MaxView

← Back to run

Log Summary

XPK Start: Mon Apr 20 19:56:56 UTC 2026
2026-04-20 19:57:00.196939: 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 00:00:1776715020.209341      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776715020.213069      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776715020.224272      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776715020.224297      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776715020.224299      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776715020.224301      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-20 19:57:19.504998: 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)
I0420 19:57:20.036060 137846163138368 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-20 19:57:29,077:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0420 19:57:29.077630 137846163138368 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-20 19:57:29,080:jax._src.distributed:157: Connecting to JAX distributed service on mt-11-optimizer-offload-nq5ks-slice-job-0-0.mt-11-optimizer-offload-nq5ks:8482
I0420 19:57:29.080070 137846163138368 distributed.py:157] Connecting to JAX distributed service on mt-11-optimizer-offload-nq5ks-slice-job-0-0.mt-11-optimizer-offload-nq5ks:8482
I0420 19:57:30.428174 137846163138368 max_utils.py:284] Jax distributed system initialized!
I0420 19:57:37.395380 137846163138368 max_utils.py:800] System Information: Jax Version: 0.8.1
I0420 19:57:37.395489 137846163138368 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0420 19:57:37.395531 137846163138368 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
I0420 19:57:37.395566 137846163138368 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 19:57:38.020086 137846163138368 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 19:57:38.020690 137846163138368 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 19:57:38.020895 137846163138368 checkpointing.py:688] Setting up checkpoint logger...
I0420 19:57:38.020954 137846163138368 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 19:57:38.021003 137846163138368 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0420 19:57:38.021349 137846163138368 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 0x7d5e43b710a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 19:57:40.925611 137846163138368 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 19:57:40.925870 137846163138368 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d4d0aba98b0>}, handler_registry=None
I0420 19:57:40.926111 137846163138368 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d4d0aba98b0>` 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`.
I0420 19:57:40.926160 137846163138368 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4d0aba9e80>` 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`.
I0420 19:57:40.926195 137846163138368 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 0x7d4d0aba98b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d4d0aba98b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4d0aba9e80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d4d0aba9e80>}).
I0420 19:57:40.926700 137846163138368 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0420 19:57:40.926781 137846163138368 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d4d0a30c180> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0420 19:57:42.344902 137846163138368 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints
I0420 19:57:42.782736 137846163138368 checkpoint_manager.py:929] [process=7][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d5d00489670>
I0420 19:57:42.782909 137846163138368 checkpointing.py:302] Checkpoint manager created!
I0420 19:57:43.679220 137846163138368 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 19:57:43.679327 137846163138368 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 19:57:44.050604 137846163138368 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 19:57:44.050713 137846163138368 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 19:57:44.066572 137846163138368 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 19:57:44.066632 137846163138368 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 19:57:44.089612 137846163138368 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 19:57:44.089697 137846163138368 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 19:57:44.105731 137846163138368 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 19:57:44.105795 137846163138368 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 19:57:44.121612 137846163138368 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 19:57:44.121683 137846163138368 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 19:57:44.146527 137846163138368 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 19:57:44.146599 137846163138368 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 19:57:44.166197 137846163138368 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 19:57:44.166264 137846163138368 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 19:57:44.356131 137846163138368 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 19:57:44.356242 137846163138368 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 19:57:46.196063 137846163138368 maxtext_utils.py:1740]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0420 19:57:46.196195 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0420 19:57:46.196256 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0420 19:57:46.196317 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   PartitionSpec('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0420 19:57:46.196369 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0420 19:57:46.196407 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0420 19:57:46.196459 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0420 19:57:46.196512 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   PartitionSpec('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0420 19:57:46.196555 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0420 19:57:46.196594 137846163138368 maxtext_utils.py:1740]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0420 19:57:46.196644 137846163138368 maxtext_utils.py:1740]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed', 'vocab')
    Physical:  ('fsdp', None)

I0420 19:57:46.196715 137846163138368 maxtext_utils.py:1740]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed')
    Physical:  (None, 'fsdp')
I0420 19:57:46.700679 137846163138368 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0420 19:57:46.700774 137846163138368 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0420 19:57:46.715621 137846163138368 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0420 19:57:46.715691 137846163138368 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0420 19:57:57.342814 137846163138368 max_utils.py:791] Total memory size: 1.3 GB, Output size: 0.1 GB, Temp size: 1.1 GB, Argument size: 0.1 GB, Host temp size: 0.0 GB.
I0420 19:57:57.343689 137846163138368 metric_logger.py:289] number parameters: 1.104 billion
I0420 19:58:08.308600 137846163138368 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 19:58:08.676005 137846163138368 checkpointing.py:798] Waited 0.36737895011901855 seconds for step 0 to finish before starting checkpointing.
I0420 19:58:08.678571 137846163138368 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 19:58:08.680414 137846163138368 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0
I0420 19:58:08.683274 137846163138368 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/0.
I0420 19:58:10.329914 137846163138368 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 19:58:10.330977 137846163138368 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0420 19:58:10.331040 137846163138368 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
I0420 19:58:10.575617 137846163138368 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245783s
I0420 19:58:10.575818 137846163138368 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.029 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5092134475708008 s) (per-host)
I0420 19:58:10.575871 137846163138368 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.509285s (batch_requests_ready=0.258545s, total_serialization_initiated=0.250661s, others=0.000079s)
I0420 19:58:10.575965 137846163138368 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.513336s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.513318)
I0420 19:58:10.576983 137727714518784 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started.
I0420 19:58:10.577162 137846163138368 async_checkpointer.py:561] Finished blocking save. Time taken: 1.896677s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/0.
I0420 19:58:10.615742 137846163138368 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 19:58:10.616060 137698455054080 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0420 19:58:10.616219 137846163138368 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776715088.6785529, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776715088.6804528, 'checkpointer_blocking_duration_secs': 1.8968164920806885, 'get_old_steps_start_time': 1776715090.577289, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776715088.676588, 'checkpoint_manager_blocking_duration_secs': 1.9395899772644043}
I0420 19:58:10.616326 137846163138368 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 19:58:10.616378 137846163138368 max_utils.py:750] 
Memstats: After params initialized:
I0420 19:58:10.616431 137846163138368 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_26(process=7,(2,6,0,0))
I0420 19:58:10.616465 137846163138368 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_27(process=7,(3,6,0,0))
I0420 19:58:10.616494 137846163138368 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_30(process=7,(2,7,0,0))
I0420 19:58:10.616520 137846163138368 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_31(process=7,(3,7,0,0))
I0420 19:58:10.877072 137846163138368 metric_logger.py:185] completed step: 0, seconds: 10.965, TFLOP/s/device: 1.239, Tokens/s/device: 186.780, total_weights: 65536, loss: 10.874
I0420 19:58:11.087154 137846163138368 metric_logger.py:185] completed step: 1, seconds: 2.567, TFLOP/s/device: 5.293, Tokens/s/device: 797.769, total_weights: 65536, loss: 10.874
I0420 19:58:11.920983 137846163138368 metric_logger.py:185] completed step: 2, seconds: 0.047, TFLOP/s/device: 290.677, Tokens/s/device: 43814.047, total_weights: 65536, loss: 10.263
I0420 19:58:12.084758 137846163138368 metric_logger.py:185] completed step: 3, seconds: 0.834, TFLOP/s/device: 16.287, Tokens/s/device: 2454.956, total_weights: 65536, loss: 9.735
I0420 19:58:12.412694 137846163138368 metric_logger.py:185] completed step: 4, seconds: 0.169, TFLOP/s/device: 80.512, Tokens/s/device: 12135.649, total_weights: 65536, loss: 9.278
I0420 19:58:12.418170 137846163138368 metric_logger.py:185] completed step: 5, seconds: 0.164, TFLOP/s/device: 82.991, Tokens/s/device: 12509.238, total_weights: 65536, loss: 8.892
I0420 19:58:30.765936    2597 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 19:58:30.947828 137846163138368 metric_logger.py:185] completed step: 6, seconds: 0.328, TFLOP/s/device: 41.402, Tokens/s/device: 6240.592, total_weights: 65536, loss: 8.592
I0420 19:58:31.112385 137846163138368 metric_logger.py:185] completed step: 7, seconds: 18.361, TFLOP/s/device: 0.740, Tokens/s/device: 111.540, total_weights: 65536, loss: 8.384
I0420 19:58:31.276049 137846163138368 metric_logger.py:185] completed step: 8, seconds: 0.176, TFLOP/s/device: 77.140, Tokens/s/device: 11627.445, total_weights: 65536, loss: 8.255
I0420 19:58:31.439256 137846163138368 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 19:58:31.439930 137846163138368 checkpointing.py:798] Waited 0.0006949901580810547 seconds for step 9 to finish before starting checkpointing.
I0420 19:58:31.443062 137846163138368 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 19:58:32.683579 137727722911488 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_7
I0420 19:59:05.083916 137727714518784 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 28.710 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 55.01727914810181 s) (per-host)
I0420 19:59:05.084033 137727714518784 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 54.506935s.
I0420 19:59:13.412055 137727714518784 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 62.834940s.
I0420 19:59:13.412348 137698455054080 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 19:59:13.412479 137698455054080 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0420 19:59:13.412529 137698455054080 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 19:59:13.414147 137698455054080 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 19:59:13.414283 137846163138368 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0420 19:59:13.414367 137846163138368 checkpoint_manager.py:1458] Waiting for previous save to complete took 41.971314 seconds. If this number is high, consider checkpointing less frequently.
I0420 19:59:13.416540 137846163138368 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9
I0420 19:59:13.419944 137846163138368 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/9.
I0420 19:59:14.385666 137846163138368 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array.
I0420 19:59:14.385764 137846163138368 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
I0420 19:59:14.429582 137846163138368 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044974s
I0420 19:59:14.429763 137846163138368 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.005 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3081827163696289 s) (per-host)
I0420 19:59:14.429815 137846163138368 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.308243s (batch_requests_ready=0.261402s, total_serialization_initiated=0.046775s, others=0.000066s)
I0420 19:59:14.429901 137846163138368 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.312449s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.312433)
I0420 19:59:14.430840 137717472007936 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started.
I0420 19:59:14.430968 137846163138368 async_checkpointer.py:561] Finished blocking save. Time taken: 1.014354s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/9.
I0420 19:59:14.441510 137846163138368 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 19:59:14.441743 137698455054080 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save.
I0420 19:59:14.441874 137846163138368 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776715111.4430325, 'wait_for_prev_duration_secs': 41.97131419181824, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776715153.4165819, 'checkpointer_blocking_duration_secs': 1.014570713043213, 'get_old_steps_start_time': 1776715154.4311726, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776715111.440164, 'checkpoint_manager_blocking_duration_secs': 43.0016770362854}
I0420 19:59:14.441977 137846163138368 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 19:59:14.442020 137846163138368 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 19:59:19.219247 137727722911488 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_7
I0420 19:59:55.871805 137717472007936 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.833 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.750181913375854 s) (per-host)
I0420 19:59:55.871929 137717472007936 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.440875s.
I0420 20:00:04.122507 137717472007936 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.691438s.
I0420 20:00:04.122819 137698455054080 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 20:00:04.122935 137698455054080 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save.
I0420 20:00:04.122992 137698455054080 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 20:00:04.124636 137698455054080 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 20:00:04.124903 137846163138368 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0420 20:00:04.125057 137846163138368 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 20:00:04.125689 137846163138368 metric_logger.py:185] completed step: 9, seconds: 0.163, TFLOP/s/device: 83.588, Tokens/s/device: 12599.355, total_weights: 65536, loss: 8.179
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 20:00:17 UTC 2026
EXIT_CODE=0