XPK Start: Wed Apr 22 17:16:33 UTC 2026 2026-04-22 17:16:37.105379: 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 00:00:1776878197.118688 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776878197.122427 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776878197.133969 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878197.133989 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878197.133992 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776878197.133994 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 17:16:56.419506: 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) I0422 17:16:56.953397 140119134377792 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 17:17:05,994:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 17:17:05.994831 140119134377792 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 17:17:05,997:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-rs-t01y2-slice-job-0-0.mt-16-zarr3-true-rs-t01y2:8482 I0422 17:17:05.997270 140119134377792 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-rs-t01y2-slice-job-0-0.mt-16-zarr3-true-rs-t01y2:8482 I0422 17:17:06.921986 140119134377792 max_utils.py:284] Jax distributed system initialized! I0422 17:17:12.689641 140119134377792 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 17:17:12.689748 140119134377792 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 17:17:12.689790 140119134377792 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 I0422 17:17:12.689828 140119134377792 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 17:17:13.311016 140119134377792 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:17:13.311645 140119134377792 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:17:13.311830 140119134377792 checkpointing.py:688] Setting up checkpoint logger... I0422 17:17:13.311878 140119134377792 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 17:17:13.311922 140119134377792 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 17:17:13.312302 140119134377792 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 0x7f6f85578a10>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 17:17:14.656874 140119134377792 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 17:17:14.657238 140119134377792 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 0x7f5d4d8c4140>}, handler_registry=None I0422 17:17:14.657481 140119134377792 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f5d4d8c4140>` 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`. I0422 17:17:14.657536 140119134377792 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5d4d3e6cf0>` 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`. I0422 17:17:14.657573 140119134377792 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 0x7f5d4d8c4140>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f5d4d8c4140>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5d4d3e6cf0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f5d4d3e6cf0>}). I0422 17:17:14.658081 140119134377792 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 17:17:14.658177 140119134377792 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f5d4cf78180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 17:17:16.522811 140001958418176 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776877956634809673, 'commit_timestamp_nsecs': 1776877998808032499, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0422 17:17:16.770704 140001958418176 checkpoint_manager.py:1877] Missing metrics for step 0 I0422 17:17:16.939348 140119134377792 checkpoint_manager.py:1818] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints I0422 17:17:17.255057 140119134377792 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/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f5d4d8c56d0> I0422 17:17:17.255275 140119134377792 checkpointing.py:302] Checkpoint manager created! I0422 17:17:18.152838 140119134377792 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:17:18.152946 140119134377792 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:17:18.527346 140119134377792 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:17:18.527441 140119134377792 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:17:18.544239 140119134377792 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:17:18.544301 140119134377792 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:17:18.567375 140119134377792 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:17:18.567447 140119134377792 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:17:18.583616 140119134377792 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:17:18.583677 140119134377792 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:17:18.599697 140119134377792 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:17:18.599758 140119134377792 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:17:18.624763 140119134377792 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 17:17:18.624834 140119134377792 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:17:18.644657 140119134377792 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 17:17:18.644726 140119134377792 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 17:17:18.844754 140119134377792 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 17:17:18.844866 140119134377792 checkpointing.py:582] restoring from this run's directory step 9 I0422 17:17:19.364568 140119134377792 checkpointer.py:304] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/9. I0422 17:17:20.774858 1390 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 17:17:26.840151 140119134377792 jax_array_handlers.py:837] [process=6] /jax/orbax/read/worker/io/requested throughput: 260.315 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.070624351501465 s) (per-host) I0422 17:17:26.844850 140119134377792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/read/gbytes_per_sec: 244.515 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.472368240356445 s) (per-host) I0422 17:17:27.516306 140119134377792 checkpointer.py:318] Finished restoring checkpoint in 8.42 seconds from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/9. I0422 17:17:27.516464 140119134377792 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776878239.0930145, 'checkpointer_duration_secs': 8.423418760299683, 'checkpoint_manager_start_time': 1776878238.8451312, 'checkpoint_manager_duration_secs': 8.671303749084473} fsdp: 32 I0422 17:17:27.931631 140119134377792 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 17:17:27.931763 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 17:17:27.931818 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 17:17:27.931879 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0422 17:17:27.931933 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0422 17:17:27.931973 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0422 17:17:27.932027 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0422 17:17:27.932081 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0422 17:17:27.932163 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0422 17:17:27.932211 140119134377792 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0422 17:17:27.932269 140119134377792 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0422 17:17:27.932321 140119134377792 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0422 17:17:28.439645 140119134377792 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 17:17:28.439738 140119134377792 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 17:17:28.454832 140119134377792 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 17:17:28.454894 140119134377792 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 17:17:39.302697 140119134377792 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. I0422 17:17:39.303668 140119134377792 metric_logger.py:289] number parameters: 1.104 billion I0422 17:17:50.316684 140119134377792 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0422 17:17:50.475752 140119134377792 checkpointing.py:798] Waited 0.15905547142028809 seconds for step 10 to finish before starting checkpointing. I0422 17:17:50.478382 140119134377792 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:17:50.480183 140119134377792 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0422 17:17:50.483263 140119134377792 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/10. I0422 17:17:51.079281 140119134377792 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 17:17:51.080237 140119134377792 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:17:51.080296 140119134377792 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 I0422 17:17:51.327105 140119134377792 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.247916s I0422 17:17:51.327278 140119134377792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.039 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5075781345367432 s) (per-host) I0422 17:17:51.327329 140119134377792 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.507640s (batch_requests_ready=0.257516s, total_serialization_initiated=0.250055s, others=0.000069s) I0422 17:17:51.327419 140119134377792 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.511796s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.511778) I0422 17:17:51.328296 140001429939968 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:17:51.328471 140119134377792 async_checkpointer.py:561] Finished blocking save. Time taken: 0.848221s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/10. I0422 17:17:51.399417 140119134377792 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:17:51.399813 139997143365376 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:17:51.399977 140119134377792 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878270.4783633, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776878270.4802198, 'checkpointer_blocking_duration_secs': 0.8483619689941406, 'get_old_steps_start_time': 1776878271.3285992, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1776878270.4762802, 'checkpoint_manager_blocking_duration_secs': 0.9236564636230469} I0422 17:17:51.400080 140119134377792 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0422 17:17:51.400143 140119134377792 max_utils.py:750] Memstats: After params initialized: I0422 17:17:51.400198 140119134377792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0422 17:17:51.400230 140119134377792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0422 17:17:51.400256 140119134377792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0422 17:17:51.400281 140119134377792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0422 17:17:51.665389 140119134377792 metric_logger.py:185] completed step: 10, seconds: 11.013, TFLOP/s/device: 1.234, Tokens/s/device: 185.964, total_weights: 65536, loss: 6.833 I0422 17:17:51.848678 140119134377792 metric_logger.py:185] completed step: 11, seconds: 1.347, TFLOP/s/device: 10.084, Tokens/s/device: 1519.941, total_weights: 65536, loss: 6.528 I0422 17:17:51.996355 140119134377792 metric_logger.py:185] completed step: 12, seconds: 0.036, TFLOP/s/device: 375.014, Tokens/s/device: 56526.179, total_weights: 65536, loss: 6.265 I0422 17:17:52.143883 140119134377792 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.838, Tokens/s/device: 13390.610, total_weights: 65536, loss: 6.045 I0422 17:17:52.148654 140119134377792 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0422 17:17:52.438383 140119134377792 checkpointing.py:798] Waited 0.2896919250488281 seconds for step 15 to finish before starting checkpointing. I0422 17:17:52.441135 140119134377792 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:17:55.981670 139997638272768 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/10/items/array_metadatas/process_6 I0422 17:18:33.042751 140001429939968 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.411 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.22301268577576 s) (per-host) I0422 17:18:33.042875 140001429939968 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.714465s. I0422 17:18:40.837198 140001429939968 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.508771s. I0422 17:18:40.837511 139997143365376 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:18:40.837635 139997143365376 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:18:40.837692 139997143365376 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:18:40.839257 139997143365376 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0422 17:18:40.839434 140119134377792 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0422 17:18:40.839581 140119134377792 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.398465 seconds. If this number is high, consider checkpointing less frequently. I0422 17:18:40.841573 140119134377792 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 15 I0422 17:18:40.845567 140119134377792 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/15. I0422 17:18:41.809582 140119134377792 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:18:41.809686 140119134377792 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 I0422 17:18:41.848682 140119134377792 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.040259s I0422 17:18:41.848853 140119134377792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.125 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3009974956512451 s) (per-host) I0422 17:18:41.848906 140119134377792 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.301059s (batch_requests_ready=0.258808s, total_serialization_initiated=0.042181s, others=0.000069s) I0422 17:18:41.848991 140119134377792 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.306452s (all_items=0.000012s, per_item={'items': '0.00001240'}, temp_paths=0.306440) I0422 17:18:41.849875 139998737192704 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:18:41.850033 140119134377792 async_checkpointer.py:561] Finished blocking save. Time taken: 1.008392s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/15. I0422 17:18:42.328587 140119134377792 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:18:42.328965 139997143365376 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:18:42.329159 140119134377792 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878272.4410918, 'wait_for_prev_duration_secs': 48.398465156555176, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776878320.8416123, 'checkpointer_blocking_duration_secs': 1.008537769317627, 'get_old_steps_start_time': 1776878321.8501678, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776878272.4390082, 'checkpoint_manager_blocking_duration_secs': 49.890114307403564} I0422 17:18:42.329269 140119134377792 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0422 17:18:42.329898 140119134377792 metric_logger.py:185] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.985, Tokens/s/device: 13864.912, total_weights: 65536, loss: 5.864 I0422 17:18:42.335994 140119134377792 metric_logger.py:185] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.081, Tokens/s/device: 13879.382, total_weights: 65536, loss: 5.719 I0422 17:19:00.102205 140119134377792 metric_logger.py:185] completed step: 16, seconds: 50.186, TFLOP/s/device: 0.271, Tokens/s/device: 40.808, total_weights: 65536, loss: 5.605 I0422 17:19:00.249556 140119134377792 metric_logger.py:185] completed step: 17, seconds: 17.620, TFLOP/s/device: 0.771, Tokens/s/device: 116.231, total_weights: 65536, loss: 5.517 I0422 17:19:00.397065 140119134377792 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.284, Tokens/s/device: 13457.924, total_weights: 65536, loss: 5.448 I0422 17:19:00.544385 140119134377792 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0422 17:19:00.545053 140119134377792 checkpointing.py:798] Waited 0.0006859302520751953 seconds for step 19 to finish before starting checkpointing. I0422 17:19:00.546951 140119134377792 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:19:03.616757 139997638272768 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/15/items/array_metadatas/process_6 I0422 17:19:41.737848 139998737192704 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 26.244 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.18995022773743 s) (per-host) I0422 17:19:41.737974 139998737192704 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.887985s. I0422 17:19:48.610730 139998737192704 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 66.760725s. I0422 17:19:48.611002 139997143365376 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:19:48.611150 139997143365376 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:19:48.611202 139997143365376 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:19:48.613789 139997143365376 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0422 17:19:48.614003 140119134377792 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0422 17:19:48.614151 140119134377792 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.067213 seconds. If this number is high, consider checkpointing less frequently. I0422 17:19:48.615652 140119134377792 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0422 17:19:48.619046 140119134377792 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/19. I0422 17:19:49.163435 140119134377792 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0422 17:19:49.163532 140119134377792 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 I0422 17:19:49.204121 140119134377792 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041653s I0422 17:19:49.204286 140119134377792 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.200 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29665589332580566 s) (per-host) I0422 17:19:49.204333 140119134377792 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296713s (batch_requests_ready=0.252903s, total_serialization_initiated=0.043748s, others=0.000062s) I0422 17:19:49.204419 140119134377792 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.300704s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.300693) I0422 17:19:49.205290 140001421547264 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:19:49.205446 140119134377792 async_checkpointer.py:561] Finished blocking save. Time taken: 0.589724s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/19. I0422 17:19:49.223593 140119134377792 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:19:49.223841 139997143365376 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:19:49.223975 140119134377792 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776878340.5469213, 'wait_for_prev_duration_secs': 48.06721329689026, 'time_between_consecutive_saves_sec': 19.70762348175049, 'checkpointer_blocking_start_time': 1776878388.6156905, 'checkpointer_blocking_duration_secs': 0.5898675918579102, 'get_old_steps_start_time': 1776878389.2055721, 'get_old_steps_duration_secs': 2.0503997802734375e-05, 'checkpoint_manager_blocking_start_time': 1776878340.5452943, 'checkpoint_manager_blocking_duration_secs': 48.678648948669434} I0422 17:19:49.224086 140119134377792 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0422 17:19:49.224144 140119134377792 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:19:54.052968 140001429939968 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/linen_xpk_feat_nnx_set_defaults_true_20260422_154647_16_zarr3_true/checkpoints/19/items/array_metadatas/process_6 I0422 17:20:31.011744 140001421547264 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.516 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.10406756401062 s) (per-host) I0422 17:20:31.011865 140001421547264 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.806466s. I0422 17:20:39.371089 140001421547264 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.165675s. I0422 17:20:39.371366 139997143365376 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:20:39.371423 139997143365376 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:20:39.371463 139997143365376 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:20:39.373058 139997143365376 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0422 17:20:39.373252 140119134377792 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0422 17:20:39.373527 140119134377792 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:20:39.374241 140119134377792 metric_logger.py:185] completed step: 19, seconds: 0.148, TFLOP/s/device: 92.021, Tokens/s/device: 13870.358, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 17:20:49 UTC 2026 EXIT_CODE=0