XPK Start: Sat Apr 25 15:56:41 UTC 2026 2026-04-25 15:56:45.314261: 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 00:00:1777132605.327272 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777132605.331048 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777132605.342878 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132605.342898 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132605.342901 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777132605.342903 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 15:57:04.538772: 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 15:57:05.070420 133117673342784 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 15:57:14,111:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 15:57:14.111952 133117673342784 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 15:57:14,114:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-8bd3h-slice-job-0-0.mt-15-ocdbt-true-sv-8bd3h:8482 I0425 15:57:14.114260 133117673342784 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-8bd3h-slice-job-0-0.mt-15-ocdbt-true-sv-8bd3h:8482 I0425 15:57:15.218405 133117673342784 max_utils.py:284] Jax distributed system initialized! I0425 15:57:22.031710 133117673342784 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 15:57:22.031819 133117673342784 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 15:57:22.031859 133117673342784 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 15:57:22.031893 133117673342784 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 15:57:22.657917 133117673342784 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 15:57:22.658537 133117673342784 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 15:57:22.658723 133117673342784 checkpointing.py:688] Setting up checkpoint logger... I0425 15:57:22.658773 133117673342784 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 15:57:22.658816 133117673342784 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 15:57:22.659196 133117673342784 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 0x791156570fb0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 15:57:25.518209 133117673342784 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 15:57:25.518565 133117673342784 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 0x78fef63e98b0>}, handler_registry=None I0425 15:57:25.518801 133117673342784 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78fef63e98b0>` 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 15:57:25.518851 133117673342784 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fef63ee510>` 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 15:57:25.518888 133117673342784 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 0x78fef63e98b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78fef63e98b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fef63ee510>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78fef63ee510>}). I0425 15:57:25.519410 133117673342784 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 15:57:25.519488 133117673342784 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78fd28330180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 15:57:26.218852 133117673342784 checkpoint_manager.py:1818] Found 0 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 15:57:26.368627 133117673342784 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78fef6ee1730> I0425 15:57:26.368808 133117673342784 checkpointing.py:302] Checkpoint manager created! I0425 15:57:27.270466 133117673342784 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 15:57:27.270578 133117673342784 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 15:57:27.641469 133117673342784 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 15:57:27.641568 133117673342784 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 15:57:27.657472 133117673342784 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 15:57:27.657532 133117673342784 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 15:57:27.680382 133117673342784 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 15:57:27.680449 133117673342784 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 15:57:27.696304 133117673342784 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 15:57:27.696367 133117673342784 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 15:57:27.712151 133117673342784 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 15:57:27.712210 133117673342784 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 15:57:27.736835 133117673342784 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 15:57:27.736907 133117673342784 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 15:57:27.756489 133117673342784 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 15:57:27.756554 133117673342784 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 15:57:27.946028 133117673342784 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 15:57:27.946158 133117673342784 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 15:57:29.381067 133117673342784 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0425 15:57:29.381210 133117673342784 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 15:57:29.381264 133117673342784 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 15:57:29.381328 133117673342784 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 15:57:29.381381 133117673342784 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 15:57:29.381419 133117673342784 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 15:57:29.381470 133117673342784 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 15:57:29.381522 133117673342784 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 15:57:29.381561 133117673342784 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 15:57:29.381598 133117673342784 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 15:57:29.381646 133117673342784 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0425 15:57:29.381692 133117673342784 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0425 15:57:29.878731 133117673342784 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 15:57:29.878826 133117673342784 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 15:57:29.893754 133117673342784 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 15:57:29.893813 133117673342784 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 15:57:40.692248 133117673342784 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 15:57:40.693177 133117673342784 metric_logger.py:289] number parameters: 1.104 billion I0425 15:57:51.954314 133117673342784 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 15:57:52.109325 133117673342784 checkpointing.py:798] Waited 0.15498566627502441 seconds for step 0 to finish before starting checkpointing. I0425 15:57:52.111672 133117673342784 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 15:57:52.113313 133117673342784 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 15:57:52.116032 133117673342784 async_checkpointer.py:452] [process=6] 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/0. I0425 15:57:52.645770 133117673342784 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 15:57:52.646831 133117673342784 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 15:57:52.646892 133117673342784 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 15:57:52.896678 133117673342784 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250991s I0425 15:57:52.896859 133117673342784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.041 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5072727203369141 s) (per-host) I0425 15:57:52.896911 133117673342784 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.507334s (batch_requests_ready=0.252554s, total_serialization_initiated=0.254710s, others=0.000070s) I0425 15:57:52.897003 133117673342784 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.511552s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.511532) I0425 15:57:52.897940 132991841986304 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 15:57:52.898123 133117673342784 async_checkpointer.py:561] Finished blocking save. Time taken: 0.784722s. 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/0. I0425 15:57:52.903692 133117673342784 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 15:57:52.903994 132991288329984 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 15:57:52.904172 133117673342784 standard_logger.py:34] {'step': 0, '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': 1777132672.111653, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777132672.1133597, 'checkpointer_blocking_duration_secs': 0.7848713397979736, 'get_old_steps_start_time': 1777132672.898251, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1777132672.1099439, 'checkpoint_manager_blocking_duration_secs': 0.7941892147064209} I0425 15:57:52.904275 133117673342784 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 15:57:52.904327 133117673342784 max_utils.py:750] Memstats: After params initialized: I0425 15:57:52.904378 133117673342784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 15:57:52.904412 133117673342784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 15:57:52.904439 133117673342784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 15:57:52.904463 133117673342784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 15:57:53.172531 133117673342784 metric_logger.py:185] completed step: 0, seconds: 11.261, TFLOP/s/device: 1.207, Tokens/s/device: 181.866, total_weights: 65536, loss: 10.874 I0425 15:57:53.337623 133117673342784 metric_logger.py:185] completed step: 1, seconds: 1.217, TFLOP/s/device: 11.165, Tokens/s/device: 1682.980, total_weights: 65536, loss: 10.874 I0425 15:57:53.765640 133117673342784 metric_logger.py:185] completed step: 2, seconds: 0.018, TFLOP/s/device: 770.857, Tokens/s/device: 116191.989, total_weights: 65536, loss: 10.560 I0425 15:57:53.913085 133117673342784 metric_logger.py:185] completed step: 3, seconds: 0.428, TFLOP/s/device: 31.719, Tokens/s/device: 4781.003, total_weights: 65536, loss: 9.980 I0425 15:57:53.917790 133117673342784 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 15:57:54.207662 133117673342784 checkpointing.py:798] Waited 0.28983473777770996 seconds for step 5 to finish before starting checkpointing. I0425 15:57:54.210342 133117673342784 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 15:57:55.600075 2609 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 15:57:58.000863 132991850379008 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0425 15:58:28.725039 132991841986304 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 43.473 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.33539700508118 s) (per-host) I0425 15:58:28.725181 132991841986304 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.827128s. I0425 15:58:37.592771 132991841986304 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.694700s. I0425 15:58:37.593072 132991288329984 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 15:58:37.593217 132991288329984 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 15:58:37.593271 132991288329984 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 15:58:37.594590 132991288329984 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 15:58:37.594767 133117673342784 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 15:58:37.594902 133117673342784 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.384567 seconds. If this number is high, consider checkpointing less frequently. I0425 15:58:37.596995 133117673342784 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 15:58:37.601165 133117673342784 async_checkpointer.py:452] [process=6] 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/5. I0425 15:58:38.140233 133117673342784 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 15:58:38.140340 133117673342784 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 15:58:38.184639 133117673342784 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045786s I0425 15:58:38.184797 133117673342784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.267 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29290151596069336 s) (per-host) I0425 15:58:38.184847 133117673342784 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292960s (batch_requests_ready=0.245305s, total_serialization_initiated=0.047590s, others=0.000065s) I0425 15:58:38.184923 133117673342784 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297971s (all_items=0.000015s, per_item={'items': '0.00001478'}, temp_paths=0.297956) I0425 15:58:38.185786 132991800022784 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 15:58:38.185943 133117673342784 async_checkpointer.py:561] Finished blocking save. Time taken: 0.588880s. 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/5. I0425 15:58:38.200149 133117673342784 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 15:58:38.200472 132991288329984 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 15:58:38.200634 133117673342784 standard_logger.py:34] {'step': 5, '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': 1777132674.2103126, 'wait_for_prev_duration_secs': 43.38456654548645, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777132717.5970345, 'checkpointer_blocking_duration_secs': 0.5890154838562012, 'get_old_steps_start_time': 1777132718.18607, 'get_old_steps_duration_secs': 3.910064697265625e-05, 'checkpoint_manager_blocking_start_time': 1777132674.208349, 'checkpoint_manager_blocking_duration_secs': 43.99225115776062} I0425 15:58:38.200736 133117673342784 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 15:58:38.201288 133117673342784 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.767, Tokens/s/device: 13379.937, total_weights: 65536, loss: 9.460 I0425 15:58:38.207124 133117673342784 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.159, Tokens/s/device: 13891.244, total_weights: 65536, loss: 8.959 I0425 15:58:57.496660 133117673342784 metric_logger.py:185] completed step: 6, seconds: 44.288, TFLOP/s/device: 0.307, Tokens/s/device: 46.243, total_weights: 65536, loss: 8.469 I0425 15:58:57.644165 133117673342784 metric_logger.py:185] completed step: 7, seconds: 19.143, TFLOP/s/device: 0.710, Tokens/s/device: 106.984, total_weights: 65536, loss: 8.003 I0425 15:58:57.791491 133117673342784 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 89.089, Tokens/s/device: 13428.452, total_weights: 65536, loss: 7.572 I0425 15:58:57.938606 133117673342784 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 15:58:57.939265 133117673342784 checkpointing.py:798] Waited 0.0006799697875976562 seconds for step 9 to finish before starting checkpointing. I0425 15:58:57.941061 133117673342784 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 15:59:01.864983 132991850379008 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0425 15:59:38.627627 132991800022784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.008 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.7356903553009 s) (per-host) I0425 15:59:38.627747 132991800022784 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.441850s. I0425 15:59:47.004085 132991800022784 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.818173s. I0425 15:59:47.004378 132991288329984 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 15:59:47.004518 132991288329984 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 15:59:47.004580 132991288329984 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 15:59:47.006072 132991288329984 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 15:59:47.006258 133117673342784 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0425 15:59:47.006383 133117673342784 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.065335 seconds. If this number is high, consider checkpointing less frequently. I0425 15:59:47.008882 133117673342784 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 15:59:47.012546 133117673342784 async_checkpointer.py:452] [process=6] 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/9. I0425 15:59:47.562186 133117673342784 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 15:59:47.562280 133117673342784 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 15:59:47.605751 133117673342784 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044483s I0425 15:59:47.605921 133117673342784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.829 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.319425106048584 s) (per-host) I0425 15:59:47.605973 133117673342784 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.319487s (batch_requests_ready=0.272916s, total_serialization_initiated=0.046504s, others=0.000067s) I0425 15:59:47.606063 133117673342784 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.323409s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.323398) I0425 15:59:47.606845 132991841986304 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 15:59:47.606997 133117673342784 async_checkpointer.py:561] Finished blocking save. Time taken: 0.598048s. 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/9. I0425 15:59:47.608702 133117673342784 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 15:59:47.608965 132991288329984 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 15:59:47.609145 133117673342784 standard_logger.py:34] {'step': 9, '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': 1777132737.9410322, 'wait_for_prev_duration_secs': 49.06533479690552, 'time_between_consecutive_saves_sec': 20.346399784088135, 'checkpointer_blocking_start_time': 1777132787.0089235, 'checkpointer_blocking_duration_secs': 0.5981881618499756, 'get_old_steps_start_time': 1777132787.6071234, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1777132737.9394941, 'checkpoint_manager_blocking_duration_secs': 49.669596433639526} I0425 15:59:47.609246 133117673342784 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 15:59:47.609288 133117673342784 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 15:59:52.292209 132991850379008 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0425 16:00:30.165323 132991841986304 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.839 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.878796339035034 s) (per-host) I0425 16:00:30.165421 132991841986304 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.558465s. I0425 16:00:37.430464 132991841986304 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.823492s. I0425 16:00:37.430749 132991288329984 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:00:37.430869 132991288329984 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:00:37.430911 132991288329984 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:00:37.432334 132991288329984 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:00:37.432508 133117673342784 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 16:00:37.432664 133117673342784 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:00:37.433353 133117673342784 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.169, Tokens/s/device: 13892.658, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 16:00:47 UTC 2026 EXIT_CODE=0