XPK Start: Sat Apr 25 17:59:28 UTC 2026 2026-04-25 17:59:32.972374: 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 00:00:1777139972.985510 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777139972.989245 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777139973.000961 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139973.000982 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139973.000985 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777139973.000987 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:59:52.373138: 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) I0425 17:59:52.908663 135096192923456 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 18:00:01,949:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 18:00:01.949718 135096192923456 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 18:00:01,952:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-rs-wdjop-slice-job-0-0.mt-16-zarr3-false-rs-wdjop:8482 I0425 18:00:01.952227 135096192923456 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-rs-wdjop-slice-job-0-0.mt-16-zarr3-false-rs-wdjop:8482 I0425 18:00:03.174004 135096192923456 max_utils.py:284] Jax distributed system initialized! I0425 18:00:10.047514 135096192923456 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 18:00:10.047629 135096192923456 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 18:00:10.047672 135096192923456 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 I0425 18:00:10.047706 135096192923456 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 18:00:10.672330 135096192923456 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 18:00:10.786651 135096192923456 checkpointing.py:688] Setting up checkpoint logger... I0425 18:00:10.786771 135096192923456 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0425 18:00:10.786815 135096192923456 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 18:00:10.787023 135096192923456 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7ade09b78ef0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 18:00:12.558618 135096192923456 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 18:00:12.558969 135096192923456 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ac988238680>}, handler_registry=None I0425 18:00:12.559220 135096192923456 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ac988238680>` 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`. I0425 18:00:12.559271 135096192923456 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac98824c170>` 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`. I0425 18:00:12.559306 135096192923456 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 0x7ac988238680>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ac988238680>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac98824c170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac98824c170>}). I0425 18:00:12.559642 135096192923456 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 18:00:12.559720 135096192923456 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ac9881b8f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 18:00:14.050004 134975319963392 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777139848930151647, 'commit_timestamp_nsecs': 1777139894550831440, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9/_CHECKPOINT_METADATA I0425 18:00:14.304114 134975319963392 checkpoint_manager.py:1877] Missing metrics for step 9 I0425 18:00:14.910552 135096192923456 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints I0425 18:00:15.349916 135096192923456 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ac988238ec0> I0425 18:00:15.350090 135096192923456 checkpointing.py:302] Checkpoint manager created! I0425 18:00:16.315011 135096192923456 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 18:00:16.315146 135096192923456 checkpointing.py:582] restoring from this run's directory step 9 I0425 18:00:16.823512 135096192923456 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 18:00:18.252273 1420 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 18:00:25.816932 135096192923456 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 208.745 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.570356845855713 s) (per-host) I0425 18:00:25.825829 135096192923456 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 197.932 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.995706081390381 s) (per-host) I0425 18:00:27.049321 135096192923456 checkpointer.py:318] Finished restoring checkpoint in 10.48 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 18:00:27.049480 135096192923456 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1777140016.5656114, 'checkpointer_duration_secs': 10.483836889266968, 'checkpoint_manager_start_time': 1777140016.3157637, 'checkpoint_manager_duration_secs': 10.733686685562134} fsdp: 32 I0425 18:00:27.130568 135096192923456 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 18:00:27.130686 135096192923456 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 18:00:27.130738 135096192923456 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 18:00:27.130782 135096192923456 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 18:00:27.130820 135096192923456 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 18:00:27.130856 135096192923456 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 18:00:27.130894 135096192923456 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 18:00:27.130931 135096192923456 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 18:00:27.130966 135096192923456 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 18:00:27.131001 135096192923456 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 18:00:27.131034 135096192923456 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 18:00:27.131068 135096192923456 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 18:00:27.533820 135096192923456 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 18:00:27.533976 135096192923456 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 18:00:27.539438 135096192923456 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 18:00:27.539531 135096192923456 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 18:00:27.692885 135096192923456 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 18:00:27.693020 135096192923456 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 18:00:27.847596 135096192923456 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 18:00:27.847732 135096192923456 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 18:00:28.010069 135096192923456 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 18:00:28.010244 135096192923456 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 18:00:28.164064 135096192923456 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 18:00:28.164219 135096192923456 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 18:00:28.318909 135096192923456 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 18:00:28.319039 135096192923456 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 18:00:28.487258 135096192923456 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 18:00:28.487396 135096192923456 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 18:00:28.510773 135096192923456 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 18:00:28.510868 135096192923456 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 18:00:39.912902 135096192923456 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. I0425 18:00:39.915311 135096192923456 metric_logger.py:289] number parameters: 1.104 billion I0425 18:00:51.249877 135096192923456 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 18:00:51.400562 135096192923456 checkpointing.py:798] Waited 0.15067100524902344 seconds for step 10 to finish before starting checkpointing. I0425 18:00:51.402856 135096192923456 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 18:00:51.404893 135096192923456 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0425 18:00:51.407587 135096192923456 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10. I0425 18:00:51.994480 135096192923456 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 18:00:51.995571 135096192923456 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 18:00:51.995629 135096192923456 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 I0425 18:00:52.247827 135096192923456 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254465s I0425 18:00:52.248002 135096192923456 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.015 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5116028785705566 s) (per-host) I0425 18:00:52.248055 135096192923456 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.511666s (batch_requests_ready=0.252474s, total_serialization_initiated=0.259120s, others=0.000071s) I0425 18:00:52.248163 135096192923456 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.515874s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.515855) I0425 18:00:52.249068 134972602038016 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 18:00:52.249256 135096192923456 async_checkpointer.py:561] Finished blocking save. Time taken: 0.844290s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10. I0425 18:00:52.292008 135096192923456 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 18:00:52.292334 134971016603392 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 18:00:52.292477 135096192923456 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777140051.402838, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777140051.4049325, 'checkpointer_blocking_duration_secs': 0.8444314002990723, 'get_old_steps_start_time': 1777140052.2493823, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1777140051.4009426, 'checkpoint_manager_blocking_duration_secs': 0.8914990425109863} I0425 18:00:52.292600 135096192923456 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 18:00:52.292677 135096192923456 max_utils.py:750] Memstats: After params initialized: I0425 18:00:52.292730 135096192923456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 18:00:52.292762 135096192923456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 18:00:52.292789 135096192923456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 18:00:52.292813 135096192923456 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 18:00:52.567328 135096192923456 metric_logger.py:185] completed step: 10, seconds: 11.333, TFLOP/s/device: 1.199, Tokens/s/device: 180.703, total_weights: 65536, loss: 6.827 I0425 18:00:52.738818 135096192923456 metric_logger.py:185] completed step: 11, seconds: 1.316, TFLOP/s/device: 10.326, Tokens/s/device: 1556.512, total_weights: 65536, loss: 6.522 I0425 18:00:52.886505 135096192923456 metric_logger.py:185] completed step: 12, seconds: 0.016, TFLOP/s/device: 863.333, Tokens/s/device: 130130.893, total_weights: 65536, loss: 6.260 I0425 18:00:53.033910 135096192923456 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 82.963, Tokens/s/device: 12505.037, total_weights: 65536, loss: 6.039 I0425 18:00:53.041698 135096192923456 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 18:00:53.329050 135096192923456 checkpointing.py:798] Waited 0.2873208522796631 seconds for step 15 to finish before starting checkpointing. I0425 18:00:53.331745 135096192923456 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 18:00:56.886888 134972610430720 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10/items/array_metadatas/process_6 I0425 18:01:34.234045 134972602038016 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.169 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.49760699272156 s) (per-host) I0425 18:01:34.234180 134972602038016 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.984981s. I0425 18:01:41.086553 134972602038016 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.837338s. I0425 18:01:41.086817 134971016603392 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 18:01:41.086874 134971016603392 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 18:01:41.086919 134971016603392 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 18:01:41.089076 134971016603392 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 18:01:41.089264 135096192923456 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0425 18:01:41.089416 135096192923456 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.757676 seconds. If this number is high, consider checkpointing less frequently. I0425 18:01:41.091148 135096192923456 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0425 18:01:41.095232 135096192923456 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15. I0425 18:01:42.409289 135096192923456 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 18:01:42.409391 135096192923456 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 I0425 18:01:42.452852 135096192923456 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045856s I0425 18:01:42.453023 135096192923456 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.088 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7388663291931152 s) (per-host) I0425 18:01:42.453077 135096192923456 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.738930s (batch_requests_ready=0.688398s, total_serialization_initiated=0.050461s, others=0.000071s) I0425 18:01:42.453190 135096192923456 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.743953s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.743939) I0425 18:01:42.454037 134972610430720 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 18:01:42.454178 135096192923456 async_checkpointer.py:561] Finished blocking save. Time taken: 1.362960s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15. I0425 18:01:42.456210 135096192923456 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 18:01:42.456471 134971016603392 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 18:01:42.456598 135096192923456 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777140053.3317142, 'wait_for_prev_duration_secs': 47.757675647735596, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777140101.0911863, 'checkpointer_blocking_duration_secs': 1.3631675243377686, 'get_old_steps_start_time': 1777140102.4543736, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777140053.3294456, 'checkpoint_manager_blocking_duration_secs': 49.12711691856384} I0425 18:01:42.456708 135096192923456 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 18:01:42.457550 135096192923456 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.356, Tokens/s/device: 13920.893, total_weights: 65536, loss: 5.858 I0425 18:01:42.465161 135096192923456 metric_logger.py:185] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.117, Tokens/s/device: 13734.090, total_weights: 65536, loss: 5.713 I0425 18:02:00.660159 135096192923456 metric_logger.py:185] completed step: 16, seconds: 49.422, TFLOP/s/device: 0.275, Tokens/s/device: 41.439, total_weights: 65536, loss: 5.599 I0425 18:02:00.807753 135096192923456 metric_logger.py:185] completed step: 17, seconds: 18.050, TFLOP/s/device: 0.753, Tokens/s/device: 113.462, total_weights: 65536, loss: 5.511 I0425 18:02:00.955160 135096192923456 metric_logger.py:185] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.751, Tokens/s/device: 13377.577, total_weights: 65536, loss: 5.442 I0425 18:02:01.103689 135096192923456 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 18:02:01.104329 135096192923456 checkpointing.py:798] Waited 0.000644683837890625 seconds for step 19 to finish before starting checkpointing. I0425 18:02:01.106687 135096192923456 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 18:02:04.234472 134971536688896 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15/items/array_metadatas/process_6 I0425 18:02:41.719577 134972610430720 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.324 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.00538182258606 s) (per-host) I0425 18:02:41.719687 134972610430720 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.265430s. I0425 18:02:52.221832 134972610430720 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.767559s. I0425 18:02:52.222126 134971016603392 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 18:02:52.222259 134971016603392 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 18:02:52.222312 134971016603392 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 18:02:52.224809 134971016603392 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 18:02:52.224993 135096192923456 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 18:02:52.225151 135096192923456 checkpoint_manager.py:1458] Waiting for previous save to complete took 51.118476 seconds. If this number is high, consider checkpointing less frequently. I0425 18:02:52.226802 135096192923456 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0425 18:02:52.230276 135096192923456 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19. I0425 18:02:52.800523 135096192923456 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 18:02:52.800626 135096192923456 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 I0425 18:02:52.848248 135096192923456 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049794s I0425 18:02:52.848422 135096192923456 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.922 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.313417911529541 s) (per-host) I0425 18:02:52.848474 135096192923456 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313480s (batch_requests_ready=0.259106s, total_serialization_initiated=0.054307s, others=0.000067s) I0425 18:02:52.848571 135096192923456 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317495s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.317483) I0425 18:02:52.849465 134975277999872 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 18:02:52.849629 135096192923456 async_checkpointer.py:561] Finished blocking save. Time taken: 0.622758s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19. I0425 18:02:53.149405 135096192923456 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 18:02:53.149773 134971016603392 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 18:02:53.149938 135096192923456 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777140121.1066597, 'wait_for_prev_duration_secs': 51.118475675582886, 'time_between_consecutive_saves_sec': 20.01752734184265, 'checkpointer_blocking_start_time': 1777140172.2268405, 'checkpointer_blocking_duration_secs': 0.6228938102722168, 'get_old_steps_start_time': 1777140172.8497481, 'get_old_steps_duration_secs': 1.9311904907226562e-05, 'checkpoint_manager_blocking_start_time': 1777140121.104564, 'checkpoint_manager_blocking_duration_secs': 52.04534077644348} I0425 18:02:53.150060 135096192923456 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 18:02:53.150118 135096192923456 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 18:02:58.532692 134975286392576 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19/items/array_metadatas/process_6 I0425 18:03:35.162185 134975277999872 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.056 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.62714219093323 s) (per-host) I0425 18:03:35.162298 134975277999872 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.312717s. I0425 18:03:42.800591 134975277999872 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.950995s. I0425 18:03:42.800838 134971016603392 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 18:03:42.800895 134971016603392 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 18:03:42.800936 134971016603392 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 18:03:42.803242 134971016603392 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 18:03:42.803421 135096192923456 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0425 18:03:42.803588 135096192923456 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 18:03:42.804188 135096192923456 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.351, Tokens/s/device: 13920.136, total_weights: 65536, loss: 5.385 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 18:03:54 UTC 2026 EXIT_CODE=0