MaxView

XPK Start: Sun Apr 19 02:34:13 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-19 02:34:37.257990: 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)
I0419 02:34:37.438394 131932836054848 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 02:34:46.479668 131932836054848 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 02:34:46.482006 131932836054848 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-75it6-slice-job-0-0.mt-15-ocdbt-false-rs-75it6:8482
I0419 02:34:48.298473 131932836054848 max_utils.py:284] Jax distributed system initialized!
I0419 02:34:54.389920 131932836054848 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 02:34:54.390027 131932836054848 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 02:34:54.390068 131932836054848 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
I0419 02:34:54.390103 131932836054848 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 02:34:55.085556 131932836054848 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 02:34:55.199432 131932836054848 checkpointing.py:688] Setting up checkpoint logger...
I0419 02:34:55.199556 131932836054848 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True
I0419 02:34:55.199601 131932836054848 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 02:34:55.199795 131932836054848 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 0x77fd63f17890>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 02:34:56.887975 131932836054848 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0419 02:34:56.888211 131932836054848 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 0x77e9085a1df0>}, handler_registry=None
I0419 02:34:56.888473 131932836054848 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77e9085a1df0>` 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`.
I0419 02:34:56.888531 131932836054848 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77e9085e6c30>` 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`.
I0419 02:34:56.888567 131932836054848 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 0x77e9085a1df0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77e9085a1df0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77e9085e6c30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77e9085e6c30>}).
I0419 02:34:56.888881 131932836054848 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 02:34:56.888950 131932836054848 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x77e9087d4fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 02:34:59.932619 131811069916928 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776565909842680444, 'commit_timestamp_nsecs': 1776565927299405425, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0419 02:35:00.180181 131811069916928 checkpoint_manager.py:1871] Missing metrics for step 0
I0419 02:35:00.626674 131932836054848 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints
I0419 02:35:01.131873 131932836054848 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x77fdb1fcf080>
I0419 02:35:01.132054 131932836054848 checkpointing.py:302] Checkpoint manager created!
I0419 02:35:02.093255 131932836054848 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 02:35:02.093379 131932836054848 checkpointing.py:582] restoring from this run's directory step 9
I0419 02:35:02.355539 131932836054848 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/9.
I0419 02:35:02.601702 131932836054848 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/9.
I0419 02:35:02.602123 131932836054848 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/9.
I0419 02:35:06.835073    1487 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 02:35:16.849652 131932836054848 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 133.486 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.838481187820435 s) (per-host)
I0419 02:35:16.859601 131932836054848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 124.442 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 12.717663764953613 s) (per-host)
I0419 02:35:17.065002 131932836054848 event_tracking.py:138] [process=6] [sync] Finished load in 14.71 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/9
I0419 02:35:17.065155 131932836054848 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776566102.3554401, 'checkpointer_duration_secs': 14.709685325622559, 'checkpoint_manager_start_time': 1776566102.0939784, 'checkpoint_manager_duration_secs': 14.971149444580078}
fsdp: 32
I0419 02:35:17.141143 131932836054848 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0419 02:35:17.141279 131932836054848 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 02:35:17.141339 131932836054848 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 02:35:17.141382 131932836054848 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0419 02:35:17.141419 131932836054848 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 02:35:17.141452 131932836054848 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 02:35:17.141487 131932836054848 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:35:17.141523 131932836054848 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0419 02:35:17.141557 131932836054848 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:35:17.141590 131932836054848 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:35:17.141622 131932836054848 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)

I0419 02:35:17.141653 131932836054848 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0419 02:35:17.541127 131932836054848 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 02:35:17.541221 131932836054848 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:35:17.546811 131932836054848 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 02:35:17.546867 131932836054848 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:35:17.563214 131932836054848 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 02:35:17.563294 131932836054848 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:35:17.578950 131932836054848 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 02:35:17.579007 131932836054848 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:35:17.602982 131932836054848 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:35:17.603052 131932836054848 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:35:17.619054 131932836054848 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:35:17.619117 131932836054848 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:35:17.634814 131932836054848 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:35:17.634873 131932836054848 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:35:17.665031 131932836054848 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 02:35:17.665101 131932836054848 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:35:17.685009 131932836054848 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 02:35:17.685075 131932836054848 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 02:35:28.919520 131932836054848 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.
I0419 02:35:28.921918 131932836054848 metric_logger.py:301] number parameters: 1.104 billion
I0419 02:35:40.233769 131932836054848 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0419 02:35:40.386693 131932836054848 checkpointing.py:798] Waited 0.1529088020324707 seconds for step 10 to finish before starting checkpointing.
I0419 02:35:40.389297 131932836054848 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 02:35:40.391041 131932836054848 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0419 02:35:40.392508 131932836054848 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/10.
I0419 02:35:40.781530 131932836054848 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 02:35:40.782410 131932836054848 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0419 02:35:40.782468 131932836054848 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
I0419 02:35:41.066271 131932836054848 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.285485s
I0419 02:35:41.066438 131932836054848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.273 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29251575469970703 s) (per-host)
I0419 02:35:41.066488 131932836054848 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292577s (batch_requests_ready=0.002753s, total_serialization_initiated=0.289754s, others=0.000070s)
I0419 02:35:41.066597 131932836054848 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.296806s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.296789)
I0419 02:35:41.067286 131932836054848 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.68 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/10.
I0419 02:35:41.067644 131808310060800 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:55:41.067604
I0419 02:35:41.069634 131932836054848 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 02:35:41.069920 131806129493760 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 02:35:41.070073 131932836054848 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776566140.3892784, 'wait_for_prev_duration_secs': 5.91278076171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776566140.3910806, 'checkpointer_blocking_duration_secs': 0.6767239570617676, 'get_old_steps_start_time': 1776566141.067827, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776566140.3870564, 'checkpoint_manager_blocking_duration_secs': 0.6829826831817627}
I0419 02:35:41.070178 131932836054848 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0419 02:35:41.070255 131932836054848 max_utils.py:750] 
Memstats: After params initialized:
I0419 02:35:41.070305 131932836054848 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0419 02:35:41.070339 131932836054848 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0419 02:35:41.070366 131932836054848 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0419 02:35:41.070389 131932836054848 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0419 02:35:41.400060 131932836054848 metric_logger.py:196] completed step: 10, seconds: 11.311, TFLOP/s/device: 1.201, Tokens/s/device: 181.067, total_weights: 65536, loss: 6.827, lm_loss: 6.827, perplexity: 922.537
I0419 02:35:41.582787 131932836054848 metric_logger.py:196] completed step: 11, seconds: 1.164, TFLOP/s/device: 11.668, Tokens/s/device: 1758.784, total_weights: 65536, loss: 6.522, lm_loss: 6.522, perplexity: 679.821
I0419 02:35:41.730740 131932836054848 metric_logger.py:196] completed step: 12, seconds: 0.026, TFLOP/s/device: 527.021, Tokens/s/device: 79438.346, total_weights: 65536, loss: 6.260, lm_loss: 6.260, perplexity: 523.013
I0419 02:35:41.878446 131932836054848 metric_logger.py:196] completed step: 13, seconds: 0.165, TFLOP/s/device: 82.259, Tokens/s/device: 12399.046, total_weights: 65536, loss: 6.039, lm_loss: 6.039, perplexity: 419.520
I0419 02:35:41.884829 131932836054848 checkpointing.py:794] Waiting for step 15 to finish before checkpoint...
I0419 02:35:42.173422 131932836054848 checkpointing.py:798] Waited 0.28856444358825684 seconds for step 15 to finish before starting checkpointing.
I0419 02:35:42.176256 131932836054848 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 02:35:49.557966 131808268097280 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_6
I0419 02:35:53.869462 131808310060800 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 120.621 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.095503568649292 s) (per-host)
I0419 02:35:53.869593 131808310060800 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 12.801834s.
I0419 02:36:01.389839 131808310060800 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 20.322064s.
I0419 02:36:01.390135 131806129493760 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 02:36:01.390287 131806129493760 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 02:36:01.390336 131806129493760 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 02:36:01.392037 131806129493760 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 02:36:01.392253 131932836054848 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.
W0419 02:36:01.392393 131932836054848 checkpoint_manager.py:1452] Waiting for previous save to complete took 19.216154 seconds. If this number is high, consider checkpointing less frequently.
I0419 02:36:01.394491 131932836054848 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0419 02:36:01.397193 131932836054848 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/15.
I0419 02:36:01.715992 131932836054848 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0419 02:36:01.716087 131932836054848 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
I0419 02:36:01.759361 131932836054848 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045022s
I0419 02:36:01.759531 131932836054848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 29.478 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05233001708984375 s) (per-host)
I0419 02:36:01.759583 131932836054848 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.052391s (batch_requests_ready=0.002774s, total_serialization_initiated=0.049548s, others=0.000069s)
I0419 02:36:01.759677 131932836054848 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.057146s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.057133)
I0419 02:36:01.760427 131932836054848 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.37 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/15.
I0419 02:36:01.760772 131806129493760 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:56:01.760734
I0419 02:36:01.764682 131932836054848 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 02:36:01.764927 131806121101056 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 02:36:01.765087 131932836054848 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776566142.1762109, 'wait_for_prev_duration_secs': 19.216154098510742, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776566161.3945305, 'checkpointer_blocking_duration_secs': 0.36638832092285156, 'get_old_steps_start_time': 1776566161.7609427, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776566142.173787, 'checkpoint_manager_blocking_duration_secs': 19.591267585754395}
I0419 02:36:01.765196 131932836054848 checkpointing.py:409] Started an asynchronous checkpoint save for step 15
I0419 02:36:01.766339 131932836054848 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.167, Tokens/s/device: 13892.375, total_weights: 65536, loss: 5.858, lm_loss: 5.858, perplexity: 350.067
I0419 02:36:01.773776 131932836054848 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.808, Tokens/s/device: 13838.305, total_weights: 65536, loss: 5.713, lm_loss: 5.713, perplexity: 302.780
I0419 02:36:06.927894 131808268097280 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_6
I0419 02:36:11.998148 131806129493760 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 153.495 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.290893316268921 s) (per-host)
I0419 02:36:11.998290 131806129493760 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.237404s.
I0419 02:36:18.518283 131806129493760 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.757380s.
I0419 02:36:18.518607 131806121101056 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 02:36:18.518729 131806121101056 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 02:36:18.518777 131806121101056 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0419 02:36:18.520135 131806121101056 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0419 02:36:23.510444 131932836054848 metric_logger.py:196] completed step: 16, seconds: 19.887, TFLOP/s/device: 0.683, Tokens/s/device: 102.980, total_weights: 65536, loss: 5.599, lm_loss: 5.599, perplexity: 270.201
I0419 02:36:23.658269 131932836054848 metric_logger.py:196] completed step: 17, seconds: 21.591, TFLOP/s/device: 0.629, Tokens/s/device: 94.853, total_weights: 65536, loss: 5.511, lm_loss: 5.511, perplexity: 247.443
I0419 02:36:23.806000 131932836054848 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.614, Tokens/s/device: 13356.812, total_weights: 65536, loss: 5.442, lm_loss: 5.442, perplexity: 230.909
I0419 02:36:23.954207 131932836054848 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0419 02:36:23.954943 131932836054848 checkpointing.py:798] Waited 0.0007483959197998047 seconds for step 19 to finish before starting checkpointing.
I0419 02:36:23.957080 131932836054848 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 02:36:23.958821 131932836054848 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0419 02:36:23.960307 131932836054848 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/19.
I0419 02:36:24.667848 131932836054848 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0419 02:36:24.668166 131932836054848 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
I0419 02:36:24.712001 131932836054848 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045578s
I0419 02:36:24.712174 131932836054848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 29.324 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05260443687438965 s) (per-host)
I0419 02:36:24.712245 131932836054848 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.052678s (batch_requests_ready=0.002843s, total_serialization_initiated=0.049756s, others=0.000079s)
I0419 02:36:24.712381 131932836054848 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.056831s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.056819)
I0419 02:36:24.713078 131932836054848 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/19.
I0419 02:36:24.713413 131806121101056 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:56:24.713375
I0419 02:36:24.715313 131932836054848 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 02:36:24.715595 131802387695360 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 02:36:24.715726 131932836054848 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776566183.9570487, 'wait_for_prev_duration_secs': 8.082389831542969e-05, 'time_between_consecutive_saves_sec': 5.436872720718384, 'checkpointer_blocking_start_time': 1776566183.9588597, 'checkpointer_blocking_duration_secs': 0.7547039985656738, 'get_old_steps_start_time': 1776566184.7135818, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776566183.955197, 'checkpoint_manager_blocking_duration_secs': 0.7604992389678955}
I0419 02:36:24.715891 131932836054848 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0419 02:36:24.715936 131932836054848 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 02:36:30.512001 131808268097280 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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_6
I0419 02:36:35.669965 131806121101056 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 143.464 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.0103759765625 s) (per-host)
I0419 02:36:35.670069 131806121101056 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.956541s.
I0419 02:36:44.272817 131806121101056 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 19.559272s.
I0419 02:36:44.273125 131802387695360 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 02:36:44.273292 131802387695360 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 02:36:44.273338 131802387695360 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 02:36:44.276132 131802387695360 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 02:36:44.276305 131932836054848 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.
I0419 02:36:44.276487 131932836054848 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 02:36:44.277387 131932836054848 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.744, Tokens/s/device: 13828.588, 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: Sun Apr 19 02:36:53 UTC 2026
EXIT_CODE=0