XPK Start: Sun Apr 19 00:52:16 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 00:52:40.705241: 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) I0419 00:52:40.885364 138839938856768 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 00:52:49.926100 138839938856768 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 00:52:49.928436 138839938856768 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-inoke-slice-job-0-0.mt-15-ocdbt-true-rs-inoke:8482 I0419 00:52:51.634078 138839938856768 max_utils.py:284] Jax distributed system initialized! I0419 00:52:57.666344 138839938856768 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 00:52:57.666445 138839938856768 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 00:52:57.666486 138839938856768 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0419 00:52:57.666522 138839938856768 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 00:52:58.709803 138839938856768 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:52:58.710399 138839938856768 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:52:58.710578 138839938856768 checkpointing.py:688] Setting up checkpoint logger... I0419 00:52:58.710627 138839938856768 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 00:52:58.710669 138839938856768 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 00:52:58.711015 138839938856768 base_pytree_checkpoint_handler.py:441] 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 0x7e4595d1cf80>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 00:53:00.775913 138839938856768 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 00:53:00.776146 138839938856768 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e446dc153d0>}, handler_registry=None I0419 00:53:00.776416 138839938856768 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e446dc153d0>` 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`. I0419 00:53:00.776470 138839938856768 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e446dc14e90>` 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`. I0419 00:53:00.776520 138839938856768 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 0x7e446dc153d0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e446dc153d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e446dc14e90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e446dc14e90>}). I0419 00:53:00.776896 138839938856768 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 00:53:00.776978 138839938856768 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e302478ac00> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 00:53:03.547272 138718107133696 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776559705404793119, 'commit_timestamp_nsecs': 1776559751829371193, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0419 00:53:03.794484 138718107133696 checkpoint_manager.py:1871] Missing metrics for step 0 I0419 00:53:04.608693 138839938856768 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints I0419 00:53:05.928735 138839938856768 checkpoint_manager.py:929] [process=0][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_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e446dcccc50> I0419 00:53:05.928911 138839938856768 checkpointing.py:302] Checkpoint manager created! I0419 00:53:06.855024 138839938856768 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 00:53:06.855127 138839938856768 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:53:07.234426 138839938856768 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:53:07.234514 138839938856768 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:53:07.250856 138839938856768 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:53:07.250915 138839938856768 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:53:07.274577 138839938856768 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:53:07.274645 138839938856768 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:53:07.290909 138839938856768 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:53:07.290970 138839938856768 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:53:07.307210 138839938856768 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:53:07.307278 138839938856768 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:53:07.332707 138839938856768 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 00:53:07.332774 138839938856768 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:53:07.353441 138839938856768 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 00:53:07.353507 138839938856768 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 00:53:07.568796 138839938856768 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 00:53:07.568919 138839938856768 checkpointing.py:582] restoring from this run's directory step 9 I0419 00:53:07.813477 138839938856768 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9. I0419 00:53:08.063671 138839938856768 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9. I0419 00:53:08.064111 138839938856768 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9. I0419 00:53:10.869600 1458 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 00:53:16.624381 138839938856768 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 274.310 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.760903358459473 s) (per-host) I0419 00:53:16.630230 138839938856768 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 257.168 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.1539411544799805 s) (per-host) I0419 00:53:19.135307 138839938856768 event_tracking.py:138] [process=0] [sync] Finished load in 11.32 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/9 I0419 00:53:19.135473 138839938856768 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776559987.8133733, 'checkpointer_duration_secs': 11.32206678390503, 'checkpoint_manager_start_time': 1776559987.5691683, 'checkpoint_manager_duration_secs': 11.566274166107178} fsdp: 32 I0419 00:53:19.548178 138839938856768 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 00:53:19.548316 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:53:19.548370 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:53:19.548426 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 00:53:19.548485 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:53:19.548523 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:53:19.548576 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 00:53:19.548629 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0419 00:53:19.548671 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0419 00:53:19.548709 138839938856768 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 00:53:19.548758 138839938856768 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 00:53:19.548817 138839938856768 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 00:53:20.037529 138839938856768 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:53:20.037623 138839938856768 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:53:20.052957 138839938856768 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:53:20.053015 138839938856768 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:53:31.011417 138839938856768 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. I0419 00:53:31.019155 138839938856768 metric_logger.py:301] number parameters: 1.104 billion I0419 00:53:42.755516 138839938856768 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 00:53:42.909696 138839938856768 checkpointing.py:798] Waited 0.154160737991333 seconds for step 10 to finish before starting checkpointing. I0419 00:53:42.912068 138839938856768 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:53:42.913763 138839938856768 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0419 00:53:42.915211 138839938856768 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10. I0419 00:53:43.647415 138839938856768 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 00:53:43.660656 138839938856768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:53:43.660771 138839938856768 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 I0419 00:53:43.918495 138717486368512 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10 I0419 00:53:43.940938 138839938856768 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281364s I0419 00:53:43.941732 138839938856768 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.393 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2860085964202881 s) (per-host) I0419 00:53:43.941817 138839938856768 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286104s (batch_requests_ready=0.002150s, total_serialization_initiated=0.283237s, others=0.000718s) I0419 00:53:43.941896 138839938856768 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290304s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.290286) I0419 00:53:43.942657 138839938856768 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10. I0419 00:53:43.942952 138717494761216 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:13:43.942920 I0419 00:53:43.948911 138839938856768 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:53:43.949162 138713216579328 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:53:43.949283 138839938856768 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776560022.9120488, 'wait_for_prev_duration_secs': 6.008148193359375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776560022.913801, 'checkpointer_blocking_duration_secs': 1.029296636581421, 'get_old_steps_start_time': 1776560023.9431233, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776560022.9102013, 'checkpoint_manager_blocking_duration_secs': 1.03904390335083} I0419 00:53:43.949393 138839938856768 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 00:53:43.949447 138839938856768 max_utils.py:750] Memstats: After params initialized: I0419 00:53:43.949496 138839938856768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 00:53:43.949527 138839938856768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 00:53:43.949553 138839938856768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 00:53:43.949578 138839938856768 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 00:53:44.298717 138839938856768 metric_logger.py:196] completed step: 10, seconds: 11.415, TFLOP/s/device: 1.190, Tokens/s/device: 179.407, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0419 00:53:44.444293 138839938856768 metric_logger.py:196] completed step: 11, seconds: 1.541, TFLOP/s/device: 8.815, Tokens/s/device: 1328.754, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0419 00:53:44.591550 138839938856768 metric_logger.py:196] completed step: 12, seconds: 0.010, TFLOP/s/device: 1334.557, Tokens/s/device: 201159.022, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0419 00:53:44.739152 138839938856768 metric_logger.py:196] completed step: 13, seconds: 0.146, TFLOP/s/device: 93.262, Tokens/s/device: 14057.438, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0419 00:53:44.747525 138839938856768 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 00:53:45.033127 138839938856768 checkpointing.py:798] Waited 0.2854588031768799 seconds for step 15 to finish before starting checkpointing. I0419 00:53:45.036014 138839938856768 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:53:47.077780 138713208186624 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776560026685341920, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776560026685341920, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0419 00:53:47.358460 138716437804800 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/items I0419 00:53:50.687922 138713705334528 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_0 I0419 00:54:30.633145 138713696941824 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.159391s (commit=40.960546s, array_metadata_write=1.198844s) I0419 00:54:30.634657 138717494761216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.624 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.97889542579651 s) (per-host) I0419 00:54:30.634853 138717494761216 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.691787s. I0419 00:54:32.361169 138717494761216 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:54:32.480843 138713208186624 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776560026685341920, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0419 00:54:32.487199 138717494761216 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000408s. I0419 00:54:33.373391 138717494761216 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.901183s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/items I0419 00:54:33.375546 138717494761216 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10/items I0419 00:54:33.879206 138717494761216 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10 I0419 00:54:36.167331 138717494761216 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10`. I0419 00:54:36.168427 138717494761216 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.25 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/10 I0419 00:54:36.169859 138717494761216 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.226792s. I0419 00:54:36.170167 138713216579328 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:54:36.170346 138713216579328 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:54:36.170471 138713216579328 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:54:36.172945 138713216579328 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 00:54:36.173270 138839938856768 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0419 00:54:36.173452 138839938856768 checkpoint_manager.py:1452] Waiting for previous save to complete took 51.137435 seconds. If this number is high, consider checkpointing less frequently. I0419 00:54:36.175289 138839938856768 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0419 00:54:36.177713 138839938856768 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15. I0419 00:54:37.325631 138839938856768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:54:37.325764 138839938856768 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 I0419 00:54:37.374508 138839938856768 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049912s I0419 00:54:37.375555 138839938856768 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 26.228 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05881357192993164 s) (per-host) I0419 00:54:37.375668 138839938856768 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.058951s (batch_requests_ready=0.004308s, total_serialization_initiated=0.053694s, others=0.000949s) I0419 00:54:37.375801 138839938856768 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.063863s (all_items=0.000023s, per_item={'items': '0.00002265'}, temp_paths=0.063841) I0419 00:54:37.376868 138839938856768 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.20 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15. I0419 00:54:37.377308 138713216579328 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:14:37.377265 I0419 00:54:37.379775 138839938856768 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:54:37.380195 138713705334528 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:54:37.380422 138839938856768 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776560025.0359714, 'wait_for_prev_duration_secs': 51.13743543624878, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776560076.1753528, 'checkpointer_blocking_duration_secs': 1.2021455764770508, 'get_old_steps_start_time': 1776560077.3775284, 'get_old_steps_duration_secs': 4.220008850097656e-05, 'checkpoint_manager_blocking_start_time': 1776560025.0337737, 'checkpoint_manager_blocking_duration_secs': 52.346596002578735} I0419 00:54:37.380566 138839938856768 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 00:54:38.005970 138717494761216 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15 I0419 00:54:41.039898 138713713727232 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15/items I0419 00:54:45.132672 138716924319488 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_0 I0419 00:54:59.486415 138839938856768 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.219, Tokens/s/device: 13900.295, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0419 00:54:59.496393 138839938856768 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.189, Tokens/s/device: 13895.768, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0419 00:54:59.641759 138839938856768 metric_logger.py:196] completed step: 16, seconds: 74.747, TFLOP/s/device: 0.182, Tokens/s/device: 27.399, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0419 00:54:59.789610 138839938856768 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1816.461, Tokens/s/device: 273796.791, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0419 00:54:59.936715 138839938856768 metric_logger.py:196] completed step: 18, seconds: 0.146, TFLOP/s/device: 92.781, Tokens/s/device: 13984.963, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0419 00:55:00.083439 138839938856768 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 00:55:00.084365 138839938856768 checkpointing.py:798] Waited 0.0008504390716552734 seconds for step 19 to finish before starting checkpointing. I0419 00:55:00.086618 138839938856768 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:55:24.696858 138716437804800 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.134503s (commit=40.507276s, array_metadata_write=1.627227s) I0419 00:55:24.698081 138713216579328 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.338 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.381324768066406 s) (per-host) I0419 00:55:24.698219 138713216579328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.320763s. I0419 00:55:27.014224 138713216579328 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:55:27.395005 138713216579328 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000255s. I0419 00:55:28.031591 138713216579328 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.041052s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15/items I0419 00:55:28.036274 138713216579328 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15/items I0419 00:55:28.990956 138713216579328 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15 I0419 00:55:30.369340 138713216579328 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15`. I0419 00:55:30.370209 138713216579328 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 54.19 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/15 I0419 00:55:30.371781 138713216579328 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.994325s. I0419 00:55:30.372034 138713705334528 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:55:30.372189 138713705334528 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:55:30.372282 138713705334528 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:55:30.373996 138713705334528 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 00:55:30.374120 138839938856768 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 00:55:30.374213 138839938856768 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.287604 seconds. If this number is high, consider checkpointing less frequently. I0419 00:55:30.375766 138839938856768 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0419 00:55:30.378005 138839938856768 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19. I0419 00:55:31.516478 138839938856768 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:55:31.516572 138839938856768 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 I0419 00:55:31.541530 138839938856768 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025775s I0419 00:55:31.542361 138839938856768 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.622 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030472278594970703 s) (per-host) I0419 00:55:31.542438 138839938856768 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030573s (batch_requests_ready=0.001970s, total_serialization_initiated=0.027827s, others=0.000776s) I0419 00:55:31.542531 138839938856768 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034548s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.034532) I0419 00:55:31.543287 138839938856768 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19. I0419 00:55:31.543619 138713705334528 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:15:31.543584 I0419 00:55:31.545932 138839938856768 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:55:31.546210 138713713727232 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:55:31.546372 138839938856768 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776560100.0865848, 'wait_for_prev_duration_secs': 30.287603855133057, 'time_between_consecutive_saves_sec': 23.913569927215576, 'checkpointer_blocking_start_time': 1776560130.375818, 'checkpointer_blocking_duration_secs': 1.1679344177246094, 'get_old_steps_start_time': 1776560131.5437725, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776560100.0848305, 'checkpoint_manager_blocking_duration_secs': 31.46150779724121} I0419 00:55:31.546483 138839938856768 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 00:55:31.546528 138839938856768 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:55:31.750621 138713216579328 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19 I0419 00:55:33.905858 138717494761216 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19/items I0419 00:55:37.230105 138716924319488 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_0 I0419 00:56:16.768215 138716437804800 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.681060s (commit=40.002946s, array_metadata_write=1.678114s) I0419 00:56:16.769196 138713705334528 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.903 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.257309436798096 s) (per-host) I0419 00:56:16.769256 138713705334528 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.225528s. I0419 00:56:18.408659 138713705334528 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:56:18.587315 138713705334528 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000237s. I0419 00:56:19.019007 138713705334528 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.438522s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19/items I0419 00:56:19.020649 138713705334528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19/items I0419 00:56:19.520238 138713705334528 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19 I0419 00:56:20.931497 138713705334528 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19`. I0419 00:56:20.932359 138713705334528 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.56 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_true/checkpoints/19 I0419 00:56:20.934648 138713705334528 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.390917s. I0419 00:56:20.934824 138713713727232 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:56:20.934890 138713713727232 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:56:20.934955 138713713727232 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:56:20.937602 138713713727232 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 00:56:20.937773 138839938856768 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0419 00:56:20.937935 138839938856768 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:56:20.938873 138839938856768 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.924, Tokens/s/device: 13855.719, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 00:56:32 UTC 2026 EXIT_CODE=0