MaxView

XPK Start: Fri Apr 24 07:39:44 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-24 07:40:09.407214: 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)
I0424 07:40:09.611532 136179606476608 max_utils.py:273] Attempting to initialize the jax distributed system...
I0424 07:40:18.654208 136179606476608 distributed.py:149] Starting JAX distributed service on [::]:8482
I0424 07:40:18.656491 136179606476608 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-vppxf-slice-job-0-0.mt-11-optimizer-offload-vppxf:8482
F0424 07:45:23.658731      10 client.h:80] Terminating process because the JAX distributed service detected fatal errors. This most likely indicates that another task died; see the other task logs for more details. Disable Python buffering, i.e. `python -u`, to be sure to see all the previous output. absl::Status: DEADLINE_EXCEEDED: Deadline Exceeded

RPC: /tensorflow.CoordinationService/RegisterTask
*** Check failure stack trace: ***
    @     0x7bda1a958464  absl::lts_20250814::log_internal::LogMessage::SendToLog()
    @     0x7bda1a9583e6  absl::lts_20250814::log_internal::LogMessage::Flush()
    @     0x7bda1a9afa3e  std::_Function_handler<>::_M_invoke()
    @     0x7bda171989d5  std::_Function_handler<>::_M_invoke()
    @     0x7bda171b2e31  xla::CoordinationServiceAgent::SetError()
    @     0x7bda171b27a2  xla::CoordinationServiceAgent::Connect()
    @     0x7bda1719743a  xla::DistributedRuntimeCoordinationServiceClient::Connect()
    @     0x7bda1a9ab413  nanobind::detail::func_create<>()::{lambda()#1}::__invoke()
    @     0x7bda170d600d  nanobind::detail::nb_func_vectorcall_simple_1()
    @     0x7bdac9676501  PyObject_Vectorcall
Fatal Python error: Aborted

Current thread 0x00007bdac911d740 (most recent call first):
  File "/usr/local/lib/python3.12/site-packages/jax/_src/distributed.py", line 173 in initialize
  File "/usr/local/lib/python3.12/site-packages/jax/_src/distributed.py", line 331 in initialize
  File "/deps/src/maxtext/utils/max_utils.py", line 275 in maybe_initialize_jax_distributed_system
  File "/deps/src/maxtext/configs/pyconfig.py", line 459 in initialize_pydantic
  File "/deps/src/maxtext/configs/pyconfig.py", line 326 in initialize
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 665 in initialize
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 736 in main
  File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 312 in _run_main
  File "/usr/local/lib/python3.12/site-packages/absl/app.py", line 367 in run
  File "/deps/src/maxtext/trainers/pre_train/train.py", line 744 in <module>
  File "<frozen runpy>", line 88 in _run_code
  File "<frozen runpy>", line 198 in _run_module_as_main

Extension modules: numpy._core._multiarray_umath, numpy._core._multiarray_tests, numpy.linalg._umath_linalg, jaxlib.cpu_feature_guard, numpy.random._common, numpy.random.bit_generator, numpy.random._bounded_integers, numpy.random._mt19937, numpy.random.mtrand, numpy.random._philox, numpy.random._pcg64, numpy.random._sfc64, numpy.random._generator, zstandard.backend_c, uvloop.loop, msgpack._cmsgpack, yaml._yaml, simplejson._speedups, google._upb._message, charset_normalizer.md, charset_normalizer.cd, requests.packages.charset_normalizer.md, requests.packages.chardet.md, requests.packages.charset_normalizer.cd, requests.packages.chardet.cd, _cffi_backend, grpc._cython.cygrpc, multidict._multidict, yarl._quoting_c, propcache._helpers_c, aiohttp._http_writer, aiohttp._http_parser, aiohttp._websocket.mask, aiohttp._websocket.reader_c, frozenlist._frozenlist, h5py._errors, h5py.defs, h5py._objects, h5py.h5, h5py.utils, h5py.h5t, h5py.h5s, h5py.h5ac, h5py.h5p, h5py.h5r, h5py._npystrings, h5py._proxy, h5py._conv, h5py.h5z, h5py.h5a, h5py.h5d, h5py.h5ds, h5py.h5g, h5py.h5i, h5py.h5o, h5py.h5f, h5py.h5fd, h5py.h5pl, h5py.h5l, h5py._selector, _cyutility, scipy._cyutility, scipy._lib._ccallback_c, scipy.sparse._sparsetools, _csparsetools, scipy.sparse._csparsetools, PIL._imaging, psutil._psutil_linux, pyarrow.lib, pandas._libs._cyutility, pandas._libs.tslibs.ccalendar, pandas._libs.tslibs.np_datetime, pandas._libs.tslibs.dtypes, pandas._libs.tslibs.base, pandas._libs.tslibs.nattype, pandas._libs.tslibs.timezones, pandas._libs.properties, pandas._libs.tslibs.fields, pandas._libs.tslibs.timedeltas, pandas._libs.tslibs.tzconversion, pandas._libs.tslibs.timestamps, pandas._libs.tslibs.offsets, pandas._libs.tslibs.strptime, pandas._libs.tslibs.parsing, pandas._libs.tslibs.conversion, pandas._libs.tslibs.period, pandas._libs.tslibs.vectorized, pandas._libs.ops_dispatch, pandas._libs.missing, pandas._libs.hashtable, pandas._libs.algos, pandas._libs.interval, pandas._libs.lib, pyarrow._compute, pandas._libs.ops, pandas._libs.hashing, pandas._libs.arrays, pandas._libs.tslib, pandas._libs.sparse, pandas._libs.internals, pandas._libs.indexing, pandas._libs.index, pandas._libs.writers, pandas._libs.join, pandas._libs.window.aggregations, pandas._libs.window.indexers, pandas._libs.reshape, pandas._libs.groupby, pandas._libs.json, pandas._libs.parsers, pandas._libs.testing, kiwisolver._cext, markupsafe._speedups, regex._regex, sentencepiece._sentencepiece, scipy.spatial._ckdtree, scipy._lib.messagestream, scipy.linalg._fblas, scipy.linalg._flapack, scipy.linalg.cython_lapack, scipy.linalg._cythonized_array_utils, scipy.linalg._solve_toeplitz, scipy.linalg._batched_linalg, scipy.linalg._decomp_lu_cython, scipy.linalg._matfuncs_schur_sqrtm, scipy.linalg._matfuncs_expm, scipy.linalg._linalg_pythran, scipy.linalg.cython_blas, scipy.linalg._decomp_update, scipy.spatial._qhull, scipy.spatial._voronoi, scipy.special._ufuncs_cxx, scipy.special._ellip_harm_2, scipy.special._special_ufuncs, scipy.special._gufuncs, scipy.special._ufuncs, scipy.special._specfun, scipy.special._comb, scipy.spatial._hausdorff, scipy.spatial._distance_wrap, scipy.spatial.transform._rotation_cy, scipy.spatial.transform._rigid_transform_cy, scipy.sparse.linalg._dsolve._superlu, scipy.sparse.linalg._eigen.arpack._arpacklib, scipy.sparse.linalg._propack, scipy.optimize._group_columns, scipy.optimize._trlib._trlib, scipy.optimize._lbfgsb, _moduleTNC, scipy.optimize._moduleTNC, scipy.optimize._slsqplib, scipy.optimize._minpack, scipy.optimize._lsq.givens_elimination, scipy.optimize._zeros, scipy._lib._uarray._uarray, scipy.linalg._decomp_interpolative, scipy.optimize._bglu_dense, scipy.optimize._lsap, scipy.optimize._direct, scipy.integrate._odepack, scipy.integrate._quadpack, scipy.integrate._vode, scipy.integrate._dop, scipy.interpolate._fitpack, scipy.interpolate._dfitpack, scipy.interpolate._dierckx, scipy.interpolate._ppoly, scipy.interpolate._interpnd, scipy.interpolate._rbfinterp_pythran, scipy.interpolate._rgi_cython, scipy.special.cython_special, scipy.stats._stats, scipy.stats._biasedurn, scipy.stats._stats_pythran, scipy.stats._levy_stable.levyst, scipy.stats._ansari_swilk_statistics, scipy.sparse.csgraph._tools, scipy.sparse.csgraph._shortest_path, scipy.sparse.csgraph._traversal, scipy.sparse.csgraph._min_spanning_tree, scipy.sparse.csgraph._flow, scipy.sparse.csgraph._matching, scipy.sparse.csgraph._reordering, scipy.stats._sobol, scipy.stats._qmc_cy, scipy.stats._rcont.rcont, scipy.stats._qmvnt_cy, scipy.ndimage._nd_image, scipy.ndimage._rank_filter_1d, _ni_label, scipy.ndimage._ni_label (total: 191)
bash: line 5:    10 Aborted                 (core dumped) python3 -m maxtext.trainers.pre_train.train src/maxtext/configs/base.yml pure_nnx=False pure_nnx_decoder=False enable_nnx=False base_output_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260424_070227 dataset_type=synthetic steps=10 learning_rate_schedule_steps=10 enable_checkpointing=True checkpoint_period=10 enable_goodput_recording=False per_device_batch_size=1 ici_fsdp_parallelism=32 log_config=False tokenizer_path=src/maxtext/assets/tokenizers/tokenizer.llama2 use_vertex_tensorboard=False vertex_tensorboard_project='' vertex_tensorboard_region='' enable_pathways_goodput=False enable_gcp_goodput_metrics=False monitor_goodput=False attention=dot_product profiler=xplane debug_sharding=True run_name=linen_xpk_main_20260424_070227_11_optimizer_offload_true optimizer_memory_host_offload=True
XPK End: Fri Apr 24 07:45:29 UTC 2026
EXIT_CODE=134

XPK Start: Fri Apr 24 12:39:13 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-24 12:39:38.615868: 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)
I0424 12:39:38.826395 136878333507392 max_utils.py:273] Attempting to initialize the jax distributed system...
I0424 12:39:47.868032 136878333507392 distributed.py:149] Starting JAX distributed service on [::]:8482
I0424 12:39:47.870410 136878333507392 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-6qpkr-slice-job-0-0.mt-11-optimizer-offload-6qpkr:8482
I0424 12:39:48.840383 136878333507392 max_utils.py:284] Jax distributed system initialized!
I0424 12:39:54.970285 136878333507392 max_utils.py:800] System Information: Jax Version: 0.9.2
I0424 12:39:54.970397 136878333507392 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0424 12:39:54.970438 136878333507392 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0424 12:39:54.970473 136878333507392 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0424 12:39:55.673484 136878333507392 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 12:39:55.674068 136878333507392 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 12:39:55.674268 136878333507392 checkpointing.py:688] Setting up checkpoint logger...
I0424 12:39:55.674319 136878333507392 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0424 12:39:55.674361 136878333507392 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0424 12:39:55.674694 136878333507392 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7c7cc0a503e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0424 12:39:58.490964 136878333507392 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0424 12:39:58.491219 136878333507392 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c689c33f6b0>}, handler_registry=None
I0424 12:39:58.491460 136878333507392 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c689c33f6b0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0424 12:39:58.491509 136878333507392 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c689c341f10>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0424 12:39:58.491544 136878333507392 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c689c33f6b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c689c33f6b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c689c341f10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c689c341f10>}).
I0424 12:39:58.491874 136878333507392 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0424 12:39:58.491943 136878333507392 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c68847c5d00> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0424 12:39:59.355143 136878333507392 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints
I0424 12:39:59.395559 136878333507392 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c689c33e990>
I0424 12:39:59.395700 136878333507392 checkpointing.py:302] Checkpoint manager created!
I0424 12:40:00.335583 136878333507392 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0424 12:40:00.335690 136878333507392 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:40:00.715566 136878333507392 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn').
I0424 12:40:00.715657 136878333507392 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:40:00.732216 136878333507392 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn').
I0424 12:40:00.732280 136878333507392 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:40:00.756291 136878333507392 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:40:00.756362 136878333507392 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:40:00.773178 136878333507392 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:40:00.773243 136878333507392 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:40:00.789762 136878333507392 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:40:00.789822 136878333507392 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:40:00.814647 136878333507392 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv').
I0424 12:40:00.814718 136878333507392 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:40:00.841934 136878333507392 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0424 12:40:00.842049 136878333507392 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0424 12:40:01.069773 136878333507392 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0424 12:40:01.069885 136878333507392 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0424 12:40:02.669616 136878333507392 maxtext_utils.py:1880]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0424 12:40:02.669750 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0424 12:40:02.669806 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0424 12:40:02.669864 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0424 12:40:02.669918 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0424 12:40:02.669963 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0424 12:40:02.670017 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0424 12:40:02.670070 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   P('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0424 12:40:02.670124 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0424 12:40:02.670164 136878333507392 maxtext_utils.py:1880]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0424 12:40:02.670214 136878333507392 maxtext_utils.py:1880]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0424 12:40:02.670266 136878333507392 maxtext_utils.py:1880]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0424 12:40:03.166999 136878333507392 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 12:40:03.167091 136878333507392 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 12:40:03.182603 136878333507392 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 12:40:03.182665 136878333507392 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 12:40:14.115612 136878333507392 max_utils.py:791] Total memory size: 1.3 GB, Output size: 0.1 GB, Temp size: 1.1 GB, Argument size: 0.1 GB, Host temp size: 0.0 GB.
I0424 12:40:14.116427 136878333507392 metric_logger.py:301] number parameters: 1.104 billion
I0424 12:40:25.444036 136878333507392 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0424 12:40:25.863033 136878333507392 checkpointing.py:798] Waited 0.4189753532409668 seconds for step 0 to finish before starting checkpointing.
I0424 12:40:25.865441 136878333507392 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 12:40:25.867114 136878333507392 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0424 12:40:25.868430 136878333507392 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/0.
I0424 12:40:26.210594 136878333507392 signaling_client.py:364] Using JaxDistributedSignalingClient
I0424 12:40:26.211540 136878333507392 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 12:40:26.211596 136878333507392 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0424 12:40:26.489023 136878333507392 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278503s
I0424 12:40:26.489216 136878333507392 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.420 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28461623191833496 s) (per-host)
I0424 12:40:26.489270 136878333507392 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.284682s (batch_requests_ready=0.002162s, total_serialization_initiated=0.282447s, others=0.000073s)
I0424 12:40:26.489365 136878333507392 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289184s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.289166)
I0424 12:40:26.490190 136878333507392 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/0.
I0424 12:40:26.490552 136731549767424 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:00:26.490514
I0424 12:40:26.503029 136878333507392 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 12:40:26.503354 136727024883456 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0424 12:40:26.503511 136878333507392 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777034425.865423, 'wait_for_prev_duration_secs': 5.936622619628906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777034425.8671508, 'checkpointer_blocking_duration_secs': 0.6235668659210205, 'get_old_steps_start_time': 1777034426.4907424, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1777034425.8635328, 'checkpoint_manager_blocking_duration_secs': 0.6399397850036621}
I0424 12:40:26.503613 136878333507392 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0424 12:40:26.503662 136878333507392 max_utils.py:750] 
Memstats: After params initialized:
I0424 12:40:26.503712 136878333507392 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_24(process=6,(0,6,0,0))
I0424 12:40:26.503745 136878333507392 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_25(process=6,(1,6,0,0))
I0424 12:40:26.503773 136878333507392 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_28(process=6,(0,7,0,0))
I0424 12:40:26.503798 136878333507392 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_29(process=6,(1,7,0,0))
I0424 12:40:26.817188 136878333507392 metric_logger.py:196] completed step: 0, seconds: 11.328, TFLOP/s/device: 1.199, Tokens/s/device: 180.799, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0424 12:40:27.017729 136878333507392 metric_logger.py:196] completed step: 1, seconds: 1.372, TFLOP/s/device: 9.905, Tokens/s/device: 1493.040, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0424 12:40:27.459156 136878333507392 metric_logger.py:196] completed step: 2, seconds: 0.038, TFLOP/s/device: 360.688, Tokens/s/device: 54366.870, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129
I0424 12:40:27.621855 136878333507392 metric_logger.py:196] completed step: 3, seconds: 0.441, TFLOP/s/device: 30.787, Tokens/s/device: 4640.487, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.047
I0424 12:40:27.947972 136878333507392 metric_logger.py:196] completed step: 4, seconds: 0.169, TFLOP/s/device: 80.396, Tokens/s/device: 12118.200, total_weights: 65536, loss: 9.278, lm_loss: 9.278, perplexity: 10694.968
I0424 12:40:27.953950 136878333507392 metric_logger.py:196] completed step: 5, seconds: 0.163, TFLOP/s/device: 83.576, Tokens/s/device: 12597.418, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.520
I0424 12:40:29.657664    2840 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0424 12:40:31.724288 136731507455744 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_6
I0424 12:40:49.687777 136878333507392 metric_logger.py:196] completed step: 6, seconds: 0.327, TFLOP/s/device: 41.592, Tokens/s/device: 6269.228, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.688
I0424 12:40:49.850501 136878333507392 metric_logger.py:196] completed step: 7, seconds: 21.571, TFLOP/s/device: 0.630, Tokens/s/device: 94.941, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4377.010
I0424 12:40:50.013180 136878333507392 metric_logger.py:196] completed step: 8, seconds: 0.168, TFLOP/s/device: 80.927, Tokens/s/device: 12198.245, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3845.943
I0424 12:40:50.175378 136878333507392 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0424 12:40:50.176022 136878333507392 checkpointing.py:798] Waited 0.0006656646728515625 seconds for step 9 to finish before starting checkpointing.
I0424 12:40:50.178027 136878333507392 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 12:41:03.349804 136731549767424 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.523 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.14517259597778 s) (per-host)
I0424 12:41:03.349914 136731549767424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.859242s.
I0424 12:41:13.312544 136731549767424 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 46.821854s.
I0424 12:41:13.312863 136727024883456 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 12:41:13.312991 136727024883456 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0424 12:41:13.313043 136727024883456 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0424 12:41:13.314686 136727024883456 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0424 12:41:13.314840 136878333507392 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0424 12:41:13.314975 136878333507392 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.136947 seconds. If this number is high, consider checkpointing less frequently.
I0424 12:41:13.317073 136878333507392 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0424 12:41:13.319058 136878333507392 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/9.
I0424 12:41:13.611880 136878333507392 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 12:41:13.611976 136878333507392 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0424 12:41:13.654818 136878333507392 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043912s
I0424 12:41:13.654983 136878333507392 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 32.616 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.047292470932006836 s) (per-host)
I0424 12:41:13.655034 136878333507392 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.047353s (batch_requests_ready=0.001839s, total_serialization_initiated=0.045445s, others=0.000069s)
I0424 12:41:13.655134 136878333507392 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051392s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.051377)
I0424 12:41:13.655821 136878333507392 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/9.
I0424 12:41:13.656137 136727024883456 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:01:13.656091
I0424 12:41:13.658135 136878333507392 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 12:41:13.658427 136731507455744 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0424 12:41:13.658596 136878333507392 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777034450.177998, 'wait_for_prev_duration_secs': 23.13694739341736, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777034473.3171222, 'checkpointer_blocking_duration_secs': 0.3391542434692383, 'get_old_steps_start_time': 1777034473.6562986, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1777034450.176259, 'checkpoint_manager_blocking_duration_secs': 23.48229742050171}
I0424 12:41:13.658699 136878333507392 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0424 12:41:13.658741 136878333507392 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 12:41:19.401590 136723078039296 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_6
I0424 12:41:56.881178 136727024883456 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.501 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.27344799041748 s) (per-host)
I0424 12:41:56.881305 136727024883456 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.225055s.
I0424 12:42:05.330310 136727024883456 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.674043s.
I0424 12:42:05.330633 136731507455744 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 12:42:05.330765 136731507455744 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0424 12:42:05.330815 136731507455744 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0424 12:42:05.332330 136731507455744 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0424 12:42:05.332516 136878333507392 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0424 12:42:05.332677 136878333507392 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 12:42:05.333645 136878333507392 metric_logger.py:196] completed step: 9, seconds: 0.163, TFLOP/s/device: 83.473, Tokens/s/device: 12582.017, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.739
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 24 12:42:13 UTC 2026
EXIT_CODE=0