MaxView

XPK Start: Sun Apr 19 18:21:38 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 18:22:02.529123: 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 18:22:03.067255 134812552341312 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 18:22:12.108246 134812552341312 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 18:22:12.110850 134812552341312 distributed.py:172] Connecting to JAX distributed service on mt-05-fp8-gj4ai-slice-job-0-0.mt-05-fp8-gj4ai:8482
I0419 18:22:13.208966 134812552341312 max_utils.py:284] Jax distributed system initialized!
I0419 18:22:18.471045 134812552341312 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 18:22:18.471147 134812552341312 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 18:22:18.471185 134812552341312 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 18:22:18.471224 134812552341312 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 18:22:19.179442 134812552341312 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 18:22:19.179717 134812552341312 checkpointing.py:677] Setting up checkpoint logger...
I0419 18:22:19.179772 134812552341312 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 18:22:19.179816 134812552341312 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 18:22:19.180158 134812552341312 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 0x7a9bdd515c10>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 18:22:22.479495 134812552341312 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0419 18:22:22.479732 134812552341312 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a88501a34d0>}, handler_registry=None
I0419 18:22:22.479968 134812552341312 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a88501a34d0>` 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 18:22:22.480015 134812552341312 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a8850100710>` 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 18:22:22.480049 134812552341312 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 0x7a88501a34d0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a88501a34d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a8850100710>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a8850100710>}).
I0419 18:22:22.480432 134812552341312 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 18:22:22.480507 134812552341312 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a882818b740> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 18:22:23.605135 134812552341312 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints
I0419 18:22:24.451371 134812552341312 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_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a88501a15e0>
I0419 18:22:24.451596 134812552341312 checkpointing.py:301] Checkpoint manager created!
I0419 18:22:25.531133 134812552341312 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 18:22:25.531251 134812552341312 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:22:25.914638 134812552341312 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 18:22:25.914734 134812552341312 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:22:25.930888 134812552341312 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 18:22:25.930944 134812552341312 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 18:22:25.983343 134812552341312 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 18:22:25.983433 134812552341312 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:22:25.999684 134812552341312 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 18:22:25.999745 134812552341312 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:22:26.015836 134812552341312 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 18:22:26.015892 134812552341312 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:22:26.040363 134812552341312 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 18:22:26.040436 134812552341312 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 18:22:26.091852 134812552341312 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 18:22:26.091931 134812552341312 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 18:22:26.531354 134812552341312 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 18:22:26.531496 134812552341312 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 18:22:28.466403 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466545 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466593 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466632 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466666 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466699 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466729 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466759 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466788 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466817 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466845 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466873 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466901 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466928 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466956 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.466983 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467012 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467041 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467069 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467096 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467123 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467149 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467182 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467208 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467235 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467262 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467288 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467314 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467339 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467366 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467393 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467432 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467460 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467495 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467522 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467549 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467576 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467602 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467630 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467655 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467682 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history
    Shape:     float32[16,1024]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467708 134812552341312 maxtext_utils.py:1654]  params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale
    Shape:     float32[16,1]
    Logical:   P()
    Physical:  ()
I0419 18:22:28.467756 134812552341312 maxtext_utils.py:1654]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0419 18:22:28.467811 134812552341312 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 18:22:28.467849 134812552341312 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 18:22:28.467897 134812552341312 maxtext_utils.py:1654]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0419 18:22:28.467942 134812552341312 maxtext_utils.py:1654]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 18:22:28.467974 134812552341312 maxtext_utils.py:1654]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 18:22:28.468021 134812552341312 maxtext_utils.py:1654]  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 18:22:28.468069 134812552341312 maxtext_utils.py:1654]  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 18:22:28.468105 134812552341312 maxtext_utils.py:1654]  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 18:22:28.468137 134812552341312 maxtext_utils.py:1654]  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 18:22:28.468187 134812552341312 maxtext_utils.py:1654]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)

I0419 18:22:28.468231 134812552341312 maxtext_utils.py:1654]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0419 18:22:30.102706 134812552341312 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 18:22:30.102799 134812552341312 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 18:22:30.117988 134812552341312 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 18:22:30.118051 134812552341312 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 18:22:43.781503 134812552341312 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 18:22:43.782392 134812552341312 metric_logger.py:301] number parameters: 1.105 billion
I0419 18:22:58.701583 134812552341312 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0419 18:22:59.026615 134812552341312 checkpointing.py:776] Waited 0.3250093460083008 seconds for step 0 to finish before starting checkpointing.
I0419 18:22:59.029251 134812552341312 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 18:22:59.031246 134812552341312 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0419 18:22:59.032649 134812552341312 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/0.
I0419 18:22:59.986831 134812552341312 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 18:22:59.987796 134812552341312 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 18:22:59.987853 134812552341312 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 18:23:00.405555 134812552341312 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.418846s
I0419 18:23:00.405735 134812552341312 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 3.608 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4276137351989746 s) (per-host)
I0419 18:23:00.405787 134812552341312 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.427678s (batch_requests_ready=0.003917s, total_serialization_initiated=0.423689s, others=0.000072s)
I0419 18:23:00.405894 134812552341312 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.431830s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.431812)
I0419 18:23:00.406746 134812552341312 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.38 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/0.
I0419 18:23:00.407054 134684078290688 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:43:00.407016
I0419 18:23:00.409034 134812552341312 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:23:00.409309 134683589338880 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:23:00.409491 134812552341312 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776622979.0292318, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776622979.0312877, 'checkpointer_blocking_duration_secs': 1.3759300708770752, 'get_old_steps_start_time': 1776622980.4072475, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776622979.027277, 'checkpoint_manager_blocking_duration_secs': 1.3821771144866943}
I0419 18:23:00.409594 134812552341312 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0419 18:23:00.409641 134812552341312 max_utils.py:750] 
Memstats: After params initialized:
I0419 18:23:00.409687 134812552341312 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_24(process=6,(0,6,0,0))
I0419 18:23:00.409718 134812552341312 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_25(process=6,(1,6,0,0))
I0419 18:23:00.409745 134812552341312 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_28(process=6,(0,7,0,0))
I0419 18:23:00.409768 134812552341312 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_29(process=6,(1,7,0,0))
I0419 18:23:00.720985 134812552341312 metric_logger.py:196] completed step: 0, seconds: 14.919, TFLOP/s/device: 0.911, Tokens/s/device: 137.274, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52776.805
I0419 18:23:00.904105 134812552341312 metric_logger.py:196] completed step: 1, seconds: 2.017, TFLOP/s/device: 6.735, Tokens/s/device: 1015.215, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52761.840
I0419 18:23:01.327475 134812552341312 metric_logger.py:196] completed step: 2, seconds: 0.026, TFLOP/s/device: 532.808, Tokens/s/device: 80310.576, total_weights: 65536, loss: 10.816, lm_loss: 10.816, perplexity: 49832.398
I0419 18:23:01.485915 134812552341312 metric_logger.py:196] completed step: 3, seconds: 0.423, TFLOP/s/device: 32.090, Tokens/s/device: 4836.988, total_weights: 65536, loss: 10.431, lm_loss: 10.431, perplexity: 33901.820
I0419 18:23:01.803240 134812552341312 metric_logger.py:196] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.647, Tokens/s/device: 12457.497, total_weights: 65536, loss: 9.992, lm_loss: 9.992, perplexity: 21847.639
I0419 18:23:01.809639 134812552341312 metric_logger.py:196] completed step: 5, seconds: 0.158, TFLOP/s/device: 85.813, Tokens/s/device: 12934.601, total_weights: 65536, loss: 9.549, lm_loss: 9.549, perplexity: 14035.006
I0419 18:23:04.081747    2869 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 18:23:06.630985 134687882909440 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/0/items/array_metadatas/process_6
I0419 18:23:21.465456 134684078290688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 73.521 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 21.487251043319702 s) (per-host)
I0419 18:23:21.465576 134684078290688 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 21.058400s.
I0419 18:23:25.913068 134812552341312 metric_logger.py:196] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.755, Tokens/s/device: 6444.548, total_weights: 65536, loss: 9.111, lm_loss: 9.111, perplexity: 9058.688
I0419 18:23:26.071532 134812552341312 metric_logger.py:196] completed step: 7, seconds: 23.946, TFLOP/s/device: 0.567, Tokens/s/device: 85.525, total_weights: 65536, loss: 8.685, lm_loss: 8.685, perplexity: 5914.720
I0419 18:23:26.230019 134812552341312 metric_logger.py:196] completed step: 8, seconds: 0.163, TFLOP/s/device: 83.113, Tokens/s/device: 12527.680, total_weights: 65536, loss: 8.281, lm_loss: 8.281, perplexity: 3950.010
I0419 18:23:26.235086 134812552341312 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0419 18:23:26.547486 134812552341312 checkpointing.py:776] Waited 0.3123619556427002 seconds for step 10 to finish before starting checkpointing.
I0419 18:23:26.550354 134812552341312 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:23:30.290239 134684078290688 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 29.883045s.
I0419 18:23:30.290531 134683589338880 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:23:30.290657 134683589338880 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:23:30.290707 134683589338880 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:23:30.292398 134683589338880 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 18:23:30.292562 134812552341312 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0419 18:23:30.292650 134812552341312 checkpoint_manager.py:1452] Waiting for previous save to complete took 3.742300 seconds. If this number is high, consider checkpointing less frequently.
I0419 18:23:30.294473 134812552341312 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0419 18:23:30.296427 134812552341312 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/10.
I0419 18:23:31.014039 134812552341312 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 18:23:31.014132 134812552341312 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 18:23:31.062368 134812552341312 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.049318s
I0419 18:23:31.062548 134812552341312 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 27.652 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05579018592834473 s) (per-host)
I0419 18:23:31.062598 134812552341312 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.055851s (batch_requests_ready=0.003323s, total_serialization_initiated=0.052460s, others=0.000068s)
I0419 18:23:31.062700 134812552341312 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.059915s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.059900)
I0419 18:23:31.063388 134812552341312 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/10.
I0419 18:23:31.063686 134683589338880 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:43:31.063653
I0419 18:23:31.065716 134812552341312 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:23:31.065917 134679839635200 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:23:31.066031 134812552341312 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776623006.550324, 'wait_for_prev_duration_secs': 3.742299795150757, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776623010.294513, 'checkpointer_blocking_duration_secs': 0.7693157196044922, 'get_old_steps_start_time': 1776623011.0638537, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776623006.5483205, 'checkpoint_manager_blocking_duration_secs': 4.5176756381988525}
I0419 18:23:31.066144 134812552341312 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0419 18:23:31.066824 134812552341312 metric_logger.py:196] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.823, Tokens/s/device: 12936.153, total_weights: 65536, loss: 7.908, lm_loss: 7.908, perplexity: 2720.116
I0419 18:23:31.073648 134812552341312 metric_logger.py:196] completed step: 10, seconds: 0.159, TFLOP/s/device: 85.708, Tokens/s/device: 12918.772, total_weights: 65536, loss: 7.568, lm_loss: 7.568, perplexity: 1936.198
I0419 18:23:31.232421 134812552341312 metric_logger.py:196] completed step: 11, seconds: 4.836, TFLOP/s/device: 2.809, Tokens/s/device: 423.456, total_weights: 65536, loss: 7.288, lm_loss: 7.288, perplexity: 1463.272
I0419 18:23:31.390531 134812552341312 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2228.860, Tokens/s/device: 335958.005, total_weights: 65536, loss: 7.032, lm_loss: 7.032, perplexity: 1132.061
I0419 18:23:31.949512 134812552341312 metric_logger.py:196] completed step: 13, seconds: 0.160, TFLOP/s/device: 85.152, Tokens/s/device: 12834.975, total_weights: 65536, loss: 6.815, lm_loss: 6.815, perplexity: 911.031
I0419 18:23:32.107986 134812552341312 metric_logger.py:196] completed step: 14, seconds: 0.554, TFLOP/s/device: 24.528, Tokens/s/device: 3697.171, total_weights: 65536, loss: 6.635, lm_loss: 6.635, perplexity: 761.080
I0419 18:23:32.266460 134812552341312 metric_logger.py:196] completed step: 15, seconds: 0.163, TFLOP/s/device: 83.276, Tokens/s/device: 12552.250, total_weights: 65536, loss: 6.492, lm_loss: 6.492, perplexity: 659.915
I0419 18:23:32.424610 134812552341312 metric_logger.py:196] completed step: 16, seconds: 0.159, TFLOP/s/device: 85.664, Tokens/s/device: 12912.256, total_weights: 65536, loss: 6.380, lm_loss: 6.380, perplexity: 589.948
I0419 18:23:32.582792 134812552341312 metric_logger.py:196] completed step: 17, seconds: 0.158, TFLOP/s/device: 85.779, Tokens/s/device: 12929.538, total_weights: 65536, loss: 6.293, lm_loss: 6.293, perplexity: 540.941
I0419 18:23:32.741100 134812552341312 metric_logger.py:196] completed step: 18, seconds: 0.158, TFLOP/s/device: 85.824, Tokens/s/device: 12936.398, total_weights: 65536, loss: 6.223, lm_loss: 6.223, perplexity: 504.108
I0419 18:23:32.898801 134812552341312 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0419 18:23:32.899832 134812552341312 checkpointing.py:776] Waited 0.0010478496551513672 seconds for step 19 to finish before starting checkpointing.
I0419 18:23:32.901898 134812552341312 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:23:38.097325 134680871413504 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/10/items/array_metadatas/process_6
I0419 18:24:13.664160 134683589338880 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.034 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.65735101699829 s) (per-host)
I0419 18:24:13.664283 134683589338880 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.600485s.
I0419 18:24:21.220284 134683589338880 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.156471s.
I0419 18:24:21.220560 134679839635200 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:24:21.220627 134679839635200 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:24:21.220674 134679839635200 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:24:21.222575 134679839635200 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0419 18:24:21.222787 134812552341312 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0419 18:24:21.222919 134812552341312 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.321026 seconds. If this number is high, consider checkpointing less frequently.
I0419 18:24:21.224387 134812552341312 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0419 18:24:21.226303 134812552341312 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/19.
I0419 18:24:21.938080 134812552341312 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array.
I0419 18:24:21.938177 134812552341312 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 18:24:21.987828 134812552341312 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.050670s
I0419 18:24:21.987974 134812552341312 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 27.035 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.05706477165222168 s) (per-host)
I0419 18:24:21.988021 134812552341312 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.057121s (batch_requests_ready=0.003303s, total_serialization_initiated=0.053757s, others=0.000061s)
I0419 18:24:21.988120 134812552341312 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.061296s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.061285)
I0419 18:24:21.988880 134812552341312 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/19.
I0419 18:24:21.989217 134679839635200 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 18:44:21.989180
I0419 18:24:21.991256 134812552341312 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 18:24:21.991482 134684078290688 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0419 18:24:21.991604 134812552341312 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776623012.90187, 'wait_for_prev_duration_secs': 48.32102560997009, 'time_between_consecutive_saves_sec': 2.6094019412994385, 'checkpointer_blocking_start_time': 1776623061.224446, 'checkpointer_blocking_duration_secs': 0.7649209499359131, 'get_old_steps_start_time': 1776623061.9893847, 'get_old_steps_duration_secs': 3.719329833984375e-05, 'checkpoint_manager_blocking_start_time': 1776623012.9001071, 'checkpoint_manager_blocking_duration_secs': 49.09146547317505}
I0419 18:24:21.991711 134812552341312 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0419 18:24:21.991753 134812552341312 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 18:24:27.662192 134686230542080 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260419_180002/linen_xpk_main_20260419_180002_05_fp8/checkpoints/19/items/array_metadatas/process_6
I0419 18:25:04.558990 134679839635200 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.059 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.62804055213928 s) (per-host)
I0419 18:25:04.559111 134679839635200 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.569782s.
I0419 18:25:12.402173 134679839635200 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.412820s.
I0419 18:25:12.402393 134684078290688 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 18:25:12.402473 134684078290688 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0419 18:25:12.402512 134684078290688 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0419 18:25:12.404230 134684078290688 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0419 18:25:12.404486 134812552341312 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0419 18:25:12.404633 134812552341312 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 18:25:12.405587 134812552341312 metric_logger.py:196] completed step: 19, seconds: 0.158, TFLOP/s/device: 85.868, Tokens/s/device: 12943.020, 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 18:25:23 UTC 2026
EXIT_CODE=0