XPK Start: Fri Apr 24 16:16:01 UTC 2026 2026-04-24 16:16:05.488109: 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 00:00:1777047365.501431 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777047365.505176 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777047365.516896 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047365.516915 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047365.516918 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777047365.516920 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 16:16:24.824614: 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) I0424 16:16:25.355512 134097730574144 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 16:16:34,397:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 16:16:34.397602 134097730574144 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 16:16:34,399:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-2oe4q-slice-job-0-0.mt-16-zarr3-true-rs-2oe4q:8482 I0424 16:16:34.399927 134097730574144 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-2oe4q-slice-job-0-0.mt-16-zarr3-true-rs-2oe4q:8482 I0424 16:16:35.293393 134097730574144 max_utils.py:284] Jax distributed system initialized! I0424 16:16:41.117593 134097730574144 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 16:16:41.117702 134097730574144 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 16:16:41.117743 134097730574144 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 I0424 16:16:41.117778 134097730574144 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 16:16:41.743547 134097730574144 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:16:41.744159 134097730574144 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 16:16:41.744342 134097730574144 checkpointing.py:688] Setting up checkpoint logger... I0424 16:16:41.744395 134097730574144 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 16:16:41.744440 134097730574144 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 16:16:41.744806 134097730574144 base_pytree_checkpoint_handler.py:415] 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 0x79f5933790d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 16:16:43.101892 134097730574144 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 16:16:43.102266 134097730574144 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 0x79e348c65b50>}, handler_registry=None I0424 16:16:43.102507 134097730574144 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e348c65b50>` 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`. I0424 16:16:43.102559 134097730574144 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e348c6ac30>` 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`. I0424 16:16:43.102598 134097730574144 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 0x79e348c65b50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79e348c65b50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e348c6ac30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79e348c6ac30>}). I0424 16:16:43.103126 134097730574144 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 16:16:43.103213 134097730574144 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79e184644180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 16:16:44.548285 133978619115264 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777047236590118433, 'commit_timestamp_nsecs': 1777047283107607999, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9/_CHECKPOINT_METADATA I0424 16:16:44.800869 133978619115264 checkpoint_manager.py:1877] Missing metrics for step 9 I0424 16:16:44.836854 134097730574144 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints I0424 16:16:45.400233 134097730574144 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79e348c660f0> I0424 16:16:45.400413 134097730574144 checkpointing.py:302] Checkpoint manager created! I0424 16:16:46.298756 134097730574144 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 16:16:46.298866 134097730574144 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:16:46.679123 134097730574144 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:16:46.679219 134097730574144 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:16:46.695523 134097730574144 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 16:16:46.695595 134097730574144 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 16:16:46.718908 134097730574144 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:16:46.718978 134097730574144 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:16:46.735366 134097730574144 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:16:46.735431 134097730574144 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:16:46.751723 134097730574144 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 16:16:46.751785 134097730574144 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:16:46.777128 134097730574144 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 16:16:46.777201 134097730574144 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 16:16:46.797165 134097730574144 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 16:16:46.797232 134097730574144 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 16:16:46.992153 134097730574144 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 16:16:46.992263 134097730574144 checkpointing.py:582] restoring from this run's directory step 9 I0424 16:16:47.498794 134097730574144 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 16:16:48.918920 1394 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 16:16:55.093801 134097730574144 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 255.702 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.180140018463135 s) (per-host) I0424 16:16:55.106529 134097730574144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 239.072 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.619741678237915 s) (per-host) I0424 16:16:55.338814 134097730574144 checkpointer.py:318] Finished restoring checkpoint in 8.10 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/9. I0424 16:16:55.338968 134097730574144 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1777047407.2430904, 'checkpointer_duration_secs': 8.095847129821777, 'checkpoint_manager_start_time': 1777047406.992515, 'checkpoint_manager_duration_secs': 8.34642481803894} fsdp: 32 I0424 16:16:55.751493 134097730574144 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0424 16:16:55.751628 134097730574144 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) I0424 16:16:55.751683 134097730574144 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) I0424 16:16:55.751744 134097730574144 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') I0424 16:16:55.751803 134097730574144 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) I0424 16:16:55.751847 134097730574144 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) I0424 16:16:55.751903 134097730574144 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) I0424 16:16:55.751958 134097730574144 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') I0424 16:16:55.751999 134097730574144 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) I0424 16:16:55.752039 134097730574144 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) I0424 16:16:55.752089 134097730574144 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0424 16:16:55.752180 134097730574144 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0424 16:16:56.263485 134097730574144 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:16:56.263590 134097730574144 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:16:56.278741 134097730574144 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 16:16:56.278803 134097730574144 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 16:17:07.106942 134097730574144 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. I0424 16:17:07.107939 134097730574144 metric_logger.py:289] number parameters: 1.104 billion I0424 16:17:18.289413 134097730574144 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0424 16:17:18.446358 134097730574144 checkpointing.py:798] Waited 0.15692687034606934 seconds for step 10 to finish before starting checkpointing. I0424 16:17:18.448922 134097730574144 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:17:18.450881 134097730574144 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0424 16:17:18.453634 134097730574144 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10. I0424 16:17:19.027218 134097730574144 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 16:17:19.029791 134097730574144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:17:19.029850 134097730574144 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 I0424 16:17:19.283256 134097730574144 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256147s I0424 16:17:19.283439 134097730574144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.074 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5018560886383057 s) (per-host) I0424 16:17:19.283493 134097730574144 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.501920s (batch_requests_ready=0.243655s, total_serialization_initiated=0.258193s, others=0.000072s) I0424 16:17:19.283584 134097730574144 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.506600s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.506580) I0424 16:17:19.284693 133977536980736 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:17:19.284807 134097730574144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.833853s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10. I0424 16:17:19.305248 134097730574144 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:17:19.305645 133974273791744 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:17:19.305802 134097730574144 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047438.4489028, 'wait_for_prev_duration_secs': 6.508827209472656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047438.4509194, 'checkpointer_blocking_duration_secs': 0.8339970111846924, 'get_old_steps_start_time': 1777047439.2849371, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1777047438.4469242, 'checkpoint_manager_blocking_duration_secs': 0.8588409423828125} I0424 16:17:19.305910 134097730574144 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0424 16:17:19.305963 134097730574144 max_utils.py:750] Memstats: After params initialized: I0424 16:17:19.306016 134097730574144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 16:17:19.306051 134097730574144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 16:17:19.306078 134097730574144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 16:17:19.306119 134097730574144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 16:17:19.585345 134097730574144 metric_logger.py:185] completed step: 10, seconds: 11.181, TFLOP/s/device: 1.215, Tokens/s/device: 183.162, total_weights: 65536, loss: 6.833 I0424 16:17:19.738571 134097730574144 metric_logger.py:185] completed step: 11, seconds: 1.295, TFLOP/s/device: 10.496, Tokens/s/device: 1582.008, total_weights: 65536, loss: 6.528 I0424 16:17:19.885834 134097730574144 metric_logger.py:185] completed step: 12, seconds: 0.006, TFLOP/s/device: 2214.691, Tokens/s/device: 333822.331, total_weights: 65536, loss: 6.265 I0424 16:17:20.033707 134097730574144 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.749, Tokens/s/device: 13377.227, total_weights: 65536, loss: 6.045 I0424 16:17:20.038336 134097730574144 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0424 16:17:20.328014 134097730574144 checkpointing.py:798] Waited 0.2896463871002197 seconds for step 15 to finish before starting checkpointing. I0424 16:17:20.331000 134097730574144 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:17:24.298931 133974290577152 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0424 16:18:00.980073 133977536980736 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.433 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.19845128059387 s) (per-host) I0424 16:18:00.980219 133977536980736 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.695462s. I0424 16:18:08.487830 133977536980736 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.203058s. I0424 16:18:08.488145 133974273791744 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:18:08.488265 133974273791744 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:18:08.488315 133974273791744 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:18:08.489620 133974273791744 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0424 16:18:08.489800 134097730574144 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. W0424 16:18:08.489942 134097730574144 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.158945 seconds. If this number is high, consider checkpointing less frequently. I0424 16:18:08.492053 134097730574144 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0424 16:18:08.496318 134097730574144 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15. I0424 16:18:09.053072 134097730574144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:18:09.053183 134097730574144 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 I0424 16:18:09.095006 134097730574144 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043015s I0424 16:18:09.095191 134097730574144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.162 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2988123893737793 s) (per-host) I0424 16:18:09.095246 134097730574144 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.298877s (batch_requests_ready=0.253886s, total_serialization_initiated=0.044919s, others=0.000072s) I0424 16:18:09.095337 134097730574144 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303888s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.303875) I0424 16:18:09.096181 133973747554048 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:18:09.096301 134097730574144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.604162s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15. I0424 16:18:09.337402 134097730574144 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:18:09.337793 133974273791744 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:18:09.337970 134097730574144 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047440.3309703, 'wait_for_prev_duration_secs': 48.15894532203674, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777047488.4921114, 'checkpointer_blocking_duration_secs': 0.6043610572814941, 'get_old_steps_start_time': 1777047489.0964925, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1777047440.3286502, 'checkpoint_manager_blocking_duration_secs': 49.009281635284424} I0424 16:18:09.338089 134097730574144 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0424 16:18:09.338702 134097730574144 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.236, Tokens/s/device: 13902.749, total_weights: 65536, loss: 5.864 I0424 16:18:09.344529 134097730574144 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.980, Tokens/s/device: 13864.161, total_weights: 65536, loss: 5.719 I0424 16:18:26.973964 134097730574144 metric_logger.py:185] completed step: 16, seconds: 49.305, TFLOP/s/device: 0.276, Tokens/s/device: 41.537, total_weights: 65536, loss: 5.605 I0424 16:18:27.121538 134097730574144 metric_logger.py:185] completed step: 17, seconds: 17.484, TFLOP/s/device: 0.777, Tokens/s/device: 117.138, total_weights: 65536, loss: 5.517 I0424 16:18:27.268980 134097730574144 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.436, Tokens/s/device: 13480.779, total_weights: 65536, loss: 5.448 I0424 16:18:27.416057 134097730574144 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0424 16:18:27.416736 134097730574144 checkpointing.py:798] Waited 0.0006988048553466797 seconds for step 19 to finish before starting checkpointing. I0424 16:18:27.418812 134097730574144 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:18:30.384963 133974290577152 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0424 16:19:07.335626 133973747554048 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.984 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 58.53920364379883 s) (per-host) I0424 16:19:07.335746 133973747554048 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 58.239362s. I0424 16:19:13.216228 133973747554048 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 64.119828s. I0424 16:19:13.216535 133974273791744 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:19:13.216661 133974273791744 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:19:13.216710 133974273791744 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:19:13.218451 133974273791744 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0424 16:19:13.218657 134097730574144 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. W0424 16:19:13.218772 134097730574144 checkpoint_manager.py:1458] Waiting for previous save to complete took 45.799974 seconds. If this number is high, consider checkpointing less frequently. I0424 16:19:13.220185 134097730574144 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0424 16:19:13.223701 134097730574144 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19. I0424 16:19:13.772197 134097730574144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0424 16:19:13.772293 134097730574144 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 I0424 16:19:13.813265 134097730574144 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042042s I0424 16:19:13.813431 134097730574144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.071 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3041861057281494 s) (per-host) I0424 16:19:13.813482 134097730574144 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304246s (batch_requests_ready=0.260078s, total_serialization_initiated=0.044102s, others=0.000066s) I0424 16:19:13.813577 134097730574144 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308303s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.308292) I0424 16:19:13.814516 133978577151744 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 16:19:13.814659 134097730574144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.594392s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19. I0424 16:19:13.818665 134097730574144 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0424 16:19:13.818969 133974273791744 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 16:19:13.819140 134097730574144 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777047507.4187822, 'wait_for_prev_duration_secs': 45.79997420310974, 'time_between_consecutive_saves_sec': 18.929121494293213, 'checkpointer_blocking_start_time': 1777047553.2202368, 'checkpointer_blocking_duration_secs': 0.5946033000946045, 'get_old_steps_start_time': 1777047553.814853, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1777047507.4169562, 'checkpoint_manager_blocking_duration_secs': 46.402151584625244} I0424 16:19:13.819246 134097730574144 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0424 16:19:13.819292 134097730574144 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 16:19:19.111756 133978585544448 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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0424 16:19:55.870607 133978577151744 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.289 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.36133050918579 s) (per-host) I0424 16:19:55.870712 133978577151744 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.055971s. I0424 16:20:01.925967 133978577151744 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.111212s. I0424 16:20:01.926268 133974273791744 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 16:20:01.926401 133974273791744 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 16:20:01.926444 133974273791744 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0424 16:20:01.928225 133974273791744 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0424 16:20:01.928418 134097730574144 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. I0424 16:20:01.928582 134097730574144 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 16:20:01.929268 134097730574144 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.107, Tokens/s/device: 13883.334, 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: Fri Apr 24 16:20:11 UTC 2026 EXIT_CODE=0