XPK Start: Thu Apr 23 16:55:06 UTC 2026 2026-04-23 16:55:10.895254: 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 00:00:1776963310.908443 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776963310.912208 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776963310.923898 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963310.923918 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963310.923921 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963310.923923 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 16:55:30.071198: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0423 16:55:30.601953 132106262218560 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 16:55:39,643:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 16:55:39.643549 132106262218560 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 16:55:39,646:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-jpg21-slice-job-0-0.mt-15-ocdbt-true-sv-jpg21:8482 I0423 16:55:39.646010 132106262218560 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-jpg21-slice-job-0-0.mt-15-ocdbt-true-sv-jpg21:8482 I0423 16:55:40.806126 132106262218560 max_utils.py:284] Jax distributed system initialized! I0423 16:55:47.579943 132106262218560 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 16:55:47.580050 132106262218560 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 16:55:47.580091 132106262218560 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0423 16:55:47.580145 132106262218560 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 16:55:48.199558 132106262218560 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 16:55:48.200155 132106262218560 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 16:55:48.200344 132106262218560 checkpointing.py:688] Setting up checkpoint logger... I0423 16:55:48.200396 132106262218560 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 16:55:48.200441 132106262218560 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 16:55:48.200803 132106262218560 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 0x7825e0f752e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 16:55:51.056471 132106262218560 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 16:55:51.056829 132106262218560 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 0x7812dbb11760>}, handler_registry=None I0423 16:55:51.057070 132106262218560 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7812dbb11760>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0423 16:55:51.057134 132106262218560 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7812dbb16de0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0423 16:55:51.057172 132106262218560 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 0x7812dbb11760>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7812dbb11760>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7812dbb16de0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7812dbb16de0>}). I0423 16:55:51.057702 132106262218560 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 16:55:51.057782 132106262218560 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7810b4698180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 16:55:52.183834 132106262218560 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints I0423 16:55:52.186079 132106262218560 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7812dbb13b60> I0423 16:55:52.186207 132106262218560 checkpointing.py:302] Checkpoint manager created! I0423 16:55:53.079542 132106262218560 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 16:55:53.079648 132106262218560 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:55:53.450599 132106262218560 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:55:53.450696 132106262218560 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:55:53.467050 132106262218560 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:55:53.467127 132106262218560 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:55:53.490182 132106262218560 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:55:53.490255 132106262218560 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:55:53.506339 132106262218560 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:55:53.506403 132106262218560 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:55:53.522362 132106262218560 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:55:53.522422 132106262218560 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:55:53.547629 132106262218560 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 16:55:53.547706 132106262218560 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:55:53.567516 132106262218560 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 16:55:53.567585 132106262218560 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 16:55:53.757444 132106262218560 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 16:55:53.757561 132106262218560 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 16:55:55.246114 132106262218560 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 16:55:55.246248 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0423 16:55:55.246302 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0423 16:55:55.246360 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0423 16:55:55.246412 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0423 16:55:55.246450 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0423 16:55:55.246508 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0423 16:55:55.246564 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0423 16:55:55.246605 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0423 16:55:55.246644 132106262218560 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0423 16:55:55.246693 132106262218560 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0423 16:55:55.246741 132106262218560 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0423 16:55:55.748468 132106262218560 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:55:55.748572 132106262218560 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0423 16:55:55.763542 132106262218560 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:55:55.763604 132106262218560 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0423 16:56:06.636886 132106262218560 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0423 16:56:06.637887 132106262218560 metric_logger.py:289] number parameters: 1.104 billion I0423 16:56:17.568171 132106262218560 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 16:56:17.735944 132106262218560 checkpointing.py:798] Waited 0.1677534580230713 seconds for step 0 to finish before starting checkpointing. I0423 16:56:17.738427 132106262218560 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:56:17.740203 132106262218560 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0423 16:56:17.742993 132106262218560 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/0. I0423 16:56:18.280663 132106262218560 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 16:56:18.281734 132106262218560 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:56:18.281794 132106262218560 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0423 16:56:18.530927 132106262218560 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250342s I0423 16:56:18.531114 132106262218560 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.058 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5044143199920654 s) (per-host) I0423 16:56:18.531169 132106262218560 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.504492s (batch_requests_ready=0.250485s, total_serialization_initiated=0.253921s, others=0.000086s) I0423 16:56:18.531263 132106262218560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.508527s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.508507) I0423 16:56:18.532247 131980503013120 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 16:56:18.532427 132106262218560 async_checkpointer.py:561] Finished blocking save. Time taken: 0.792151s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/0. I0423 16:56:18.550518 132106262218560 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:56:18.550812 131979395700480 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:56:18.550972 132106262218560 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776963377.7384098, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776963377.7402418, 'checkpointer_blocking_duration_secs': 0.7922959327697754, 'get_old_steps_start_time': 1776963378.5325584, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776963377.736541, 'checkpoint_manager_blocking_duration_secs': 0.8143913745880127} I0423 16:56:18.551079 132106262218560 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 16:56:18.551146 132106262218560 max_utils.py:750] Memstats: After params initialized: I0423 16:56:18.551200 132106262218560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 16:56:18.551240 132106262218560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 16:56:18.551269 132106262218560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 16:56:18.551295 132106262218560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 16:56:18.822860 132106262218560 metric_logger.py:185] completed step: 0, seconds: 10.930, TFLOP/s/device: 1.243, Tokens/s/device: 187.371, total_weights: 65536, loss: 10.874 I0423 16:56:18.991693 132106262218560 metric_logger.py:185] completed step: 1, seconds: 1.253, TFLOP/s/device: 10.840, Tokens/s/device: 1633.992, total_weights: 65536, loss: 10.874 I0423 16:56:19.438644 132106262218560 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 621.041, Tokens/s/device: 93610.019, total_weights: 65536, loss: 10.560 I0423 16:56:19.586079 132106262218560 metric_logger.py:185] completed step: 3, seconds: 0.447, TFLOP/s/device: 30.382, Tokens/s/device: 4579.463, total_weights: 65536, loss: 9.980 I0423 16:56:19.590238 132106262218560 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0423 16:56:19.880567 132106262218560 checkpointing.py:798] Waited 0.290302038192749 seconds for step 5 to finish before starting checkpointing. I0423 16:56:19.883292 132106262218560 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:56:21.973008 2628 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 16:56:23.909826 131980511405824 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0423 16:56:55.940028 131980503013120 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 41.663 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.91330075263977 s) (per-host) I0423 16:56:55.940162 131980503013120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.407795s. I0423 16:57:03.529042 131980503013120 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.996660s. I0423 16:57:03.529384 131979395700480 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:57:03.529504 131979395700480 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:57:03.529557 131979395700480 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:57:03.531038 131979395700480 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 16:57:03.531231 132106262218560 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. W0423 16:57:03.531368 132106262218560 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.648077 seconds. If this number is high, consider checkpointing less frequently. I0423 16:57:03.533325 132106262218560 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0423 16:57:03.537299 132106262218560 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/5. I0423 16:57:04.099879 132106262218560 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:57:04.099982 132106262218560 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0423 16:57:04.152532 132106262218560 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.053805s I0423 16:57:04.152692 132106262218560 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.982 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30965423583984375 s) (per-host) I0423 16:57:04.152742 132106262218560 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.309714s (batch_requests_ready=0.254006s, total_serialization_initiated=0.055640s, others=0.000067s) I0423 16:57:04.152820 132106262218560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314605s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.314590) I0423 16:57:04.153718 131979941840640 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 16:57:04.153871 132106262218560 async_checkpointer.py:561] Finished blocking save. Time taken: 0.620475s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/5. I0423 16:57:04.158677 132106262218560 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:57:04.158907 131979395700480 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:57:04.159033 132106262218560 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776963379.883265, 'wait_for_prev_duration_secs': 43.64807748794556, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776963423.533364, 'checkpointer_blocking_duration_secs': 0.6206154823303223, 'get_old_steps_start_time': 1776963424.153999, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776963379.8811643, 'checkpoint_manager_blocking_duration_secs': 44.27783489227295} I0423 16:57:04.159159 132106262218560 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0423 16:57:04.159746 132106262218560 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.204, Tokens/s/device: 13445.731, total_weights: 65536, loss: 9.460 I0423 16:57:04.165563 132106262218560 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.323, Tokens/s/device: 13915.974, total_weights: 65536, loss: 8.959 I0423 16:57:23.317352 132106262218560 metric_logger.py:185] completed step: 6, seconds: 44.574, TFLOP/s/device: 0.305, Tokens/s/device: 45.946, total_weights: 65536, loss: 8.469 I0423 16:57:23.464903 132106262218560 metric_logger.py:185] completed step: 7, seconds: 19.006, TFLOP/s/device: 0.715, Tokens/s/device: 107.757, total_weights: 65536, loss: 8.003 I0423 16:57:23.612125 132106262218560 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.387, Tokens/s/device: 13473.418, total_weights: 65536, loss: 7.572 I0423 16:57:23.759291 132106262218560 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 16:57:23.759968 132106262218560 checkpointing.py:798] Waited 0.0006985664367675781 seconds for step 9 to finish before starting checkpointing. I0423 16:57:23.761877 132106262218560 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:57:27.773078 131980511405824 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0423 16:58:04.960060 131979941840640 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.845 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.11698365211487 s) (per-host) I0423 16:58:04.960186 131979941840640 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.806362s. I0423 16:58:12.162522 131979941840640 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.008683s. I0423 16:58:12.162798 131979395700480 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:58:12.162939 131979395700480 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:58:12.162991 131979395700480 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:58:12.164580 131979395700480 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0423 16:58:12.164753 132106262218560 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. W0423 16:58:12.164894 132106262218560 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.403029 seconds. If this number is high, consider checkpointing less frequently. I0423 16:58:12.166379 132106262218560 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0423 16:58:12.169834 132106262218560 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/9. I0423 16:58:13.155897 132106262218560 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:58:13.155992 132106262218560 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0423 16:58:13.202946 132106262218560 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048035s I0423 16:58:13.203121 132106262218560 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.963 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3108174800872803 s) (per-host) I0423 16:58:13.203171 132106262218560 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.310893s (batch_requests_ready=0.260774s, total_serialization_initiated=0.050038s, others=0.000081s) I0423 16:58:13.203266 132106262218560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.314975s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.314964) I0423 16:58:13.204115 131980503013120 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 16:58:13.204271 132106262218560 async_checkpointer.py:561] Finished blocking save. Time taken: 1.037822s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/9. I0423 16:58:13.634745 132106262218560 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:58:13.635147 131979395700480 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:58:13.635319 132106262218560 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776963443.7618487, 'wait_for_prev_duration_secs': 48.40302896499634, 'time_between_consecutive_saves_sec': 20.230770111083984, 'checkpointer_blocking_start_time': 1776963492.166418, 'checkpointer_blocking_duration_secs': 1.0379629135131836, 'get_old_steps_start_time': 1776963493.2043953, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1776963443.7602131, 'checkpoint_manager_blocking_duration_secs': 49.87507247924805} I0423 16:58:13.635437 132106262218560 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 16:58:13.635476 132106262218560 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:58:18.576371 131980511405824 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0423 16:58:54.820940 131980503013120 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.673 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.92861294746399 s) (per-host) I0423 16:58:54.821052 131980503013120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.616828s. I0423 16:59:03.704046 131980503013120 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.499807s. I0423 16:59:03.704334 131979395700480 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:59:03.704458 131979395700480 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:59:03.704500 131979395700480 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:59:03.706002 131979395700480 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 16:59:03.706186 132106262218560 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. I0423 16:59:03.706332 132106262218560 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:59:03.707035 132106262218560 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.115, Tokens/s/device: 13884.557, 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: Thu Apr 23 16:59:12 UTC 2026 EXIT_CODE=0