XPK Start: Sun Apr 19 06:19:46 UTC 2026 2026-04-19 06:19:50.771666: 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 00:00:1776579590.784674 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776579590.788405 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776579590.799903 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579590.799925 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579590.799927 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579590.799929 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 06:20:10.581749: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0419 06:20:11.110556 134495522674496 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 06:20:20,151:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 06:20:20.151274 134495522674496 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 06:20:20,153:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-5jzdq-slice-job-0-0.mt-16-zarr3-true-rs-5jzdq:8482 I0419 06:20:20.153680 134495522674496 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-5jzdq-slice-job-0-0.mt-16-zarr3-true-rs-5jzdq:8482 I0419 06:20:21.208975 134495522674496 max_utils.py:284] Jax distributed system initialized! I0419 06:20:27.026752 134495522674496 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 06:20:27.026859 134495522674496 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 06:20:27.026901 134495522674496 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0419 06:20:27.026936 134495522674496 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 06:20:27.645065 134495522674496 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 06:20:27.763492 134495522674496 checkpointing.py:688] Setting up checkpoint logger... I0419 06:20:27.763610 134495522674496 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 06:20:27.763660 134495522674496 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 06:20:27.763862 134495522674496 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 0x7a5227779280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 06:20:29.475681 134495522674496 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 06:20:29.476023 134495522674496 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 0x7a3fb6155b80>}, handler_registry=None I0419 06:20:29.476279 134495522674496 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3fb6155b80>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0419 06:20:29.476337 134495522674496 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3fb48e0380>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0419 06:20:29.476375 134495522674496 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 0x7a3fb6155b80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3fb6155b80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3fb48e0380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3fb48e0380>}). I0419 06:20:29.476714 134495522674496 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 06:20:29.476792 134495522674496 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a3fb4855260> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 06:20:31.466965 134376289457920 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776579412008150123, 'commit_timestamp_nsecs': 1776579460285160509, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA I0419 06:20:31.717598 134376289457920 checkpoint_manager.py:1877] Missing metrics for step 5 I0419 06:20:32.891301 134495522674496 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints I0419 06:20:32.964744 134495522674496 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a3fb48a5a60> I0419 06:20:32.964920 134495522674496 checkpointing.py:302] Checkpoint manager created! I0419 06:20:33.903140 134495522674496 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 06:20:33.903273 134495522674496 checkpointing.py:582] restoring from this run's directory step 9 I0419 06:20:34.399656 134495522674496 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/9. I0419 06:20:36.670541 1399 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 06:20:43.469654 134495522674496 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 232.238 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.8045690059661865 s) (per-host) I0419 06:20:43.477896 134495522674496 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 219.613 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 7.206334829330444 s) (per-host) I0419 06:20:45.255761 134495522674496 checkpointer.py:318] Finished restoring checkpoint in 11.10 seconds from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/9. I0419 06:20:45.255930 134495522674496 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776579634.1567748, 'checkpointer_duration_secs': 11.099126100540161, 'checkpoint_manager_start_time': 1776579633.9039178, 'checkpoint_manager_duration_secs': 11.351984977722168} fsdp: 32 I0419 06:20:45.338964 134495522674496 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 06:20:45.339082 134495522674496 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:20:45.339136 134495522674496 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:20:45.339181 134495522674496 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 06:20:45.339219 134495522674496 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:20:45.339269 134495522674496 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:20:45.339319 134495522674496 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:20:45.339359 134495522674496 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 06:20:45.339396 134495522674496 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:20:45.339431 134495522674496 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:20:45.339464 134495522674496 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 06:20:45.339497 134495522674496 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 06:20:45.744663 134495522674496 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:20:45.744810 134495522674496 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:20:45.750268 134495522674496 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:20:45.750386 134495522674496 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:20:45.903971 134495522674496 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:20:45.904107 134495522674496 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:20:46.060162 134495522674496 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:20:46.060311 134495522674496 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:20:46.224534 134495522674496 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:20:46.224678 134495522674496 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:20:46.376940 134495522674496 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:20:46.377068 134495522674496 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:20:46.528421 134495522674496 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 06:20:46.528557 134495522674496 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:20:46.691939 134495522674496 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 06:20:46.692073 134495522674496 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:20:46.715145 134495522674496 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 06:20:46.715253 134495522674496 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 06:20:57.917775 134495522674496 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0419 06:20:57.920473 134495522674496 metric_logger.py:289] number parameters: 1.104 billion I0419 06:21:09.103682 134495522674496 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0419 06:21:09.254981 134495522674496 checkpointing.py:798] Waited 0.15128445625305176 seconds for step 10 to finish before starting checkpointing. I0419 06:21:09.257332 134495522674496 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:21:09.259395 134495522674496 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0419 06:21:09.262058 134495522674496 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/10. I0419 06:21:10.237029 134495522674496 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 06:21:10.238272 134495522674496 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:21:10.238333 134495522674496 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 06:21:10.481058 134495522674496 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245215s I0419 06:21:10.481225 134495522674496 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.119 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4946482181549072 s) (per-host) I0419 06:21:10.481299 134495522674496 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.494733s (batch_requests_ready=0.244615s, total_serialization_initiated=0.250025s, others=0.000092s) I0419 06:21:10.481398 134495522674496 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.498765s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=0.498744) I0419 06:21:10.482278 134373009516288 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:21:10.482408 134495522674496 async_checkpointer.py:561] Finished blocking save. Time taken: 1.222942s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/10. I0419 06:21:10.923495 134495522674496 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:21:10.923896 134371893810944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:21:10.924062 134495522674496 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579669.257314, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776579669.259434, 'checkpointer_blocking_duration_secs': 1.2231614589691162, 'get_old_steps_start_time': 1776579670.4826155, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776579669.2553864, 'checkpoint_manager_blocking_duration_secs': 1.6686375141143799} I0419 06:21:10.924180 134495522674496 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0419 06:21:10.924274 134495522674496 max_utils.py:750] Memstats: After params initialized: I0419 06:21:10.924332 134495522674496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 06:21:10.924367 134495522674496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 06:21:10.924393 134495522674496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 06:21:10.924419 134495522674496 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 06:21:11.196163 134495522674496 metric_logger.py:185] completed step: 10, seconds: 11.182, TFLOP/s/device: 1.215, Tokens/s/device: 183.149, total_weights: 65536, loss: 6.827 I0419 06:21:11.366439 134495522674496 metric_logger.py:185] completed step: 11, seconds: 2.091, TFLOP/s/device: 6.499, Tokens/s/device: 979.570, total_weights: 65536, loss: 6.522 I0419 06:21:11.514068 134495522674496 metric_logger.py:185] completed step: 12, seconds: 0.015, TFLOP/s/device: 900.466, Tokens/s/device: 135728.014, total_weights: 65536, loss: 6.260 I0419 06:21:11.661881 134495522674496 metric_logger.py:185] completed step: 13, seconds: 0.164, TFLOP/s/device: 82.742, Tokens/s/device: 12471.759, total_weights: 65536, loss: 6.039 I0419 06:21:11.668384 134495522674496 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0419 06:21:11.956860 134495522674496 checkpointing.py:798] Waited 0.2884492874145508 seconds for step 15 to finish before starting checkpointing. I0419 06:21:11.959520 134495522674496 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:21:16.103389 134371910596352 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0419 06:21:53.381209 134373009516288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.401 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.39459586143494 s) (per-host) I0419 06:21:53.381339 134373009516288 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.898853s. I0419 06:21:59.912651 134373009516288 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.430149s. I0419 06:21:59.912967 134371893810944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:21:59.913092 134371893810944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:21:59.913140 134371893810944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:21:59.915616 134371893810944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0419 06:21:59.915829 134495522674496 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0419 06:21:59.915965 134495522674496 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.956449 seconds. If this number is high, consider checkpointing less frequently. I0419 06:21:59.917890 134495522674496 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0419 06:21:59.922015 134495522674496 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/15. I0419 06:22:01.322480 134495522674496 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:22:01.322577 134495522674496 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 06:22:01.366330 134495522674496 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045922s I0419 06:22:01.366501 134495522674496 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.093 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7368636131286621 s) (per-host) I0419 06:22:01.366551 134495522674496 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.736923s (batch_requests_ready=0.686214s, total_serialization_initiated=0.050641s, others=0.000068s) I0419 06:22:01.366645 134495522674496 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.741828s (all_items=0.000013s, per_item={'items': '0.00001311'}, temp_paths=0.741815) I0419 06:22:01.367497 134369754720000 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:22:01.367669 134495522674496 async_checkpointer.py:561] Finished blocking save. Time taken: 1.449708s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/15. I0419 06:22:01.818001 134495522674496 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:22:01.818427 134371893810944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:22:01.818604 134495522674496 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579671.9594924, 'wait_for_prev_duration_secs': 47.956448793411255, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776579719.917929, 'checkpointer_blocking_duration_secs': 1.4498474597930908, 'get_old_steps_start_time': 1776579721.3677945, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776579671.957255, 'checkpoint_manager_blocking_duration_secs': 49.861311197280884} I0419 06:22:01.818722 134495522674496 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0419 06:22:01.819508 134495522674496 metric_logger.py:185] completed step: 14, seconds: 0.146, TFLOP/s/device: 92.784, Tokens/s/device: 13985.441, total_weights: 65536, loss: 5.858 I0419 06:22:01.827342 134495522674496 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.836, Tokens/s/device: 13842.514, total_weights: 65536, loss: 5.713 I0419 06:22:20.074114 134495522674496 metric_logger.py:185] completed step: 16, seconds: 50.157, TFLOP/s/device: 0.271, Tokens/s/device: 40.831, total_weights: 65536, loss: 5.599 I0419 06:22:20.221893 134495522674496 metric_logger.py:185] completed step: 17, seconds: 18.102, TFLOP/s/device: 0.751, Tokens/s/device: 113.138, total_weights: 65536, loss: 5.511 I0419 06:22:20.369517 134495522674496 metric_logger.py:185] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.461, Tokens/s/device: 13333.767, total_weights: 65536, loss: 5.442 I0419 06:22:20.518021 134495522674496 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0419 06:22:20.518657 134495522674496 checkpointing.py:798] Waited 0.0006458759307861328 seconds for step 19 to finish before starting checkpointing. I0419 06:22:20.520597 134495522674496 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:22:25.402838 134371910596352 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0419 06:23:01.057814 134369754720000 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.140 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.428136587142944 s) (per-host) I0419 06:23:01.057939 134369754720000 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.690319s. I0419 06:23:10.188587 134369754720000 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.820939s. I0419 06:23:10.188808 134371893810944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:23:10.188876 134371893810944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:23:10.188923 134371893810944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:23:10.191966 134371893810944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0419 06:23:10.192126 134495522674496 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0419 06:23:10.192277 134495522674496 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.671693 seconds. If this number is high, consider checkpointing less frequently. I0419 06:23:10.193825 134495522674496 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0419 06:23:10.197500 134495522674496 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/19. I0419 06:23:11.586098 134495522674496 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:23:11.586197 134495522674496 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0419 06:23:11.628856 134495522674496 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045055s I0419 06:23:11.629023 134495522674496 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.155 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2992095947265625 s) (per-host) I0419 06:23:11.629073 134495522674496 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299268s (batch_requests_ready=0.249557s, total_serialization_initiated=0.049648s, others=0.000063s) I0419 06:23:11.629171 134495522674496 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303305s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.303292) I0419 06:23:11.630024 134375123429120 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:23:11.630180 134495522674496 async_checkpointer.py:561] Finished blocking save. Time taken: 1.436289s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/19. I0419 06:23:12.505936 134495522674496 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:23:12.506289 134371893810944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:23:12.506461 134495522674496 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579740.520568, 'wait_for_prev_duration_secs': 49.67169260978699, 'time_between_consecutive_saves_sec': 20.604912281036377, 'checkpointer_blocking_start_time': 1776579790.193865, 'checkpointer_blocking_duration_secs': 1.4364371299743652, 'get_old_steps_start_time': 1776579791.6303158, 'get_old_steps_duration_secs': 2.1696090698242188e-05, 'checkpoint_manager_blocking_start_time': 1776579740.5188944, 'checkpoint_manager_blocking_duration_secs': 51.98753213882446} I0419 06:23:12.506592 134495522674496 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0419 06:23:12.506636 134495522674496 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:23:19.366750 134375685478144 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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0419 06:23:54.606158 134375123429120 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.500 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.2763032913208 s) (per-host) I0419 06:23:54.606294 134375123429120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.976157s. I0419 06:24:02.911108 134375123429120 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.280957s. I0419 06:24:02.911323 134371893810944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:24:02.911391 134371893810944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:24:02.911436 134371893810944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:24:02.913683 134371893810944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0419 06:24:02.913856 134495522674496 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0419 06:24:02.914029 134495522674496 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:24:02.914773 134495522674496 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.333, Tokens/s/device: 13917.487, total_weights: 65536, loss: 5.385 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 06:24:11 UTC 2026 EXIT_CODE=0