XPK Start: Sun Apr 19 04:47:39 UTC 2026 2026-04-19 04:47:43.675663: 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 00:00:1776574063.688193 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776574063.692028 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776574063.703117 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776574063.703136 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776574063.703139 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776574063.703141 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 04:48:03.703731: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0419 04:48:04.233399 139461108422464 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 04:48:13,274:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 04:48:13.274840 139461108422464 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 04:48:13,277:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-cbnxe-slice-job-0-0.mt-16-zarr3-false-sv-cbnxe:8482 I0419 04:48:13.277341 139461108422464 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-cbnxe-slice-job-0-0.mt-16-zarr3-false-sv-cbnxe:8482 I0419 04:48:14.266807 139461108422464 max_utils.py:284] Jax distributed system initialized! I0419 04:48:21.197294 139461108422464 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 04:48:21.197401 139461108422464 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 04:48:21.197441 139461108422464 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 I0419 04:48:21.197476 139461108422464 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 04:48:21.818655 139461108422464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:48:21.819266 139461108422464 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 04:48:21.819454 139461108422464 checkpointing.py:688] Setting up checkpoint logger... I0419 04:48:21.819510 139461108422464 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0419 04:48:21.819555 139461108422464 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 04:48:21.819901 139461108422464 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 0x7ed643779280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 04:48:25.512111 139461108422464 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 04:48:25.512482 139461108422464 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 0x7ec3adbe9970>}, handler_registry=None I0419 04:48:25.512725 139461108422464 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ec3adbe9970>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0419 04:48:25.512774 139461108422464 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec3adbeee40>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0419 04:48:25.512813 139461108422464 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 0x7ec3adbe9970>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ec3adbe9970>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec3adbeee40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ec3adbeee40>}). I0419 04:48:25.513330 139461108422464 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 04:48:25.513410 139461108422464 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ec3ad1cc180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 04:48:26.618385 139461108422464 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints I0419 04:48:27.518934 139461108422464 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ec3adbe8950> I0419 04:48:27.519110 139461108422464 checkpointing.py:302] Checkpoint manager created! I0419 04:48:28.403334 139461108422464 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 04:48:28.403439 139461108422464 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:48:28.772814 139461108422464 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:48:28.772910 139461108422464 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:48:28.788811 139461108422464 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 04:48:28.788870 139461108422464 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 04:48:28.811828 139461108422464 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:48:28.811899 139461108422464 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:48:28.827835 139461108422464 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:48:28.827896 139461108422464 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:48:28.843991 139461108422464 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 04:48:28.844055 139461108422464 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:48:28.868978 139461108422464 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 04:48:28.869048 139461108422464 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 04:48:28.888650 139461108422464 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 04:48:28.888715 139461108422464 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 04:48:29.079711 139461108422464 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 04:48:29.079821 139461108422464 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 04:48:30.518885 139461108422464 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0419 04:48:30.519020 139461108422464 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) I0419 04:48:30.519077 139461108422464 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) I0419 04:48:30.519136 139461108422464 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') I0419 04:48:30.519189 139461108422464 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) I0419 04:48:30.519229 139461108422464 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) I0419 04:48:30.519315 139461108422464 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) I0419 04:48:30.519374 139461108422464 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') I0419 04:48:30.519416 139461108422464 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) I0419 04:48:30.519453 139461108422464 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) I0419 04:48:30.519507 139461108422464 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0419 04:48:30.519557 139461108422464 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0419 04:48:31.040408 139461108422464 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:48:31.040503 139461108422464 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:48:31.055553 139461108422464 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 04:48:31.055611 139461108422464 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 04:48:41.794898 139461108422464 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0419 04:48:41.796113 139461108422464 metric_logger.py:289] number parameters: 1.104 billion I0419 04:48:52.765517 139461108422464 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 04:48:53.184466 139461108422464 checkpointing.py:798] Waited 0.41893672943115234 seconds for step 0 to finish before starting checkpointing. I0419 04:48:53.186826 139461108422464 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:48:53.188535 139461108422464 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0419 04:48:53.191344 139461108422464 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/0. I0419 04:48:54.639030 139461108422464 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 04:48:54.640052 139461108422464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:48:54.640109 139461108422464 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 04:48:54.890974 139461108422464 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252024s I0419 04:48:54.891141 139461108422464 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.865 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5383481979370117 s) (per-host) I0419 04:48:54.891192 139461108422464 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.538408s (batch_requests_ready=0.282681s, total_serialization_initiated=0.255660s, others=0.000067s) I0419 04:48:54.891297 139461108422464 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.542435s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.542415) I0419 04:48:54.892284 139338218182400 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:48:54.892457 139461108422464 async_checkpointer.py:561] Finished blocking save. Time taken: 1.703851s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/0. I0419 04:48:55.304086 139461108422464 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:48:55.304502 139333403129600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:48:55.304671 139461108422464 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776574133.186808, 'wait_for_prev_duration_secs': 6.270408630371094e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776574133.1885748, 'checkpointer_blocking_duration_secs': 1.7039895057678223, 'get_old_steps_start_time': 1776574134.8925889, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1776574133.185065, 'checkpoint_manager_blocking_duration_secs': 2.1195638179779053} I0419 04:48:55.304778 139461108422464 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 04:48:55.304831 139461108422464 max_utils.py:750] Memstats: After params initialized: I0419 04:48:55.304890 139461108422464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 04:48:55.304926 139461108422464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 04:48:55.304959 139461108422464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 04:48:55.304983 139461108422464 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 04:48:55.568681 139461108422464 metric_logger.py:185] completed step: 0, seconds: 10.969, TFLOP/s/device: 1.239, Tokens/s/device: 186.705, total_weights: 65536, loss: 10.874 I0419 04:48:55.738358 139461108422464 metric_logger.py:185] completed step: 1, seconds: 2.802, TFLOP/s/device: 4.849, Tokens/s/device: 730.944, total_weights: 65536, loss: 10.874 I0419 04:48:56.145744 139461108422464 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 608.225, Tokens/s/device: 91678.231, total_weights: 65536, loss: 10.560 I0419 04:48:56.293315 139461108422464 metric_logger.py:185] completed step: 3, seconds: 0.408, TFLOP/s/device: 33.293, Tokens/s/device: 5018.304, total_weights: 65536, loss: 9.980 I0419 04:48:56.297933 139461108422464 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0419 04:48:56.587849 139461108422464 checkpointing.py:798] Waited 0.28988170623779297 seconds for step 5 to finish before starting checkpointing. I0419 04:48:56.590751 139461108422464 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:48:57.925441 2607 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 04:49:00.607411 139338226575104 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0419 04:49:34.782530 139338218182400 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 39.070 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.42969346046448 s) (per-host) I0419 04:49:34.782666 139338218182400 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.890265s. I0419 04:49:46.822765 139338218182400 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.930348s. I0419 04:49:46.823039 139333403129600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:49:46.823162 139333403129600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:49:46.823212 139333403129600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:49:46.825142 139333403129600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 04:49:46.825315 139461108422464 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 04:49:46.825443 139461108422464 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.234698 seconds. If this number is high, consider checkpointing less frequently. I0419 04:49:46.827291 139461108422464 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0419 04:49:46.831530 139461108422464 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/5. I0419 04:49:48.234196 139461108422464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:49:48.234313 139461108422464 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 04:49:48.277848 139461108422464 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044813s I0419 04:49:48.278010 139461108422464 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.093 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7368812561035156 s) (per-host) I0419 04:49:48.278062 139461108422464 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.736943s (batch_requests_ready=0.690243s, total_serialization_initiated=0.046630s, others=0.000069s) I0419 04:49:48.278145 139461108422464 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.741910s (all_items=0.000015s, per_item={'items': '0.00001478'}, temp_paths=0.741895) I0419 04:49:48.278980 139338226575104 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:49:48.279144 139461108422464 async_checkpointer.py:561] Finished blocking save. Time taken: 1.451780s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/5. I0419 04:49:48.284945 139461108422464 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:49:48.285182 139333403129600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:49:48.285310 139461108422464 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776574136.590722, 'wait_for_prev_duration_secs': 50.23469805717468, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776574186.8273332, 'checkpointer_blocking_duration_secs': 1.4519307613372803, 'get_old_steps_start_time': 1776574188.2792857, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776574136.5884714, 'checkpoint_manager_blocking_duration_secs': 51.696802616119385} I0419 04:49:48.285474 139461108422464 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0419 04:49:48.285988 139461108422464 metric_logger.py:185] completed step: 4, seconds: 0.152, TFLOP/s/device: 89.144, Tokens/s/device: 13436.821, total_weights: 65536, loss: 9.460 I0419 04:49:48.291499 139461108422464 metric_logger.py:185] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.054, Tokens/s/device: 13875.433, total_weights: 65536, loss: 8.959 I0419 04:50:07.735779 139461108422464 metric_logger.py:185] completed step: 6, seconds: 51.992, TFLOP/s/device: 0.261, Tokens/s/device: 39.390, total_weights: 65536, loss: 8.469 I0419 04:50:07.883633 139461108422464 metric_logger.py:185] completed step: 7, seconds: 19.298, TFLOP/s/device: 0.704, Tokens/s/device: 106.126, total_weights: 65536, loss: 8.003 I0419 04:50:08.031117 139461108422464 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.198, Tokens/s/device: 13444.849, total_weights: 65536, loss: 7.572 I0419 04:50:08.178214 139461108422464 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 04:50:08.178906 139461108422464 checkpointing.py:798] Waited 0.0007138252258300781 seconds for step 9 to finish before starting checkpointing. I0419 04:50:08.180972 139461108422464 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:50:12.513043 139332337780480 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0419 04:50:47.522533 139338226575104 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.335 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.98137164115906 s) (per-host) I0419 04:50:47.522650 139338226575104 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.243560s. I0419 04:50:57.142393 139338226575104 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.863287s. I0419 04:50:57.142723 139333403129600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:50:57.142859 139333403129600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:50:57.142910 139333403129600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:50:57.144525 139333403129600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0419 04:50:57.144713 139461108422464 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0419 04:50:57.144853 139461108422464 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.963892 seconds. If this number is high, consider checkpointing less frequently. I0419 04:50:57.146389 139461108422464 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0419 04:50:57.149682 139461108422464 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/9. I0419 04:50:58.122151 139461108422464 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0419 04:50:58.122264 139461108422464 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 04:50:58.163230 139461108422464 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042265s I0419 04:50:58.163410 139461108422464 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.315 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2902078628540039 s) (per-host) I0419 04:50:58.163464 139461108422464 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290272s (batch_requests_ready=0.245892s, total_serialization_initiated=0.044311s, others=0.000070s) I0419 04:50:58.163553 139461108422464 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294361s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.294349) I0419 04:50:58.164421 139333965715200 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 04:50:58.164579 139461108422464 async_checkpointer.py:561] Finished blocking save. Time taken: 1.018120s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/9. I0419 04:50:58.614708 139461108422464 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 04:50:58.615068 139333403129600 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 04:50:58.615249 139461108422464 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776574208.1809444, 'wait_for_prev_duration_secs': 48.963892459869385, 'time_between_consecutive_saves_sec': 21.355762004852295, 'checkpointer_blocking_start_time': 1776574257.1464279, 'checkpointer_blocking_duration_secs': 1.0182621479034424, 'get_old_steps_start_time': 1776574258.1647034, 'get_old_steps_duration_secs': 1.9311904907226562e-05, 'checkpoint_manager_blocking_start_time': 1776574208.1791298, 'checkpoint_manager_blocking_duration_secs': 50.4360716342926} I0419 04:50:58.615355 139461108422464 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 04:50:58.615398 139461108422464 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 04:51:03.820530 139338218182400 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_20260419_025828/linen_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0419 04:51:39.019958 139333965715200 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.389 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.146716594696045 s) (per-host) I0419 04:51:39.020079 139333965715200 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.855549s. I0419 04:51:48.663141 139333965715200 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.498595s. I0419 04:51:48.663476 139333403129600 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 04:51:48.663596 139333403129600 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 04:51:48.663637 139333403129600 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 04:51:48.665390 139333403129600 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 04:51:48.665561 139461108422464 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 04:51:48.665702 139461108422464 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 04:51:48.666380 139461108422464 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.928, Tokens/s/device: 13856.375, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 04:52:00 UTC 2026 EXIT_CODE=0