XPK Start: Sat Apr 18 01:53:15 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-18 01:53:39.254207: 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)
I0418 01:53:39.789937 138627631273792 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 01:53:48.829694 138627631273792 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 01:53:48.832076 138627631273792 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-nxn2o-slice-job-0-0.mt-16-zarr3-false-sv-nxn2o:8482
I0418 01:53:49.913203 138627631273792 max_utils.py:284] Jax distributed system initialized!
I0418 01:53:55.852209 138627631273792 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 01:53:55.852310 138627631273792 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 01:53:55.852349 138627631273792 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
I0418 01:53:55.852384 138627631273792 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 01:53:56.532265 138627631273792 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 01:53:56.532536 138627631273792 checkpointing.py:677] Setting up checkpoint logger...
I0418 01:53:56.532588 138627631273792 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0418 01:53:56.532632 138627631273792 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 01:53:56.532983 138627631273792 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e140b310e90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 01:54:02.832321 138627631273792 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 01:54:02.832554 138627631273792 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dfff4296810>}, handler_registry=None
I0418 01:54:02.832796 138627631273792 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dfff4296810>` 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`.
I0418 01:54:02.832864 138627631273792 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dfff42a2060>` 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`.
I0418 01:54:02.832902 138627631273792 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 0x7dfff4296810>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dfff4296810>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dfff42a2060>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dfff42a2060>}).
I0418 01:54:02.833287 138627631273792 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 01:54:02.833358 138627631273792 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dff145b3c40> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 01:54:04.388395 138627631273792 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints
I0418 01:54:05.230921 138627631273792 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dfff42976b0>
I0418 01:54:05.231092 138627631273792 checkpointing.py:301] Checkpoint manager created!
I0418 01:54:06.134897 138627631273792 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 01:54:06.135005 138627631273792 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:54:06.517283 138627631273792 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 01:54:06.517374 138627631273792 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:54:06.533970 138627631273792 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 01:54:06.534032 138627631273792 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 01:54:06.558952 138627631273792 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 01:54:06.559021 138627631273792 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:54:06.576908 138627631273792 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 01:54:06.576969 138627631273792 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:54:06.593233 138627631273792 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 01:54:06.593292 138627631273792 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:54:06.618650 138627631273792 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 01:54:06.618723 138627631273792 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 01:54:06.639658 138627631273792 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 01:54:06.639720 138627631273792 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 01:54:06.856893 138627631273792 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 01:54:06.856999 138627631273792 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 01:54:08.270377 138627631273792 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 01:54:08.270534 138627631273792 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)
I0418 01:54:08.270612 138627631273792 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)
I0418 01:54:08.270684 138627631273792 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')
I0418 01:54:08.270739 138627631273792 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)
I0418 01:54:08.270778 138627631273792 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)
I0418 01:54:08.270829 138627631273792 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)
I0418 01:54:08.270919 138627631273792 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')
I0418 01:54:08.270966 138627631273792 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)
I0418 01:54:08.271005 138627631273792 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)
I0418 01:54:08.271054 138627631273792 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 01:54:08.271109 138627631273792 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 01:54:08.759948 138627631273792 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 01:54:08.760040 138627631273792 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 01:54:08.775321 138627631273792 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 01:54:08.775382 138627631273792 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 01:54:19.821353 138627631273792 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.
I0418 01:54:19.829136 138627631273792 metric_logger.py:301] number parameters: 1.104 billion
I0418 01:54:31.358527 138627631273792 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 01:54:31.511605 138627631273792 checkpointing.py:776] Waited 0.15305805206298828 seconds for step 0 to finish before starting checkpointing.
I0418 01:54:31.514196 138627631273792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 01:54:31.516150 138627631273792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 01:54:31.517600 138627631273792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0.
I0418 01:54:32.277048 138627631273792 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 01:54:32.291256 138627631273792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:54:32.291324 138627631273792 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
I0418 01:54:32.564001 138627631273792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273451s
I0418 01:54:32.564873 138627631273792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.512 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2798607349395752 s) (per-host)
I0418 01:54:32.564956 138627631273792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.279970s (batch_requests_ready=0.002200s, total_serialization_initiated=0.276969s, others=0.000801s)
I0418 01:54:32.565041 138627631273792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283988s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.283969)
I0418 01:54:32.565863 138627631273792 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.05 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0.
I0418 01:54:32.566164 138500223608576 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:14:32.566130
I0418 01:54:32.577774 138627631273792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:54:32.578086 138498078127872 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:54:32.578244 138627631273792 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776477271.514174, 'wait_for_prev_duration_secs': 7.963180541992188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776477271.5161974, 'checkpointer_blocking_duration_secs': 1.0501089096069336, 'get_old_steps_start_time': 1776477272.5663345, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1776477271.5121605, 'checkpoint_manager_blocking_duration_secs': 1.0660440921783447}
I0418 01:54:32.578346 138627631273792 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 01:54:32.578398 138627631273792 max_utils.py:750]
Memstats: After params initialized:
I0418 01:54:32.578447 138627631273792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 01:54:32.578480 138627631273792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 01:54:32.578508 138627631273792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 01:54:32.578537 138627631273792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 01:54:32.912986 138627631273792 metric_logger.py:196] completed step: 0, seconds: 11.241, TFLOP/s/device: 1.209, Tokens/s/device: 182.198, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 01:54:32.916627 138627631273792 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/tensorboard/'
I0418 01:54:32.955909 138499743282944 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0
I0418 01:54:33.326379 138627631273792 metric_logger.py:196] completed step: 1, seconds: 1.553, TFLOP/s/device: 8.751, Tokens/s/device: 1319.051, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875
I0418 01:54:33.472273 138627631273792 metric_logger.py:196] completed step: 2, seconds: 0.414, TFLOP/s/device: 32.814, Tokens/s/device: 4946.143, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719
I0418 01:54:33.621965 138627631273792 metric_logger.py:196] completed step: 3, seconds: 0.010, TFLOP/s/device: 1428.720, Tokens/s/device: 215352.261, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506
I0418 01:54:33.631523 138627631273792 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0418 01:54:33.916150 138627631273792 checkpointing.py:776] Waited 0.2845885753631592 seconds for step 5 to finish before starting checkpointing.
I0418 01:54:33.919191 138627631273792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:54:35.099777 138499726497536 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/items
I0418 01:54:35.253098 138496477415168 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776477274848008789, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776477274848008789, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:54:36.221679 2793 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 01:54:38.054665 138498094913280 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/items/array_metadatas/process_0
I0418 01:55:18.360315 138498086520576 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.137510s (commit=40.896615s, array_metadata_write=1.240895s)
I0418 01:55:18.361877 138500223608576 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.282 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.076823711395264 s) (per-host)
I0418 01:55:18.362045 138500223608576 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.795768s.
I0418 01:55:19.063691 138500223608576 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776477274848008789, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:55:20.668560 138496477415168 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776477274848008789, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA
I0418 01:55:21.433372 138500223608576 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000429s.
I0418 01:55:22.223286 138500223608576 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.603240s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/items
I0418 01:55:22.225479 138500223608576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0/items
I0418 01:55:22.753366 138500223608576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0
I0418 01:55:24.112576 138500223608576 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0`.
I0418 01:55:24.113534 138500223608576 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.60 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/0
I0418 01:55:24.115079 138500223608576 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.548800s.
I0418 01:55:24.115369 138498078127872 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:55:24.115533 138498078127872 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:55:24.115647 138498078127872 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:55:24.117385 138498078127872 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 01:55:24.117701 138627631273792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0418 01:55:24.117905 138627631273792 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.198717 seconds. If this number is high, consider checkpointing less frequently.
I0418 01:55:24.119639 138627631273792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 5
I0418 01:55:24.122094 138627631273792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5.
I0418 01:55:25.256670 138627631273792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:55:25.256803 138627631273792 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
I0418 01:55:25.295669 138627631273792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040030s
I0418 01:55:25.296560 138627631273792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 31.578 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.048850297927856445 s) (per-host)
I0418 01:55:25.296668 138627631273792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.048984s (batch_requests_ready=0.003996s, total_serialization_initiated=0.044217s, others=0.000771s)
I0418 01:55:25.296798 138627631273792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.053850s (all_items=0.000026s, per_item={'items': '0.00002599'}, temp_paths=0.053824)
I0418 01:55:25.297881 138627631273792 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.18 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5.
I0418 01:55:25.298341 138498078127872 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:15:25.298299
I0418 01:55:25.300852 138627631273792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:55:25.301181 138498124277504 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:55:25.301406 138627631273792 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776477273.9191425, 'wait_for_prev_duration_secs': 50.19871735572815, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776477324.119707, 'checkpointer_blocking_duration_secs': 1.1788206100463867, 'get_old_steps_start_time': 1776477325.2985554, 'get_old_steps_duration_secs': 4.100799560546875e-05, 'checkpoint_manager_blocking_start_time': 1776477273.9168215, 'checkpoint_manager_blocking_duration_secs': 51.38453245162964}
I0418 01:55:25.301555 138627631273792 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0418 01:55:25.919929 138500223608576 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5
I0418 01:55:28.049238 138499206412032 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5/items
I0418 01:55:31.977519 138498662283008 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5/items/array_metadatas/process_0
I0418 01:55:47.512397 138627631273792 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.592, Tokens/s/device: 13956.468, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355
I0418 01:55:47.521319 138627631273792 metric_logger.py:196] completed step: 5, seconds: 0.150, TFLOP/s/device: 90.633, Tokens/s/device: 13661.166, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467
I0418 01:55:47.666938 138627631273792 metric_logger.py:196] completed step: 6, seconds: 73.888, TFLOP/s/device: 0.184, Tokens/s/device: 27.717, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740
I0418 01:55:47.814275 138627631273792 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1866.108, Tokens/s/device: 281280.044, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151
I0418 01:55:47.961514 138627631273792 metric_logger.py:196] completed step: 8, seconds: 0.146, TFLOP/s/device: 92.797, Tokens/s/device: 13987.351, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098
I0418 01:55:48.108145 138627631273792 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 01:55:48.108916 138627631273792 checkpointing.py:776] Waited 0.0007917881011962891 seconds for step 9 to finish before starting checkpointing.
I0418 01:55:48.110928 138627631273792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:56:12.973746 138498653890304 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.417711s (commit=40.944474s, array_metadata_write=2.473237s)
I0418 01:56:12.974795 138498078127872 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.096 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.727089643478394 s) (per-host)
I0418 01:56:12.974904 138498078127872 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.676417s.
I0418 01:56:15.720406 138498078127872 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000229s.
I0418 01:56:16.194086 138498078127872 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.908014s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5/items
I0418 01:56:16.202383 138498078127872 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5/items
I0418 01:56:16.795641 138498078127872 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5
I0418 01:56:18.185379 138498078127872 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5`.
I0418 01:56:18.186225 138498078127872 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 54.07 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/5
I0418 01:56:18.187586 138498078127872 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.889097s.
I0418 01:56:18.187781 138498124277504 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:56:18.187928 138498124277504 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:56:18.188004 138498124277504 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:56:18.189928 138498124277504 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0418 01:56:18.190094 138627631273792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0418 01:56:18.190221 138627631273792 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.079301 seconds. If this number is high, consider checkpointing less frequently.
I0418 01:56:18.191611 138627631273792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 01:56:18.193463 138627631273792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9.
I0418 01:56:18.912457 138627631273792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 01:56:18.912553 138627631273792 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
I0418 01:56:18.939172 138627631273792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027613s
I0418 01:56:18.940077 138627631273792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.096 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.032073020935058594 s) (per-host)
I0418 01:56:18.940156 138627631273792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032173s (batch_requests_ready=0.001771s, total_serialization_initiated=0.029581s, others=0.000821s)
I0418 01:56:18.940238 138627631273792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036134s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.036116)
I0418 01:56:18.940959 138627631273792 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9.
I0418 01:56:18.941261 138498124277504 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:16:18.941225
I0418 01:56:18.948887 138627631273792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 01:56:18.949173 138498662283008 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 01:56:18.949336 138627631273792 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776477348.1108928, 'wait_for_prev_duration_secs': 30.07930064201355, 'time_between_consecutive_saves_sec': 23.993435621261597, 'checkpointer_blocking_start_time': 1776477378.191649, 'checkpointer_blocking_duration_secs': 0.7497427463531494, 'get_old_steps_start_time': 1776477378.9414105, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776477348.1091416, 'checkpoint_manager_blocking_duration_secs': 30.840160369873047}
I0418 01:56:18.949457 138627631273792 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 01:56:18.949502 138627631273792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 01:56:19.148940 138498078127872 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9
I0418 01:56:21.756809 138500223608576 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9/items
I0418 01:56:25.513459 138499198555904 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9/items/array_metadatas/process_0
I0418 01:57:05.566116 138499172865792 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.236758s (commit=40.191262s, array_metadata_write=2.045496s)
I0418 01:57:05.567210 138498124277504 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.854 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.65917444229126 s) (per-host)
I0418 01:57:05.567344 138498124277504 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.625975s.
I0418 01:57:07.877566 138498124277504 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000236s.
I0418 01:57:08.633234 138498124277504 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.185752s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9/items
I0418 01:57:08.634912 138498124277504 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9/items
I0418 01:57:09.147499 138498124277504 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9
I0418 01:57:10.525426 138498124277504 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9`.
I0418 01:57:10.526303 138498124277504 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.33 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_16_zarr3_false/checkpoints/9
I0418 01:57:10.527605 138498124277504 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.586233s.
I0418 01:57:10.527817 138498662283008 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 01:57:10.527948 138498662283008 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 01:57:10.528004 138498662283008 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 01:57:10.530406 138498662283008 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 01:57:10.530548 138627631273792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0418 01:57:10.530654 138627631273792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 01:57:10.531640 138627631273792 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.247, Tokens/s/device: 13904.542, 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: Sat Apr 18 01:57:21 UTC 2026
EXIT_CODE=0