XPK Start: Wed Apr 22 18:57:17 UTC 2026 2026-04-22 18:57:21.627164: 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 00:00:1776884241.640276 9 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776884241.644020 9 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776884241.655602 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776884241.655620 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776884241.655622 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776884241.655624 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 18:57:41.057263: 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 18:57:41.590616 135140206384960 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 18:57:50,632:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 18:57:50.632563 135140206384960 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 18:57:50,635:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-ob9se-slice-job-0-0.mt-16-zarr3-false-sv-ob9se:8482 I0422 18:57:50.635053 135140206384960 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-ob9se-slice-job-0-0.mt-16-zarr3-false-sv-ob9se:8482 I0422 18:57:51.470377 135140206384960 max_utils.py:284] Jax distributed system initialized! I0422 18:57:58.432627 135140206384960 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:57:58.432735 135140206384960 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:57:58.432776 135140206384960 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 18:57:58.432810 135140206384960 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:57:59.050127 135140206384960 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:57:59.163992 135140206384960 checkpointing.py:688] Setting up checkpoint logger... I0422 18:57:59.164149 135140206384960 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0422 18:57:59.164200 135140206384960 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 18:57:59.164405 135140206384960 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 0x7ae841179310>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:58:02.587676 135140206384960 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 18:58:02.588026 135140206384960 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 0x7ad3dc22f260>}, handler_registry=None I0422 18:58:02.588276 135140206384960 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad3dc22f260>` 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 18:58:02.588325 135140206384960 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad3b42d3d70>` 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 18:58:02.588361 135140206384960 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 0x7ad3dc22f260>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad3dc22f260>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad3b42d3d70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad3b42d3d70>}). I0422 18:58:02.588691 135140206384960 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 18:58:02.588765 135140206384960 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ad3b43fcf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:58:03.265934 135140206384960 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints I0422 18:58:03.727672 135140206384960 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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ad3dc4feb70> I0422 18:58:03.727846 135140206384960 checkpointing.py:302] Checkpoint manager created! I0422 18:58:04.686929 135140206384960 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:58:04.687045 135140206384960 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 18:58:06.137718 135140206384960 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 18:58:06.137840 135140206384960 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:58:06.137894 135140206384960 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 18:58:06.137939 135140206384960 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 18:58:06.137977 135140206384960 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:58:06.138011 135140206384960 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 18:58:06.138050 135140206384960 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:58:06.138087 135140206384960 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 18:58:06.138139 135140206384960 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:58:06.138175 135140206384960 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 18:58:06.138211 135140206384960 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 18:58:06.138244 135140206384960 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 18:58:06.533168 135140206384960 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:58:06.533321 135140206384960 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:58:06.538676 135140206384960 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:58:06.538774 135140206384960 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:58:06.689978 135140206384960 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:58:06.690133 135140206384960 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:58:06.839187 135140206384960 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:58:06.839329 135140206384960 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 18:58:06.997201 135140206384960 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 18:58:06.997340 135140206384960 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:58:07.145146 135140206384960 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 18:58:07.145282 135140206384960 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:58:07.290493 135140206384960 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 18:58:07.290626 135140206384960 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:58:07.445732 135140206384960 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:58:07.445872 135140206384960 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 18:58:07.469588 135140206384960 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:58:07.469683 135140206384960 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 18:58:18.179466 135140206384960 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 18:58:18.182139 135140206384960 metric_logger.py:289] number parameters: 1.104 billion I0422 18:58:29.570559 135140206384960 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 18:58:29.723033 135140206384960 checkpointing.py:798] Waited 0.15245914459228516 seconds for step 0 to finish before starting checkpointing. I0422 18:58:29.725659 135140206384960 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:58:29.727519 135140206384960 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 18:58:29.730657 135140206384960 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0. I0422 18:58:30.715130 135140206384960 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 18:58:30.716481 135140206384960 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:58:30.716541 135140206384960 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 18:58:30.963732 135140206384960 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.249740s I0422 18:58:30.963910 135140206384960 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.108 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4962899684906006 s) (per-host) I0422 18:58:30.963965 135140206384960 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.496356s (batch_requests_ready=0.239933s, total_serialization_initiated=0.256349s, others=0.000074s) I0422 18:58:30.964063 135140206384960 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.500744s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.500725) I0422 18:58:30.965139 135018017953536 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:58:30.965312 135140206384960 async_checkpointer.py:561] Finished blocking save. Time taken: 1.237723s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0. I0422 18:58:31.020377 135140206384960 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:58:31.020730 135013773342464 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:58:31.020886 135140206384960 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_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': 1776884309.7256408, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776884309.727558, 'checkpointer_blocking_duration_secs': 1.2378597259521484, 'get_old_steps_start_time': 1776884310.9654396, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1776884309.7234585, 'checkpoint_manager_blocking_duration_secs': 1.2973871231079102} I0422 18:58:31.020995 135140206384960 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 18:58:31.021070 135140206384960 max_utils.py:750] Memstats: After params initialized: I0422 18:58:31.021140 135140206384960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 18:58:31.021177 135140206384960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 18:58:31.021202 135140206384960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 18:58:31.021224 135140206384960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 18:58:31.292338 135140206384960 metric_logger.py:185] completed step: 0, seconds: 11.387, TFLOP/s/device: 1.193, Tokens/s/device: 179.849, total_weights: 65536, loss: 10.874 I0422 18:58:31.455829 135140206384960 metric_logger.py:185] completed step: 1, seconds: 1.720, TFLOP/s/device: 7.900, Tokens/s/device: 1190.721, total_weights: 65536, loss: 10.874 I0422 18:58:31.906645 135140206384960 metric_logger.py:185] completed step: 2, seconds: 0.018, TFLOP/s/device: 762.508, Tokens/s/device: 114933.498, total_weights: 65536, loss: 10.560 I0422 18:58:32.054143 135140206384960 metric_logger.py:185] completed step: 3, seconds: 0.441, TFLOP/s/device: 30.777, Tokens/s/device: 4639.047, total_weights: 65536, loss: 9.979 I0422 18:58:32.060040 135140206384960 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0422 18:58:32.349037 135140206384960 checkpointing.py:798] Waited 0.2889678478240967 seconds for step 5 to finish before starting checkpointing. I0422 18:58:32.351624 135140206384960 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:58:34.079709 2639 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 18:58:36.040363 135018026346240 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0422 18:59:12.365535 135018017953536 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.701 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.89786076545715 s) (per-host) I0422 18:59:12.365650 135018017953536 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.400397s. I0422 18:59:20.215109 135018017953536 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.249824s. I0422 18:59:20.215399 135013773342464 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:59:20.215531 135013773342464 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:59:20.215583 135013773342464 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:59:20.221966 135013773342464 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 18:59:20.222154 135140206384960 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 18:59:20.222281 135140206384960 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.870661 seconds. If this number is high, consider checkpointing less frequently. I0422 18:59:20.224624 135140206384960 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0422 18:59:20.228816 135140206384960 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5. I0422 18:59:20.787397 135140206384960 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 18:59:20.787552 135140206384960 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 18:59:20.832437 135140206384960 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047060s I0422 18:59:20.832608 135140206384960 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.036 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3063333034515381 s) (per-host) I0422 18:59:20.832659 135140206384960 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306393s (batch_requests_ready=0.254132s, total_serialization_initiated=0.052192s, others=0.000068s) I0422 18:59:20.832752 135140206384960 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311314s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.311298) I0422 18:59:20.833652 135018026346240 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 18:59:20.833814 135140206384960 async_checkpointer.py:561] Finished blocking save. Time taken: 0.609071s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5. I0422 18:59:21.274219 135140206384960 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:59:21.274587 135013773342464 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:59:21.274762 135140206384960 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_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': 1776884312.3515964, 'wait_for_prev_duration_secs': 47.87066102027893, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776884360.2247088, 'checkpointer_blocking_duration_secs': 0.6092145442962646, 'get_old_steps_start_time': 1776884360.833944, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1776884312.3494346, 'checkpoint_manager_blocking_duration_secs': 48.925291538238525} I0422 18:59:21.274876 135140206384960 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0422 18:59:21.275683 135140206384960 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.307, Tokens/s/device: 12556.945, total_weights: 65536, loss: 9.457 I0422 18:59:21.283118 135140206384960 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.385, Tokens/s/device: 13925.342, total_weights: 65536, loss: 8.955 I0422 18:59:42.110110 135140206384960 metric_logger.py:185] completed step: 6, seconds: 49.221, TFLOP/s/device: 0.276, Tokens/s/device: 41.608, total_weights: 65536, loss: 8.465 I0422 18:59:42.257440 135140206384960 metric_logger.py:185] completed step: 7, seconds: 20.681, TFLOP/s/device: 0.657, Tokens/s/device: 99.026, total_weights: 65536, loss: 7.998 I0422 18:59:42.404953 135140206384960 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.487, Tokens/s/device: 13337.762, total_weights: 65536, loss: 7.567 I0422 18:59:42.553275 135140206384960 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 18:59:42.553891 135140206384960 checkpointing.py:798] Waited 0.0006284713745117188 seconds for step 9 to finish before starting checkpointing. I0422 18:59:42.556335 135140206384960 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:59:45.950833 135011055417088 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0422 19:00:22.799062 135018026346240 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.366 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 62.27274298667908 s) (per-host) I0422 19:00:22.799190 135018026346240 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.965428s. I0422 19:00:30.469068 135018026346240 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 69.635291s. I0422 19:00:30.469360 135013773342464 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 19:00:30.469498 135013773342464 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 19:00:30.469549 135013773342464 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0422 19:00:30.471085 135013773342464 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0422 19:00:30.471315 135140206384960 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 19:00:30.471446 135140206384960 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.915120 seconds. If this number is high, consider checkpointing less frequently. I0422 19:00:30.472904 135140206384960 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 19:00:30.476579 135140206384960 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9. I0422 19:00:31.403848 135140206384960 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 19:00:31.403949 135140206384960 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 19:00:31.450407 135140206384960 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048642s I0422 19:00:31.450576 135140206384960 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.018 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3074359893798828 s) (per-host) I0422 19:00:31.450624 135140206384960 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307495s (batch_requests_ready=0.254244s, total_serialization_initiated=0.053182s, others=0.000069s) I0422 19:00:31.450729 135140206384960 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.311562s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=0.311552) I0422 19:00:31.451682 135013211293440 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 19:00:31.451813 135140206384960 async_checkpointer.py:561] Finished blocking save. Time taken: 0.978841s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9. I0422 19:00:31.873710 135140206384960 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 19:00:31.874136 135013773342464 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 19:00:31.874307 135140206384960 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_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': 1776884382.5563095, 'wait_for_prev_duration_secs': 47.915120363235474, 'time_between_consecutive_saves_sec': 22.33430004119873, 'checkpointer_blocking_start_time': 1776884430.4729419, 'checkpointer_blocking_duration_secs': 0.9790477752685547, 'get_old_steps_start_time': 1776884431.4520042, 'get_old_steps_duration_secs': 1.8835067749023438e-05, 'checkpoint_manager_blocking_start_time': 1776884382.5541234, 'checkpoint_manager_blocking_duration_secs': 49.320148944854736} I0422 19:00:31.874416 135140206384960 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 19:00:31.874458 135140206384960 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 19:00:36.428163 135018017953536 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0422 19:01:13.159166 135013211293440 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.595 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.016000747680664 s) (per-host) I0422 19:01:13.159266 135013211293440 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.707371s. I0422 19:01:20.273369 135013211293440 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.821458s. I0422 19:01:20.273662 135013773342464 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 19:01:20.273783 135013773342464 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 19:01:20.273823 135013773342464 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 19:01:20.275169 135013773342464 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 19:01:20.275327 135140206384960 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 19:01:20.275500 135140206384960 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 19:01:20.276132 135140206384960 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.400, Tokens/s/device: 13927.520, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 19:01:33 UTC 2026 EXIT_CODE=0