XPK Start: Mon Apr 20 21:35:56 UTC 2026 2026-04-20 21:36:01.061924: 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 00:00:1776720961.075220 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776720961.079019 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776720961.090463 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720961.090484 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720961.090486 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776720961.090488 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:36:20.208388: 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 21:36:20.737090 134416556930880 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:36:29,777:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:36:29.777755 134416556930880 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:36:29,780:jax._src.distributed:157: Connecting to JAX distributed service on mt-13-scan-layers-true-apl6s-slice-job-0-0.mt-13-scan-layers-true-apl6s:8482 I0420 21:36:29.780083 134416556930880 distributed.py:157] Connecting to JAX distributed service on mt-13-scan-layers-true-apl6s-slice-job-0-0.mt-13-scan-layers-true-apl6s:8482 I0420 21:36:31.162470 134416556930880 max_utils.py:284] Jax distributed system initialized! I0420 21:36:38.177278 134416556930880 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:36:38.177384 134416556930880 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:36:38.177426 134416556930880 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 21:36:38.177462 134416556930880 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:36:38.794673 134416556930880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:36:38.908871 134416556930880 checkpointing.py:688] Setting up checkpoint logger... I0420 21:36:38.908992 134416556930880 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:36:38.909037 134416556930880 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:36:38.909240 134416556930880 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 0x7a3fcbd79460>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:36:41.809346 134416556930880 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:36:41.809710 134416556930880 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a2b640bd130>}, handler_registry=None I0420 21:36:41.809948 134416556930880 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a2b640bd130>` 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 21:36:41.809997 134416556930880 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2b64273500>` 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 21:36:41.810033 134416556930880 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 0x7a2b640bd130>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a2b640bd130>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2b64273500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a2b64273500>}). I0420 21:36:41.810365 134416556930880 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:36:41.810441 134416556930880 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a2b641b0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:36:42.617832 134416556930880 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_13_scan_layers_true/checkpoints I0420 21:36:42.620028 134416556930880 checkpoint_manager.py:929] [process=7][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a2b64231220> I0420 21:36:42.620141 134416556930880 checkpointing.py:302] Checkpoint manager created! I0420 21:36:43.558097 134416556930880 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:36:43.558217 134416556930880 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 21:36:44.966390 134416556930880 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:36:44.966513 134416556930880 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:36:44.966565 134416556930880 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:36:44.966609 134416556930880 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:36:44.966647 134416556930880 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:36:44.966698 134416556930880 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:36:44.966738 134416556930880 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:36:44.966776 134416556930880 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:36:44.966812 134416556930880 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:36:44.966846 134416556930880 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:36:44.966880 134416556930880 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:36:44.966914 134416556930880 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:36:45.358587 134416556930880 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:36:45.358749 134416556930880 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:36:45.364186 134416556930880 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:36:45.364289 134416556930880 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:36:45.519121 134416556930880 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:36:45.519259 134416556930880 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:36:45.671765 134416556930880 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:36:45.671895 134416556930880 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 21:36:45.835469 134416556930880 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 21:36:45.835601 134416556930880 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:36:45.986265 134416556930880 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 21:36:45.986399 134416556930880 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:36:46.137346 134416556930880 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 21:36:46.137481 134416556930880 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:36:46.298862 134416556930880 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:36:46.299004 134416556930880 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 21:36:46.321760 134416556930880 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:36:46.321857 134416556930880 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 21:36:57.158455 134416556930880 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 21:36:57.160857 134416556930880 metric_logger.py:289] number parameters: 1.104 billion I0420 21:37:08.457408 134416556930880 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 21:37:08.609899 134416556930880 checkpointing.py:798] Waited 0.15247702598571777 seconds for step 0 to finish before starting checkpointing. I0420 21:37:08.612535 134416556930880 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:37:08.614387 134416556930880 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 21:37:08.617306 134416556930880 async_checkpointer.py:452] [process=7] 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_13_scan_layers_true/checkpoints/0. I0420 21:37:09.601206 134416556930880 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:37:09.602309 134416556930880 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:37:09.602368 134416556930880 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 21:37:09.857542 134416556930880 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.257452s I0420 21:37:09.857724 134416556930880 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.864 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5385580062866211 s) (per-host) I0420 21:37:09.857779 134416556930880 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.538623s (batch_requests_ready=0.274584s, total_serialization_initiated=0.263966s, others=0.000072s) I0420 21:37:09.857874 134416556930880 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.542836s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=0.542816) I0420 21:37:09.858896 134294081099520 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:37:09.859025 134416556930880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.244568s. 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_13_scan_layers_true/checkpoints/0. I0420 21:37:10.251991 134416556930880 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:37:10.252379 134290188785408 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:37:10.252552 134416556930880 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_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776721028.6125169, 'wait_for_prev_duration_secs': 6.532669067382812e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776721028.6144245, 'checkpointer_blocking_duration_secs': 1.2447905540466309, 'get_old_steps_start_time': 1776721029.8592353, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776721028.6103098, 'checkpoint_manager_blocking_duration_secs': 1.6421983242034912} I0420 21:37:10.252673 134416556930880 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 21:37:10.252749 134416556930880 max_utils.py:750] Memstats: After params initialized: I0420 21:37:10.252810 134416556930880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 21:37:10.252840 134416556930880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 21:37:10.252864 134416556930880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 21:37:10.252891 134416556930880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 21:37:10.522606 134416556930880 metric_logger.py:185] completed step: 0, seconds: 11.296, TFLOP/s/device: 1.203, Tokens/s/device: 181.311, total_weights: 65536, loss: 10.874 I0420 21:37:10.689817 134416556930880 metric_logger.py:185] completed step: 1, seconds: 2.063, TFLOP/s/device: 6.585, Tokens/s/device: 992.574, total_weights: 65536, loss: 10.874 I0420 21:37:11.093358 134416556930880 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 640.479, Tokens/s/device: 96540.021, total_weights: 65536, loss: 10.262 I0420 21:37:11.240740 134416556930880 metric_logger.py:185] completed step: 3, seconds: 0.394, TFLOP/s/device: 34.455, Tokens/s/device: 5193.422, total_weights: 65536, loss: 9.732 I0420 21:37:11.536410 134416556930880 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.156, Tokens/s/device: 12534.120, total_weights: 65536, loss: 9.274 I0420 21:37:11.543778 134416556930880 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.423, Tokens/s/device: 13930.930, total_weights: 65536, loss: 8.888 I0420 21:37:30.475919 134416556930880 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.887, Tokens/s/device: 6916.652, total_weights: 65536, loss: 8.589 I0420 21:37:30.623335 134416556930880 metric_logger.py:185] completed step: 7, seconds: 18.786, TFLOP/s/device: 0.723, Tokens/s/device: 109.017, total_weights: 65536, loss: 8.380 I0420 21:37:30.770809 134416556930880 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.509, Tokens/s/device: 13340.977, total_weights: 65536, loss: 8.251 I0420 21:37:30.919376 134416556930880 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 21:37:30.919986 134416556930880 checkpointing.py:798] Waited 0.0006208419799804688 seconds for step 9 to finish before starting checkpointing. I0420 21:37:30.921866 134416556930880 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:37:31.350059 2624 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:37:33.230396 134294089492224 array_metadata_store.py:203] [process=7][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_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_7 I0420 21:38:05.532390 134294081099520 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 28.100 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 56.21318459510803 s) (per-host) I0420 21:38:05.532511 134294081099520 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.673402s. I0420 21:38:13.352454 134294081099520 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 63.493330s. I0420 21:38:13.352778 134290188785408 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:38:13.352905 134290188785408 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:38:13.352955 134290188785408 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:38:13.354537 134290188785408 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 21:38:13.354753 134416556930880 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 21:38:13.354900 134416556930880 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.433037 seconds. If this number is high, consider checkpointing less frequently. I0420 21:38:13.357948 134416556930880 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 21:38:13.361477 134416556930880 async_checkpointer.py:452] [process=7] 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_13_scan_layers_true/checkpoints/9. I0420 21:38:14.317003 134416556930880 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:38:14.317162 134416556930880 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 21:38:14.366616 134416556930880 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051648s I0420 21:38:14.366805 134416556930880 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.990 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3091139793395996 s) (per-host) I0420 21:38:14.366857 134416556930880 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.309176s (batch_requests_ready=0.252772s, total_serialization_initiated=0.056334s, others=0.000070s) I0420 21:38:14.366950 134416556930880 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.313231s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.313215) I0420 21:38:14.367839 134290868266752 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:38:14.368005 134416556930880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.009986s. 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_13_scan_layers_true/checkpoints/9. I0420 21:38:14.374639 134416556930880 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:38:14.374969 134290188785408 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:38:14.375141 134416556930880 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_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776721050.921837, 'wait_for_prev_duration_secs': 42.43303728103638, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776721093.3579855, 'checkpointer_blocking_duration_secs': 1.0101358890533447, 'get_old_steps_start_time': 1776721094.3681412, 'get_old_steps_duration_secs': 2.384185791015625e-05, 'checkpoint_manager_blocking_start_time': 1776721050.920223, 'checkpoint_manager_blocking_duration_secs': 43.45488405227661} I0420 21:38:14.375251 134416556930880 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 21:38:14.375293 134416556930880 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:38:19.143233 134294089492224 array_metadata_store.py:203] [process=7][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_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_7 I0420 21:38:56.301642 134290868266752 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.392 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.24391174316406 s) (per-host) I0420 21:38:56.301787 134290868266752 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.933831s. I0420 21:39:03.105177 134290868266752 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 48.737127s. I0420 21:39:03.105487 134290188785408 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:39:03.105606 134290188785408 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:39:03.105668 134290188785408 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:39:03.106923 134290188785408 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 21:39:03.107133 134416556930880 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 21:39:03.107303 134416556930880 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:39:03.108071 134416556930880 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.499, Tokens/s/device: 13942.501, 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 21:39:12 UTC 2026 EXIT_CODE=0