MaxView

← Back to run

Log Summary

XPK Start: Mon Apr 20 21:05:45 UTC 2026
2026-04-20 21:05:49.120656: 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 00:00:1776719149.133577      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776719149.137328      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776719149.148891      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776719149.148915      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776719149.148917      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776719149.148919      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-20 21:06:08.433179: 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)
I0420 21:06:08.958531 140286415558464 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-20 21:06:18,000:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0420 21:06:18.000266 140286415558464 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-20 21:06:18,002:jax._src.distributed:157: Connecting to JAX distributed service on mt-06-grad-accum-3g4oe-slice-job-0-0.mt-06-grad-accum-3g4oe:8482
I0420 21:06:18.002737 140286415558464 distributed.py:157] Connecting to JAX distributed service on mt-06-grad-accum-3g4oe-slice-job-0-0.mt-06-grad-accum-3g4oe:8482
I0420 21:06:18.955282 140286415558464 max_utils.py:284] Jax distributed system initialized!
I0420 21:06:24.944762 140286415558464 max_utils.py:800] System Information: Jax Version: 0.8.1
I0420 21:06:24.944870 140286415558464 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0420 21:06:24.944915 140286415558464 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
I0420 21:06:24.944950 140286415558464 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 21:06:25.559774 140286415558464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 21:06:25.675076 140286415558464 checkpointing.py:688] Setting up checkpoint logger...
I0420 21:06:25.675204 140286415558464 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 21:06:25.675249 140286415558464 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0420 21:06:25.675455 140286415558464 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 0x7f9670979640>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 21:06:28.644792 140286415558464 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 21:06:28.645165 140286415558464 checkpoint_manager.py:709] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f82a8154680>}, handler_registry=None
I0420 21:06:28.645402 140286415558464 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f82a8154680>` 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`.
I0420 21:06:28.645452 140286415558464 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f82a8705c10>` 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`.
I0420 21:06:28.645487 140286415558464 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 0x7f82a8154680>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f82a8154680>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f82a8705c10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f82a8705c10>}).
I0420 21:06:28.645818 140286415558464 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0420 21:06:28.645896 140286415558464 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f82a82a4f40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0420 21:06:29.400132 140286415558464 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints
I0420 21:06:29.840401 140286415558464 checkpoint_manager.py:929] [process=0][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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f82a870e090>
I0420 21:06:29.840585 140286415558464 checkpointing.py:302] Checkpoint manager created!
I0420 21:06:31.542440 140286415558464 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 21:06:31.542561 140286415558464 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 21:06:32.934767 140286415558464 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0420 21:06:32.934888 140286415558464 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 21:06:32.934944 140286415558464 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 21:06:32.934990 140286415558464 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0420 21:06:32.935029 140286415558464 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 21:06:32.935066 140286415558464 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 21:06:32.935118 140286415558464 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 21:06:32.935158 140286415558464 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0420 21:06:32.935196 140286415558464 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 21:06:32.935231 140286415558464 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 21:06:32.935265 140286415558464 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0420 21:06:32.935298 140286415558464 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0420 21:06:33.323781 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0420 21:06:33.323876 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0420 21:06:33.338124 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0420 21:06:33.338181 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0420 21:06:33.366533 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0420 21:06:33.366600 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0420 21:06:33.380644 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0420 21:06:33.380715 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0420 21:06:33.408839 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0420 21:06:33.408908 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0420 21:06:33.423041 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0420 21:06:33.423119 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0420 21:06:33.465531 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0420 21:06:33.465609 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0420 21:06:33.479712 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0420 21:06:33.479770 140286415558464 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0420 21:06:33.702251 140286415558464 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 21:06:33.702344 140286415558464 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 21:06:33.706314 140286415558464 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 21:06:33.706368 140286415558464 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 21:06:33.722594 140286415558464 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 21:06:33.722654 140286415558464 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 21:06:33.738539 140286415558464 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 21:06:33.738604 140286415558464 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 21:06:33.762246 140286415558464 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 21:06:33.762312 140286415558464 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 21:06:33.778459 140286415558464 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 21:06:33.778516 140286415558464 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 21:06:33.794572 140286415558464 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 21:06:33.794631 140286415558464 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 21:06:33.819986 140286415558464 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 21:06:33.820054 140286415558464 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 21:06:33.839560 140286415558464 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 21:06:33.839637 140286415558464 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 21:06:45.314276 140286415558464 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0420 21:06:45.321662 140286415558464 metric_logger.py:289] number parameters: 1.104 billion
I0420 21:06:47.540905 140286415558464 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 21:06:57.193763 140286415558464 checkpointing.py:798] Waited 9.652843952178955 seconds for step 0 to finish before starting checkpointing.
I0420 21:06:57.195971 140286415558464 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 21:06:57.197629 140286415558464 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 0
I0420 21:06:57.200385 140286415558464 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0.
I0420 21:06:57.919549 140286415558464 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 21:06:58.170145 140163103586048 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0
I0420 21:06:58.198415 140286415558464 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 21:06:58.198511 140286415558464 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
I0420 21:06:58.442313 140286415558464 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245949s
I0420 21:06:58.443273 140286415558464 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.990 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5158870220184326 s) (per-host)
I0420 21:06:58.443350 140286415558464 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.515984s (batch_requests_ready=0.261669s, total_serialization_initiated=0.253437s, others=0.000878s)
I0420 21:06:58.443438 140286415558464 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.519936s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.519917)
I0420 21:06:58.444412 140162525964032 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0420 21:06:58.444599 140286415558464 async_checkpointer.py:561] Finished blocking save. Time taken: 1.246869s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0.
I0420 21:06:58.472505 140286415558464 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 21:06:58.472819 140164132988672 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 21:06:58.472952 140286415558464 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776719217.1959534, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776719217.1976702, 'checkpointer_blocking_duration_secs': 1.2470369338989258, 'get_old_steps_start_time': 1776719218.4447289, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776719217.1941752, 'checkpoint_manager_blocking_duration_secs': 1.278738260269165}
I0420 21:06:58.473058 140286415558464 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 21:06:58.473152 140286415558464 max_utils.py:750] 
Memstats: After params initialized:
I0420 21:06:58.473204 140286415558464 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0420 21:06:58.473237 140286415558464 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0420 21:06:58.473264 140286415558464 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0420 21:06:58.473289 140286415558464 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0420 21:06:58.785172 140286415558464 metric_logger.py:185] completed step: 0, seconds: 1.931, TFLOP/s/device: 28.150, Tokens/s/device: 4243.073, total_weights: 262144, loss: 10.876
I0420 21:06:58.788045 140286415558464 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/tensorboard/'
I0420 21:06:59.358011 140286415558464 metric_logger.py:185] completed step: 1, seconds: 11.241, TFLOP/s/device: 4.835, Tokens/s/device: 728.774, total_weights: 262144, loss: 10.876
I0420 21:06:59.936652 140286415558464 metric_logger.py:185] completed step: 2, seconds: 0.433, TFLOP/s/device: 125.392, Tokens/s/device: 18900.486, total_weights: 262144, loss: 10.561
I0420 21:06:59.953849 140162556217088 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/items
I0420 21:07:00.140366 140161987921664 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776719219706741092, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776719219706741092, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 21:07:00.514926 140286415558464 metric_logger.py:185] completed step: 3, seconds: 0.153, TFLOP/s/device: 354.097, Tokens/s/device: 53373.294, total_weights: 262144, loss: 10.270
I0420 21:07:01.119478    2363 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 21:07:38.853685 140286415558464 metric_logger.py:185] completed step: 4, seconds: 0.580, TFLOP/s/device: 93.654, Tokens/s/device: 14116.617, total_weights: 262144, loss: 10.021
I0420 21:07:38.886327 140286415558464 metric_logger.py:185] completed step: 5, seconds: 0.577, TFLOP/s/device: 94.208, Tokens/s/device: 14200.035, total_weights: 262144, loss: 9.819
I0420 21:07:39.461791 140286415558464 metric_logger.py:185] completed step: 6, seconds: 38.359, TFLOP/s/device: 1.417, Tokens/s/device: 213.560, total_weights: 262144, loss: 9.666
I0420 21:07:39.622473 140167390160640 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/items/array_metadatas/process_0
I0420 21:07:40.040257 140286415558464 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 6610.937, Tokens/s/device: 996472.449, total_weights: 262144, loss: 9.560
I0420 21:07:40.618509 140286415558464 metric_logger.py:185] completed step: 8, seconds: 0.578, TFLOP/s/device: 94.071, Tokens/s/device: 14179.413, total_weights: 262144, loss: 9.495
I0420 21:07:41.197728 140286415558464 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 21:07:41.198403 140286415558464 checkpointing.py:798] Waited 0.0006866455078125 seconds for step 9 to finish before starting checkpointing.
I0420 21:07:41.200292 140286415558464 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 21:08:13.812007 140167381767936 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 72.717538s (commit=71.626960s, array_metadata_write=1.090578s)
I0420 21:08:13.813024 140162525964032 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 20.816 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 75.88563704490662 s) (per-host)
I0420 21:08:13.813117 140162525964032 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 75.368595s.
I0420 21:08:14.569985 140162525964032 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776719219706741092, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 21:08:15.324943 140161987921664 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776719219706741092, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 21:08:15.380333 140162525964032 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000441s.
I0420 21:08:16.156876 140162525964032 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 21:08:17.114045 140162525964032 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.418252s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/items
I0420 21:08:17.115650 140162525964032 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0/items
I0420 21:08:17.520286 140162525964032 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0
I0420 21:08:18.833293 140162525964032 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0`.
I0420 21:08:18.834206 140162525964032 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 81.636508s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/0
I0420 21:08:18.835534 140162525964032 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 80.391009s.
I0420 21:08:18.835725 140164132988672 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 21:08:18.835847 140164132988672 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 21:08:18.835913 140164132988672 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 21:08:18.838042 140164132988672 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 21:08:18.838186 140286415558464 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0420 21:08:18.838273 140286415558464 checkpoint_manager.py:1458] Waiting for previous save to complete took 37.637988 seconds. If this number is high, consider checkpointing less frequently.
I0420 21:08:18.840673 140286415558464 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 9
I0420 21:08:18.844028 140286415558464 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9.
I0420 21:08:19.390830 140162525964032 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9
I0420 21:08:19.401673 140286415558464 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0420 21:08:19.401770 140286415558464 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
I0420 21:08:19.431411 140286415558464 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.031956s
I0420 21:08:19.432307 140286415558464 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.407 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28527235984802246 s) (per-host)
I0420 21:08:19.432383 140286415558464 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285366s (batch_requests_ready=0.246855s, total_serialization_initiated=0.037694s, others=0.000817s)
I0420 21:08:19.432467 140286415558464 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289427s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.289411)
I0420 21:08:19.433311 140160947709696 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0420 21:08:19.433461 140286415558464 async_checkpointer.py:561] Finished blocking save. Time taken: 0.592715s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9.
I0420 21:08:19.465210 140286415558464 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 21:08:19.465513 140164132988672 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 21:08:19.465686 140286415558464 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776719261.2002625, 'wait_for_prev_duration_secs': 37.63798785209656, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776719298.8407147, 'checkpointer_blocking_duration_secs': 0.592853307723999, 'get_old_steps_start_time': 1776719299.4335938, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776719261.1986454, 'checkpoint_manager_blocking_duration_secs': 38.267000913619995}
I0420 21:08:19.465825 140286415558464 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 21:08:19.465868 140286415558464 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 21:08:21.625174 140161484580608 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9/items
I0420 21:08:24.468311 140161492973312 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9/items/array_metadatas/process_0
I0420 21:09:04.940520 140161442617088 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.150283s (commit=41.171901s, array_metadata_write=0.978382s)
I0420 21:09:04.942510 140160947709696 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 34.493 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.79544496536255 s) (per-host)
I0420 21:09:04.942633 140160947709696 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.509213s.
I0420 21:09:06.699011 140160947709696 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000396s.
I0420 21:09:07.137844 140160947709696 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 21:09:08.105722 140160947709696 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.371417s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9/items
I0420 21:09:08.107381 140160947709696 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9/items
I0420 21:09:08.540897 140160947709696 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9
I0420 21:09:09.873105 140160947709696 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9`.
I0420 21:09:09.873905 140160947709696 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 51.033163s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_06_grad_accum/checkpoints/9
I0420 21:09:09.875324 140160947709696 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 50.441903s.
I0420 21:09:09.875488 140164132988672 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 21:09:09.875545 140164132988672 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 21:09:09.875615 140164132988672 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 21:09:09.877505 140164132988672 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 21:09:09.877661 140286415558464 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0420 21:09:09.877816 140286415558464 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 21:09:09.878465 140286415558464 metric_logger.py:185] completed step: 9, seconds: 0.577, TFLOP/s/device: 94.116, Tokens/s/device: 14186.117, total_weights: 262144, loss: 9.457
Per train step:
 Total TFLOPs: 54.35 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 21:09:20 UTC 2026
EXIT_CODE=0