XPK Start: Sun Apr 19 06:25:15 UTC 2026 2026-04-19 06:25:19.097940: 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 00:00:1776579919.110464 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776579919.114166 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776579919.125357 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579919.125375 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579919.125378 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776579919.125380 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 06:25:38.694715: 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 06:25:39.223948 133695950047040 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 06:25:48,266:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 06:25:48.266123 133695950047040 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 06:25:48,268:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-sejoa-slice-job-0-0.mt-16-zarr3-false-sv-sejoa:8482 I0419 06:25:48.268656 133695950047040 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-sejoa-slice-job-0-0.mt-16-zarr3-false-sv-sejoa:8482 I0419 06:25:49.526720 133695950047040 max_utils.py:284] Jax distributed system initialized! I0419 06:25:56.344845 133695950047040 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 06:25:56.344956 133695950047040 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 06:25:56.344996 133695950047040 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 06:25:56.345031 133695950047040 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 06:25:56.962143 133695950047040 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 06:25:57.077338 133695950047040 checkpointing.py:688] Setting up checkpoint logger... I0419 06:25:57.077458 133695950047040 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0419 06:25:57.077505 133695950047040 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 06:25:57.077714 133695950047040 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7997fa375280>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 06:26:00.800823 133695950047040 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 06:26:00.801069 133695950047040 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 0x79850ca3dee0>}, handler_registry=None I0419 06:26:00.801324 133695950047040 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79850ca3dee0>` 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 06:26:00.801376 133695950047040 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79843f268d70>` 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 06:26:00.801413 133695950047040 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 0x79850ca3dee0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79850ca3dee0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79843f268d70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79843f268d70>}). I0419 06:26:00.801753 133695950047040 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 06:26:00.801832 133695950047040 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79843f200f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 06:26:01.550580 133695950047040 checkpoint_manager.py:1818] Found 0 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_false/checkpoints I0419 06:26:01.971681 133695950047040 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_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79843f1bb320> I0419 06:26:01.971859 133695950047040 checkpointing.py:302] Checkpoint manager created! I0419 06:26:02.896462 133695950047040 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 06:26:02.896581 133695950047040 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 06:26:04.345038 133695950047040 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 06:26:04.345166 133695950047040 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:26:04.345221 133695950047040 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 06:26:04.345288 133695950047040 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 06:26:04.345331 133695950047040 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:26:04.345366 133695950047040 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 06:26:04.345403 133695950047040 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:26:04.345441 133695950047040 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 06:26:04.345476 133695950047040 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:26:04.345512 133695950047040 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 06:26:04.345547 133695950047040 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 06:26:04.345581 133695950047040 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 06:26:04.740486 133695950047040 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:26:04.740635 133695950047040 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:26:04.746010 133695950047040 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 06:26:04.746107 133695950047040 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:26:04.901595 133695950047040 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:26:04.901731 133695950047040 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:26:05.055983 133695950047040 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 06:26:05.056120 133695950047040 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 06:26:05.220368 133695950047040 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 06:26:05.220507 133695950047040 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:26:05.371381 133695950047040 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 06:26:05.371520 133695950047040 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:26:05.525016 133695950047040 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 06:26:05.525154 133695950047040 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:26:05.687655 133695950047040 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 06:26:05.687799 133695950047040 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 06:26:05.711904 133695950047040 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 06:26:05.712004 133695950047040 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 06:26:16.814818 133695950047040 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 06:26:16.817440 133695950047040 metric_logger.py:289] number parameters: 1.104 billion I0419 06:26:27.698106 133695950047040 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 06:26:27.849855 133695950047040 checkpointing.py:798] Waited 0.1517353057861328 seconds for step 0 to finish before starting checkpointing. I0419 06:26:27.852181 133695950047040 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:26:27.853910 133695950047040 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0419 06:26:27.856645 133695950047040 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_false/checkpoints/0. I0419 06:26:28.836768 133695950047040 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 06:26:28.837993 133695950047040 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:26:28.838052 133695950047040 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 06:26:29.084965 133695950047040 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.249309s I0419 06:26:29.085142 133695950047040 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.007 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5129995346069336 s) (per-host) I0419 06:26:29.085194 133695950047040 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.513061s (batch_requests_ready=0.257182s, total_serialization_initiated=0.255810s, others=0.000070s) I0419 06:26:29.085315 133695950047040 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.517246s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=0.517225) I0419 06:26:29.086303 133571989722880 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:26:29.086477 133695950047040 async_checkpointer.py:561] Finished blocking save. Time taken: 1.232492s. 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_false/checkpoints/0. I0419 06:26:29.538470 133695950047040 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:26:29.538847 133567174670080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:26:29.539015 133695950047040 standard_logger.py:34] {'step': 0, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579987.8521636, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776579987.85395, 'checkpointer_blocking_duration_secs': 1.2326381206512451, 'get_old_steps_start_time': 1776579989.0866067, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776579987.850281, 'checkpoint_manager_blocking_duration_secs': 1.6886913776397705} I0419 06:26:29.539128 133695950047040 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 06:26:29.539203 133695950047040 max_utils.py:750] Memstats: After params initialized: I0419 06:26:29.539276 133695950047040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0419 06:26:29.539319 133695950047040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0419 06:26:29.539347 133695950047040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0419 06:26:29.539372 133695950047040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0419 06:26:29.805852 133695950047040 metric_logger.py:185] completed step: 0, seconds: 10.880, TFLOP/s/device: 1.249, Tokens/s/device: 188.242, total_weights: 65536, loss: 10.874 I0419 06:26:29.972311 133695950047040 metric_logger.py:185] completed step: 1, seconds: 2.106, TFLOP/s/device: 6.452, Tokens/s/device: 972.473, total_weights: 65536, loss: 10.874 I0419 06:26:30.807825 133695950047040 metric_logger.py:185] completed step: 2, seconds: 0.020, TFLOP/s/device: 663.110, Tokens/s/device: 99951.196, total_weights: 65536, loss: 10.560 I0419 06:26:30.955296 133695950047040 metric_logger.py:185] completed step: 3, seconds: 0.827, TFLOP/s/device: 16.431, Tokens/s/device: 2476.633, total_weights: 65536, loss: 9.979 I0419 06:26:30.961176 133695950047040 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0419 06:26:31.250507 133695950047040 checkpointing.py:798] Waited 0.28930115699768066 seconds for step 5 to finish before starting checkpointing. I0419 06:26:31.253589 133695950047040 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:26:33.023054 2641 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 06:26:34.837354 133571998115584 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_false/checkpoints/0/items/array_metadatas/process_6 I0419 06:27:11.072832 133571989722880 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.166 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.500629901885986 s) (per-host) I0419 06:27:11.072955 133571989722880 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.986538s. I0419 06:27:20.622830 133571989722880 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 51.536397s. I0419 06:27:20.623120 133567174670080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:27:20.623275 133567174670080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:27:20.623329 133567174670080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:27:20.625267 133567174670080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 06:27:20.625441 133695950047040 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. W0419 06:27:20.625580 133695950047040 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.371995 seconds. If this number is high, consider checkpointing less frequently. I0419 06:27:20.627549 133695950047040 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0419 06:27:20.631893 133695950047040 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_false/checkpoints/5. I0419 06:27:22.476155 133695950047040 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:27:22.476324 133695950047040 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 06:27:22.517263 133695950047040 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043121s I0419 06:27:22.517442 133695950047040 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.070 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7453463077545166 s) (per-host) I0419 06:27:22.517493 133695950047040 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.745406s (batch_requests_ready=0.697074s, total_serialization_initiated=0.048265s, others=0.000067s) I0419 06:27:22.517584 133695950047040 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.750460s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.750443) I0419 06:27:22.518516 133568737920768 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:27:22.518646 133695950047040 async_checkpointer.py:561] Finished blocking save. Time taken: 1.891025s. 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_false/checkpoints/5. I0419 06:27:22.521595 133695950047040 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:27:22.521850 133567174670080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:27:22.522010 133695950047040 standard_logger.py:34] {'step': 5, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776579991.2535613, 'wait_for_prev_duration_secs': 49.371994972229004, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776580040.6275878, 'checkpointer_blocking_duration_secs': 1.8912312984466553, 'get_old_steps_start_time': 1776580042.518838, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776579991.2508612, 'checkpoint_manager_blocking_duration_secs': 51.271113872528076} I0419 06:27:22.522117 133695950047040 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0419 06:27:22.522881 133695950047040 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.604, Tokens/s/device: 12601.681, total_weights: 65536, loss: 9.457 I0419 06:27:22.530444 133695950047040 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.353, Tokens/s/device: 13920.420, total_weights: 65536, loss: 8.955 I0419 06:27:42.191184 133695950047040 metric_logger.py:185] completed step: 6, seconds: 51.567, TFLOP/s/device: 0.263, Tokens/s/device: 39.715, total_weights: 65536, loss: 8.465 I0419 06:27:42.338758 133695950047040 metric_logger.py:185] completed step: 7, seconds: 19.515, TFLOP/s/device: 0.696, Tokens/s/device: 104.942, total_weights: 65536, loss: 7.998 I0419 06:27:42.486383 133695950047040 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.412, Tokens/s/device: 13326.479, total_weights: 65536, loss: 7.567 I0419 06:27:42.634937 133695950047040 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 06:27:42.635562 133695950047040 checkpointing.py:798] Waited 0.000637054443359375 seconds for step 9 to finish before starting checkpointing. I0419 06:27:42.638323 133695950047040 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:27:46.560396 133566610515712 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_false/checkpoints/5/items/array_metadatas/process_6 I0419 06:28:23.014161 133568737920768 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.793 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.24202370643616 s) (per-host) I0419 06:28:23.014292 133568737920768 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.495566s. I0419 06:28:32.730089 133568737920768 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.211349s. I0419 06:28:32.730404 133567174670080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:28:32.730563 133567174670080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:28:32.730615 133567174670080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:28:32.732193 133567174670080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0419 06:28:32.732388 133695950047040 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0419 06:28:32.732522 133695950047040 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.094210 seconds. If this number is high, consider checkpointing less frequently. I0419 06:28:32.734122 133695950047040 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0419 06:28:32.737549 133695950047040 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_false/checkpoints/9. I0419 06:28:34.135189 133695950047040 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 06:28:34.135302 133695950047040 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 06:28:34.173896 133695950047040 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040676s I0419 06:28:34.174074 133695950047040 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.167 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29854607582092285 s) (per-host) I0419 06:28:34.174136 133695950047040 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.298617s (batch_requests_ready=0.253385s, total_serialization_initiated=0.045155s, others=0.000077s) I0419 06:28:34.174269 133695950047040 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.302772s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.302760) I0419 06:28:34.175153 133568207255296 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0419 06:28:34.175288 133695950047040 async_checkpointer.py:561] Finished blocking save. Time taken: 1.441099s. 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_false/checkpoints/9. I0419 06:28:34.630708 133695950047040 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 06:28:34.631082 133567174670080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0419 06:28:34.631255 133695950047040 standard_logger.py:34] {'step': 9, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776580062.6382966, 'wait_for_prev_duration_secs': 50.09420967102051, 'time_between_consecutive_saves_sec': 22.012990474700928, 'checkpointer_blocking_start_time': 1776580112.734162, 'checkpointer_blocking_duration_secs': 1.4412941932678223, 'get_old_steps_start_time': 1776580114.1754692, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1776580062.6357787, 'checkpoint_manager_blocking_duration_secs': 51.9954195022583} I0419 06:28:34.631421 133695950047040 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 06:28:34.631466 133695950047040 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 06:28:40.153608 133571990259456 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_false/checkpoints/9/items/array_metadatas/process_6 I0419 06:29:16.767613 133568207255296 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 36.827 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.8920636177063 s) (per-host) I0419 06:29:16.767712 133568207255296 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.592347s. I0419 06:29:24.437887 133568207255296 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.262506s. I0419 06:29:24.438195 133567174670080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 06:29:24.438338 133567174670080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0419 06:29:24.438389 133567174670080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 06:29:24.440651 133567174670080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 06:29:24.440818 133695950047040 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. I0419 06:29:24.440992 133695950047040 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 06:29:24.441660 133695950047040 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.387, Tokens/s/device: 13925.626, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 06:29:35 UTC 2026 EXIT_CODE=0