XPK Start: Thu Apr 23 16:07:38 UTC 2026 2026-04-23 16:07:42.861275: 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 00:00:1776960462.873826 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776960462.877510 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776960462.889026 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776960462.889046 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776960462.889048 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776960462.889050 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 16:08:02.057059: 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) I0423 16:08:02.582735 131970275329856 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 16:08:11,624:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 16:08:11.624266 131970275329856 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 16:08:11,626:jax._src.distributed:157: Connecting to JAX distributed service on mt-06-grad-accum-bb3d1-slice-job-0-0.mt-06-grad-accum-bb3d1:8482 I0423 16:08:11.626745 131970275329856 distributed.py:157] Connecting to JAX distributed service on mt-06-grad-accum-bb3d1-slice-job-0-0.mt-06-grad-accum-bb3d1:8482 I0423 16:08:13.451816 131970275329856 max_utils.py:284] Jax distributed system initialized! I0423 16:08:20.314239 131970275329856 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 16:08:20.314351 131970275329856 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 16:08:20.314392 131970275329856 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 I0423 16:08:20.314427 131970275329856 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 16:08:20.942213 131970275329856 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 16:08:20.942806 131970275329856 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 16:08:20.943005 131970275329856 checkpointing.py:688] Setting up checkpoint logger... I0423 16:08:20.943057 131970275329856 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 16:08:20.943100 131970275329856 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 16:08:20.943458 131970275329856 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 0x780644975400>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 16:08:23.901315 131970275329856 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 16:08:23.901573 131970275329856 checkpoint_manager.py:709] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77f1e06ad0a0>}, handler_registry=None I0423 16:08:23.901833 131970275329856 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77f1e06ad0a0>` 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`. I0423 16:08:23.901904 131970275329856 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77f1e06b1fa0>` 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`. I0423 16:08:23.901961 131970275329856 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 0x77f1e06ad0a0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x77f1e06ad0a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77f1e06b1fa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x77f1e06b1fa0>}). I0423 16:08:23.902491 131970275329856 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 16:08:23.902594 131970275329856 async_checkpointer.py:177] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x77f1e0474180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 16:08:25.021498 131970275329856 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints I0423 16:08:25.495502 131970275329856 checkpoint_manager.py:929] [process=4][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=10), 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_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7804f3fff170> I0423 16:08:25.495745 131970275329856 checkpointing.py:302] Checkpoint manager created! I0423 16:08:27.166017 131970275329856 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 16:08:27.166121 131970275329856 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:08:27.537654 131970275329856 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:08:27.537746 131970275329856 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:08:27.553774 131970275329856 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:08:27.553835 131970275329856 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 16:08:27.576771 131970275329856 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:08:27.576840 131970275329856 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:08:27.592775 131970275329856 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:08:27.592840 131970275329856 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:08:27.608635 131970275329856 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:08:27.608692 131970275329856 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:08:27.633626 131970275329856 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 16:08:27.633699 131970275329856 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 16:08:27.653239 131970275329856 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 16:08:27.653307 131970275329856 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 16:08:27.843360 131970275329856 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 16:08:27.843471 131970275329856 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 16:08:29.300206 131970275329856 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 16:08:29.300353 131970275329856 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) I0423 16:08:29.300411 131970275329856 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) I0423 16:08:29.300474 131970275329856 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') I0423 16:08:29.300532 131970275329856 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) I0423 16:08:29.300572 131970275329856 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) I0423 16:08:29.300627 131970275329856 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) I0423 16:08:29.300683 131970275329856 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') I0423 16:08:29.300724 131970275329856 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) I0423 16:08:29.300762 131970275329856 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) I0423 16:08:29.300837 131970275329856 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0423 16:08:29.300926 131970275329856 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0423 16:08:29.322785 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown. I0423 16:08:29.322851 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,). I0423 16:08:29.337127 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown. I0423 16:08:29.337186 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None). I0423 16:08:29.365899 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown. I0423 16:08:29.365962 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp'). I0423 16:08:29.380091 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown. I0423 16:08:29.380146 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None). I0423 16:08:29.408715 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown. I0423 16:08:29.408788 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None). I0423 16:08:29.423058 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown. I0423 16:08:29.423116 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp'). I0423 16:08:29.465615 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown. I0423 16:08:29.465683 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None). I0423 16:08:29.479939 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown. I0423 16:08:29.479995 131970275329856 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp'). I0423 16:08:30.152130 131970275329856 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:08:30.152225 131970275329856 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0423 16:08:30.167691 131970275329856 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:08:30.167751 131970275329856 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0423 16:08:41.614627 131970275329856 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0423 16:08:41.615569 131970275329856 metric_logger.py:289] number parameters: 1.104 billion I0423 16:08:43.689956 131970275329856 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 16:08:53.673574 131970275329856 checkpointing.py:798] Waited 9.98359751701355 seconds for step 0 to finish before starting checkpointing. I0423 16:08:53.676070 131970275329856 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:08:53.677850 131970275329856 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 0 I0423 16:08:53.680588 131970275329856 async_checkpointer.py:452] [process=4] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/0. I0423 16:08:54.307507 131970275329856 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 16:08:54.308543 131970275329856 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:08:54.308601 131970275329856 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 I0423 16:08:54.555363 131970275329856 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.247935s I0423 16:08:54.555535 131970275329856 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.868 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5378119945526123 s) (per-host) I0423 16:08:54.555588 131970275329856 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.537874s (batch_requests_ready=0.286342s, total_serialization_initiated=0.251461s, others=0.000071s) I0423 16:08:54.555683 131970275329856 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.541964s (all_items=0.000021s, per_item={'items': '0.00002122'}, temp_paths=0.541943) I0423 16:08:54.556581 131848348874496 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0423 16:08:54.556757 131970275329856 async_checkpointer.py:561] Finished blocking save. Time taken: 0.878819s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/0. I0423 16:08:54.571500 131970275329856 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:08:54.571805 131844615956224 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:08:54.571989 131970275329856 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776960533.6760519, 'wait_for_prev_duration_secs': 6.532669067382812e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776960533.6779046, 'checkpointer_blocking_duration_secs': 0.878960132598877, 'get_old_steps_start_time': 1776960534.556896, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776960533.6741018, 'checkpoint_manager_blocking_duration_secs': 0.8978455066680908} I0423 16:08:54.572093 131970275329856 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 16:08:54.572144 131970275329856 max_utils.py:750] Memstats: After params initialized: I0423 16:08:54.572193 131970275329856 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_16(process=4,(0,4,0,0)) I0423 16:08:54.572227 131970275329856 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0)) I0423 16:08:54.572254 131970275329856 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0)) I0423 16:08:54.572285 131970275329856 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0)) I0423 16:08:54.835682 131970275329856 metric_logger.py:185] completed step: 0, seconds: 2.074, TFLOP/s/device: 26.201, Tokens/s/device: 3949.322, total_weights: 262144, loss: 10.877 I0423 16:08:55.436945 131970275329856 metric_logger.py:185] completed step: 1, seconds: 11.144, TFLOP/s/device: 4.877, Tokens/s/device: 735.074, total_weights: 262144, loss: 10.877 I0423 16:08:56.015413 131970275329856 metric_logger.py:185] completed step: 2, seconds: 0.023, TFLOP/s/device: 2369.676, Tokens/s/device: 357183.344, total_weights: 262144, loss: 10.563 I0423 16:08:56.593684 131970275329856 metric_logger.py:185] completed step: 3, seconds: 0.584, TFLOP/s/device: 93.038, Tokens/s/device: 14023.699, total_weights: 262144, loss: 10.272 I0423 16:08:57.639853 2400 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 16:08:57.750732 131970275329856 metric_logger.py:185] completed step: 4, seconds: 0.578, TFLOP/s/device: 93.981, Tokens/s/device: 14165.829, total_weights: 262144, loss: 10.022 I0423 16:08:57.756964 131970275329856 metric_logger.py:185] completed step: 5, seconds: 0.578, TFLOP/s/device: 93.958, Tokens/s/device: 14162.352, total_weights: 262144, loss: 9.820 I0423 16:08:59.564423 131848357267200 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/0/items/array_metadatas/process_4 I0423 16:09:32.792567 131848348874496 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 40.738 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.77480387687683 s) (per-host) I0423 16:09:32.792695 131848348874496 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.235999s. I0423 16:09:36.228007 131970275329856 metric_logger.py:185] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.942, Tokens/s/device: 7075.586, total_weights: 262144, loss: 9.667 I0423 16:09:36.806478 131970275329856 metric_logger.py:185] completed step: 7, seconds: 37.894, TFLOP/s/device: 1.434, Tokens/s/device: 216.184, total_weights: 262144, loss: 9.561 I0423 16:09:37.384733 131970275329856 metric_logger.py:185] completed step: 8, seconds: 0.583, TFLOP/s/device: 93.232, Tokens/s/device: 14052.880, total_weights: 262144, loss: 9.496 I0423 16:09:37.962851 131970275329856 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 16:09:37.963505 131970275329856 checkpointing.py:798] Waited 0.0006759166717529297 seconds for step 9 to finish before starting checkpointing. I0423 16:09:37.965562 131970275329856 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:10:16.506475 131848348874496 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 81.949762s. I0423 16:10:16.506784 131844615956224 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:10:16.506936 131844615956224 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:10:16.506990 131844615956224 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:10:16.508644 131844615956224 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 16:10:16.508886 131970275329856 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0423 16:10:16.509018 131970275329856 checkpoint_manager.py:1458] Waiting for previous save to complete took 38.543463 seconds. If this number is high, consider checkpointing less frequently. I0423 16:10:16.510746 131970275329856 checkpoint_manager.py:1518] [process=4] Saving checkpoint at step 9 I0423 16:10:16.514154 131970275329856 async_checkpointer.py:452] [process=4] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/9. I0423 16:10:17.100784 131970275329856 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:10:17.100894 131970275329856 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 I0423 16:10:17.151371 131970275329856 base_pytree_checkpoint_handler.py:153] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051529s I0423 16:10:17.151541 131970275329856 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.644 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.332167387008667 s) (per-host) I0423 16:10:17.151593 131970275329856 base_pytree_checkpoint_handler.py:737] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.332228s (batch_requests_ready=0.278721s, total_serialization_initiated=0.053438s, others=0.000069s) I0423 16:10:17.151683 131970275329856 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.336377s (all_items=0.000021s, per_item={'items': '0.00002146'}, temp_paths=0.336356) I0423 16:10:17.152608 131844607563520 async_checkpointer.py:79] [process=4][thread=async_save] Background save thread started. I0423 16:10:17.152767 131970275329856 async_checkpointer.py:561] Finished blocking save. Time taken: 0.641950s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/9. I0423 16:10:17.154696 131970275329856 checkpoint_manager.py:1566] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:10:17.154941 131844615956224 async_checkpointer.py:265] [process=4][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:10:17.155100 131970275329856 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776960577.9655318, 'wait_for_prev_duration_secs': 38.54346299171448, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776960616.5107853, 'checkpointer_blocking_duration_secs': 0.6421072483062744, 'get_old_steps_start_time': 1776960617.1529117, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776960577.9637187, 'checkpoint_manager_blocking_duration_secs': 39.19134759902954} I0423 16:10:17.155202 131970275329856 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 16:10:17.155247 131970275329856 checkpoint_manager.py:2024] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:10:21.914113 131848357267200 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_06_grad_accum/checkpoints/9/items/array_metadatas/process_4 I0423 16:10:59.697092 131844607563520 base_pytree_checkpoint_handler.py:129] [process=4] /jax/checkpoint/write/gbytes_per_sec: 36.840 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.87767577171326 s) (per-host) I0423 16:10:59.697214 131844607563520 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.544496s. I0423 16:11:07.708178 131844607563520 async_checkpointer.py:144] [process=4][thread=async_save] Background save thread done. Time taken: 50.555445s. I0423 16:11:07.708486 131844615956224 async_checkpointer.py:273] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:11:07.708602 131844615956224 async_checkpointer.py:283] [process=4][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:11:07.708652 131844615956224 checkpoint_manager.py:2133] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:11:07.710134 131844615956224 checkpoint_manager.py:2142] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 16:11:07.710319 131970275329856 checkpoint_manager.py:2036] [process=4][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0423 16:11:07.710460 131970275329856 checkpoint_manager.py:2013] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:11:07.711196 131970275329856 metric_logger.py:185] completed step: 9, seconds: 0.578, TFLOP/s/device: 93.969, Tokens/s/device: 14163.992, total_weights: 262144, loss: 9.457 Per train step: Total TFLOPs: 54.35 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 16:11:19 UTC 2026 EXIT_CODE=0