XPK Start: Sat Apr 25 16:30:24 UTC 2026 2026-04-25 16:30:28.255891: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1777134628.268673 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777134628.272433 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777134628.284135 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134628.284152 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134628.284155 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134628.284156 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:30:47.586781: 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) I0425 16:30:48.122634 133149995796288 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:30:57,164:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:30:57.164837 133149995796288 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:30:57,167:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-rs-mqggs-slice-job-0-0.mt-16-zarr3-false-rs-mqggs:8482 I0425 16:30:57.167183 133149995796288 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-rs-mqggs-slice-job-0-0.mt-16-zarr3-false-rs-mqggs:8482 I0425 16:30:58.075513 133149995796288 max_utils.py:284] Jax distributed system initialized! I0425 16:31:05.474978 133149995796288 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:31:05.475085 133149995796288 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:31:05.475144 133149995796288 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0425 16:31:05.475180 133149995796288 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:31:06.161931 133149995796288 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:31:06.162566 133149995796288 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:31:06.162757 133149995796288 checkpointing.py:688] Setting up checkpoint logger... I0425 16:31:06.162808 133149995796288 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0425 16:31:06.162851 133149995796288 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:31:06.163213 133149995796288 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7918e1d75220>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:31:07.896912 133149995796288 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:31:07.897168 133149995796288 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x790444695bb0>}, handler_registry=None I0425 16:31:07.897417 133149995796288 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x790444695bb0>` 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`. I0425 16:31:07.897469 133149995796288 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79044469af30>` 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`. I0425 16:31:07.897505 133149995796288 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 0x790444695bb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x790444695bb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79044469af30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79044469af30>}). I0425 16:31:07.898013 133149995796288 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:31:07.898107 133149995796288 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x790444480180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:31:09.385701 133029053835008 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777134502111907538, 'commit_timestamp_nsecs': 1777134548902287294, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9/_CHECKPOINT_METADATA I0425 16:31:09.639721 133029053835008 checkpoint_manager.py:1877] Missing metrics for step 9 I0425 16:31:10.050245 133149995796288 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints I0425 16:31:10.929496 133149995796288 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7917a1bad730> I0425 16:31:10.929691 133149995796288 checkpointing.py:302] Checkpoint manager created! I0425 16:31:11.841499 133149995796288 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:31:11.841609 133149995796288 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:31:12.219149 133149995796288 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:31:12.219247 133149995796288 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:31:12.236119 133149995796288 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:31:12.236186 133149995796288 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:31:12.260122 133149995796288 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:31:12.260190 133149995796288 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:31:12.277037 133149995796288 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:31:12.277111 133149995796288 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:31:12.293820 133149995796288 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:31:12.293879 133149995796288 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:31:12.319724 133149995796288 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:31:12.319789 133149995796288 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:31:12.340608 133149995796288 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:31:12.340678 133149995796288 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:31:12.542642 133149995796288 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:31:12.542759 133149995796288 checkpointing.py:582] restoring from this run's directory step 9 I0425 16:31:13.041898 133149995796288 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 16:31:14.457484 1432 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:31:20.391780 133149995796288 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 266.055 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.939647197723389 s) (per-host) I0425 16:31:20.405699 133149995796288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 249.378 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.346176385879517 s) (per-host) I0425 16:31:20.929368 133149995796288 checkpointer.py:318] Finished restoring checkpoint in 8.13 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 16:31:20.929530 133149995796288 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1777134672.8007874, 'checkpointer_duration_secs': 8.128712892532349, 'checkpoint_manager_start_time': 1777134672.5430114, 'checkpoint_manager_duration_secs': 8.386491060256958} fsdp: 32 I0425 16:31:21.352114 133149995796288 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0425 16:31:21.352254 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 16:31:21.352311 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 16:31:21.352374 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0425 16:31:21.352428 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0425 16:31:21.352469 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0425 16:31:21.352523 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0425 16:31:21.352578 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0425 16:31:21.352626 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0425 16:31:21.352665 133149995796288 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0425 16:31:21.352713 133149995796288 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0425 16:31:21.352760 133149995796288 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0425 16:31:21.861926 133149995796288 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:31:21.862021 133149995796288 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:31:21.877244 133149995796288 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:31:21.877304 133149995796288 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:31:32.952649 133149995796288 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. I0425 16:31:32.953582 133149995796288 metric_logger.py:289] number parameters: 1.104 billion I0425 16:31:44.216252 133149995796288 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 16:31:44.374009 133149995796288 checkpointing.py:798] Waited 0.15773940086364746 seconds for step 10 to finish before starting checkpointing. I0425 16:31:44.376456 133149995796288 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:31:44.378362 133149995796288 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0425 16:31:44.381059 133149995796288 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10. I0425 16:31:45.332460 133149995796288 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:31:45.333697 133149995796288 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:31:45.333755 133149995796288 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 I0425 16:31:45.578250 133149995796288 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245882s I0425 16:31:45.578436 133149995796288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.139 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4914119243621826 s) (per-host) I0425 16:31:45.578490 133149995796288 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.491478s (batch_requests_ready=0.243468s, total_serialization_initiated=0.247936s, others=0.000074s) I0425 16:31:45.578582 133149995796288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.495570s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.495551) I0425 16:31:45.579470 133028961580800 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:31:45.579613 133149995796288 async_checkpointer.py:561] Finished blocking save. Time taken: 1.201180s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10. I0425 16:31:45.911406 133149995796288 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:31:45.911805 133023098664704 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:31:45.911967 133149995796288 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134704.3764365, 'wait_for_prev_duration_secs': 6.723403930664062e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777134704.3784018, 'checkpointer_blocking_duration_secs': 1.2013897895812988, 'get_old_steps_start_time': 1777134705.5798087, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777134704.3745651, 'checkpoint_manager_blocking_duration_secs': 1.537365436553955} I0425 16:31:45.912082 133149995796288 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 16:31:45.912151 133149995796288 max_utils.py:750] Memstats: After params initialized: I0425 16:31:45.912209 133149995796288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 16:31:45.912243 133149995796288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 16:31:45.912270 133149995796288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 16:31:45.912296 133149995796288 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 16:31:46.186260 133149995796288 metric_logger.py:185] completed step: 10, seconds: 11.263, TFLOP/s/device: 1.206, Tokens/s/device: 181.842, total_weights: 65536, loss: 6.833 I0425 16:31:46.344985 133149995796288 metric_logger.py:185] completed step: 11, seconds: 1.969, TFLOP/s/device: 6.902, Tokens/s/device: 1040.312, total_weights: 65536, loss: 6.528 I0425 16:31:46.492806 133149995796288 metric_logger.py:185] completed step: 12, seconds: 0.012, TFLOP/s/device: 1179.643, Tokens/s/device: 177808.647, total_weights: 65536, loss: 6.265 I0425 16:31:46.640037 133149995796288 metric_logger.py:185] completed step: 13, seconds: 0.154, TFLOP/s/device: 88.457, Tokens/s/device: 13333.247, total_weights: 65536, loss: 6.045 I0425 16:31:46.644721 133149995796288 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 16:31:46.934774 133149995796288 checkpointing.py:798] Waited 0.29001760482788086 seconds for step 15 to finish before starting checkpointing. I0425 16:31:46.938316 133149995796288 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:31:50.433145 133024280712960 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/10/items/array_metadatas/process_6 I0425 16:32:27.321150 133028961580800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.401 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.23407435417175 s) (per-host) I0425 16:32:27.321273 133028961580800 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.741580s. I0425 16:32:34.491853 133028961580800 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.912144s. I0425 16:32:34.492167 133023098664704 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:32:34.492291 133023098664704 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:32:34.492341 133023098664704 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:32:34.493762 133023098664704 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 16:32:34.493939 133149995796288 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0425 16:32:34.494073 133149995796288 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.555759 seconds. If this number is high, consider checkpointing less frequently. I0425 16:32:34.495851 133149995796288 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0425 16:32:34.500006 133149995796288 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15. I0425 16:32:35.477039 133149995796288 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:32:35.477152 133149995796288 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 I0425 16:32:35.518387 133149995796288 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042475s I0425 16:32:35.518568 133149995796288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.064 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.304610013961792 s) (per-host) I0425 16:32:35.518628 133149995796288 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304676s (batch_requests_ready=0.260240s, total_serialization_initiated=0.044362s, others=0.000073s) I0425 16:32:35.518746 133149995796288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.309464s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.309451) I0425 16:32:35.519603 133022024189696 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:32:35.519714 133149995796288 async_checkpointer.py:561] Finished blocking save. Time taken: 1.023792s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15. I0425 16:32:35.766553 133149995796288 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:32:35.766949 133023098664704 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:32:35.767131 133149995796288 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134706.9382882, 'wait_for_prev_duration_secs': 47.55575942993164, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777134754.4958901, 'checkpointer_blocking_duration_secs': 1.0239410400390625, 'get_old_steps_start_time': 1777134755.519849, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1777134706.9354458, 'checkpoint_manager_blocking_duration_secs': 48.831636905670166} I0425 16:32:35.767248 133149995796288 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 16:32:35.767838 133149995796288 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.188, Tokens/s/device: 13895.580, total_weights: 65536, loss: 5.864 I0425 16:32:35.773811 133149995796288 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.293, Tokens/s/device: 13911.342, total_weights: 65536, loss: 5.719 I0425 16:32:53.517028 133149995796288 metric_logger.py:185] completed step: 16, seconds: 49.128, TFLOP/s/device: 0.277, Tokens/s/device: 41.687, total_weights: 65536, loss: 5.605 I0425 16:32:53.664403 133149995796288 metric_logger.py:185] completed step: 17, seconds: 17.597, TFLOP/s/device: 0.772, Tokens/s/device: 116.383, total_weights: 65536, loss: 5.517 I0425 16:32:53.811869 133149995796288 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.254, Tokens/s/device: 13453.327, total_weights: 65536, loss: 5.448 I0425 16:32:53.958915 133149995796288 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 16:32:53.959619 133149995796288 checkpointing.py:798] Waited 0.0007240772247314453 seconds for step 19 to finish before starting checkpointing. I0425 16:32:53.961560 133149995796288 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:32:57.395173 133024280712960 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/15/items/array_metadatas/process_6 I0425 16:33:34.201429 133022024189696 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.779 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 58.98742985725403 s) (per-host) I0425 16:33:34.201560 133022024189696 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 58.681892s. I0425 16:33:40.179494 133022024189696 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 64.659811s. I0425 16:33:40.179787 133023098664704 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:33:40.179925 133023098664704 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:33:40.179975 133023098664704 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:33:40.181489 133023098664704 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 16:33:40.181672 133149995796288 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 16:33:40.181809 133149995796288 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.220263 seconds. If this number is high, consider checkpointing less frequently. I0425 16:33:40.183393 133149995796288 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0425 16:33:40.186774 133149995796288 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19. I0425 16:33:40.732152 133149995796288 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:33:40.732252 133149995796288 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 I0425 16:33:40.776143 133149995796288 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044986s I0425 16:33:40.776309 133149995796288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.944 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3120238780975342 s) (per-host) I0425 16:33:40.776355 133149995796288 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.312081s (batch_requests_ready=0.264948s, total_serialization_initiated=0.047071s, others=0.000062s) I0425 16:33:40.776444 133149995796288 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.315982s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.315970) I0425 16:33:40.777348 133028953188096 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:33:40.777509 133149995796288 async_checkpointer.py:561] Finished blocking save. Time taken: 0.594047s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19. I0425 16:33:40.803626 133149995796288 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:33:40.803895 133023098664704 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:33:40.804056 133149995796288 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134773.9615295, 'wait_for_prev_duration_secs': 46.22026300430298, 'time_between_consecutive_saves_sec': 19.467724323272705, 'checkpointer_blocking_start_time': 1777134820.1834311, 'checkpointer_blocking_duration_secs': 0.5941839218139648, 'get_old_steps_start_time': 1777134820.7776282, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1777134773.959847, 'checkpoint_manager_blocking_duration_secs': 46.844178676605225} I0425 16:33:40.804179 133149995796288 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 16:33:40.804225 133149995796288 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:33:45.537413 133028961580800 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/19/items/array_metadatas/process_6 I0425 16:34:22.354598 133028953188096 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.708 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.890259742736816 s) (per-host) I0425 16:34:22.354717 133028953188096 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.577253s. I0425 16:34:27.858705 133028953188096 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.081227s. I0425 16:34:27.859008 133023098664704 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:34:27.859151 133023098664704 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:34:27.859194 133023098664704 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:34:27.861949 133023098664704 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 16:34:27.862138 133149995796288 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0425 16:34:27.862285 133149995796288 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:34:27.863005 133149995796288 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.245, Tokens/s/device: 13904.164, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 16:34:39 UTC 2026 EXIT_CODE=0