XPK Start: Fri Apr 24 15:10:17 UTC 2026 2026-04-24 15:10:21.165724: 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 00:00:1777043421.178680 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777043421.182577 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777043421.194290 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777043421.194307 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777043421.194309 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777043421.194312 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-24 15:10:40.325920: 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) I0424 15:10:40.857112 139207004628800 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-24 15:10:49,898:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0424 15:10:49.898820 139207004628800 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-24 15:10:49,901:jax._src.distributed:157: Connecting to JAX distributed service on mt-05-fp8-lwsjy-slice-job-0-0.mt-05-fp8-lwsjy:8482 I0424 15:10:49.901062 139207004628800 distributed.py:157] Connecting to JAX distributed service on mt-05-fp8-lwsjy-slice-job-0-0.mt-05-fp8-lwsjy:8482 I0424 15:10:51.342446 139207004628800 max_utils.py:284] Jax distributed system initialized! I0424 15:10:58.410708 139207004628800 max_utils.py:800] System Information: Jax Version: 0.8.1 I0424 15:10:58.410816 139207004628800 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0424 15:10:58.410857 139207004628800 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 I0424 15:10:58.410892 139207004628800 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 15:10:59.032209 139207004628800 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 15:10:59.032807 139207004628800 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 15:10:59.032996 139207004628800 checkpointing.py:688] Setting up checkpoint logger... I0424 15:10:59.033046 139207004628800 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 15:10:59.033089 139207004628800 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0424 15:10:59.033464 139207004628800 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 0x7e9b2636d190>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 15:11:01.974281 139207004628800 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 15:11:01.974529 139207004628800 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 0x7e86a00d2120>}, handler_registry=None I0424 15:11:01.974763 139207004628800 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e86a00d2120>` 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`. I0424 15:11:01.974812 139207004628800 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e86a00d7320>` 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`. I0424 15:11:01.974849 139207004628800 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 0x7e86a00d2120>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e86a00d2120>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e86a00d7320>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e86a00d7320>}). I0424 15:11:01.975372 139207004628800 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0424 15:11:01.975450 139207004628800 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e86784f8180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0424 15:11:02.760739 139207004628800 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints I0424 15:11:03.074345 139207004628800 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=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_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e86a00d2720> I0424 15:11:03.074530 139207004628800 checkpointing.py:302] Checkpoint manager created! I0424 15:11:04.152691 139207004628800 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 15:11:04.152796 139207004628800 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:11:04.527741 139207004628800 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 15:11:04.527846 139207004628800 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:11:04.543785 139207004628800 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0424 15:11:04.543847 139207004628800 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 15:11:04.603645 139207004628800 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:11:04.603736 139207004628800 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:11:04.619684 139207004628800 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:11:04.619746 139207004628800 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:11:04.635583 139207004628800 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 15:11:04.635658 139207004628800 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:11:04.660147 139207004628800 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0424 15:11:04.660218 139207004628800 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 15:11:04.723989 139207004628800 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 15:11:04.724087 139207004628800 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 15:11:05.190023 139207004628800 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 15:11:05.190149 139207004628800 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 15:11:07.247844 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.247974 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248024 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248063 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248112 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248151 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248185 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248223 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248256 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248291 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248326 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248357 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248390 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248423 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248455 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248487 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248519 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248548 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248577 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248606 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248633 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248660 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248689 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248736 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248768 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248796 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248823 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248851 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248879 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248907 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248934 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.248961 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.248988 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.249015 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.249041 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.249070 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.249120 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.249152 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.249181 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.249208 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.249235 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: PartitionSpec() Physical: () I0424 15:11:07.249262 139207004628800 maxtext_utils.py:1740] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: PartitionSpec() Physical: () I0424 15:11:07.249319 139207004628800 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0424 15:11:07.249375 139207004628800 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) I0424 15:11:07.249415 139207004628800 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) I0424 15:11:07.249464 139207004628800 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') I0424 15:11:07.249509 139207004628800 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) I0424 15:11:07.249542 139207004628800 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) I0424 15:11:07.249593 139207004628800 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) I0424 15:11:07.249643 139207004628800 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') I0424 15:11:07.249680 139207004628800 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) I0424 15:11:07.249714 139207004628800 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) I0424 15:11:07.249762 139207004628800 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0424 15:11:07.249806 139207004628800 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0424 15:11:09.037130 139207004628800 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 15:11:09.037228 139207004628800 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 15:11:09.053445 139207004628800 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 15:11:09.053511 139207004628800 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 15:11:22.314065 139207004628800 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. I0424 15:11:22.314990 139207004628800 metric_logger.py:289] number parameters: 1.105 billion I0424 15:11:37.188952 139207004628800 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 15:11:37.356742 139207004628800 checkpointing.py:798] Waited 0.1677711009979248 seconds for step 0 to finish before starting checkpointing. I0424 15:11:37.359737 139207004628800 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 15:11:37.361813 139207004628800 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0424 15:11:37.364826 139207004628800 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/0. I0424 15:11:37.963902 139207004628800 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 15:11:37.964905 139207004628800 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array. I0424 15:11:37.965024 139207004628800 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 I0424 15:11:38.323291 139207004628800 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.359502s I0424 15:11:38.323461 139207004628800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.469 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.6247844696044922 s) (per-host) I0424 15:11:38.323513 139207004628800 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.624846s (batch_requests_ready=0.260007s, total_serialization_initiated=0.364769s, others=0.000070s) I0424 15:11:38.323618 139207004628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.629014s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.628995) I0424 15:11:38.324600 139080469821184 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:11:38.324772 139207004628800 async_checkpointer.py:561] Finished blocking save. Time taken: 0.962886s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/0. I0424 15:11:38.364601 139207004628800 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:11:38.364939 139077902898944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:11:38.365111 139207004628800 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777043497.3597186, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777043497.3618531, 'checkpointer_blocking_duration_secs': 0.9630258083343506, 'get_old_steps_start_time': 1777043498.3248997, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777043497.357592, 'checkpoint_manager_blocking_duration_secs': 1.007469892501831} I0424 15:11:38.365220 139207004628800 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 15:11:38.365273 139207004628800 max_utils.py:750] Memstats: After params initialized: I0424 15:11:38.365333 139207004628800 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_24(process=6,(0,6,0,0)) I0424 15:11:38.365369 139207004628800 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_25(process=6,(1,6,0,0)) I0424 15:11:38.365396 139207004628800 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_28(process=6,(0,7,0,0)) I0424 15:11:38.365420 139207004628800 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_29(process=6,(1,7,0,0)) I0424 15:11:38.633650 139207004628800 metric_logger.py:185] completed step: 0, seconds: 14.874, TFLOP/s/device: 0.913, Tokens/s/device: 137.691, total_weights: 65536, loss: 10.874 I0424 15:11:38.820266 139207004628800 metric_logger.py:185] completed step: 1, seconds: 1.443, TFLOP/s/device: 9.417, Tokens/s/device: 1419.420, total_weights: 65536, loss: 10.874 I0424 15:11:39.236539 139207004628800 metric_logger.py:185] completed step: 2, seconds: 0.029, TFLOP/s/device: 464.247, Tokens/s/device: 69976.424, total_weights: 65536, loss: 10.816 I0424 15:11:39.394783 139207004628800 metric_logger.py:185] completed step: 3, seconds: 0.417, TFLOP/s/device: 32.611, Tokens/s/device: 4915.479, total_weights: 65536, loss: 10.431 I0424 15:11:39.711903 139207004628800 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.971, Tokens/s/device: 12506.336, total_weights: 65536, loss: 9.992 I0424 15:11:39.718361 139207004628800 metric_logger.py:185] completed step: 5, seconds: 0.158, TFLOP/s/device: 86.017, Tokens/s/device: 12965.472, total_weights: 65536, loss: 9.549 I0424 15:12:00.550866 139207004628800 metric_logger.py:185] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.742, Tokens/s/device: 6442.581, total_weights: 65536, loss: 9.111 I0424 15:12:00.709054 139207004628800 metric_logger.py:185] completed step: 7, seconds: 20.675, TFLOP/s/device: 0.657, Tokens/s/device: 99.055, total_weights: 65536, loss: 8.685 I0424 15:12:00.867355 139207004628800 metric_logger.py:185] completed step: 8, seconds: 0.163, TFLOP/s/device: 83.273, Tokens/s/device: 12551.788, total_weights: 65536, loss: 8.281 I0424 15:12:00.872332 139207004628800 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0424 15:12:01.183709 139207004628800 checkpointing.py:798] Waited 0.31134581565856934 seconds for step 10 to finish before starting checkpointing. I0424 15:12:01.186349 139207004628800 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:12:01.669564 2696 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 15:12:03.616331 139080478213888 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/0/items/array_metadatas/process_6 I0424 15:12:18.735277 139080469821184 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.496 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.036542892456055 s) (per-host) I0424 15:12:18.735409 139080469821184 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.410698s. I0424 15:12:26.025252 139080469821184 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.700525s. I0424 15:12:26.025552 139077902898944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:12:26.025677 139077902898944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:12:26.025729 139077902898944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:12:26.027400 139077902898944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 15:12:26.027577 139207004628800 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. W0424 15:12:26.027712 139207004628800 checkpoint_manager.py:1458] Waiting for previous save to complete took 24.841366 seconds. If this number is high, consider checkpointing less frequently. I0424 15:12:26.029577 139207004628800 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 10 I0424 15:12:26.033015 139207004628800 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/10. I0424 15:12:27.049784 139207004628800 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array. I0424 15:12:27.049892 139207004628800 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 I0424 15:12:27.107625 139207004628800 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.058954s I0424 15:12:27.107794 139207004628800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.372 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.6502835750579834 s) (per-host) I0424 15:12:27.107847 139207004628800 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.650344s (batch_requests_ready=0.587847s, total_serialization_initiated=0.062428s, others=0.000069s) I0424 15:12:27.107945 139207004628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.654396s (all_items=0.000015s, per_item={'items': '0.00001478'}, temp_paths=0.654382) I0424 15:12:27.108836 139076841338624 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:12:27.108994 139207004628800 async_checkpointer.py:561] Finished blocking save. Time taken: 1.079340s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/10. I0424 15:12:27.110958 139207004628800 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:12:27.111200 139077902898944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:12:27.111345 139207004628800 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777043521.1863198, 'wait_for_prev_duration_secs': 24.841366052627563, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777043546.0296223, 'checkpointer_blocking_duration_secs': 1.079507827758789, 'get_old_steps_start_time': 1777043547.1091516, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1777043521.1846004, 'checkpoint_manager_blocking_duration_secs': 25.92670965194702} I0424 15:12:27.111455 139207004628800 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0424 15:12:27.111857 139207004628800 metric_logger.py:185] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.959, Tokens/s/device: 12956.695, total_weights: 65536, loss: 7.908 I0424 15:12:27.118242 139207004628800 metric_logger.py:185] completed step: 10, seconds: 0.158, TFLOP/s/device: 85.903, Tokens/s/device: 12948.258, total_weights: 65536, loss: 7.568 I0424 15:12:27.276962 139207004628800 metric_logger.py:185] completed step: 11, seconds: 26.244, TFLOP/s/device: 0.518, Tokens/s/device: 78.037, total_weights: 65536, loss: 7.288 I0424 15:12:27.435068 139207004628800 metric_logger.py:185] completed step: 12, seconds: 0.006, TFLOP/s/device: 2453.879, Tokens/s/device: 369875.384, total_weights: 65536, loss: 7.032 I0424 15:12:27.593373 139207004628800 metric_logger.py:185] completed step: 13, seconds: 0.160, TFLOP/s/device: 84.876, Tokens/s/device: 12793.363, total_weights: 65536, loss: 6.815 I0424 15:12:27.751538 139207004628800 metric_logger.py:185] completed step: 14, seconds: 0.158, TFLOP/s/device: 85.978, Tokens/s/device: 12959.565, total_weights: 65536, loss: 6.635 I0424 15:12:27.909724 139207004628800 metric_logger.py:185] completed step: 15, seconds: 0.158, TFLOP/s/device: 85.840, Tokens/s/device: 12938.768, total_weights: 65536, loss: 6.492 I0424 15:12:28.067944 139207004628800 metric_logger.py:185] completed step: 16, seconds: 0.158, TFLOP/s/device: 85.801, Tokens/s/device: 12932.885, total_weights: 65536, loss: 6.380 I0424 15:12:28.226047 139207004628800 metric_logger.py:185] completed step: 17, seconds: 0.158, TFLOP/s/device: 86.043, Tokens/s/device: 12969.331, total_weights: 65536, loss: 6.293 I0424 15:12:28.384249 139207004628800 metric_logger.py:185] completed step: 18, seconds: 0.159, TFLOP/s/device: 85.625, Tokens/s/device: 12906.316, total_weights: 65536, loss: 6.223 I0424 15:12:28.542189 139207004628800 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0424 15:12:28.543092 139207004628800 checkpointing.py:798] Waited 0.0009226799011230469 seconds for step 19 to finish before starting checkpointing. I0424 15:12:28.545253 139207004628800 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:12:31.456674 139077919684352 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/10/items/array_metadatas/process_6 I0424 15:13:08.630832 139076841338624 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.459 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.17327952384949 s) (per-host) I0424 15:13:08.630957 139076841338624 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.522011s. I0424 15:13:16.652892 139076841338624 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.543929s. I0424 15:13:16.653145 139077902898944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:13:16.653216 139077902898944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:13:16.653274 139077902898944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:13:16.654901 139077902898944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0424 15:13:16.655123 139207004628800 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0424 15:13:16.655256 139207004628800 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.110014 seconds. If this number is high, consider checkpointing less frequently. I0424 15:13:16.656628 139207004628800 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 19 I0424 15:13:16.660212 139207004628800 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/19. I0424 15:13:17.202187 139207004628800 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array. I0424 15:13:17.202285 139207004628800 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 I0424 15:13:17.258490 139207004628800 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.057307s I0424 15:13:17.258659 139207004628800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.881 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3160686492919922 s) (per-host) I0424 15:13:17.258708 139207004628800 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.316127s (batch_requests_ready=0.255394s, total_serialization_initiated=0.060670s, others=0.000063s) I0424 15:13:17.258810 139207004628800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.320113s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.320102) I0424 15:13:17.259665 139077919684352 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0424 15:13:17.259822 139207004628800 async_checkpointer.py:561] Finished blocking save. Time taken: 0.603123s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/19. I0424 15:13:17.270597 139207004628800 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0424 15:13:17.270905 139077902898944 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 15:13:17.271064 139207004628800 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777043548.5452251, 'wait_for_prev_duration_secs': 48.11001396179199, 'time_between_consecutive_saves_sec': 2.5177865028381348, 'checkpointer_blocking_start_time': 1777043596.6566677, 'checkpointer_blocking_duration_secs': 0.6032602787017822, 'get_old_steps_start_time': 1777043597.259942, 'get_old_steps_duration_secs': 1.9073486328125e-05, 'checkpoint_manager_blocking_start_time': 1777043548.543392, 'checkpoint_manager_blocking_duration_secs': 48.7276406288147} I0424 15:13:17.271187 139207004628800 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0424 15:13:17.271229 139207004628800 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 15:13:22.540012 139080469821184 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260424_145744/linen_xpk_feat_nnx_set_defaults_true_20260424_145744_05_fp8/checkpoints/19/items/array_metadatas/process_6 I0424 15:13:58.567823 139077919684352 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.952 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.6251859664917 s) (per-host) I0424 15:13:58.567942 139077919684352 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.308168s. I0424 15:14:10.649667 139077919684352 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 53.389878s. I0424 15:14:10.649937 139077902898944 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 15:14:10.650053 139077902898944 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 15:14:10.650135 139077902898944 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0424 15:14:10.651337 139077902898944 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0424 15:14:10.651524 139207004628800 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0424 15:14:10.651670 139207004628800 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 15:14:10.652307 139207004628800 metric_logger.py:185] completed step: 19, seconds: 0.158, TFLOP/s/device: 86.093, Tokens/s/device: 12976.891, total_weights: 65536, loss: 6.167 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 15:14:22 UTC 2026 EXIT_CODE=0