XPK Start: Mon Apr 20 17:02:21 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 17:02:45.762136: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 17:02:45.942435 135758522918720 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 17:02:54.983600 135758522918720 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 17:02:54.986165 135758522918720 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-qj4tq-slice-job-0-0.mt-15-ocdbt-true-rs-qj4tq:8482 I0420 17:02:56.391419 135758522918720 max_utils.py:284] Jax distributed system initialized! I0420 17:03:01.563190 135758522918720 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 17:03:01.563293 135758522918720 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 17:03:01.563334 135758522918720 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 I0420 17:03:01.563369 135758522918720 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 17:03:02.605303 135758522918720 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 17:03:02.605880 135758522918720 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 17:03:02.606055 135758522918720 checkpointing.py:688] Setting up checkpoint logger... I0420 17:03:02.606119 135758522918720 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 17:03:02.606166 135758522918720 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 17:03:02.606494 135758522918720 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 0x7b782d1144d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 17:03:03.955471 135758522918720 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 17:03:03.955706 135758522918720 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 0x7b76fa8f4860>}, handler_registry=None I0420 17:03:03.955944 135758522918720 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b76fa8f4860>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 17:03:03.955991 135758522918720 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b76fa96bd10>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 17:03:03.956026 135758522918720 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 0x7b76fa8f4860>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b76fa8f4860>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b76fa96bd10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b76fa96bd10>}). I0420 17:03:03.956371 135758522918720 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 17:03:03.956442 135758522918720 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b64787e2c00> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 17:03:05.801523 135639949367040 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776704417644190895, 'commit_timestamp_nsecs': 1776704466077287440, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/9/_CHECKPOINT_METADATA I0420 17:03:06.060660 135639949367040 checkpoint_manager.py:1871] Missing metrics for step 9 I0420 17:03:06.507550 135758522918720 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints I0420 17:03:07.218849 135758522918720 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b76fa773e30> I0420 17:03:07.219025 135758522918720 checkpointing.py:302] Checkpoint manager created! I0420 17:03:08.128804 135758522918720 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 17:03:08.128914 135758522918720 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:03:08.510363 135758522918720 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 17:03:08.510454 135758522918720 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:03:08.526536 135758522918720 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 17:03:08.526592 135758522918720 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 17:03:08.549951 135758522918720 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 17:03:08.550018 135758522918720 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:03:08.566245 135758522918720 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 17:03:08.566307 135758522918720 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:03:08.582510 135758522918720 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 17:03:08.582570 135758522918720 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:03:08.607670 135758522918720 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 17:03:08.607741 135758522918720 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 17:03:08.628461 135758522918720 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 17:03:08.628530 135758522918720 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 17:03:08.842468 135758522918720 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 17:03:08.842571 135758522918720 checkpointing.py:582] restoring from this run's directory step 9 I0420 17:03:09.098541 135758522918720 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/9. I0420 17:03:09.351647 135758522918720 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/9. I0420 17:03:09.352070 135758522918720 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/9. I0420 17:03:10.843074 1415 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 17:03:17.041230 135758522918720 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 254.719 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.20399808883667 s) (per-host) I0420 17:03:17.046224 135758522918720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 237.722 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.657335519790649 s) (per-host) I0420 17:03:18.876724 135758522918720 event_tracking.py:138] [process=0] [sync] Finished load in 9.78 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/9 I0420 17:03:18.876886 135758522918720 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776704589.0984433, 'checkpointer_duration_secs': 9.778404235839844, 'checkpoint_manager_start_time': 1776704588.8428094, 'checkpoint_manager_duration_secs': 10.034040212631226} fsdp: 32 I0420 17:03:19.291106 135758522918720 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 17:03:19.291233 135758522918720 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) I0420 17:03:19.291285 135758522918720 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) I0420 17:03:19.291343 135758522918720 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') I0420 17:03:19.291395 135758522918720 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) I0420 17:03:19.291433 135758522918720 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) I0420 17:03:19.291487 135758522918720 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) I0420 17:03:19.291541 135758522918720 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') I0420 17:03:19.291580 135758522918720 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) I0420 17:03:19.291617 135758522918720 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) I0420 17:03:19.291666 135758522918720 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 17:03:19.291712 135758522918720 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 17:03:19.776674 135758522918720 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 17:03:19.776764 135758522918720 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 17:03:19.792191 135758522918720 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 17:03:19.792254 135758522918720 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 17:03:30.706095 135758522918720 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 17:03:30.713877 135758522918720 metric_logger.py:301] number parameters: 1.104 billion I0420 17:03:42.173332 135758522918720 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 17:03:42.327173 135758522918720 checkpointing.py:798] Waited 0.1538240909576416 seconds for step 10 to finish before starting checkpointing. I0420 17:03:42.329634 135758522918720 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 17:03:42.331428 135758522918720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0420 17:03:42.332829 135758522918720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10. I0420 17:03:43.101872 135758522918720 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 17:03:43.114809 135758522918720 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:03:43.114875 135758522918720 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 17:03:43.384963 135758522918720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271121s I0420 17:03:43.385888 135758522918720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.595 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27570056915283203 s) (per-host) I0420 17:03:43.385970 135758522918720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.275804s (batch_requests_ready=0.001966s, total_serialization_initiated=0.272995s, others=0.000842s) I0420 17:03:43.386051 135758522918720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.280088s (all_items=0.000032s, per_item={'items': '0.00003171'}, temp_paths=0.280056) I0420 17:03:43.386759 135758522918720 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10. I0420 17:03:43.387058 135639160846080 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:23:43.387035 I0420 17:03:43.405147 135758522918720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:03:43.405411 135635297883904 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:03:43.405534 135758522918720 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704622.3296149, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776704622.3314667, 'checkpointer_blocking_duration_secs': 1.0556728839874268, 'get_old_steps_start_time': 1776704623.3871636, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776704622.3276634, 'checkpoint_manager_blocking_duration_secs': 1.077834129333496} I0420 17:03:43.405648 135758522918720 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 17:03:43.405698 135758522918720 max_utils.py:750] Memstats: After params initialized: I0420 17:03:43.405747 135758522918720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0420 17:03:43.405779 135758522918720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0420 17:03:43.405804 135758522918720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0420 17:03:43.405828 135758522918720 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0420 17:03:43.770284 135758522918720 metric_logger.py:196] completed step: 10, seconds: 11.151, TFLOP/s/device: 1.218, Tokens/s/device: 183.657, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0420 17:03:43.834179 135635872511744 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10 I0420 17:03:43.915366 135758522918720 metric_logger.py:196] completed step: 11, seconds: 1.595, TFLOP/s/device: 8.518, Tokens/s/device: 1283.993, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0420 17:03:44.063156 135758522918720 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1231.499, Tokens/s/device: 185624.943, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0420 17:03:44.210645 135758522918720 metric_logger.py:196] completed step: 13, seconds: 0.145, TFLOP/s/device: 93.628, Tokens/s/device: 14112.653, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0420 17:03:44.219782 135758522918720 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0420 17:03:44.504990 135758522918720 checkpointing.py:798] Waited 0.28505444526672363 seconds for step 15 to finish before starting checkpointing. I0420 17:03:44.508000 135758522918720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:03:45.688000 135635841955584 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/items I0420 17:03:45.847916 135635289491200 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776704625435448657, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776704625435448657, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0420 17:03:49.118051 135635825170176 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_0 I0420 17:04:29.178828 135635315025664 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.875293s (commit=40.211766s, array_metadata_write=1.663527s) I0420 17:04:29.180203 135639160846080 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.287 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.06997561454773 s) (per-host) I0420 17:04:29.180371 135639160846080 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.792708s. I0420 17:04:31.219572 135635289491200 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776704625435448657, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA I0420 17:04:31.616307 135639160846080 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000412s. I0420 17:04:31.828703 135639160846080 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 17:04:32.805986 135639160846080 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.243614s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/items I0420 17:04:32.808468 135639160846080 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10/items I0420 17:04:33.387544 135639160846080 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10 I0420 17:04:34.830368 135639160846080 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10`. I0420 17:04:34.831510 135639160846080 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.50 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/10 I0420 17:04:34.833101 135639160846080 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.445420s. I0420 17:04:34.833402 135635297883904 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:04:34.833563 135635297883904 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:04:34.833673 135635297883904 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:04:34.835327 135635297883904 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 17:04:34.835646 135758522918720 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. W0420 17:04:34.835829 135758522918720 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.327828 seconds. If this number is high, consider checkpointing less frequently. I0420 17:04:34.837536 135758522918720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0420 17:04:34.839924 135758522918720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15. I0420 17:04:35.556406 135758522918720 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:04:35.556543 135758522918720 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 17:04:35.587551 135758522918720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.032179s I0420 17:04:35.588547 135758522918720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 37.789 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04082036018371582 s) (per-host) I0420 17:04:35.588645 135758522918720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040941s (batch_requests_ready=0.004028s, total_serialization_initiated=0.036031s, others=0.000881s) I0420 17:04:35.588766 135758522918720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045975s (all_items=0.000022s, per_item={'items': '0.00002241'}, temp_paths=0.045953) I0420 17:04:35.589735 135758522918720 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15. I0420 17:04:35.590075 135635297883904 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:24:35.590047 I0420 17:04:35.592507 135758522918720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:04:35.592842 135635825170176 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:04:35.593072 135758522918720 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704624.507956, 'wait_for_prev_duration_secs': 50.3278284072876, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776704674.8376064, 'checkpointer_blocking_duration_secs': 0.7526133060455322, 'get_old_steps_start_time': 1776704675.5902476, 'get_old_steps_duration_secs': 4.363059997558594e-05, 'checkpoint_manager_blocking_start_time': 1776704624.5056627, 'checkpoint_manager_blocking_duration_secs': 51.08735680580139} I0420 17:04:35.593226 135758522918720 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0420 17:04:36.219478 135639160846080 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15 I0420 17:04:38.031729 135635864119040 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15/items I0420 17:04:41.058145 135635841955584 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_0 I0420 17:04:56.207714 135758522918720 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.069, Tokens/s/device: 13877.689, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0420 17:04:56.216940 135758522918720 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.070, Tokens/s/device: 13877.783, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0420 17:04:56.362322 135758522918720 metric_logger.py:196] completed step: 16, seconds: 71.996, TFLOP/s/device: 0.189, Tokens/s/device: 28.446, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0420 17:04:56.509816 135758522918720 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1872.279, Tokens/s/device: 282210.280, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0420 17:04:56.657186 135758522918720 metric_logger.py:196] completed step: 18, seconds: 0.146, TFLOP/s/device: 92.748, Tokens/s/device: 13979.999, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0420 17:04:56.803770 135758522918720 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 17:04:56.804531 135758522918720 checkpointing.py:798] Waited 0.0007812976837158203 seconds for step 19 to finish before starting checkpointing. I0420 17:04:56.806535 135758522918720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:05:21.988013 135635833562880 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.768070s (commit=41.490635s, array_metadata_write=1.277435s) I0420 17:05:21.989154 135635297883904 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.013 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.44140577316284 s) (per-host) I0420 17:05:21.989283 135635297883904 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.399092s. I0420 17:05:23.602361 135635297883904 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000273s. I0420 17:05:23.795454 135635297883904 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 17:05:24.916870 135635297883904 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.003859s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15/items I0420 17:05:24.918568 135635297883904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15/items I0420 17:05:25.494566 135635297883904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15 I0420 17:05:27.027290 135635297883904 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15`. I0420 17:05:27.027991 135635297883904 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.19 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/15 I0420 17:05:27.029408 135635297883904 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.439216s. I0420 17:05:27.029527 135635825170176 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:05:27.029593 135635825170176 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:05:27.029657 135635825170176 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:05:27.031432 135635825170176 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0420 17:05:27.031540 135758522918720 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. W0420 17:05:27.031603 135758522918720 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.225084 seconds. If this number is high, consider checkpointing less frequently. I0420 17:05:27.033347 135758522918720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0420 17:05:27.035514 135758522918720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19. I0420 17:05:27.750105 135758522918720 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0420 17:05:27.750201 135758522918720 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 17:05:27.775156 135758522918720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025925s I0420 17:05:27.775887 135758522918720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.997 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030248641967773438 s) (per-host) I0420 17:05:27.775954 135758522918720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030333s (batch_requests_ready=0.001781s, total_serialization_initiated=0.027883s, others=0.000670s) I0420 17:05:27.776026 135758522918720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034365s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.034349) I0420 17:05:27.776772 135758522918720 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19. I0420 17:05:27.777060 135635825170176 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:25:27.777027 I0420 17:05:27.782926 135758522918720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 17:05:27.783211 135635864119040 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0420 17:05:27.783370 135758522918720 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776704696.8065004, 'wait_for_prev_duration_secs': 30.225083827972412, 'time_between_consecutive_saves_sec': 21.971107006072998, 'checkpointer_blocking_start_time': 1776704727.0333865, 'checkpointer_blocking_duration_secs': 0.7438194751739502, 'get_old_steps_start_time': 1776704727.7772257, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776704696.8047645, 'checkpoint_manager_blocking_duration_secs': 30.97857403755188} I0420 17:05:27.783478 135758522918720 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 17:05:27.783520 135758522918720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 17:05:28.230635 135635297883904 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19 I0420 17:05:30.108047 135635841955584 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19/items I0420 17:05:33.992879 135639152453376 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_0 I0420 17:06:13.590056 135635872511744 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.860140s (commit=40.584250s, array_metadata_write=1.275890s) I0420 17:06:13.591183 135635825170176 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.455 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.845513105392456 s) (per-host) I0420 17:06:13.591303 135635825170176 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.814120s. I0420 17:06:15.544251 135635825170176 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0420 17:06:15.996240 135635825170176 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000248s. I0420 17:06:16.210554 135635825170176 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.684942s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19/items I0420 17:06:16.213104 135635825170176 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19/items I0420 17:06:16.806798 135635825170176 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19 I0420 17:06:18.231244 135635825170176 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_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19`. I0420 17:06:18.232068 135635825170176 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.20 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_15_ocdbt_true/checkpoints/19 I0420 17:06:18.233640 135635825170176 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.456456s. I0420 17:06:18.233831 135635864119040 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 17:06:18.233946 135635864119040 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0420 17:06:18.233998 135635864119040 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 17:06:18.236354 135635864119040 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 17:06:18.236524 135758522918720 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. I0420 17:06:18.236675 135758522918720 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 17:06:18.237705 135758522918720 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.147, Tokens/s/device: 13889.454, 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: Mon Apr 20 17:06:27 UTC 2026 EXIT_CODE=0