XPK Start: Sun Apr 19 18:33:42 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 18:34:06.628193: 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 18:34:06.808355 138110966900544 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 18:34:15.851010 138110966900544 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 18:34:15.853479 138110966900544 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--oiszg-slice-job-0-0.mt-08-checkpoint-async--oiszg:8482
I0419 18:34:16.902940 138110966900544 max_utils.py:284] Jax distributed system initialized!
I0419 18:34:23.145822 138110966900544 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 18:34:23.145922 138110966900544 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 18:34:23.145962 138110966900544 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 18:34:23.145996 138110966900544 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 18:34:23.837243 138110966900544 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 18:34:23.837536 138110966900544 checkpointing.py:677] Setting up checkpoint logger...
I0419 18:34:23.837591 138110966900544 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 18:34:23.837635 138110966900544 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 18:34:23.837963 138110966900544 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 0x7d9be3d0a6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 18:34:27.180629 138110966900544 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 18:34:27.180860 138110966900544 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 0x7d9ab3069c40>}, handler_registry=None
I0419 18:34:27.181097 138110966900544 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d9ab3069c40>` 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 18:34:27.181144 138110966900544 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d87b82c84d0>` 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 18:34:27.181180 138110966900544 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 0x7d9ab3069c40>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d9ab3069c40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d87b82c84d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d87b82c84d0>}).
I0419 18:34:27.181574 138110966900544 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 18:34:27.181662 138110966900544 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d87a05e3b00> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 18:34:28.700597 138110966900544 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints
I0419 18:34:28.705960 138110966900544 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_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d87b82c6120>
I0419 18:34:28.706071 138110966900544 checkpointing.py:301] Checkpoint manager created!
I0419 18:34:29.632742 138110966900544 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 18:34:29.632857 138110966900544 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:34:30.011479 138110966900544 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 18:34:30.011567 138110966900544 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:34:30.027735 138110966900544 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 18:34:30.027793 138110966900544 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:34:30.051373 138110966900544 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 18:34:30.051449 138110966900544 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:34:30.068067 138110966900544 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 18:34:30.068127 138110966900544 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:34:30.084837 138110966900544 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 18:34:30.084897 138110966900544 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:34:30.109255 138110966900544 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 18:34:30.109323 138110966900544 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:34:30.131086 138110966900544 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 18:34:30.131153 138110966900544 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 18:34:30.351735 138110966900544 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 18:34:30.351853 138110966900544 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 18:34:31.774167 138110966900544 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0419 18:34:31.774289 138110966900544 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 18:34:31.774342 138110966900544 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 18:34:31.774395 138110966900544 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 18:34:31.774478 138110966900544 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 18:34:31.774528 138110966900544 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 18:34:31.774580 138110966900544 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 18:34:31.774630 138110966900544 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 18:34:31.774670 138110966900544 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 18:34:31.774707 138110966900544 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 18:34:31.774753 138110966900544 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0419 18:34:31.774800 138110966900544 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0419 18:34:32.259287 138110966900544 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 18:34:32.259378 138110966900544 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 18:34:32.274610 138110966900544 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 18:34:32.274669 138110966900544 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 18:34:43.323460 138110966900544 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 18:34:43.324369 138110966900544 metric_logger.py:301] number parameters: 1.104 billion
I0419 18:34:54.808897 138110966900544 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0419 18:34:54.962532 138110966900544 checkpointing.py:776] Waited 0.15360665321350098 seconds for step 0 to finish before starting checkpointing.
I0419 18:34:54.965231 138110966900544 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 18:34:54.966844 138110966900544 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0419 18:34:54.968603 138110966900544 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/0.
I0419 18:34:55.710258 138110966900544 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 18:34:55.711325 138110966900544 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 18:34:55.711381 138110966900544 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 18:34:55.997565 138110966900544 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.287385s
I0419 18:34:55.997739 138110966900544 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.269 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2927560806274414 s) (per-host)
I0419 18:34:55.997792 138110966900544 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292819s (batch_requests_ready=0.002182s, total_serialization_initiated=0.290567s, others=0.000071s)
I0419 18:34:55.997886 138110966900544 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.296880s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.296862)
I0419 18:34:55.998660 138110966900544 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/0.
I0419 18:34:55.998997 137982266832640 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:54:55.998960
I0419 18:34:56.000966 138110966900544 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:34:56.001252 137981184698112 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:34:56.001419 138110966900544 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776623694.9652107, 'wait_for_prev_duration_secs': 6.747245788574219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776623694.9669297, 'checkpointer_blocking_duration_secs': 1.0322203636169434, 'get_old_steps_start_time': 1776623695.999177, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776623694.9631479, 'checkpoint_manager_blocking_duration_secs': 1.038222312927246}
I0419 18:34:56.001519 138110966900544 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0419 18:34:56.001568 138110966900544 max_utils.py:750]
Memstats: After params initialized:
I0419 18:34:56.001614 138110966900544 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0419 18:34:56.001646 138110966900544 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0419 18:34:56.001672 138110966900544 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0419 18:34:56.001695 138110966900544 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0419 18:34:56.309712 138110966900544 metric_logger.py:196] completed step: 0, seconds: 11.484, TFLOP/s/device: 1.183, Tokens/s/device: 178.329, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 18:34:56.495509 138110966900544 metric_logger.py:196] completed step: 1, seconds: 1.499, TFLOP/s/device: 9.062, Tokens/s/device: 1365.960, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0419 18:34:56.895529 138110966900544 metric_logger.py:196] completed step: 2, seconds: 0.038, TFLOP/s/device: 354.644, Tokens/s/device: 53455.836, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0419 18:34:57.043120 138110966900544 metric_logger.py:196] completed step: 3, seconds: 0.400, TFLOP/s/device: 33.936, Tokens/s/device: 5115.204, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0419 18:34:57.047825 138110966900544 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0419 18:34:57.337495 138110966900544 checkpointing.py:776] Waited 0.28963303565979004 seconds for step 5 to finish before starting checkpointing.
I0419 18:34:57.340481 138110966900544 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:34:59.438126 2771 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 18:35:02.485780 137981193090816 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_08_checkpoint_async_true/checkpoints/0/items/array_metadatas/process_6
I0419 18:35:33.084134 137982266832640 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 42.259 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.37909507751465 s) (per-host)
I0419 18:35:33.084259 137982266832640 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.085149s.
I0419 18:35:42.200750 137982266832640 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 46.201624s.
I0419 18:35:42.201056 137981184698112 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:35:42.201178 137981184698112 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:35:42.201243 137981184698112 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:35:42.203770 137981184698112 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 18:35:42.203969 138110966900544 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 18:35:42.204108 138110966900544 checkpoint_manager.py:1452] Waiting for previous save to complete took 44.863627 seconds. If this number is high, consider checkpointing less frequently.
I0419 18:35:42.206065 138110966900544 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5
I0419 18:35:42.208588 138110966900544 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/5.
I0419 18:35:42.505311 138110966900544 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 18:35:42.505424 138110966900544 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 18:35:42.547727 138110966900544 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043561s
I0419 18:35:42.547879 138110966900544 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 32.887 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.046905517578125 s) (per-host)
I0419 18:35:42.547927 138110966900544 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.046963s (batch_requests_ready=0.001738s, total_serialization_initiated=0.045162s, others=0.000063s)
I0419 18:35:42.548026 138110966900544 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051934s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.051918)
I0419 18:35:42.548765 138110966900544 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/5.
I0419 18:35:42.549047 137981184698112 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:55:42.549016
I0419 18:35:42.551030 138110966900544 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:35:42.551255 137982233261824 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:35:42.551370 138110966900544 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776623697.3404505, 'wait_for_prev_duration_secs': 44.8636269569397, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776623742.2061052, 'checkpointer_blocking_duration_secs': 0.34308767318725586, 'get_old_steps_start_time': 1776623742.5492177, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776623697.3380804, 'checkpoint_manager_blocking_duration_secs': 45.21325755119324}
I0419 18:35:42.551489 138110966900544 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0419 18:35:42.552213 138110966900544 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.622, Tokens/s/device: 13358.119, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0419 18:35:42.558030 138110966900544 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.112, Tokens/s/device: 13884.087, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0419 18:35:49.891100 137981193090816 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_08_checkpoint_async_true/checkpoints/5/items/array_metadatas/process_6
I0419 18:36:05.171798 138110966900544 metric_logger.py:196] completed step: 6, seconds: 45.509, TFLOP/s/device: 0.299, Tokens/s/device: 45.002, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0419 18:36:05.319247 138110966900544 metric_logger.py:196] completed step: 7, seconds: 22.468, TFLOP/s/device: 0.605, Tokens/s/device: 91.153, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0419 18:36:05.466523 138110966900544 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.312, Tokens/s/device: 13462.082, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0419 18:36:05.613392 138110966900544 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0419 18:36:05.614067 138110966900544 checkpointing.py:776] Waited 0.0006964206695556641 seconds for step 9 to finish before starting checkpointing.
I0419 18:36:05.616121 138110966900544 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:36:25.756672 137981184698112 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.518 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.25564646720886 s) (per-host)
I0419 18:36:25.756780 137981184698112 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.207621s.
I0419 18:36:33.284728 137981184698112 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.735553s.
I0419 18:36:33.284947 137982233261824 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:36:33.285030 137982233261824 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:36:33.285079 137982233261824 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:36:33.286736 137982233261824 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0419 18:36:33.286897 138110966900544 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 18:36:33.287024 138110966900544 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.670910 seconds. If this number is high, consider checkpointing less frequently.
I0419 18:36:33.288511 138110966900544 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0419 18:36:33.290595 138110966900544 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints/9.
I0419 18:36:34.007402 138110966900544 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 18:36:34.007516 138110966900544 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 18:36:34.048959 138110966900544 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042489s
I0419 18:36:34.049117 138110966900544 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.645 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04584860801696777 s) (per-host)
I0419 18:36:34.049168 138110966900544 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045910s (batch_requests_ready=0.001596s, total_serialization_initiated=0.044248s, others=0.000066s)
I0419 18:36:34.049255 138110966900544 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050059s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.050047)
I0419 18:36:34.050018 138110966900544 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_08_checkpoint_async_true/checkpoints/9.
I0419 18:36:34.050324 137982233261824 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:56:34.050287
I0419 18:36:34.052311 138110966900544 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:36:34.052523 137987065100032 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:36:34.052641 138110966900544 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776623765.6160917, 'wait_for_prev_duration_secs': 27.670909643173218, 'time_between_consecutive_saves_sec': 23.412280321121216, 'checkpointer_blocking_start_time': 1776623793.2885494, 'checkpointer_blocking_duration_secs': 0.7618856430053711, 'get_old_steps_start_time': 1776623794.0504546, 'get_old_steps_duration_secs': 2.2649765014648438e-05, 'checkpoint_manager_blocking_start_time': 1776623765.614292, 'checkpoint_manager_blocking_duration_secs': 28.438319444656372}
I0419 18:36:34.052748 138110966900544 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0419 18:36:34.052789 138110966900544 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:36:39.110991 137981193090816 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_08_checkpoint_async_true/checkpoints/9/items/array_metadatas/process_6
I0419 18:37:15.913486 137982233261824 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.690 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.910178661346436 s) (per-host)
I0419 18:37:15.913603 137982233261824 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.862678s.
I0419 18:37:25.121691 137982233261824 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.070750s.
I0419 18:37:25.121917 137987065100032 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:37:25.121972 137987065100032 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:37:25.122009 137987065100032 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:37:25.123468 137987065100032 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 18:37:25.123642 138110966900544 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 18:37:25.123805 138110966900544 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 18:37:25.124760 138110966900544 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.170, Tokens/s/device: 13892.940, 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 18:37:37 UTC 2026
EXIT_CODE=0