XPK Start: Wed Apr 22 17:21:51 UTC 2026 2026-04-22 17:21:55.459267: 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 00:00:1776878515.471921 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776878515.475911 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776878515.487462 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878515.487482 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878515.487485 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878515.487487 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 17:22:14.763776: 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) I0422 17:22:15.293379 138428558899008 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 17:22:24,333:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 17:22:24.333256 138428558899008 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 17:22:24,335:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-malz4-slice-job-0-0.mt-16-zarr3-false-sv-malz4:8482 I0422 17:22:24.335610 138428558899008 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-malz4-slice-job-0-0.mt-16-zarr3-false-sv-malz4:8482 I0422 17:22:25.179870 138428558899008 max_utils.py:284] Jax distributed system initialized! I0422 17:22:31.111905 138428558899008 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 17:22:31.112010 138428558899008 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 17:22:31.112051 138428558899008 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 I0422 17:22:31.112085 138428558899008 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 17:22:31.748682 138428558899008 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:22:31.749300 138428558899008 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:22:31.749483 138428558899008 checkpointing.py:688] Setting up checkpoint logger... I0422 17:22:31.749534 138428558899008 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0422 17:22:31.749577 138428558899008 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 17:22:31.749928 138428558899008 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 0x7de5e23793a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 17:22:35.064393 138428558899008 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 17:22:35.064631 138428558899008 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 0x7dd3e4149c70>}, handler_registry=None I0422 17:22:35.064869 138428558899008 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dd3e4149c70>` 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`. I0422 17:22:35.064918 138428558899008 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dd376f3adb0>` 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`. I0422 17:22:35.064955 138428558899008 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 0x7dd3e4149c70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dd3e4149c70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dd376f3adb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dd376f3adb0>}). I0422 17:22:35.065491 138428558899008 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 17:22:35.065572 138428558899008 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dd22c578180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 17:22:35.785812 138428558899008 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints I0422 17:22:36.218661 138428558899008 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dd376f36570> I0422 17:22:36.218839 138428558899008 checkpointing.py:302] Checkpoint manager created! I0422 17:22:37.117188 138428558899008 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:22:37.117295 138428558899008 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:22:37.485643 138428558899008 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:22:37.485743 138428558899008 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:22:37.501746 138428558899008 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:22:37.501808 138428558899008 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:22:37.524686 138428558899008 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:22:37.524755 138428558899008 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:22:37.540931 138428558899008 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:22:37.540994 138428558899008 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:22:37.557004 138428558899008 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:22:37.557067 138428558899008 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:22:37.581872 138428558899008 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 17:22:37.581943 138428558899008 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:22:37.601574 138428558899008 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 17:22:37.601640 138428558899008 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 17:22:37.792718 138428558899008 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 17:22:37.792828 138428558899008 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 17:22:39.238083 138428558899008 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 17:22:39.238234 138428558899008 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) I0422 17:22:39.238287 138428558899008 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) I0422 17:22:39.238348 138428558899008 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') I0422 17:22:39.238402 138428558899008 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) I0422 17:22:39.238442 138428558899008 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) I0422 17:22:39.238502 138428558899008 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) I0422 17:22:39.238559 138428558899008 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') I0422 17:22:39.238600 138428558899008 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) I0422 17:22:39.238640 138428558899008 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) I0422 17:22:39.238690 138428558899008 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0422 17:22:39.238737 138428558899008 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0422 17:22:39.743564 138428558899008 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 17:22:39.743658 138428558899008 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 17:22:39.758885 138428558899008 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 17:22:39.758947 138428558899008 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 17:22:50.499462 138428558899008 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. I0422 17:22:50.500369 138428558899008 metric_logger.py:289] number parameters: 1.104 billion I0422 17:23:01.564144 138428558899008 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 17:23:01.761770 138428558899008 checkpointing.py:798] Waited 0.1976027488708496 seconds for step 0 to finish before starting checkpointing. I0422 17:23:01.764327 138428558899008 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:23:01.766028 138428558899008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 17:23:01.768733 138428558899008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0. I0422 17:23:02.364077 138428558899008 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 17:23:02.365082 138428558899008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:23:02.365154 138428558899008 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 I0422 17:23:02.619660 138428558899008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.255663s I0422 17:23:02.619829 138428558899008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.724 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5662133693695068 s) (per-host) I0422 17:23:02.619880 138428558899008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.566273s (batch_requests_ready=0.306979s, total_serialization_initiated=0.259227s, others=0.000067s) I0422 17:23:02.619967 138428558899008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.570418s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.570399) I0422 17:23:02.620962 138306327144192 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:23:02.621103 138428558899008 async_checkpointer.py:561] Finished blocking save. Time taken: 0.854994s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0. I0422 17:23:02.623064 138428558899008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:23:02.623311 138305219831552 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:23:02.623465 138428558899008 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878581.7643027, 'wait_for_prev_duration_secs': 7.009506225585938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776878581.7660658, 'checkpointer_blocking_duration_secs': 0.8552379608154297, 'get_old_steps_start_time': 1776878582.6213253, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776878581.7623315, 'checkpoint_manager_blocking_duration_secs': 0.8610961437225342} I0422 17:23:02.623567 138428558899008 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 17:23:02.623617 138428558899008 max_utils.py:750] Memstats: After params initialized: I0422 17:23:02.623664 138428558899008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 17:23:02.623696 138428558899008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 17:23:02.623722 138428558899008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 17:23:02.623746 138428558899008 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 17:23:02.889797 138428558899008 metric_logger.py:185] completed step: 0, seconds: 11.064, TFLOP/s/device: 1.228, Tokens/s/device: 185.111, total_weights: 65536, loss: 10.874 I0422 17:23:03.068161 138428558899008 metric_logger.py:185] completed step: 1, seconds: 1.324, TFLOP/s/device: 10.259, Tokens/s/device: 1546.385, total_weights: 65536, loss: 10.874 I0422 17:23:03.513814 138428558899008 metric_logger.py:185] completed step: 2, seconds: 0.031, TFLOP/s/device: 435.234, Tokens/s/device: 65603.178, total_weights: 65536, loss: 10.560 I0422 17:23:03.661514 138428558899008 metric_logger.py:185] completed step: 3, seconds: 0.446, TFLOP/s/device: 30.480, Tokens/s/device: 4594.328, total_weights: 65536, loss: 9.980 I0422 17:23:03.665939 138428558899008 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 17:23:03.956029 138428558899008 checkpointing.py:798] Waited 0.2900552749633789 seconds for step 5 to finish before starting checkpointing. I0422 17:23:03.958966 138428558899008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:23:10.371322 2555 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 17:23:12.335130 138306335536896 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0422 17:23:48.832430 138306327144192 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 33.767 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.77876567840576 s) (per-host) I0422 17:23:48.832552 138306327144192 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.211354s. I0422 17:23:58.316591 138306327144192 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 55.695378s. I0422 17:23:58.316917 138305219831552 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:23:58.317045 138305219831552 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:23:58.317121 138305219831552 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:23:58.319569 138305219831552 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 17:23:58.319753 138428558899008 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. W0422 17:23:58.319880 138428558899008 checkpoint_manager.py:1458] Waiting for previous save to complete took 54.360917 seconds. If this number is high, consider checkpointing less frequently. I0422 17:23:58.321715 138428558899008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0422 17:23:58.325791 138428558899008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5. I0422 17:23:58.856377 138428558899008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:23:58.856475 138428558899008 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 I0422 17:23:58.901875 138428558899008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046729s I0422 17:23:58.902027 138428558899008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.158 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2990725040435791 s) (per-host) I0422 17:23:58.902078 138428558899008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299132s (batch_requests_ready=0.250533s, total_serialization_initiated=0.048534s, others=0.000066s) I0422 17:23:58.902172 138428558899008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303926s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.303911) I0422 17:23:58.903034 138306335536896 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:23:58.903173 138428558899008 async_checkpointer.py:561] Finished blocking save. Time taken: 0.581387s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5. I0422 17:23:59.342157 138428558899008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:23:59.342565 138305219831552 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:23:59.342741 138428558899008 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878583.9589386, 'wait_for_prev_duration_secs': 54.36091661453247, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776878638.3217537, 'checkpointer_blocking_duration_secs': 0.5815927982330322, 'get_old_steps_start_time': 1776878638.9033685, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776878583.956653, 'checkpoint_manager_blocking_duration_secs': 55.38605046272278} I0422 17:23:59.342854 138428558899008 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 17:23:59.343460 138428558899008 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.970, Tokens/s/device: 13410.601, total_weights: 65536, loss: 9.460 I0422 17:23:59.349210 138428558899008 metric_logger.py:185] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.039, Tokens/s/device: 13873.083, total_weights: 65536, loss: 8.959 I0422 17:24:18.784665 138428558899008 metric_logger.py:185] completed step: 6, seconds: 55.682, TFLOP/s/device: 0.244, Tokens/s/device: 36.780, total_weights: 65536, loss: 8.469 I0422 17:24:18.931991 138428558899008 metric_logger.py:185] completed step: 7, seconds: 19.289, TFLOP/s/device: 0.704, Tokens/s/device: 106.175, total_weights: 65536, loss: 8.003 I0422 17:24:19.079328 138428558899008 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.190, Tokens/s/device: 13443.701, total_weights: 65536, loss: 7.572 I0422 17:24:19.226376 138428558899008 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 17:24:19.227021 138428558899008 checkpointing.py:798] Waited 0.0006670951843261719 seconds for step 9 to finish before starting checkpointing. I0422 17:24:19.229007 138428558899008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:24:22.819397 138305749186304 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0422 17:24:58.546893 138306335536896 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.351 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 59.943899154663086 s) (per-host) I0422 17:24:58.547008 138306335536896 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.643756s. I0422 17:25:07.436266 138306335536896 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.532999s. I0422 17:25:07.436556 138305219831552 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:25:07.436697 138305219831552 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:25:07.436747 138305219831552 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:25:07.438122 138305219831552 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 17:25:07.438308 138428558899008 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. W0422 17:25:07.438445 138428558899008 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.209452 seconds. If this number is high, consider checkpointing less frequently. I0422 17:25:07.440014 138428558899008 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 17:25:07.443498 138428558899008 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9. I0422 17:25:08.425218 138428558899008 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:25:08.425318 138428558899008 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 I0422 17:25:08.469662 138428558899008 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045496s I0422 17:25:08.469828 138428558899008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.012 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3077969551086426 s) (per-host) I0422 17:25:08.469879 138428558899008 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307857s (batch_requests_ready=0.260315s, total_serialization_initiated=0.047477s, others=0.000065s) I0422 17:25:08.469967 138428558899008 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312049s (all_items=0.000010s, per_item={'items': '0.00001001'}, temp_paths=0.312039) I0422 17:25:08.470863 138305757579008 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:25:08.470991 138428558899008 async_checkpointer.py:561] Finished blocking save. Time taken: 1.030907s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9. I0422 17:25:08.474662 138428558899008 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:25:08.474886 138305219831552 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:25:08.475006 138428558899008 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878659.228978, 'wait_for_prev_duration_secs': 48.2094521522522, 'time_between_consecutive_saves_sec': 20.909366369247437, 'checkpointer_blocking_start_time': 1776878707.440054, 'checkpointer_blocking_duration_secs': 1.0311172008514404, 'get_old_steps_start_time': 1776878708.4711854, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1776878659.227257, 'checkpoint_manager_blocking_duration_secs': 49.24771976470947} I0422 17:25:08.475119 138428558899008 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 17:25:08.475161 138428558899008 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:25:13.268382 138306327144192 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_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0422 17:25:49.723699 138305757579008 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.006 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.56162095069885 s) (per-host) I0422 17:25:49.723822 138305757579008 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.252754s. I0422 17:25:56.796453 138305757579008 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.325369s. I0422 17:25:56.796744 138305219831552 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:25:56.796863 138305219831552 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:25:56.796905 138305219831552 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:25:56.798944 138305219831552 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 17:25:56.799194 138428558899008 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. I0422 17:25:56.799359 138428558899008 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:25:56.800024 138428558899008 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.254, Tokens/s/device: 13905.581, 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: Wed Apr 22 17:26:06 UTC 2026 EXIT_CODE=0