XPK Start: Sat Apr 25 16:01:49 UTC 2026 2026-04-25 16:01:53.587979: 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 00:00:1777132913.600861 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777132913.604619 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777132913.616370 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132913.616388 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132913.616391 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132913.616393 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:02:12.761010: 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) I0425 16:02:13.288326 133703275628352 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:02:22,329:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:02:22.329252 133703275628352 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:02:22,331:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-rs-a3boq-slice-job-0-0.mt-15-ocdbt-true-rs-a3boq:8482 I0425 16:02:22.331696 133703275628352 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-a3boq-slice-job-0-0.mt-15-ocdbt-true-rs-a3boq:8482 I0425 16:02:23.370716 133703275628352 max_utils.py:284] Jax distributed system initialized! I0425 16:02:30.206854 133703275628352 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:02:30.206979 133703275628352 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:02:30.207024 133703275628352 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 I0425 16:02:30.207058 133703275628352 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:02:30.827004 133703275628352 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:02:30.827643 133703275628352 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:02:30.827848 133703275628352 checkpointing.py:688] Setting up checkpoint logger... I0425 16:02:30.827910 133703275628352 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 16:02:30.827984 133703275628352 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:02:30.828356 133703275628352 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 0x7999b1575280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:02:32.100668 133703275628352 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:02:32.100915 133703275628352 checkpoint_manager.py:709] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7986a5ee57c0>}, handler_registry=None I0425 16:02:32.101214 133703275628352 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7986a5ee57c0>` 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`. I0425 16:02:32.101279 133703275628352 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7986a5eee540>` 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`. I0425 16:02:32.101330 133703275628352 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 0x7986a5ee57c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7986a5ee57c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7986a5eee540>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7986a5eee540>}). I0425 16:02:32.101921 133703275628352 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:02:32.102034 133703275628352 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7986a54a0180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:02:33.846708 133582718748416 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777132674234110685, 'commit_timestamp_nsecs': 1777132716427117504, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0425 16:02:34.104060 133582735533824 checkpoint_manager.py:1877] Missing metrics for step 5 I0425 16:02:34.311442 133703275628352 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints I0425 16:02:34.490906 133703275628352 checkpoint_manager.py:929] [process=3][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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7986a5ee4f80> I0425 16:02:34.491122 133703275628352 checkpointing.py:302] Checkpoint manager created! I0425 16:02:35.407726 133703275628352 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:02:35.407836 133703275628352 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:02:35.779960 133703275628352 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:02:35.780055 133703275628352 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:02:35.796327 133703275628352 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:02:35.796390 133703275628352 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:02:35.820201 133703275628352 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:02:35.820271 133703275628352 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:02:35.836413 133703275628352 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:02:35.836483 133703275628352 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:02:35.852815 133703275628352 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:02:35.852887 133703275628352 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:02:35.877945 133703275628352 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:02:35.878016 133703275628352 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:02:35.897708 133703275628352 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:02:35.897778 133703275628352 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:02:36.089518 133703275628352 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:02:36.089651 133703275628352 checkpointing.py:582] restoring from this run's directory step 9 I0425 16:02:36.576557 133703275628352 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9. I0425 16:02:38.394797 1417 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:02:44.710478 133703275628352 jax_array_handlers.py:837] [process=3] /jax/orbax/read/worker/io/requested throughput: 250.009 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.320866823196411 s) (per-host) I0425 16:02:44.725048 133703275628352 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/read/gbytes_per_sec: 235.283 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.726330995559692 s) (per-host) I0425 16:02:44.786332 133703275628352 checkpointer.py:318] Finished restoring checkpoint in 8.46 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9. I0425 16:02:44.786478 133703275628352 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1777132956.328582, 'checkpointer_duration_secs': 8.457866430282593, 'checkpoint_manager_start_time': 1777132956.0899153, 'checkpoint_manager_duration_secs': 8.696535348892212} fsdp: 32 I0425 16:02:45.188447 133703275628352 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0425 16:02:45.188579 133703275628352 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) I0425 16:02:45.188635 133703275628352 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) I0425 16:02:45.188698 133703275628352 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') I0425 16:02:45.188759 133703275628352 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) I0425 16:02:45.188802 133703275628352 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) I0425 16:02:45.188860 133703275628352 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) I0425 16:02:45.188915 133703275628352 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') I0425 16:02:45.189088 133703275628352 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) I0425 16:02:45.189199 133703275628352 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) I0425 16:02:45.189288 133703275628352 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0425 16:02:45.189350 133703275628352 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0425 16:02:45.692070 133703275628352 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:02:45.692168 133703275628352 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:02:45.707190 133703275628352 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:02:45.707252 133703275628352 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:02:56.749171 133703275628352 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. I0425 16:02:56.750086 133703275628352 metric_logger.py:289] number parameters: 1.104 billion I0425 16:03:07.980726 133703275628352 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 16:03:08.133428 133703275628352 checkpointing.py:798] Waited 0.15268492698669434 seconds for step 10 to finish before starting checkpointing. I0425 16:03:08.135862 133703275628352 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:03:08.137971 133703275628352 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 10 I0425 16:03:08.140667 133703275628352 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10. I0425 16:03:08.732122 133703275628352 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:03:08.733086 133703275628352 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:03:08.733145 133703275628352 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 I0425 16:03:08.982197 133703275628352 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250211s I0425 16:03:08.982368 133703275628352 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.987 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5165109634399414 s) (per-host) I0425 16:03:08.982424 133703275628352 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.516572s (batch_requests_ready=0.264194s, total_serialization_initiated=0.252309s, others=0.000068s) I0425 16:03:08.982514 133703275628352 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.520955s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.520935) I0425 16:03:08.983357 133582693570304 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0425 16:03:08.983540 133703275628352 async_checkpointer.py:561] Finished blocking save. Time taken: 0.845490s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10. I0425 16:03:09.386907 133703275628352 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:03:09.387298 133577886910208 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:03:09.387461 133703275628352 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777132988.1358433, 'wait_for_prev_duration_secs': 6.723403930664062e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777132988.138011, 'checkpointer_blocking_duration_secs': 0.8456401824951172, 'get_old_steps_start_time': 1777132988.983669, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777132988.1339526, 'checkpoint_manager_blocking_duration_secs': 1.2534699440002441} I0425 16:03:09.387579 133703275628352 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 16:03:09.387631 133703275628352 max_utils.py:750] Memstats: After params initialized: I0425 16:03:09.387685 133703275628352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0425 16:03:09.387721 133703275628352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0425 16:03:09.387747 133703275628352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0425 16:03:09.387774 133703275628352 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0425 16:03:09.654666 133703275628352 metric_logger.py:185] completed step: 10, seconds: 11.231, TFLOP/s/device: 1.210, Tokens/s/device: 182.360, total_weights: 65536, loss: 6.833 I0425 16:03:09.837319 133703275628352 metric_logger.py:185] completed step: 11, seconds: 1.673, TFLOP/s/device: 8.123, Tokens/s/device: 1224.456, total_weights: 65536, loss: 6.528 I0425 16:03:09.984689 133703275628352 metric_logger.py:185] completed step: 12, seconds: 0.035, TFLOP/s/device: 385.035, Tokens/s/device: 58036.726, total_weights: 65536, loss: 6.265 I0425 16:03:10.131971 133703275628352 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.579, Tokens/s/device: 13351.587, total_weights: 65536, loss: 6.045 I0425 16:03:10.136823 133703275628352 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 16:03:10.426469 133703275628352 checkpointing.py:798] Waited 0.28961873054504395 seconds for step 15 to finish before starting checkpointing. I0425 16:03:10.429525 133703275628352 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:03:13.513921 133577903695616 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_3 I0425 16:03:49.773241 133582693570304 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 38.240 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.30734443664551 s) (per-host) I0425 16:03:49.773374 133582693570304 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.789897s. I0425 16:04:02.786366 133582693570304 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 53.802875s. I0425 16:04:02.786678 133577886910208 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:04:02.786802 133577886910208 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:04:02.786850 133577886910208 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:04:02.788501 133577886910208 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 16:04:02.788687 133703275628352 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0425 16:04:02.788836 133703275628352 checkpoint_manager.py:1458] Waiting for previous save to complete took 52.359314 seconds. If this number is high, consider checkpointing less frequently. I0425 16:04:02.790873 133703275628352 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 15 I0425 16:04:02.795087 133703275628352 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15. I0425 16:04:03.738610 133703275628352 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:04:03.738709 133703275628352 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 I0425 16:04:03.784974 133703275628352 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047638s I0425 16:04:03.785145 133703275628352 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.280 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29218029975891113 s) (per-host) I0425 16:04:03.785196 133703275628352 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292241s (batch_requests_ready=0.242646s, total_serialization_initiated=0.049527s, others=0.000068s) I0425 16:04:03.785287 133703275628352 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297162s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.297150) I0425 16:04:03.786223 133579505915648 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0425 16:04:03.786352 133703275628352 async_checkpointer.py:561] Finished blocking save. Time taken: 0.995406s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15. I0425 16:04:03.917120 133703275628352 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:04:03.917481 133577886910208 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:04:03.917653 133703275628352 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777132990.4294956, 'wait_for_prev_duration_secs': 52.35931396484375, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777133042.7909136, 'checkpointer_blocking_duration_secs': 0.9956164360046387, 'get_old_steps_start_time': 1777133043.7865474, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1777132990.427074, 'checkpoint_manager_blocking_duration_secs': 53.49054408073425} I0425 16:04:03.917772 133703275628352 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 16:04:03.918380 133703275628352 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.112, Tokens/s/device: 13884.087, total_weights: 65536, loss: 5.864 I0425 16:04:03.924121 133703275628352 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.296, Tokens/s/device: 13911.909, total_weights: 65536, loss: 5.719 I0425 16:04:21.969381 133703275628352 metric_logger.py:185] completed step: 16, seconds: 53.786, TFLOP/s/device: 0.253, Tokens/s/device: 38.077, total_weights: 65536, loss: 5.605 I0425 16:04:22.116952 133703275628352 metric_logger.py:185] completed step: 17, seconds: 17.899, TFLOP/s/device: 0.759, Tokens/s/device: 114.419, total_weights: 65536, loss: 5.517 I0425 16:04:22.264230 133703275628352 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.249, Tokens/s/device: 13452.620, total_weights: 65536, loss: 5.448 I0425 16:04:22.411290 133703275628352 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 16:04:22.411958 133703275628352 checkpointing.py:798] Waited 0.0006878376007080078 seconds for step 19 to finish before starting checkpointing. I0425 16:04:22.413998 133703275628352 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:04:26.609338 133582207055616 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_3 I0425 16:05:03.301168 133579505915648 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 26.411 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.808159589767456 s) (per-host) I0425 16:05:03.301282 133579505915648 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.514852s. I0425 16:05:10.467836 133579505915648 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 66.681391s. I0425 16:05:10.468151 133577886910208 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:05:10.468279 133577886910208 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:05:10.468326 133577886910208 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:05:10.470045 133577886910208 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 16:05:10.470221 133703275628352 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 16:05:10.470357 133703275628352 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.056369 seconds. If this number is high, consider checkpointing less frequently. I0425 16:05:10.471885 133703275628352 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 19 I0425 16:05:10.475962 133703275628352 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19. I0425 16:05:11.035537 133703275628352 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:05:11.035634 133703275628352 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 I0425 16:05:11.080800 133703275628352 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046188s I0425 16:05:11.080977 133703275628352 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.917 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31372570991516113 s) (per-host) I0425 16:05:11.081027 133703275628352 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313786s (batch_requests_ready=0.265480s, total_serialization_initiated=0.048240s, others=0.000066s) I0425 16:05:11.081114 133703275628352 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318478s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.318467) I0425 16:05:11.082052 133582207055616 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0425 16:05:11.082213 133703275628352 async_checkpointer.py:561] Finished blocking save. Time taken: 0.610226s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19. I0425 16:05:11.084087 133703275628352 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:05:11.084306 133577886910208 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:05:11.084433 133703275628352 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777133062.4139707, 'wait_for_prev_duration_secs': 48.0563690662384, 'time_between_consecutive_saves_sec': 19.625428438186646, 'checkpointer_blocking_start_time': 1777133110.4719558, 'checkpointer_blocking_duration_secs': 0.6103677749633789, 'get_old_steps_start_time': 1777133111.0823362, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1777133062.412188, 'checkpoint_manager_blocking_duration_secs': 48.67221522331238} I0425 16:05:11.084545 133703275628352 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 16:05:11.084591 133703275628352 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:05:15.849976 133582693570304 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_3 I0425 16:05:52.179063 133582207055616 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 38.144 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.41176986694336 s) (per-host) I0425 16:05:52.179187 133582207055616 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.097022s. I0425 16:05:59.642035 133582207055616 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 48.559856s. I0425 16:05:59.642322 133577886910208 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:05:59.642453 133577886910208 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:05:59.642506 133577886910208 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:05:59.644271 133577886910208 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 16:05:59.644452 133703275628352 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0425 16:05:59.644618 133703275628352 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:05:59.645257 133703275628352 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.181, Tokens/s/device: 13894.543, 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: Sat Apr 25 16:06:10 UTC 2026 EXIT_CODE=0