XPK Start: Sat Apr 18 19:23:06 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 19:23:30.639634: 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) I0418 19:23:30.823367 134933364119360 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 19:23:39.864414 134933364119360 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 19:23:39.866739 134933364119360 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-sche3-slice-job-0-0.mt-15-ocdbt-true-sv-sche3:8482 I0418 19:23:41.187144 134933364119360 max_utils.py:284] Jax distributed system initialized! I0418 19:23:47.207521 134933364119360 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 19:23:47.207620 134933364119360 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 19:23:47.207660 134933364119360 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 I0418 19:23:47.207703 134933364119360 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 19:23:47.901968 134933364119360 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 19:23:47.902239 134933364119360 checkpointing.py:677] Setting up checkpoint logger... I0418 19:23:47.902292 134933364119360 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 19:23:47.902336 134933364119360 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 19:23:47.902680 134933364119360 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 0x7ab805b1c770>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 19:23:50.779143 134933364119360 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 19:23:50.779376 134933364119360 checkpoint_manager.py:708] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab6db341640>}, handler_registry=None I0418 19:23:50.779617 134933364119360 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab6db341640>` 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`. I0418 19:23:50.779666 134933364119360 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa3ac13d8b0>` 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`. I0418 19:23:50.779701 134933364119360 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 0x7ab6db341640>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ab6db341640>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa3ac13d8b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7aa3ac13d8b0>}). I0418 19:23:50.780017 134933364119360 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 19:23:50.780089 134933364119360 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7aa2d85d7d80> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 19:23:52.954433 134933364119360 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints I0418 19:23:52.968473 134933364119360 checkpoint_manager.py:929] [process=5][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ab853dd3b00> I0418 19:23:52.968593 134933364119360 checkpointing.py:301] Checkpoint manager created! I0418 19:23:53.896764 134933364119360 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 19:23:53.896876 134933364119360 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 19:23:54.283757 134933364119360 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 19:23:54.283848 134933364119360 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 19:23:54.300388 134933364119360 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 19:23:54.300448 134933364119360 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 19:23:54.324420 134933364119360 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 19:23:54.324504 134933364119360 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 19:23:54.341092 134933364119360 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 19:23:54.341150 134933364119360 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 19:23:54.357709 134933364119360 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 19:23:54.357770 134933364119360 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 19:23:54.382603 134933364119360 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 19:23:54.382675 134933364119360 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 19:23:54.404820 134933364119360 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 19:23:54.404887 134933364119360 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 19:23:54.624800 134933364119360 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 19:23:54.624905 134933364119360 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 19:23:56.064494 134933364119360 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 19:23:56.064620 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 19:23:56.064676 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 19:23:56.064733 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0418 19:23:56.064786 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 19:23:56.064824 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 19:23:56.064876 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 19:23:56.064925 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0418 19:23:56.064965 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0418 19:23:56.065000 134933364119360 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 19:23:56.065045 134933364119360 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 19:23:56.065092 134933364119360 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 19:23:56.555333 134933364119360 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 19:23:56.555429 134933364119360 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 19:23:56.570938 134933364119360 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 19:23:56.570997 134933364119360 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 19:24:07.478537 134933364119360 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. I0418 19:24:07.479367 134933364119360 metric_logger.py:301] number parameters: 1.104 billion I0418 19:24:18.751158 134933364119360 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 19:24:18.905253 134933364119360 checkpointing.py:776] Waited 0.15407061576843262 seconds for step 0 to finish before starting checkpointing. I0418 19:24:18.907759 134933364119360 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 19:24:18.909306 134933364119360 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0 I0418 19:24:18.910834 134933364119360 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/0. I0418 19:24:19.649217 134933364119360 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 19:24:19.650246 134933364119360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 19:24:19.650303 134933364119360 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 I0418 19:24:19.925670 134933364119360 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276525s I0418 19:24:19.925849 134933364119360 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.471 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28196263313293457 s) (per-host) I0418 19:24:19.925904 134933364119360 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282029s (batch_requests_ready=0.002223s, total_serialization_initiated=0.279731s, others=0.000075s) I0418 19:24:19.925995 134933364119360 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286178s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.286160) I0418 19:24:19.926762 134933364119360 event_tracking.py:125] [process=5] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/0. I0418 19:24:19.927036 134805006583552 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 19:44:19.927009 I0418 19:24:19.932327 134933364119360 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 19:24:19.932640 134802908899072 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 19:24:19.932803 134933364119360 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776540258.9077408, 'wait_for_prev_duration_secs': 5.888938903808594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776540258.9093437, 'checkpointer_blocking_duration_secs': 1.0178053379058838, 'get_old_steps_start_time': 1776540259.9271746, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776540258.9058003, 'checkpoint_manager_blocking_duration_secs': 1.026960849761963} I0418 19:24:19.932911 134933364119360 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 19:24:19.932966 134933364119360 max_utils.py:750] Memstats: After params initialized: I0418 19:24:19.933016 134933364119360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0418 19:24:19.933049 134933364119360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0418 19:24:19.933077 134933364119360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0418 19:24:19.933102 134933364119360 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0418 19:24:20.246558 134933364119360 metric_logger.py:196] completed step: 0, seconds: 11.272, TFLOP/s/device: 1.205, Tokens/s/device: 181.694, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 19:24:20.426385 134933364119360 metric_logger.py:196] completed step: 1, seconds: 1.494, TFLOP/s/device: 9.095, Tokens/s/device: 1370.918, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 19:24:20.832238 134933364119360 metric_logger.py:196] completed step: 2, seconds: 0.033, TFLOP/s/device: 416.476, Tokens/s/device: 62775.871, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0418 19:24:20.979521 134933364119360 metric_logger.py:196] completed step: 3, seconds: 0.406, TFLOP/s/device: 33.467, Tokens/s/device: 5044.447, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0418 19:24:20.984130 134933364119360 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0418 19:24:21.273825 134933364119360 checkpointing.py:776] Waited 0.289658784866333 seconds for step 5 to finish before starting checkpointing. I0418 19:24:21.277671 134933364119360 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 19:24:23.661223 2843 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 19:24:26.604809 134802917291776 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_5 I0418 19:24:57.214982 134805006583552 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 42.043 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.57106590270996 s) (per-host) I0418 19:24:57.215091 134805006583552 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.287974s. I0418 19:25:06.635062 134805006583552 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 46.707928s. I0418 19:25:06.635329 134802908899072 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 19:25:06.635485 134802908899072 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 19:25:06.635551 134802908899072 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 19:25:06.637347 134802908899072 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 19:25:06.637534 134933364119360 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 19:25:06.637664 134933364119360 checkpoint_manager.py:1452] Waiting for previous save to complete took 45.359991 seconds. If this number is high, consider checkpointing less frequently. I0418 19:25:06.639805 134933364119360 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 5 I0418 19:25:06.642309 134933364119360 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/5. I0418 19:25:07.353446 134933364119360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 19:25:07.353552 134933364119360 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 I0418 19:25:07.387119 134933364119360 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035031s I0418 19:25:07.387284 134933364119360 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 40.162 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03840899467468262 s) (per-host) I0418 19:25:07.387337 134933364119360 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038471s (batch_requests_ready=0.001765s, total_serialization_initiated=0.036638s, others=0.000069s) I0418 19:25:07.387427 134933364119360 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043586s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.043571) I0418 19:25:07.388031 134933364119360 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/5. I0418 19:25:07.388357 134779991205632 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 19:45:07.388320 I0418 19:25:07.397394 134933364119360 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0418 19:25:07.397693 134802908899072 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 19:25:07.397853 134933364119360 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776540261.2776434, 'wait_for_prev_duration_secs': 45.35999131202698, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776540306.639845, 'checkpointer_blocking_duration_secs': 0.7486262321472168, 'get_old_steps_start_time': 1776540307.3884933, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776540261.274504, 'checkpoint_manager_blocking_duration_secs': 46.12331748008728} I0418 19:25:07.397960 134933364119360 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0418 19:25:07.398689 134933364119360 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.765, Tokens/s/device: 13379.674, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0418 19:25:07.404643 134933364119360 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.241, Tokens/s/device: 13903.598, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0418 19:25:15.597193 134802917291776 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_5 I0418 19:25:31.276177 134933364119360 metric_logger.py:196] completed step: 6, seconds: 46.419, TFLOP/s/device: 0.293, Tokens/s/device: 44.120, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0418 19:25:31.423468 134933364119360 metric_logger.py:196] completed step: 7, seconds: 23.726, TFLOP/s/device: 0.573, Tokens/s/device: 86.321, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0418 19:25:31.571054 134933364119360 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.226, Tokens/s/device: 13449.175, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0418 19:25:31.717764 134933364119360 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 19:25:31.718441 134933364119360 checkpointing.py:776] Waited 0.0006983280181884766 seconds for step 9 to finish before starting checkpointing. I0418 19:25:31.720485 134933364119360 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 19:25:50.782321 134779991205632 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 36.368 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.433417320251465 s) (per-host) I0418 19:25:50.782428 134779991205632 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.393535s. I0418 19:25:58.847206 134779991205632 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 51.458296s. I0418 19:25:58.847511 134802908899072 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 19:25:58.847633 134802908899072 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 19:25:58.847686 134802908899072 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0418 19:25:58.849294 134802908899072 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0418 19:25:58.849491 134933364119360 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0418 19:25:58.849631 134933364119360 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.129163 seconds. If this number is high, consider checkpointing less frequently. I0418 19:25:58.851060 134933364119360 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9 I0418 19:25:58.853163 134933364119360 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/9. I0418 19:25:59.166241 134933364119360 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 19:25:59.166333 134933364119360 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 I0418 19:25:59.200920 134933364119360 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035667s I0418 19:25:59.201094 134933364119360 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 39.477 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03907489776611328 s) (per-host) I0418 19:25:59.201145 134933364119360 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039136s (batch_requests_ready=0.001625s, total_serialization_initiated=0.037445s, others=0.000066s) I0418 19:25:59.201235 134933364119360 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043216s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=0.043205) I0418 19:25:59.201933 134933364119360 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/9. I0418 19:25:59.202247 134805006583552 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 19:45:59.202206 I0418 19:25:59.209620 134933364119360 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 19:25:59.209883 134802908899072 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 19:25:59.210036 134933364119360 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776540331.720442, 'wait_for_prev_duration_secs': 27.129162788391113, 'time_between_consecutive_saves_sec': 25.08305549621582, 'checkpointer_blocking_start_time': 1776540358.851101, 'checkpointer_blocking_duration_secs': 0.35129857063293457, 'get_old_steps_start_time': 1776540359.202415, 'get_old_steps_duration_secs': 2.3603439331054688e-05, 'checkpoint_manager_blocking_start_time': 1776540331.7186885, 'checkpoint_manager_blocking_duration_secs': 27.491317749023438} I0418 19:25:59.210142 134933364119360 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 19:25:59.210185 134933364119360 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 19:26:04.772058 134809317779200 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_180002/linen_xpk_main_20260418_180002_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_5 I0418 19:26:41.491776 134805006583552 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.316 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.32971978187561 s) (per-host) I0418 19:26:41.491901 134805006583552 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.289536s. I0418 19:26:50.073544 134805006583552 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 50.871163s. I0418 19:26:50.073831 134802908899072 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 19:26:50.073962 134802908899072 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 19:26:50.074003 134802908899072 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 19:26:50.075994 134802908899072 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 19:26:50.076162 134933364119360 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 19:26:50.076310 134933364119360 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 19:26:50.077215 134933364119360 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.273, Tokens/s/device: 13908.414, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 19:27:01 UTC 2026 EXIT_CODE=0