XPK Start: Sat Apr 25 16:07:12 UTC 2026 2026-04-25 16:07:16.697429: 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 00:00:1777133236.710236 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777133236.714220 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777133236.725892 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777133236.725912 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777133236.725915 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777133236.725917 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:07:35.898396: 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 16:07:36.420494 138979489498944 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:07:45,466:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:07:45.466567 138979489498944 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:07:45,468:jax._src.distributed:157: Connecting to JAX distributed service on mt-15-ocdbt-false-sv-n1fh1-slice-job-0-0.mt-15-ocdbt-false-sv-n1fh1:8482 I0425 16:07:45.468897 138979489498944 distributed.py:157] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-n1fh1-slice-job-0-0.mt-15-ocdbt-false-sv-n1fh1:8482 I0425 16:07:46.487485 138979489498944 max_utils.py:284] Jax distributed system initialized! I0425 16:07:53.568215 138979489498944 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:07:53.568320 138979489498944 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:07:53.568361 138979489498944 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 16:07:53.568397 138979489498944 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:07:54.188718 138979489498944 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:07:54.189330 138979489498944 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:07:54.189517 138979489498944 checkpointing.py:688] Setting up checkpoint logger... I0425 16:07:54.189567 138979489498944 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0425 16:07:54.189609 138979489498944 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:07:54.189951 138979489498944 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 0x7e6620b751c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:07:57.265717 138979489498944 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:07:57.265963 138979489498944 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 0x7e51c4131880>}, handler_registry=None I0425 16:07:57.266208 138979489498944 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e51c4131880>` 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 16:07:57.266259 138979489498944 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e51c41368d0>` 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 16:07:57.266296 138979489498944 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 0x7e51c4131880>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e51c4131880>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e51c41368d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e51c41368d0>}). I0425 16:07:57.266812 138979489498944 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:07:57.266892 138979489498944 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e5174628180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:07:59.141613 138979489498944 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_15_ocdbt_false/checkpoints I0425 16:07:59.143837 138979489498944 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_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e51c429d6d0> I0425 16:07:59.143959 138979489498944 checkpointing.py:302] Checkpoint manager created! I0425 16:08:00.048706 138979489498944 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:08:00.048813 138979489498944 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:08:00.420840 138979489498944 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:08:00.420933 138979489498944 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:08:00.436969 138979489498944 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:08:00.437031 138979489498944 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 16:08:00.460060 138979489498944 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 16:08:00.460136 138979489498944 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:08:00.476172 138979489498944 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 16:08:00.476233 138979489498944 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:08:00.492343 138979489498944 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 16:08:00.492406 138979489498944 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:08:00.517573 138979489498944 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:08:00.517646 138979489498944 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 16:08:00.537213 138979489498944 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:08:00.537281 138979489498944 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 16:08:00.727999 138979489498944 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:08:00.728120 138979489498944 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 16:08:02.175552 138979489498944 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0425 16:08:02.175686 138979489498944 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 16:08:02.175742 138979489498944 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 16:08:02.175801 138979489498944 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 16:08:02.175853 138979489498944 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 16:08:02.175892 138979489498944 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 16:08:02.175945 138979489498944 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 16:08:02.175997 138979489498944 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 16:08:02.176037 138979489498944 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 16:08:02.176074 138979489498944 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 16:08:02.176168 138979489498944 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0425 16:08:02.176238 138979489498944 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0425 16:08:02.674819 138979489498944 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:08:02.674917 138979489498944 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:08:02.689912 138979489498944 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 16:08:02.689979 138979489498944 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 16:08:13.358657 138979489498944 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 16:08:13.359532 138979489498944 metric_logger.py:289] number parameters: 1.104 billion I0425 16:08:24.677349 138979489498944 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 16:08:24.832023 138979489498944 checkpointing.py:798] Waited 0.15465426445007324 seconds for step 0 to finish before starting checkpointing. I0425 16:08:24.834449 138979489498944 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:08:24.836125 138979489498944 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 16:08:24.839013 138979489498944 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_15_ocdbt_false/checkpoints/0. I0425 16:08:25.837024 138979489498944 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:08:25.838110 138979489498944 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:08:25.838169 138979489498944 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 16:08:26.091573 138979489498944 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.254632s I0425 16:08:26.091749 138979489498944 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.849 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5414230823516846 s) (per-host) I0425 16:08:26.091800 138979489498944 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.541484s (batch_requests_ready=0.283213s, total_serialization_initiated=0.258203s, others=0.000068s) I0425 16:08:26.091889 138979489498944 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.545470s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.545451) I0425 16:08:26.092867 138856611436288 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:08:26.093041 138979489498944 async_checkpointer.py:561] Finished blocking save. Time taken: 1.256845s. 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_15_ocdbt_false/checkpoints/0. I0425 16:08:26.097442 138979489498944 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:08:26.097702 138851779598080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:08:26.097828 138979489498944 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777133304.8344316, 'wait_for_prev_duration_secs': 7.033348083496094e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777133304.8361642, 'checkpointer_blocking_duration_secs': 1.2569985389709473, 'get_old_steps_start_time': 1777133306.0931818, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777133304.8325949, 'checkpoint_manager_blocking_duration_secs': 1.2651984691619873} I0425 16:08:26.097929 138979489498944 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 16:08:26.097981 138979489498944 max_utils.py:750] Memstats: After params initialized: I0425 16:08:26.098029 138979489498944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 16:08:26.098064 138979489498944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 16:08:26.098092 138979489498944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 16:08:26.098133 138979489498944 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 16:08:26.365267 138979489498944 metric_logger.py:185] completed step: 0, seconds: 11.318, TFLOP/s/device: 1.201, Tokens/s/device: 180.955, total_weights: 65536, loss: 10.874 I0425 16:08:26.536564 138979489498944 metric_logger.py:185] completed step: 1, seconds: 1.687, TFLOP/s/device: 8.056, Tokens/s/device: 1214.274, total_weights: 65536, loss: 10.874 I0425 16:08:26.932943 138979489498944 metric_logger.py:185] completed step: 2, seconds: 0.024, TFLOP/s/device: 560.895, Tokens/s/device: 84544.254, total_weights: 65536, loss: 10.560 I0425 16:08:27.080404 138979489498944 metric_logger.py:185] completed step: 3, seconds: 0.397, TFLOP/s/device: 34.254, Tokens/s/device: 5163.086, total_weights: 65536, loss: 9.980 I0425 16:08:27.084828 138979489498944 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 16:08:27.374801 138979489498944 checkpointing.py:798] Waited 0.28993725776672363 seconds for step 5 to finish before starting checkpointing. I0425 16:08:27.377559 138979489498944 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:08:28.648120 2607 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:08:30.537809 138856619828992 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_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6 I0425 16:08:36.334970 138856611436288 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 146.468 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.784610271453857 s) (per-host) I0425 16:08:36.335076 138856611436288 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.242095s. I0425 16:08:42.312911 138856611436288 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.219914s. I0425 16:08:42.313188 138851779598080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:08:42.313246 138851779598080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:08:42.313292 138851779598080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:08:42.314961 138851779598080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 16:08:42.315149 138979489498944 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 16:08:42.315287 138979489498944 checkpoint_manager.py:1458] Waiting for previous save to complete took 14.937727 seconds. If this number is high, consider checkpointing less frequently. I0425 16:08:42.317229 138979489498944 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0425 16:08:42.321212 138979489498944 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_15_ocdbt_false/checkpoints/5. I0425 16:08:42.924638 138979489498944 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:08:42.924738 138979489498944 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 16:08:42.969079 138979489498944 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045713s I0425 16:08:42.969276 138979489498944 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.268 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2928454875946045 s) (per-host) I0425 16:08:42.969328 138979489498944 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.292909s (batch_requests_ready=0.245270s, total_serialization_initiated=0.047568s, others=0.000071s) I0425 16:08:42.969415 138979489498944 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297806s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=0.297791) I0425 16:08:42.970328 138846935537408 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:08:42.970458 138979489498944 async_checkpointer.py:561] Finished blocking save. Time taken: 0.653153s. 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_15_ocdbt_false/checkpoints/5. I0425 16:08:42.991863 138979489498944 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:08:42.992143 138851779598080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:08:42.992318 138979489498944 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777133307.3775306, 'wait_for_prev_duration_secs': 14.937726736068726, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777133322.3172681, 'checkpointer_blocking_duration_secs': 0.6533603668212891, 'get_old_steps_start_time': 1777133322.9706469, 'get_old_steps_duration_secs': 2.384185791015625e-05, 'checkpoint_manager_blocking_start_time': 1777133307.3753572, 'checkpoint_manager_blocking_duration_secs': 15.616926431655884} I0425 16:08:42.992428 138979489498944 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 16:08:42.992974 138979489498944 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.987, Tokens/s/device: 13413.149, total_weights: 65536, loss: 9.460 I0425 16:08:42.998582 138979489498944 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.304, Tokens/s/device: 13913.043, total_weights: 65536, loss: 8.959 I0425 16:09:02.298756 138979489498944 metric_logger.py:185] completed step: 6, seconds: 15.913, TFLOP/s/device: 0.854, Tokens/s/device: 128.704, total_weights: 65536, loss: 8.469 I0425 16:09:02.446111 138979489498944 metric_logger.py:185] completed step: 7, seconds: 19.154, TFLOP/s/device: 0.709, Tokens/s/device: 106.924, total_weights: 65536, loss: 8.003 I0425 16:09:02.593290 138979489498944 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.261, Tokens/s/device: 13454.388, total_weights: 65536, loss: 7.572 I0425 16:09:02.740396 138979489498944 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 16:09:02.741043 138979489498944 checkpointing.py:798] Waited 0.0006670951843261719 seconds for step 9 to finish before starting checkpointing. I0425 16:09:02.743384 138979489498944 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:09:06.947346 138856619828992 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_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6 I0425 16:09:13.039328 138846935537408 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 52.024 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 30.362829446792603 s) (per-host) I0425 16:09:13.039448 138846935537408 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.068910s. I0425 16:09:19.229089 138846935537408 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 36.258537s. I0425 16:09:19.229344 138851779598080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:09:19.229415 138851779598080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:09:19.229462 138851779598080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:09:19.231011 138851779598080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 16:09:19.231201 138979489498944 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 16:09:19.231351 138979489498944 checkpoint_manager.py:1458] Waiting for previous save to complete took 16.487979 seconds. If this number is high, consider checkpointing less frequently. I0425 16:09:19.232636 138979489498944 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 16:09:19.236038 138979489498944 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_15_ocdbt_false/checkpoints/9. I0425 16:09:19.840986 138979489498944 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0425 16:09:19.841087 138979489498944 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 16:09:19.885449 138979489498944 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045439s I0425 16:09:19.885622 138979489498944 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.149 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2995870113372803 s) (per-host) I0425 16:09:19.885671 138979489498944 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299646s (batch_requests_ready=0.252129s, total_serialization_initiated=0.047453s, others=0.000064s) I0425 16:09:19.885759 138979489498944 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303429s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.303418) I0425 16:09:19.886572 138849592391424 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:09:19.886698 138979489498944 async_checkpointer.py:561] Finished blocking save. Time taken: 0.653992s. 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_15_ocdbt_false/checkpoints/9. I0425 16:09:19.894132 138979489498944 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:09:19.894429 138851779598080 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:09:19.894595 138979489498944 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_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777133342.7433558, 'wait_for_prev_duration_secs': 16.487979412078857, 'time_between_consecutive_saves_sec': 20.428351640701294, 'checkpointer_blocking_start_time': 1777133359.2326782, 'checkpointer_blocking_duration_secs': 0.6541900634765625, 'get_old_steps_start_time': 1777133359.8868797, 'get_old_steps_duration_secs': 2.1457672119140625e-05, 'checkpoint_manager_blocking_start_time': 1777133342.741299, 'checkpoint_manager_blocking_duration_secs': 17.153264045715332} I0425 16:09:19.894697 138979489498944 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 16:09:19.894739 138979489498944 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:09:24.542628 138856619828992 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_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6 I0425 16:09:31.275420 138849592391424 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 135.131 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.689341306686401 s) (per-host) I0425 16:09:31.275529 138849592391424 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 11.388751s. I0425 16:09:36.056492 138849592391424 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 16.169698s. I0425 16:09:36.056792 138851779598080 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:09:36.056961 138851779598080 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:09:36.057011 138851779598080 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:09:36.058595 138851779598080 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:09:36.058780 138979489498944 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 16:09:36.058937 138979489498944 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:09:36.059579 138979489498944 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.165, Tokens/s/device: 13892.092, 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 16:09:48 UTC 2026 EXIT_CODE=0