MaxView

XPK Start: Wed Apr 22 16:32:27 UTC 2026
2026-04-22 16:32:32.013514: 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 00:00:1776875552.026031      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776875552.029718      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776875552.040928      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776875552.040947      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776875552.040950      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776875552.040952      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-22 16:32:51.241453: 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)
I0422 16:32:51.767068 139326327740224 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-22 16:33:00,806:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0422 16:33:00.806908 139326327740224 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-22 16:33:00,809:jax._src.distributed:157: Connecting to JAX distributed service on mt-11-optimizer-offload-08mh1-slice-job-0-0.mt-11-optimizer-offload-08mh1:8482
I0422 16:33:00.809265 139326327740224 distributed.py:157] Connecting to JAX distributed service on mt-11-optimizer-offload-08mh1-slice-job-0-0.mt-11-optimizer-offload-08mh1:8482
I0422 16:33:01.656550 139326327740224 max_utils.py:284] Jax distributed system initialized!
I0422 16:33:07.687360 139326327740224 max_utils.py:800] System Information: Jax Version: 0.8.1
I0422 16:33:07.687465 139326327740224 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0422 16:33:07.687508 139326327740224 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
I0422 16:33:07.687547 139326327740224 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 16:33:08.302102 139326327740224 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 16:33:08.302703 139326327740224 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 16:33:08.302888 139326327740224 checkpointing.py:688] Setting up checkpoint logger...
I0422 16:33:08.302939 139326327740224 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 16:33:08.302984 139326327740224 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0422 16:33:08.303334 139326327740224 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 0x7eb6dd779340>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 16:33:11.217854 139326327740224 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 16:33:11.218204 139326327740224 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 0x7ea58353e0c0>}, handler_registry=None
I0422 16:33:11.218438 139326327740224 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ea58353e0c0>` 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`.
I0422 16:33:11.218488 139326327740224 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ea5820f8ef0>` 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`.
I0422 16:33:11.218524 139326327740224 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 0x7ea58353e0c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ea58353e0c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ea5820f8ef0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ea5820f8ef0>}).
I0422 16:33:11.219044 139326327740224 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0422 16:33:11.219125 139326327740224 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ea581050180> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0422 16:33:12.793515 139326327740224 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints
I0422 16:33:12.795762 139326327740224 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ea5820f8bf0>
I0422 16:33:12.795877 139326327740224 checkpointing.py:302] Checkpoint manager created!
I0422 16:33:13.683178 139326327740224 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 16:33:13.683291 139326327740224 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 16:33:14.055547 139326327740224 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 16:33:14.055641 139326327740224 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 16:33:14.071460 139326327740224 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 16:33:14.071519 139326327740224 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 16:33:14.094317 139326327740224 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 16:33:14.094382 139326327740224 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 16:33:14.110200 139326327740224 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 16:33:14.110267 139326327740224 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 16:33:14.125961 139326327740224 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 16:33:14.126027 139326327740224 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 16:33:14.150824 139326327740224 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 16:33:14.150895 139326327740224 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 16:33:14.170633 139326327740224 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 16:33:14.170710 139326327740224 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 16:33:14.359454 139326327740224 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 16:33:14.359562 139326327740224 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0422 16:33:16.235725 139326327740224 maxtext_utils.py:1740]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0422 16:33:16.235858 139326327740224 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)
I0422 16:33:16.235914 139326327740224 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)
I0422 16:33:16.235976 139326327740224 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')
I0422 16:33:16.236035 139326327740224 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)
I0422 16:33:16.236075 139326327740224 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)
I0422 16:33:16.236129 139326327740224 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)
I0422 16:33:16.236183 139326327740224 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')
I0422 16:33:16.236223 139326327740224 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)
I0422 16:33:16.236261 139326327740224 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)
I0422 16:33:16.236317 139326327740224 maxtext_utils.py:1740]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed', 'vocab')
    Physical:  ('fsdp', None)

I0422 16:33:16.236365 139326327740224 maxtext_utils.py:1740]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed')
    Physical:  (None, 'fsdp')
I0422 16:33:16.752885 139326327740224 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 16:33:16.752980 139326327740224 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 16:33:16.767840 139326327740224 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 16:33:16.767897 139326327740224 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 16:33:27.518843 139326327740224 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.
I0422 16:33:27.519724 139326327740224 metric_logger.py:289] number parameters: 1.104 billion
I0422 16:33:38.612798 139326327740224 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0422 16:33:38.782290 139326327740224 checkpointing.py:798] Waited 0.16946101188659668 seconds for step 0 to finish before starting checkpointing.
I0422 16:33:38.784857 139326327740224 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 16:33:38.786742 139326327740224 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0
I0422 16:33:38.789630 139326327740224 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/0.
I0422 16:33:39.366373 139326327740224 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 16:33:39.367523 139326327740224 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0422 16:33:39.367581 139326327740224 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
I0422 16:33:39.613302 139326327740224 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.247008s
I0422 16:33:39.613470 139326327740224 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.068 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5027756690979004 s) (per-host)
I0422 16:33:39.613523 139326327740224 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.502840s (batch_requests_ready=0.250992s, total_serialization_initiated=0.251776s, others=0.000072s)
I0422 16:33:39.613612 139326327740224 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.507246s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.507228)
I0422 16:33:39.614605 139196794660608 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started.
I0422 16:33:39.614751 139326327740224 async_checkpointer.py:561] Finished blocking save. Time taken: 0.827935s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/0.
I0422 16:33:39.658407 139326327740224 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 16:33:39.658753 139193816692480 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0422 16:33:39.658912 139326327740224 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776875618.7848387, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776875618.786782, 'checkpointer_blocking_duration_secs': 0.8281588554382324, 'get_old_steps_start_time': 1776875619.6149614, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776875618.782879, 'checkpoint_manager_blocking_duration_secs': 0.8759944438934326}
I0422 16:33:39.659017 139326327740224 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0422 16:33:39.659069 139326327740224 max_utils.py:750] 
Memstats: After params initialized:
I0422 16:33:39.659121 139326327740224 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_18(process=5,(2,4,0,0))
I0422 16:33:39.659154 139326327740224 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_19(process=5,(3,4,0,0))
I0422 16:33:39.659188 139326327740224 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_22(process=5,(2,5,0,0))
I0422 16:33:39.659214 139326327740224 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_23(process=5,(3,5,0,0))
I0422 16:33:39.925573 139326327740224 metric_logger.py:185] completed step: 0, seconds: 11.093, TFLOP/s/device: 1.225, Tokens/s/device: 184.622, total_weights: 65536, loss: 10.874
I0422 16:33:40.118255 139326327740224 metric_logger.py:185] completed step: 1, seconds: 1.311, TFLOP/s/device: 10.360, Tokens/s/device: 1561.643, total_weights: 65536, loss: 10.874
I0422 16:33:40.535148 139326327740224 metric_logger.py:185] completed step: 2, seconds: 0.029, TFLOP/s/device: 464.517, Tokens/s/device: 70017.094, total_weights: 65536, loss: 10.263
I0422 16:33:40.699130 139326327740224 metric_logger.py:185] completed step: 3, seconds: 0.417, TFLOP/s/device: 32.565, Tokens/s/device: 4908.575, total_weights: 65536, loss: 9.735
I0422 16:33:41.026746 139326327740224 metric_logger.py:185] completed step: 4, seconds: 0.169, TFLOP/s/device: 80.341, Tokens/s/device: 12109.816, total_weights: 65536, loss: 9.278
I0422 16:33:41.032489 139326327740224 metric_logger.py:185] completed step: 5, seconds: 0.164, TFLOP/s/device: 82.917, Tokens/s/device: 12498.093, total_weights: 65536, loss: 8.892
I0422 16:33:59.854337 139326327740224 metric_logger.py:185] completed step: 6, seconds: 0.328, TFLOP/s/device: 41.419, Tokens/s/device: 6243.065, total_weights: 65536, loss: 8.592
I0422 16:34:00.018007 139326327740224 metric_logger.py:185] completed step: 7, seconds: 18.659, TFLOP/s/device: 0.728, Tokens/s/device: 109.760, total_weights: 65536, loss: 8.384
I0422 16:34:00.181713 139326327740224 metric_logger.py:185] completed step: 8, seconds: 0.169, TFLOP/s/device: 80.620, Tokens/s/device: 12151.851, total_weights: 65536, loss: 8.255
I0422 16:34:00.345055 139326327740224 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0422 16:34:00.345735 139326327740224 checkpointing.py:798] Waited 0.0007116794586181641 seconds for step 9 to finish before starting checkpointing.
I0422 16:34:00.348067 139326327740224 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 16:34:00.965762    2542 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 16:34:02.916765 139196803053312 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_5
I0422 16:34:34.859040 139196794660608 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 28.333 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.74830746650696 s) (per-host)
I0422 16:34:34.859161 139196794660608 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.244334s.
I0422 16:34:43.204429 139196794660608 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 63.589586s.
I0422 16:34:43.204730 139193816692480 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 16:34:43.204857 139193816692480 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0422 16:34:43.204908 139193816692480 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0422 16:34:43.206640 139193816692480 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0422 16:34:43.206838 139326327740224 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.
W0422 16:34:43.206967 139326327740224 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.858904 seconds. If this number is high, consider checkpointing less frequently.
I0422 16:34:43.208682 139326327740224 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9
I0422 16:34:43.212108 139326327740224 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/9.
I0422 16:34:44.086892 139326327740224 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0422 16:34:44.086991 139326327740224 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
I0422 16:34:44.129862 139326327740224 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043991s
I0422 16:34:44.130036 139326327740224 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.976 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3099837303161621 s) (per-host)
I0422 16:34:44.130090 139326327740224 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.310053s (batch_requests_ready=0.264254s, total_serialization_initiated=0.045723s, others=0.000077s)
I0422 16:34:44.130176 139326327740224 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314129s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.314114)
I0422 16:34:44.131045 139196266182400 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started.
I0422 16:34:44.131206 139326327740224 async_checkpointer.py:561] Finished blocking save. Time taken: 0.922448s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/9.
I0422 16:34:44.136088 139326327740224 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 16:34:44.136380 139193816692480 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0422 16:34:44.136540 139326327740224 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776875640.3480387, 'wait_for_prev_duration_secs': 42.858904123306274, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776875683.2087233, 'checkpointer_blocking_duration_secs': 0.9225926399230957, 'get_old_steps_start_time': 1776875684.1313355, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776875640.3459554, 'checkpoint_manager_blocking_duration_secs': 43.79055070877075}
I0422 16:34:44.136645 139326327740224 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0422 16:34:44.136703 139326327740224 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 16:34:48.892793 139196803053312 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_5
I0422 16:35:26.077581 139196266182400 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 37.379 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.25749063491821 s) (per-host)
I0422 16:35:26.077717 139196266182400 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.946558s.
I0422 16:35:32.043538 139196266182400 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 47.912364s.
I0422 16:35:32.043842 139193816692480 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 16:35:32.043964 139193816692480 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0422 16:35:32.044011 139193816692480 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0422 16:35:32.046572 139193816692480 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0422 16:35:32.046767 139326327740224 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.
I0422 16:35:32.046914 139326327740224 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 16:35:32.047530 139326327740224 metric_logger.py:185] completed step: 9, seconds: 0.164, TFLOP/s/device: 83.038, Tokens/s/device: 12516.348, 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: Wed Apr 22 16:35:43 UTC 2026
EXIT_CODE=0