Bump burn to 0.12

Cargo.toml

@@ -6,4 +6,4 @@ members = [
 ]
 
 [workspace.dependencies]
-serde = "1.0.192"
+serde = "1"

burn-rl/Cargo.toml

@@ -6,7 +6,7 @@ publish = false
 
 [dependencies]
 rand = "0.8.5"
-burn = { version = "0.11.1", features = ["ndarray", "autodiff"] }
+burn = { version = "0.12.0", features = ["ndarray", "autodiff"] }
 gym-rs = { git = "https://github.com/MathisWellmann/gym-rs.git" }
 ringbuffer = "0.15.0"
 serde = { workspace = true }

burn-rl/src/agent/ppo/agent.rs

@@ -96,6 +96,7 @@ impl<E: Environment, B: AutodiffBackend, M: PPOModel<B> + AutodiffModule<B>> PPO
                             .map(|x| *x as i32)
                             .collect::<Vec<_>>()
                             .as_slice(),
+                        &Default::default(),
                     );
 
                     let state_batch =
@@ -199,7 +200,8 @@ pub(crate) fn get_gae<B: Backend>(
     }
 
     Some(GAEOutput::new(
-        Tensor::from_floats(returns.as_slice()).reshape([returns.len(), 1]),
-        Tensor::from_floats(advantages.as_slice()).reshape([advantages.len(), 1]),
+        Tensor::from_floats(returns.as_slice(), &Default::default()).reshape([returns.len(), 1]),
+        Tensor::from_floats(advantages.as_slice(), &Default::default())
+            .reshape([advantages.len(), 1]),
     ))
 }

burn-rl/src/agent/sac/model.rs

@@ -17,7 +17,7 @@ pub struct SACTemperature<B: Backend> {
 impl<B: Backend> Default for SACTemperature<B> {
     fn default() -> Self {
         Self {
-            temperature: Param::from(Tensor::zeros([1, 1])),
+            temperature: Param::from(Tensor::zeros([1, 1], &Default::default())),
         }
     }
 }

burn-rl/src/base/memory.rs

@@ -160,7 +160,7 @@ mod tests {
         type Data = [ElemType; 2];
 
         fn to_tensor<B: Backend>(&self) -> Tensor<B, 1> {
-            Tensor::<B, 1>::from_floats(self.data)
+            Tensor::<B, 1>::from_floats(self.data, &Default::default())
         }
         fn size() -> usize {
             1
@@ -217,7 +217,7 @@ mod tests {
         });
         assert_eq!(state_batch.shape(), Shape::new([5, 2]));
         let state_sample = state_batch
-            .select(0, Tensor::from_ints([0, 1]))
+            .select(0, Tensor::from_ints([0, 1], &Default::default()))
             .to_data()
             .value;
         assert_eq!(state_sample[0] * 2.0, state_sample[1]);
@@ -226,7 +226,7 @@ mod tests {
             get_batch(memory.next_states(), &sample_indices, ref_to_state_tensor);
         assert_eq!(next_state_batch.shape(), Shape::new([5, 2]));
         let next_state_sample = next_state_batch
-            .select(0, Tensor::from_ints([0, 1]))
+            .select(0, Tensor::from_ints([0, 1], &Default::default()))
             .to_data()
             .value;
         assert_eq!(next_state_sample[0] * -3.0, next_state_sample[1]);

burn-rl/src/environment/cart_pole.rs

@@ -31,7 +31,7 @@ impl From<CartPoleObservation> for CartPoleState {
 impl State for CartPoleState {
     type Data = StateData;
     fn to_tensor<B: Backend>(&self) -> Tensor<B, 1> {
-        Tensor::<B, 1>::from_floats(self.data)
+        Tensor::<B, 1>::from_floats(self.data, &Default::default())
     }
 
     fn size() -> usize {

burn-rl/src/environment/mountain_car.rs

@@ -26,7 +26,7 @@ impl From<MountainCarObservation> for MountainCarState {
 impl State for MountainCarState {
     type Data = StateData;
     fn to_tensor<B: Backend>(&self) -> Tensor<B, 1> {
-        Tensor::<B, 1>::from_floats(self.data)
+        Tensor::<B, 1>::from_floats(self.data, &Default::default())
     }
 
     fn size() -> usize {

burn-rl/src/utils/mod.rs

@@ -28,15 +28,15 @@ pub(crate) fn convert_tenor_to_action<A: Action, B: Backend>(output: Tensor<B, 2
 }
 
 pub(crate) fn to_action_tensor<A: Action, B: Backend>(action: A) -> Tensor<B, 1, Int> {
-    Tensor::<B, 1, Int>::from_ints([action.into() as i32])
+    Tensor::<B, 1, Int>::from_ints([action.into() as i32], &Default::default())
 }
 
 pub(crate) fn ref_to_action_tensor<A: Action, B: Backend>(action: &A) -> Tensor<B, 1, Int> {
     to_action_tensor(*action)
 }
 
 pub(crate) fn to_reward_tensor<B: Backend>(reward: impl Into<ElemType> + Clone) -> Tensor<B, 1> {
-    Tensor::from_floats([reward.into()])
+    Tensor::from_floats([reward.into()], &Default::default())
 }
 
 pub(crate) fn ref_to_reward_tensor<B: Backend>(
@@ -45,7 +45,7 @@ pub(crate) fn ref_to_reward_tensor<B: Backend>(
     to_reward_tensor(reward.clone())
 }
 pub(crate) fn to_not_done_tensor<B: Backend>(done: bool) -> Tensor<B, 1> {
-    Tensor::from_floats([if done { 0.0 } else { 1.0 }])
+    Tensor::from_floats([if done { 0.0 } else { 1.0 }], &Default::default())
 }
 
 pub(crate) fn ref_to_not_done_tensor<B: Backend>(done: &bool) -> Tensor<B, 1> {

examples/Cargo.toml

@@ -5,8 +5,8 @@ edition = "2021"
 publish = false
 
 [dependencies]
-burn = { version = "0.11.1", features = ["ndarray", "autodiff"] }
-burn-autodiff = "0.11.1"
+burn = { version = "0.12.0", features = ["ndarray", "autodiff"] }
+burn-autodiff = "0.12.0"
 serde = { workspace = true }
 
 burn-rl = { path = "../burn-rl" }

examples/src/dqn.rs

@@ -20,9 +20,9 @@ impl<B: Backend> Net<B> {
     #[allow(unused)]
     pub fn new(input_size: usize, dense_size: usize, output_size: usize) -> Self {
         Self {
-            linear_0: LinearConfig::new(input_size, dense_size).init(),
-            linear_1: LinearConfig::new(dense_size, dense_size).init(),
-            linear_2: LinearConfig::new(dense_size, output_size).init(),
+            linear_0: LinearConfig::new(input_size, dense_size).init(&Default::default()),
+            linear_1: LinearConfig::new(dense_size, dense_size).init(&Default::default()),
+            linear_2: LinearConfig::new(dense_size, output_size).init(&Default::default()),
         }
     }
 }

examples/src/ppo.rs

@@ -21,13 +21,13 @@ impl<B: Backend> Net<B> {
         Self {
             linear: LinearConfig::new(input_size, dense_size)
                 .with_initializer(initializer.clone())
-                .init(),
+                .init(&Default::default()),
             linear_actor: LinearConfig::new(dense_size, output_size)
                 .with_initializer(initializer.clone())
-                .init(),
+                .init(&Default::default()),
             linear_critic: LinearConfig::new(dense_size, 1)
                 .with_initializer(initializer)
-                .init(),
+                .init(&Default::default()),
         }
     }
 }

examples/src/sac.rs

@@ -18,9 +18,9 @@ pub struct Actor<B: Backend> {
 impl<B: Backend> Actor<B> {
     pub fn new(input_size: usize, dense_size: usize, output_size: usize) -> Self {
         Self {
-            linear_0: LinearConfig::new(input_size, dense_size).init(),
-            linear_1: LinearConfig::new(dense_size, dense_size).init(),
-            linear_2: LinearConfig::new(dense_size, output_size).init(),
+            linear_0: LinearConfig::new(input_size, dense_size).init(&Default::default()),
+            linear_1: LinearConfig::new(dense_size, dense_size).init(&Default::default()),
+            linear_2: LinearConfig::new(dense_size, output_size).init(&Default::default()),
         }
     }
 }
@@ -50,9 +50,9 @@ pub struct Critic<B: Backend> {
 impl<B: Backend> Critic<B> {
     pub fn new(input_size: usize, dense_size: usize, output_size: usize) -> Self {
         Self {
-            linear_0: LinearConfig::new(input_size, dense_size).init(),
-            linear_1: LinearConfig::new(dense_size, dense_size).init(),
-            linear_2: LinearConfig::new(dense_size, output_size).init(),
+            linear_0: LinearConfig::new(input_size, dense_size).init(&Default::default()),
+            linear_1: LinearConfig::new(dense_size, dense_size).init(&Default::default()),
+            linear_2: LinearConfig::new(dense_size, output_size).init(&Default::default()),
         }
     }
 }