XPK Start: Sat Apr 18 06:08:51 UTC 2026 2026-04-18 06:09:43.247087: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0418 06:09:43.463802 133843971372864 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-18 06:09:52,505:jax._src.distributed:166: Connecting to JAX distributed service on mt-01-sft-smoke-sgsgn-slice-job-0-0.mt-01-sft-smoke-sgsgn:8482 I0418 06:09:52.505791 133843971372864 distributed.py:166] Connecting to JAX distributed service on mt-01-sft-smoke-sgsgn-slice-job-0-0.mt-01-sft-smoke-sgsgn:8482 I0418 06:10:15.091982 133843971372864 max_utils.py:284] Jax distributed system initialized! I0418 06:10:22.107359 133843971372864 max_utils.py:800] System Information: Jax Version: 0.8.3 I0418 06:10:22.107472 133843971372864 max_utils.py:801] System Information: Jaxlib Version: 0.8.3 I0418 06:10:22.107512 133843971372864 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Dec 15 2025 14:03:46 (1765836226) cl/844590465 I0418 06:10:22.110838 133843971372864 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 06:10:22.306889 133843971372864 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) Traceback (most recent call last): File "<frozen runpy>", line 198, in _run_module_as_main File "<frozen runpy>", line 88, in _run_code File "/deps/src/maxtext/trainers/post_train/sft/train_sft.py", line 216, in <module> app.run(main) File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 316, in run _run_main(main, args) File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 261, in _run_main sys.exit(main(argv)) ^^^^^^^^^^ File "/deps/src/maxtext/trainers/post_train/sft/train_sft.py", line 212, in main train(mt_config, goodput_recorder) File "/deps/src/maxtext/trainers/post_train/sft/train_sft.py", line 186, in train trainer, mesh = setup_trainer_state(mt_config, goodput_recorder) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/deps/src/maxtext/trainers/post_train/sft/train_sft.py", line 149, in setup_trainer_state model, mesh = model_creation_utils.create_nnx_model(mt_config) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/deps/src/maxtext/utils/model_creation_utils.py", line 337, in create_nnx_model model = create_nnx_sharded_model_hybrid(config, mesh, devices, model_mode, rng_key) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/deps/src/maxtext/utils/model_creation_utils.py", line 314, in create_nnx_sharded_model_hybrid sharded_state = create_sharded_state() ^^^^^^^^^^^^^^^^^^^^^^ ValueError: One of pjit outputs was given the sharding of NamedSharding(mesh=Mesh('diloco': 1, 'data': 1, 'stage': 1, 'fsdp': 32, 'fsdp_transpose': 1, 'sequence': 1, 'context': 1, 'context_autoregressive': 1, 'tensor': 1, 'tensor_transpose': 1, 'tensor_sequence': 1, 'expert': 1, 'autoregressive': 1, axis_types=(Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto, Auto)), spec=PartitionSpec(('fsdp', 'sequence', 'tensor_transpose', 'context', 'expert'), 'stage', ('fsdp_transpose', 'tensor', 'tensor_sequence', 'autoregressive')), memory_kind=device), which implies that the global size of its dimension 0 should be divisible by 32, but it is equal to 16 (full shape: (16, 1, 64)) -------------------- For simplicity, JAX has removed its internal frames from the traceback of the following exception. Set JAX_TRACEBACK_FILTERING=off to include these. XPK End: Sat Apr 18 06:10:32 UTC 2026 EXIT_CODE=1