MaxView

XPK Start: Sun Apr 19 22:06:06 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-19 22:06:30.346824: 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)
I0419 22:06:30.527052 139884318066496 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 22:06:39.566859 139884318066496 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 22:06:39.569205 139884318066496 distributed.py:172] Connecting to JAX distributed service on mt-05-fp8-jdkrh-slice-job-0-0.mt-05-fp8-jdkrh:8482
I0419 22:06:41.865244 139884318066496 max_utils.py:284] Jax distributed system initialized!
I0419 22:06:46.846958 139884318066496 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 22:06:46.847058 139884318066496 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 22:06:46.847096 139884318066496 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0419 22:06:46.847132 139884318066496 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 22:06:47.896193 139884318066496 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 22:06:47.896855 139884318066496 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 22:06:47.897055 139884318066496 checkpointing.py:688] Setting up checkpoint logger...
I0419 22:06:47.897110 139884318066496 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 22:06:47.897163 139884318066496 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 22:06:47.897547 139884318066496 base_pytree_checkpoint_handler.py:441] 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 0x7f38ba3eb680>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 22:06:52.169749 139884318066496 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0419 22:06:52.169978 139884318066496 checkpoint_manager.py:708] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3797309910>}, handler_registry=None
I0419 22:06:52.170212 139884318066496 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3797309910>` 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`.
I0419 22:06:52.170258 139884318066496 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3797197650>` 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`.
I0419 22:06:52.170293 139884318066496 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 0x7f3797309910>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f3797309910>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3797197650>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f3797197650>}).
I0419 22:06:52.170623 139884318066496 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 22:06:52.170694 139884318066496 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f24fc273740> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 22:06:54.184252 139884318066496 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints
I0419 22:06:54.588885 139884318066496 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_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f37973086e0>
I0419 22:06:54.589050 139884318066496 checkpointing.py:302] Checkpoint manager created!
I0419 22:06:55.672910 139884318066496 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 22:06:55.673017 139884318066496 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 22:06:56.064443 139884318066496 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 22:06:56.064532 139884318066496 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 22:06:56.080780 139884318066496 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 22:06:56.080843 139884318066496 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 22:06:56.133995 139884318066496 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 22:06:56.134074 139884318066496 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 22:06:56.150250 139884318066496 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 22:06:56.150310 139884318066496 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 22:06:56.166423 139884318066496 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 22:06:56.166488 139884318066496 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 22:06:56.191161 139884318066496 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 22:06:56.191228 139884318066496 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 22:06:56.241957 139884318066496 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 22:06:56.242038 139884318066496 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 22:06:56.679614 139884318066496 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 22:06:56.679726 139884318066496 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 22:06:58.587203 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587325 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587393 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587434 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587472 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587510 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587542 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587572 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587602 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587632 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587661 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587691 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587719 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587748 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587776 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587811 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587841 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587870 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587898 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587925 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587952 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.587979 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588005 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588032 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588060 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588087 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588114 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588145 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588172 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588199 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588226 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588253 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588279 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588306 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588399 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588440 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588470 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588500 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588544 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588593 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588641 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588673 139884318066496 maxtext_utils.py:1821]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 22:06:58.588726 139884318066496 maxtext_utils.py:1821]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0419 22:06:58.588781 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 22:06:58.588823 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 22:06:58.588871 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0419 22:06:58.588917 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 22:06:58.588950 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 22:06:58.588997 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0419 22:06:58.589044 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   P('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0419 22:06:58.589083 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0419 22:06:58.589117 139884318066496 maxtext_utils.py:1821]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0419 22:06:58.589163 139884318066496 maxtext_utils.py:1821]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)

I0419 22:06:58.589207 139884318066496 maxtext_utils.py:1821]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0419 22:07:00.214017 139884318066496 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 22:07:00.214114 139884318066496 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 22:07:00.229362 139884318066496 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 22:07:00.229419 139884318066496 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 22:07:13.857553 139884318066496 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.
I0419 22:07:13.858424 139884318066496 metric_logger.py:301] number parameters: 1.105 billion
I0419 22:07:28.925310 139884318066496 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0419 22:07:29.097482 139884318066496 checkpointing.py:798] Waited 0.17215251922607422 seconds for step 0 to finish before starting checkpointing.
I0419 22:07:29.099792 139884318066496 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 22:07:29.101737 139884318066496 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 0
I0419 22:07:29.103169 139884318066496 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/0.
I0419 22:07:30.291501 139884318066496 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 22:07:30.293855 139884318066496 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 22:07:30.293978 139884318066496 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
I0419 22:07:30.699257 139884318066496 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.407877s
I0419 22:07:30.699445 139884318066496 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 3.706 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.41632509231567383 s) (per-host)
I0419 22:07:30.699496 139884318066496 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.416385s (batch_requests_ready=0.003643s, total_serialization_initiated=0.412674s, others=0.000067s)
I0419 22:07:30.699595 139884318066496 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.420398s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.420380)
I0419 22:07:30.700358 139884318066496 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.60 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/0.
I0419 22:07:30.700707 139757304698624 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 22:27:30.700670
I0419 22:07:30.716753 139884318066496 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 22:07:30.717067 139755729798912 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 22:07:30.717223 139884318066496 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776636449.0997748, 'wait_for_prev_duration_secs': 6.532669067382812e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776636449.1017756, 'checkpointer_blocking_duration_secs': 1.599083662033081, 'get_old_steps_start_time': 1776636450.7008853, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776636449.0980814, 'checkpoint_manager_blocking_duration_secs': 1.6191027164459229}
I0419 22:07:30.717349 139884318066496 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0419 22:07:30.717402 139884318066496 max_utils.py:750] 
Memstats: After params initialized:
I0419 22:07:30.717454 139884318066496 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_16(process=4,(0,4,0,0))
I0419 22:07:30.717488 139884318066496 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_17(process=4,(1,4,0,0))
I0419 22:07:30.717514 139884318066496 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_20(process=4,(0,5,0,0))
I0419 22:07:30.717541 139884318066496 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_21(process=4,(1,5,0,0))
I0419 22:07:31.028622 139884318066496 metric_logger.py:196] completed step: 0, seconds: 15.067, TFLOP/s/device: 0.902, Tokens/s/device: 135.928, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52776.805
I0419 22:07:31.225169 139884318066496 metric_logger.py:196] completed step: 1, seconds: 2.101, TFLOP/s/device: 6.466, Tokens/s/device: 974.669, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52761.840
I0419 22:07:32.061665 139884318066496 metric_logger.py:196] completed step: 2, seconds: 0.039, TFLOP/s/device: 349.643, Tokens/s/device: 52702.007, total_weights: 65536, loss: 10.816, lm_loss: 10.816, perplexity: 49832.398
I0419 22:07:32.220037 139884318066496 metric_logger.py:196] completed step: 3, seconds: 0.837, TFLOP/s/device: 16.236, Tokens/s/device: 2447.240, total_weights: 65536, loss: 10.431, lm_loss: 10.431, perplexity: 33901.820
I0419 22:07:32.537569 139884318066496 metric_logger.py:196] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.765, Tokens/s/device: 12475.254, total_weights: 65536, loss: 9.992, lm_loss: 9.992, perplexity: 21847.639
I0419 22:07:32.544048 139884318066496 metric_logger.py:196] completed step: 5, seconds: 0.158, TFLOP/s/device: 85.890, Tokens/s/device: 12946.293, total_weights: 65536, loss: 9.549, lm_loss: 9.549, perplexity: 14035.006
I0419 22:07:38.127371    2851 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 22:07:42.801949 139755738191616 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/0/items/array_metadatas/process_4
I0419 22:07:54.738780 139757304698624 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 64.597 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 24.45561909675598 s) (per-host)
I0419 22:07:54.738910 139757304698624 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 24.038093s.
I0419 22:07:59.461738 139884318066496 metric_logger.py:196] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.712, Tokens/s/device: 6438.085, total_weights: 65536, loss: 9.111, lm_loss: 9.111, perplexity: 9058.688
I0419 22:07:59.620244 139884318066496 metric_logger.py:196] completed step: 7, seconds: 26.760, TFLOP/s/device: 0.508, Tokens/s/device: 76.532, total_weights: 65536, loss: 8.685, lm_loss: 8.685, perplexity: 5914.720
I0419 22:07:59.778567 139884318066496 metric_logger.py:196] completed step: 8, seconds: 0.164, TFLOP/s/device: 82.858, Tokens/s/device: 12489.328, total_weights: 65536, loss: 8.281, lm_loss: 8.281, perplexity: 3950.010
I0419 22:07:59.783723 139884318066496 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0419 22:08:00.095485 139884318066496 checkpointing.py:798] Waited 0.31172823905944824 seconds for step 10 to finish before starting checkpointing.
I0419 22:08:00.098279 139884318066496 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 22:08:06.585425 139757304698624 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 35.884594s.
I0419 22:08:06.585695 139755729798912 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 22:08:06.585820 139755729798912 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 22:08:06.585871 139755729798912 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 22:08:06.587763 139755729798912 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 22:08:06.587942 139884318066496 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0419 22:08:06.588076 139884318066496 checkpoint_manager.py:1452] Waiting for previous save to complete took 6.489795 seconds. If this number is high, consider checkpointing less frequently.
I0419 22:08:06.589899 139884318066496 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 10
I0419 22:08:06.591952 139884318066496 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/10.
I0419 22:08:07.736996 139884318066496 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 22:08:07.737088 139884318066496 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
I0419 22:08:07.788445 139884318066496 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.052385s
I0419 22:08:07.788619 139884318066496 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 26.234 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0588076114654541 s) (per-host)
I0419 22:08:07.788668 139884318066496 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.058867s (batch_requests_ready=0.003320s, total_serialization_initiated=0.055480s, others=0.000067s)
I0419 22:08:07.788769 139884318066496 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.062826s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.062809)
I0419 22:08:07.789461 139884318066496 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.20 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/10.
I0419 22:08:07.789817 139755729798912 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 22:28:07.789779
I0419 22:08:07.791813 139884318066496 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 22:08:07.792087 139753043248896 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 22:08:07.792241 139884318066496 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776636480.0982502, 'wait_for_prev_duration_secs': 6.489795207977295, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776636486.589938, 'checkpointer_blocking_duration_secs': 1.2000272274017334, 'get_old_steps_start_time': 1776636487.789987, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776636480.0962522, 'checkpoint_manager_blocking_duration_secs': 7.695955991744995}
I0419 22:08:07.792371 139884318066496 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0419 22:08:07.793067 139884318066496 metric_logger.py:196] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.821, Tokens/s/device: 12935.826, total_weights: 65536, loss: 7.908, lm_loss: 7.908, perplexity: 2720.116
I0419 22:08:07.800196 139884318066496 metric_logger.py:196] completed step: 10, seconds: 0.158, TFLOP/s/device: 85.724, Tokens/s/device: 12921.217, total_weights: 65536, loss: 7.568, lm_loss: 7.568, perplexity: 1936.198
I0419 22:08:07.958708 139884318066496 metric_logger.py:196] completed step: 11, seconds: 8.014, TFLOP/s/device: 1.695, Tokens/s/device: 255.542, total_weights: 65536, loss: 7.288, lm_loss: 7.288, perplexity: 1463.272
I0419 22:08:08.116961 139884318066496 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2114.399, Tokens/s/device: 318705.260, total_weights: 65536, loss: 7.032, lm_loss: 7.032, perplexity: 1132.061
I0419 22:08:08.275280 139884318066496 metric_logger.py:196] completed step: 13, seconds: 0.159, TFLOP/s/device: 85.314, Tokens/s/device: 12859.475, total_weights: 65536, loss: 6.815, lm_loss: 6.815, perplexity: 911.031
I0419 22:08:08.838699 139884318066496 metric_logger.py:196] completed step: 14, seconds: 0.158, TFLOP/s/device: 85.854, Tokens/s/device: 12940.894, total_weights: 65536, loss: 6.635, lm_loss: 6.635, perplexity: 761.080
I0419 22:08:08.997172 139884318066496 metric_logger.py:196] completed step: 15, seconds: 0.558, TFLOP/s/device: 24.336, Tokens/s/device: 3668.180, total_weights: 65536, loss: 6.492, lm_loss: 6.492, perplexity: 659.915
I0419 22:08:09.155529 139884318066496 metric_logger.py:196] completed step: 16, seconds: 0.163, TFLOP/s/device: 83.154, Tokens/s/device: 12533.813, total_weights: 65536, loss: 6.380, lm_loss: 6.380, perplexity: 589.948
I0419 22:08:09.313760 139884318066496 metric_logger.py:196] completed step: 17, seconds: 0.158, TFLOP/s/device: 85.747, Tokens/s/device: 12924.724, total_weights: 65536, loss: 6.293, lm_loss: 6.293, perplexity: 540.941
I0419 22:08:09.472198 139884318066496 metric_logger.py:196] completed step: 18, seconds: 0.158, TFLOP/s/device: 85.832, Tokens/s/device: 12937.461, total_weights: 65536, loss: 6.223, lm_loss: 6.223, perplexity: 504.108
I0419 22:08:09.629990 139884318066496 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0419 22:08:09.631012 139884318066496 checkpointing.py:798] Waited 0.001043081283569336 seconds for step 19 to finish before starting checkpointing.
I0419 22:08:09.633132 139884318066496 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 22:08:17.594070 139754133776128 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/10/items/array_metadatas/process_4
I0419 22:08:53.353439 139755729798912 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 34.626 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.623586654663086 s) (per-host)
I0419 22:08:53.353569 139755729798912 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.563631s.
I0419 22:09:02.055244 139755729798912 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 54.265293s.
I0419 22:09:02.055579 139753043248896 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 22:09:02.055698 139753043248896 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 22:09:02.055745 139753043248896 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 22:09:02.057432 139753043248896 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 22:09:02.057616 139884318066496 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0419 22:09:02.057741 139884318066496 checkpoint_manager.py:1452] Waiting for previous save to complete took 52.424615 seconds. If this number is high, consider checkpointing less frequently.
I0419 22:09:02.059296 139884318066496 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 19
I0419 22:09:02.061526 139884318066496 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/19.
I0419 22:09:02.778877 139884318066496 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 22:09:02.778969 139884318066496 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
I0419 22:09:02.816738 139884318066496 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038776s
I0419 22:09:02.816875 139884318066496 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 34.051 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04530668258666992 s) (per-host)
I0419 22:09:02.816922 139884318066496 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045364s (batch_requests_ready=0.003313s, total_serialization_initiated=0.041989s, others=0.000062s)
I0419 22:09:02.817002 139884318066496 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049476s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.049464)
I0419 22:09:02.817700 139884318066496 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/19.
I0419 22:09:02.818024 139753043248896 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 22:29:02.817988
I0419 22:09:02.822954 139884318066496 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 22:09:02.823239 139754133776128 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0419 22:09:02.823407 139884318066496 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776636489.6331046, 'wait_for_prev_duration_secs': 52.424614667892456, 'time_between_consecutive_saves_sec': 3.045297384262085, 'checkpointer_blocking_start_time': 1776636542.05935, 'checkpointer_blocking_duration_secs': 0.7588133811950684, 'get_old_steps_start_time': 1776636542.8181825, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776636489.6312842, 'checkpoint_manager_blocking_duration_secs': 53.19209003448486}
I0419 22:09:02.823518 139884318066496 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0419 22:09:02.823561 139884318066496 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 22:09:10.951644 139754116990720 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525/linen_xpk_feat_nnx_trainstate_and_training_loop_20260419_214525_05_fp8/checkpoints/19/items/array_metadatas/process_4
I0419 22:09:46.777021 139753043248896 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 35.899 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.0054132938385 s) (per-host)
I0419 22:09:46.777145 139753043248896 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.959011s.
I0419 22:09:54.528488 139753043248896 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 51.710339s.
I0419 22:09:54.528743 139754133776128 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 22:09:54.528867 139754133776128 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0419 22:09:54.528908 139754133776128 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 22:09:54.530508 139754133776128 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 22:09:54.530689 139884318066496 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0419 22:09:54.530834 139884318066496 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 22:09:54.531841 139884318066496 metric_logger.py:196] completed step: 19, seconds: 0.158, TFLOP/s/device: 85.764, Tokens/s/device: 12927.253, total_weights: 65536, loss: 6.167, lm_loss: 6.167, perplexity: 476.947
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 22:10:06 UTC 2026
EXIT_CODE=0