XPK Start: Sun Apr 19 19:31:45 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 19:32:09.683389: 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 19:32:10.229752 137823634863936 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 19:32:19.270903 137823634863936 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 19:32:19.273329 137823634863936 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-sv-iioax-slice-job-0-0.mt-16-zarr3-true-sv-iioax:8482
I0419 19:32:20.646674 137823634863936 max_utils.py:284] Jax distributed system initialized!
I0419 19:32:26.804899 137823634863936 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 19:32:26.805002 137823634863936 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 19:32:26.805041 137823634863936 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 19:32:26.805076 137823634863936 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 19:32:27.500904 137823634863936 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 19:32:27.501186 137823634863936 checkpointing.py:677] Setting up checkpoint logger...
I0419 19:32:27.501238 137823634863936 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 19:32:27.501282 137823634863936 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 19:32:27.501627 137823634863936 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 0x7d58f2d1cef0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 19:32:30.913811 137823634863936 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 19:32:30.914044 137823634863936 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c011ea80>}, handler_registry=None
I0419 19:32:30.914348 137823634863936 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c011ea80>` 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 19:32:30.914404 137823634863936 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44a8552b10>` 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 19:32:30.914456 137823634863936 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 0x7d44c011ea80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c011ea80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44a8552b10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44a8552b10>}).
I0419 19:32:30.914779 137823634863936 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 19:32:30.914848 137823634863936 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d4480147880> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 19:32:32.017926 137823634863936 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints
I0419 19:32:32.449746 137823634863936 checkpoint_manager.py:929] [process=6][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_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d44a854e9c0>
I0419 19:32:32.449913 137823634863936 checkpointing.py:301] Checkpoint manager created!
I0419 19:32:33.373218 137823634863936 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 19:32:33.373335 137823634863936 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:32:33.753434 137823634863936 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 19:32:33.753528 137823634863936 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:32:33.769494 137823634863936 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 19:32:33.769550 137823634863936 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 19:32:33.793866 137823634863936 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 19:32:33.793934 137823634863936 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:32:33.809976 137823634863936 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 19:32:33.810038 137823634863936 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:32:33.825955 137823634863936 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 19:32:33.826016 137823634863936 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:32:33.850273 137823634863936 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 19:32:33.850340 137823634863936 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 19:32:33.876587 137823634863936 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 19:32:33.876675 137823634863936 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 19:32:34.090352 137823634863936 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 19:32:34.090475 137823634863936 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 19:32:35.508308 137823634863936 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0419 19:32:35.508445 137823634863936 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 19:32:35.508501 137823634863936 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 19:32:35.508568 137823634863936 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 19:32:35.508622 137823634863936 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 19:32:35.508662 137823634863936 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 19:32:35.508715 137823634863936 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 19:32:35.508767 137823634863936 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 19:32:35.508807 137823634863936 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 19:32:35.508843 137823634863936 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 19:32:35.508890 137823634863936 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0419 19:32:35.508938 137823634863936 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0419 19:32:35.995195 137823634863936 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 19:32:35.995286 137823634863936 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 19:32:36.010395 137823634863936 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 19:32:36.010468 137823634863936 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 19:32:47.082849 137823634863936 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0419 19:32:47.085221 137823634863936 metric_logger.py:301] number parameters: 1.104 billion
I0419 19:32:58.322808 137823634863936 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0419 19:32:58.476639 137823634863936 checkpointing.py:776] Waited 0.1538095474243164 seconds for step 0 to finish before starting checkpointing.
I0419 19:32:58.479174 137823634863936 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 19:32:58.480872 137823634863936 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0419 19:32:58.482418 137823634863936 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/0.
I0419 19:32:59.632148 137823634863936 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 19:32:59.634631 137823634863936 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:32:59.634693 137823634863936 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 19:32:59.912763 137823634863936 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.280690s
I0419 19:32:59.912929 137823634863936 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.392 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2861008644104004 s) (per-host)
I0419 19:32:59.912980 137823634863936 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286161s (batch_requests_ready=0.002192s, total_serialization_initiated=0.283901s, others=0.000068s)
I0419 19:32:59.913070 137823634863936 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290339s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.290321)
I0419 19:32:59.913823 137823634863936 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.43 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/0.
I0419 19:32:59.914152 137694073870080 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:52:59.914114
I0419 19:32:59.920456 137823634863936 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:32:59.920701 137692989871872 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:32:59.920830 137823634863936 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627178.479157, 'wait_for_prev_duration_secs': 5.888938903808594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776627178.4809113, 'checkpointer_blocking_duration_secs': 1.433396339416504, 'get_old_steps_start_time': 1776627179.9143307, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776627178.4771924, 'checkpoint_manager_blocking_duration_secs': 1.4435997009277344}
I0419 19:32:59.920933 137823634863936 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0419 19:32:59.920981 137823634863936 max_utils.py:750]
Memstats: After params initialized:
I0419 19:32:59.921029 137823634863936 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0419 19:32:59.921061 137823634863936 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0419 19:32:59.921087 137823634863936 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0419 19:32:59.921111 137823634863936 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0419 19:33:00.233738 137823634863936 metric_logger.py:196] completed step: 0, seconds: 11.237, TFLOP/s/device: 1.209, Tokens/s/device: 182.247, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 19:33:00.409892 137823634863936 metric_logger.py:196] completed step: 1, seconds: 1.909, TFLOP/s/device: 7.116, Tokens/s/device: 1072.560, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 19:33:00.823810 137823634863936 metric_logger.py:196] completed step: 2, seconds: 0.029, TFLOP/s/device: 471.464, Tokens/s/device: 71064.228, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0419 19:33:00.971394 137823634863936 metric_logger.py:196] completed step: 3, seconds: 0.414, TFLOP/s/device: 32.801, Tokens/s/device: 4944.137, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0419 19:33:00.976372 137823634863936 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0419 19:33:01.265759 137823634863936 checkpointing.py:776] Waited 0.289370059967041 seconds for step 5 to finish before starting checkpointing.
I0419 19:33:01.268749 137823634863936 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 19:33:02.890375 2822 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 19:33:04.904756 137692998264576 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/0/items/array_metadatas/process_6
I0419 19:33:36.788778 137694073870080 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.506 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.16189765930176 s) (per-host)
I0419 19:33:36.788904 137694073870080 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.874640s.
I0419 19:33:45.405695 137694073870080 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 45.491415s.
I0419 19:33:45.405996 137692989871872 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:33:45.406117 137692989871872 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:33:45.406167 137692989871872 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:33:45.408519 137692989871872 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 19:33:45.408733 137823634863936 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0419 19:33:45.408886 137823634863936 checkpoint_manager.py:1452] Waiting for previous save to complete took 44.140138 seconds. If this number is high, consider checkpointing less frequently.
I0419 19:33:45.410853 137823634863936 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5
I0419 19:33:45.413306 137823634863936 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/5.
I0419 19:33:46.127642 137823634863936 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:33:46.127742 137823634863936 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 19:33:46.174167 137823634863936 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047646s
I0419 19:33:46.174336 137823634863936 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 30.212 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.051058053970336914 s) (per-host)
I0419 19:33:46.174388 137823634863936 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.051118s (batch_requests_ready=0.001764s, total_serialization_initiated=0.049287s, others=0.000068s)
I0419 19:33:46.174490 137823634863936 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.056270s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.056252)
I0419 19:33:46.175149 137823634863936 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/5.
I0419 19:33:46.175518 137692989871872 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:53:46.175483
I0419 19:33:46.177430 137823634863936 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:33:46.177655 137693564499712 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:33:46.177759 137823634863936 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627181.2687197, 'wait_for_prev_duration_secs': 44.140137910842896, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776627225.4108927, 'checkpointer_blocking_duration_secs': 0.7647695541381836, 'get_old_steps_start_time': 1776627226.1756847, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776627181.2663517, 'checkpoint_manager_blocking_duration_secs': 44.91137647628784}
I0419 19:33:46.177909 137823634863936 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0419 19:33:46.178679 137823634863936 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.827, Tokens/s/device: 13389.034, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0419 19:33:46.184940 137823634863936 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 91.816, Tokens/s/device: 13839.521, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0419 19:33:52.185277 137692998264576 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/5/items/array_metadatas/process_6
I0419 19:34:09.201487 137823634863936 metric_logger.py:196] completed step: 6, seconds: 45.207, TFLOP/s/device: 0.301, Tokens/s/device: 45.303, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0419 19:34:09.349103 137823634863936 metric_logger.py:196] completed step: 7, seconds: 22.870, TFLOP/s/device: 0.594, Tokens/s/device: 89.549, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0419 19:34:09.496183 137823634863936 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.110, Tokens/s/device: 13431.622, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0419 19:34:09.642821 137823634863936 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0419 19:34:09.643502 137823634863936 checkpointing.py:776] Waited 0.0007004737854003906 seconds for step 9 to finish before starting checkpointing.
I0419 19:34:09.645768 137823634863936 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 19:34:28.338988 137692989871872 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.417 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.21566843986511 s) (per-host)
I0419 19:34:28.339117 137692989871872 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.163483s.
I0419 19:34:37.674808 137692989871872 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.499160s.
I0419 19:34:37.675138 137693564499712 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:34:37.675264 137693564499712 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:34:37.675315 137693564499712 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:34:37.676929 137693564499712 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0419 19:34:37.677116 137823634863936 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0419 19:34:37.677247 137823634863936 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.031486 seconds. If this number is high, consider checkpointing less frequently.
I0419 19:34:37.678765 137823634863936 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0419 19:34:37.680819 137823634863936 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9.
I0419 19:34:38.409606 137823634863936 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 19:34:38.409699 137823634863936 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 19:34:38.457394 137823634863936 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048799s
I0419 19:34:38.457571 137823634863936 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 29.551 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.052199602127075195 s) (per-host)
I0419 19:34:38.457620 137823634863936 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.052258s (batch_requests_ready=0.001636s, total_serialization_initiated=0.050558s, others=0.000064s)
I0419 19:34:38.457704 137823634863936 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.056595s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.056583)
I0419 19:34:38.458432 137823634863936 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.78 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9.
I0419 19:34:38.458693 137693564499712 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 19:54:38.458671
I0419 19:34:38.460757 137823634863936 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 19:34:38.461064 137693539321600 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 19:34:38.461252 137823634863936 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776627249.645738, 'wait_for_prev_duration_secs': 28.03148579597473, 'time_between_consecutive_saves_sec': 24.23717761039734, 'checkpointer_blocking_start_time': 1776627277.6788042, 'checkpointer_blocking_duration_secs': 0.7799906730651855, 'get_old_steps_start_time': 1776627278.4588108, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776627249.6437259, 'checkpoint_manager_blocking_duration_secs': 28.817485332489014}
I0419 19:34:38.461371 137823634863936 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0419 19:34:38.461443 137823634863936 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 19:34:44.021340 137692998264576 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_16_zarr3_true/checkpoints/9/items/array_metadatas/process_6
I0419 19:35:19.989723 137693564499712 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.985 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.584311723709106 s) (per-host)
I0419 19:35:19.989849 137693564499712 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.531078s.
I0419 19:35:28.170758 137693564499712 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.711972s.
I0419 19:35:28.171049 137693539321600 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 19:35:28.171169 137693539321600 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 19:35:28.171210 137693539321600 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 19:35:28.172729 137693539321600 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 19:35:28.172911 137823634863936 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0419 19:35:28.173074 137823634863936 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 19:35:28.173969 137823634863936 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.041, Tokens/s/device: 13873.459, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 19:35:40 UTC 2026
EXIT_CODE=0