XPK Start: Sun Apr 19 04:23:38 UTC 2026 2026-04-19 04:23:42.372134: 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 00:00:1776572622.384755 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776572622.388456 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776572622.399647 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776572622.399666 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776572622.399668 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776572622.399670 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 04:24:02.224514: 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) I0419 04:24:02.755206 136702025779008 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 04:24:11,796:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 04:24:11.796297 136702025779008 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 04:24:11,798:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-gw7fd-slice-job-0-0.mt-15-ocdbt-true-rs-gw7fd:8482 I0419 04:24:11.798675 136702025779008 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-gw7fd-slice-job-0-0.mt-15-ocdbt-true-rs-gw7fd:8482 I0419 04:24:12.638802 136702025779008 max_utils.py:284] Jax distributed system initialized! I0419 04:24:19.580165 136702025779008 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 04:24:19.580291 136702025779008 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 04:24:19.580337 136702025779008 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 I0419 04:24:19.580376 136702025779008 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 04:24:20.199879 136702025779008 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:24:20.200512 136702025779008 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:24:20.200697 136702025779008 checkpointing.py:688] Setting up checkpoint logger... I0419 04:24:20.200747 136702025779008 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 04:24:20.200791 136702025779008 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 04:24:20.201134 136702025779008 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 0x7c53dd575160>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 04:24:21.854838 136702025779008 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 04:24:21.855078 136702025779008 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 0x7c417d16dd00>}, handler_registry=None I0419 04:24:21.855340 136702025779008 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c417d16dd00>` 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`. I0419 04:24:21.855396 136702025779008 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c417d176c60>` 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`. I0419 04:24:21.855434 136702025779008 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 0x7c417d16dd00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c417d16dd00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c417d176c60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c417d176c60>}). I0419 04:24:21.855937 136702025779008 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 04:24:21.856018 136702025779008 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c4026dcc180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 04:24:25.072187 136581419620096 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776572492728717919, 'commit_timestamp_nsecs': 1776572540884904347, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/9/_CHECKPOINT_METADATA I0419 04:24:25.315388 136581419620096 checkpoint_manager.py:1877] Missing metrics for step 9 I0419 04:24:26.631343 136702025779008 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints I0419 04:24:26.633707 136702025779008 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c417d16e390> I0419 04:24:26.633821 136702025779008 checkpointing.py:302] Checkpoint manager created! I0419 04:24:27.529006 136702025779008 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 04:24:27.529113 136702025779008 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:24:27.900071 136702025779008 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:24:27.900166 136702025779008 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:24:27.916134 136702025779008 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:24:27.916195 136702025779008 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:24:27.939128 136702025779008 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:24:27.939196 136702025779008 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:24:27.961304 136702025779008 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:24:27.961409 136702025779008 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:24:27.977572 136702025779008 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:24:27.977635 136702025779008 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:24:28.002866 136702025779008 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 04:24:28.002938 136702025779008 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:24:28.022699 136702025779008 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 04:24:28.022767 136702025779008 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 04:24:28.213196 136702025779008 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 04:24:28.213330 136702025779008 checkpointing.py:582] restoring from this run's directory step 9 I0419 04:24:28.714907 136702025779008 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/9. I0419 04:24:33.656545 1418 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 04:24:40.191477 136702025779008 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 241.627 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.540148019790649 s) (per-host) I0419 04:24:40.205757 136702025779008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 202.782 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.804402589797974 s) (per-host) I0419 04:24:40.207971 136702025779008 checkpointer.py:318] Finished restoring checkpoint in 11.75 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/9. I0419 04:24:40.208123 136702025779008 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776572668.459927, 'checkpointer_duration_secs': 11.748164653778076, 'checkpoint_manager_start_time': 1776572668.2135873, 'checkpoint_manager_duration_secs': 11.9945068359375} fsdp: 32 I0419 04:24:40.617892 136702025779008 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0419 04:24:40.618034 136702025779008 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) I0419 04:24:40.618088 136702025779008 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) I0419 04:24:40.618151 136702025779008 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') I0419 04:24:40.618206 136702025779008 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) I0419 04:24:40.618267 136702025779008 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) I0419 04:24:40.618330 136702025779008 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) I0419 04:24:40.618390 136702025779008 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') I0419 04:24:40.618433 136702025779008 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) I0419 04:24:40.618473 136702025779008 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) I0419 04:24:40.618522 136702025779008 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0419 04:24:40.618573 136702025779008 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0419 04:24:41.115698 136702025779008 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:24:41.115797 136702025779008 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:24:41.130847 136702025779008 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:24:41.130908 136702025779008 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:24:51.886750 136702025779008 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. I0419 04:24:51.887646 136702025779008 metric_logger.py:289] number parameters: 1.104 billion I0419 04:25:03.107010 136702025779008 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 04:25:03.260674 136702025779008 checkpointing.py:798] Waited 0.15364313125610352 seconds for step 10 to finish before starting checkpointing. I0419 04:25:03.263464 136702025779008 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:25:03.265110 136702025779008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0419 04:25:03.267862 136702025779008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10. I0419 04:25:04.719085 136702025779008 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 04:25:04.720181 136702025779008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:25:04.720250 136702025779008 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 I0419 04:25:04.964966 136702025779008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245972s I0419 04:25:04.965150 136702025779008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.609 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9585683345794678 s) (per-host) I0419 04:25:04.965216 136702025779008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.958643s (batch_requests_ready=0.710586s, total_serialization_initiated=0.247970s, others=0.000087s) I0419 04:25:04.965359 136702025779008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.962934s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.962914) I0419 04:25:04.966381 136581394441984 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:25:04.966519 136702025779008 async_checkpointer.py:561] Finished blocking save. Time taken: 1.701330s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10. I0419 04:25:05.371909 136702025779008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:25:05.372304 136577032365824 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:25:05.372473 136702025779008 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776572703.263446, 'wait_for_prev_duration_secs': 6.413459777832031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776572703.26515, 'checkpointer_blocking_duration_secs': 1.701556921005249, 'get_old_steps_start_time': 1776572704.9667275, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776572703.2613084, 'checkpoint_manager_blocking_duration_secs': 2.111124038696289} I0419 04:25:05.372588 136702025779008 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 04:25:05.372642 136702025779008 max_utils.py:750] Memstats: After params initialized: I0419 04:25:05.372703 136702025779008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 04:25:05.372738 136702025779008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 04:25:05.372766 136702025779008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 04:25:05.372791 136702025779008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 04:25:05.631147 136702025779008 metric_logger.py:185] completed step: 10, seconds: 11.219, TFLOP/s/device: 1.211, Tokens/s/device: 182.543, total_weights: 65536, loss: 6.833 I0419 04:25:05.898689 136702025779008 metric_logger.py:185] completed step: 11, seconds: 2.523, TFLOP/s/device: 5.386, Tokens/s/device: 811.808, total_weights: 65536, loss: 6.528 I0419 04:25:06.046158 136702025779008 metric_logger.py:185] completed step: 12, seconds: 0.121, TFLOP/s/device: 112.578, Tokens/s/device: 16968.954, total_weights: 65536, loss: 6.265 I0419 04:25:06.194040 136702025779008 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.630, Tokens/s/device: 13359.252, total_weights: 65536, loss: 6.045 I0419 04:25:06.199286 136702025779008 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 04:25:06.488718 136702025779008 checkpointing.py:798] Waited 0.2893948554992676 seconds for step 15 to finish before starting checkpointing. I0419 04:25:06.491632 136702025779008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:25:09.845802 136577049151232 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0419 04:25:45.811325 136581394441984 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.785 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.804707765579224 s) (per-host) I0419 04:25:45.811430 136581394441984 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.844826s. I0419 04:25:53.683858 136581394441984 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.717237s. I0419 04:25:53.684152 136577032365824 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:25:53.684292 136577032365824 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:25:53.684351 136577032365824 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:25:53.685662 136577032365824 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 04:25:53.685849 136702025779008 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. W0419 04:25:53.685986 136702025779008 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.194359 seconds. If this number is high, consider checkpointing less frequently. I0419 04:25:53.687947 136702025779008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0419 04:25:53.692119 136702025779008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15. I0419 04:25:55.489258 136702025779008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:25:55.489358 136702025779008 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 I0419 04:25:55.534344 136702025779008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046268s I0419 04:25:55.534520 136702025779008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.133 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.723339319229126 s) (per-host) I0419 04:25:55.534572 136702025779008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.723403s (batch_requests_ready=0.675208s, total_serialization_initiated=0.048122s, others=0.000072s) I0419 04:25:55.534678 136702025779008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.728164s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.728152) I0419 04:25:55.535619 136575455323904 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:25:55.535748 136702025779008 async_checkpointer.py:561] Finished blocking save. Time taken: 1.847730s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15. I0419 04:25:55.543857 136702025779008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:25:55.544138 136577032365824 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:25:55.544311 136702025779008 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776572706.4916031, 'wait_for_prev_duration_secs': 47.194358587265015, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776572753.6879869, 'checkpointer_blocking_duration_secs': 1.8479366302490234, 'get_old_steps_start_time': 1776572755.535942, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776572706.4894094, 'checkpoint_manager_blocking_duration_secs': 49.05486798286438} I0419 04:25:55.544415 136702025779008 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 04:25:55.544924 136702025779008 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.078, Tokens/s/device: 13879.006, total_weights: 65536, loss: 5.864 I0419 04:25:55.550673 136702025779008 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.909, Tokens/s/device: 13853.564, total_weights: 65536, loss: 5.719 I0419 04:26:13.533333 136702025779008 metric_logger.py:185] completed step: 16, seconds: 49.350, TFLOP/s/device: 0.275, Tokens/s/device: 41.499, total_weights: 65536, loss: 5.605 I0419 04:26:13.681011 136702025779008 metric_logger.py:185] completed step: 17, seconds: 17.836, TFLOP/s/device: 0.762, Tokens/s/device: 114.822, total_weights: 65536, loss: 5.517 I0419 04:26:13.828216 136702025779008 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.116, Tokens/s/device: 13432.591, total_weights: 65536, loss: 5.448 I0419 04:26:13.975411 136702025779008 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 04:26:13.976081 136702025779008 checkpointing.py:798] Waited 0.0006909370422363281 seconds for step 19 to finish before starting checkpointing. I0419 04:26:13.978179 136702025779008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:26:20.606186 136577049151232 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0419 04:26:56.265743 136575455323904 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.704 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.45452332496643 s) (per-host) I0419 04:26:56.265868 136575455323904 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.730037s. I0419 04:27:05.903834 136575455323904 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.367989s. I0419 04:27:05.904121 136577032365824 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:27:05.904269 136577032365824 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:27:05.904329 136577032365824 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:27:05.906007 136577032365824 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 04:27:05.906184 136702025779008 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. W0419 04:27:05.906341 136702025779008 checkpoint_manager.py:1458] Waiting for previous save to complete took 51.928175 seconds. If this number is high, consider checkpointing less frequently. I0419 04:27:05.907891 136702025779008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0419 04:27:05.911225 136702025779008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19. I0419 04:27:06.853195 136702025779008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:27:06.853309 136702025779008 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 I0419 04:27:06.895803 136702025779008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043604s I0419 04:27:06.895971 136702025779008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.203 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29648423194885254 s) (per-host) I0419 04:27:06.896018 136702025779008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296542s (batch_requests_ready=0.250851s, total_serialization_initiated=0.045629s, others=0.000063s) I0419 04:27:06.896108 136702025779008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.300544s (all_items=0.000010s, per_item={'items': '0.00000954'}, temp_paths=0.300534) I0419 04:27:06.897227 136581386049280 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:27:06.897400 136702025779008 async_checkpointer.py:561] Finished blocking save. Time taken: 0.989440s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19. I0419 04:27:07.336463 136702025779008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:27:07.336891 136577032365824 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:27:07.337063 136702025779008 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776572773.9781492, 'wait_for_prev_duration_secs': 51.92817544937134, 'time_between_consecutive_saves_sec': 20.2924485206604, 'checkpointer_blocking_start_time': 1776572825.90793, 'checkpointer_blocking_duration_secs': 0.9895756244659424, 'get_old_steps_start_time': 1776572826.897519, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776572773.9763322, 'checkpoint_manager_blocking_duration_secs': 53.36069679260254} I0419 04:27:07.337178 136702025779008 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 04:27:07.337232 136702025779008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:27:13.247198 136581394441984 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0419 04:27:49.935945 136581386049280 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.450 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.336416721343994 s) (per-host) I0419 04:27:49.936074 136581386049280 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.038719s. I0419 04:27:58.074093 136581386049280 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.176724s. I0419 04:27:58.074344 136577032365824 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:27:58.074403 136577032365824 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:27:58.074443 136577032365824 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:27:58.076886 136577032365824 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 04:27:58.077060 136702025779008 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. I0419 04:27:58.077216 136702025779008 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:27:58.077892 136702025779008 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.049, Tokens/s/device: 13874.587, 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: Sun Apr 19 04:28:09 UTC 2026 EXIT_CODE=0