XPK Start: Fri Apr 17 14:35:22 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-17 14:35:46.298279: 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)
I0417 14:35:46.479363 132248673806144 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 14:35:55.520973 132248673806144 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 14:35:55.523222 132248673806144 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-xdvnt-slice-job-0-0.mt-15-ocdbt-true-rs-xdvnt:8482
I0417 14:35:56.991189 132248673806144 max_utils.py:284] Jax distributed system initialized!
I0417 14:36:03.359941 132248673806144 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 14:36:03.360048 132248673806144 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 14:36:03.360088 132248673806144 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
I0417 14:36:03.360123 132248673806144 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 14:36:04.380255 132248673806144 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 14:36:04.380548 132248673806144 checkpointing.py:677] Setting up checkpoint logger...
I0417 14:36:04.380617 132248673806144 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 14:36:04.380681 132248673806144 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 14:36:04.381054 132248673806144 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 0x7846d3648110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 14:36:05.721307 132248673806144 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 14:36:05.721536 132248673806144 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 0x7832a414bbc0>}, handler_registry=None
I0417 14:36:05.721777 132248673806144 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7832a414bbc0>` 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`.
I0417 14:36:05.721824 132248673806144 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7845c7ba79b0>` 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`.
I0417 14:36:05.721877 132248673806144 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 0x7832a414bbc0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7832a414bbc0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7845c7ba79b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7845c7ba79b0>}).
I0417 14:36:05.722200 132248673806144 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 14:36:05.722273 132248673806144 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x783274204180> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 14:36:08.127220 132126565439232 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436306772144251, 'commit_timestamp_nsecs': 1776436346198688496, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA
I0417 14:36:08.385149 132126565439232 checkpoint_manager.py:1871] Missing metrics for step 0
I0417 14:36:08.749201 132248673806144 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints
I0417 14:36:10.070961 132248673806144 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=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7845c7ba64e0>
I0417 14:36:10.071134 132248673806144 checkpointing.py:301] Checkpoint manager created!
I0417 14:36:10.985860 132248673806144 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 14:36:10.985978 132248673806144 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:36:11.364179 132248673806144 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:36:11.364266 132248673806144 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:36:11.380367 132248673806144 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 14:36:11.380425 132248673806144 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 14:36:11.403790 132248673806144 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 14:36:11.403882 132248673806144 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:36:11.419966 132248673806144 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 14:36:11.420025 132248673806144 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:36:11.436113 132248673806144 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0417 14:36:11.436172 132248673806144 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:36:11.460455 132248673806144 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 14:36:11.460522 132248673806144 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 14:36:11.481973 132248673806144 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 14:36:11.482038 132248673806144 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 14:36:11.694168 132248673806144 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 14:36:11.694276 132248673806144 checkpointing.py:581] restoring from this run's directory step 9
I0417 14:36:11.965641 132248673806144 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9.
I0417 14:36:12.664413 132248673806144 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9.
I0417 14:36:12.664864 132248673806144 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.70 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9.
I0417 14:36:14.895667 1472 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 14:36:20.532952 132248673806144 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 280.025 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.6433329582214355 s) (per-host)
I0417 14:36:20.538039 132248673806144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 261.769 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.045757293701172 s) (per-host)
I0417 14:36:21.254078 132248673806144 event_tracking.py:138] [process=0] [sync] Finished load in 9.29 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/9
I0417 14:36:21.254246 132248673806144 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1776436571.9655445, 'checkpointer_duration_secs': 9.288662672042847, 'checkpoint_manager_start_time': 1776436571.6945164, 'checkpoint_manager_duration_secs': 9.559693574905396}
fsdp: 32
I0417 14:36:21.671576 132248673806144 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 14:36:21.671697 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 14:36:21.671748 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 14:36:21.671803 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0417 14:36:21.671875 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 14:36:21.671922 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 14:36:21.671975 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/key/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0417 14:36:21.672027 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: P('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0417 14:36:21.672067 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0417 14:36:21.672102 132248673806144 maxtext_utils.py:1651] params/params/decoder/layers/self_attention/value/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0417 14:36:21.672154 132248673806144 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 14:36:21.672201 132248673806144 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 14:36:22.159625 132248673806144 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:36:22.159720 132248673806144 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:36:22.174907 132248673806144 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 14:36:22.174965 132248673806144 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 14:36:33.173193 132248673806144 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.
I0417 14:36:33.180599 132248673806144 metric_logger.py:301] number parameters: 1.104 billion
I0417 14:36:44.775641 132248673806144 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0417 14:36:44.928627 132248673806144 checkpointing.py:776] Waited 0.15296578407287598 seconds for step 10 to finish before starting checkpointing.
I0417 14:36:44.931139 132248673806144 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:36:44.933279 132248673806144 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10
I0417 14:36:44.934801 132248673806144 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10.
I0417 14:36:45.258623 132248673806144 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 14:36:45.271245 132248673806144 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:36:45.271312 132248673806144 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
I0417 14:36:45.525721 132124401202944 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10
I0417 14:36:45.539018 132248673806144 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.268516s
I0417 14:36:45.539945 132248673806144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.633 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27382826805114746 s) (per-host)
I0417 14:36:45.540024 132248673806144 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.273930s (batch_requests_ready=0.002289s, total_serialization_initiated=0.270801s, others=0.000840s)
I0417 14:36:45.540107 132248673806144 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.277957s (all_items=0.000021s, per_item={'items': '0.00002122'}, temp_paths=0.277936)
I0417 14:36:45.540911 132248673806144 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10.
I0417 14:36:45.541209 132125474944768 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:56:45.541178
I0417 14:36:45.553461 132248673806144 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:36:45.553730 132121127462656 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:36:45.553876 132248673806144 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436604.9311185, 'wait_for_prev_duration_secs': 6.961822509765625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776436604.933319, 'checkpointer_blocking_duration_secs': 0.6080329418182373, 'get_old_steps_start_time': 1776436605.5413737, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776436604.9290922, 'checkpoint_manager_blocking_duration_secs': 0.6247348785400391}
I0417 14:36:45.553986 132248673806144 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0417 14:36:45.554037 132248673806144 max_utils.py:750]
Memstats: After params initialized:
I0417 14:36:45.554087 132248673806144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 14:36:45.554119 132248673806144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 14:36:45.554146 132248673806144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 14:36:45.554171 132248673806144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 14:36:45.888622 132248673806144 metric_logger.py:196] completed step: 10, seconds: 11.285, TFLOP/s/device: 1.204, Tokens/s/device: 181.484, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0417 14:36:46.033741 132248673806144 metric_logger.py:196] completed step: 11, seconds: 1.111, TFLOP/s/device: 12.229, Tokens/s/device: 1843.248, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0417 14:36:46.181377 132248673806144 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1257.369, Tokens/s/device: 189524.338, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0417 14:36:46.328935 132248673806144 metric_logger.py:196] completed step: 13, seconds: 0.145, TFLOP/s/device: 93.546, Tokens/s/device: 14100.313, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0417 14:36:46.338327 132248673806144 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0417 14:36:46.623268 132248673806144 checkpointing.py:776] Waited 0.2847862243652344 seconds for step 15 to finish before starting checkpointing.
I0417 14:36:46.626274 132248673806144 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:36:47.582583 132122203363072 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/items
I0417 14:36:47.792475 132120643106560 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436607328535654, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776436607328535654, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA
I0417 14:36:50.904041 132121680955136 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_0
I0417 14:37:30.233940 132121135855360 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.552852s (commit=40.354274s, array_metadata_write=1.198578s)
I0417 14:37:30.235432 132125474944768 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.126 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.969279289245605 s) (per-host)
I0417 14:37:30.235613 132125474944768 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.694287s.
I0417 14:37:31.705030 132125474944768 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000410s.
I0417 14:37:31.777622 132120643106560 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776436607328535654, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/_CHECKPOINT_METADATA
I0417 14:37:32.344011 132125474944768 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:37:33.298302 132125474944768 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.178705s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/items
I0417 14:37:33.300414 132125474944768 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10/items
I0417 14:37:33.832882 132125474944768 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10
I0417 14:37:35.640498 132125474944768 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10`.
I0417 14:37:35.641695 132125474944768 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.71 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/10
I0417 14:37:35.643298 132125474944768 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.101972s.
I0417 14:37:35.643609 132121127462656 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:37:35.643736 132121127462656 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:37:35.643870 132121127462656 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:37:35.645507 132121127462656 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0417 14:37:35.645790 132248673806144 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.
W0417 14:37:35.645929 132248673806144 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.019662 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:37:35.647669 132248673806144 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15
I0417 14:37:35.650003 132248673806144 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15.
I0417 14:37:35.954169 132248673806144 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:37:35.954299 132248673806144 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
I0417 14:37:35.981851 132248673806144 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028752s
I0417 14:37:35.982752 132248673806144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.578 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.038015127182006836 s) (per-host)
I0417 14:37:35.982870 132248673806144 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038155s (batch_requests_ready=0.004358s, total_serialization_initiated=0.032953s, others=0.000844s)
I0417 14:37:35.982997 132248673806144 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042976s (all_items=0.000024s, per_item={'items': '0.00002432'}, temp_paths=0.042952)
I0417 14:37:35.983941 132248673806144 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15.
I0417 14:37:35.984348 132121127462656 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:57:35.984308
I0417 14:37:35.986801 132248673806144 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:37:35.987149 132121680955136 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:37:35.987335 132248673806144 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436606.6262279, 'wait_for_prev_duration_secs': 49.01966190338135, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776436655.6477358, 'checkpointer_blocking_duration_secs': 0.3367955684661865, 'get_old_steps_start_time': 1776436655.9845583, 'get_old_steps_duration_secs': 4.38690185546875e-05, 'checkpoint_manager_blocking_start_time': 1776436606.623969, 'checkpoint_manager_blocking_duration_secs': 49.3633131980896}
I0417 14:37:35.987485 132248673806144 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0417 14:37:36.181565 132125474944768 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15
I0417 14:37:37.961256 132121689347840 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15/items
I0417 14:37:40.829470 132124392810240 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_0
I0417 14:37:56.685623 132248673806144 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.049, Tokens/s/device: 13874.681, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0417 14:37:56.695214 132248673806144 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 91.922, Tokens/s/device: 13855.438, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0417 14:37:56.840722 132248673806144 metric_logger.py:196] completed step: 16, seconds: 70.355, TFLOP/s/device: 0.193, Tokens/s/device: 29.109, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0417 14:37:56.988198 132248673806144 metric_logger.py:196] completed step: 17, seconds: 0.008, TFLOP/s/device: 1677.423, Tokens/s/device: 252839.506, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0417 14:37:57.135388 132248673806144 metric_logger.py:196] completed step: 18, seconds: 0.146, TFLOP/s/device: 92.822, Tokens/s/device: 13991.078, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0417 14:37:57.282660 132248673806144 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0417 14:37:57.283434 132248673806144 checkpointing.py:776] Waited 0.0007948875427246094 seconds for step 19 to finish before starting checkpointing.
I0417 14:37:57.285519 132248673806144 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:38:21.493580 132122203363072 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.427094s (commit=40.751017s, array_metadata_write=1.676077s)
I0417 14:38:21.495680 132121127462656 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.678 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.550925731658936 s) (per-host)
I0417 14:38:21.495810 132121127462656 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.511318s.
I0417 14:38:22.998508 132121127462656 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000205s.
I0417 14:38:23.646928 132121127462656 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:38:24.592785 132121127462656 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.188986s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15/items
I0417 14:38:24.599385 132121127462656 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15/items
I0417 14:38:25.117669 132121127462656 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15
I0417 14:38:26.485248 132121127462656 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15`.
I0417 14:38:26.486124 132121127462656 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.84 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/15
I0417 14:38:26.487705 132121127462656 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.503211s.
I0417 14:38:26.487928 132121680955136 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:38:26.488053 132121680955136 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:38:26.488130 132121680955136 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:38:26.489745 132121680955136 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0417 14:38:26.489948 132248673806144 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0417 14:38:26.490077 132248673806144 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.204563 seconds. If this number is high, consider checkpointing less frequently.
I0417 14:38:26.491536 132248673806144 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19
I0417 14:38:26.493564 132248673806144 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19.
I0417 14:38:26.792490 132248673806144 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 14:38:26.792585 132248673806144 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
I0417 14:38:26.818023 132248673806144 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026277s
I0417 14:38:26.818782 132248673806144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.023 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030837535858154297 s) (per-host)
I0417 14:38:26.818854 132248673806144 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030916s (batch_requests_ready=0.001887s, total_serialization_initiated=0.028321s, others=0.000708s)
I0417 14:38:26.818923 132248673806144 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034868s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.034851)
I0417 14:38:26.819608 132248673806144 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19.
I0417 14:38:26.819926 132121680955136 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 14:58:26.819889
I0417 14:38:26.821916 132248673806144 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 14:38:26.822180 132122203363072 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 14:38:26.822318 132248673806144 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776436677.2854846, 'wait_for_prev_duration_secs': 29.20456314086914, 'time_between_consecutive_saves_sec': 21.63988971710205, 'checkpointer_blocking_start_time': 1776436706.4915755, 'checkpointer_blocking_duration_secs': 0.3284893035888672, 'get_old_steps_start_time': 1776436706.8200848, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776436677.283665, 'checkpoint_manager_blocking_duration_secs': 29.53861951828003}
I0417 14:38:26.822472 132248673806144 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0417 14:38:26.822519 132248673806144 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 14:38:27.026764 132121127462656 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19
I0417 14:38:28.795910 132124392810240 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19/items
I0417 14:38:32.053946 132125466552064 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_0
I0417 14:39:11.195189 132124401202944 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.874493s (commit=39.598953s, array_metadata_write=1.275539s)
I0417 14:39:11.196253 132121680955136 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.570 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.40826892852783 s) (per-host)
I0417 14:39:11.196376 132121680955136 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.376338s.
I0417 14:39:13.287406 132121680955136 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 14:39:13.318069 132121680955136 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000254s.
I0417 14:39:14.316568 132121680955136 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.905790s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19/items
I0417 14:39:14.318186 132121680955136 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19/items
I0417 14:39:14.860030 132121680955136 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19
I0417 14:39:16.705511 132121680955136 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19`.
I0417 14:39:16.706379 132121680955136 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.21 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_15_ocdbt_true/checkpoints/19
I0417 14:39:16.707688 132121680955136 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.887647s.
I0417 14:39:16.707890 132122203363072 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 14:39:16.708005 132122203363072 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 14:39:16.708056 132122203363072 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0417 14:39:16.710103 132122203363072 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0417 14:39:16.710279 132248673806144 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.
I0417 14:39:16.710425 132248673806144 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 14:39:16.711302 132248673806144 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.354, Tokens/s/device: 13920.609, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 17 14:39:28 UTC 2026
EXIT_CODE=0