MaxView

XPK Start: Thu Apr 23 14:30:58 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-23 14:31:22.789785: 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)
I0423 14:31:23.002042 132720895014720 max_utils.py:273] Attempting to initialize the jax distributed system...
I0423 14:31:32.044381 132720895014720 distributed.py:149] Starting JAX distributed service on [::]:8482
I0423 14:31:32.046726 132720895014720 distributed.py:172] Connecting to JAX distributed service on mt-05-fp8-0b9f4-slice-job-0-0.mt-05-fp8-0b9f4:8482
I0423 14:31:33.401512 132720895014720 max_utils.py:284] Jax distributed system initialized!
I0423 14:31:39.475560 132720895014720 max_utils.py:800] System Information: Jax Version: 0.9.2
I0423 14:31:39.475681 132720895014720 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0423 14:31:39.475726 132720895014720 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
I0423 14:31:39.475764 132720895014720 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 14:31:40.158612 132720895014720 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 14:31:40.592018 132720895014720 checkpointing.py:688] Setting up checkpoint logger...
I0423 14:31:40.592137 132720895014720 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0423 14:31:40.592182 132720895014720 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0423 14:31:40.592406 132720895014720 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 0x78b4dad18890>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 14:31:43.491353 132720895014720 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0423 14:31:43.491607 132720895014720 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a0544dcfe0>}, handler_registry=None
I0423 14:31:43.491856 132720895014720 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a0544dcfe0>` 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`.
I0423 14:31:43.491905 132720895014720 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a0544d9ac0>` 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`.
I0423 14:31:43.491942 132720895014720 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 0x78a0544dcfe0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78a0544dcfe0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a0544d9ac0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78a0544d9ac0>}).
I0423 14:31:43.492269 132720895014720 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0423 14:31:43.492341 132720895014720 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78a0a40b6fc0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0423 14:31:44.697677 132720895014720 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints
I0423 14:31:45.105427 132720895014720 checkpoint_manager.py:929] [process=0][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/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78a0a40fbf20>
I0423 14:31:45.105626 132720895014720 checkpointing.py:302] Checkpoint manager created!
I0423 14:31:46.354678 132720895014720 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 14:31:46.354807 132720895014720 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0423 14:31:48.449262 132720895014720 maxtext_utils.py:1889]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0423 14:31:48.449380 132720895014720 maxtext_utils.py:1889]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0423 14:31:48.449435 132720895014720 maxtext_utils.py:1889]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0423 14:31:48.449479 132720895014720 maxtext_utils.py:1889]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0423 14:31:48.449517 132720895014720 maxtext_utils.py:1889]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0423 14:31:48.449552 132720895014720 maxtext_utils.py:1889]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0423 14:31:48.449600 132720895014720 maxtext_utils.py:1889]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:31:48.449636 132720895014720 maxtext_utils.py:1889]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0423 14:31:48.449671 132720895014720 maxtext_utils.py:1889]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:31:48.449703 132720895014720 maxtext_utils.py:1889]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:31:48.449735 132720895014720 maxtext_utils.py:1889]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]

    Physical:  ('fsdp', None)
I0423 14:31:48.449768 132720895014720 maxtext_utils.py:1889]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0423 14:31:48.866006 132720895014720 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 14:31:48.866109 132720895014720 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:31:49.268905 132720895014720 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 14:31:49.269010 132720895014720 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:31:49.290550 132720895014720 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 14:31:49.290630 132720895014720 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:31:49.307496 132720895014720 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 14:31:49.307562 132720895014720 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:31:49.368539 132720895014720 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 14:31:49.368632 132720895014720 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:31:49.385608 132720895014720 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 14:31:49.385672 132720895014720 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:31:49.402663 132720895014720 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 14:31:49.402725 132720895014720 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:31:49.434544 132720895014720 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 14:31:49.434627 132720895014720 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:31:49.494872 132720895014720 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 14:31:49.494955 132720895014720 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 14:32:02.928919 132720895014720 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.
I0423 14:32:02.940051 132720895014720 metric_logger.py:301] number parameters: 1.104 billion
I0423 14:32:17.340161 132720895014720 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0423 14:32:17.501348 132720895014720 checkpointing.py:798] Waited 0.16117238998413086 seconds for step 0 to finish before starting checkpointing.
I0423 14:32:17.503795 132720895014720 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 14:32:17.505926 132720895014720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0423 14:32:17.507377 132720895014720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0.
I0423 14:32:17.837808 132720895014720 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 14:32:17.867572 132720895014720 jax_array_handlers.py:360] Scheduling D2H of 153 prioritized jax.Array.
I0423 14:32:17.867670 132720895014720 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
I0423 14:32:18.101064 132591000741632 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0
I0423 14:32:18.722691 132720895014720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.857497s
I0423 14:32:18.723456 132720895014720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 1.757 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.8780510425567627 s) (per-host)
I0423 14:32:18.723529 132720895014720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.878144s (batch_requests_ready=0.006089s, total_serialization_initiated=0.871370s, others=0.000684s)
I0423 14:32:18.723630 132720895014720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.882142s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.882125)
I0423 14:32:18.724358 132720895014720 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.22 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0.
I0423 14:32:18.724667 132591009134336 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 14:52:18.724633
I0423 14:32:18.726746 132720895014720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 14:32:18.726978 132590472263424 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 14:32:18.727097 132720895014720 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776954737.5037768, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776954737.505964, 'checkpointer_blocking_duration_secs': 1.2188405990600586, 'get_old_steps_start_time': 1776954738.7248318, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776954737.5018582, 'checkpoint_manager_blocking_duration_secs': 1.22520112991333}
I0423 14:32:18.727210 132720895014720 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0423 14:32:18.727286 132720895014720 max_utils.py:750] 
Memstats: After params initialized:
I0423 14:32:18.727342 132720895014720 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_0(process=0,(0,0,0,0))
I0423 14:32:18.727372 132720895014720 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_1(process=0,(1,0,0,0))
I0423 14:32:18.727397 132720895014720 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_4(process=0,(0,1,0,0))
I0423 14:32:18.727421 132720895014720 max_utils.py:756] 	Using (GB) 0.45 / 31.25 (1.440000%) on TPU_5(process=0,(1,1,0,0))
I0423 14:32:19.085392 132720895014720 metric_logger.py:196] completed step: 0, seconds: 13.976, TFLOP/s/device: 0.972, Tokens/s/device: 146.533, total_weights: 65536, loss: 10.872, lm_loss: 10.872, perplexity: 52680.742
I0423 14:32:19.088994 132720895014720 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/tensorboard/'
I0423 14:32:19.530663 132720895014720 metric_logger.py:196] completed step: 1, seconds: 1.739, TFLOP/s/device: 7.815, Tokens/s/device: 1177.894, total_weights: 65536, loss: 10.872, lm_loss: 10.872, perplexity: 52680.742
I0423 14:32:19.676761 132720895014720 metric_logger.py:196] completed step: 2, seconds: 0.447, TFLOP/s/device: 30.403, Tokens/s/device: 4582.681, total_weights: 65536, loss: 10.856, lm_loss: 10.856, perplexity: 51864.762
I0423 14:32:19.833719 132720895014720 metric_logger.py:196] completed step: 3, seconds: 0.018, TFLOP/s/device: 751.708, Tokens/s/device: 113305.671, total_weights: 65536, loss: 10.824, lm_loss: 10.824, perplexity: 50225.512
I0423 14:32:20.437756 132590445537024 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776954739993613427, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776954739993613427, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/_CHECKPOINT_METADATA
I0423 14:32:20.670472 132590983956224 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/items
I0423 14:32:21.824865    2852 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 14:32:23.715182 132590967170816 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 153 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/items/array_metadatas/process_0
I0423 14:32:41.205755 132590958778112 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 19.385699s (commit=18.174843s, array_metadata_write=1.210855s)
I0423 14:32:41.207141 132591009134336 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 67.622 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 23.361706733703613 s) (per-host)
I0423 14:32:41.207263 132591009134336 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 22.482489s.
I0423 14:32:42.353886 132591009134336 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776954739993613427, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/_CHECKPOINT_METADATA
I0423 14:32:43.755728 132591009134336 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0423 14:32:43.892026 132720895014720 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.365, Tokens/s/device: 13922.312, total_weights: 65536, loss: 10.793, lm_loss: 10.793, perplexity: 48689.156
I0423 14:32:43.906951 132720895014720 metric_logger.py:196] completed step: 5, seconds: 0.156, TFLOP/s/device: 86.922, Tokens/s/device: 13101.833, total_weights: 65536, loss: 10.763, lm_loss: 10.763, perplexity: 47229.012
I0423 14:32:44.058422 132720895014720 metric_logger.py:196] completed step: 6, seconds: 24.058, TFLOP/s/device: 0.565, Tokens/s/device: 85.129, total_weights: 65536, loss: 10.733, lm_loss: 10.733, perplexity: 45864.258
I0423 14:32:44.124404 132591009134336 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.001087s.
I0423 14:32:44.131034 132590445537024 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776954739993613427, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/_CHECKPOINT_METADATA
I0423 14:32:44.215494 132720895014720 metric_logger.py:196] completed step: 7, seconds: 0.013, TFLOP/s/device: 1031.203, Tokens/s/device: 155434.123, total_weights: 65536, loss: 10.706, lm_loss: 10.706, perplexity: 44620.242
I0423 14:32:44.372512 132720895014720 metric_logger.py:196] completed step: 8, seconds: 0.151, TFLOP/s/device: 90.131, Tokens/s/device: 13585.496, total_weights: 65536, loss: 10.680, lm_loss: 10.680, perplexity: 43477.055
I0423 14:32:44.382974 132720895014720 checkpointing.py:794] Waiting for step 10 to finish before checkpoint...
I0423 14:32:44.686815 132720895014720 checkpointing.py:798] Waited 0.30381107330322266 seconds for step 10 to finish before starting checkpointing.
I0423 14:32:44.689173 132720895014720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 14:32:44.716846 132591009134336 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.242672s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/items
I0423 14:32:44.718629 132591009134336 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0/items
I0423 14:32:45.235283 132591009134336 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0
I0423 14:32:46.604382 132591009134336 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0`.
I0423 14:32:46.605243 132591009134336 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 29.10 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/0
I0423 14:32:46.606785 132591009134336 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 27.882009s.
I0423 14:32:46.606990 132590472263424 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 14:32:46.607110 132590472263424 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 14:32:46.607172 132590472263424 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0423 14:32:46.609765 132590472263424 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0423 14:32:46.609906 132720895014720 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0423 14:32:46.609991 132720895014720 checkpoint_manager.py:1452] Waiting for previous save to complete took 1.920824 seconds. If this number is high, consider checkpointing less frequently.
I0423 14:32:46.612265 132720895014720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10
I0423 14:32:46.614403 132720895014720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10.
I0423 14:32:47.342008 132720895014720 jax_array_handlers.py:360] Scheduling D2H of 153 prioritized jax.Array.
I0423 14:32:47.342109 132720895014720 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
I0423 14:32:47.382052 132720895014720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042416s
I0423 14:32:47.382783 132720895014720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 25.533 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.060422658920288086 s) (per-host)
I0423 14:32:47.382848 132720895014720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.060506s (batch_requests_ready=0.005744s, total_serialization_initiated=0.054096s, others=0.000666s)
I0423 14:32:47.382929 132720895014720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.064428s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.064412)
I0423 14:32:47.383587 132720895014720 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10.
I0423 14:32:47.383923 132590472263424 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 14:52:47.383883
I0423 14:32:47.385946 132720895014720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 14:32:47.386253 132589896443648 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 14:32:47.386394 132720895014720 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776954764.6891396, 'wait_for_prev_duration_secs': 1.9208238124847412, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776954766.6123054, 'checkpointer_blocking_duration_secs': 0.7717578411102295, 'get_old_steps_start_time': 1776954767.384088, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776954764.687279, 'checkpoint_manager_blocking_duration_secs': 2.6990807056427}
I0423 14:32:47.386566 132720895014720 checkpointing.py:409] Started an asynchronous checkpoint save for step 10
I0423 14:32:47.387208 132720895014720 metric_logger.py:196] completed step: 9, seconds: 0.157, TFLOP/s/device: 86.595, Tokens/s/device: 13052.567, total_weights: 65536, loss: 10.656, lm_loss: 10.656, perplexity: 42467.320
I0423 14:32:47.402743 132720895014720 metric_logger.py:196] completed step: 10, seconds: 0.157, TFLOP/s/device: 86.690, Tokens/s/device: 13066.891, total_weights: 65536, loss: 10.636, lm_loss: 10.636, perplexity: 41601.941
I0423 14:32:47.877994 132591009134336 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10
I0423 14:32:47.890155 132720895014720 metric_logger.py:196] completed step: 11, seconds: 3.017, TFLOP/s/device: 4.504, Tokens/s/device: 678.892, total_weights: 65536, loss: 10.618, lm_loss: 10.618, perplexity: 40849.570
I0423 14:32:48.042129 132720895014720 metric_logger.py:196] completed step: 12, seconds: 0.489, TFLOP/s/device: 27.764, Tokens/s/device: 4184.853, total_weights: 65536, loss: 10.602, lm_loss: 10.602, perplexity: 40203.926
I0423 14:32:48.198847 132720895014720 metric_logger.py:196] completed step: 13, seconds: 0.010, TFLOP/s/device: 1306.078, Tokens/s/device: 196866.289, total_weights: 65536, loss: 10.588, lm_loss: 10.588, perplexity: 39652.145
I0423 14:32:48.355726 132720895014720 metric_logger.py:196] completed step: 14, seconds: 0.155, TFLOP/s/device: 87.843, Tokens/s/device: 13240.580, total_weights: 65536, loss: 10.577, lm_loss: 10.577, perplexity: 39212.879
I0423 14:32:48.512539 132720895014720 metric_logger.py:196] completed step: 15, seconds: 0.156, TFLOP/s/device: 87.057, Tokens/s/device: 13122.233, total_weights: 65536, loss: 10.568, lm_loss: 10.568, perplexity: 38879.223
I0423 14:32:48.669502 132720895014720 metric_logger.py:196] completed step: 16, seconds: 0.157, TFLOP/s/device: 86.809, Tokens/s/device: 13084.757, total_weights: 65536, loss: 10.562, lm_loss: 10.562, perplexity: 38619.105
I0423 14:32:48.826304 132720895014720 metric_logger.py:196] completed step: 17, seconds: 0.156, TFLOP/s/device: 86.841, Tokens/s/device: 13089.608, total_weights: 65536, loss: 10.556, lm_loss: 10.556, perplexity: 38395.816
I0423 14:32:48.983341 132720895014720 metric_logger.py:196] completed step: 18, seconds: 0.157, TFLOP/s/device: 86.587, Tokens/s/device: 13051.403, total_weights: 65536, loss: 10.552, lm_loss: 10.552, perplexity: 38236.652
I0423 14:32:49.142069 132720895014720 checkpointing.py:794] Waiting for step 19 to finish before checkpoint...
I0423 14:32:49.143494 132720895014720 checkpointing.py:798] Waited 0.0014369487762451172 seconds for step 19 to finish before starting checkpointing.
I0423 14:32:49.146095 132720895014720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 14:32:50.455962 132590460815104 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10/items
I0423 14:32:53.477093 132590420358912 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 153 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10/items/array_metadatas/process_0
I0423 14:33:12.316863 132589904836352 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 20.655986s (commit=19.444351s, array_metadata_write=1.211634s)
I0423 14:33:12.318139 132590472263424 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 63.202 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 24.99575662612915 s) (per-host)
I0423 14:33:12.318251 132590472263424 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 24.934216s.
I0423 14:33:13.942189 132590472263424 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000878s.
I0423 14:33:14.089160 132590472263424 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0423 14:33:15.029169 132590472263424 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.924794s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10/items
I0423 14:33:15.030869 132590472263424 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10/items
I0423 14:33:15.574783 132590472263424 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10
I0423 14:33:17.384059 132590472263424 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10`.
I0423 14:33:17.384788 132590472263424 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 30.77 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/10
I0423 14:33:17.386013 132590472263424 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 30.001976s.
I0423 14:33:17.386227 132589896443648 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 14:33:17.386343 132589896443648 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 14:33:17.386406 132589896443648 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0423 14:33:17.389103 132589896443648 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0423 14:33:17.389282 132720895014720 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0423 14:33:17.389414 132720895014720 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.243324 seconds. If this number is high, consider checkpointing less frequently.
I0423 14:33:17.390945 132720895014720 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19
I0423 14:33:17.393070 132720895014720 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19.
I0423 14:33:17.696691 132720895014720 jax_array_handlers.py:360] Scheduling D2H of 153 prioritized jax.Array.
I0423 14:33:17.696848 132720895014720 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
I0423 14:33:17.745778 132720895014720 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051428s
I0423 14:33:17.746572 132720895014720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 22.404 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.06886076927185059 s) (per-host)
I0423 14:33:17.746647 132720895014720 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.068954s (batch_requests_ready=0.005248s, total_serialization_initiated=0.062974s, others=0.000732s)
I0423 14:33:17.746735 132720895014720 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.072836s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.072825)
I0423 14:33:17.747433 132720895014720 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.36 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19.
I0423 14:33:17.747805 132589904836352 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 14:53:17.747761
I0423 14:33:17.923546 132590472263424 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19
I0423 14:33:18.186308 132720895014720 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 14:33:18.186753 132589896443648 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 14:33:18.186930 132720895014720 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776954769.146067, 'wait_for_prev_duration_secs': 28.243324279785156, 'time_between_consecutive_saves_sec': 2.536262273788452, 'checkpointer_blocking_start_time': 1776954797.3909845, 'checkpointer_blocking_duration_secs': 0.3569784164428711, 'get_old_steps_start_time': 1776954797.7479818, 'get_old_steps_duration_secs': 2.3603439331054688e-05, 'checkpoint_manager_blocking_start_time': 1776954769.1438427, 'checkpoint_manager_blocking_duration_secs': 29.04305362701416}
I0423 14:33:18.187051 132720895014720 checkpointing.py:409] Started an asynchronous checkpoint save for step 19
I0423 14:33:18.187102 132720895014720 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 14:33:19.981174 132591009134336 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19/items
I0423 14:33:22.956787 132590460815104 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 153 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19/items/array_metadatas/process_0
I0423 14:33:41.579477 132590420358912 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 20.438482s (commit=19.236643s, array_metadata_write=1.201839s)
I0423 14:33:41.580878 132589904836352 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 66.091 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 23.90313744544983 s) (per-host)
I0423 14:33:41.581003 132589904836352 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 23.833071s.
I0423 14:33:42.991832 132589904836352 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000575s.
I0423 14:33:43.349599 132589904836352 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0423 14:33:44.471175 132589904836352 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.067461s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19/items
I0423 14:33:44.472885 132589904836352 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19/items
I0423 14:33:44.990915 132589904836352 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19
I0423 14:33:46.378654 132589904836352 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19`.
I0423 14:33:46.379489 132589904836352 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 28.99 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_05_fp8/checkpoints/19
I0423 14:33:46.380855 132589904836352 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 28.632919s.
I0423 14:33:46.381013 132589896443648 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 14:33:46.381069 132589896443648 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 14:33:46.381126 132589896443648 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0423 14:33:46.383719 132589896443648 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0423 14:33:46.383899 132720895014720 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0423 14:33:46.384090 132720895014720 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 14:33:46.385076 132720895014720 metric_logger.py:196] completed step: 19, seconds: 0.157, TFLOP/s/device: 86.596, Tokens/s/device: 13052.650, total_weights: 65536, loss: 10.549, lm_loss: 10.549, perplexity: 38125.074
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Thu Apr 23 14:33:58 UTC 2026
EXIT_CODE=0