XPK Start: Tue Apr 21 01:03:48 UTC 2026
2026-04-21 01:03:52.293074: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1776733432.305495 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776733432.309271 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776733432.320629 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776733432.320648 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776733432.320650 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776733432.320652 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-21 01:04:11.417390: 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)
I0421 01:04:11.925359 134491761944384 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-21 01:04:20,965:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0421 01:04:20.965377 134491761944384 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-21 01:04:20,967:jax._src.distributed:157: Connecting to JAX distributed service on mt-05-fp8-kwmgs-slice-job-0-0.mt-05-fp8-kwmgs:8482
I0421 01:04:20.967814 134491761944384 distributed.py:157] Connecting to JAX distributed service on mt-05-fp8-kwmgs-slice-job-0-0.mt-05-fp8-kwmgs:8482
I0421 01:04:22.444115 134491761944384 max_utils.py:284] Jax distributed system initialized!
I0421 01:04:29.545162 134491761944384 max_utils.py:800] System Information: Jax Version: 0.8.1
I0421 01:04:29.545264 134491761944384 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0421 01:04:29.545306 134491761944384 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709
I0421 01:04:29.545342 134491761944384 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 01:04:30.164262 134491761944384 maxtext_utils.py:1398] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 01:04:30.164580 134491761944384 checkpointing.py:677] Setting up checkpoint logger...
I0421 01:04:30.164640 134491761944384 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0421 01:04:30.164684 134491761944384 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0421 01:04:30.165028 134491761944384 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a5150d71370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 01:04:33.117168 134491761944384 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0421 01:04:33.117537 134491761944384 checkpoint_manager.py:709] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3de583a030>}, handler_registry=None
I0421 01:04:33.117774 134491761944384 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3de583a030>` 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`.
I0421 01:04:33.117822 134491761944384 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3de65aec00>` 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`.
I0421 01:04:33.117861 134491761944384 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 0x7a3de583a030>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a3de583a030>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3de65aec00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a3de65aec00>}).
I0421 01:04:33.118360 134491761944384 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0421 01:04:33.118451 134491761944384 async_checkpointer.py:177] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a3de6238040> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0421 01:04:34.229017 134491761944384 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints
I0421 01:04:34.622460 134491761944384 checkpoint_manager.py:929] [process=1][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_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a500879bcb0>
I0421 01:04:34.622630 134491761944384 checkpointing.py:301] Checkpoint manager created!
I0421 01:04:35.692895 134491761944384 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 01:04:35.693018 134491761944384 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 01:04:36.066143 134491761944384 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 01:04:36.066240 134491761944384 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 01:04:36.082044 134491761944384 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 01:04:36.082104 134491761944384 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 01:04:36.140876 134491761944384 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 01:04:36.140961 134491761944384 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 01:04:36.156931 134491761944384 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 01:04:36.156998 134491761944384 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 01:04:36.172799 134491761944384 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 01:04:36.172861 134491761944384 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 01:04:36.196606 134491761944384 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 01:04:36.196678 134491761944384 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 01:04:36.256871 134491761944384 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 01:04:36.256955 134491761944384 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 01:04:36.719732 134491761944384 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 01:04:36.719848 134491761944384 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0421 01:04:38.743701 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.743826 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.743876 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.743930 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.743968 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744001 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744033 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744064 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744094 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744124 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744153 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744184 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744214 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744244 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744273 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744302 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744331 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744361 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744389 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744433 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744463 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744490 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744518 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744545 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744573 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744601 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744628 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744655 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744681 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744708 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744735 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744761 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744788 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744814 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744841 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744868 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744898 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744925 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744952 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.744978 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.745005 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history
Shape: float32[16,1024]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.745031 134491761944384 maxtext_utils.py:1501] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale
Shape: float32[16,1]
Logical: PartitionSpec()
Physical: ()
I0421 01:04:38.745079 134491761944384 maxtext_utils.py:1501] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: PartitionSpec('norm',)
Physical: (None,)
I0421 01:04:38.745142 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: PartitionSpec('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0421 01:04:38.745181 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: PartitionSpec('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0421 01:04:38.745228 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: PartitionSpec('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0421 01:04:38.745271 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: PartitionSpec('norm', 'layers')
Physical: (None, None)
I0421 01:04:38.745304 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: PartitionSpec('norm', 'layers')
Physical: (None, None)
I0421 01:04:38.745351 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/key/kernel
Shape: float32[2048,16,16,128]
Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0421 01:04:38.745421 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: PartitionSpec('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0421 01:04:38.745465 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0421 01:04:38.745501 134491761944384 maxtext_utils.py:1501] params/params/decoder/layers/self_attention/value/kernel
Shape: float32[2048,16,16,128]
Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0421 01:04:38.745549 134491761944384 maxtext_utils.py:1501] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: PartitionSpec('embed', 'vocab')
Physical: ('fsdp', None)
I0421 01:04:38.745592 134491761944384 maxtext_utils.py:1501] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: PartitionSpec('vocab', 'embed')
Physical: (None, 'fsdp')
I0421 01:04:40.511505 134491761944384 train.py:156] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 01:04:40.511602 134491761944384 train.py:156] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 01:04:40.526652 134491761944384 train.py:163] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 01:04:40.526712 134491761944384 train.py:163] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 01:04:53.479868 134491761944384 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.
I0421 01:04:53.480835 134491761944384 metric_logger.py:289] number parameters: 1.105 billion
I0421 01:05:08.087339 134491761944384 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0421 01:05:08.354638 134491761944384 checkpointing.py:776] Waited 0.26728057861328125 seconds for step 0 to finish before starting checkpointing.
I0421 01:05:08.357353 134491761944384 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 01:05:08.359317 134491761944384 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 0
I0421 01:05:08.362063 134491761944384 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/0.
I0421 01:05:09.847257 134491761944384 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 01:05:09.848316 134491761944384 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0421 01:05:09.848376 134491761944384 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
I0421 01:05:10.216958 134491761944384 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.369826s
I0421 01:05:10.217127 134491761944384 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.446 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 1.0665411949157715 s) (per-host)
I0421 01:05:10.217177 134491761944384 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 1.066599s (batch_requests_ready=0.691461s, total_serialization_initiated=0.375073s, others=0.000065s)
I0421 01:05:10.217281 134491761944384 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.070696s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=1.070676)
I0421 01:05:10.218262 134368882304768 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0421 01:05:10.218446 134491761944384 async_checkpointer.py:561] Finished blocking save. Time taken: 1.859056s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/0.
I0421 01:05:10.220293 134491761944384 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 01:05:10.220535 134361936553728 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0421 01:05:10.220663 134491761944384 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733508.3573349, 'wait_for_prev_duration_secs': 7.82012939453125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733508.3593576, 'checkpointer_blocking_duration_secs': 1.8591952323913574, 'get_old_steps_start_time': 1776733510.2185729, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776733508.3555343, 'checkpoint_manager_blocking_duration_secs': 1.8650915622711182}
I0421 01:05:10.220827 134491761944384 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0421 01:05:10.220883 134491761944384 max_utils.py:750]
Memstats: After params initialized:
I0421 01:05:10.220932 134491761944384 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_2(process=1,(2,0,0,0))
I0421 01:05:10.220968 134491761944384 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_3(process=1,(3,0,0,0))
I0421 01:05:10.220996 134491761944384 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_6(process=1,(2,1,0,0))
I0421 01:05:10.221020 134491761944384 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_7(process=1,(3,1,0,0))
I0421 01:05:10.483204 134491761944384 metric_logger.py:185] completed step: 0, seconds: 14.606, TFLOP/s/device: 0.930, Tokens/s/device: 140.213, total_weights: 65536, loss: 10.874
I0421 01:05:10.665257 134491761944384 metric_logger.py:185] completed step: 1, seconds: 2.394, TFLOP/s/device: 5.676, Tokens/s/device: 855.505, total_weights: 65536, loss: 10.874
I0421 01:05:11.076025 134491761944384 metric_logger.py:185] completed step: 2, seconds: 0.025, TFLOP/s/device: 547.449, Tokens/s/device: 82517.426, total_weights: 65536, loss: 10.816
I0421 01:05:11.234155 134491761944384 metric_logger.py:185] completed step: 3, seconds: 0.411, TFLOP/s/device: 33.030, Tokens/s/device: 4978.692, total_weights: 65536, loss: 10.431
I0421 01:05:11.551558 134491761944384 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.157, Tokens/s/device: 12534.273, total_weights: 65536, loss: 9.992
I0421 01:05:11.557825 134491761944384 metric_logger.py:185] completed step: 5, seconds: 0.158, TFLOP/s/device: 85.924, Tokens/s/device: 12951.369, total_weights: 65536, loss: 9.549
I0421 01:05:32.230993 134491761944384 metric_logger.py:185] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.754, Tokens/s/device: 6444.345, total_weights: 65536, loss: 9.111
I0421 01:05:32.389210 134491761944384 metric_logger.py:185] completed step: 7, seconds: 20.516, TFLOP/s/device: 0.662, Tokens/s/device: 99.822, total_weights: 65536, loss: 8.685
I0421 01:05:32.547502 134491761944384 metric_logger.py:185] completed step: 8, seconds: 0.163, TFLOP/s/device: 83.501, Tokens/s/device: 12586.192, total_weights: 65536, loss: 8.281
I0421 01:05:32.552586 134491761944384 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0421 01:05:32.864268 134491761944384 checkpointing.py:776] Waited 0.3116493225097656 seconds for step 10 to finish before starting checkpointing.
I0421 01:05:32.867048 134491761944384 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 01:05:33.764613 2683 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 01:05:35.620198 134368890697472 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/0/items/array_metadatas/process_1
I0421 01:05:50.810943 134368882304768 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 37.920 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.660295486450195 s) (per-host)
I0421 01:05:50.811062 134368882304768 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.592688s.
I0421 01:05:58.549211 134368882304768 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 48.330820s.
I0421 01:05:58.549502 134361936553728 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 01:05:58.549620 134361936553728 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0421 01:05:58.549674 134361936553728 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0421 01:05:58.551312 134361936553728 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0421 01:05:58.551497 134491761944384 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0421 01:05:58.551627 134491761944384 checkpoint_manager.py:1458] Waiting for previous save to complete took 25.684584 seconds. If this number is high, consider checkpointing less frequently.
I0421 01:05:58.553565 134491761944384 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 10
I0421 01:05:58.556993 134491761944384 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/10.
I0421 01:05:59.093441 134491761944384 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0421 01:05:59.093557 134491761944384 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
I0421 01:05:59.156429 134491761944384 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.064004s
I0421 01:05:59.156593 134491761944384 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.835 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31908440589904785 s) (per-host)
I0421 01:05:59.156642 134491761944384 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.319142s (batch_requests_ready=0.251564s, total_serialization_initiated=0.067513s, others=0.000066s)
I0421 01:05:59.156738 134491761944384 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.323179s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.323164)
I0421 01:05:59.157602 134368890697472 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0421 01:05:59.157756 134491761944384 async_checkpointer.py:561] Finished blocking save. Time taken: 0.604119s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/10.
I0421 01:05:59.159629 134491761944384 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 01:05:59.159927 134361936553728 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0421 01:05:59.160060 134491761944384 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733532.8670187, 'wait_for_prev_duration_secs': 25.684584140777588, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776733558.5536044, 'checkpointer_blocking_duration_secs': 0.6042587757110596, 'get_old_steps_start_time': 1776733559.1578832, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776733532.865192, 'checkpoint_manager_blocking_duration_secs': 26.294833660125732}
I0421 01:05:59.160236 134491761944384 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0421 01:05:59.160602 134491761944384 metric_logger.py:185] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.876, Tokens/s/device: 12944.166, total_weights: 65536, loss: 7.908
I0421 01:05:59.167113 134491761944384 metric_logger.py:185] completed step: 10, seconds: 0.158, TFLOP/s/device: 85.873, Tokens/s/device: 12943.757, total_weights: 65536, loss: 7.568
I0421 01:05:59.325665 134491761944384 metric_logger.py:185] completed step: 11, seconds: 26.613, TFLOP/s/device: 0.511, Tokens/s/device: 76.956, total_weights: 65536, loss: 7.288
I0421 01:05:59.483928 134491761944384 metric_logger.py:185] completed step: 12, seconds: 0.006, TFLOP/s/device: 2339.785, Tokens/s/device: 352677.803, total_weights: 65536, loss: 7.032
I0421 01:05:59.642014 134491761944384 metric_logger.py:185] completed step: 13, seconds: 0.160, TFLOP/s/device: 85.160, Tokens/s/device: 12836.182, total_weights: 65536, loss: 6.815
I0421 01:05:59.800252 134491761944384 metric_logger.py:185] completed step: 14, seconds: 0.158, TFLOP/s/device: 85.801, Tokens/s/device: 12932.804, total_weights: 65536, loss: 6.635
I0421 01:05:59.958416 134491761944384 metric_logger.py:185] completed step: 15, seconds: 0.158, TFLOP/s/device: 85.938, Tokens/s/device: 12953.581, total_weights: 65536, loss: 6.492
I0421 01:06:00.116539 134491761944384 metric_logger.py:185] completed step: 16, seconds: 0.158, TFLOP/s/device: 85.949, Tokens/s/device: 12955.138, total_weights: 65536, loss: 6.380
I0421 01:06:00.274785 134491761944384 metric_logger.py:185] completed step: 17, seconds: 0.158, TFLOP/s/device: 85.913, Tokens/s/device: 12949.813, total_weights: 65536, loss: 6.293
I0421 01:06:00.432933 134491761944384 metric_logger.py:185] completed step: 18, seconds: 0.158, TFLOP/s/device: 85.883, Tokens/s/device: 12945.229, total_weights: 65536, loss: 6.223
I0421 01:06:00.590941 134491761944384 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0421 01:06:00.591879 134491761944384 checkpointing.py:776] Waited 0.0009567737579345703 seconds for step 19 to finish before starting checkpointing.
I0421 01:06:00.593913 134491761944384 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 01:06:03.926673 134362448246528 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/10/items/array_metadatas/process_1
I0421 01:06:40.248553 134368890697472 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 38.148 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.410995960235596 s) (per-host)
I0421 01:06:40.248675 134368890697472 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.090968s.
I0421 01:06:50.366649 134368890697472 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 51.208927s.
I0421 01:06:50.366952 134361936553728 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 01:06:50.367071 134361936553728 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0421 01:06:50.367118 134361936553728 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0421 01:06:50.368885 134361936553728 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0421 01:06:50.369086 134491761944384 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0421 01:06:50.369223 134491761944384 checkpoint_manager.py:1458] Waiting for previous save to complete took 49.775323 seconds. If this number is high, consider checkpointing less frequently.
I0421 01:06:50.371089 134491761944384 checkpoint_manager.py:1518] [process=1] Saving checkpoint at step 19
I0421 01:06:50.374665 134491761944384 async_checkpointer.py:452] [process=1] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/19.
I0421 01:06:50.910393 134491761944384 jax_array_handlers.py:358] Scheduling D2H of 81 prioritized jax.Array.
I0421 01:06:50.910507 134491761944384 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
I0421 01:06:50.971938 134491761944384 base_pytree_checkpoint_handler.py:153] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.062489s
I0421 01:06:50.972099 134491761944384 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.792 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3219606876373291 s) (per-host)
I0421 01:06:50.972146 134491761944384 base_pytree_checkpoint_handler.py:737] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.322017s (batch_requests_ready=0.256068s, total_serialization_initiated=0.065888s, others=0.000061s)
I0421 01:06:50.972244 134491761944384 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.326064s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.326054)
I0421 01:06:50.973068 134362448246528 async_checkpointer.py:79] [process=1][thread=async_save] Background save thread started.
I0421 01:06:50.973228 134491761944384 async_checkpointer.py:561] Finished blocking save. Time taken: 0.602061s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/19.
I0421 01:06:50.999935 134491761944384 checkpoint_manager.py:1566] [process=1][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 01:06:51.000257 134361936553728 async_checkpointer.py:265] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0421 01:06:51.000430 134491761944384 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776733560.5938828, 'wait_for_prev_duration_secs': 49.77532339096069, 'time_between_consecutive_saves_sec': 2.042530059814453, 'checkpointer_blocking_start_time': 1776733610.371135, 'checkpointer_blocking_duration_secs': 0.6022000312805176, 'get_old_steps_start_time': 1776733610.9733493, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1776733560.5921562, 'checkpoint_manager_blocking_duration_secs': 50.408231258392334}
I0421 01:06:51.000545 134491761944384 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0421 01:06:51.000588 134491761944384 checkpoint_manager.py:2024] [process=1][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 01:06:55.669014 134364654446336 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260421_005327/linen_xpk_test_pipeline_scan_nnx_20260421_005327_05_fp8/checkpoints/19/items/array_metadatas/process_1
I0421 01:07:33.224014 134362448246528 base_pytree_checkpoint_handler.py:129] [process=1] /jax/checkpoint/write/gbytes_per_sec: 37.106 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.57384157180786 s) (per-host)
I0421 01:07:33.224117 134362448246528 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.250937s.
I0421 01:07:40.182493 134362448246528 async_checkpointer.py:144] [process=1][thread=async_save] Background save thread done. Time taken: 49.209296s.
I0421 01:07:40.182772 134361936553728 async_checkpointer.py:273] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 01:07:40.182903 134361936553728 async_checkpointer.py:283] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0421 01:07:40.182954 134361936553728 checkpoint_manager.py:2133] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0421 01:07:40.184307 134361936553728 checkpoint_manager.py:2142] [process=1][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0421 01:07:40.184512 134491761944384 checkpoint_manager.py:2036] [process=1][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0421 01:07:40.184656 134491761944384 checkpoint_manager.py:2013] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 01:07:40.185268 134491761944384 metric_logger.py:185] completed step: 19, seconds: 0.158, TFLOP/s/device: 85.906, Tokens/s/device: 12948.749, total_weights: 65536, loss: 6.167
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Tue Apr 21 01:07:49 UTC 2026
EXIT_CODE=0