XPK Start: Mon Apr 20 20:34:31 UTC 2026 2026-04-20 20:34:35.704693: 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 00:00:1776717275.717702 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776717275.721444 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776717275.732680 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717275.732702 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717275.732705 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717275.732707 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 20:34:54.915155: 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) I0420 20:34:55.436902 135426393175872 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 20:35:04,479:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 20:35:04.479280 135426393175872 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 20:35:04,481:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-sv-gzvb4-slice-job-0-0.mt-16-zarr3-true-sv-gzvb4:8482 I0420 20:35:04.481698 135426393175872 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-sv-gzvb4-slice-job-0-0.mt-16-zarr3-true-sv-gzvb4:8482 I0420 20:35:05.348736 135426393175872 max_utils.py:284] Jax distributed system initialized! I0420 20:35:11.271448 135426393175872 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 20:35:11.271554 135426393175872 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 20:35:11.271595 135426393175872 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 I0420 20:35:11.271629 135426393175872 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 20:35:11.893594 135426393175872 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:35:11.894225 135426393175872 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:35:11.894429 135426393175872 checkpointing.py:688] Setting up checkpoint logger... I0420 20:35:11.894496 135426393175872 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 20:35:11.894560 135426393175872 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 20:35:11.894935 135426393175872 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 0x7b2ade375040>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 20:35:14.815028 135426393175872 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 20:35:14.815375 135426393175872 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b1844b39730>}, handler_registry=None I0420 20:35:14.815610 135426393175872 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b1844b39730>` 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`. I0420 20:35:14.815671 135426393175872 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b1844b3e6f0>` 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`. I0420 20:35:14.815708 135426393175872 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 0x7b1844b39730>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b1844b39730>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b1844b3e6f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b1844b3e6f0>}). I0420 20:35:14.816251 135426393175872 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 20:35:14.816331 135426393175872 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b16f0410180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 20:35:15.553971 135426393175872 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints I0420 20:35:15.985222 135426393175872 checkpoint_manager.py:929] [process=7][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_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b1844b3b1d0> I0420 20:35:15.985394 135426393175872 checkpointing.py:302] Checkpoint manager created! I0420 20:35:16.877164 135426393175872 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 20:35:16.877276 135426393175872 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:35:17.246869 135426393175872 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:35:17.246965 135426393175872 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:35:17.262723 135426393175872 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:35:17.262784 135426393175872 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:35:17.285349 135426393175872 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:35:17.285419 135426393175872 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:35:17.301234 135426393175872 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:35:17.301294 135426393175872 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:35:17.317021 135426393175872 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:35:17.317082 135426393175872 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:35:17.341635 135426393175872 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 20:35:17.341715 135426393175872 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:35:17.361022 135426393175872 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 20:35:17.361089 135426393175872 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 20:35:17.548648 135426393175872 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 20:35:17.548766 135426393175872 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 20:35:19.006975 135426393175872 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0420 20:35:19.007113 135426393175872 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) I0420 20:35:19.007169 135426393175872 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) I0420 20:35:19.007230 135426393175872 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') I0420 20:35:19.007284 135426393175872 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) I0420 20:35:19.007323 135426393175872 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) I0420 20:35:19.007375 135426393175872 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) I0420 20:35:19.007427 135426393175872 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') I0420 20:35:19.007467 135426393175872 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) I0420 20:35:19.007506 135426393175872 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) I0420 20:35:19.007556 135426393175872 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0420 20:35:19.007605 135426393175872 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0420 20:35:19.504562 135426393175872 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:35:19.504669 135426393175872 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:35:19.519610 135426393175872 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:35:19.519680 135426393175872 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:35:30.430575 135426393175872 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. I0420 20:35:30.431415 135426393175872 metric_logger.py:289] number parameters: 1.104 billion I0420 20:35:41.618473 135426393175872 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 20:35:41.771194 135426393175872 checkpointing.py:798] Waited 0.15270471572875977 seconds for step 0 to finish before starting checkpointing. I0420 20:35:41.773648 135426393175872 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:35:41.775549 135426393175872 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 20:35:41.778381 135426393175872 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/0. I0420 20:35:42.794337 135426393175872 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 20:35:42.795334 135426393175872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:35:42.795394 135426393175872 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 I0420 20:35:43.051083 135426393175872 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.256823s I0420 20:35:43.051256 135426393175872 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.843 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5426106452941895 s) (per-host) I0420 20:35:43.051308 135426393175872 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.542672s (batch_requests_ready=0.282212s, total_serialization_initiated=0.260390s, others=0.000070s) I0420 20:35:43.051398 135426393175872 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.546933s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.546914) I0420 20:35:43.052323 135305554290432 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:35:43.052496 135426393175872 async_checkpointer.py:561] Finished blocking save. Time taken: 1.276879s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/0. I0420 20:35:43.074258 135426393175872 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:35:43.074511 135301225752320 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:35:43.074633 135426393175872 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776717341.77363, 'wait_for_prev_duration_secs': 7.891654968261719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776717341.775588, 'checkpointer_blocking_duration_secs': 1.277015209197998, 'get_old_steps_start_time': 1776717343.0526245, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776717341.771816, 'checkpoint_manager_blocking_duration_secs': 1.302778720855713} I0420 20:35:43.074749 135426393175872 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 20:35:43.074799 135426393175872 max_utils.py:750] Memstats: After params initialized: I0420 20:35:43.074849 135426393175872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 20:35:43.074881 135426393175872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 20:35:43.074908 135426393175872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 20:35:43.074932 135426393175872 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 20:35:43.342088 135426393175872 metric_logger.py:185] completed step: 0, seconds: 11.187, TFLOP/s/device: 1.215, Tokens/s/device: 183.070, total_weights: 65536, loss: 10.874 I0420 20:35:43.522459 135426393175872 metric_logger.py:185] completed step: 1, seconds: 1.722, TFLOP/s/device: 7.889, Tokens/s/device: 1189.135, total_weights: 65536, loss: 10.874 I0420 20:35:43.943585 135426393175872 metric_logger.py:185] completed step: 2, seconds: 0.033, TFLOP/s/device: 406.557, Tokens/s/device: 61280.670, total_weights: 65536, loss: 10.560 I0420 20:35:44.091103 135426393175872 metric_logger.py:185] completed step: 3, seconds: 0.422, TFLOP/s/device: 32.231, Tokens/s/device: 4858.157, total_weights: 65536, loss: 9.980 I0420 20:35:44.095832 135426393175872 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 20:35:44.385869 135426393175872 checkpointing.py:798] Waited 0.29000353813171387 seconds for step 5 to finish before starting checkpointing. I0420 20:35:44.388731 135426393175872 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:35:46.222838 2572 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 20:35:48.131727 135305562683136 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/0/items/array_metadatas/process_7 I0420 20:36:20.087028 135305554290432 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 42.035 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.5783257484436 s) (per-host) I0420 20:36:20.087154 135305554290432 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.034718s. I0420 20:36:28.018297 135305554290432 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 44.965843s. I0420 20:36:28.018545 135301225752320 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:36:28.018614 135301225752320 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:36:28.018699 135301225752320 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:36:28.020328 135301225752320 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 20:36:28.020522 135426393175872 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 20:36:28.020697 135426393175872 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.631969 seconds. If this number is high, consider checkpointing less frequently. I0420 20:36:28.022579 135426393175872 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 5 I0420 20:36:28.026670 135426393175872 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/5. I0420 20:36:28.588153 135426393175872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:36:28.588260 135426393175872 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 I0420 20:36:28.637923 135426393175872 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051553s I0420 20:36:28.638089 135426393175872 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.853 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31785154342651367 s) (per-host) I0420 20:36:28.638142 135426393175872 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.317912s (batch_requests_ready=0.264450s, total_serialization_initiated=0.053394s, others=0.000068s) I0420 20:36:28.638242 135426393175872 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.322788s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.322773) I0420 20:36:28.638987 135301267715840 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:36:28.639142 135426393175872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.616500s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/5. I0420 20:36:29.058922 135426393175872 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:36:29.059300 135301225752320 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:36:29.059476 135426393175872 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776717344.3887029, 'wait_for_prev_duration_secs': 43.63196897506714, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776717388.0226204, 'checkpointer_blocking_duration_secs': 0.6165950298309326, 'get_old_steps_start_time': 1776717388.6392329, 'get_old_steps_duration_secs': 2.288818359375e-05, 'checkpoint_manager_blocking_start_time': 1776717344.3864667, 'checkpoint_manager_blocking_duration_secs': 44.67297315597534} I0420 20:36:29.059582 135426393175872 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 20:36:29.060216 135426393175872 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 89.044, Tokens/s/device: 13421.675, total_weights: 65536, loss: 9.460 I0420 20:36:29.066456 135426393175872 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.170, Tokens/s/device: 13892.940, total_weights: 65536, loss: 8.959 I0420 20:36:48.153185 135426393175872 metric_logger.py:185] completed step: 6, seconds: 44.969, TFLOP/s/device: 0.302, Tokens/s/device: 45.542, total_weights: 65536, loss: 8.469 I0420 20:36:48.300810 135426393175872 metric_logger.py:185] completed step: 7, seconds: 18.941, TFLOP/s/device: 0.717, Tokens/s/device: 108.128, total_weights: 65536, loss: 8.003 I0420 20:36:48.448193 135426393175872 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.265, Tokens/s/device: 13454.918, total_weights: 65536, loss: 7.572 I0420 20:36:48.595278 135426393175872 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 20:36:48.595979 135426393175872 checkpointing.py:798] Waited 0.0007193088531494141 seconds for step 9 to finish before starting checkpointing. I0420 20:36:48.598072 135426393175872 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:36:52.583528 135305562683136 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/5/items/array_metadatas/process_7 I0420 20:37:29.353746 135301267715840 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 25.881 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.033466815948486 s) (per-host) I0420 20:37:29.353866 135301267715840 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.714761s. I0420 20:37:37.571894 135301267715840 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 68.932772s. I0420 20:37:37.572192 135301225752320 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:37:37.572326 135301225752320 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:37:37.572374 135301225752320 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:37:37.573907 135301225752320 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 20:37:37.574102 135426393175872 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 20:37:37.574246 135426393175872 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.976176 seconds. If this number is high, consider checkpointing less frequently. I0420 20:37:37.575854 135426393175872 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 20:37:37.579243 135426393175872 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/9. I0420 20:37:38.150811 135426393175872 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:37:38.150911 135426393175872 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 I0420 20:37:38.201493 135426393175872 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051802s I0420 20:37:38.201677 135426393175872 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.921 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31348276138305664 s) (per-host) I0420 20:37:38.201726 135426393175872 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.313561s (batch_requests_ready=0.259602s, total_serialization_initiated=0.053876s, others=0.000083s) I0420 20:37:38.201816 135426393175872 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.317879s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.317868) I0420 20:37:38.202700 135305554290432 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:37:38.202858 135426393175872 async_checkpointer.py:561] Finished blocking save. Time taken: 0.626937s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/9. I0420 20:37:38.222340 135426393175872 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:37:38.222569 135301225752320 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:37:38.222715 135426393175872 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776717408.5980444, 'wait_for_prev_duration_secs': 48.9761757850647, 'time_between_consecutive_saves_sec': 20.577677488327026, 'checkpointer_blocking_start_time': 1776717457.5758939, 'checkpointer_blocking_duration_secs': 0.6270713806152344, 'get_old_steps_start_time': 1776717458.2029793, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776717408.5962067, 'checkpoint_manager_blocking_duration_secs': 49.62647795677185} I0420 20:37:38.222820 135426393175872 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 20:37:38.222861 135426393175872 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:37:43.137279 135305562683136 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_true/checkpoints/9/items/array_metadatas/process_7 I0420 20:38:19.636641 135305554290432 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.836 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.74842357635498 s) (per-host) I0420 20:38:19.636776 135305554290432 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.433964s. I0420 20:38:27.605625 135305554290432 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.402797s. I0420 20:38:27.605928 135301225752320 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:38:27.606050 135301225752320 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:38:27.606091 135301225752320 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:38:27.607745 135301225752320 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 20:38:27.607925 135426393175872 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 20:38:27.608066 135426393175872 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:38:27.608776 135426393175872 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.147, Tokens/s/device: 13889.454, 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: Mon Apr 20 20:38:37 UTC 2026 EXIT_CODE=0