XPK Start: Mon Apr 20 21:27:12 UTC 2026 2026-04-20 21:27:16.239513: 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 00:00:1776720436.252471 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776720436.256397 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776720436.267943 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720436.267964 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720436.267967 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720436.267969 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:27:35.318828: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 21:27:35.844442 140655584429888 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:27:44,886:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:27:44.886203 140655584429888 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:27:44,888:jax._src.distributed:157: Connecting to JAX distributed service on mt-10-shardy-false-h4sgx-slice-job-0-0.mt-10-shardy-false-h4sgx:8482 I0420 21:27:44.888515 140655584429888 distributed.py:157] Connecting to JAX distributed service on mt-10-shardy-false-h4sgx-slice-job-0-0.mt-10-shardy-false-h4sgx:8482 I0420 21:27:46.317173 140655584429888 max_utils.py:284] Jax distributed system initialized! I0420 21:27:53.505311 140655584429888 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:27:53.505421 140655584429888 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:27:53.505492 140655584429888 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 21:27:53.505530 140655584429888 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:27:54.126465 140655584429888 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:27:54.240180 140655584429888 checkpointing.py:688] Setting up checkpoint logger... I0420 21:27:54.240299 140655584429888 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:27:54.240350 140655584429888 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:27:54.240571 140655584429888 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 0x7fec76d50da0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:27:57.167879 140655584429888 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:27:57.168223 140655584429888 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 0x7fd9c8c43590>}, handler_registry=None I0420 21:27:57.168516 140655584429888 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fd9c8c43590>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 21:27:57.168571 140655584429888 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd7fc6c3f80>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 21:27:57.168611 140655584429888 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 0x7fd9c8c43590>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fd9c8c43590>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd7fc6c3f80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd7fc6c3f80>}). I0420 21:27:57.168946 140655584429888 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:27:57.169024 140655584429888 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7fd7fc675260> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:27:58.752143 140655584429888 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints I0420 21:27:58.754504 140655584429888 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=10), 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7fd7fc6bd6d0> I0420 21:27:58.754618 140655584429888 checkpointing.py:302] Checkpoint manager created! I0420 21:27:59.685154 140655584429888 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:27:59.685270 140655584429888 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 21:28:01.099834 140655584429888 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:28:01.099953 140655584429888 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:28:01.100007 140655584429888 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:28:01.100052 140655584429888 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:28:01.100090 140655584429888 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:28:01.100125 140655584429888 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:28:01.100162 140655584429888 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:28:01.100198 140655584429888 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:28:01.100234 140655584429888 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:28:01.100268 140655584429888 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:28:01.100300 140655584429888 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:28:01.100339 140655584429888 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:28:01.494972 140655584429888 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:28:01.495121 140655584429888 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:28:01.500451 140655584429888 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:28:01.500547 140655584429888 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:28:01.648825 140655584429888 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:28:01.648957 140655584429888 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:28:01.797278 140655584429888 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:28:01.797415 140655584429888 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:28:01.951683 140655584429888 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:28:01.951817 140655584429888 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:28:02.097354 140655584429888 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:28:02.097497 140655584429888 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:28:02.242937 140655584429888 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:28:02.243069 140655584429888 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:28:02.398271 140655584429888 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:28:02.398413 140655584429888 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:28:02.420695 140655584429888 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:28:02.420786 140655584429888 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 21:28:13.443026 140655584429888 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 21:28:13.445739 140655584429888 metric_logger.py:289] number parameters: 1.104 billion I0420 21:28:24.533208 140655584429888 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 21:28:24.685506 140655584429888 checkpointing.py:798] Waited 0.15224528312683105 seconds for step 0 to finish before starting checkpointing. I0420 21:28:24.688915 140655584429888 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:28:24.690595 140655584429888 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0420 21:28:24.693250 140655584429888 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/0. I0420 21:28:25.713064 140655584429888 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:28:25.714126 140655584429888 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:28:25.714182 140655584429888 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 21:28:25.957328 140655584429888 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245406s I0420 21:28:25.957512 140655584429888 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.804 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5500659942626953 s) (per-host) I0420 21:28:25.957577 140655584429888 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.550135s (batch_requests_ready=0.298162s, total_serialization_initiated=0.251895s, others=0.000078s) I0420 21:28:25.957675 140655584429888 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.554106s (all_items=0.000021s, per_item={'items': '0.00002050'}, temp_paths=0.554086) I0420 21:28:25.958651 140533519337216 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0420 21:28:25.958819 140655584429888 async_checkpointer.py:561] Finished blocking save. Time taken: 1.268152s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/0. I0420 21:28:25.962395 140655584429888 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:28:25.962668 140530818197248 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:28:25.962789 140655584429888 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776720504.6888976, 'wait_for_prev_duration_secs': 6.413459777832031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776720504.6906333, 'checkpointer_blocking_duration_secs': 1.26829195022583, 'get_old_steps_start_time': 1776720505.9589455, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776720504.6859007, 'checkpoint_manager_blocking_duration_secs': 1.276850938796997} I0420 21:28:25.962898 140655584429888 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 21:28:25.962972 140655584429888 max_utils.py:750] Memstats: After params initialized: I0420 21:28:25.963022 140655584429888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0420 21:28:25.963054 140655584429888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0420 21:28:25.963081 140655584429888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0420 21:28:25.963105 140655584429888 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0420 21:28:26.229787 140655584429888 metric_logger.py:185] completed step: 0, seconds: 11.086, TFLOP/s/device: 1.226, Tokens/s/device: 184.731, total_weights: 65536, loss: 10.874 I0420 21:28:26.395079 140655584429888 metric_logger.py:185] completed step: 1, seconds: 1.695, TFLOP/s/device: 8.017, Tokens/s/device: 1208.440, total_weights: 65536, loss: 10.874 I0420 21:28:26.823326 140655584429888 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 660.531, Tokens/s/device: 99562.470, total_weights: 65536, loss: 10.262 I0420 21:28:26.970822 140655584429888 metric_logger.py:185] completed step: 3, seconds: 0.418, TFLOP/s/device: 32.537, Tokens/s/device: 4904.344, total_weights: 65536, loss: 9.732 I0420 21:28:27.266314 140655584429888 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.107, Tokens/s/device: 12526.837, total_weights: 65536, loss: 9.274 I0420 21:28:27.273499 140655584429888 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.397, Tokens/s/device: 13927.141, total_weights: 65536, loss: 8.888 I0420 21:28:46.908375 140655584429888 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.920, Tokens/s/device: 6921.561, total_weights: 65536, loss: 8.589 I0420 21:28:47.055709 140655584429888 metric_logger.py:185] completed step: 7, seconds: 19.489, TFLOP/s/device: 0.697, Tokens/s/device: 105.086, total_weights: 65536, loss: 8.380 I0420 21:28:47.203065 140655584429888 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.655, Tokens/s/device: 13363.000, total_weights: 65536, loss: 8.251 I0420 21:28:47.351690 140655584429888 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 21:28:47.352302 140655584429888 checkpointing.py:798] Waited 0.0006835460662841797 seconds for step 9 to finish before starting checkpointing. I0420 21:28:47.354207 140655584429888 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:28:48.062627 2608 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:28:50.921506 140533527729920 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/0/items/array_metadatas/process_6 I0420 21:29:21.068809 140533519337216 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 28.379 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.66132116317749 s) (per-host) I0420 21:29:21.068931 140533519337216 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.110168s. I0420 21:29:30.732310 140533519337216 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 64.773531s. I0420 21:29:30.732570 140530818197248 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:29:30.732644 140530818197248 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:29:30.732691 140530818197248 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:29:30.734461 140530818197248 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 21:29:30.734635 140655584429888 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. W0420 21:29:30.734776 140655584429888 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.380573 seconds. If this number is high, consider checkpointing less frequently. I0420 21:29:30.736710 140655584429888 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0420 21:29:30.740074 140655584429888 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/9. I0420 21:29:31.696184 140655584429888 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:29:31.696344 140655584429888 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 21:29:31.723308 140655584429888 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029111s I0420 21:29:31.723509 140655584429888 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.388 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28631591796875 s) (per-host) I0420 21:29:31.723570 140655584429888 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.286387s (batch_requests_ready=0.252516s, total_serialization_initiated=0.033788s, others=0.000083s) I0420 21:29:31.723669 140655584429888 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290470s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.290454) I0420 21:29:31.724545 140530281666304 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0420 21:29:31.724717 140655584429888 async_checkpointer.py:561] Finished blocking save. Time taken: 0.987938s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/9. I0420 21:29:31.754623 140655584429888 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:29:31.754926 140530818197248 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:29:31.755070 140655584429888 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776720527.3541775, 'wait_for_prev_duration_secs': 43.3805730342865, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776720570.7367482, 'checkpointer_blocking_duration_secs': 0.9880754947662354, 'get_old_steps_start_time': 1776720571.7248435, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776720527.352538, 'checkpoint_manager_blocking_duration_secs': 44.4024977684021} I0420 21:29:31.755181 140655584429888 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 21:29:31.755225 140655584429888 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:29:36.979732 140530290059008 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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_10_shardy_false/checkpoints/9/items/array_metadatas/process_6 I0420 21:30:12.856366 140530281666304 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.137 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.41913461685181 s) (per-host) I0420 21:30:12.856526 140530281666304 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.131863s. I0420 21:30:21.526072 140530281666304 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.801392s. I0420 21:30:21.526383 140530818197248 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:30:21.526530 140530818197248 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:30:21.526580 140530818197248 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:30:21.528193 140530818197248 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 21:30:21.528368 140655584429888 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. I0420 21:30:21.528567 140655584429888 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:30:21.529214 140655584429888 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.417, Tokens/s/device: 13930.078, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 21:30:33 UTC 2026 EXIT_CODE=0