MaxView

XPK Start: Thu Apr 23 08:33:12 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. 
2026-04-23 08:33:37.348107: 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)
I0423 08:33:37.564880 138516683810624 max_utils.py:273] Attempting to initialize the jax distributed system...
I0423 08:33:46.607956 138516683810624 distributed.py:149] Starting JAX distributed service on [::]:8482
I0423 08:33:46.610293 138516683810624 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-385hi-slice-job-0-0.mt-15-ocdbt-false-rs-385hi:8482
I0423 08:33:47.691051 138516683810624 max_utils.py:284] Jax distributed system initialized!
I0423 08:33:53.744961 138516683810624 max_utils.py:800] System Information: Jax Version: 0.9.2
I0423 08:33:53.745068 138516683810624 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0423 08:33:53.745124 138516683810624 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
I0423 08:33:53.745162 138516683810624 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 08:33:54.434655 138516683810624 maxtext_utils.py:1604] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 08:33:54.434937 138516683810624 checkpointing.py:677] Setting up checkpoint logger...
I0423 08:33:54.434989 138516683810624 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0423 08:33:54.435033 138516683810624 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0423 08:33:54.435387 138516683810624 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7dfa4cb146b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 08:33:56.151742 138516683810624 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0423 08:33:56.151973 138516683810624 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 0x7de6340beab0>}, handler_registry=None
I0423 08:33:56.152223 138516683810624 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7de6340beab0>` 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`.
I0423 08:33:56.152282 138516683810624 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de6340c2bd0>` 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`.
I0423 08:33:56.152318 138516683810624 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 0x7de6340beab0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7de6340beab0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de6340c2bd0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7de6340c2bd0>}).
I0423 08:33:56.152648 138516683810624 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0423 08:33:56.152721 138516683810624 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7de61c5a5940> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0423 08:33:57.621878 138396613711616 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776933084257717177, 'commit_timestamp_nsecs': 1776933098681534237, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/5/_CHECKPOINT_METADATA
I0423 08:33:57.870288 138396613711616 checkpoint_manager.py:1871] Missing metrics for step 5
I0423 08:33:58.339720 138516683810624 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints
I0423 08:33:58.743520 138516683810624 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7de6340bcbf0>
I0423 08:33:58.743696 138516683810624 checkpointing.py:301] Checkpoint manager created!
I0423 08:33:59.674792 138516683810624 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 08:33:59.674927 138516683810624 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 08:34:00.053963 138516683810624 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn').
I0423 08:34:00.054055 138516683810624 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 08:34:00.070577 138516683810624 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn').
I0423 08:34:00.070637 138516683810624 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 08:34:00.094520 138516683810624 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 08:34:00.094589 138516683810624 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 08:34:00.110951 138516683810624 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 08:34:00.111012 138516683810624 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 08:34:00.127358 138516683810624 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 08:34:00.127420 138516683810624 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 08:34:00.152135 138516683810624 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv').
I0423 08:34:00.152208 138516683810624 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 08:34:00.174081 138516683810624 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 08:34:00.174159 138516683810624 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 08:34:00.385687 138516683810624 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 08:34:00.385799 138516683810624 checkpointing.py:581] restoring from this run's directory step 9
I0423 08:34:00.637878 138516683810624 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/9.
I0423 08:34:00.882969 138516683810624 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/9.
I0423 08:34:00.883424 138516683810624 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/9.
I0423 08:34:03.937086    1497 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 08:34:13.704370 138516683810624 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 145.262 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.878753185272217 s) (per-host)
I0423 08:34:13.711288 138516683810624 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 140.435 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.26926589012146 s) (per-host)
I0423 08:34:14.533459 138516683810624 event_tracking.py:138] [process=6] [sync] Finished load in 13.90 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/9
I0423 08:34:14.533616 138516683810624 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776933240.6377778, 'checkpointer_duration_secs': 13.895805597305298, 'checkpoint_manager_start_time': 1776933240.3860433, 'checkpoint_manager_duration_secs': 14.147542238235474}
fsdp: 32
I0423 08:34:14.949476 138516683810624 maxtext_utils.py:1707]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0423 08:34:14.949619 138516683810624 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 08:34:14.949674 138516683810624 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 08:34:14.949733 138516683810624 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0423 08:34:14.949786 138516683810624 maxtext_utils.py:1707]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0423 08:34:14.949825 138516683810624 maxtext_utils.py:1707]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0423 08:34:14.949877 138516683810624 maxtext_utils.py:1707]  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)
I0423 08:34:14.949929 138516683810624 maxtext_utils.py:1707]  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')
I0423 08:34:14.949969 138516683810624 maxtext_utils.py:1707]  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)
I0423 08:34:14.950005 138516683810624 maxtext_utils.py:1707]  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)
I0423 08:34:14.950051 138516683810624 maxtext_utils.py:1707]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0423 08:34:14.950114 138516683810624 maxtext_utils.py:1707]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0423 08:34:15.432203 138516683810624 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 08:34:15.432297 138516683810624 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 08:34:15.447924 138516683810624 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0423 08:34:15.447983 138516683810624 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0423 08:34:26.552638 138516683810624 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.
I0423 08:34:26.553461 138516683810624 metric_logger.py:301] number parameters: 1.104 billion
I0423 08:34:38.364846 138516683810624 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0423 08:34:38.518566 138516683810624 checkpointing.py:776] Waited 0.15369844436645508 seconds for step 10 to finish before starting checkpointing.
I0423 08:34:38.521017 138516683810624 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 08:34:38.522570 138516683810624 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0423 08:34:38.524086 138516683810624 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/10.
I0423 08:34:39.284461 138516683810624 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 08:34:39.285296 138516683810624 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 08:34:39.285353 138516683810624 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
I0423 08:34:39.571571 138516683810624 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.287186s
I0423 08:34:39.571739 138516683810624 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.296 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2912459373474121 s) (per-host)
I0423 08:34:39.571792 138516683810624 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.291310s (batch_requests_ready=0.002312s, total_serialization_initiated=0.288925s, others=0.000073s)
I0423 08:34:39.571883 138516683810624 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.295414s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.295396)
I0423 08:34:39.572621 138516683810624 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.05 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/10.
I0423 08:34:39.572963 138393375733504 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 08:54:39.572924
I0423 08:34:39.577538 138516683810624 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 08:34:39.577821 138392278320896 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 08:34:39.577970 138516683810624 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776933278.5209987, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776933278.5226083, 'checkpointer_blocking_duration_secs': 1.050529956817627, 'get_old_steps_start_time': 1776933279.5731626, 'get_old_steps_duration_secs': 3.361701965332031e-05, 'checkpoint_manager_blocking_start_time': 1776933278.5190597, 'checkpoint_manager_blocking_duration_secs': 1.0588748455047607}
I0423 08:34:39.578075 138516683810624 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0423 08:34:39.578131 138516683810624 max_utils.py:750] 
Memstats: After params initialized:
I0423 08:34:39.578179 138516683810624 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0423 08:34:39.578212 138516683810624 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0423 08:34:39.578239 138516683810624 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0423 08:34:39.578263 138516683810624 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0423 08:34:39.889642 138516683810624 metric_logger.py:196] completed step: 10, seconds: 11.811, TFLOP/s/device: 1.150, Tokens/s/device: 173.394, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0423 08:34:40.071686 138516683810624 metric_logger.py:196] completed step: 11, seconds: 1.523, TFLOP/s/device: 8.920, Tokens/s/device: 1344.536, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0423 08:34:40.219126 138516683810624 metric_logger.py:196] completed step: 12, seconds: 0.035, TFLOP/s/device: 391.673, Tokens/s/device: 59037.187, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0423 08:34:40.366203 138516683810624 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.585, Tokens/s/device: 13352.545, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0423 08:34:40.370813 138516683810624 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0423 08:34:40.660383 138516683810624 checkpointing.py:776] Waited 0.289539098739624 seconds for step 15 to finish before starting checkpointing.
I0423 08:34:40.663406 138516683810624 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 08:34:44.235119 138392286713600 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6
I0423 08:34:50.206439 138393375733504 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 144.573 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.925909042358398 s) (per-host)
I0423 08:34:50.206566 138393375733504 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.633492s.
I0423 08:34:55.871254 138393375733504 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.298163s.
I0423 08:34:55.871551 138392278320896 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 08:34:55.871666 138392278320896 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 08:34:55.871713 138392278320896 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0423 08:34:55.873430 138392278320896 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0423 08:34:55.873590 138516683810624 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.
W0423 08:34:55.873720 138516683810624 checkpoint_manager.py:1452] Waiting for previous save to complete took 15.210313 seconds. If this number is high, consider checkpointing less frequently.
I0423 08:34:55.875687 138516683810624 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0423 08:34:55.878235 138516683810624 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/15.
I0423 08:34:56.601117 138516683810624 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 08:34:56.601211 138516683810624 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
I0423 08:34:56.627542 138516683810624 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028239s
I0423 08:34:56.627685 138516683810624 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 48.464 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0318293571472168 s) (per-host)
I0423 08:34:56.627737 138516683810624 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031894s (batch_requests_ready=0.001918s, total_serialization_initiated=0.029904s, others=0.000072s)
I0423 08:34:56.627826 138516683810624 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036769s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.036758)
I0423 08:34:56.628533 138516683810624 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/15.
I0423 08:34:56.628829 138392278320896 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 08:54:56.628794
I0423 08:34:56.638857 138516683810624 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 08:34:56.639163 138388535502592 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 08:34:56.639325 138516683810624 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776933280.6633759, 'wait_for_prev_duration_secs': 15.210313320159912, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776933295.8757277, 'checkpointer_blocking_duration_secs': 0.7532479763031006, 'get_old_steps_start_time': 1776933296.6289976, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776933280.6610115, 'checkpoint_manager_blocking_duration_secs': 15.97827696800232}
I0423 08:34:56.639430 138516683810624 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0423 08:34:56.640311 138516683810624 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.144, Tokens/s/device: 13888.889, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0423 08:34:56.646792 138516683810624 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.343, Tokens/s/device: 13919.001, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0423 08:35:02.123327 138392286713600 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6
I0423 08:35:07.339658 138392278320896 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 147.024 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.743781089782715 s) (per-host)
I0423 08:35:07.339765 138392278320896 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.710825s.
I0423 08:35:13.071119 138392278320896 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.442142s.
I0423 08:35:13.071401 138388535502592 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 08:35:13.071513 138388535502592 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 08:35:13.071563 138388535502592 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0423 08:35:13.073046 138388535502592 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0423 08:35:17.775368 138516683810624 metric_logger.py:196] completed step: 16, seconds: 16.274, TFLOP/s/device: 0.835, Tokens/s/device: 125.841, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0423 08:35:17.923108 138516683810624 metric_logger.py:196] completed step: 17, seconds: 20.982, TFLOP/s/device: 0.648, Tokens/s/device: 97.605, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0423 08:35:18.070327 138516683810624 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.975, Tokens/s/device: 13411.216, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0423 08:35:18.216834 138516683810624 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0423 08:35:18.217529 138516683810624 checkpointing.py:776] Waited 0.0007145404815673828 seconds for step 19 to finish before starting checkpointing.
I0423 08:35:18.219589 138516683810624 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 08:35:18.221171 138516683810624 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0423 08:35:18.222685 138516683810624 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/19.
I0423 08:35:18.929416 138516683810624 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0423 08:35:18.929516 138516683810624 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
I0423 08:35:18.953633 138516683810624 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025253s
I0423 08:35:18.953759 138516683810624 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 53.422 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.028875350952148438 s) (per-host)
I0423 08:35:18.953805 138516683810624 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.028930s (batch_requests_ready=0.001929s, total_serialization_initiated=0.026942s, others=0.000059s)
I0423 08:35:18.953884 138516683810624 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.032929s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.032917)
I0423 08:35:18.954620 138516683810624 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.73 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/19.
I0423 08:35:18.954885 138388535502592 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 08:55:18.954862
I0423 08:35:18.960535 138516683810624 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 08:35:18.960750 138387952482048 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0423 08:35:18.960865 138516683810624 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776933318.2195566, 'wait_for_prev_duration_secs': 7.224082946777344e-05, 'time_between_consecutive_saves_sec': 5.146472692489624, 'checkpointer_blocking_start_time': 1776933318.2212083, 'checkpointer_blocking_duration_secs': 0.7337803840637207, 'get_old_steps_start_time': 1776933318.955007, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776933318.2177682, 'checkpoint_manager_blocking_duration_secs': 0.7430620193481445}
I0423 08:35:18.960983 138516683810624 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0423 08:35:18.961030 138516683810624 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 08:35:24.046237 138392286713600 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260423_071538/linen_xpk_main_20260423_071538_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6
I0423 08:35:29.116051 138388535502592 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 154.997 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.191127061843872 s) (per-host)
I0423 08:35:29.116193 138388535502592 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.161227s.
I0423 08:35:34.740356 138388535502592 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 15.785375s.
I0423 08:35:34.740575 138387952482048 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 08:35:34.740641 138387952482048 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0423 08:35:34.740690 138387952482048 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0423 08:35:34.743121 138387952482048 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0423 08:35:34.743291 138516683810624 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.
I0423 08:35:34.743443 138516683810624 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 08:35:34.744546 138516683810624 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.164, Tokens/s/device: 13891.998, 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: Thu Apr 23 08:35:47 UTC 2026
EXIT_CODE=0