MaxView

XPK Start: Wed Apr 22 22:58:21 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. 
2026-04-22 22:58:46.744134: 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)
I0422 22:58:46.956295 140227760072512 max_utils.py:273] Attempting to initialize the jax distributed system...
I0422 22:58:55.998512 140227760072512 distributed.py:149] Starting JAX distributed service on [::]:8482
I0422 22:58:56.000867 140227760072512 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-k5b86-slice-job-0-0.mt-16-zarr3-false-rs-k5b86:8482
I0422 22:58:56.770401 140227760072512 max_utils.py:284] Jax distributed system initialized!
I0422 22:59:03.043307 140227760072512 max_utils.py:800] System Information: Jax Version: 0.9.2
I0422 22:59:03.043420 140227760072512 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0422 22:59:03.043462 140227760072512 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
I0422 22:59:03.043500 140227760072512 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 22:59:03.747885 140227760072512 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 22:59:03.748187 140227760072512 checkpointing.py:677] Setting up checkpoint logger...
I0422 22:59:03.748245 140227760072512 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False
I0422 22:59:03.748290 140227760072512 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0422 22:59:03.748655 140227760072512 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 0x7f889a2480e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 22:59:05.020761 140227760072512 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0422 22:59:05.021001 140227760072512 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 0x7f745c2979e0>}, handler_registry=None
I0422 22:59:05.021259 140227760072512 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f745c2979e0>` 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`.
I0422 22:59:05.021311 140227760072512 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f745c298f50>` 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`.
I0422 22:59:05.021353 140227760072512 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 0x7f745c2979e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f745c2979e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f745c298f50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f745c298f50>}).
I0422 22:59:05.021680 140227760072512 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0422 22:59:05.021750 140227760072512 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f745c0cb560> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0422 22:59:07.626848 140106312701696 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776898524951548322, 'commit_timestamp_nsecs': 1776898573136110208, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA
I0422 22:59:07.871170 140106312701696 checkpoint_manager.py:1871] Missing metrics for step 5
I0422 22:59:08.435749 140227760072512 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints
I0422 22:59:09.211269 140227760072512 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_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f745c2966c0>
I0422 22:59:09.211450 140227760072512 checkpointing.py:301] Checkpoint manager created!
I0422 22:59:10.142543 140227760072512 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 22:59:10.142664 140227760072512 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 22:59:10.527256 140227760072512 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 22:59:10.527354 140227760072512 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 22:59:10.544089 140227760072512 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 22:59:10.544162 140227760072512 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 22:59:10.567974 140227760072512 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 22:59:10.568040 140227760072512 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 22:59:10.584804 140227760072512 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 22:59:10.584868 140227760072512 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 22:59:10.601369 140227760072512 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 22:59:10.601434 140227760072512 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 22:59:10.626415 140227760072512 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 22:59:10.626483 140227760072512 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 22:59:10.648693 140227760072512 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 22:59:10.648765 140227760072512 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 22:59:10.860306 140227760072512 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 22:59:10.860422 140227760072512 checkpointing.py:581] restoring from this run's directory step 9
I0422 22:59:11.106711 140227760072512 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/9.
I0422 22:59:11.778502 140227760072512 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/9.
I0422 22:59:11.778938 140227760072512 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.67 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/9.
I0422 22:59:13.201628    1457 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 22:59:19.619686 140227760072512 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 245.989 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.424152851104736 s) (per-host)
I0422 22:59:19.633271 140227760072512 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 230.606 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.8627495765686035 s) (per-host)
I0422 22:59:21.611058 140227760072512 event_tracking.py:138] [process=6] [sync] Finished load in 10.50 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/9
I0422 22:59:21.611271 140227760072512 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776898751.106615, 'checkpointer_duration_secs': 10.504614114761353, 'checkpoint_manager_start_time': 1776898750.860669, 'checkpoint_manager_duration_secs': 10.750563144683838}
fsdp: 32
I0422 22:59:22.032927 140227760072512 maxtext_utils.py:1668]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0422 22:59:22.033055 140227760072512 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)
I0422 22:59:22.033121 140227760072512 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)
I0422 22:59:22.033185 140227760072512 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')
I0422 22:59:22.033240 140227760072512 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)
I0422 22:59:22.033286 140227760072512 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)
I0422 22:59:22.033344 140227760072512 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)
I0422 22:59:22.033397 140227760072512 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')
I0422 22:59:22.033437 140227760072512 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)
I0422 22:59:22.033474 140227760072512 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)
I0422 22:59:22.033521 140227760072512 maxtext_utils.py:1668]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0422 22:59:22.033568 140227760072512 maxtext_utils.py:1668]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0422 22:59:22.518821 140227760072512 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 22:59:22.518912 140227760072512 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 22:59:22.534468 140227760072512 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 22:59:22.534525 140227760072512 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 22:59:33.558912 140227760072512 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.
I0422 22:59:33.559714 140227760072512 metric_logger.py:301] number parameters: 1.104 billion
I0422 22:59:45.713625 140227760072512 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0422 22:59:45.866360 140227760072512 checkpointing.py:776] Waited 0.15271306037902832 seconds for step 10 to finish before starting checkpointing.
I0422 22:59:45.868734 140227760072512 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 22:59:45.870640 140227760072512 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0422 22:59:45.872346 140227760072512 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/10.
I0422 22:59:46.620854 140227760072512 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 22:59:46.621738 140227760072512 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 22:59:46.621796 140227760072512 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
I0422 22:59:46.897997 140227760072512 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277217s
I0422 22:59:46.898181 140227760072512 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.485 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28125810623168945 s) (per-host)
I0422 22:59:46.898242 140227760072512 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281326s (batch_requests_ready=0.002331s, total_serialization_initiated=0.278918s, others=0.000076s)
I0422 22:59:46.898343 140227760072512 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285411s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.285395)
I0422 22:59:46.899081 140227760072512 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/10.
I0422 22:59:46.899387 140105524180736 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 23:19:46.899354
I0422 22:59:46.903714 140227760072512 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 22:59:46.904005 140102235846400 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 22:59:46.904170 140227760072512 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776898785.868711, 'wait_for_prev_duration_secs': 7.724761962890625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776898785.870683, 'checkpointer_blocking_duration_secs': 1.0288596153259277, 'get_old_steps_start_time': 1776898786.8995686, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776898785.8667948, 'checkpoint_manager_blocking_duration_secs': 1.0373389720916748}
I0422 22:59:46.904274 140227760072512 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0422 22:59:46.904323 140227760072512 max_utils.py:750] 
Memstats: After params initialized:
I0422 22:59:46.904375 140227760072512 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0422 22:59:46.904406 140227760072512 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0422 22:59:46.904433 140227760072512 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0422 22:59:46.904457 140227760072512 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0422 22:59:47.220299 140227760072512 metric_logger.py:196] completed step: 10, seconds: 12.154, TFLOP/s/device: 1.118, Tokens/s/device: 168.507, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0422 22:59:47.394074 140227760072512 metric_logger.py:196] completed step: 11, seconds: 1.505, TFLOP/s/device: 9.028, Tokens/s/device: 1360.739, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0422 22:59:47.541487 140227760072512 metric_logger.py:196] completed step: 12, seconds: 0.026, TFLOP/s/device: 514.197, Tokens/s/device: 77505.298, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0422 22:59:47.688797 140227760072512 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.544, Tokens/s/device: 13346.367, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0422 22:59:47.693486 140227760072512 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0422 22:59:47.982926 140227760072512 checkpointing.py:776] Waited 0.28940343856811523 seconds for step 15 to finish before starting checkpointing.
I0422 22:59:47.986016 140227760072512 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 22:59:52.248658 140102722361088 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/10/items/array_metadatas/process_6
I0422 23:00:28.241824 140105524180736 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.948 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.624866247177124 s) (per-host)
I0422 23:00:28.241951 140105524180736 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.342452s.
I0422 23:00:36.620470 140105524180736 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.720954s.
I0422 23:00:36.620784 140102235846400 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 23:00:36.620928 140102235846400 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 23:00:36.620988 140102235846400 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0422 23:00:36.622571 140102235846400 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0422 23:00:36.622761 140227760072512 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0422 23:00:36.622840 140227760072512 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.636821 seconds. If this number is high, consider checkpointing less frequently.
I0422 23:00:36.624904 140227760072512 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0422 23:00:36.627337 140227760072512 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/15.
I0422 23:00:37.345247 140227760072512 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 23:00:37.345341 140227760072512 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
I0422 23:00:37.371284 140227760072512 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027842s
I0422 23:00:37.371454 140227760072512 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 49.055 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03144550323486328 s) (per-host)
I0422 23:00:37.371510 140227760072512 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031506s (batch_requests_ready=0.001914s, total_serialization_initiated=0.029526s, others=0.000067s)
I0422 23:00:37.371608 140227760072512 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036216s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.036204)
I0422 23:00:37.372267 140227760072512 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/15.
I0422 23:00:37.372596 140102722361088 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 23:20:37.372548
I0422 23:00:37.385213 140227760072512 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 23:00:37.385446 140102235846400 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 23:00:37.385582 140227760072512 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776898787.9859874, 'wait_for_prev_duration_secs': 48.636821031570435, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776898836.6249428, 'checkpointer_blocking_duration_secs': 0.7477641105651855, 'get_old_steps_start_time': 1776898837.3727283, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776898787.9835763, 'checkpoint_manager_blocking_duration_secs': 49.401973724365234}
I0422 23:00:37.385691 140227760072512 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0422 23:00:37.386503 140227760072512 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.113, Tokens/s/device: 13884.275, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0422 23:00:37.392537 140227760072512 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.234, Tokens/s/device: 13902.560, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0422 23:00:44.062180 140100683949824 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/15/items/array_metadatas/process_6
I0422 23:00:58.424198 140227760072512 metric_logger.py:196] completed step: 16, seconds: 49.698, TFLOP/s/device: 0.273, Tokens/s/device: 41.209, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0422 23:00:58.571683 140227760072512 metric_logger.py:196] completed step: 17, seconds: 20.886, TFLOP/s/device: 0.651, Tokens/s/device: 98.058, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0422 23:00:58.718967 140227760072512 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.172, Tokens/s/device: 13440.966, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0422 23:00:58.865652 140227760072512 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0422 23:00:58.866376 140227760072512 checkpointing.py:776] Waited 0.0007455348968505859 seconds for step 19 to finish before starting checkpointing.
I0422 23:00:58.868406 140227760072512 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 23:01:20.306394 140102722361088 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.764 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.9663450717926 s) (per-host)
I0422 23:01:20.306522 140102722361088 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.933383s.
I0422 23:01:28.722824 140102722361088 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.349669s.
I0422 23:01:28.723120 140102235846400 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 23:01:28.723240 140102235846400 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 23:01:28.723289 140102235846400 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0422 23:01:28.724953 140102235846400 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0422 23:01:28.725133 140227760072512 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0422 23:01:28.725253 140227760072512 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.856856 seconds. If this number is high, consider checkpointing less frequently.
I0422 23:01:28.726753 140227760072512 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0422 23:01:28.728846 140227760072512 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/19.
I0422 23:01:29.439836 140227760072512 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 23:01:29.439930 140227760072512 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
I0422 23:01:29.465652 140227760072512 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026695s
I0422 23:01:29.465778 140227760072512 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 50.923 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030292272567749023 s) (per-host)
I0422 23:01:29.465824 140227760072512 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030348s (batch_requests_ready=0.001804s, total_serialization_initiated=0.028484s, others=0.000060s)
I0422 23:01:29.465894 140227760072512 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034323s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.034310)
I0422 23:01:29.466590 140227760072512 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/19.
I0422 23:01:29.466941 140102235846400 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 23:21:29.466901
I0422 23:01:29.470342 140227760072512 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 23:01:29.470595 140099494209280 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 23:01:29.470731 140227760072512 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776898858.868375, 'wait_for_prev_duration_secs': 29.85685634613037, 'time_between_consecutive_saves_sec': 22.245704174041748, 'checkpointer_blocking_start_time': 1776898888.7267933, 'checkpointer_blocking_duration_secs': 0.7403147220611572, 'get_old_steps_start_time': 1776898889.4671247, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776898858.8666198, 'checkpoint_manager_blocking_duration_secs': 30.6040780544281}
I0422 23:01:29.470849 140227760072512 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0422 23:01:29.470892 140227760072512 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 23:01:35.113003 140100683949824 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_16_zarr3_false/checkpoints/19/items/array_metadatas/process_6
I0422 23:02:11.926273 140102235846400 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.175 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.490745306015015 s) (per-host)
I0422 23:02:11.926391 140102235846400 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.459326s.
I0422 23:02:20.039464 140102235846400 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.572384s.
I0422 23:02:20.039709 140099494209280 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 23:02:20.039768 140099494209280 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 23:02:20.039812 140099494209280 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0422 23:02:20.041829 140099494209280 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0422 23:02:20.042016 140227760072512 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0422 23:02:20.042187 140227760072512 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 23:02:20.043239 140227760072512 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.236, Tokens/s/device: 13902.843, 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: Wed Apr 22 23:02:32 UTC 2026
EXIT_CODE=0