MaxView

XPK Start: Thu Apr 23 15:56:39 UTC 2026
2026-04-23 15:56:43.998291: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1776959804.010843      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776959804.014644      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776959804.025709      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776959804.025729      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776959804.025731      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776959804.025733      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-23 15:57:03.149391: 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)
I0423 15:57:03.675909 135968039954240 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-23 15:57:12,717:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0423 15:57:12.717846 135968039954240 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-23 15:57:12,720:jax._src.distributed:157: Connecting to JAX distributed service on mt-02-synthetic-agzb1-slice-job-0-0.mt-02-synthetic-agzb1:8482
I0423 15:57:12.720075 135968039954240 distributed.py:157] Connecting to JAX distributed service on mt-02-synthetic-agzb1-slice-job-0-0.mt-02-synthetic-agzb1:8482
I0423 15:57:14.132560 135968039954240 max_utils.py:284] Jax distributed system initialized!
I0423 15:57:20.090382 135968039954240 max_utils.py:800] System Information: Jax Version: 0.8.1
I0423 15:57:20.090488 135968039954240 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0423 15:57:20.090528 135968039954240 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709
I0423 15:57:20.090563 135968039954240 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 15:57:20.715950 135968039954240 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 15:57:20.716552 135968039954240 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 15:57:20.716747 135968039954240 checkpointing.py:688] Setting up checkpoint logger...
I0423 15:57:20.716801 135968039954240 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0423 15:57:20.716846 135968039954240 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0423 15:57:20.717205 135968039954240 base_pytree_checkpoint_handler.py:415] 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 0x7ba90ed712e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 15:57:23.617870 135968039954240 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0423 15:57:23.618215 135968039954240 checkpoint_manager.py:709] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b96d8831a00>}, handler_registry=None
I0423 15:57:23.618448 135968039954240 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b96d8831a00>` 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`.
I0423 15:57:23.618496 135968039954240 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b96d8836690>` 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`.
I0423 15:57:23.618532 135968039954240 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 0x7b96d8831a00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b96d8831a00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b96d8836690>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b96d8836690>}).
I0423 15:57:23.619045 135968039954240 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0423 15:57:23.619126 135968039954240 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b9641c10180> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0423 15:57:24.307330 135968039954240 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints
I0423 15:57:24.747810 135968039954240 checkpoint_manager.py:929] [process=5][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_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b96d8830e30>
I0423 15:57:24.747998 135968039954240 checkpointing.py:302] Checkpoint manager created!
I0423 15:57:25.640162 135968039954240 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 15:57:25.640273 135968039954240 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 15:57:26.013946 135968039954240 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 15:57:26.014047 135968039954240 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 15:57:26.030028 135968039954240 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 15:57:26.030086 135968039954240 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 15:57:26.052927 135968039954240 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 15:57:26.052992 135968039954240 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 15:57:26.068976 135968039954240 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 15:57:26.069038 135968039954240 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 15:57:26.084898 135968039954240 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 15:57:26.084960 135968039954240 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 15:57:26.109853 135968039954240 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 15:57:26.109922 135968039954240 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 15:57:26.129600 135968039954240 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 15:57:26.129671 135968039954240 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 15:57:26.319694 135968039954240 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 15:57:26.319803 135968039954240 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0423 15:57:27.750918 135968039954240 maxtext_utils.py:1740]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0423 15:57:27.751084 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 15:57:27.751142 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 15:57:27.751202 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   PartitionSpec('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0423 15:57:27.751255 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 15:57:27.751297 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 15:57:27.751350 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 15:57:27.751403 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   PartitionSpec('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0423 15:57:27.751444 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0423 15:57:27.751482 135968039954240 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 15:57:27.751531 135968039954240 maxtext_utils.py:1740]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed', 'vocab')
    Physical:  ('fsdp', None)
I0423 15:57:27.751580 135968039954240 maxtext_utils.py:1740]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed')
    Physical:  (None, 'fsdp')

I0423 15:57:28.254366 135968039954240 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 15:57:28.254456 135968039954240 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 15:57:28.269431 135968039954240 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 15:57:28.269494 135968039954240 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 15:57:39.063272 135968039954240 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0423 15:57:39.064183 135968039954240 metric_logger.py:289] number parameters: 1.104 billion
I0423 15:57:50.069550 135968039954240 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0423 15:57:50.224200 135968039954240 checkpointing.py:798] Waited 0.15463495254516602 seconds for step 0 to finish before starting checkpointing.
I0423 15:57:50.226741 135968039954240 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 15:57:50.228468 135968039954240 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0
I0423 15:57:50.231221 135968039954240 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/0.
I0423 15:57:50.765551 135968039954240 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 15:57:50.766497 135968039954240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0423 15:57:50.766555 135968039954240 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
I0423 15:57:51.014900 135968039954240 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.249428s
I0423 15:57:51.015091 135968039954240 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.077 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5012760162353516 s) (per-host)
I0423 15:57:51.015150 135968039954240 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.501342s (batch_requests_ready=0.248168s, total_serialization_initiated=0.253099s, others=0.000075s)
I0423 15:57:51.015276 135968039954240 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.505737s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.505718)
I0423 15:57:51.016279 135845562009344 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started.
I0423 15:57:51.016447 135968039954240 async_checkpointer.py:561] Finished blocking save. Time taken: 0.787907s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/0.
I0423 15:57:51.027016 135968039954240 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 15:57:51.027320 135843945604864 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0423 15:57:51.027477 135968039954240 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776959870.2267225, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776959870.228507, 'checkpointer_blocking_duration_secs': 0.7880513668060303, 'get_old_steps_start_time': 1776959871.0165792, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776959870.2246833, 'checkpoint_manager_blocking_duration_secs': 0.8027539253234863}
I0423 15:57:51.027580 135968039954240 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0423 15:57:51.027633 135968039954240 max_utils.py:750] 
Memstats: After params initialized:
I0423 15:57:51.027700 135968039954240 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0))
I0423 15:57:51.027739 135968039954240 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0))
I0423 15:57:51.027768 135968039954240 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0))
I0423 15:57:51.027794 135968039954240 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0))
I0423 15:57:51.296432 135968039954240 metric_logger.py:185] completed step: 0, seconds: 11.005, TFLOP/s/device: 1.235, Tokens/s/device: 186.093, total_weights: 65536, loss: 10.874
I0423 15:57:51.462313 135968039954240 metric_logger.py:185] completed step: 1, seconds: 1.226, TFLOP/s/device: 11.087, Tokens/s/device: 1671.137, total_weights: 65536, loss: 10.874
I0423 15:57:51.852342 135968039954240 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 718.782, Tokens/s/device: 108342.591, total_weights: 65536, loss: 10.263
I0423 15:57:51.999868 135968039954240 metric_logger.py:185] completed step: 3, seconds: 0.390, TFLOP/s/device: 34.835, Tokens/s/device: 5250.717, total_weights: 65536, loss: 9.734
I0423 15:57:52.295595 135968039954240 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.852, Tokens/s/device: 13392.799, total_weights: 65536, loss: 9.278
I0423 15:57:52.301281 135968039954240 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.176, Tokens/s/device: 13893.788, total_weights: 65536, loss: 8.892
I0423 15:58:11.193688 135968039954240 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.876, Tokens/s/device: 6914.947, total_weights: 65536, loss: 8.592
I0423 15:58:11.340974 135968039954240 metric_logger.py:185] completed step: 7, seconds: 18.746, TFLOP/s/device: 0.725, Tokens/s/device: 109.252, total_weights: 65536, loss: 8.384
I0423 15:58:11.488498 135968039954240 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.447, Tokens/s/device: 13482.377, total_weights: 65536, loss: 8.255
I0423 15:58:11.635582 135968039954240 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0423 15:58:11.636275 135968039954240 checkpointing.py:798] Waited 0.0007112026214599609 seconds for step 9 to finish before starting checkpointing.
I0423 15:58:11.638273 135968039954240 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 15:58:13.000176    2574 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 15:58:14.929063 135845570402048 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/0/items/array_metadatas/process_5
I0423 15:58:46.437114 135845562009344 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 28.246 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.923261404037476 s) (per-host)
I0423 15:58:46.437227 135845562009344 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.420837s.
I0423 15:58:54.677908 135845562009344 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 63.661503s.
I0423 15:58:54.678197 135843945604864 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 15:58:54.678336 135843945604864 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0423 15:58:54.678405 135843945604864 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0423 15:58:54.680913 135843945604864 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0423 15:58:54.681128 135968039954240 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0423 15:58:54.681274 135968039954240 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.043005 seconds. If this number is high, consider checkpointing less frequently.
I0423 15:58:54.683097 135968039954240 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9
I0423 15:58:54.686360 135968039954240 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/9.
I0423 15:58:55.641834 135968039954240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0423 15:58:55.641937 135968039954240 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
I0423 15:58:55.686937 135968039954240 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046503s
I0423 15:58:55.687099 135968039954240 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.130 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30069804191589355 s) (per-host)
I0423 15:58:55.687150 135968039954240 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.300759s (batch_requests_ready=0.252352s, total_serialization_initiated=0.048340s, others=0.000067s)
I0423 15:58:55.687233 135968039954240 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.304920s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.304905)
I0423 15:58:55.688092 135845535782656 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started.
I0423 15:58:55.688256 135968039954240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.005090s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/9.
I0423 15:58:55.902173 135968039954240 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 15:58:55.902549 135843945604864 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0423 15:58:55.902741 135968039954240 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776959891.6382444, 'wait_for_prev_duration_secs': 43.04300546646118, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776959934.6831353, 'checkpointer_blocking_duration_secs': 1.0052311420440674, 'get_old_steps_start_time': 1776959935.6883855, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776959891.6364973, 'checkpoint_manager_blocking_duration_secs': 44.2662079334259}
I0423 15:58:55.902846 135968039954240 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0423 15:58:55.902894 135968039954240 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 15:59:00.872358 135845570664192 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_02_synthetic/checkpoints/9/items/array_metadatas/process_5
I0423 15:59:38.373237 135845535782656 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 36.746 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.98679542541504 s) (per-host)
I0423 15:59:38.373360 135845535782656 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.685158s.
I0423 15:59:45.420404 135845535782656 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 49.732188s.
I0423 15:59:45.420660 135843945604864 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 15:59:45.420777 135843945604864 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0423 15:59:45.420825 135843945604864 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0423 15:59:45.422301 135843945604864 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0423 15:59:45.422457 135968039954240 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0423 15:59:45.422597 135968039954240 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 15:59:45.423235 135968039954240 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.216, Tokens/s/device: 13899.824, total_weights: 65536, loss: 8.179
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Thu Apr 23 15:59:55 UTC 2026
EXIT_CODE=0