XPK Start: Mon Apr 20 20:16:28 UTC 2026 2026-04-20 20:16:32.702747: 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 00:00:1776716192.716421 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776716192.720206 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776716192.731788 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716192.731807 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716192.731810 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776716192.731812 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 20:16:51.969508: 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 20:16:52.492342 140677625362240 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 20:17:01,532:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 20:17:01.532844 140677625362240 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 20:17:01,535:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-true-sv-ab4s5-slice-job-0-0.mt-15-ocdbt-true-sv-ab4s5:8482 I0420 20:17:01.535168 140677625362240 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-ab4s5-slice-job-0-0.mt-15-ocdbt-true-sv-ab4s5:8482 I0420 20:17:02.301230 140677625362240 max_utils.py:284] Jax distributed system initialized! I0420 20:17:09.492793 140677625362240 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 20:17:09.492898 140677625362240 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 20:17:09.492937 140677625362240 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 20:17:09.492976 140677625362240 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 20:17:10.203926 140677625362240 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:17:10.204519 140677625362240 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:17:10.204722 140677625362240 checkpointing.py:688] Setting up checkpoint logger... I0420 20:17:10.204775 140677625362240 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 20:17:10.204818 140677625362240 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 20:17:10.205158 140677625362240 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 0x7ff193179760>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 20:17:13.291598 140677625362240 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 20:17:13.291856 140677625362240 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 0x7fdd0c0cdf70>}, handler_registry=None I0420 20:17:13.292089 140677625362240 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fdd0c0cdf70>` 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 20:17:13.292142 140677625362240 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdd0c0d2d50>` 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 20:17:13.292176 140677625362240 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 0x7fdd0c0cdf70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fdd0c0cdf70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdd0c0d2d50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fdd0c0d2d50>}). I0420 20:17:13.292773 140677625362240 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 20:17:13.292852 140677625362240 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7fdcf4500180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 20:17:13.970893 140677625362240 checkpoint_manager.py:1818] Found 0 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_15_ocdbt_true/checkpoints I0420 20:17:14.434025 140677625362240 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_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7fdd0c2ed730> I0420 20:17:14.434200 140677625362240 checkpointing.py:302] Checkpoint manager created! I0420 20:17:15.324588 140677625362240 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 20:17:15.324710 140677625362240 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:17:15.701150 140677625362240 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:17:15.701246 140677625362240 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:17:15.716809 140677625362240 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:17:15.716869 140677625362240 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:17:15.739382 140677625362240 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 20:17:15.739450 140677625362240 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:17:15.755076 140677625362240 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 20:17:15.755136 140677625362240 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:17:15.770700 140677625362240 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 20:17:15.770760 140677625362240 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:17:15.795146 140677625362240 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 20:17:15.795214 140677625362240 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:17:15.814330 140677625362240 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 20:17:15.814393 140677625362240 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 20:17:16.000197 140677625362240 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 20:17:16.000308 140677625362240 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 20:17:17.419972 140677625362240 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0420 20:17:17.420108 140677625362240 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 20:17:17.420162 140677625362240 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 20:17:17.420220 140677625362240 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 20:17:17.420283 140677625362240 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 20:17:17.420323 140677625362240 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 20:17:17.420375 140677625362240 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 20:17:17.420427 140677625362240 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 20:17:17.420467 140677625362240 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 20:17:17.420505 140677625362240 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 20:17:17.420553 140677625362240 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0420 20:17:17.420601 140677625362240 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0420 20:17:17.923674 140677625362240 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:17:17.923772 140677625362240 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:17:17.938541 140677625362240 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:17:17.938600 140677625362240 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:17:28.673401 140677625362240 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 20:17:28.674306 140677625362240 metric_logger.py:289] number parameters: 1.104 billion I0420 20:17:39.857265 140677625362240 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 20:17:40.012805 140677625362240 checkpointing.py:798] Waited 0.1555314064025879 seconds for step 0 to finish before starting checkpointing. I0420 20:17:40.015308 140677625362240 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:17:40.017006 140677625362240 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 20:17:40.019742 140677625362240 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_15_ocdbt_true/checkpoints/0. I0420 20:17:41.015068 140677625362240 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 20:17:41.016235 140677625362240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:17:41.016294 140677625362240 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 20:17:41.272767 140677625362240 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.257777s I0420 20:17:41.272938 140677625362240 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.765 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5578265190124512 s) (per-host) I0420 20:17:41.272990 140677625362240 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.557888s (batch_requests_ready=0.296461s, total_serialization_initiated=0.261357s, others=0.000070s) I0420 20:17:41.273084 140677625362240 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.562028s (all_items=0.000021s, per_item={'items': '0.00002098'}, temp_paths=0.562007) I0420 20:17:41.274077 140554809632512 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:17:41.274247 140677625362240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.257171s. 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_15_ocdbt_true/checkpoints/0. I0420 20:17:41.689716 140677625362240 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:17:41.690079 140549432530688 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:17:41.690249 140677625362240 standard_logger.py:34] {'step': 0, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716260.0152895, 'wait_for_prev_duration_secs': 6.508827209472656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776716260.0170445, 'checkpointer_blocking_duration_secs': 1.257310152053833, 'get_old_steps_start_time': 1776716261.2743742, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776716260.0133255, 'checkpoint_manager_blocking_duration_secs': 1.6768825054168701} I0420 20:17:41.690357 140677625362240 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 20:17:41.690410 140677625362240 max_utils.py:750] Memstats: After params initialized: I0420 20:17:41.690468 140677625362240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 20:17:41.690503 140677625362240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 20:17:41.690531 140677625362240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 20:17:41.690556 140677625362240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 20:17:41.958309 140677625362240 metric_logger.py:185] completed step: 0, seconds: 11.183, TFLOP/s/device: 1.215, Tokens/s/device: 183.138, total_weights: 65536, loss: 10.874 I0420 20:17:42.131572 140677625362240 metric_logger.py:185] completed step: 1, seconds: 2.100, TFLOP/s/device: 6.471, Tokens/s/device: 975.379, total_weights: 65536, loss: 10.874 I0420 20:17:42.542570 140677625362240 metric_logger.py:185] completed step: 2, seconds: 0.026, TFLOP/s/device: 518.672, Tokens/s/device: 78179.875, total_weights: 65536, loss: 10.560 I0420 20:17:42.690201 140677625362240 metric_logger.py:185] completed step: 3, seconds: 0.412, TFLOP/s/device: 33.008, Tokens/s/device: 4975.294, total_weights: 65536, loss: 9.980 I0420 20:17:42.694872 140677625362240 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 20:17:42.984960 140677625362240 checkpointing.py:798] Waited 0.2900533676147461 seconds for step 5 to finish before starting checkpointing. I0420 20:17:42.987686 140677625362240 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:17:43.994306 2629 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 20:17:46.905004 140554818025216 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_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_7 I0420 20:18:17.715708 140554809632512 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 42.691 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.000524044036865 s) (per-host) I0420 20:18:17.715835 140554809632512 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.441646s. I0420 20:18:26.589722 140554809632512 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 45.315518s. I0420 20:18:26.590017 140549432530688 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:18:26.590134 140549432530688 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:18:26.590183 140549432530688 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:18:26.591933 140549432530688 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 20:18:26.592115 140677625362240 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 20:18:26.592271 140677625362240 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.604596 seconds. If this number is high, consider checkpointing less frequently. I0420 20:18:26.594327 140677625362240 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 5 I0420 20:18:26.598526 140677625362240 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_15_ocdbt_true/checkpoints/5. I0420 20:18:27.599869 140677625362240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:18:27.599968 140677625362240 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 20:18:27.649237 140677625362240 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.050667s I0420 20:18:27.649402 140677625362240 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.907 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3143622875213623 s) (per-host) I0420 20:18:27.649455 140677625362240 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.314425s (batch_requests_ready=0.261866s, total_serialization_initiated=0.052489s, others=0.000070s) I0420 20:18:27.649533 140677625362240 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.319475s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.319458) I0420 20:18:27.650411 140548887238400 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:18:27.650575 140677625362240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.056184s. 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_15_ocdbt_true/checkpoints/5. I0420 20:18:28.084589 140677625362240 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:18:28.085007 140549432530688 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:18:28.085181 140677625362240 standard_logger.py:34] {'step': 5, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716262.9876409, 'wait_for_prev_duration_secs': 43.60459566116333, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776716306.5943654, 'checkpointer_blocking_duration_secs': 1.0563349723815918, 'get_old_steps_start_time': 1776716307.65072, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776716262.9855719, 'checkpoint_manager_blocking_duration_secs': 45.09957313537598} I0420 20:18:28.085291 140677625362240 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 20:18:28.085915 140677625362240 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 89.051, Tokens/s/device: 13422.731, total_weights: 65536, loss: 9.460 I0420 20:18:28.091974 140677625362240 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.178, Tokens/s/device: 13894.071, total_weights: 65536, loss: 8.959 I0420 20:18:47.371850 140677625362240 metric_logger.py:185] completed step: 6, seconds: 45.396, TFLOP/s/device: 0.299, Tokens/s/device: 45.114, total_weights: 65536, loss: 8.469 I0420 20:18:47.519418 140677625362240 metric_logger.py:185] completed step: 7, seconds: 19.134, TFLOP/s/device: 0.710, Tokens/s/device: 107.035, total_weights: 65536, loss: 8.003 I0420 20:18:47.667039 140677625362240 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.153, Tokens/s/device: 13438.144, total_weights: 65536, loss: 7.572 I0420 20:18:47.814219 140677625362240 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 20:18:47.814839 140677625362240 checkpointing.py:798] Waited 0.0006380081176757812 seconds for step 9 to finish before starting checkpointing. I0420 20:18:47.816955 140677625362240 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:18:50.857824 140554818025216 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_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_7 I0420 20:19:30.974392 140548887238400 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 24.821 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 63.63930368423462 s) (per-host) I0420 20:19:30.974511 140548887238400 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 63.323981s. I0420 20:19:35.820446 140548887238400 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 68.169901s. I0420 20:19:35.820749 140549432530688 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:19:35.820868 140549432530688 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:19:35.820915 140549432530688 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:19:35.822533 140549432530688 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 20:19:35.822711 140677625362240 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 20:19:35.822856 140677625362240 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.005914 seconds. If this number is high, consider checkpointing less frequently. I0420 20:19:35.824453 140677625362240 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 20:19:35.828207 140677625362240 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_15_ocdbt_true/checkpoints/9. I0420 20:19:36.819508 140677625362240 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:19:36.819604 140677625362240 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 20:19:36.867031 140677625362240 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048678s I0420 20:19:36.867197 140677625362240 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.075 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7434065341949463 s) (per-host) I0420 20:19:36.867248 140677625362240 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.743467s (batch_requests_ready=0.692741s, total_serialization_initiated=0.050660s, others=0.000066s) I0420 20:19:36.867337 140677625362240 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.747527s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.747514) I0420 20:19:36.868228 140554809632512 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:19:36.868387 140677625362240 async_checkpointer.py:561] Finished blocking save. Time taken: 1.043863s. 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_15_ocdbt_true/checkpoints/9. I0420 20:19:36.870283 140677625362240 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:19:36.870506 140549432530688 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:19:36.870608 140677625362240 standard_logger.py:34] {'step': 9, '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_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776716327.8169262, 'wait_for_prev_duration_secs': 48.005913734436035, 'time_between_consecutive_saves_sec': 21.224953174591064, 'checkpointer_blocking_start_time': 1776716375.8244944, 'checkpointer_blocking_duration_secs': 1.044001817703247, 'get_old_steps_start_time': 1776716376.8685098, 'get_old_steps_duration_secs': 1.8596649169921875e-05, 'checkpoint_manager_blocking_start_time': 1776716327.8150969, 'checkpoint_manager_blocking_duration_secs': 49.05548167228699} I0420 20:19:36.870771 140677625362240 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 20:19:36.870812 140677625362240 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:19:41.174234 140554818025216 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_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_7 I0420 20:20:18.499948 140554809632512 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.276 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.3761191368103 s) (per-host) I0420 20:20:18.500068 140554809632512 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.631731s. I0420 20:20:26.219639 140554809632512 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.351286s. I0420 20:20:26.219869 140549432530688 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:20:26.219928 140549432530688 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:20:26.219968 140549432530688 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:20:26.221610 140549432530688 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 20:20:26.221793 140677625362240 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 20:20:26.221944 140677625362240 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:20:26.222618 140677625362240 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.224, Tokens/s/device: 13900.956, 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: Mon Apr 20 20:20:39 UTC 2026 EXIT_CODE=0