XPK Start: Thu Apr 23 17:00:15 UTC 2026 2026-04-23 17:00:19.324925: 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 00:00:1776963619.337550 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776963619.341274 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776963619.352895 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963619.352914 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963619.352916 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963619.352918 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 17:00:38.598714: 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 17:00:39.125649 132322610181952 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 17:00:48,166:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 17:00:48.166728 132322610181952 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 17:00:48,169:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-vfnae-slice-job-0-0.mt-15-ocdbt-true-rs-vfnae:8482 I0423 17:00:48.169087 132322610181952 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-vfnae-slice-job-0-0.mt-15-ocdbt-true-rs-vfnae:8482 I0423 17:00:49.070316 132322610181952 max_utils.py:284] Jax distributed system initialized! I0423 17:00:55.260560 132322610181952 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 17:00:55.260665 132322610181952 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 17:00:55.260707 132322610181952 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 17:00:55.260741 132322610181952 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 17:00:55.897132 132322610181952 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 17:00:55.897724 132322610181952 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 17:00:55.897906 132322610181952 checkpointing.py:688] Setting up checkpoint logger... I0423 17:00:55.897957 132322610181952 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 17:00:55.897999 132322610181952 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 17:00:55.898387 132322610181952 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 0x785845d752b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 17:00:57.289085 132322610181952 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 17:00:57.289469 132322610181952 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 0x78446c585580>}, handler_registry=None I0423 17:00:57.289720 132322610181952 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78446c585580>` 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 17:00:57.289770 132322610181952 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78446c58a3c0>` 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 17:00:57.289807 132322610181952 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 0x78446c585580>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78446c585580>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78446c58a3c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78446c58a3c0>}). I0423 17:00:57.290350 132322610181952 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 17:00:57.290431 132322610181952 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78446c354180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 17:00:58.937276 132205787469568 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776963380201528208, 'commit_timestamp_nsecs': 1776963422331197555, 'custom_metadata': {}} from 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/_CHECKPOINT_METADATA I0423 17:00:59.190111 132205787469568 checkpoint_manager.py:1877] Missing metrics for step 0 I0423 17:00:59.191939 132322610181952 checkpoint_manager.py:1818] Found 3 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 17:00:59.681455 132322610181952 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 0x78446c585d00> I0423 17:00:59.681639 132322610181952 checkpointing.py:302] Checkpoint manager created! I0423 17:01:00.591630 132322610181952 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 17:01:00.591754 132322610181952 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:01:00.964872 132322610181952 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 17:01:00.964969 132322610181952 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:01:00.986071 132322610181952 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 17:01:00.986199 132322610181952 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:01:01.009422 132322610181952 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 17:01:01.009493 132322610181952 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:01:01.025686 132322610181952 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 17:01:01.025748 132322610181952 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:01:01.041887 132322610181952 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 17:01:01.041949 132322610181952 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:01:01.066814 132322610181952 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 17:01:01.066885 132322610181952 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:01:01.086740 132322610181952 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 17:01:01.086808 132322610181952 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 17:01:01.290215 132322610181952 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 17:01:01.290331 132322610181952 checkpointing.py:582] restoring from this run's directory step 9 I0423 17:01:01.807063 132322610181952 checkpointer.py:304] Restoring checkpoint from 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 17:01:03.251241 1390 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 17:01:09.706901 132322610181952 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 244.589 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.460941314697266 s) (per-host) I0423 17:01:09.718694 132322610181952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 229.850 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.885341167449951 s) (per-host) I0423 17:01:10.238655 132322610181952 checkpointer.py:318] Finished restoring checkpoint in 8.69 seconds from 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 17:01:10.238816 132322610181952 standard_logger.py:34] {'step': 9, 'event_type': 'restore', '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', 'checkpointer_start_time': 1776963661.5502994, 'checkpointer_duration_secs': 8.688484191894531, 'checkpoint_manager_start_time': 1776963661.2905831, 'checkpoint_manager_duration_secs': 8.948202133178711} fsdp: 32 I0423 17:01:10.648759 132322610181952 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 17:01:10.648886 132322610181952 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 17:01:10.648940 132322610181952 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 17:01:10.649003 132322610181952 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 17:01:10.649056 132322610181952 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 17:01:10.649125 132322610181952 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 17:01:10.649200 132322610181952 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 17:01:10.649262 132322610181952 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 17:01:10.649305 132322610181952 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 17:01:10.649343 132322610181952 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 17:01:10.649390 132322610181952 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0423 17:01:10.649436 132322610181952 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0423 17:01:11.149224 132322610181952 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 17:01:11.149319 132322610181952 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0423 17:01:11.164305 132322610181952 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 17:01:11.164365 132322610181952 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0423 17:01:22.156044 132322610181952 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 17:01:22.156979 132322610181952 metric_logger.py:289] number parameters: 1.104 billion I0423 17:01:33.011637 132322610181952 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0423 17:01:33.905770 132322610181952 checkpointing.py:798] Waited 0.8941135406494141 seconds for step 10 to finish before starting checkpointing. I0423 17:01:33.908286 132322610181952 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 17:01:33.910345 132322610181952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0423 17:01:33.912959 132322610181952 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/10. I0423 17:01:34.510209 132322610181952 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 17:01:34.511218 132322610181952 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:01:34.511276 132322610181952 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 17:01:34.756170 132322610181952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.246061s I0423 17:01:34.756344 132322610181952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.071 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.502230167388916 s) (per-host) I0423 17:01:34.756398 132322610181952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.502296s (batch_requests_ready=0.254104s, total_serialization_initiated=0.248119s, others=0.000073s) I0423 17:01:34.756488 132322610181952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.506599s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.506580) I0423 17:01:34.757359 132202465564416 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:01:34.757528 132322610181952 async_checkpointer.py:561] Finished blocking save. Time taken: 0.847112s. 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/10. I0423 17:01:34.795212 132322610181952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:01:34.795566 132200871737088 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:01:34.795726 132322610181952 standard_logger.py:34] {'step': 10, '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': 1776963693.9082696, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776963693.910382, 'checkpointer_blocking_duration_secs': 0.8472576141357422, 'get_old_steps_start_time': 1776963694.7576578, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776963693.90633, 'checkpoint_manager_blocking_duration_secs': 0.8893554210662842} I0423 17:01:34.795851 132322610181952 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0423 17:01:34.795904 132322610181952 max_utils.py:750] Memstats: After params initialized: I0423 17:01:34.795958 132322610181952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 17:01:34.795991 132322610181952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 17:01:34.796018 132322610181952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 17:01:34.796042 132322610181952 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 17:01:35.063983 132322610181952 metric_logger.py:185] completed step: 10, seconds: 10.855, TFLOP/s/device: 1.252, Tokens/s/device: 188.677, total_weights: 65536, loss: 6.833 I0423 17:01:35.231087 132322610181952 metric_logger.py:185] completed step: 11, seconds: 2.051, TFLOP/s/device: 6.624, Tokens/s/device: 998.506, total_weights: 65536, loss: 6.528 I0423 17:01:35.378476 132322610181952 metric_logger.py:185] completed step: 12, seconds: 0.020, TFLOP/s/device: 681.332, Tokens/s/device: 102697.824, total_weights: 65536, loss: 6.265 I0423 17:01:35.525874 132322610181952 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.845, Tokens/s/device: 13391.660, total_weights: 65536, loss: 6.045 I0423 17:01:35.530396 132322610181952 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0423 17:01:35.820455 132322610181952 checkpointing.py:798] Waited 0.2900221347808838 seconds for step 15 to finish before starting checkpointing. I0423 17:01:35.823307 132322610181952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:01:39.906979 132205191882496 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/10/items/array_metadatas/process_6 I0423 17:02:17.093931 132202465564416 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.872 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.839781284332275 s) (per-host) I0423 17:02:17.094055 132202465564416 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.336583s. I0423 17:02:25.308832 132202465564416 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.551343s. I0423 17:02:25.309163 132200871737088 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:02:25.309293 132200871737088 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:02:25.309342 132200871737088 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:02:25.314316 132200871737088 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0423 17:02:25.314489 132322610181952 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. W0423 17:02:25.314628 132322610181952 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.491323 seconds. If this number is high, consider checkpointing less frequently. I0423 17:02:25.316687 132322610181952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0423 17:02:25.321154 132322610181952 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/15. I0423 17:02:25.883581 132322610181952 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:02:25.883680 132322610181952 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 17:02:25.924170 132322610181952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041775s I0423 17:02:25.924359 132322610181952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.144 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2998676300048828 s) (per-host) I0423 17:02:25.924423 132322610181952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299938s (batch_requests_ready=0.256205s, total_serialization_initiated=0.043653s, others=0.000081s) I0423 17:02:25.924540 132322610181952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.304834s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.304821) I0423 17:02:25.925449 132198749431552 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:02:25.925577 132322610181952 async_checkpointer.py:561] Finished blocking save. Time taken: 0.608820s. 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/15. I0423 17:02:25.942247 132322610181952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:02:25.942544 132200871737088 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:02:25.942704 132322610181952 standard_logger.py:34] {'step': 15, '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': 1776963695.8232803, 'wait_for_prev_duration_secs': 49.491323471069336, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776963745.3167279, 'checkpointer_blocking_duration_secs': 0.6090176105499268, 'get_old_steps_start_time': 1776963745.9257662, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776963695.821127, 'checkpoint_manager_blocking_duration_secs': 50.1215443611145} I0423 17:02:25.942810 132322610181952 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0423 17:02:25.943385 132322610181952 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.150, Tokens/s/device: 13889.925, total_weights: 65536, loss: 5.864 I0423 17:02:25.949169 132322610181952 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.261, Tokens/s/device: 13906.525, total_weights: 65536, loss: 5.719 I0423 17:02:43.743475 132322610181952 metric_logger.py:185] completed step: 16, seconds: 50.418, TFLOP/s/device: 0.269, Tokens/s/device: 40.621, total_weights: 65536, loss: 5.605 I0423 17:02:43.891171 132322610181952 metric_logger.py:185] completed step: 17, seconds: 17.648, TFLOP/s/device: 0.770, Tokens/s/device: 116.045, total_weights: 65536, loss: 5.517 I0423 17:02:44.038389 132322610181952 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.391, Tokens/s/device: 13473.950, total_weights: 65536, loss: 5.448 I0423 17:02:44.185475 132322610181952 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0423 17:02:44.186165 132322610181952 checkpointing.py:798] Waited 0.0007116794586181641 seconds for step 19 to finish before starting checkpointing. I0423 17:02:44.188180 132322610181952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:02:48.214770 132205191882496 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/15/items/array_metadatas/process_6 I0423 17:03:25.029883 132198749431552 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.590 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.40536904335022 s) (per-host) I0423 17:03:25.029985 132198749431552 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.104331s. I0423 17:03:33.035141 132198749431552 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 67.109470s. I0423 17:03:33.035431 132200871737088 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:03:33.035568 132200871737088 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:03:33.035622 132200871737088 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:03:33.037182 132200871737088 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0423 17:03:33.037371 132322610181952 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. W0423 17:03:33.037535 132322610181952 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.849365 seconds. If this number is high, consider checkpointing less frequently. I0423 17:03:33.038992 132322610181952 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0423 17:03:33.042382 132322610181952 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/19. I0423 17:03:34.045324 132322610181952 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:03:34.045417 132322610181952 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 17:03:34.088585 132322610181952 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044214s I0423 17:03:34.088752 132322610181952 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.024 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30706143379211426 s) (per-host) I0423 17:03:34.088800 132322610181952 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307121s (batch_requests_ready=0.260799s, total_serialization_initiated=0.046257s, others=0.000064s) I0423 17:03:34.088886 132322610181952 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311129s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.311118) I0423 17:03:34.089809 132205183489792 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:03:34.089966 132322610181952 async_checkpointer.py:561] Finished blocking save. Time taken: 1.050904s. 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/19. I0423 17:03:34.101608 132322610181952 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:03:34.101869 132200871737088 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:03:34.102022 132322610181952 standard_logger.py:34] {'step': 19, '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': 1776963764.1881542, 'wait_for_prev_duration_secs': 48.84936451911926, 'time_between_consecutive_saves_sec': 18.87378692626953, 'checkpointer_blocking_start_time': 1776963813.0390317, 'checkpointer_blocking_duration_secs': 1.0510425567626953, 'get_old_steps_start_time': 1776963814.0900884, 'get_old_steps_duration_secs': 3.3855438232421875e-05, 'checkpoint_manager_blocking_start_time': 1776963764.1864016, 'checkpoint_manager_blocking_duration_secs': 49.91559076309204} I0423 17:03:34.102132 132322610181952 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0423 17:03:34.102175 132322610181952 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:03:38.921475 132205191882496 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/19/items/array_metadatas/process_6 I0423 17:04:15.850653 132205183489792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.548 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.068922996520996 s) (per-host) I0423 17:04:15.850773 132205183489792 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.760853s. I0423 17:04:24.309853 132205183489792 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.219919s. I0423 17:04:24.310121 132200871737088 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:04:24.310178 132200871737088 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:04:24.310216 132200871737088 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:04:24.312441 132200871737088 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0423 17:04:24.312608 132322610181952 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. I0423 17:04:24.312753 132322610181952 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 17:04:24.313329 132322610181952 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.104, Tokens/s/device: 13882.957, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 17:04:35 UTC 2026 EXIT_CODE=0