XPK Start: Fri Apr 24 16:10:47 UTC 2026 2026-04-24 16:10:51.217926: 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 00:00:1777047051.231608 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777047051.235370 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777047051.247511 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047051.247532 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047051.247535 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047051.247537 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 16:11:10.485676: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0424 16:11:11.016133 134109413422912 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 16:11:20,058:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 16:11:20.058889 134109413422912 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 16:11:20,061:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-sv-4snvy-slice-job-0-0.mt-16-zarr3-true-sv-4snvy:8482 I0424 16:11:20.061401 134109413422912 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-sv-4snvy-slice-job-0-0.mt-16-zarr3-true-sv-4snvy:8482 I0424 16:11:21.080760 134109413422912 max_utils.py:284] Jax distributed system initialized! I0424 16:11:28.000532 134109413422912 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 16:11:28.000646 134109413422912 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 16:11:28.000687 134109413422912 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 I0424 16:11:28.000721 134109413422912 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 16:11:28.621681 134109413422912 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:11:28.622297 134109413422912 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:11:28.622483 134109413422912 checkpointing.py:688] Setting up checkpoint logger... I0424 16:11:28.622533 134109413422912 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 16:11:28.622582 134109413422912 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 16:11:28.622929 134109413422912 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 0x79f8379793d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 16:11:31.576720 134109413422912 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 16:11:31.576961 134109413422912 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 0x79e3bc205a60>}, handler_registry=None I0424 16:11:31.577203 134109413422912 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e3bc205a60>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0424 16:11:31.577259 134109413422912 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e3bc20a810>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0424 16:11:31.577297 134109413422912 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 0x79e3bc205a60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e3bc205a60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e3bc20a810>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e3bc20a810>}). I0424 16:11:31.577807 134109413422912 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 16:11:31.577887 134109413422912 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79e3ac360180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 16:11:33.091238 134109413422912 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints I0424 16:11:33.114082 134109413422912 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/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79e3bc204b30> I0424 16:11:33.114230 134109413422912 checkpointing.py:302] Checkpoint manager created! I0424 16:11:34.012640 134109413422912 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 16:11:34.012751 134109413422912 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:11:34.379750 134109413422912 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:11:34.379848 134109413422912 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:11:34.395783 134109413422912 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:11:34.395851 134109413422912 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:11:34.418608 134109413422912 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:11:34.418675 134109413422912 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:11:34.434631 134109413422912 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:11:34.434692 134109413422912 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:11:34.450451 134109413422912 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:11:34.450515 134109413422912 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:11:34.475387 134109413422912 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 16:11:34.475459 134109413422912 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:11:34.495043 134109413422912 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 16:11:34.495119 134109413422912 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 16:11:34.684341 134109413422912 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 16:11:34.684451 134109413422912 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 16:11:36.121187 134109413422912 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0424 16:11:36.121322 134109413422912 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) I0424 16:11:36.121380 134109413422912 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) I0424 16:11:36.121439 134109413422912 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') I0424 16:11:36.121499 134109413422912 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) I0424 16:11:36.121541 134109413422912 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) I0424 16:11:36.121597 134109413422912 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) I0424 16:11:36.121653 134109413422912 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') I0424 16:11:36.121695 134109413422912 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) I0424 16:11:36.121733 134109413422912 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) I0424 16:11:36.121782 134109413422912 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0424 16:11:36.121833 134109413422912 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0424 16:11:36.625916 134109413422912 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:11:36.626013 134109413422912 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:11:36.640963 134109413422912 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:11:36.641028 134109413422912 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:11:47.598143 134109413422912 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. I0424 16:11:47.599026 134109413422912 metric_logger.py:289] number parameters: 1.104 billion I0424 16:11:58.898459 134109413422912 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 16:11:59.056835 134109413422912 checkpointing.py:798] Waited 0.1583561897277832 seconds for step 0 to finish before starting checkpointing. I0424 16:11:59.059262 134109413422912 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:11:59.061030 134109413422912 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 16:11:59.063834 134109413422912 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/0. I0424 16:11:59.676022 134109413422912 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 16:11:59.677021 134109413422912 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:11:59.677080 134109413422912 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 16:11:59.926200 134109413422912 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250258s I0424 16:11:59.926377 134109413422912 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.991 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5157909393310547 s) (per-host) I0424 16:11:59.926434 134109413422912 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.515858s (batch_requests_ready=0.261917s, total_serialization_initiated=0.253865s, others=0.000075s) I0424 16:11:59.926538 134109413422912 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.519897s (all_items=0.000026s, per_item={'items': '0.00002646'}, temp_paths=0.519870) I0424 16:11:59.927640 133984033658624 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:11:59.927815 134109413422912 async_checkpointer.py:561] Finished blocking save. Time taken: 0.866714s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/0. I0424 16:11:59.952633 134109413422912 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:11:59.952935 133982930835200 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:11:59.953109 134109413422912 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047119.0592391, 'wait_for_prev_duration_secs': 6.818771362304688e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047119.0610685, 'checkpointer_blocking_duration_secs': 0.866856575012207, 'get_old_steps_start_time': 1777047119.9279463, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777047119.0574799, 'checkpoint_manager_blocking_duration_secs': 0.895576000213623} I0424 16:11:59.953212 134109413422912 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 16:11:59.953264 134109413422912 max_utils.py:750] Memstats: After params initialized: I0424 16:11:59.953316 134109413422912 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 16:11:59.953349 134109413422912 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 16:11:59.953376 134109413422912 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 16:11:59.953401 134109413422912 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 16:12:00.220563 134109413422912 metric_logger.py:185] completed step: 0, seconds: 11.299, TFLOP/s/device: 1.202, Tokens/s/device: 181.250, total_weights: 65536, loss: 10.874 I0424 16:12:00.389956 134109413422912 metric_logger.py:185] completed step: 1, seconds: 1.321, TFLOP/s/device: 10.289, Tokens/s/device: 1550.822, total_weights: 65536, loss: 10.874 I0424 16:12:00.803232 134109413422912 metric_logger.py:185] completed step: 2, seconds: 0.023, TFLOP/s/device: 600.271, Tokens/s/device: 90479.346, total_weights: 65536, loss: 10.560 I0424 16:12:00.950623 134109413422912 metric_logger.py:185] completed step: 3, seconds: 0.414, TFLOP/s/device: 32.834, Tokens/s/device: 4949.059, total_weights: 65536, loss: 9.980 I0424 16:12:00.955395 134109413422912 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 16:12:01.245127 134109413422912 checkpointing.py:798] Waited 0.2896866798400879 seconds for step 5 to finish before starting checkpointing. I0424 16:12:01.249318 134109413422912 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:12:02.630401 2593 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 16:12:04.568796 133984042051328 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/0/items/array_metadatas/process_6 I0424 16:12:36.684584 133984033658624 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 42.378 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.27395701408386 s) (per-host) I0424 16:12:36.684702 133984033658624 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.756950s. I0424 16:12:44.410761 133984033658624 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.482993s. I0424 16:12:44.411072 133982930835200 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:12:44.411203 133982930835200 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:12:44.411260 133982930835200 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:12:44.412871 133982930835200 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 16:12:44.413042 134109413422912 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0424 16:12:44.413214 134109413422912 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.163899 seconds. If this number is high, consider checkpointing less frequently. I0424 16:12:44.415242 134109413422912 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 16:12:44.419339 134109413422912 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/5. I0424 16:12:44.964170 134109413422912 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:12:44.964269 134109413422912 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 16:12:45.005051 134109413422912 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042130s I0424 16:12:45.005223 134109413422912 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.066 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3044743537902832 s) (per-host) I0424 16:12:45.005284 134109413422912 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304540s (batch_requests_ready=0.260509s, total_serialization_initiated=0.043958s, others=0.000073s) I0424 16:12:45.005401 134109413422912 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.309503s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.309488) I0424 16:12:45.006284 133979738576640 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:12:45.006452 134109413422912 async_checkpointer.py:561] Finished blocking save. Time taken: 0.591140s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/5. I0424 16:12:45.030414 134109413422912 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:12:45.030686 133982930835200 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:12:45.030821 134109413422912 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047121.2492905, 'wait_for_prev_duration_secs': 43.163898944854736, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047164.41528, 'checkpointer_blocking_duration_secs': 0.5912823677062988, 'get_old_steps_start_time': 1777047165.0065813, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1777047121.245749, 'checkpoint_manager_blocking_duration_secs': 43.785035848617554} I0424 16:12:45.030935 134109413422912 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 16:12:45.031548 134109413422912 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.155, Tokens/s/device: 13438.408, total_weights: 65536, loss: 9.460 I0424 16:12:45.037275 134109413422912 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.184, Tokens/s/device: 13895.014, total_weights: 65536, loss: 8.959 I0424 16:13:04.894353 134109413422912 metric_logger.py:185] completed step: 6, seconds: 44.081, TFLOP/s/device: 0.308, Tokens/s/device: 46.460, total_weights: 65536, loss: 8.469 I0424 16:13:05.041858 134109413422912 metric_logger.py:185] completed step: 7, seconds: 19.711, TFLOP/s/device: 0.689, Tokens/s/device: 103.902, total_weights: 65536, loss: 8.003 I0424 16:13:05.189287 134109413422912 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.273, Tokens/s/device: 13456.244, total_weights: 65536, loss: 7.572 I0424 16:13:05.336343 134109413422912 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 16:13:05.337010 134109413422912 checkpointing.py:798] Waited 0.0006873607635498047 seconds for step 9 to finish before starting checkpointing. I0424 16:13:05.339065 134109413422912 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:13:09.329333 133984042051328 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/5/items/array_metadatas/process_6 I0424 16:13:46.744781 133979738576640 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.459 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 62.0439817905426 s) (per-host) I0424 16:13:46.744903 133979738576640 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.738500s. I0424 16:13:54.492637 133979738576640 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.486219s. I0424 16:13:54.492934 133982930835200 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:13:54.493067 133982930835200 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:13:54.493134 133982930835200 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:13:54.494532 133982930835200 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 16:13:54.494704 134109413422912 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0424 16:13:54.494834 134109413422912 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.155778 seconds. If this number is high, consider checkpointing less frequently. I0424 16:13:54.496328 134109413422912 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 16:13:54.499525 134109413422912 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 16:13:55.083720 134109413422912 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:13:55.083817 134109413422912 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0424 16:13:55.129151 134109413422912 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046487s I0424 16:13:55.129320 134109413422912 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.859 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3174619674682617 s) (per-host) I0424 16:13:55.129368 134109413422912 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.317519s (batch_requests_ready=0.268930s, total_serialization_initiated=0.048527s, others=0.000062s) I0424 16:13:55.129458 134109413422912 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.321649s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.321637) I0424 16:13:55.130365 133984037025536 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:13:55.130524 134109413422912 async_checkpointer.py:561] Finished blocking save. Time taken: 0.634126s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 16:13:55.132444 134109413422912 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:13:55.132654 133982930835200 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:13:55.132809 134109413422912 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047185.339039, 'wait_for_prev_duration_secs': 49.15577840805054, 'time_between_consecutive_saves_sec': 20.92612886428833, 'checkpointer_blocking_start_time': 1777047234.4963672, 'checkpointer_blocking_duration_secs': 0.634270429611206, 'get_old_steps_start_time': 1777047235.1306517, 'get_old_steps_duration_secs': 1.8596649169921875e-05, 'checkpoint_manager_blocking_start_time': 1777047185.337259, 'checkpoint_manager_blocking_duration_secs': 49.795520305633545} I0424 16:13:55.132912 134109413422912 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 16:13:55.132954 134109413422912 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:13:59.931396 133984045418240 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9/items/array_metadatas/process_6 I0424 16:14:37.251847 133984037025536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.220 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.43994688987732 s) (per-host) I0424 16:14:37.251972 133984037025536 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.121494s. I0424 16:14:44.284887 133984037025536 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.154394s. I0424 16:14:44.285131 133982930835200 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:14:44.285199 133982930835200 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:14:44.285255 133982930835200 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:14:44.286938 133982930835200 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 16:14:44.287122 134109413422912 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0424 16:14:44.287269 134109413422912 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:14:44.287962 134109413422912 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.955, Tokens/s/device: 13860.502, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 16:14:57 UTC 2026 EXIT_CODE=0