Arthur Brussee
Arthur Brussee
(First OSS contribution, so apologies if I mess up the format, roll-ups, or err the rest! Picked a random issue to see what it's like) # Objective Simplify implementing some...
### Feature description I'm writing some tests to check gradients again a reference implementation. This works great for leaf nodes, but I can't atm seem to get gradients of intermediate...
### Changes According to [wgpu docs](https://docs.rs/wgpu/latest/wgpu/struct.BufferSlice.html#method.map_async), the device only needs to be polled for the map_async to complete at some point. Waiting for the channel does the actual waiting (async...
**Describe the bug** If some operations use a tensor multiple times, Chrome can crash complaining it's a violation of the WebGPU specification to bind aliased buffers. **To Reproduce** ``` let...
## Changes Currently, working with multiple different learning rates in Burn can be a bit of a hassle. The easiest way is to create a different GradientParams for each set...
Hiya, Zed on Windows is looking rather washed out when using HDR: https://github.com/zed-industries/zed/issues/16182 I think this is due to how blade creates the swapchain. In vulkan/init.rs, it checks supported_formats.rs for...
**Description** After updating an egui + burn app to wgpu master, I observed random crashes with a generic "validation failed: parent device lost" error. Nothing in particular seems to cause...
on_submitted_work_done works for native platforms but is [unimplemented!()](https://github.com/gfx-rs/wgpu/blob/73764fdc6a52e4078a6d5e3429fde1a25c12e059/wgpu/src/backend/webgpu.rs#L2773) for WebGPU. See also https://github.com/gfx-rs/wgpu/discussions/5122, https://github.com/gfx-rs/wgpu/pull/5012