XPK Start: Sun Apr 19 01:09:40 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 01:10:04.863046: 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 01:10:05.041539 135010788489024 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 01:10:14.083085 135010788489024 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 01:10:14.085403 135010788489024 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-7zqmc-slice-job-0-0.mt-16-zarr3-true-rs-7zqmc:8482 I0419 01:10:15.692782 135010788489024 max_utils.py:284] Jax distributed system initialized! I0419 01:10:21.904133 135010788489024 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 01:10:21.904237 135010788489024 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 01:10:21.904278 135010788489024 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 01:10:21.904313 135010788489024 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 01:10:22.591049 135010788489024 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 01:10:22.591629 135010788489024 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 01:10:22.591817 135010788489024 checkpointing.py:688] Setting up checkpoint logger... I0419 01:10:22.591866 135010788489024 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 01:10:22.591907 135010788489024 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 01:10:22.592232 135010788489024 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 0x7ac9ee650290>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 01:10:24.343408 135010788489024 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 01:10:24.343645 135010788489024 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 0x7ab574591a60>}, handler_registry=None I0419 01:10:24.343892 135010788489024 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab574591a60>` 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 01:10:24.343940 135010788489024 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab574590d70>` 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 01:10:24.343976 135010788489024 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 0x7ab574591a60>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab574591a60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab574590d70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ab574590d70>}). I0419 01:10:24.344286 135010788489024 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 01:10:24.344356 135010788489024 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ab5741cb740> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 01:10:27.281619 134889647089408 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776560806693760557, 'commit_timestamp_nsecs': 1776560854262009590, '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_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA I0419 01:10:27.526737 134889647089408 checkpoint_manager.py:1871] Missing metrics for step 5 I0419 01:10:27.972214 135010788489024 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_16_zarr3_true/checkpoints I0419 01:10:28.522316 135010788489024 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_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ab574591f70> I0419 01:10:28.522487 135010788489024 checkpointing.py:302] Checkpoint manager created! I0419 01:10:29.435702 135010788489024 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 01:10:29.435998 135010788489024 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:10:29.816960 135010788489024 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:10:29.817047 135010788489024 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:10:29.833233 135010788489024 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 01:10:29.833293 135010788489024 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 01:10:29.857138 135010788489024 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 01:10:29.857208 135010788489024 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:10:29.873670 135010788489024 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 01:10:29.873731 135010788489024 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:10:29.890003 135010788489024 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 01:10:29.890063 135010788489024 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:10:29.914617 135010788489024 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 01:10:29.914690 135010788489024 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 01:10:29.935603 135010788489024 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 01:10:29.935666 135010788489024 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 01:10:30.152048 135010788489024 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 01:10:30.152158 135010788489024 checkpointing.py:582] restoring from this run's directory step 9 I0419 01:10:30.826844 135010788489024 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_16_zarr3_true/checkpoints/9. I0419 01:10:31.070566 135010788489024 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_16_zarr3_true/checkpoints/9. I0419 01:10:31.071013 135010788489024 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.24 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_16_zarr3_true/checkpoints/9. I0419 01:10:32.465848 1453 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 01:10:45.739971 135010788489024 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 118.994 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.280232429504395 s) (per-host) I0419 01:10:45.745725 135010788489024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 115.670 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.68195915222168 s) (per-host) I0419 01:10:45.747878 135010788489024 event_tracking.py:138] [process=0] [sync] Finished load in 14.92 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_16_zarr3_true/checkpoints/9 I0419 01:10:45.747974 135010788489024 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_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776561030.8267334, 'checkpointer_duration_secs': 14.921212196350098, 'checkpoint_manager_start_time': 1776561030.1524127, 'checkpoint_manager_duration_secs': 15.595535278320312} fsdp: 32 I0419 01:10:46.161142 135010788489024 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 01:10:46.161267 135010788489024 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 01:10:46.161320 135010788489024 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 01:10:46.161379 135010788489024 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 01:10:46.161431 135010788489024 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 01:10:46.161493 135010788489024 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 01:10:46.161549 135010788489024 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 01:10:46.161619 135010788489024 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 01:10:46.161668 135010788489024 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 01:10:46.161705 135010788489024 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 01:10:46.161753 135010788489024 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 01:10:46.161798 135010788489024 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 01:10:46.648718 135010788489024 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 01:10:46.648820 135010788489024 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 01:10:46.664069 135010788489024 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 01:10:46.664131 135010788489024 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 01:10:57.579159 135010788489024 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 01:10:57.587241 135010788489024 metric_logger.py:301] number parameters: 1.104 billion I0419 01:11:09.009692 135010788489024 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 01:11:09.163541 135010788489024 checkpointing.py:798] Waited 0.15383052825927734 seconds for step 10 to finish before starting checkpointing. I0419 01:11:09.166027 135010788489024 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:11:09.167747 135010788489024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0419 01:11:09.169232 135010788489024 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_16_zarr3_true/checkpoints/10. I0419 01:11:09.895586 135010788489024 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 01:11:09.908286 135010788489024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 01:11:09.908356 135010788489024 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 01:11:10.181900 134885293405952 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_16_zarr3_true/checkpoints/10 I0419 01:11:10.183320 135010788489024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275934s I0419 01:11:10.184098 135010788489024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.498 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2805767059326172 s) (per-host) I0419 01:11:10.184161 135010788489024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.280656s (batch_requests_ready=0.002140s, total_serialization_initiated=0.277832s, others=0.000684s) I0419 01:11:10.184237 135010788489024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.284630s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.284612) I0419 01:11:10.184961 135010788489024 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 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_16_zarr3_true/checkpoints/10. I0419 01:11:10.185301 134883675096832 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:31:10.185259 I0419 01:11:10.196318 135010788489024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:11:10.196578 134885301798656 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:11:10.196696 135010788489024 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776561069.1660075, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776561069.1677837, 'checkpointer_blocking_duration_secs': 1.0176734924316406, 'get_old_steps_start_time': 1776561070.1854813, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776561069.1640303, 'checkpoint_manager_blocking_duration_secs': 1.0326306819915771} I0419 01:11:10.196825 135010788489024 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 01:11:10.196877 135010788489024 max_utils.py:750] Memstats: After params initialized: I0419 01:11:10.196928 135010788489024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 01:11:10.196961 135010788489024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 01:11:10.196988 135010788489024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 01:11:10.197013 135010788489024 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 01:11:10.539776 135010788489024 metric_logger.py:196] completed step: 10, seconds: 11.113, TFLOP/s/device: 1.223, Tokens/s/device: 184.288, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0419 01:11:10.685197 135010788489024 metric_logger.py:196] completed step: 11, seconds: 1.528, TFLOP/s/device: 8.890, Tokens/s/device: 1340.048, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0419 01:11:10.832769 135010788489024 metric_logger.py:196] completed step: 12, seconds: 0.010, TFLOP/s/device: 1295.740, Tokens/s/device: 195308.030, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0419 01:11:10.980120 135010788489024 metric_logger.py:196] completed step: 13, seconds: 0.146, TFLOP/s/device: 93.310, Tokens/s/device: 14064.678, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0419 01:11:10.988910 135010788489024 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 01:11:11.274266 135010788489024 checkpointing.py:798] Waited 0.28521251678466797 seconds for step 15 to finish before starting checkpointing. I0419 01:11:11.277178 135010788489024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:11:12.392598 134884775601920 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_16_zarr3_true/checkpoints/10/items I0419 01:11:12.957877 134883666704128 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776561072140281897, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776561072140281897, "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_16_zarr3_true/checkpoints/10/_CHECKPOINT_METADATA I0419 01:11:15.239383 134884745086720 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_16_zarr3_true/checkpoints/10/items/array_metadatas/process_0 I0419 01:11:55.637464 134884736694016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.144766s (commit=40.955538s, array_metadata_write=1.189228s) I0419 01:11:55.638837 134883675096832 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.538 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.735284090042114 s) (per-host) I0419 01:11:55.638950 134883675096832 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.453532s. I0419 01:11:57.519585 134883666704128 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776561072140281897, '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_16_zarr3_true/checkpoints/10/_CHECKPOINT_METADATA I0419 01:11:57.773865 134883675096832 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000408s. I0419 01:11:58.162611 134883675096832 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 01:11:59.084527 134883675096832 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.175472s. 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_16_zarr3_true/checkpoints/10/items I0419 01:11:59.086685 134883675096832 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_16_zarr3_true/checkpoints/10/items I0419 01:11:59.618183 134883675096832 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_16_zarr3_true/checkpoints/10 I0419 01:12:00.984762 134883675096832 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_16_zarr3_true/checkpoints/10`. I0419 01:12:00.985913 134883675096832 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.82 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_16_zarr3_true/checkpoints/10 I0419 01:12:00.987677 134883675096832 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.802255s. I0419 01:12:00.987912 134885301798656 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:12:00.988037 134885301798656 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:12:00.988150 134885301798656 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:12:00.989927 134885301798656 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 01:12:00.990239 135010788489024 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 01:12:00.990440 135010788489024 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.713251 seconds. If this number is high, consider checkpointing less frequently. I0419 01:12:00.992184 135010788489024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0419 01:12:00.994586 135010788489024 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_16_zarr3_true/checkpoints/15. I0419 01:12:02.132090 135010788489024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 01:12:02.132224 135010788489024 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 01:12:02.177696 135010788489024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046551s I0419 01:12:02.178778 135010788489024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 27.559 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05597352981567383 s) (per-host) I0419 01:12:02.178899 135010788489024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.056120s (batch_requests_ready=0.004328s, total_serialization_initiated=0.050821s, others=0.000970s) I0419 01:12:02.179031 135010788489024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.060903s (all_items=0.000022s, per_item={'items': '0.00002193'}, temp_paths=0.060881) I0419 01:12:02.180056 135010788489024 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.19 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_16_zarr3_true/checkpoints/15. I0419 01:12:02.180429 134885301798656 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:32:02.180400 I0419 01:12:02.182854 135010788489024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:12:02.183256 134884745086720 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:12:02.183512 135010788489024 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776561071.2771347, 'wait_for_prev_duration_secs': 49.7132511138916, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776561120.992254, 'checkpointer_blocking_duration_secs': 1.1883149147033691, 'get_old_steps_start_time': 1776561122.1805995, 'get_old_steps_duration_secs': 4.363059997558594e-05, 'checkpoint_manager_blocking_start_time': 1776561071.2749264, 'checkpoint_manager_blocking_duration_secs': 50.9085259437561} I0419 01:12:02.183678 135010788489024 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 01:12:02.382175 134883675096832 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_16_zarr3_true/checkpoints/15 I0419 01:12:04.577293 134884767209216 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_16_zarr3_true/checkpoints/15/items I0419 01:12:08.063489 134884783994624 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_16_zarr3_true/checkpoints/15/items/array_metadatas/process_0 I0419 01:12:23.739838 135010788489024 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.206, Tokens/s/device: 13898.220, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0419 01:12:23.748737 135010788489024 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.230, Tokens/s/device: 13901.899, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0419 01:12:23.894517 135010788489024 metric_logger.py:196] completed step: 16, seconds: 72.758, TFLOP/s/device: 0.187, Tokens/s/device: 28.148, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0419 01:12:24.041705 135010788489024 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1823.042, Tokens/s/device: 274788.676, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0419 01:12:24.189173 135010788489024 metric_logger.py:196] completed step: 18, seconds: 0.149, TFLOP/s/device: 91.285, Tokens/s/device: 13759.465, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0419 01:12:24.335761 135010788489024 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 01:12:24.336365 135010788489024 checkpointing.py:798] Waited 0.0006191730499267578 seconds for step 19 to finish before starting checkpointing. I0419 01:12:24.338707 135010788489024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:12:47.117129 134884775601920 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.237166s (commit=40.004460s, array_metadata_write=1.232706s) I0419 01:12:47.118140 134885301798656 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.106 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.995315074920654 s) (per-host) I0419 01:12:47.118283 134885301798656 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.937745s. I0419 01:12:49.253120 134885301798656 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 01:12:49.429269 134885301798656 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000247s. I0419 01:12:50.178200 134885301798656 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.179626s. 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_16_zarr3_true/checkpoints/15/items I0419 01:12:50.179907 134885301798656 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_16_zarr3_true/checkpoints/15/items I0419 01:12:50.699695 134885301798656 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_16_zarr3_true/checkpoints/15 I0419 01:12:52.127765 134885301798656 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_16_zarr3_true/checkpoints/15`. I0419 01:12:52.128524 134885301798656 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.14 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_16_zarr3_true/checkpoints/15 I0419 01:12:52.130039 134885301798656 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.949500s. I0419 01:12:52.130251 134884745086720 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:12:52.130368 134884745086720 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:12:52.130441 134884745086720 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:12:52.132136 134884745086720 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 01:12:52.132308 135010788489024 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 01:12:52.132436 135010788489024 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.793753 seconds. If this number is high, consider checkpointing less frequently. I0419 01:12:52.133926 135010788489024 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0419 01:12:52.136065 135010788489024 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_16_zarr3_true/checkpoints/19. I0419 01:12:52.842004 135010788489024 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 01:12:52.842096 135010788489024 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 01:12:52.867423 135010788489024 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026375s I0419 01:12:52.868289 135010788489024 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.960 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03087639808654785 s) (per-host) I0419 01:12:52.868372 135010788489024 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030982s (batch_requests_ready=0.001886s, total_serialization_initiated=0.028308s, others=0.000787s) I0419 01:12:52.868463 135010788489024 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034978s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.034962) I0419 01:12:52.869162 135010788489024 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 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_16_zarr3_true/checkpoints/19. I0419 01:12:52.869464 134884775601920 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 01:32:52.869425 I0419 01:12:52.871865 135010788489024 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 01:12:52.872142 134884745086720 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 01:12:52.872311 135010788489024 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776561144.3386543, 'wait_for_prev_duration_secs': 27.793753385543823, 'time_between_consecutive_saves_sec': 23.348663806915283, 'checkpointer_blocking_start_time': 1776561172.1339664, 'checkpointer_blocking_duration_secs': 0.735630989074707, 'get_old_steps_start_time': 1776561172.869616, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776561144.3366, 'checkpoint_manager_blocking_duration_secs': 28.535676956176758} I0419 01:12:52.872427 135010788489024 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 01:12:52.872472 135010788489024 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 01:12:53.078927 134885301798656 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_16_zarr3_true/checkpoints/19 I0419 01:12:55.308480 134884783994624 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_16_zarr3_true/checkpoints/19/items I0419 01:12:58.562852 134887956788992 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_16_zarr3_true/checkpoints/19/items/array_metadatas/process_0 I0419 01:13:39.099241 134885293405952 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.236620s (commit=41.006488s, array_metadata_write=1.230133s) I0419 01:13:39.100376 134884775601920 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.144 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.26293587684631 s) (per-host) I0419 01:13:39.100509 134884775601920 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.230934s. I0419 01:13:40.850920 134884775601920 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 01:13:40.988714 134884775601920 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000237s. I0419 01:13:41.904520 134884775601920 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.920830s. 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_16_zarr3_true/checkpoints/19/items I0419 01:13:41.906034 134884775601920 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_16_zarr3_true/checkpoints/19/items I0419 01:13:42.451366 134884775601920 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_16_zarr3_true/checkpoints/19 I0419 01:13:43.879602 134884775601920 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_16_zarr3_true/checkpoints/19`. I0419 01:13:43.880414 134884775601920 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.75 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_16_zarr3_true/checkpoints/19 I0419 01:13:43.881726 134884775601920 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.012149s. I0419 01:13:43.881904 134884745086720 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 01:13:43.881981 134884745086720 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 01:13:43.882042 134884745086720 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 01:13:43.883834 134884745086720 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 01:13:43.883956 135010788489024 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 01:13:43.884051 135010788489024 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 01:13:43.884917 135010788489024 metric_logger.py:196] completed step: 19, seconds: 0.145, TFLOP/s/device: 93.598, Tokens/s/device: 14108.084, 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 01:13:55 UTC 2026 EXIT_CODE=0