XPK Start: Sun Apr 19 04:42:11 UTC 2026 2026-04-19 04:42:15.150113: 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 00:00:1776573735.163267 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776573735.167012 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776573735.178376 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776573735.178395 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776573735.178397 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776573735.178399 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 04:42:34.916691: 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 04:42:35.440502 135942732764992 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 04:42:44,479:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 04:42:44.479432 135942732764992 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 04:42:44,481:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-zzn91-slice-job-0-0.mt-16-zarr3-true-rs-zzn91:8482 I0419 04:42:44.481714 135942732764992 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-zzn91-slice-job-0-0.mt-16-zarr3-true-rs-zzn91:8482 I0419 04:42:45.593563 135942732764992 max_utils.py:284] Jax distributed system initialized! I0419 04:42:52.507030 135942732764992 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 04:42:52.507136 135942732764992 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 04:42:52.507178 135942732764992 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 04:42:52.507211 135942732764992 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 04:42:53.127839 135942732764992 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:42:53.128458 135942732764992 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:42:53.128651 135942732764992 checkpointing.py:688] Setting up checkpoint logger... I0419 04:42:53.128703 135942732764992 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 04:42:53.128747 135942732764992 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 04:42:53.129089 135942732764992 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 0x7ba31df79280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 04:42:55.682955 135942732764992 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 04:42:55.683323 135942732764992 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 0x7b8ebc3c5eb0>}, handler_registry=None I0419 04:42:55.683562 135942732764992 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8ebc3c5eb0>` 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 04:42:55.683614 135942732764992 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8ebc3ca900>` 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 04:42:55.683651 135942732764992 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 0x7b8ebc3c5eb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b8ebc3c5eb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8ebc3ca900>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b8ebc3ca900>}). I0419 04:42:55.684153 135942732764992 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 04:42:55.684232 135942732764992 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b8e8c6ac180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 04:42:57.159513 135822409975552 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776573557937034614, 'commit_timestamp_nsecs': 1776573604846198718, '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_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA I0419 04:42:57.413124 135822409975552 checkpoint_manager.py:1877] Missing metrics for step 5 I0419 04:42:58.962724 135942732764992 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_16_zarr3_true/checkpoints I0419 04:42:59.075026 135942732764992 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_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ba1dcfd98b0> I0419 04:42:59.075206 135942732764992 checkpointing.py:302] Checkpoint manager created! I0419 04:42:59.972485 135942732764992 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 04:42:59.972601 135942732764992 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:43:00.346675 135942732764992 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:43:00.346772 135942732764992 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:43:00.362902 135942732764992 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:43:00.362964 135942732764992 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:43:00.386153 135942732764992 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 04:43:00.386222 135942732764992 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:43:00.408618 135942732764992 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 04:43:00.408719 135942732764992 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:43:00.424863 135942732764992 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 04:43:00.424926 135942732764992 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:43:00.450208 135942732764992 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 04:43:00.450329 135942732764992 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:43:00.470107 135942732764992 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 04:43:00.470174 135942732764992 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 04:43:00.663803 135942732764992 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 04:43:00.663914 135942732764992 checkpointing.py:582] restoring from this run's directory step 9 I0419 04:43:01.147110 135942732764992 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_16_zarr3_true/checkpoints/9. I0419 04:43:03.294354 1417 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 04:43:10.030128 135942732764992 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 234.425 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.741071462631226 s) (per-host) I0419 04:43:10.043359 135942732764992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 208.874 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.576770305633545 s) (per-host) I0419 04:43:10.323595 135942732764992 checkpointer.py:318] Finished restoring checkpoint in 9.42 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/9. I0419 04:43:10.323753 135942732764992 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_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776573780.905761, 'checkpointer_duration_secs': 9.41796088218689, 'checkpoint_manager_start_time': 1776573780.6641667, 'checkpoint_manager_duration_secs': 9.659557342529297} fsdp: 32 I0419 04:43:10.744148 135942732764992 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0419 04:43:10.744304 135942732764992 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 04:43:10.744364 135942732764992 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 04:43:10.744428 135942732764992 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 04:43:10.744482 135942732764992 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 04:43:10.744524 135942732764992 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 04:43:10.744584 135942732764992 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 04:43:10.744639 135942732764992 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 04:43:10.744681 135942732764992 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 04:43:10.744722 135942732764992 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 04:43:10.744771 135942732764992 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0419 04:43:10.744819 135942732764992 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0419 04:43:11.254123 135942732764992 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:43:11.254219 135942732764992 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:43:11.275430 135942732764992 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:43:11.275532 135942732764992 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:43:22.069112 135942732764992 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 04:43:22.070013 135942732764992 metric_logger.py:289] number parameters: 1.104 billion I0419 04:43:33.274719 135942732764992 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 04:43:33.430500 135942732764992 checkpointing.py:798] Waited 0.1557626724243164 seconds for step 10 to finish before starting checkpointing. I0419 04:43:33.433053 135942732764992 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:43:33.435104 135942732764992 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0419 04:43:33.437758 135942732764992 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_16_zarr3_true/checkpoints/10. I0419 04:43:35.277804 135942732764992 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 04:43:35.279028 135942732764992 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:43:35.279141 135942732764992 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 04:43:35.530437 135942732764992 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252728s I0419 04:43:35.530608 135942732764992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.636 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9426836967468262 s) (per-host) I0419 04:43:35.530661 135942732764992 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.942747s (batch_requests_ready=0.687881s, total_serialization_initiated=0.254794s, others=0.000072s) I0419 04:43:35.530753 135942732764992 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.946915s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.946895) I0419 04:43:35.531678 135821898282752 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:43:35.531843 135942732764992 async_checkpointer.py:561] Finished blocking save. Time taken: 2.096671s. 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_16_zarr3_true/checkpoints/10. I0419 04:43:35.535436 135942732764992 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:43:35.535734 135817035888384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:43:35.535902 135942732764992 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776573813.4330354, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776573813.4351432, 'checkpointer_blocking_duration_secs': 2.096806526184082, 'get_old_steps_start_time': 1776573815.5319676, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776573813.4310641, 'checkpoint_manager_blocking_duration_secs': 2.104804039001465} I0419 04:43:35.536002 135942732764992 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 04:43:35.536052 135942732764992 max_utils.py:750] Memstats: After params initialized: I0419 04:43:35.536097 135942732764992 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 04:43:35.536130 135942732764992 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 04:43:35.536157 135942732764992 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 04:43:35.536181 135942732764992 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 04:43:35.800831 135942732764992 metric_logger.py:185] completed step: 10, seconds: 11.205, TFLOP/s/device: 1.213, Tokens/s/device: 182.782, total_weights: 65536, loss: 6.833 I0419 04:43:35.972494 135942732764992 metric_logger.py:185] completed step: 11, seconds: 2.525, TFLOP/s/device: 5.381, Tokens/s/device: 811.158, total_weights: 65536, loss: 6.528 I0419 04:43:36.120217 135942732764992 metric_logger.py:185] completed step: 12, seconds: 0.025, TFLOP/s/device: 554.509, Tokens/s/device: 83581.602, total_weights: 65536, loss: 6.265 I0419 04:43:36.267837 135942732764992 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.815, Tokens/s/device: 13387.108, total_weights: 65536, loss: 6.045 I0419 04:43:36.272379 135942732764992 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 04:43:36.562556 135942732764992 checkpointing.py:798] Waited 0.2901418209075928 seconds for step 15 to finish before starting checkpointing. I0419 04:43:36.565439 135942732764992 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:43:41.237315 135817569744640 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_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0419 04:44:17.369219 135821898282752 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.923 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.78127098083496 s) (per-host) I0419 04:44:17.369340 135821898282752 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.837554s. I0419 04:44:26.273431 135821898282752 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.741630s. I0419 04:44:26.273716 135817035888384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:44:26.273839 135817035888384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:44:26.273887 135817035888384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:44:26.275635 135817035888384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 04:44:26.275827 135942732764992 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 04:44:26.275958 135942732764992 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.710523 seconds. If this number is high, consider checkpointing less frequently. I0419 04:44:26.277961 135942732764992 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0419 04:44:26.282171 135942732764992 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_16_zarr3_true/checkpoints/15. I0419 04:44:28.524206 135942732764992 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:44:28.524329 135942732764992 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 04:44:28.565588 135942732764992 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042744s I0419 04:44:28.565753 135942732764992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.338 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 1.1533167362213135 s) (per-host) I0419 04:44:28.565803 135942732764992 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 1.153376s (batch_requests_ready=1.108639s, total_serialization_initiated=0.044669s, others=0.000067s) I0419 04:44:28.565891 135942732764992 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.158235s (all_items=0.000014s, per_item={'items': '0.00001359'}, temp_paths=1.158222) I0419 04:44:28.566828 135819180357376 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:44:28.566988 135942732764992 async_checkpointer.py:561] Finished blocking save. Time taken: 2.288956s. 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_16_zarr3_true/checkpoints/15. I0419 04:44:28.568912 135942732764992 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:44:28.569123 135817035888384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:44:28.569262 135942732764992 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776573816.565411, 'wait_for_prev_duration_secs': 49.71052289009094, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776573866.278, 'checkpointer_blocking_duration_secs': 2.2890987396240234, 'get_old_steps_start_time': 1776573868.5671186, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1776573816.5631495, 'checkpoint_manager_blocking_duration_secs': 52.00607180595398} I0419 04:44:28.569416 135942732764992 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 04:44:28.569926 135942732764992 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.891, Tokens/s/device: 13850.753, total_weights: 65536, loss: 5.864 I0419 04:44:28.575492 135942732764992 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.178, Tokens/s/device: 13894.071, total_weights: 65536, loss: 5.719 I0419 04:44:46.005293 135942732764992 metric_logger.py:185] completed step: 16, seconds: 52.302, TFLOP/s/device: 0.260, Tokens/s/device: 39.157, total_weights: 65536, loss: 5.605 I0419 04:44:46.152937 135942732764992 metric_logger.py:185] completed step: 17, seconds: 17.284, TFLOP/s/device: 0.786, Tokens/s/device: 118.494, total_weights: 65536, loss: 5.517 I0419 04:44:46.300360 135942732764992 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.265, Tokens/s/device: 13454.918, total_weights: 65536, loss: 5.448 I0419 04:44:46.447640 135942732764992 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 04:44:46.448338 135942732764992 checkpointing.py:798] Waited 0.0007162094116210938 seconds for step 19 to finish before starting checkpointing. I0419 04:44:46.450213 135942732764992 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:44:50.677981 135821327841024 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_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0419 04:45:26.711678 135819180357376 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.638 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.29920148849487 s) (per-host) I0419 04:45:26.711794 135819180357376 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 58.144855s. I0419 04:45:34.312304 135819180357376 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 65.745343s. I0419 04:45:34.312594 135817035888384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:45:34.312727 135817035888384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:45:34.312781 135817035888384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:45:34.314299 135817035888384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 04:45:34.314489 135942732764992 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 04:45:34.314620 135942732764992 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.864420 seconds. If this number is high, consider checkpointing less frequently. I0419 04:45:34.316136 135942732764992 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0419 04:45:34.319554 135942732764992 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_16_zarr3_true/checkpoints/19. I0419 04:45:35.690046 135942732764992 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:45:35.690138 135942732764992 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 04:45:35.732359 135942732764992 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043460s I0419 04:45:35.732519 135942732764992 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.172 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7103519439697266 s) (per-host) I0419 04:45:35.732568 135942732764992 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.710410s (batch_requests_ready=0.664823s, total_serialization_initiated=0.045525s, others=0.000062s) I0419 04:45:35.732658 135942732764992 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.714355s (all_items=0.000009s, per_item={'items': '0.00000930'}, temp_paths=0.714345) I0419 04:45:35.733523 135821889890048 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:45:35.733681 135942732764992 async_checkpointer.py:561] Finished blocking save. Time taken: 1.417476s. 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_16_zarr3_true/checkpoints/19. I0419 04:45:35.759457 135942732764992 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:45:35.759727 135817035888384 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:45:35.759864 135942732764992 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776573886.4501834, 'wait_for_prev_duration_secs': 47.86442017555237, 'time_between_consecutive_saves_sec': 20.174508810043335, 'checkpointer_blocking_start_time': 1776573934.3161743, 'checkpointer_blocking_duration_secs': 1.4176173210144043, 'get_old_steps_start_time': 1776573935.7338042, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1776573886.4485674, 'checkpoint_manager_blocking_duration_secs': 49.31126427650452} I0419 04:45:35.759967 135942732764992 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 04:45:35.760008 135942732764992 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:45:42.180938 135821898282752 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_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0419 04:46:18.426832 135821889890048 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.392 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.404624700546265 s) (per-host) I0419 04:46:18.426954 135821889890048 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.693319s. I0419 04:46:27.471908 135821889890048 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.738258s. I0419 04:46:27.472151 135817035888384 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:46:27.472220 135817035888384 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:46:27.472273 135817035888384 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:46:27.473974 135817035888384 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 04:46:27.474131 135942732764992 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 04:46:27.474288 135942732764992 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:46:27.474959 135942732764992 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.076, Tokens/s/device: 13878.630, 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 04:46:37 UTC 2026 EXIT_CODE=0