() &Tensor[int] {
 // as_string casts the Tensor to a Tensor of string values.
 // If the original Tensor is not a Tensor of strings, then each value is cast to a string,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_string[T]() &Tensor[string] {
 	$if T is string {
 		return t
@@ -143,7 +143,7 @@ pub fn (t &Tensor[T]) as_string[T]() &Tensor[string] {
 // as_u8 casts the Tensor to a Tensor of u8 values.
 // If the original Tensor is not a Tensor of u8s, then each value is cast to a u8,
 // otherwise the original Tensor is returned.
-[inline]
+@[inline]
 pub fn (t &Tensor[T]) as_u8[T]() &Tensor[u8] {
 	$if T is u8 {
 		return t