XPK Start: Sun Apr 19 02:25:14 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 02:25:38.530446: 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 02:25:38.710796 134427774138176 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 02:25:47.752480 134427774138176 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 02:25:47.757038 134427774138176 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-remfl-slice-job-0-0.mt-15-ocdbt-true-rs-remfl:8482 I0419 02:25:49.032471 134427774138176 max_utils.py:284] Jax distributed system initialized! I0419 02:25:55.073554 134427774138176 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 02:25:55.073660 134427774138176 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 02:25:55.073701 134427774138176 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 02:25:55.073736 134427774138176 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 02:25:55.781445 134427774138176 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 02:25:55.896194 134427774138176 checkpointing.py:688] Setting up checkpoint logger... I0419 02:25:55.896327 134427774138176 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 02:25:55.896375 134427774138176 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 02:25:55.896572 134427774138176 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a424c31d130>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 02:25:58.428809 134427774138176 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 02:25:58.429093 134427774138176 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 0x7a2d3441d250>}, handler_registry=None I0419 02:25:58.429406 134427774138176 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a2d3441d250>` 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 02:25:58.429470 134427774138176 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2d344364e0>` 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 02:25:58.429517 134427774138176 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 0x7a2d3441d250>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a2d3441d250>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2d344364e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2d344364e0>}). I0419 02:25:58.429924 134427774138176 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 02:25:58.430015 134427774138176 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a2e20108fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 02:26:00.252418 134306840176384 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776565291627892306, 'commit_timestamp_nsecs': 1776565333801437029, '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_true/checkpoints/0/_CHECKPOINT_METADATA I0419 02:26:00.916172 134306823390976 checkpoint_manager.py:1871] Missing metrics for step 9 I0419 02:26:01.358527 134427774138176 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_true/checkpoints I0419 02:26:02.640932 134427774138176 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_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a4124b97f80> I0419 02:26:02.641108 134427774138176 checkpointing.py:302] Checkpoint manager created! I0419 02:26:03.617347 134427774138176 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 02:26:03.617468 134427774138176 checkpointing.py:582] restoring from this run's directory step 9 I0419 02:26:03.865051 134427774138176 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_true/checkpoints/9. I0419 02:26:04.111482 134427774138176 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_true/checkpoints/9. I0419 02:26:04.111916 134427774138176 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_true/checkpoints/9. I0419 02:26:06.817774 1444 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 02:26:15.349140 134427774138176 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 185.081 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 8.53831148147583 s) (per-host) I0419 02:26:15.357901 134427774138176 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 161.312 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 9.810811042785645 s) (per-host) I0419 02:26:15.771861 134427774138176 event_tracking.py:138] [process=6] [sync] Finished load in 11.91 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_true/checkpoints/9 I0419 02:26:15.772011 134427774138176 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_true/checkpoints', 'checkpointer_start_time': 1776565563.8649502, 'checkpointer_duration_secs': 11.907029628753662, 'checkpoint_manager_start_time': 1776565563.6180646, 'checkpoint_manager_duration_secs': 12.153916835784912} fsdp: 32 I0419 02:26:15.847696 134427774138176 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 02:26:15.847805 134427774138176 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 02:26:15.847856 134427774138176 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 02:26:15.847898 134427774138176 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 02:26:15.847935 134427774138176 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 02:26:15.847969 134427774138176 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 02:26:15.848003 134427774138176 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:26:15.848037 134427774138176 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 02:26:15.848072 134427774138176 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:26:15.848105 134427774138176 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 02:26:15.848138 134427774138176 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 02:26:15.848171 134427774138176 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 02:26:16.247886 134427774138176 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 02:26:16.247976 134427774138176 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:26:16.253445 134427774138176 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 02:26:16.253507 134427774138176 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:26:16.269785 134427774138176 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 02:26:16.269845 134427774138176 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:26:16.285408 134427774138176 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 02:26:16.285480 134427774138176 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 02:26:16.309485 134427774138176 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 02:26:16.309552 134427774138176 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:26:16.325485 134427774138176 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 02:26:16.325544 134427774138176 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:26:16.341271 134427774138176 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 02:26:16.341330 134427774138176 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:26:16.378921 134427774138176 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 02:26:16.379026 134427774138176 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 02:26:16.399582 134427774138176 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 02:26:16.399645 134427774138176 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 02:26:27.989908 134427774138176 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 02:26:27.992270 134427774138176 metric_logger.py:301] number parameters: 1.104 billion I0419 02:26:39.236604 134427774138176 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 02:26:39.387968 134427774138176 checkpointing.py:798] Waited 0.1513524055480957 seconds for step 10 to finish before starting checkpointing. I0419 02:26:39.390594 134427774138176 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 02:26:39.392516 134427774138176 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0419 02:26:39.394174 134427774138176 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_true/checkpoints/10. I0419 02:26:40.155474 134427774138176 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 02:26:40.156420 134427774138176 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 02:26:40.156476 134427774138176 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 02:26:40.435300 134427774138176 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.280602s I0419 02:26:40.435469 134427774138176 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.356 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2879958152770996 s) (per-host) I0419 02:26:40.435521 134427774138176 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.288059s (batch_requests_ready=0.003113s, total_serialization_initiated=0.284874s, others=0.000072s) I0419 02:26:40.435624 134427774138176 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.292039s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.292022) I0419 02:26:40.436394 134427774138176 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.04 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_true/checkpoints/10. I0419 02:26:40.436708 134305732863744 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:46:40.436674 I0419 02:26:40.441876 134427774138176 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 02:26:40.442182 134303041582848 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 02:26:40.442351 134427774138176 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776565599.3905754, 'wait_for_prev_duration_secs': 6.604194641113281e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776565599.3925545, 'checkpointer_blocking_duration_secs': 1.044311761856079, 'get_old_steps_start_time': 1776565600.436893, 'get_old_steps_duration_secs': 3.409385681152344e-05, 'checkpoint_manager_blocking_start_time': 1776565599.3883445, 'checkpoint_manager_blocking_duration_secs': 1.0539703369140625} I0419 02:26:40.442457 134427774138176 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 02:26:40.442524 134427774138176 max_utils.py:750] Memstats: After params initialized: I0419 02:26:40.442573 134427774138176 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 02:26:40.442610 134427774138176 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 02:26:40.442635 134427774138176 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 02:26:40.442657 134427774138176 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 02:26:40.763123 134427774138176 metric_logger.py:196] completed step: 10, seconds: 11.243, TFLOP/s/device: 1.208, Tokens/s/device: 182.153, total_weights: 65536, loss: 6.827, lm_loss: 6.827, perplexity: 922.537 I0419 02:26:40.940358 134427774138176 metric_logger.py:196] completed step: 11, seconds: 1.525, TFLOP/s/device: 8.912, Tokens/s/device: 1343.301, total_weights: 65536, loss: 6.522, lm_loss: 6.522, perplexity: 679.821 I0419 02:26:41.088224 134427774138176 metric_logger.py:196] completed step: 12, seconds: 0.020, TFLOP/s/device: 670.208, Tokens/s/device: 101021.062, total_weights: 65536, loss: 6.260, lm_loss: 6.260, perplexity: 523.013 I0419 02:26:41.235991 134427774138176 metric_logger.py:196] completed step: 13, seconds: 0.165, TFLOP/s/device: 82.296, Tokens/s/device: 12404.528, total_weights: 65536, loss: 6.039, lm_loss: 6.039, perplexity: 419.520 I0419 02:26:41.243149 134427774138176 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 02:26:41.530930 134427774138176 checkpointing.py:798] Waited 0.28775477409362793 seconds for step 15 to finish before starting checkpointing. I0419 02:26:41.533703 134427774138176 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 02:26:47.931015 134303065294592 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_true/checkpoints/10/items/array_metadatas/process_6 I0419 02:27:22.157202 134305732863744 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.601 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.00969696044922 s) (per-host) I0419 02:27:22.157329 134305732863744 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.720508s. I0419 02:27:31.662408 134305732863744 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.225567s. I0419 02:27:31.662775 134303041582848 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 02:27:31.662900 134303041582848 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 02:27:31.662952 134303041582848 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 02:27:31.664825 134303041582848 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 02:27:31.664990 134427774138176 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 02:27:31.665124 134427774138176 checkpoint_manager.py:1452] Waiting for previous save to complete took 50.131423 seconds. If this number is high, consider checkpointing less frequently. I0419 02:27:31.667382 134427774138176 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0419 02:27:31.669813 134427774138176 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_true/checkpoints/15. I0419 02:27:32.810110 134427774138176 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 02:27:32.810204 134427774138176 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 02:27:32.837307 134427774138176 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028725s I0419 02:27:32.837474 134427774138176 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 43.157 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.035742998123168945 s) (per-host) I0419 02:27:32.837524 134427774138176 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.035803s (batch_requests_ready=0.002780s, total_serialization_initiated=0.032956s, others=0.000066s) I0419 02:27:32.837621 134427774138176 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.040749s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.040736) I0419 02:27:32.838310 134427774138176 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.17 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_true/checkpoints/15. I0419 02:27:32.838642 134303065294592 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:47:32.838606 I0419 02:27:32.860716 134427774138176 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 02:27:32.860998 134303041582848 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 02:27:32.861148 134427774138176 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776565601.5336735, 'wait_for_prev_duration_secs': 50.13142275810242, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776565651.6674225, 'checkpointer_blocking_duration_secs': 1.1713626384735107, 'get_old_steps_start_time': 1776565652.8388085, 'get_old_steps_duration_secs': 3.1948089599609375e-05, 'checkpoint_manager_blocking_start_time': 1776565601.5313082, 'checkpoint_manager_blocking_duration_secs': 51.32980465888977} I0419 02:27:32.861274 134427774138176 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 02:27:32.862376 134427774138176 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.179, Tokens/s/device: 13894.166, total_weights: 65536, loss: 5.858, lm_loss: 5.858, perplexity: 350.067 I0419 02:27:32.870426 134427774138176 metric_logger.py:196] completed step: 15, seconds: 0.149, TFLOP/s/device: 91.260, Tokens/s/device: 13755.676, total_weights: 65536, loss: 5.713, lm_loss: 5.713, perplexity: 302.780 I0419 02:27:39.715932 134300917810944 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_true/checkpoints/15/items/array_metadatas/process_6 I0419 02:27:54.591838 134427774138176 metric_logger.py:196] completed step: 16, seconds: 51.625, TFLOP/s/device: 0.263, Tokens/s/device: 39.670, total_weights: 65536, loss: 5.599, lm_loss: 5.599, perplexity: 270.201 I0419 02:27:54.739387 134427774138176 metric_logger.py:196] completed step: 17, seconds: 21.576, TFLOP/s/device: 0.630, Tokens/s/device: 94.919, total_weights: 65536, loss: 5.511, lm_loss: 5.511, perplexity: 247.443 I0419 02:27:54.886927 134427774138176 metric_logger.py:196] completed step: 18, seconds: 0.153, TFLOP/s/device: 88.760, Tokens/s/device: 13378.800, total_weights: 65536, loss: 5.442, lm_loss: 5.442, perplexity: 230.909 I0419 02:27:55.035462 134427774138176 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 02:27:55.036142 134427774138176 checkpointing.py:798] Waited 0.0006935596466064453 seconds for step 19 to finish before starting checkpointing. I0419 02:27:55.038089 134427774138176 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 02:28:15.497919 134303065294592 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.996 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.69614815711975 s) (per-host) I0419 02:28:15.498046 134303065294592 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.659290s. I0419 02:28:23.995447 134303065294592 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.156675s. I0419 02:28:23.995744 134303041582848 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 02:28:23.995862 134303041582848 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 02:28:23.995911 134303041582848 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 02:28:23.997464 134303041582848 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 02:28:23.997650 134427774138176 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 02:28:23.997774 134427774138176 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.959689 seconds. If this number is high, consider checkpointing less frequently. I0419 02:28:24.001278 134427774138176 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0419 02:28:24.003283 134427774138176 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_true/checkpoints/19. I0419 02:28:24.724098 134427774138176 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0419 02:28:24.724193 134427774138176 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 02:28:24.750244 134427774138176 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027632s I0419 02:28:24.750417 134427774138176 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 44.376 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03476119041442871 s) (per-host) I0419 02:28:24.750469 134427774138176 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.034823s (batch_requests_ready=0.002887s, total_serialization_initiated=0.031868s, others=0.000068s) I0419 02:28:24.750571 134427774138176 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.038864s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.038853) I0419 02:28:24.751286 134427774138176 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_true/checkpoints/19. I0419 02:28:24.751606 134303041582848 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:48:24.751570 I0419 02:28:24.755087 134427774138176 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 02:28:24.755345 134299315590912 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 02:28:24.755484 134427774138176 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_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776565675.0380597, 'wait_for_prev_duration_secs': 28.959688663482666, 'time_between_consecutive_saves_sec': 23.373191833496094, 'checkpointer_blocking_start_time': 1776565704.0013177, 'checkpointer_blocking_duration_secs': 0.7504386901855469, 'get_old_steps_start_time': 1776565704.7517753, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776565675.036396, 'checkpoint_manager_blocking_duration_secs': 29.719053983688354} I0419 02:28:24.755604 134427774138176 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 02:28:24.755649 134427774138176 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 02:28:32.344297 134300917810944 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_true/checkpoints/19/items/array_metadatas/process_6 I0419 02:29:08.003315 134303041582848 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.491 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.28762102127075 s) (per-host) I0419 02:29:08.003423 134303041582848 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.251701s. I0419 02:29:16.667190 134303041582848 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.915447s. I0419 02:29:16.667536 134299315590912 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 02:29:16.667673 134299315590912 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 02:29:16.667721 134299315590912 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 02:29:16.669628 134299315590912 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 02:29:16.669755 134427774138176 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 02:29:16.669865 134427774138176 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 02:29:16.670837 134427774138176 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.845, Tokens/s/device: 13843.824, 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 02:29:29 UTC 2026 EXIT_CODE=0