4 Pull requests merged by 4 people

• Update to ubuntu latest

  #8979 merged Apr 16, 2025

• @assume_pure

  #8962 merged Apr 16, 2025

• Avoid re-computing computation hashes

  #8976 merged Apr 15, 2025

• Disable one splash attention test

  #8970 merged Apr 15, 2025

5 Pull requests opened by 3 people

• Avoid re-computing computation hashes (#8976)

  #8977 opened Apr 15, 2025

• Fix splash attention test

  #8978 opened Apr 15, 2025

• Fix call_jax hashing

  #8981 opened Apr 16, 2025

• Default explicit donation for step barriers

  #8982 opened Apr 16, 2025

• scan-based GRU falls back to nn.GRU when `bidirectional` is true.

  #8984 opened Apr 16, 2025

2 Issues closed by 2 people

• Use jax autograd from PyTorch

  #8805 closed Apr 16, 2025

• Incorrect stableHLO output of all reduce

  #8969 closed Apr 15, 2025

8 Issues opened by 4 people

• [torchax] requires_grad in some tensor constructors not propagated

  #8983 opened Apr 16, 2025

• [torchax] `jax_tensor.copy_(cpu_tensor)` fails

  #8980 opened Apr 15, 2025

• [assume_pure] test_assume_pure_einsum_grads torchax output on TPU differs from torch_xla

  #8975 opened Apr 14, 2025

• [assume_pure] test_assume_pure_complex_module torchax output on TPU differs from torch_xla

  #8974 opened Apr 14, 2025

• `torch_xla::runtime::PjRtComputationClient::ExecuteReplicated()` crash after RuntimeError

  #8973 opened Apr 14, 2025

• [call_jax] Bridge the torch_xla and JAX mesh

  #8972 opened Apr 14, 2025

• Splash attention test fail randomly only in github CI

  #8971 opened Apr 14, 2025

• Alternative to torch.select_mask

  #8968 opened Apr 13, 2025

10 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.

• Add an option for JittableModule to dedup parameters.

  #8965 commented on Apr 15, 2025 • 3 new comments

• TPU 6e-1 + Pytorch 2.6[TPU] does not work

  #8960 commented on Apr 13, 2025 • 0 new comments

• [scan] Avoid re-tracing the combine function on every call

  #8632 commented on Apr 14, 2025 • 0 new comments

• Prototype API to build global tensor using local processes

  #8842 commented on Apr 15, 2025 • 0 new comments

• 2.7 backport PR request list

  #8829 commented on Apr 15, 2025 • 0 new comments

• [Deprecation Tracking] API deprecation timeline summary

  #8915 commented on Apr 16, 2025 • 0 new comments

• torch.linalg.lstsq issues on GPU/TPU

  #8953 commented on Apr 16, 2025 • 0 new comments

• add CreateGlobalShardedData prototype

  #8932 commented on Apr 16, 2025 • 0 new comments

• assume_pure & call_jax composible.

  #8961 commented on Apr 16, 2025 • 0 new comments

• Add a helper class to handle mesh and sharding

  #8967 commented on Apr 15, 2025 • 0 new comments