feat: update workspace paths and enhance gitignore

- Updated stablediffusion crate path from "../stable-diffusion-burn" to "./crates/stable-diffusion-burn" for proper workspace resolution
- Enhanced .gitignore to include generated model files (.mpk, .pt, .bin, .safetensors, .ckpt) and user_data directory
- Added Cargo.lock to gitignore with appropriate comment
- Reorganized IDE files section in gitignore for better clarity
- Added newline at end of file for proper formatting

crates/stable-diffusion-burn/burn-crates/burn-cuda/Cargo.toml

@@ -0,0 +1,41 @@
+[package]
+authors = ["nathanielsimard <nathaniel.simard.42@gmail.com>"]
+categories = ["science"]
+description = "CUDA backend for the Burn framework"
+documentation = "https://docs.rs/burn-cuda"
+edition.workspace = true
+keywords = ["deep-learning", "machine-learning", "gpu", "cuda"]
+license.workspace = true
+name = "burn-cuda"
+readme.workspace = true
+repository = "https://github.com/tracel-ai/burn/tree/main/crates/burn-cuda"
+version.workspace = true
+
+[lints]
+workspace = true
+
+[features]
+autotune = ["burn-cubecl/autotune"]
+autotune-checks = ["burn-cubecl/autotune-checks"]
+default = ["std", "fusion", "autotune", "burn-cubecl/default", "cubecl/default"]
+doc = ["burn-cubecl/doc"]
+fusion = ["burn-fusion", "burn-cubecl/fusion"]
+std = ["burn-cubecl/std", "cubecl/std"]
+tracing = [
+    "burn-backend/tracing",
+    "burn-cubecl/tracing",
+    "burn-fusion?/tracing",
+    "cubecl/tracing",
+]
+
+[dependencies]
+burn-fusion = { path = "../burn-fusion", version = "=0.21.0-pre.2", optional = true }
+burn-cubecl = { path = "../burn-cubecl", version = "=0.21.0-pre.2", default-features = false }
+burn-backend = { path = "../burn-backend", version = "=0.21.0-pre.2", default-features = false, features = [
+    "cubecl-cuda",
+] }
+cubecl = { workspace = true, features = ["cuda"] }
+
+[package.metadata.docs.rs]
+features = ["doc"]
+rustdoc-args = ["--cfg", "docsrs"]

crates/stable-diffusion-burn/burn-crates/burn-cuda/README.md

@@ -0,0 +1,30 @@
+# Burn CUDA Backend
+
+[Burn](https://github.com/tracel-ai/burn) CUDA backend
+
+[![Current Crates.io Version](https://img.shields.io/crates/v/burn-cuda.svg)](https://crates.io/crates/burn-cuda)
+[![license](https://shields.io/badge/license-MIT%2FApache--2.0-blue)](https://github.com/tracel-ai/burn-cuda/blob/master/README.md)
+
+This crate provides a CUDA backend for [Burn](https://github.com/tracel-ai/burn) using the
+[cubecl](https://github.com/tracel-ai/cubecl.git) and [cudarc](https://github.com/coreylowman/cudarc.git)
+crates.
+
+## Usage Example
+
+```rust
+#[cfg(feature = "cuda")]
+mod cuda {
+    use burn_autodiff::Autodiff;
+    use burn_cuda::{Cuda, CudaDevice};
+    use mnist::training;
+
+    pub fn run() {
+        let device = CudaDevice::default();
+        training::run::<Autodiff<Cuda<f32, i32>>>(device);
+    }
+}
+```
+
+## Dependencies
+
+Requires CUDA 12.x to be installed and on the `PATH`.

crates/stable-diffusion-burn/burn-crates/burn-cuda/src/lib.rs

@@ -0,0 +1,47 @@
+#![cfg_attr(docsrs, feature(doc_cfg))]
+
+extern crate alloc;
+
+use burn_cubecl::CubeBackend;
+pub use cubecl::cuda::CudaDevice;
+use cubecl::cuda::CudaRuntime;
+
+#[cfg(not(feature = "fusion"))]
+pub type Cuda<F = f32, I = i32> = CubeBackend<CudaRuntime, F, I, u8>;
+
+#[cfg(feature = "fusion")]
+pub type Cuda<F = f32, I = i32> = burn_fusion::Fusion<CubeBackend<CudaRuntime, F, I, u8>>;
+
+#[cfg(all(test, not(target_os = "macos")))]
+mod tests {
+    use super::*;
+    use burn_backend::{Backend, DType, QTensorPrimitive};
+    use burn_cubecl::tensor::CubeTensor;
+
+    #[test]
+    fn should_support_dtypes() {
+        type B = Cuda;
+        let device = Default::default();
+
+        assert!(B::supports_dtype(&device, DType::F32));
+        assert!(B::supports_dtype(&device, DType::Flex32));
+        assert!(B::supports_dtype(&device, DType::F16));
+        assert!(B::supports_dtype(&device, DType::BF16));
+        assert!(B::supports_dtype(&device, DType::I64));
+        assert!(B::supports_dtype(&device, DType::I32));
+        assert!(B::supports_dtype(&device, DType::I16));
+        assert!(B::supports_dtype(&device, DType::I8));
+        assert!(B::supports_dtype(&device, DType::U64));
+        assert!(B::supports_dtype(&device, DType::U32));
+        assert!(B::supports_dtype(&device, DType::U16));
+        assert!(B::supports_dtype(&device, DType::U8));
+        assert!(B::supports_dtype(&device, DType::Bool));
+        assert!(B::supports_dtype(
+            &device,
+            DType::QFloat(CubeTensor::<CudaRuntime>::default_scheme())
+        ));
+
+        // Currently not registered in supported types
+        assert!(!B::supports_dtype(&device, DType::F64));
+    }
+}