XPK Start: Mon Apr 20 20:50:11 UTC 2026 2026-04-20 20:50:15.299291: 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 00:00:1776718215.312047 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776718215.315738 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776718215.326938 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718215.326958 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718215.326960 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718215.326962 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 20:50:34.644919: 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) I0420 20:50:35.175297 138702098118464 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 20:50:44,216:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 20:50:44.216993 138702098118464 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 20:50:44,219:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-rs-xxx86-slice-job-0-0.mt-16-zarr3-false-rs-xxx86:8482 I0420 20:50:44.219250 138702098118464 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-rs-xxx86-slice-job-0-0.mt-16-zarr3-false-rs-xxx86:8482 I0420 20:50:45.201885 138702098118464 max_utils.py:284] Jax distributed system initialized! I0420 20:50:51.945537 138702098118464 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 20:50:51.945642 138702098118464 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 20:50:51.945782 138702098118464 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 I0420 20:50:51.945821 138702098118464 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 20:50:52.566581 138702098118464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:50:52.567192 138702098118464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:50:52.567376 138702098118464 checkpointing.py:688] Setting up checkpoint logger... I0420 20:50:52.567426 138702098118464 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0420 20:50:52.567469 138702098118464 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 20:50:52.567824 138702098118464 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e259c750830>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 20:50:53.866960 138702098118464 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 20:50:53.867207 138702098118464 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e131b8e21e0>}, handler_registry=None I0420 20:50:53.867438 138702098118464 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e131b8e21e0>` 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`. I0420 20:50:53.867487 138702098118464 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e131b8e6c60>` 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`. I0420 20:50:53.867522 138702098118464 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 0x7e131b8e21e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e131b8e21e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e131b8e6c60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e131b8e6c60>}). I0420 20:50:53.868045 138702098118464 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 20:50:53.868123 138702098118464 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e1319aac180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 20:50:55.738221 138586070427392 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776718039743841398, 'commit_timestamp_nsecs': 1776718088232853616, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA I0420 20:50:55.993591 138586078820096 checkpoint_manager.py:1877] Missing metrics for step 0 I0420 20:50:56.406763 138702098118464 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints I0420 20:50:56.409291 138702098118464 checkpoint_manager.py:929] [process=7][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_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e131be6c770> I0420 20:50:56.409409 138702098118464 checkpointing.py:302] Checkpoint manager created! I0420 20:50:57.309755 138702098118464 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 20:50:57.309861 138702098118464 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:50:57.682169 138702098118464 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:50:57.682265 138702098118464 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:50:57.698438 138702098118464 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:50:57.698500 138702098118464 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:50:57.721493 138702098118464 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:50:57.721563 138702098118464 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:50:57.737813 138702098118464 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:50:57.737875 138702098118464 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:50:57.754074 138702098118464 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:50:57.754134 138702098118464 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:50:57.779484 138702098118464 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 20:50:57.779554 138702098118464 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:50:57.799476 138702098118464 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 20:50:57.799543 138702098118464 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 20:50:57.989948 138702098118464 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 20:50:57.990061 138702098118464 checkpointing.py:582] restoring from this run's directory step 9 I0420 20:50:58.506432 138702098118464 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9. I0420 20:50:59.909560 1410 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 20:51:06.227555 138702098118464 jax_array_handlers.py:837] [process=7] /jax/orbax/read/worker/io/requested throughput: 249.917 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.323195934295654 s) (per-host) I0420 20:51:06.233179 138702098118464 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/read/gbytes_per_sec: 235.433 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.7220447063446045 s) (per-host) I0420 20:51:06.416700 138702098118464 checkpointer.py:318] Finished restoring checkpoint in 8.16 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9. I0420 20:51:06.416863 138702098118464 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776718258.2569096, 'checkpointer_duration_secs': 8.15991997718811, 'checkpoint_manager_start_time': 1776718257.9903061, 'checkpoint_manager_duration_secs': 8.426526069641113} fsdp: 32 I0420 20:51:06.825357 138702098118464 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0420 20:51:06.825483 138702098118464 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) I0420 20:51:06.825536 138702098118464 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) I0420 20:51:06.825596 138702098118464 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') I0420 20:51:06.825650 138702098118464 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) I0420 20:51:06.825708 138702098118464 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) I0420 20:51:06.825764 138702098118464 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) I0420 20:51:06.825819 138702098118464 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') I0420 20:51:06.825859 138702098118464 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) I0420 20:51:06.825897 138702098118464 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) I0420 20:51:06.825944 138702098118464 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0420 20:51:06.825996 138702098118464 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0420 20:51:07.332119 138702098118464 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:51:07.332215 138702098118464 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:51:07.347396 138702098118464 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:51:07.347456 138702098118464 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:51:18.081012 138702098118464 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. I0420 20:51:18.081891 138702098118464 metric_logger.py:289] number parameters: 1.104 billion I0420 20:51:29.398817 138702098118464 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 20:51:29.554120 138702098118464 checkpointing.py:798] Waited 0.15528559684753418 seconds for step 10 to finish before starting checkpointing. I0420 20:51:29.556560 138702098118464 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:51:29.558294 138702098118464 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 10 I0420 20:51:29.561044 138702098118464 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10. I0420 20:51:30.171149 138702098118464 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 20:51:30.172164 138702098118464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:51:30.172223 138702098118464 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 I0420 20:51:30.422865 138702098118464 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251801s I0420 20:51:30.423025 138702098118464 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.928 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.52685546875 s) (per-host) I0420 20:51:30.423075 138702098118464 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.526916s (batch_requests_ready=0.273021s, total_serialization_initiated=0.253828s, others=0.000067s) I0420 20:51:30.423164 138702098118464 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.531076s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.531058) I0420 20:51:30.424060 138585491625728 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:51:30.424240 138702098118464 async_checkpointer.py:561] Finished blocking save. Time taken: 0.865875s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10. I0420 20:51:30.829515 138702098118464 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:51:30.829930 138581699958528 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:51:30.830093 138702098118464 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718289.5565412, 'wait_for_prev_duration_secs': 6.365776062011719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776718289.5583313, 'checkpointer_blocking_duration_secs': 0.8660159111022949, 'get_old_steps_start_time': 1776718290.424366, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776718289.5546865, 'checkpoint_manager_blocking_duration_secs': 1.2753655910491943} I0420 20:51:30.830200 138702098118464 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 20:51:30.830253 138702098118464 max_utils.py:750] Memstats: After params initialized: I0420 20:51:30.830312 138702098118464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 20:51:30.830347 138702098118464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 20:51:30.830375 138702098118464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 20:51:30.830400 138702098118464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 20:51:31.096936 138702098118464 metric_logger.py:185] completed step: 10, seconds: 11.317, TFLOP/s/device: 1.201, Tokens/s/device: 180.970, total_weights: 65536, loss: 6.833 I0420 20:51:31.268105 138702098118464 metric_logger.py:185] completed step: 11, seconds: 1.697, TFLOP/s/device: 8.008, Tokens/s/device: 1207.033, total_weights: 65536, loss: 6.528 I0420 20:51:31.416281 138702098118464 metric_logger.py:185] completed step: 12, seconds: 0.024, TFLOP/s/device: 566.225, Tokens/s/device: 85347.558, total_weights: 65536, loss: 6.265 I0420 20:51:31.564259 138702098118464 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.651, Tokens/s/device: 13362.477, total_weights: 65536, loss: 6.045 I0420 20:51:31.568979 138702098118464 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 20:51:31.859095 138702098118464 checkpointing.py:798] Waited 0.2900819778442383 seconds for step 15 to finish before starting checkpointing. I0420 20:51:31.862003 138702098118464 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:51:35.484708 138581716743936 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/10/items/array_metadatas/process_7 I0420 20:52:12.030699 138585491625728 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.489 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.13448452949524 s) (per-host) I0420 20:52:12.030805 138585491625728 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.606630s. I0420 20:52:19.984842 138585491625728 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.560649s. I0420 20:52:19.985193 138581699958528 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:52:19.985314 138581699958528 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:52:19.985363 138581699958528 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:52:19.986794 138581699958528 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 20:52:19.986986 138702098118464 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0420 20:52:19.987140 138702098118464 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.125141 seconds. If this number is high, consider checkpointing less frequently. I0420 20:52:19.992375 138702098118464 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 15 I0420 20:52:19.996373 138702098118464 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15. I0420 20:52:20.544641 138702098118464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:52:20.544747 138702098118464 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 I0420 20:52:20.594964 138702098118464 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051518s I0420 20:52:20.595119 138702098118464 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.918 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.313687801361084 s) (per-host) I0420 20:52:20.595169 138702098118464 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313747s (batch_requests_ready=0.260297s, total_serialization_initiated=0.053384s, others=0.000066s) I0420 20:52:20.595256 138702098118464 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.318665s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.318653) I0420 20:52:20.596160 138581716743936 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:52:20.596322 138702098118464 async_checkpointer.py:561] Finished blocking save. Time taken: 0.603873s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15. I0420 20:52:20.755514 138702098118464 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:52:20.755907 138581699958528 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:52:20.756072 138702098118464 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718291.861973, 'wait_for_prev_duration_secs': 48.12514138221741, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776718339.9924176, 'checkpointer_blocking_duration_secs': 0.6040112972259521, 'get_old_steps_start_time': 1776718340.5964496, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776718291.8597898, 'checkpoint_manager_blocking_duration_secs': 48.89624547958374} I0420 20:52:20.756196 138702098118464 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 20:52:20.756813 138702098118464 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.732, Tokens/s/device: 13826.907, total_weights: 65536, loss: 5.864 I0420 20:52:20.762747 138702098118464 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.833, Tokens/s/device: 13842.047, total_weights: 65536, loss: 5.719 I0420 20:52:39.142284 138702098118464 metric_logger.py:185] completed step: 16, seconds: 49.193, TFLOP/s/device: 0.276, Tokens/s/device: 41.632, total_weights: 65536, loss: 5.605 I0420 20:52:39.289783 138702098118464 metric_logger.py:185] completed step: 17, seconds: 18.233, TFLOP/s/device: 0.745, Tokens/s/device: 112.322, total_weights: 65536, loss: 5.517 I0420 20:52:39.437156 138702098118464 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.143, Tokens/s/device: 13436.557, total_weights: 65536, loss: 5.448 I0420 20:52:39.584402 138702098118464 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 20:52:39.585106 138702098118464 checkpointing.py:798] Waited 0.0007238388061523438 seconds for step 19 to finish before starting checkpointing. I0420 20:52:39.587112 138702098118464 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:52:43.658070 138582287185664 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/15/items/array_metadatas/process_7 I0420 20:53:20.711682 138581716743936 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 26.139 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.4301974773407 s) (per-host) I0420 20:53:20.711804 138581716743936 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.115530s. I0420 20:53:30.484292 138581716743936 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 69.888004s. I0420 20:53:30.484575 138581699958528 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:53:30.484725 138581699958528 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:53:30.484775 138581699958528 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:53:30.486537 138581699958528 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 20:53:30.486752 138702098118464 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0420 20:53:30.486882 138702098118464 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.899782 seconds. If this number is high, consider checkpointing less frequently. I0420 20:53:30.488238 138702098118464 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 19 I0420 20:53:30.491936 138702098118464 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19. I0420 20:53:31.498426 138702098118464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:53:31.498520 138702098118464 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 I0420 20:53:31.552114 138702098118464 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.054833s I0420 20:53:31.552278 138702098118464 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.954 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3113596439361572 s) (per-host) I0420 20:53:31.552327 138702098118464 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.311418s (batch_requests_ready=0.254374s, total_serialization_initiated=0.056981s, others=0.000064s) I0420 20:53:31.552416 138702098118464 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.315593s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.315581) I0420 20:53:31.553317 138584921184000 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:53:31.553474 138702098118464 async_checkpointer.py:561] Finished blocking save. Time taken: 1.065160s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19. I0420 20:53:31.555375 138702098118464 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:53:31.555603 138581699958528 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:53:31.555773 138702098118464 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718359.587083, 'wait_for_prev_duration_secs': 50.899781942367554, 'time_between_consecutive_saves_sec': 19.60024666786194, 'checkpointer_blocking_start_time': 1776718410.4882848, 'checkpointer_blocking_duration_secs': 1.0652985572814941, 'get_old_steps_start_time': 1776718411.5535965, 'get_old_steps_duration_secs': 2.3365020751953125e-05, 'checkpoint_manager_blocking_start_time': 1776718359.585335, 'checkpoint_manager_blocking_duration_secs': 51.97040796279907} I0420 20:53:31.555879 138702098118464 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 20:53:31.555919 138702098118464 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:53:37.663820 138585483233024 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/19/items/array_metadatas/process_7 I0420 20:54:14.452060 138584921184000 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 36.555 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.211101055145264 s) (per-host) I0420 20:54:14.452180 138584921184000 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.898752s. I0420 20:54:20.080528 138584921184000 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 48.527085s. I0420 20:54:20.080840 138581699958528 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:54:20.080957 138581699958528 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:54:20.080999 138581699958528 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:54:20.083853 138581699958528 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 20:54:20.084041 138702098118464 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0420 20:54:20.084191 138702098118464 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:54:20.084866 138702098118464 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.167, Tokens/s/device: 13892.469, 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: Mon Apr 20 20:54:32 UTC 2026 EXIT_CODE=0