XPK Start: Fri Apr 17 13:16:50 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 13:17:14.860575: 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) I0417 13:17:15.037145 136911100684096 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 13:17:24.077505 136911100684096 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 13:17:24.079935 136911100684096 distributed.py:172] Connecting to JAX distributed service on mt-04-int8-gh3xi-slice-job-0-0.mt-04-int8-gh3xi:8482 I0417 13:17:25.101885 136911100684096 max_utils.py:284] Jax distributed system initialized! I0417 13:17:30.123748 136911100684096 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 13:17:30.123904 136911100684096 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 13:17:30.123955 136911100684096 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 I0417 13:17:30.123997 136911100684096 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 13:17:30.794727 136911100684096 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 13:17:30.795156 136911100684096 checkpointing.py:677] Setting up checkpoint logger... I0417 13:17:30.795213 136911100684096 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 13:17:30.795257 136911100684096 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 13:17:30.795590 136911100684096 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 0x7c84865efce0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 13:17:35.004638 136911100684096 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0417 13:17:35.004888 136911100684096 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 0x7c6f6c269c70>}, handler_registry=None I0417 13:17:35.005121 136911100684096 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c6f6c269c70>` 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`. I0417 13:17:35.005168 136911100684096 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6f8410c2f0>` 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`. I0417 13:17:35.005203 136911100684096 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 0x7c6f6c269c70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c6f6c269c70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6f8410c2f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c6f8410c2f0>}). I0417 13:17:35.005517 136911100684096 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 13:17:35.005595 136911100684096 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c6f444989a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 13:17:36.564445 136911100684096 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints I0417 13:17:37.060180 136911100684096 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=10), 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_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c6f6c269e20> I0417 13:17:37.060349 136911100684096 checkpointing.py:301] Checkpoint manager created! I0417 13:17:38.179550 136911100684096 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 13:17:38.179667 136911100684096 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:17:38.562083 136911100684096 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:17:38.562176 136911100684096 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:17:38.584795 136911100684096 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:17:38.584909 136911100684096 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 13:17:38.650583 136911100684096 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 13:17:38.650669 136911100684096 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:17:38.667029 136911100684096 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 13:17:38.667090 136911100684096 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:17:38.683313 136911100684096 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 13:17:38.683373 136911100684096 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:17:38.707883 136911100684096 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 13:17:38.707948 136911100684096 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 13:17:38.774066 136911100684096 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 13:17:38.774151 136911100684096 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 13:17:39.298849 136911100684096 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0417 13:17:39.298956 136911100684096 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 13:17:41.754566 136911100684096 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 13:17:41.754689 136911100684096 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 13:17:41.754740 136911100684096 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 13:17:41.754814 136911100684096 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 13:17:41.754937 136911100684096 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 13:17:41.754989 136911100684096 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 13:17:41.755050 136911100684096 maxtext_utils.py:1651] 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) I0417 13:17:41.755109 136911100684096 maxtext_utils.py:1651] 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') I0417 13:17:41.755151 136911100684096 maxtext_utils.py:1651] 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) I0417 13:17:41.755188 136911100684096 maxtext_utils.py:1651] 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) I0417 13:17:41.755236 136911100684096 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 13:17:41.755281 136911100684096 maxtext_utils.py:1651] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 13:17:43.144079 136911100684096 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:17:43.144178 136911100684096 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0417 13:17:43.159570 136911100684096 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:17:43.159630 136911100684096 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0417 13:17:56.083297 136911100684096 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0417 13:17:56.091024 136911100684096 metric_logger.py:301] number parameters: 1.104 billion I0417 13:18:11.254796 136911100684096 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0417 13:18:11.495007 136911100684096 checkpointing.py:776] Waited 0.2401902675628662 seconds for step 0 to finish before starting checkpointing. I0417 13:18:11.497621 136911100684096 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:18:11.499387 136911100684096 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 13:18:11.500789 136911100684096 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0. I0417 13:18:12.229871 136911100684096 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 13:18:12.243941 136911100684096 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 13:18:12.244008 136911100684096 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 I0417 13:18:12.521682 136911100684096 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278389s I0417 13:18:12.522586 136911100684096 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.414 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2849290370941162 s) (per-host) I0417 13:18:12.522660 136911100684096 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285022s (batch_requests_ready=0.002236s, total_serialization_initiated=0.281965s, others=0.000820s) I0417 13:18:12.522750 136911100684096 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289170s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.289152) I0417 13:18:12.523582 136911100684096 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0. I0417 13:18:12.523875 136779499767552 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:38:12.523825 I0417 13:18:12.536273 136911100684096 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:18:12.536546 136778428307200 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:18:12.536698 136911100684096 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776431891.497601, 'wait_for_prev_duration_secs': 6.29425048828125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776431891.4994242, 'checkpointer_blocking_duration_secs': 1.0245981216430664, 'get_old_steps_start_time': 1776431892.524048, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776431891.4955482, 'checkpoint_manager_blocking_duration_secs': 1.0411121845245361} I0417 13:18:12.536802 136911100684096 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0417 13:18:12.536908 136911100684096 max_utils.py:750] Memstats: After params initialized: I0417 13:18:12.536957 136911100684096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 13:18:12.536989 136911100684096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 13:18:12.537015 136911100684096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 13:18:12.537039 136911100684096 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 13:18:12.890802 136911100684096 metric_logger.py:196] completed step: 0, seconds: 14.776, TFLOP/s/device: 0.920, Tokens/s/device: 138.602, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277 I0417 13:18:12.894242 136911100684096 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/tensorboard/' I0417 13:18:12.906439 136779491374848 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0 I0417 13:18:13.304884 136911100684096 metric_logger.py:196] completed step: 1, seconds: 1.634, TFLOP/s/device: 8.315, Tokens/s/device: 1253.291, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277 I0417 13:18:13.532444 136911100684096 metric_logger.py:196] completed step: 2, seconds: 0.415, TFLOP/s/device: 32.756, Tokens/s/device: 4937.343, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28634.820 I0417 13:18:13.763596 136911100684096 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1417.689, Tokens/s/device: 213689.482, total_weights: 65536, loss: 9.731, lm_loss: 9.731, perplexity: 16827.199 I0417 13:18:15.752606 136777364924160 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776431895320032983, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776431895320032983, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/_CHECKPOINT_METADATA I0417 13:18:16.024335 136778996508416 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/items I0417 13:18:18.018750 2946 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 13:18:21.086152 136778949236480 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/items/array_metadatas/process_0 I0417 13:18:45.214119 136911100684096 metric_logger.py:196] completed step: 4, seconds: 0.229, TFLOP/s/device: 59.353, Tokens/s/device: 8946.357, total_weights: 65536, loss: 9.272, lm_loss: 9.272, perplexity: 10638.531 I0417 13:18:45.223440 136911100684096 metric_logger.py:196] completed step: 5, seconds: 0.231, TFLOP/s/device: 58.777, Tokens/s/device: 8859.434, total_weights: 65536, loss: 8.887, lm_loss: 8.887, perplexity: 7233.941 I0417 13:18:45.450922 136911100684096 metric_logger.py:196] completed step: 6, seconds: 31.449, TFLOP/s/device: 0.432, Tokens/s/device: 65.121, total_weights: 65536, loss: 8.588, lm_loss: 8.588, perplexity: 5365.468 I0417 13:18:45.681490 136911100684096 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2008.445, Tokens/s/device: 302734.664, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4358.779 I0417 13:18:45.912388 136911100684096 metric_logger.py:196] completed step: 8, seconds: 0.229, TFLOP/s/device: 59.437, Tokens/s/device: 8959.037, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.453 I0417 13:18:46.143023 136911100684096 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0417 13:18:46.143779 136911100684096 checkpointing.py:776] Waited 0.0007755756378173828 seconds for step 9 to finish before starting checkpointing. I0417 13:18:46.146000 136911100684096 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:18:50.797075 136778468030208 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 32.779366s (commit=31.529243s, array_metadata_write=1.250123s) I0417 13:18:50.798159 136779499767552 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 40.964 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.56048107147217 s) (per-host) I0417 13:18:50.798271 136779499767552 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.274286s. I0417 13:18:52.014190 136779499767552 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776431895320032983, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/_CHECKPOINT_METADATA I0417 13:18:52.742428 136777364924160 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776431895320032983, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/_CHECKPOINT_METADATA I0417 13:18:53.170710 136779499767552 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000261s. I0417 13:18:53.401574 136779499767552 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:18:54.190804 136779499767552 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.046608s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/items I0417 13:18:54.192547 136779499767552 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0/items I0417 13:18:55.183963 136779499767552 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0 I0417 13:18:56.571767 136779499767552 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0`. I0417 13:18:56.572638 136779499767552 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 45.07 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/0 I0417 13:18:56.574158 136779499767552 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 44.050169s. I0417 13:18:56.574357 136778428307200 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:18:56.574477 136778428307200 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:18:56.574554 136778428307200 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:18:56.576285 136778428307200 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 13:18:56.576448 136911100684096 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0417 13:18:56.576584 136911100684096 checkpoint_manager.py:1452] Waiting for previous save to complete took 10.430583 seconds. If this number is high, consider checkpointing less frequently. I0417 13:18:56.578395 136911100684096 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 13:18:56.580496 136911100684096 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9. I0417 13:18:57.296801 136911100684096 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 13:18:57.296912 136911100684096 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 I0417 13:18:57.322878 136911100684096 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026813s I0417 13:18:57.323623 136911100684096 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.182 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03136491775512695 s) (per-host) I0417 13:18:57.323691 136911100684096 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031445s (batch_requests_ready=0.001829s, total_serialization_initiated=0.028922s, others=0.000693s) I0417 13:18:57.323759 136911100684096 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035556s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.035540) I0417 13:18:57.324535 136911100684096 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9. I0417 13:18:57.324903 136777905981184 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:38:57.324866 I0417 13:18:57.332738 136911100684096 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:18:57.333027 136778428307200 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:18:57.333194 136911100684096 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776431926.1459668, 'wait_for_prev_duration_secs': 10.430583477020264, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776431936.5784328, 'checkpointer_blocking_duration_secs': 0.7466137409210205, 'get_old_steps_start_time': 1776431937.3250766, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776431926.1440487, 'checkpoint_manager_blocking_duration_secs': 11.189110040664673} I0417 13:18:57.333314 136911100684096 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0417 13:18:57.333359 136911100684096 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:18:57.531098 136779499767552 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9 I0417 13:18:59.739459 136778971330304 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9/items I0417 13:19:05.207055 136777931159296 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9/items/array_metadatas/process_0 I0417 13:19:44.304072 136777914373888 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.735878s (commit=39.670100s, array_metadata_write=2.065778s) I0417 13:19:44.305195 136777905981184 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.599 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.012895584106445 s) (per-host) I0417 13:19:44.305321 136777905981184 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.980304s. I0417 13:19:46.631736 136777905981184 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:19:46.651472 136777905981184 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000236s. I0417 13:19:47.678611 136777905981184 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.461058s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9/items I0417 13:19:47.680194 136777905981184 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9/items I0417 13:19:48.194157 136777905981184 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9 I0417 13:19:49.622429 136777905981184 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9`. I0417 13:19:49.623274 136777905981184 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.04 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_04_int8/checkpoints/9 I0417 13:19:49.624647 136777905981184 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.299627s. I0417 13:19:49.624877 136778428307200 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:19:49.624991 136778428307200 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:19:49.625107 136778428307200 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:19:49.629621 136778428307200 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 13:19:49.629797 136911100684096 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0417 13:19:49.629954 136911100684096 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:19:49.630933 136911100684096 metric_logger.py:196] completed step: 9, seconds: 0.230, TFLOP/s/device: 58.947, Tokens/s/device: 8885.110, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3552.333 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 17 13:20:00 UTC 2026 EXIT_CODE=0