11 Pull requests merged by 6 people

• Fix/autotune error handling

  #2670 merged Jan 8, 2025

• [burnbench] Import code from github-device-flow crate

  #2667 merged Jan 8, 2025

• Fix output float dtype in fusion

  #2671 merged Jan 8, 2025

• Fix load_file usage to keep using model

  #2672 merged Jan 8, 2025

• Refactor unary + binary kernels

  #2665 merged Jan 7, 2025

• add fscore, reorganize mod and change error in doc.

  #2648 merged Jan 7, 2025

• Update cubecl

  #2662 merged Jan 6, 2025

• Bump libc from 0.2.168 to 0.2.169

  #2656 merged Jan 6, 2025

• Combined PRs

  #2661 merged Jan 6, 2025

• Refactor/cubecl

  #2654 merged Jan 5, 2025

• Change pad value w/ ElementConversion

  #2653 merged Jan 3, 2025

4 Pull requests opened by 3 people

• Wasserstein Generative Adversarial Network

  #2660 opened Jan 6, 2025

• Add cumsum tensor op

  #2664 opened Jan 6, 2025

• Merge reduce

  #2673 opened Jan 8, 2025

• Add checks for even padding when kernel size is even

  #2677 opened Jan 9, 2025

6 Issues closed by 3 people

• PaddingConfig::Same doesn't return the same input shape when kernel_size = [2, 2]

  #2675 closed Jan 9, 2025

• RUSTSEC-2024-0375: `atty` is unmaintained

  #2546 closed Jan 8, 2025

• RUSTSEC-2021-0145: Potential unaligned read

  #1305 closed Jan 8, 2025

• RUSTSEC-2024-0370: proc-macro-error is unmaintained

  #2548 closed Jan 8, 2025

• Tensor concatenation and appending

  #2666 closed Jan 8, 2025

• Inquiry: Support for `self.training` in Burn Modules or Alternative Solutions

  #2652 closed Jan 3, 2025

3 Issues opened by 3 people

• Support asymmetric padding for "same" padding with even kernel size

  #2676 opened Jan 9, 2025

• Implement `grid_sample` for sampling from a spatial tensor at arbitrary locations, similar to PyTorch

  #2674 opened Jan 8, 2025

• Improve `PyTorchFileRecorder` enum deserializer error message

  #2663 opened Jan 6, 2025

5 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Feature add new one hot function meeting multi-dimensions (ranks)

  #2613 commented on Jan 8, 2025 • 15 new comments

• 4 bit / 8 bit model training / inference capabilities

  #464 commented on Jan 3, 2025 • 0 new comments

• expected `*mut u8`, found `*mut i8`

  #2596 commented on Jan 6, 2025 • 0 new comments

• Feature Request: Full Gemm operation / node

  #2049 commented on Jan 6, 2025 • 0 new comments

• Module derive relies on `B` generics for Backend

  #452 commented on Jan 8, 2025 • 0 new comments