MaxView

XPK Start: Wed Apr 22 15:23:23 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-22 15:23:47.985979: 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 15:23:48.168314 132726446835520 max_utils.py:273] Attempting to initialize the jax distributed system...
I0422 15:23:57.210385 132726446835520 distributed.py:149] Starting JAX distributed service on [::]:8482
I0422 15:23:57.212872 132726446835520 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-8n5n9-slice-job-0-0.mt-15-ocdbt-false-rs-8n5n9:8482
I0422 15:23:58.744604 132726446835520 max_utils.py:284] Jax distributed system initialized!
I0422 15:24:04.040840 132726446835520 max_utils.py:800] System Information: Jax Version: 0.9.2
I0422 15:24:04.040943 132726446835520 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0422 15:24:04.040982 132726446835520 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 15:24:04.041015 132726446835520 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 15:24:04.745777 132726446835520 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 15:24:04.857776 132726446835520 checkpointing.py:688] Setting up checkpoint logger...
I0422 15:24:04.857891 132726446835520 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True
I0422 15:24:04.857937 132726446835520 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0422 15:24:04.858151 132726446835520 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 0x78b610e4c050>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 15:24:06.164223 132726446835520 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 15:24:06.164466 132726446835520 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 0x78a2a41197c0>}, handler_registry=None
I0422 15:24:06.164707 132726446835520 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a2a41197c0>` 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 15:24:06.164753 132726446835520 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a2a47a0200>` 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 15:24:06.164788 132726446835520 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 0x78a2a41197c0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a2a41197c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a2a47a0200>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a2a47a0200>}).
I0422 15:24:06.165116 132726446835520 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0422 15:24:06.165190 132726446835520 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78a2a469dda0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0422 15:24:07.813812 132608113501952 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776871276583689195, 'commit_timestamp_nsecs': 1776871291530614583, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0422 15:24:08.057898 132608113501952 checkpoint_manager.py:1871] Missing metrics for step 0
I0422 15:24:08.477690 132726446835520 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints
I0422 15:24:08.588393 132726446835520 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/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78a2a40fc500>
I0422 15:24:08.588566 132726446835520 checkpointing.py:302] Checkpoint manager created!
I0422 15:24:09.555056 132726446835520 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 15:24:09.555186 132726446835520 checkpointing.py:582] restoring from this run's directory step 9
I0422 15:24:09.835521 132726446835520 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/9.
I0422 15:24:10.083691 132726446835520 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/9.
I0422 15:24:10.084125 132726446835520 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/9.
I0422 15:24:13.483865    1438 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 15:24:23.225707 132726446835520 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 141.560 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.163331270217896 s) (per-host)
I0422 15:24:23.235657 132726446835520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 131.683 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.018321752548218 s) (per-host)
I0422 15:24:23.943707 132726446835520 event_tracking.py:138] [process=6] [sync] Finished load in 14.11 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/9
I0422 15:24:23.943862 132726446835520 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776871449.835424, 'checkpointer_duration_secs': 14.108405590057373, 'checkpoint_manager_start_time': 1776871449.5558102, 'checkpoint_manager_duration_secs': 14.388021230697632}
fsdp: 32
I0422 15:24:24.019759 132726446835520 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0422 15:24:24.019870 132726446835520 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 15:24:24.019920 132726446835520 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 15:24:24.019961 132726446835520 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0422 15:24:24.019996 132726446835520 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 15:24:24.020029 132726446835520 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 15:24:24.020064 132726446835520 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 15:24:24.020110 132726446835520 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0422 15:24:24.020147 132726446835520 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 15:24:24.020179 132726446835520 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 15:24:24.020211 132726446835520 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0422 15:24:24.020243 132726446835520 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0422 15:24:24.422121 132726446835520 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 15:24:24.422214 132726446835520 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 15:24:24.427607 132726446835520 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 15:24:24.427662 132726446835520 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 15:24:24.444290 132726446835520 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 15:24:24.444354 132726446835520 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 15:24:24.460457 132726446835520 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 15:24:24.460516 132726446835520 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 15:24:24.484651 132726446835520 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 15:24:24.484718 132726446835520 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 15:24:24.501085 132726446835520 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 15:24:24.501206 132726446835520 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 15:24:24.517414 132726446835520 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 15:24:24.517472 132726446835520 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 15:24:24.547787 132726446835520 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 15:24:24.547856 132726446835520 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 15:24:24.567828 132726446835520 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 15:24:24.567890 132726446835520 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 15:24:35.767439 132726446835520 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 15:24:35.770376 132726446835520 metric_logger.py:301] number parameters: 1.104 billion
I0422 15:24:47.194297 132726446835520 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0422 15:24:47.345467 132726446835520 checkpointing.py:798] Waited 0.1511554718017578 seconds for step 10 to finish before starting checkpointing.
I0422 15:24:47.347640 132726446835520 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 15:24:47.349635 132726446835520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0422 15:24:47.350990 132726446835520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/10.
I0422 15:24:47.680057 132726446835520 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 15:24:47.680979 132726446835520 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 15:24:47.681035 132726446835520 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 15:24:47.966620 132726446835520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.287323s
I0422 15:24:47.966794 132726446835520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.241 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2943277359008789 s) (per-host)
I0422 15:24:47.966846 132726446835520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.294389s (batch_requests_ready=0.002716s, total_serialization_initiated=0.291605s, others=0.000069s)
I0422 15:24:47.966941 132726446835520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.298483s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.298466)
I0422 15:24:47.967696 132726446835520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.62 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/10.
I0422 15:24:47.967985 132604279904000 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:44:47.967953
I0422 15:24:47.969950 132726446835520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 15:24:47.970256 132604247799552 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 15:24:47.970409 132726446835520 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776871487.3476212, 'wait_for_prev_duration_secs': 6.008148193359375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776871487.3496728, 'checkpointer_blocking_duration_secs': 0.6184742450714111, 'get_old_steps_start_time': 1776871487.968169, 'get_old_steps_duration_secs': 3.457069396972656e-05, 'checkpoint_manager_blocking_start_time': 1776871487.3458478, 'checkpoint_manager_blocking_duration_secs': 0.6245260238647461}
I0422 15:24:47.970515 132726446835520 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0422 15:24:47.970593 132726446835520 max_utils.py:750] 
Memstats: After params initialized:
I0422 15:24:47.970640 132726446835520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0422 15:24:47.970671 132726446835520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0422 15:24:47.970696 132726446835520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0422 15:24:47.970720 132726446835520 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0422 15:24:48.286004 132726446835520 metric_logger.py:196] completed step: 10, seconds: 11.423, TFLOP/s/device: 1.189, Tokens/s/device: 179.290, total_weights: 65536, loss: 6.827, lm_loss: 6.827, perplexity: 922.537
I0422 15:24:48.473323 132726446835520 metric_logger.py:196] completed step: 11, seconds: 1.090, TFLOP/s/device: 12.467, Tokens/s/device: 1879.104, total_weights: 65536, loss: 6.522, lm_loss: 6.522, perplexity: 679.821
I0422 15:24:48.621116 132726446835520 metric_logger.py:196] completed step: 12, seconds: 0.031, TFLOP/s/device: 445.392, Tokens/s/device: 67134.334, total_weights: 65536, loss: 6.260, lm_loss: 6.260, perplexity: 523.013
I0422 15:24:48.768718 132726446835520 metric_logger.py:196] completed step: 13, seconds: 0.165, TFLOP/s/device: 82.449, Tokens/s/device: 12427.637, total_weights: 65536, loss: 6.039, lm_loss: 6.039, perplexity: 419.520
I0422 15:24:48.775059 132726446835520 checkpointing.py:794] Waiting for step 15 to finish before checkpoint...
I0422 15:24:49.063617 132726446835520 checkpointing.py:798] Waited 0.2885305881500244 seconds for step 15 to finish before starting checkpointing.
I0422 15:24:49.066347 132726446835520 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 15:24:52.833682 132604271511296 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6
I0422 15:24:58.361452 132604279904000 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 147.779 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.688945770263672 s) (per-host)
I0422 15:24:58.361597 132604279904000 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.393502s.
I0422 15:25:05.357151 132604279904000 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 17.389038s.
I0422 15:25:05.357439 132604247799552 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 15:25:05.357570 132604247799552 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 15:25:05.357620 132604247799552 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0422 15:25:05.359274 132604247799552 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0422 15:25:05.359450 132726446835520 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 15:25:05.359598 132726446835520 checkpoint_manager.py:1452] Waiting for previous save to complete took 16.293250 seconds. If this number is high, consider checkpointing less frequently.
I0422 15:25:05.361710 132726446835520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0422 15:25:05.364201 132726446835520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/15.
I0422 15:25:05.807505 132726446835520 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 15:25:05.807611 132726446835520 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 15:25:05.834625 132726446835520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028722s
I0422 15:25:05.834761 132726446835520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 42.627 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03618764877319336 s) (per-host)
I0422 15:25:05.834811 132726446835520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.036245s (batch_requests_ready=0.003233s, total_serialization_initiated=0.032949s, others=0.000063s)
I0422 15:25:05.834895 132726446835520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041164s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.041152)
I0422 15:25:05.835597 132726446835520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.47 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/15.
I0422 15:25:05.835882 132604247799552 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:45:05.835849
I0422 15:25:05.855825 132726446835520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 15:25:05.856065 132600496629504 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 15:25:05.856212 132726446835520 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776871489.0663183, 'wait_for_prev_duration_secs': 16.29325008392334, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776871505.3617485, 'checkpointer_blocking_duration_secs': 0.4742715358734131, 'get_old_steps_start_time': 1776871505.8360436, 'get_old_steps_duration_secs': 3.3855438232421875e-05, 'checkpoint_manager_blocking_start_time': 1776871489.0639803, 'checkpoint_manager_blocking_duration_secs': 16.792198419570923}
I0422 15:25:05.856326 132726446835520 checkpointing.py:409] Started an asynchronous checkpoint save for step 15
I0422 15:25:05.857440 132726446835520 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.158, Tokens/s/device: 13891.056, total_weights: 65536, loss: 5.858, lm_loss: 5.858, perplexity: 350.067
I0422 15:25:05.864723 132726446835520 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.884, Tokens/s/device: 13849.816, total_weights: 65536, loss: 5.713, lm_loss: 5.713, perplexity: 302.780
I0422 15:25:11.671897 132604271511296 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6
I0422 15:25:16.858770 132604247799552 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 142.818 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.060174703598022 s) (per-host)
I0422 15:25:16.858881 132604247799552 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.022889s.
I0422 15:25:22.901673 132604247799552 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 17.065665s.
I0422 15:25:22.901911 132600496629504 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 15:25:22.901967 132600496629504 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 15:25:22.902014 132600496629504 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0422 15:25:22.903711 132600496629504 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0422 15:25:28.882081 132726446835520 metric_logger.py:196] completed step: 16, seconds: 17.088, TFLOP/s/device: 0.795, Tokens/s/device: 119.851, total_weights: 65536, loss: 5.599, lm_loss: 5.599, perplexity: 270.201
I0422 15:25:29.029625 132726446835520 metric_logger.py:196] completed step: 17, seconds: 22.872, TFLOP/s/device: 0.594, Tokens/s/device: 89.542, total_weights: 65536, loss: 5.511, lm_loss: 5.511, perplexity: 247.443
I0422 15:25:29.177359 132726446835520 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.677, Tokens/s/device: 13366.401, total_weights: 65536, loss: 5.442, lm_loss: 5.442, perplexity: 230.909
I0422 15:25:29.325868 132726446835520 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0422 15:25:29.326549 132726446835520 checkpointing.py:798] Waited 0.0006933212280273438 seconds for step 19 to finish before starting checkpointing.
I0422 15:25:29.328658 132726446835520 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 15:25:29.330292 132726446835520 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0422 15:25:29.331588 132726446835520 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/19.
I0422 15:25:29.654864 132726446835520 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 15:25:29.655012 132726446835520 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 15:25:29.681438 132726446835520 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028030s
I0422 15:25:29.681617 132726446835520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 43.553 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03541851043701172 s) (per-host)
I0422 15:25:29.681671 132726446835520 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.035483s (batch_requests_ready=0.002903s, total_serialization_initiated=0.032511s, others=0.000070s)
I0422 15:25:29.681774 132726446835520 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.039757s (all_items=0.000012s, per_item={'items': '0.00001216'}, temp_paths=0.039744)
I0422 15:25:29.682504 132726446835520 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/19.
I0422 15:25:29.682762 132600496629504 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:45:29.682735
I0422 15:25:29.685318 132726446835520 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 15:25:29.685588 132599464851200 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 15:25:29.685750 132726446835520 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776871529.328628, 'wait_for_prev_duration_secs': 7.224082946777344e-05, 'time_between_consecutive_saves_sec': 6.424877166748047, 'checkpointer_blocking_start_time': 1776871529.330333, 'checkpointer_blocking_duration_secs': 0.3525524139404297, 'get_old_steps_start_time': 1776871529.6829062, 'get_old_steps_duration_secs': 3.3855438232421875e-05, 'checkpoint_manager_blocking_start_time': 1776871529.3267884, 'checkpoint_manager_blocking_duration_secs': 0.35892748832702637}
I0422 15:25:29.685860 132726446835520 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0422 15:25:29.685906 132726446835520 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 15:25:35.028360 132604271511296 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6
I0422 15:25:39.990853 132600496629504 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 152.697 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.344619035720825 s) (per-host)
I0422 15:25:39.990962 132600496629504 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.308109s.
I0422 15:25:45.346246 132600496629504 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 15.663375s.
I0422 15:25:45.346567 132599464851200 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 15:25:45.346683 132599464851200 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 15:25:45.346725 132599464851200 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0422 15:25:45.348443 132599464851200 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0422 15:25:45.348619 132726446835520 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 15:25:45.348782 132726446835520 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 15:25:45.349810 132726446835520 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.910, Tokens/s/device: 13853.751, total_weights: 65536, loss: 5.385, lm_loss: 5.385, perplexity: 218.045
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Wed Apr 22 15:25:58 UTC 2026
EXIT_CODE=0