XPK Start: Sat Apr 25 16:25:02 UTC 2026 2026-04-25 16:25:06.355519: 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 00:00:1777134306.368356 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777134306.372126 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777134306.383735 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134306.383757 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134306.383759 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777134306.383761 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:25:25.756164: 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 16:25:26.282805 134298243360576 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:25:35,324:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:25:35.324359 134298243360576 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:25:35,326:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-8vo61-slice-job-0-0.mt-16-zarr3-false-sv-8vo61:8482 I0425 16:25:35.326689 134298243360576 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-8vo61-slice-job-0-0.mt-16-zarr3-false-sv-8vo61:8482 I0425 16:25:36.285314 134298243360576 max_utils.py:284] Jax distributed system initialized! I0425 16:25:43.079403 134298243360576 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:25:43.079508 134298243360576 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:25:43.079550 134298243360576 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 16:25:43.079585 134298243360576 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:25:43.701262 134298243360576 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:25:43.701861 134298243360576 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:25:43.702044 134298243360576 checkpointing.py:688] Setting up checkpoint logger... I0425 16:25:43.702119 134298243360576 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0425 16:25:43.702164 134298243360576 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:25:43.702515 134298243360576 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 0x7a242e3710a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:25:46.598587 134298243360576 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:25:46.598827 134298243360576 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 0x7a0f98445af0>}, handler_registry=None I0425 16:25:46.599062 134298243360576 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a0f98445af0>` 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 16:25:46.599123 134298243360576 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0f9844ac60>` 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 16:25:46.599159 134298243360576 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 0x7a0f98445af0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a0f98445af0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0f9844ac60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a0f9844ac60>}). I0425 16:25:46.599665 134298243360576 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:25:46.599744 134298243360576 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a0f98228180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:25:47.696020 134298243360576 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints I0425 16:25:48.156717 134298243360576 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a11d5d8aff0> I0425 16:25:48.156892 134298243360576 checkpointing.py:302] Checkpoint manager created! I0425 16:25:49.047830 134298243360576 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:25:49.047939 134298243360576 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:25:49.417230 134298243360576 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:25:49.417323 134298243360576 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:25:49.433130 134298243360576 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:25:49.433189 134298243360576 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:25:49.455961 134298243360576 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 16:25:49.456029 134298243360576 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:25:49.471827 134298243360576 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 16:25:49.471897 134298243360576 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:25:49.487714 134298243360576 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 16:25:49.487776 134298243360576 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:25:49.512761 134298243360576 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:25:49.512830 134298243360576 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:25:49.532421 134298243360576 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:25:49.532489 134298243360576 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:25:49.721299 134298243360576 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:25:49.721410 134298243360576 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 16:25:51.161247 134298243360576 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0425 16:25:51.161387 134298243360576 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) I0425 16:25:51.161448 134298243360576 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) I0425 16:25:51.161509 134298243360576 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') I0425 16:25:51.161566 134298243360576 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) I0425 16:25:51.161608 134298243360576 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) I0425 16:25:51.161662 134298243360576 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) I0425 16:25:51.161715 134298243360576 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') I0425 16:25:51.161756 134298243360576 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) I0425 16:25:51.161794 134298243360576 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) I0425 16:25:51.161843 134298243360576 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0425 16:25:51.161892 134298243360576 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0425 16:25:51.675220 134298243360576 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:25:51.675314 134298243360576 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:25:51.690262 134298243360576 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:25:51.690321 134298243360576 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:26:02.638749 134298243360576 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 16:26:02.639567 134298243360576 metric_logger.py:289] number parameters: 1.104 billion I0425 16:26:14.074505 134298243360576 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 16:26:14.228742 134298243360576 checkpointing.py:798] Waited 0.15421652793884277 seconds for step 0 to finish before starting checkpointing. I0425 16:26:14.231592 134298243360576 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:26:14.233385 134298243360576 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 16:26:14.236709 134298243360576 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0. I0425 16:26:14.824907 134298243360576 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:26:14.825901 134298243360576 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:26:14.825957 134298243360576 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 16:26:15.076142 134298243360576 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251315s I0425 16:26:15.076313 134298243360576 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.944 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.523956298828125 s) (per-host) I0425 16:26:15.076364 134298243360576 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.524016s (batch_requests_ready=0.269074s, total_serialization_initiated=0.254875s, others=0.000067s) I0425 16:26:15.076453 134298243360576 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.528147s (all_items=0.000025s, per_item={'items': '0.00002527'}, temp_paths=0.528121) I0425 16:26:15.077463 134172144293632 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:26:15.077639 134298243360576 async_checkpointer.py:561] Finished blocking save. Time taken: 0.844180s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0. I0425 16:26:15.537436 134298243360576 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:26:15.537815 134171036980992 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:26:15.537985 134298243360576 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134374.2315724, 'wait_for_prev_duration_secs': 6.508827209472656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777134374.2334251, 'checkpointer_blocking_duration_secs': 0.8443214893341064, 'get_old_steps_start_time': 1777134375.0777686, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777134374.2294505, 'checkpoint_manager_blocking_duration_secs': 1.30849289894104} I0425 16:26:15.538091 134298243360576 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 16:26:15.538160 134298243360576 max_utils.py:750] Memstats: After params initialized: I0425 16:26:15.538216 134298243360576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 16:26:15.538251 134298243360576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 16:26:15.538286 134298243360576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 16:26:15.538314 134298243360576 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 16:26:15.804695 134298243360576 metric_logger.py:185] completed step: 0, seconds: 11.435, TFLOP/s/device: 1.188, Tokens/s/device: 179.102, total_weights: 65536, loss: 10.874 I0425 16:26:15.982061 134298243360576 metric_logger.py:185] completed step: 1, seconds: 1.729, TFLOP/s/device: 7.859, Tokens/s/device: 1184.555, total_weights: 65536, loss: 10.874 I0425 16:26:16.376713 134298243360576 metric_logger.py:185] completed step: 2, seconds: 0.030, TFLOP/s/device: 449.250, Tokens/s/device: 67715.911, total_weights: 65536, loss: 10.560 I0425 16:26:16.523986 134298243360576 metric_logger.py:185] completed step: 3, seconds: 0.395, TFLOP/s/device: 34.415, Tokens/s/device: 5187.371, total_weights: 65536, loss: 9.980 I0425 16:26:16.528990 134298243360576 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 16:26:16.818566 134298243360576 checkpointing.py:798] Waited 0.2895371913909912 seconds for step 5 to finish before starting checkpointing. I0425 16:26:16.821470 134298243360576 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:26:17.737190 2602 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:26:19.657360 134172152686336 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0425 16:26:54.750960 134172144293632 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 39.295 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.19857597351074 s) (per-host) I0425 16:26:54.751063 134172144293632 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.673488s. I0425 16:27:08.568705 134172144293632 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 53.491112s. I0425 16:27:08.568984 134171036980992 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:27:08.569132 134171036980992 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:27:08.569182 134171036980992 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:27:08.572319 134171036980992 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 16:27:08.572510 134298243360576 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 16:27:08.572653 134298243360576 checkpoint_manager.py:1458] Waiting for previous save to complete took 51.751188 seconds. If this number is high, consider checkpointing less frequently. I0425 16:27:08.574586 134298243360576 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 16:27:08.578644 134298243360576 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5. I0425 16:27:09.124941 134298243360576 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:27:09.125040 134298243360576 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 16:27:09.171382 134298243360576 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047565s I0425 16:27:09.171547 134298243360576 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.013 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30773472785949707 s) (per-host) I0425 16:27:09.171600 134298243360576 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.307796s (batch_requests_ready=0.258323s, total_serialization_initiated=0.049403s, others=0.000069s) I0425 16:27:09.171689 134298243360576 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312669s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.312655) I0425 16:27:09.172488 134172152686336 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:27:09.172645 134298243360576 async_checkpointer.py:561] Finished blocking save. Time taken: 0.597997s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5. I0425 16:27:09.185035 134298243360576 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:27:09.185285 134171036980992 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:27:09.185412 134298243360576 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134376.8214402, 'wait_for_prev_duration_secs': 51.75118827819824, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777134428.5746262, 'checkpointer_blocking_duration_secs': 0.5980875492095947, 'get_old_steps_start_time': 1777134429.1727307, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1777134376.8192666, 'checkpoint_manager_blocking_duration_secs': 52.366111755371094} I0425 16:27:09.185513 134298243360576 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 16:27:09.186056 134298243360576 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.771, Tokens/s/device: 13380.549, total_weights: 65536, loss: 9.460 I0425 16:27:09.191456 134298243360576 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.201, Tokens/s/device: 13897.465, total_weights: 65536, loss: 8.959 I0425 16:27:29.188180 134298243360576 metric_logger.py:185] completed step: 6, seconds: 52.661, TFLOP/s/device: 0.258, Tokens/s/device: 38.890, total_weights: 65536, loss: 8.469 I0425 16:27:29.335713 134298243360576 metric_logger.py:185] completed step: 7, seconds: 19.851, TFLOP/s/device: 0.684, Tokens/s/device: 103.170, total_weights: 65536, loss: 8.003 I0425 16:27:29.483074 134298243360576 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.377, Tokens/s/device: 13471.912, total_weights: 65536, loss: 7.572 I0425 16:27:29.630088 134298243360576 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 16:27:29.630805 134298243360576 checkpointing.py:798] Waited 0.0007371902465820312 seconds for step 9 to finish before starting checkpointing. I0425 16:27:29.632779 134298243360576 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:27:33.909921 134169459939072 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0425 16:28:09.697614 134172152686336 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.966 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.8337607383728 s) (per-host) I0425 16:28:09.697720 134172152686336 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.525111s. I0425 16:28:19.922339 134172152686336 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 70.749713s. I0425 16:28:19.922566 134171036980992 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:28:19.922640 134171036980992 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:28:19.922692 134171036980992 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:28:19.924470 134171036980992 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 16:28:19.924671 134298243360576 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 16:28:19.924808 134298243360576 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.292043 seconds. If this number is high, consider checkpointing less frequently. I0425 16:28:19.926301 134298243360576 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 16:28:19.929683 134298243360576 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 16:28:20.614808 134298243360576 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:28:20.614904 134298243360576 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 16:28:20.659788 134298243360576 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045949s I0425 16:28:20.659952 134298243360576 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.029 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3067331314086914 s) (per-host) I0425 16:28:20.660003 134298243360576 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306793s (batch_requests_ready=0.258783s, total_serialization_initiated=0.047945s, others=0.000065s) I0425 16:28:20.660105 134298243360576 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.310797s (all_items=0.000010s, per_item={'items': '0.00001049'}, temp_paths=0.310787) I0425 16:28:20.662029 134171087337216 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:28:20.662154 134298243360576 async_checkpointer.py:561] Finished blocking save. Time taken: 0.735782s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9. I0425 16:28:20.689176 134298243360576 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:28:20.689433 134171036980992 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:28:20.689602 134298243360576 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777134449.6327486, 'wait_for_prev_duration_secs': 50.29204297065735, 'time_between_consecutive_saves_sec': 21.060389757156372, 'checkpointer_blocking_start_time': 1777134499.9263418, 'checkpointer_blocking_duration_secs': 0.7359225749969482, 'get_old_steps_start_time': 1777134500.6622784, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1777134449.63103, 'checkpoint_manager_blocking_duration_secs': 51.05853605270386} I0425 16:28:20.689704 134298243360576 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 16:28:20.689747 134298243360576 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:28:26.315350 134172144293632 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_20260425_150412/linen_xpk_feat_nnx_set_defaults_true_20260425_150412_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0425 16:29:01.467187 134171087337216 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.420 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.11392641067505 s) (per-host) I0425 16:29:01.467316 134171087337216 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.805220s. I0425 16:29:10.065069 134171087337216 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.402957s. I0425 16:29:10.065368 134171036980992 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:29:10.065500 134171036980992 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:29:10.065552 134171036980992 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:29:10.067693 134171036980992 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:29:10.067864 134298243360576 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 16:29:10.068004 134298243360576 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:29:10.068652 134298243360576 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.125, Tokens/s/device: 13886.158, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 16:29:23 UTC 2026 EXIT_CODE=0