XPK Start: Wed Apr 22 16:53:26 UTC 2026 2026-04-22 16:53:30.436368: 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 00:00:1776876810.449687 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776876810.453436 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776876810.465007 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776876810.465026 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776876810.465029 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776876810.465030 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 16:53:49.735593: 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) I0422 16:53:50.261458 132219947800384 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 16:53:59,303:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 16:53:59.303387 132219947800384 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 16:53:59,306:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-za2ul-slice-job-0-0.mt-15-ocdbt-true-sv-za2ul:8482 I0422 16:53:59.306050 132219947800384 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-za2ul-slice-job-0-0.mt-15-ocdbt-true-sv-za2ul:8482 I0422 16:54:00.671415 132219947800384 max_utils.py:284] Jax distributed system initialized! I0422 16:54:06.576746 132219947800384 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 16:54:06.576856 132219947800384 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 16:54:06.576897 132219947800384 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 I0422 16:54:06.576933 132219947800384 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 16:54:07.196955 132219947800384 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 16:54:07.197560 132219947800384 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 16:54:07.197747 132219947800384 checkpointing.py:688] Setting up checkpoint logger... I0422 16:54:07.197797 132219947800384 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 16:54:07.197839 132219947800384 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 16:54:07.198218 132219947800384 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 0x78404fb790a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 16:54:10.149447 132219947800384 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 16:54:10.149689 132219947800384 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 0x782c3c625a60>}, handler_registry=None I0422 16:54:10.149924 132219947800384 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x782c3c625a60>` 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`. I0422 16:54:10.149972 132219947800384 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782c3c78b380>` 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`. I0422 16:54:10.150010 132219947800384 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 0x782c3c625a60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x782c3c625a60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782c3c78b380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x782c3c78b380>}). I0422 16:54:10.150545 132219947800384 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 16:54:10.150623 132219947800384 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x782c3c2f44a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 16:54:11.337678 132219947800384 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints I0422 16:54:11.339791 132219947800384 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x782c3c367e30> I0422 16:54:11.339900 132219947800384 checkpointing.py:302] Checkpoint manager created! I0422 16:54:12.244284 132219947800384 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 16:54:12.244392 132219947800384 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 16:54:12.619075 132219947800384 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 16:54:12.619185 132219947800384 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 16:54:12.634943 132219947800384 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 16:54:12.635001 132219947800384 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 16:54:12.658606 132219947800384 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 16:54:12.658677 132219947800384 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 16:54:12.674673 132219947800384 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 16:54:12.674735 132219947800384 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 16:54:12.690588 132219947800384 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 16:54:12.690649 132219947800384 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 16:54:12.714604 132219947800384 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 16:54:12.714673 132219947800384 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 16:54:12.734329 132219947800384 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 16:54:12.734395 132219947800384 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 16:54:12.924685 132219947800384 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 16:54:12.924797 132219947800384 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 16:54:14.372234 132219947800384 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 16:54:14.372366 132219947800384 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) I0422 16:54:14.372420 132219947800384 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) I0422 16:54:14.372479 132219947800384 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') I0422 16:54:14.372532 132219947800384 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) I0422 16:54:14.372571 132219947800384 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) I0422 16:54:14.372629 132219947800384 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) I0422 16:54:14.372681 132219947800384 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') I0422 16:54:14.372721 132219947800384 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) I0422 16:54:14.372757 132219947800384 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) I0422 16:54:14.372804 132219947800384 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0422 16:54:14.372851 132219947800384 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0422 16:54:14.879474 132219947800384 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 16:54:14.879576 132219947800384 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 16:54:14.894873 132219947800384 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 16:54:14.894933 132219947800384 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 16:54:25.717862 132219947800384 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. I0422 16:54:25.718720 132219947800384 metric_logger.py:289] number parameters: 1.104 billion I0422 16:54:37.066287 132219947800384 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 16:54:37.219089 132219947800384 checkpointing.py:798] Waited 0.1527996063232422 seconds for step 0 to finish before starting checkpointing. I0422 16:54:37.221620 132219947800384 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 16:54:37.223359 132219947800384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 16:54:37.226081 132219947800384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0. I0422 16:54:37.780179 132219947800384 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 16:54:37.781191 132219947800384 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 16:54:37.781246 132219947800384 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 I0422 16:54:38.032054 132219947800384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251969s I0422 16:54:38.032241 132219947800384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.923 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5276789665222168 s) (per-host) I0422 16:54:38.032298 132219947800384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.527741s (batch_requests_ready=0.272130s, total_serialization_initiated=0.255541s, others=0.000070s) I0422 16:54:38.032389 132219947800384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.531936s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.531917) I0422 16:54:38.033298 132095703656192 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 16:54:38.033466 132219947800384 async_checkpointer.py:561] Finished blocking save. Time taken: 0.810039s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0. I0422 16:54:38.083729 132219947800384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 16:54:38.084082 132094630012672 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 16:54:38.084253 132219947800384 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776876877.2216017, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776876877.2233999, 'checkpointer_blocking_duration_secs': 0.8101704120635986, 'get_old_steps_start_time': 1776876878.0335972, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776876877.2196112, 'checkpoint_manager_blocking_duration_secs': 0.8646042346954346} I0422 16:54:38.084359 132219947800384 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 16:54:38.084413 132219947800384 max_utils.py:750] Memstats: After params initialized: I0422 16:54:38.084466 132219947800384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 16:54:38.084501 132219947800384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 16:54:38.084528 132219947800384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 16:54:38.084553 132219947800384 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 16:54:38.351743 132219947800384 metric_logger.py:185] completed step: 0, seconds: 11.347, TFLOP/s/device: 1.197, Tokens/s/device: 180.481, total_weights: 65536, loss: 10.874 I0422 16:54:38.519894 132219947800384 metric_logger.py:185] completed step: 1, seconds: 1.284, TFLOP/s/device: 10.581, Tokens/s/device: 1594.876, total_weights: 65536, loss: 10.874 I0422 16:54:38.958566 132219947800384 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 646.114, Tokens/s/device: 97389.320, total_weights: 65536, loss: 10.560 I0422 16:54:39.105957 132219947800384 metric_logger.py:185] completed step: 3, seconds: 0.439, TFLOP/s/device: 30.939, Tokens/s/device: 4663.438, total_weights: 65536, loss: 9.980 I0422 16:54:39.110624 132219947800384 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 16:54:39.400379 132219947800384 checkpointing.py:798] Waited 0.28972506523132324 seconds for step 5 to finish before starting checkpointing. I0422 16:54:39.403797 132219947800384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 16:54:40.719330 2594 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 16:54:42.900994 132095712048896 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0422 16:55:14.450336 132095703656192 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 42.755 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.94573140144348 s) (per-host) I0422 16:55:14.450450 132095703656192 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.417041s. I0422 16:55:22.853076 132095703656192 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.819649s. I0422 16:55:22.853389 132094630012672 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 16:55:22.853511 132094630012672 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 16:55:22.853562 132094630012672 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 16:55:22.855376 132094630012672 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 16:55:22.855539 132219947800384 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. W0422 16:55:22.855678 132219947800384 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.451884 seconds. If this number is high, consider checkpointing less frequently. I0422 16:55:22.857851 132219947800384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0422 16:55:22.861922 132219947800384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5. I0422 16:55:23.836643 132219947800384 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 16:55:23.836743 132219947800384 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 I0422 16:55:23.877353 132219947800384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041951s I0422 16:55:23.877520 132219947800384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.867 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31691813468933105 s) (per-host) I0422 16:55:23.877570 132219947800384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.316978s (batch_requests_ready=0.273086s, total_serialization_initiated=0.043825s, others=0.000067s) I0422 16:55:23.877659 132219947800384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.321832s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.321817) I0422 16:55:23.878572 132090351830784 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 16:55:23.878729 132219947800384 async_checkpointer.py:561] Finished blocking save. Time taken: 1.020808s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5. I0422 16:55:24.286041 132219947800384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 16:55:24.286424 132094630012672 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 16:55:24.286603 132219947800384 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776876879.403769, 'wait_for_prev_duration_secs': 43.4518837928772, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776876922.857889, 'checkpointer_blocking_duration_secs': 1.0209476947784424, 'get_old_steps_start_time': 1776876923.8788555, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776876879.4010053, 'checkpoint_manager_blocking_duration_secs': 44.885558128356934} I0422 16:55:24.286715 132219947800384 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 16:55:24.287335 132219947800384 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 89.055, Tokens/s/device: 13423.347, total_weights: 65536, loss: 9.460 I0422 16:55:24.293119 132219947800384 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.200, Tokens/s/device: 13897.371, total_weights: 65536, loss: 8.959 I0422 16:55:43.460407 132219947800384 metric_logger.py:185] completed step: 6, seconds: 45.181, TFLOP/s/device: 0.301, Tokens/s/device: 45.328, total_weights: 65536, loss: 8.469 I0422 16:55:43.607955 132219947800384 metric_logger.py:185] completed step: 7, seconds: 19.021, TFLOP/s/device: 0.714, Tokens/s/device: 107.670, total_weights: 65536, loss: 8.003 I0422 16:55:43.755203 132219947800384 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.483, Tokens/s/device: 13487.793, total_weights: 65536, loss: 7.572 I0422 16:55:43.902251 132219947800384 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 16:55:43.902901 132219947800384 checkpointing.py:798] Waited 0.0006699562072753906 seconds for step 9 to finish before starting checkpointing. I0422 16:55:43.904989 132219947800384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 16:55:47.833577 132095712048896 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0422 16:56:24.104365 132090351830784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.090 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.543720960617065 s) (per-host) I0422 16:56:24.104485 132090351830784 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.225803s. I0422 16:56:31.370411 132090351830784 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 67.491713s. I0422 16:56:31.370703 132094630012672 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 16:56:31.370830 132094630012672 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 16:56:31.370878 132094630012672 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 16:56:31.372202 132094630012672 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 16:56:31.372385 132219947800384 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. W0422 16:56:31.372523 132219947800384 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.467545 seconds. If this number is high, consider checkpointing less frequently. I0422 16:56:31.374366 132219947800384 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 16:56:31.377759 132219947800384 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 16:56:32.352787 132219947800384 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 16:56:32.352883 132219947800384 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 I0422 16:56:32.391219 132219947800384 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039602s I0422 16:56:32.391390 132219947800384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.143 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.719881534576416 s) (per-host) I0422 16:56:32.391440 132219947800384 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.719943s (batch_requests_ready=0.678261s, total_serialization_initiated=0.041616s, others=0.000066s) I0422 16:56:32.391527 132219947800384 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.723817s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.723806) I0422 16:56:32.392403 132094621619968 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 16:56:32.392558 132219947800384 async_checkpointer.py:561] Finished blocking save. Time taken: 1.018121s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9. I0422 16:56:32.394427 132219947800384 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 16:56:32.394640 132094630012672 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 16:56:32.394737 132219947800384 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776876943.9049556, 'wait_for_prev_duration_secs': 47.46754503250122, 'time_between_consecutive_saves_sec': 21.049541234970093, 'checkpointer_blocking_start_time': 1776876991.3744085, 'checkpointer_blocking_duration_secs': 1.0182609558105469, 'get_old_steps_start_time': 1776876992.392684, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1776876943.9031568, 'checkpoint_manager_blocking_duration_secs': 48.491551876068115} I0422 16:56:32.394902 132219947800384 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 16:56:32.394944 132219947800384 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 16:56:37.948300 132094638405376 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0422 16:57:14.603800 132094621619968 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.793 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.93225145339966 s) (per-host) I0422 16:57:14.603919 132094621619968 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.211407s. I0422 16:57:21.725422 132094621619968 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.332894s. I0422 16:57:21.725645 132094630012672 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 16:57:21.725703 132094630012672 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 16:57:21.725746 132094630012672 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 16:57:21.728543 132094630012672 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 16:57:21.728708 132219947800384 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. I0422 16:57:21.728858 132219947800384 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 16:57:21.729533 132219947800384 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.053, Tokens/s/device: 13875.245, 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: Wed Apr 22 16:57:31 UTC 2026 EXIT_CODE=0