XPK Start: Sat Apr 25 17:36:09 UTC 2026 2026-04-25 17:36:13.258650: 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 00:00:1777138573.271406 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777138573.275279 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777138573.286896 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138573.286914 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138573.286917 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777138573.286919 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 17:36:32.461136: 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) I0425 17:36:32.990272 133283523098432 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 17:36:42,031:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 17:36:42.031922 133283523098432 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 17:36:42,034:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-sv-nplg3-slice-job-0-0.mt-15-ocdbt-false-sv-nplg3:8482 I0425 17:36:42.034279 133283523098432 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-nplg3-slice-job-0-0.mt-15-ocdbt-false-sv-nplg3:8482 I0425 17:36:43.019597 133283523098432 max_utils.py:284] Jax distributed system initialized! I0425 17:36:49.993511 133283523098432 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 17:36:49.993618 133283523098432 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 17:36:49.993660 133283523098432 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 I0425 17:36:49.993696 133283523098432 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 17:36:50.610895 133283523098432 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 17:36:50.725772 133283523098432 checkpointing.py:688] Setting up checkpoint logger... I0425 17:36:50.725891 133283523098432 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0425 17:36:50.725935 133283523098432 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 17:36:50.726199 133283523098432 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7937fe375100>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 17:36:53.665980 133283523098432 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 17:36:53.666347 133283523098432 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 0x792358737c80>}, handler_registry=None I0425 17:36:53.666588 133283523098432 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x792358737c80>` 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`. I0425 17:36:53.666637 133283523098432 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7923587d0170>` 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`. I0425 17:36:53.666672 133283523098432 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 0x792358737c80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x792358737c80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7923587d0170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7923587d0170>}). I0425 17:36:53.667010 133283523098432 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 17:36:53.667087 133283523098432 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x792358764f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 17:36:54.770808 133283523098432 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints I0425 17:36:55.163446 133283523098432 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7923587b6f90> I0425 17:36:55.163643 133283523098432 checkpointing.py:302] Checkpoint manager created! I0425 17:36:56.099637 133283523098432 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 17:36:56.099759 133283523098432 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 17:36:57.526879 133283523098432 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 17:36:57.527006 133283523098432 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:36:57.527058 133283523098432 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 17:36:57.527116 133283523098432 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 17:36:57.527161 133283523098432 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:36:57.527200 133283523098432 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 17:36:57.527239 133283523098432 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:36:57.527278 133283523098432 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 17:36:57.527320 133283523098432 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:36:57.527356 133283523098432 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 17:36:57.527390 133283523098432 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 17:36:57.527423 133283523098432 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 17:36:57.922659 133283523098432 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:36:57.922812 133283523098432 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:36:57.928211 133283523098432 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 17:36:57.928315 133283523098432 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:36:58.074558 133283523098432 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:36:58.074691 133283523098432 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:36:58.217866 133283523098432 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 17:36:58.218000 133283523098432 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 17:36:58.372172 133283523098432 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:36:58.372310 133283523098432 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:36:58.512890 133283523098432 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:36:58.513024 133283523098432 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:36:58.650711 133283523098432 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 17:36:58.650848 133283523098432 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:36:58.799284 133283523098432 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 17:36:58.799432 133283523098432 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 17:36:58.823595 133283523098432 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 17:36:58.823692 133283523098432 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 17:37:09.964169 133283523098432 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. I0425 17:37:09.966850 133283523098432 metric_logger.py:289] number parameters: 1.104 billion I0425 17:37:21.294171 133283523098432 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 17:37:21.444801 133283523098432 checkpointing.py:798] Waited 0.15061402320861816 seconds for step 0 to finish before starting checkpointing. I0425 17:37:21.447631 133283523098432 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:37:21.449202 133283523098432 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 17:37:21.451970 133283523098432 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/0. I0425 17:37:22.488242 133283523098432 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 17:37:22.489329 133283523098432 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:37:22.489391 133283523098432 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 I0425 17:37:22.741776 133283523098432 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254689s I0425 17:37:22.741952 133283523098432 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.956 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.521796703338623 s) (per-host) I0425 17:37:22.742019 133283523098432 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.521876s (batch_requests_ready=0.260540s, total_serialization_initiated=0.261249s, others=0.000087s) I0425 17:37:22.742161 133283523098432 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.525751s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.525731) I0425 17:37:22.743203 133157533771520 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:37:22.743382 133283523098432 async_checkpointer.py:561] Finished blocking save. Time taken: 1.294108s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/0. I0425 17:37:23.147151 133283523098432 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:37:23.147544 133160729818880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:37:23.147712 133283523098432 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138641.4476016, 'wait_for_prev_duration_secs': 0.0001239776611328125, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138641.44924, 'checkpointer_blocking_duration_secs': 1.2942488193511963, 'get_old_steps_start_time': 1777138642.7435093, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777138641.445355, 'checkpoint_manager_blocking_duration_secs': 1.7023143768310547} I0425 17:37:23.147823 133283523098432 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 17:37:23.147900 133283523098432 max_utils.py:750] Memstats: After params initialized: I0425 17:37:23.147959 133283523098432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 17:37:23.147995 133283523098432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 17:37:23.148024 133283523098432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 17:37:23.148049 133283523098432 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 17:37:23.421984 133283523098432 metric_logger.py:185] completed step: 0, seconds: 11.326, TFLOP/s/device: 1.200, Tokens/s/device: 180.819, total_weights: 65536, loss: 10.874 I0425 17:37:23.586571 133283523098432 metric_logger.py:185] completed step: 1, seconds: 2.126, TFLOP/s/device: 6.391, Tokens/s/device: 963.347, total_weights: 65536, loss: 10.874 I0425 17:37:24.002091 133283523098432 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 723.913, Tokens/s/device: 109116.096, total_weights: 65536, loss: 10.560 I0425 17:37:24.149512 133283523098432 metric_logger.py:185] completed step: 3, seconds: 0.407, TFLOP/s/device: 33.407, Tokens/s/device: 5035.529, total_weights: 65536, loss: 9.979 I0425 17:37:24.155648 133283523098432 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 17:37:24.444501 133283523098432 checkpointing.py:798] Waited 0.28882527351379395 seconds for step 5 to finish before starting checkpointing. I0425 17:37:24.447580 133283523098432 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:37:25.635968 2634 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 17:37:27.545174 133160721426176 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6 I0425 17:37:32.719323 133157533771520 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 150.450 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.499127626419067 s) (per-host) I0425 17:37:32.719444 133157533771520 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 9.976129s. I0425 17:37:40.696393 133157533771520 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 17.953062s. I0425 17:37:40.696685 133160729818880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:37:40.696811 133160729818880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:37:40.696863 133160729818880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:37:40.698202 133160729818880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 17:37:40.698386 133283523098432 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. W0425 17:37:40.698514 133283523098432 checkpoint_manager.py:1458] Waiting for previous save to complete took 16.250940 seconds. If this number is high, consider checkpointing less frequently. I0425 17:37:40.700664 133283523098432 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 17:37:40.704718 133283523098432 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/5. I0425 17:37:41.254591 133283523098432 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:37:41.254744 133283523098432 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 I0425 17:37:41.298660 133283523098432 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046080s I0425 17:37:41.298822 133283523098432 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.856 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3176710605621338 s) (per-host) I0425 17:37:41.298871 133283523098432 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.317728s (batch_requests_ready=0.266390s, total_serialization_initiated=0.051273s, others=0.000065s) I0425 17:37:41.298962 133283523098432 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.322392s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.322376) I0425 17:37:41.299826 133151633528576 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:37:41.299985 133283523098432 async_checkpointer.py:561] Finished blocking save. Time taken: 0.599248s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/5. I0425 17:37:41.717158 133283523098432 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:37:41.717531 133160729818880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:37:41.717679 133283523098432 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138644.4475505, 'wait_for_prev_duration_secs': 16.250940084457397, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777138660.7007039, 'checkpointer_blocking_duration_secs': 0.5993835926055908, 'get_old_steps_start_time': 1777138661.300122, 'get_old_steps_duration_secs': 2.7894973754882812e-05, 'checkpoint_manager_blocking_start_time': 1777138644.4448578, 'checkpoint_manager_blocking_duration_secs': 17.272785425186157} I0425 17:37:41.717854 133283523098432 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 17:37:41.718755 133283523098432 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.479, Tokens/s/device: 12582.867, total_weights: 65536, loss: 9.457 I0425 17:37:41.726062 133283523098432 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.266, Tokens/s/device: 13907.280, total_weights: 65536, loss: 8.955 I0425 17:38:01.477593 133283523098432 metric_logger.py:185] completed step: 6, seconds: 17.569, TFLOP/s/device: 0.773, Tokens/s/device: 116.570, total_weights: 65536, loss: 8.465 I0425 17:38:01.624963 133283523098432 metric_logger.py:185] completed step: 7, seconds: 19.606, TFLOP/s/device: 0.693, Tokens/s/device: 104.457, total_weights: 65536, loss: 7.998 I0425 17:38:01.772405 133283523098432 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.458, Tokens/s/device: 13333.420, total_weights: 65536, loss: 7.567 I0425 17:38:01.920842 133283523098432 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 17:38:01.921534 133283523098432 checkpointing.py:798] Waited 0.0007016658782958984 seconds for step 9 to finish before starting checkpointing. I0425 17:38:01.923914 133283523098432 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:38:04.809567 133160721426176 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6 I0425 17:38:10.633116 133151633528576 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 53.271 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 29.651895761489868 s) (per-host) I0425 17:38:10.633238 133151633528576 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.333301s. I0425 17:38:16.764460 133151633528576 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 35.464506s. I0425 17:38:16.764760 133160729818880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:38:16.764877 133160729818880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:38:16.764928 133160729818880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:38:16.766313 133160729818880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 17:38:16.766498 133283523098432 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. W0425 17:38:16.766626 133283523098432 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.842723 seconds. If this number is high, consider checkpointing less frequently. I0425 17:38:16.768309 133283523098432 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 17:38:16.771796 133283523098432 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9. I0425 17:38:17.310986 133283523098432 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 17:38:17.311091 133283523098432 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 I0425 17:38:17.354996 133283523098432 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046083s I0425 17:38:17.355183 133283523098432 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.940 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3122427463531494 s) (per-host) I0425 17:38:17.355235 133283523098432 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.312304s (batch_requests_ready=0.261342s, total_serialization_initiated=0.050895s, others=0.000067s) I0425 17:38:17.355332 133283523098432 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.316345s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.316333) I0425 17:38:17.356226 133153238804224 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 17:38:17.356385 133283523098432 async_checkpointer.py:561] Finished blocking save. Time taken: 0.587999s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9. I0425 17:38:17.784964 133283523098432 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 17:38:17.785353 133160729818880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 17:38:17.785527 133283523098432 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777138681.9238865, 'wait_for_prev_duration_secs': 14.842723369598389, 'time_between_consecutive_saves_sec': 21.22563862800598, 'checkpointer_blocking_start_time': 1777138696.7683551, 'checkpointer_blocking_duration_secs': 0.5881423950195312, 'get_old_steps_start_time': 1777138697.3565116, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1777138681.9217756, 'checkpoint_manager_blocking_duration_secs': 15.86371636390686} I0425 17:38:17.785662 133283523098432 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 17:38:17.785706 133283523098432 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 17:38:23.114836 133160721426176 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6 I0425 17:38:27.691999 133153238804224 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 148.333 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.649022817611694 s) (per-host) I0425 17:38:27.692127 133153238804224 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.335788s. I0425 17:38:34.152519 133153238804224 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.796166s. I0425 17:38:34.152805 133160729818880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 17:38:34.152931 133160729818880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 17:38:34.152973 133160729818880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 17:38:34.154539 133160729818880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 17:38:34.154783 133283523098432 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. I0425 17:38:34.154949 133283523098432 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 17:38:34.155624 133283523098432 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.637, Tokens/s/device: 13963.224, 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: Sat Apr 25 17:38:44 UTC 2026 EXIT_CODE=0