XPK Start: Fri Apr 24 15:53:01 UTC 2026 2026-04-24 15:53:05.589371: 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 00:00:1777045985.602397 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777045985.606384 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777045985.618018 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777045985.618038 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777045985.618041 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777045985.618043 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 15:53:24.911567: 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) I0424 15:53:25.449444 138399127938880 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 15:53:34,492:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 15:53:34.492203 138399127938880 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 15:53:34,494:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-mr1h4-slice-job-0-0.mt-15-ocdbt-true-sv-mr1h4:8482 I0424 15:53:34.494477 138399127938880 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-mr1h4-slice-job-0-0.mt-15-ocdbt-true-sv-mr1h4:8482 I0424 15:53:35.334250 138399127938880 max_utils.py:284] Jax distributed system initialized! I0424 15:53:41.154718 138399127938880 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 15:53:41.154824 138399127938880 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 15:53:41.154863 138399127938880 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 I0424 15:53:41.154897 138399127938880 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 15:53:41.784160 138399127938880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 15:53:41.784777 138399127938880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 15:53:41.784963 138399127938880 checkpointing.py:688] Setting up checkpoint logger... I0424 15:53:41.785016 138399127938880 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 15:53:41.785060 138399127938880 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 15:53:41.785422 138399127938880 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 0x7ddf07f71130>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 15:53:44.956857 138399127938880 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 15:53:44.957108 138399127938880 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 0x7dcce0aaef90>}, handler_registry=None I0424 15:53:44.957346 138399127938880 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dcce0aaef90>` 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`. I0424 15:53:44.957394 138399127938880 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dcce05623c0>` 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`. I0424 15:53:44.957429 138399127938880 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 0x7dcce0aaef90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dcce0aaef90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dcce05623c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dcce05623c0>}). I0424 15:53:44.957949 138399127938880 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 15:53:44.958029 138399127938880 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dcc47110180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 15:53:45.738280 138399127938880 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints I0424 15:53:46.179234 138399127938880 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dcce0560e00> I0424 15:53:46.179415 138399127938880 checkpointing.py:302] Checkpoint manager created! I0424 15:53:47.082353 138399127938880 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 15:53:47.082462 138399127938880 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:53:47.461566 138399127938880 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 15:53:47.461668 138399127938880 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:53:47.478065 138399127938880 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 15:53:47.478148 138399127938880 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:53:47.501568 138399127938880 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:53:47.501639 138399127938880 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:53:47.518051 138399127938880 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:53:47.518126 138399127938880 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:53:47.534383 138399127938880 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:53:47.534446 138399127938880 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:53:47.559748 138399127938880 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 15:53:47.559819 138399127938880 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:53:47.579853 138399127938880 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 15:53:47.579919 138399127938880 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 15:53:47.775607 138399127938880 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 15:53:47.775720 138399127938880 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 15:53:49.226601 138399127938880 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0424 15:53:49.226735 138399127938880 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) I0424 15:53:49.226792 138399127938880 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) I0424 15:53:49.226853 138399127938880 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') I0424 15:53:49.226907 138399127938880 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) I0424 15:53:49.226946 138399127938880 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) I0424 15:53:49.226999 138399127938880 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) I0424 15:53:49.227055 138399127938880 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') I0424 15:53:49.227112 138399127938880 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) I0424 15:53:49.227157 138399127938880 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) I0424 15:53:49.227210 138399127938880 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0424 15:53:49.227261 138399127938880 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0424 15:53:49.760549 138399127938880 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 15:53:49.760641 138399127938880 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 15:53:49.775736 138399127938880 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 15:53:49.775795 138399127938880 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 15:54:00.232152 138399127938880 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. I0424 15:54:00.233133 138399127938880 metric_logger.py:289] number parameters: 1.104 billion I0424 15:54:11.510188 138399127938880 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 15:54:11.825771 138399127938880 checkpointing.py:798] Waited 0.3155686855316162 seconds for step 0 to finish before starting checkpointing. I0424 15:54:11.828303 138399127938880 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 15:54:11.830170 138399127938880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 15:54:11.832865 138399127938880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0. I0424 15:54:12.681222 138399127938880 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 15:54:12.682259 138399127938880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 15:54:12.682317 138399127938880 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 I0424 15:54:12.940982 138399127938880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.259843s I0424 15:54:12.941159 138399127938880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.870 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5374338626861572 s) (per-host) I0424 15:54:12.941211 138399127938880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.537494s (batch_requests_ready=0.273862s, total_serialization_initiated=0.263565s, others=0.000068s) I0424 15:54:12.941308 138399127938880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.541725s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.541705) I0424 15:54:12.942211 138279995291392 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:54:12.942382 138399127938880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.112139s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0. I0424 15:54:12.958740 138399127938880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:54:12.959004 138275163453184 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:54:12.959147 138399127938880 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777046051.8282833, 'wait_for_prev_duration_secs': 6.604194641113281e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777046051.8302085, 'checkpointer_blocking_duration_secs': 1.112281084060669, 'get_old_steps_start_time': 1777046052.9425104, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777046051.8263495, 'checkpoint_manager_blocking_duration_secs': 1.1327383518218994} I0424 15:54:12.959251 138399127938880 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 15:54:12.959309 138399127938880 max_utils.py:750] Memstats: After params initialized: I0424 15:54:12.959357 138399127938880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 15:54:12.959391 138399127938880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 15:54:12.959418 138399127938880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 15:54:12.959443 138399127938880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 15:54:13.227856 138399127938880 metric_logger.py:185] completed step: 0, seconds: 11.277, TFLOP/s/device: 1.205, Tokens/s/device: 181.610, total_weights: 65536, loss: 10.874 I0424 15:54:13.395150 138399127938880 metric_logger.py:185] completed step: 1, seconds: 1.716, TFLOP/s/device: 7.917, Tokens/s/device: 1193.274, total_weights: 65536, loss: 10.874 I0424 15:54:13.810195 138399127938880 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 674.400, Tokens/s/device: 101652.852, total_weights: 65536, loss: 10.560 I0424 15:54:13.957582 138399127938880 metric_logger.py:185] completed step: 3, seconds: 0.415, TFLOP/s/device: 32.702, Tokens/s/device: 4929.179, total_weights: 65536, loss: 9.980 I0424 15:54:13.961951 138399127938880 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 15:54:14.252215 138399127938880 checkpointing.py:798] Waited 0.29022979736328125 seconds for step 5 to finish before starting checkpointing. I0424 15:54:14.254976 138399127938880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:54:15.668526 2555 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 15:54:17.741686 138280003684096 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0424 15:54:48.937921 138279995291392 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 43.236 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.53415513038635 s) (per-host) I0424 15:54:48.938045 138279995291392 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.995722s. I0424 15:54:57.031751 138279995291392 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 44.089414s. I0424 15:54:57.031965 138275163453184 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:54:57.032039 138275163453184 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:54:57.032107 138275163453184 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:54:57.033788 138275163453184 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 15:54:57.033967 138399127938880 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. W0424 15:54:57.034122 138399127938880 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.779135 seconds. If this number is high, consider checkpointing less frequently. I0424 15:54:57.036219 138399127938880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0424 15:54:57.040071 138399127938880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5. I0424 15:54:57.602041 138399127938880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 15:54:57.602153 138399127938880 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 I0424 15:54:57.648572 138399127938880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048424s I0424 15:54:57.648739 138399127938880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.944 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.312014102935791 s) (per-host) I0424 15:54:57.648788 138399127938880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.312072s (batch_requests_ready=0.261628s, total_serialization_initiated=0.050379s, others=0.000065s) I0424 15:54:57.648877 138399127938880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.316993s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.316977) I0424 15:54:57.649655 138274098104064 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:54:57.649837 138399127938880 async_checkpointer.py:561] Finished blocking save. Time taken: 0.613556s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5. I0424 15:54:57.654622 138399127938880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:54:57.654846 138275163453184 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:54:57.655001 138399127938880 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777046054.254946, 'wait_for_prev_duration_secs': 42.77913451194763, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777046097.036258, 'checkpointer_blocking_duration_secs': 0.6136481761932373, 'get_old_steps_start_time': 1777046097.6499245, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1777046054.2528281, 'checkpoint_manager_blocking_duration_secs': 43.40214014053345} I0424 15:54:57.655112 138399127938880 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 15:54:57.655623 138399127938880 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 89.045, Tokens/s/device: 13421.763, total_weights: 65536, loss: 9.460 I0424 15:54:57.661247 138399127938880 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.335, Tokens/s/device: 13917.771, total_weights: 65536, loss: 8.959 I0424 15:55:16.692911 138399127938880 metric_logger.py:185] completed step: 6, seconds: 43.698, TFLOP/s/device: 0.311, Tokens/s/device: 46.867, total_weights: 65536, loss: 8.469 I0424 15:55:16.840337 138399127938880 metric_logger.py:185] completed step: 7, seconds: 18.886, TFLOP/s/device: 0.719, Tokens/s/device: 108.443, total_weights: 65536, loss: 8.003 I0424 15:55:16.987685 138399127938880 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.315, Tokens/s/device: 13462.524, total_weights: 65536, loss: 7.572 I0424 15:55:17.134699 138399127938880 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 15:55:17.135399 138399127938880 checkpointing.py:798] Waited 0.0007224082946777344 seconds for step 9 to finish before starting checkpointing. I0424 15:55:17.137398 138399127938880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:55:21.357921 138280003684096 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0424 15:55:58.808928 138274098104064 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.696 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.472156047821045 s) (per-host) I0424 15:55:58.809046 138274098104064 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.159253s. I0424 15:56:06.007730 138274098104064 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.357924s. I0424 15:56:06.008022 138275163453184 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:56:06.008162 138275163453184 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:56:06.008212 138275163453184 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:56:06.009712 138275163453184 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 15:56:06.009898 138399127938880 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. W0424 15:56:06.010024 138399127938880 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.872641 seconds. If this number is high, consider checkpointing less frequently. I0424 15:56:06.011564 138399127938880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0424 15:56:06.015719 138399127938880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9. I0424 15:56:06.558844 138399127938880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 15:56:06.558940 138399127938880 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 I0424 15:56:06.607148 138399127938880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049292s I0424 15:56:06.607311 138399127938880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.929 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31295156478881836 s) (per-host) I0424 15:56:06.607362 138399127938880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313012s (batch_requests_ready=0.261613s, total_serialization_initiated=0.051333s, others=0.000066s) I0424 15:56:06.607454 138399127938880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317081s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.317069) I0424 15:56:06.608297 138279995291392 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:56:06.608458 138399127938880 async_checkpointer.py:561] Finished blocking save. Time taken: 0.596825s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9. I0424 15:56:06.610282 138399127938880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:56:06.610548 138275163453184 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:56:06.610679 138399127938880 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777046117.1373672, 'wait_for_prev_duration_secs': 48.87264060974121, 'time_between_consecutive_saves_sec': 20.103540658950806, 'checkpointer_blocking_start_time': 1777046166.0116029, 'checkpointer_blocking_duration_secs': 0.5969662666320801, 'get_old_steps_start_time': 1777046166.608583, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1777046117.1356375, 'checkpoint_manager_blocking_duration_secs': 49.475011348724365} I0424 15:56:06.610833 138399127938880 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 15:56:06.610876 138399127938880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:56:12.285207 138280003684096 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0424 15:56:49.689224 138279995291392 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.401 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.39483642578125 s) (per-host) I0424 15:56:49.689328 138279995291392 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.080917s. I0424 15:56:54.763066 138279995291392 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.154639s. I0424 15:56:54.763364 138275163453184 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:56:54.763484 138275163453184 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:56:54.763525 138275163453184 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:56:54.764951 138275163453184 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 15:56:54.765123 138399127938880 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. I0424 15:56:54.765273 138399127938880 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 15:56:54.765879 138399127938880 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.182, Tokens/s/device: 13894.731, 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: Fri Apr 24 15:57:06 UTC 2026 EXIT_CODE=0