XPK Start: Thu Apr 23 17:05:37 UTC 2026 2026-04-23 17:05:41.862440: 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 00:00:1776963941.875270 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776963941.879268 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776963941.890658 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963941.890679 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963941.890681 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776963941.890683 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 17:06:01.102510: 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) I0423 17:06:01.635366 138857496962880 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 17:06:10,677:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 17:06:10.677407 138857496962880 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 17:06:10,679:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-sv-5qzhr-slice-job-0-0.mt-15-ocdbt-false-sv-5qzhr:8482 I0423 17:06:10.679697 138857496962880 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-5qzhr-slice-job-0-0.mt-15-ocdbt-false-sv-5qzhr:8482 I0423 17:06:11.892691 138857496962880 max_utils.py:284] Jax distributed system initialized! I0423 17:06:17.651313 138857496962880 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 17:06:17.651420 138857496962880 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 17:06:17.651464 138857496962880 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 I0423 17:06:17.651502 138857496962880 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 17:06:18.273605 138857496962880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 17:06:18.274233 138857496962880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 17:06:18.274416 138857496962880 checkpointing.py:688] Setting up checkpoint logger... I0423 17:06:18.274468 138857496962880 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0423 17:06:18.274519 138857496962880 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 17:06:18.274869 138857496962880 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e49c3f71490>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 17:06:21.286827 138857496962880 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 17:06:21.287198 138857496962880 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 0x7e3775fea120>}, handler_registry=None I0423 17:06:21.287441 138857496962880 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e3775fea120>` 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`. I0423 17:06:21.287490 138857496962880 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3775fef3e0>` 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`. I0423 17:06:21.287524 138857496962880 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 0x7e3775fea120>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e3775fea120>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3775fef3e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3775fef3e0>}). I0423 17:06:21.288027 138857496962880 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 17:06:21.288119 138857496962880 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e37755a0180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 17:06:22.291534 138857496962880 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints I0423 17:06:22.330430 138857496962880 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e3775fea3f0> I0423 17:06:22.330570 138857496962880 checkpointing.py:302] Checkpoint manager created! I0423 17:06:23.220540 138857496962880 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 17:06:23.220643 138857496962880 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:06:23.595676 138857496962880 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 17:06:23.595766 138857496962880 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:06:23.611798 138857496962880 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 17:06:23.611857 138857496962880 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 17:06:23.634761 138857496962880 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 17:06:23.634836 138857496962880 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:06:23.650808 138857496962880 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 17:06:23.650870 138857496962880 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:06:23.666806 138857496962880 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 17:06:23.666867 138857496962880 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:06:23.691727 138857496962880 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 17:06:23.691797 138857496962880 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 17:06:23.711535 138857496962880 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 17:06:23.711601 138857496962880 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 17:06:23.901228 138857496962880 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 17:06:23.901338 138857496962880 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 17:06:25.335138 138857496962880 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 17:06:25.335273 138857496962880 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) I0423 17:06:25.335334 138857496962880 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) I0423 17:06:25.335394 138857496962880 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') I0423 17:06:25.335448 138857496962880 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) I0423 17:06:25.335488 138857496962880 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) I0423 17:06:25.335540 138857496962880 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) I0423 17:06:25.335594 138857496962880 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') I0423 17:06:25.335636 138857496962880 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) I0423 17:06:25.335674 138857496962880 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) I0423 17:06:25.335723 138857496962880 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0423 17:06:25.335772 138857496962880 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0423 17:06:25.836527 138857496962880 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 17:06:25.836626 138857496962880 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0423 17:06:25.851927 138857496962880 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 17:06:25.851988 138857496962880 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0423 17:06:36.543013 138857496962880 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. I0423 17:06:36.543853 138857496962880 metric_logger.py:289] number parameters: 1.104 billion I0423 17:06:47.689880 138857496962880 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 17:06:48.115461 138857496962880 checkpointing.py:798] Waited 0.42556095123291016 seconds for step 0 to finish before starting checkpointing. I0423 17:06:48.120142 138857496962880 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 17:06:48.121896 138857496962880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0423 17:06:48.125143 138857496962880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/0. I0423 17:06:49.121013 138857496962880 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 17:06:49.122047 138857496962880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:06:49.122121 138857496962880 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 I0423 17:06:49.382904 138857496962880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.261971s I0423 17:06:49.383076 138857496962880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.766 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.557624340057373 s) (per-host) I0423 17:06:49.383153 138857496962880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.557710s (batch_requests_ready=0.292033s, total_serialization_initiated=0.265583s, others=0.000094s) I0423 17:06:49.383247 138857496962880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.561875s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.561856) I0423 17:06:49.384236 138738147505920 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:06:49.384365 138857496962880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.262395s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/0. I0423 17:06:49.386257 138857496962880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:06:49.386565 138733349238528 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:06:49.386696 138857496962880 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776964008.1201231, 'wait_for_prev_duration_secs': 6.246566772460938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776964008.1219375, 'checkpointer_blocking_duration_secs': 1.2626111507415771, 'get_old_steps_start_time': 1776964009.3845692, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776964008.116051, 'checkpoint_manager_blocking_duration_secs': 1.2706072330474854} I0423 17:06:49.386846 138857496962880 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 17:06:49.386897 138857496962880 max_utils.py:750] Memstats: After params initialized: I0423 17:06:49.386945 138857496962880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 17:06:49.386981 138857496962880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 17:06:49.387010 138857496962880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 17:06:49.387034 138857496962880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 17:06:49.659039 138857496962880 metric_logger.py:185] completed step: 0, seconds: 11.146, TFLOP/s/device: 1.219, Tokens/s/device: 183.744, total_weights: 65536, loss: 10.874 I0423 17:06:49.830410 138857496962880 metric_logger.py:185] completed step: 1, seconds: 1.968, TFLOP/s/device: 6.904, Tokens/s/device: 1040.711, total_weights: 65536, loss: 10.874 I0423 17:06:50.279155 138857496962880 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 607.653, Tokens/s/device: 91592.129, total_weights: 65536, loss: 10.560 I0423 17:06:50.426657 138857496962880 metric_logger.py:185] completed step: 3, seconds: 0.451, TFLOP/s/device: 30.122, Tokens/s/device: 4540.265, total_weights: 65536, loss: 9.980 I0423 17:06:50.431025 138857496962880 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0423 17:06:50.721410 138857496962880 checkpointing.py:798] Waited 0.29034852981567383 seconds for step 5 to finish before starting checkpointing. I0423 17:06:50.724209 138857496962880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:06:52.037119 2541 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 17:06:53.983337 138738155898624 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6 I0423 17:06:59.503010 138738147505920 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 147.937 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.677516460418701 s) (per-host) I0423 17:06:59.503149 138738147505920 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.118702s. I0423 17:07:05.386812 138738147505920 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.002351s. I0423 17:07:05.387155 138733349238528 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:07:05.387279 138733349238528 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:07:05.387331 138733349238528 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:07:05.389574 138733349238528 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 17:07:05.389760 138857496962880 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. W0423 17:07:05.389885 138857496962880 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.665682 seconds. If this number is high, consider checkpointing less frequently. I0423 17:07:05.391800 138857496962880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0423 17:07:05.395967 138857496962880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/5. I0423 17:07:05.948922 138857496962880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:07:05.949022 138857496962880 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 I0423 17:07:05.994898 138857496962880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047246s I0423 17:07:05.995064 138857496962880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.891 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3153989315032959 s) (per-host) I0423 17:07:05.995130 138857496962880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.315473s (batch_requests_ready=0.266247s, total_serialization_initiated=0.049143s, others=0.000082s) I0423 17:07:05.995222 138857496962880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.320336s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.320321) I0423 17:07:05.996121 138729557571328 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:07:05.996248 138857496962880 async_checkpointer.py:561] Finished blocking save. Time taken: 0.604372s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/5. I0423 17:07:06.015073 138857496962880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:07:06.015367 138733349238528 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:07:06.015541 138857496962880 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776964010.7241805, 'wait_for_prev_duration_secs': 14.665681838989258, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776964025.3918397, 'checkpointer_blocking_duration_secs': 0.6045761108398438, 'get_old_steps_start_time': 1776964025.9964359, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776964010.7220144, 'checkpoint_manager_blocking_duration_secs': 15.293489217758179} I0423 17:07:06.015659 138857496962880 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0423 17:07:06.016257 138857496962880 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.177, Tokens/s/device: 13441.760, total_weights: 65536, loss: 9.460 I0423 17:07:06.022114 138857496962880 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.202, Tokens/s/device: 13897.748, total_weights: 65536, loss: 8.959 I0423 17:07:25.185122 138857496962880 metric_logger.py:185] completed step: 6, seconds: 15.590, TFLOP/s/device: 0.872, Tokens/s/device: 131.367, total_weights: 65536, loss: 8.469 I0423 17:07:25.332612 138857496962880 metric_logger.py:185] completed step: 7, seconds: 19.017, TFLOP/s/device: 0.714, Tokens/s/device: 107.694, total_weights: 65536, loss: 8.003 I0423 17:07:25.479906 138857496962880 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.313, Tokens/s/device: 13462.171, total_weights: 65536, loss: 7.572 I0423 17:07:25.627010 138857496962880 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 17:07:25.627676 138857496962880 checkpointing.py:798] Waited 0.0006847381591796875 seconds for step 9 to finish before starting checkpointing. I0423 17:07:25.629607 138857496962880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:07:30.656883 138738155898624 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6 I0423 17:07:36.405584 138729557571328 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 51.409 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 30.725874662399292 s) (per-host) I0423 17:07:36.405703 138729557571328 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.409374s. I0423 17:07:41.334238 138729557571328 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 35.337895s. I0423 17:07:41.334548 138733349238528 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:07:41.334666 138733349238528 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:07:41.334715 138733349238528 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:07:41.336155 138733349238528 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0423 17:07:41.336329 138857496962880 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. W0423 17:07:41.336448 138857496962880 checkpoint_manager.py:1458] Waiting for previous save to complete took 15.706855 seconds. If this number is high, consider checkpointing less frequently. I0423 17:07:41.337970 138857496962880 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0423 17:07:41.341533 138857496962880 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/9. I0423 17:07:42.319788 138857496962880 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 17:07:42.319885 138857496962880 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 I0423 17:07:42.361990 138857496962880 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043141s I0423 17:07:42.362153 138857496962880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.123 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7264947891235352 s) (per-host) I0423 17:07:42.362202 138857496962880 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.726552s (batch_requests_ready=0.681366s, total_serialization_initiated=0.045123s, others=0.000062s) I0423 17:07:42.362297 138857496962880 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.730726s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.730714) I0423 17:07:42.363170 138731688269568 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 17:07:42.363330 138857496962880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.025290s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/9. I0423 17:07:42.365284 138857496962880 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 17:07:42.365523 138733349238528 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 17:07:42.365644 138857496962880 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776964045.6295772, 'wait_for_prev_duration_secs': 15.706855297088623, 'time_between_consecutive_saves_sec': 20.239963054656982, 'checkpointer_blocking_start_time': 1776964061.3380096, 'checkpointer_blocking_duration_secs': 1.025430679321289, 'get_old_steps_start_time': 1776964062.3634534, 'get_old_steps_duration_secs': 1.9550323486328125e-05, 'checkpoint_manager_blocking_start_time': 1776964045.6279032, 'checkpoint_manager_blocking_duration_secs': 16.737709045410156} I0423 17:07:42.365750 138857496962880 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 17:07:42.365793 138857496962880 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 17:07:47.488057 138738155898624 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6 I0423 17:07:53.417191 138731688269568 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 134.074 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.781492233276367 s) (per-host) I0423 17:07:53.417296 138731688269568 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.054015s. I0423 17:07:58.658773 138731688269568 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.295476s. I0423 17:07:58.659023 138733349238528 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 17:07:58.659079 138733349238528 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 17:07:58.659129 138733349238528 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 17:07:58.660537 138733349238528 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 17:07:58.660729 138857496962880 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. I0423 17:07:58.660884 138857496962880 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 17:07:58.661707 138857496962880 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.101, Tokens/s/device: 13882.393, 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: Thu Apr 23 17:08:08 UTC 2026 EXIT_CODE=0