MaxView

XPK Start: Tue Apr 21 07:37:28 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-21 07:37:53.094045: 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)
I0421 07:37:53.305858 139210259830592 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 07:38:02.345610 139210259830592 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 07:38:02.347949 139210259830592 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-os7zz-slice-job-0-0.mt-16-zarr3-true-rs-os7zz:8482
I0421 07:38:03.604861 139210259830592 max_utils.py:284] Jax distributed system initialized!
I0421 07:38:09.593822 139210259830592 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 07:38:09.593927 139210259830592 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 07:38:09.593966 139210259830592 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
I0421 07:38:09.594003 139210259830592 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 07:38:10.288410 139210259830592 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 07:38:10.288691 139210259830592 checkpointing.py:677] Setting up checkpoint logger...
I0421 07:38:10.288760 139210259830592 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0421 07:38:10.288807 139210259830592 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 07:38:10.289144 139210259830592 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 0x7e9bb2650560>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 07:38:11.533843 139210259830592 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0421 07:38:11.534083 139210259830592 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e8780517c20>}, handler_registry=None
I0421 07:38:11.534323 139210259830592 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e8780517c20>` 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`.
I0421 07:38:11.534372 139210259830592 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8780517c80>` 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`.
I0421 07:38:11.534407 139210259830592 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 0x7e8780517c20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e8780517c20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8780517c80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e8780517c80>}).
I0421 07:38:11.534724 139210259830592 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 07:38:11.534811 139210259830592 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e87502f1800> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 07:38:14.367647 139088506099456 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776756831487135447, 'commit_timestamp_nsecs': 1776756874442900067, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA
I0421 07:38:14.613161 139088506099456 checkpoint_manager.py:1871] Missing metrics for step 0
I0421 07:38:15.495881 139210259830592 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints
I0421 07:38:15.498354 139210259830592 checkpoint_manager.py:929] [process=7][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_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e8798074470>
I0421 07:38:15.498474 139210259830592 checkpointing.py:301] Checkpoint manager created!
I0421 07:38:16.432080 139210259830592 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 07:38:16.432201 139210259830592 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:38:16.817317 139210259830592 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 07:38:16.817411 139210259830592 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:38:16.833639 139210259830592 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 07:38:16.833700 139210259830592 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 07:38:16.857258 139210259830592 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 07:38:16.857327 139210259830592 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:38:16.873632 139210259830592 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 07:38:16.873691 139210259830592 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:38:16.890016 139210259830592 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 07:38:16.890081 139210259830592 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:38:16.916206 139210259830592 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 07:38:16.916286 139210259830592 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 07:38:16.938888 139210259830592 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 07:38:16.938960 139210259830592 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 07:38:17.148363 139210259830592 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 07:38:17.148473 139210259830592 checkpointing.py:581] restoring from this run's directory step 9
I0421 07:38:17.400612 139210259830592 event_tracking.py:70] [process=7] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/9.
I0421 07:38:17.633437 139210259830592 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/9.
I0421 07:38:17.633878 139210259830592 event_tracking.py:125] [process=7] [sync] Finished blocking load in 0.23 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/9.
I0421 07:38:19.038239    1460 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 07:38:25.569197 139210259830592 jax_array_handlers.py:843] [process=7] /jax/orbax/read/worker/io/requested throughput: 241.741 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.537045001983643 s) (per-host)
I0421 07:38:25.574864 139210259830592 base_pytree_checkpoint_handler.py:130] [process=7] /jax/checkpoint/read/gbytes_per_sec: 227.104 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.968593120574951 s) (per-host)
I0421 07:38:27.989347 139210259830592 event_tracking.py:138] [process=7] [sync] Finished load in 10.59 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/9
I0421 07:38:27.989508 139210259830592 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776757097.400512, 'checkpointer_duration_secs': 10.588960886001587, 'checkpoint_manager_start_time': 1776757097.1487112, 'checkpoint_manager_duration_secs': 10.840764999389648}
fsdp: 32
I0421 07:38:28.407865 139210259830592 maxtext_utils.py:1668]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0421 07:38:28.407994 139210259830592 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 07:38:28.408051 139210259830592 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 07:38:28.408117 139210259830592 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0421 07:38:28.408169 139210259830592 maxtext_utils.py:1668]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 07:38:28.408207 139210259830592 maxtext_utils.py:1668]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 07:38:28.408269 139210259830592 maxtext_utils.py:1668]  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)
I0421 07:38:28.408322 139210259830592 maxtext_utils.py:1668]  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')
I0421 07:38:28.408363 139210259830592 maxtext_utils.py:1668]  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)
I0421 07:38:28.408399 139210259830592 maxtext_utils.py:1668]  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)
I0421 07:38:28.408445 139210259830592 maxtext_utils.py:1668]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0421 07:38:28.408490 139210259830592 maxtext_utils.py:1668]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0421 07:38:28.894848 139210259830592 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 07:38:28.894944 139210259830592 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 07:38:28.910339 139210259830592 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 07:38:28.910401 139210259830592 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 07:38:39.999827 139210259830592 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.
I0421 07:38:40.000630 139210259830592 metric_logger.py:301] number parameters: 1.104 billion
I0421 07:38:51.635978 139210259830592 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0421 07:38:51.788712 139210259830592 checkpointing.py:776] Waited 0.1527111530303955 seconds for step 10 to finish before starting checkpointing.
I0421 07:38:51.791626 139210259830592 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 07:38:51.793580 139210259830592 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 10
I0421 07:38:51.795045 139210259830592 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/10.
I0421 07:38:52.545326 139210259830592 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 07:38:52.546365 139210259830592 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:38:52.546425 139210259830592 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
I0421 07:38:52.819751 139210259830592 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274489s
I0421 07:38:52.819935 139210259830592 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.538 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27855730056762695 s) (per-host)
I0421 07:38:52.819991 139210259830592 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278625s (batch_requests_ready=0.002317s, total_serialization_initiated=0.276231s, others=0.000077s)
I0421 07:38:52.820082 139210259830592 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282726s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.282709)
I0421 07:38:52.820834 139210259830592 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/10.
I0421 07:38:52.821186 139084280481536 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 07:58:52.821147
I0421 07:38:52.831244 139210259830592 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:38:52.831522 139083757381376 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:38:52.831675 139210259830592 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757131.7916076, 'wait_for_prev_duration_secs': 6.079673767089844e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776757131.7936182, 'checkpointer_blocking_duration_secs': 1.027721643447876, 'get_old_steps_start_time': 1776757132.821363, 'get_old_steps_duration_secs': 3.1948089599609375e-05, 'checkpoint_manager_blocking_start_time': 1776757131.7892175, 'checkpoint_manager_blocking_duration_secs': 1.0424220561981201}
I0421 07:38:52.831784 139210259830592 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0421 07:38:52.831833 139210259830592 max_utils.py:750] 
Memstats: After params initialized:
I0421 07:38:52.831882 139210259830592 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0))
I0421 07:38:52.831914 139210259830592 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0))
I0421 07:38:52.831940 139210259830592 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0))
I0421 07:38:52.831963 139210259830592 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0))
I0421 07:38:53.139723 139210259830592 metric_logger.py:196] completed step: 10, seconds: 11.635, TFLOP/s/device: 1.168, Tokens/s/device: 176.017, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0421 07:38:53.735582 139210259830592 metric_logger.py:196] completed step: 11, seconds: 1.502, TFLOP/s/device: 9.045, Tokens/s/device: 1363.321, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0421 07:38:53.882830 139210259830592 metric_logger.py:196] completed step: 12, seconds: 0.449, TFLOP/s/device: 30.235, Tokens/s/device: 4557.339, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0421 07:38:54.030095 139210259830592 metric_logger.py:196] completed step: 13, seconds: 0.152, TFLOP/s/device: 89.237, Tokens/s/device: 13450.765, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0421 07:38:54.036261 139210259830592 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0421 07:38:54.324036 139210259830592 checkpointing.py:776] Waited 0.2877357006072998 seconds for step 15 to finish before starting checkpointing.
I0421 07:38:54.327223 139210259830592 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 07:38:58.094651 139083765774080 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/10/items/array_metadatas/process_7
I0421 07:39:34.341952 139084280481536 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.789 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.80055093765259 s) (per-host)
I0421 07:39:34.342057 139084280481536 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.520761s.
I0421 07:39:43.129957 139084280481536 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 50.308643s.
I0421 07:39:43.130239 139083757381376 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:39:43.130356 139083757381376 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:39:43.130404 139083757381376 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:39:43.132312 139083757381376 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0421 07:39:43.132511 139210259830592 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0421 07:39:43.132651 139210259830592 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.805427 seconds. If this number is high, consider checkpointing less frequently.
I0421 07:39:43.134757 139210259830592 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 15
I0421 07:39:43.137090 139210259830592 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/15.
I0421 07:39:43.442634 139210259830592 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:39:43.442727 139210259830592 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
I0421 07:39:43.468593 139210259830592 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027686s
I0421 07:39:43.468721 139210259830592 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 49.328 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031271934509277344 s) (per-host)
I0421 07:39:43.468780 139210259830592 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031340s (batch_requests_ready=0.001936s, total_serialization_initiated=0.029329s, others=0.000074s)
I0421 07:39:43.468870 139210259830592 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036329s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.036315)
I0421 07:39:43.469549 139210259830592 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/15.
I0421 07:39:43.469829 139083757381376 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 07:59:43.469807
I0421 07:39:43.479806 139210259830592 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:39:43.480050 139083765774080 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:39:43.480177 139210259830592 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757134.3271952, 'wait_for_prev_duration_secs': 48.805426836013794, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776757183.134794, 'checkpointer_blocking_duration_secs': 0.33513641357421875, 'get_old_steps_start_time': 1776757183.4699543, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776757134.3246515, 'checkpoint_manager_blocking_duration_secs': 49.15549159049988}
I0421 07:39:43.480288 139210259830592 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0421 07:39:43.481127 139210259830592 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.256, Tokens/s/device: 13905.769, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0421 07:39:43.487373 139210259830592 metric_logger.py:196] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.155, Tokens/s/device: 13739.895, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0421 07:39:49.438943 139081087530752 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/15/items/array_metadatas/process_7
I0421 07:40:04.406790 139210259830592 metric_logger.py:196] completed step: 16, seconds: 49.450, TFLOP/s/device: 0.275, Tokens/s/device: 41.416, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0421 07:40:04.554435 139210259830592 metric_logger.py:196] completed step: 17, seconds: 20.774, TFLOP/s/device: 0.654, Tokens/s/device: 98.587, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0421 07:40:04.701830 139210259830592 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.206, Tokens/s/device: 13446.084, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0421 07:40:04.848519 139210259830592 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0421 07:40:04.849286 139210259830592 checkpointing.py:776] Waited 0.0007872581481933594 seconds for step 19 to finish before starting checkpointing.
I0421 07:40:04.851244 139210259830592 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 07:40:26.024533 139083757381376 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.091 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.58704423904419 s) (per-host)
I0421 07:40:26.024662 139083757381376 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.554755s.
I0421 07:40:32.835092 139083757381376 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 49.365158s.
I0421 07:40:32.835319 139083765774080 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:40:32.835383 139083765774080 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:40:32.835431 139083765774080 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:40:32.838062 139083765774080 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0421 07:40:32.838226 139210259830592 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0421 07:40:32.838362 139210259830592 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.987122 seconds. If this number is high, consider checkpointing less frequently.
I0421 07:40:32.839883 139210259830592 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 19
I0421 07:40:32.841850 139210259830592 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/19.
I0421 07:40:33.561542 139210259830592 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 07:40:33.561640 139210259830592 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
I0421 07:40:33.587650 139210259830592 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026949s
I0421 07:40:33.587842 139210259830592 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 50.329 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030649662017822266 s) (per-host)
I0421 07:40:33.587895 139210259830592 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030713s (batch_requests_ready=0.001819s, total_serialization_initiated=0.028826s, others=0.000068s)
I0421 07:40:33.587990 139210259830592 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034862s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=0.034852)
I0421 07:40:33.588634 139210259830592 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/19.
I0421 07:40:33.588919 139084280481536 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 08:00:33.588885
I0421 07:40:33.592180 139210259830592 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 07:40:33.592463 139083765774080 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0421 07:40:33.592619 139210259830592 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776757204.851215, 'wait_for_prev_duration_secs': 27.98712182044983, 'time_between_consecutive_saves_sec': 21.71885895729065, 'checkpointer_blocking_start_time': 1776757232.8399212, 'checkpointer_blocking_duration_secs': 0.749152660369873, 'get_old_steps_start_time': 1776757233.5890884, 'get_old_steps_duration_secs': 2.288818359375e-05, 'checkpoint_manager_blocking_start_time': 1776757204.8495262, 'checkpoint_manager_blocking_duration_secs': 28.7430636882782}
I0421 07:40:33.592725 139210259830592 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0421 07:40:33.592779 139210259830592 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 07:40:39.942575 139081087530752 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_16_zarr3_true/checkpoints/19/items/array_metadatas/process_7
I0421 07:41:15.154110 139084280481536 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.974 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.59686994552612 s) (per-host)
I0421 07:41:15.154232 139084280481536 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.565198s.
I0421 07:41:24.943087 139084280481536 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 51.354036s.
I0421 07:41:24.943358 139083765774080 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 07:41:24.943475 139083765774080 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0421 07:41:24.943520 139083765774080 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0421 07:41:24.946284 139083765774080 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0421 07:41:24.946467 139210259830592 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0421 07:41:24.946617 139210259830592 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 07:41:24.947481 139210259830592 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.107, Tokens/s/device: 13883.428, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Tue Apr 21 07:41:33 UTC 2026
EXIT_CODE=0