Search results

Running the test_run_model test results in a segmentation fault error

Status: Open.

#377 In AI-Hypercomputer/torchprime;

· erichuang-cienet opened on Sep 2, 2025

Some tests will hang when running tests with pytest on a TPU VM

Status: Open.

#375 In AI-Hypercomputer/torchprime;

· erichuang-cienet opened on Aug 28, 2025

Jobs hang with fsdp only (1d) sharding with >=64 tpu chips

Status: Open.

#363 In AI-Hypercomputer/torchprime;

· jialei777 opened on Aug 13, 2025

different SyncTensorsGraph for different training steps

Status: Open.

#360 In AI-Hypercomputer/torchprime;

· jialei777 opened on Aug 12, 2025

Bug in torch_xla._XLAC._xla_tensors_from_aten

Status: Open.

#353 In AI-Hypercomputer/torchprime;

· liurupeng opened on Jul 29, 2025

Add backwards UnitTests to existing models

Status: Open.

#352 In AI-Hypercomputer/torchprime;

· vlasenkoalexey opened on Jul 25, 2025

Update e2e_testing/update_step_time.py tool to preserve existing values

Status: Open.

#348 In AI-Hypercomputer/torchprime;

· vlasenkoalexey opened on Jul 23, 2025

Llama 8.0 8B SFT and <code>assume_pure</code> errors

bugSomething isn't workingSomething isn't working

Status: Open.

Bug

#343 In AI-Hypercomputer/torchprime;

· pgmoka opened on Jul 16, 2025

replace hardcoded string with enum

Status: Open.

#333 In AI-Hypercomputer/torchprime;

· liurupeng opened on Jul 8, 2025

Allow passing entire config to attention layer to dedup parameters spec in model config and sharding config

Status: Open.

#330 In AI-Hypercomputer/torchprime;

· liurupeng opened on Jun 30, 2025

Refactor test directory structure

Status: Open.

#325 In AI-Hypercomputer/torchprime;

· yaoshiang opened on Jun 24, 2025

Inquiry on Torch XLA Scalability and Performance Comparison with JAX on TPU

Status: Open.

#313 In AI-Hypercomputer/torchprime;

· Rayjryang opened on Jun 15, 2025