XPK Start: Sun Apr 19 18:55:07 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 18:55:31.402410: 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 18:55:31.581049 134333930518336 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 18:55:40.620642 134333930518336 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 18:55:40.622901 134333930518336 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-drp0x-slice-job-0-0.mt-11-optimizer-offload-drp0x:8482 I0419 18:55:42.366620 134333930518336 max_utils.py:284] Jax distributed system initialized! I0419 18:55:48.379421 134333930518336 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 18:55:48.379540 134333930518336 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 18:55:48.379587 134333930518336 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 18:55:48.379621 134333930518336 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 18:55:49.073504 134333930518336 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 18:55:49.073798 134333930518336 checkpointing.py:677] Setting up checkpoint logger... I0419 18:55:49.073848 134333930518336 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 18:55:49.073890 134333930518336 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 18:55:49.074217 134333930518336 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 0x7a2c5684c290>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 18:55:52.849397 134333930518336 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0419 18:55:52.849642 134333930518336 checkpoint_manager.py:708] [process=2][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a184c2d2630>}, handler_registry=None I0419 18:55:52.849878 134333930518336 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a184c2d2630>` 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 18:55:52.849925 134333930518336 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a184c1262a0>` 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 18:55:52.849960 134333930518336 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 0x7a184c2d2630>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a184c2d2630>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a184c1262a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a184c1262a0>}). I0419 18:55:52.850279 134333930518336 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 18:55:52.850349 134333930518336 async_checkpointer.py:192] [process=2][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a181c16b9c0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 18:55:53.968688 134333930518336 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints I0419 18:55:54.430735 134333930518336 checkpoint_manager.py:929] [process=2][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_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a184c123b60> I0419 18:55:54.430901 134333930518336 checkpointing.py:301] Checkpoint manager created! I0419 18:55:55.355355 134333930518336 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 18:55:55.355467 134333930518336 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:55:55.733359 134333930518336 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 18:55:55.733445 134333930518336 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:55:55.749675 134333930518336 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 18:55:55.749733 134333930518336 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 18:55:55.773298 134333930518336 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 18:55:55.773364 134333930518336 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:55:55.789689 134333930518336 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 18:55:55.789748 134333930518336 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:55:55.805875 134333930518336 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 18:55:55.805942 134333930518336 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:55:55.830442 134333930518336 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 18:55:55.830509 134333930518336 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 18:55:55.852295 134333930518336 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 18:55:55.852360 134333930518336 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 18:55:56.069683 134333930518336 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0419 18:55:56.069788 134333930518336 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 18:55:57.661075 134333930518336 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 18:55:57.661244 134333930518336 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 18:55:57.661300 134333930518336 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 18:55:57.661364 134333930518336 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 18:55:57.661420 134333930518336 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 18:55:57.661460 134333930518336 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 18:55:57.661530 134333930518336 maxtext_utils.py:1654] 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 18:55:57.661589 134333930518336 maxtext_utils.py:1654] 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 18:55:57.661630 134333930518336 maxtext_utils.py:1654] 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 18:55:57.661665 134333930518336 maxtext_utils.py:1654] 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 18:55:57.661712 134333930518336 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 18:55:57.661762 134333930518336 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 18:55:58.157408 134333930518336 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 18:55:58.157503 134333930518336 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 18:55:58.172597 134333930518336 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 18:55:58.172656 134333930518336 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 18:56:09.142300 134333930518336 max_utils.py:791] Total memory size: 1.3 GB, Output size: 0.1 GB, Temp size: 1.1 GB, Argument size: 0.1 GB, Host temp size: 0.0 GB. I0419 18:56:09.143101 134333930518336 metric_logger.py:301] number parameters: 1.104 billion I0419 18:56:20.559033 134333930518336 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0419 18:56:20.727851 134333930518336 checkpointing.py:776] Waited 0.16879558563232422 seconds for step 0 to finish before starting checkpointing. I0419 18:56:20.730431 134333930518336 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 18:56:20.732317 134333930518336 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 0 I0419 18:56:20.733846 134333930518336 event_tracking.py:70] [process=2] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/0. I0419 18:56:21.494362 134333930518336 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 18:56:21.495411 134333930518336 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 18:56:21.495473 134333930518336 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 18:56:21.771713 134333930518336 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277423s I0419 18:56:21.771883 134333930518336 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 5.432 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2839484214782715 s) (per-host) I0419 18:56:21.771936 134333930518336 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.284012s (batch_requests_ready=0.002350s, total_serialization_initiated=0.281590s, others=0.000072s) I0419 18:56:21.772031 134333930518336 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.288113s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.288095) I0419 18:56:21.772803 134333930518336 event_tracking.py:125] [process=2] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/0. I0419 18:56:21.773076 134198294800128 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:16:21.773051 I0419 18:56:21.775052 134333930518336 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 18:56:21.775367 134187143108352 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save. I0419 18:56:21.775534 134333930518336 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776624980.7304132, 'wait_for_prev_duration_secs': 5.936622619628906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776624980.732356, 'checkpointer_blocking_duration_secs': 1.0408313274383545, 'get_old_steps_start_time': 1776624981.773213, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776624980.7284143, 'checkpoint_manager_blocking_duration_secs': 1.0470709800720215} I0419 18:56:21.775638 134333930518336 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0419 18:56:21.775688 134333930518336 max_utils.py:750] Memstats: After params initialized: I0419 18:56:21.775741 134333930518336 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_8(process=2,(0,2,0,0)) I0419 18:56:21.775775 134333930518336 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_9(process=2,(1,2,0,0)) I0419 18:56:21.775804 134333930518336 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_12(process=2,(0,3,0,0)) I0419 18:56:21.775829 134333930518336 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_13(process=2,(1,3,0,0)) I0419 18:56:22.092431 134333930518336 metric_logger.py:196] completed step: 0, seconds: 11.416, TFLOP/s/device: 1.190, Tokens/s/device: 179.400, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 18:56:22.278023 134333930518336 metric_logger.py:196] completed step: 1, seconds: 1.532, TFLOP/s/device: 8.870, Tokens/s/device: 1336.922, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 18:56:22.688673 134333930518336 metric_logger.py:196] completed step: 2, seconds: 0.023, TFLOP/s/device: 595.378, Tokens/s/device: 89741.904, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0419 18:56:22.851616 134333930518336 metric_logger.py:196] completed step: 3, seconds: 0.411, TFLOP/s/device: 33.066, Tokens/s/device: 4984.133, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.047 I0419 18:56:23.177617 134333930518336 metric_logger.py:196] completed step: 4, seconds: 0.168, TFLOP/s/device: 80.725, Tokens/s/device: 12167.734, total_weights: 65536, loss: 9.278, lm_loss: 9.278, perplexity: 10694.968 I0419 18:56:23.183505 134333930518336 metric_logger.py:196] completed step: 5, seconds: 0.163, TFLOP/s/device: 83.396, Tokens/s/device: 12570.355, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.520 I0419 18:56:26.007192 2822 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 18:56:28.946665 134198278014720 array_metadata_store.py:203] [process=2][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_2 I0419 18:56:45.187382 134333930518336 metric_logger.py:196] completed step: 6, seconds: 0.327, TFLOP/s/device: 41.595, Tokens/s/device: 6269.650, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.688 I0419 18:56:45.350227 134333930518336 metric_logger.py:196] completed step: 7, seconds: 21.842, TFLOP/s/device: 0.622, Tokens/s/device: 93.766, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4377.010 I0419 18:56:45.512997 134333930518336 metric_logger.py:196] completed step: 8, seconds: 0.167, TFLOP/s/device: 81.123, Tokens/s/device: 12227.669, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3845.943 I0419 18:56:45.675218 134333930518336 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0419 18:56:45.675885 134333930518336 checkpointing.py:776] Waited 0.0006866455078125 seconds for step 9 to finish before starting checkpointing. I0419 18:56:45.678245 134333930518336 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 18:56:59.413893 134198294800128 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 41.648 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.92591905593872 s) (per-host) I0419 18:56:59.414015 134198294800128 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.640858s. I0419 18:57:08.864100 134198294800128 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 47.090927s. I0419 18:57:08.864392 134187143108352 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 18:57:08.864545 134187143108352 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save. I0419 18:57:08.864600 134187143108352 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 18:57:08.866557 134187143108352 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 18:57:08.866727 134333930518336 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 18:57:08.866862 134333930518336 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.188619 seconds. If this number is high, consider checkpointing less frequently. I0419 18:57:08.869740 134333930518336 checkpoint_manager.py:1512] [process=2] Saving checkpoint at step 9 I0419 18:57:08.872255 134333930518336 event_tracking.py:70] [process=2] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/9. I0419 18:57:09.580241 134333930518336 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 18:57:09.580331 134333930518336 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 18:57:09.613481 134333930518336 base_pytree_checkpoint_handler.py:154] [process=2][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034159s I0419 18:57:09.613660 134333930518336 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/blocking_gbytes_per_sec: 40.219 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03835248947143555 s) (per-host) I0419 18:57:09.613709 134333930518336 base_pytree_checkpoint_handler.py:768] [process=2][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038410s (batch_requests_ready=0.002590s, total_serialization_initiated=0.035755s, others=0.000066s) I0419 18:57:09.613798 134333930518336 composite_checkpoint_handler.py:715] [process=2][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042489s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.042474) I0419 18:57:09.614459 134333930518336 event_tracking.py:125] [process=2] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/9. I0419 18:57:09.614828 134182378387200 async_checkpointer.py:76] [process=2][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:17:09.614791 I0419 18:57:09.616845 134333930518336 checkpoint_manager.py:1560] [process=2][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 18:57:09.617095 134187143108352 async_checkpointer.py:280] [process=2][thread=save_finalize] Waiting for background save thread=async_save. I0419 18:57:09.617235 134333930518336 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776625005.6782157, 'wait_for_prev_duration_secs': 23.188619375228882, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776625028.8697803, 'checkpointer_blocking_duration_secs': 0.7452003955841064, 'get_old_steps_start_time': 1776625029.6150053, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776625005.6761122, 'checkpoint_manager_blocking_duration_secs': 23.94108819961548} I0419 18:57:09.617340 134333930518336 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0419 18:57:09.617384 134333930518336 checkpoint_manager.py:2020] [process=2][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 18:57:16.617254 134198278014720 array_metadata_store.py:203] [process=2][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_2 I0419 18:57:51.654101 134182378387200 base_pytree_checkpoint_handler.py:130] [process=2] /jax/orbax/write/gbytes_per_sec: 37.538 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.07875299453735 s) (per-host) I0419 18:57:51.654243 134182378387200 async_checkpointer.py:90] [process=2][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.039297s. I0419 18:57:59.596351 134182378387200 async_checkpointer.py:160] [process=2][thread=async_save] Background save thread done. Time taken: 49.981391s. I0419 18:57:59.596647 134187143108352 async_checkpointer.py:288] [process=2][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 18:57:59.596767 134187143108352 async_checkpointer.py:298] [process=2][thread=save_finalize] No errors found in background save thread=async_save. I0419 18:57:59.596818 134187143108352 checkpoint_manager.py:2137] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 18:57:59.598210 134187143108352 checkpoint_manager.py:2146] [process=2][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 18:57:59.600697 134333930518336 checkpoint_manager.py:2032] [process=2][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 18:57:59.600841 134333930518336 checkpoint_manager.py:2009] [process=2][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 18:57:59.601827 134333930518336 metric_logger.py:196] completed step: 9, seconds: 0.163, TFLOP/s/device: 83.427, Tokens/s/device: 12575.064, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.739 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 18:58:12 UTC 2026 EXIT_CODE=0