MaxView

XPK Start: Sun Apr 19 00:01:44 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-19 00:02:08.231807: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0419 00:02:08.409271 138973401782080 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 00:02:17.448552 138973401782080 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 00:02:17.450850 138973401782080 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-78ixe-slice-job-0-0.mt-06-grad-accum-78ixe:8482
I0419 00:02:19.686976 138973401782080 max_utils.py:284] Jax distributed system initialized!
I0419 00:02:25.938467 138973401782080 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 00:02:25.938567 138973401782080 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 00:02:25.938606 138973401782080 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0419 00:02:25.938642 138973401782080 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 00:02:27.010752 138973401782080 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 00:02:27.011358 138973401782080 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 00:02:27.011535 138973401782080 checkpointing.py:688] Setting up checkpoint logger...
I0419 00:02:27.011585 138973401782080 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 00:02:27.011628 138973401782080 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 00:02:27.011962 138973401782080 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 0x7e648fc28710>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 00:02:30.815753 138973401782080 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0419 00:02:30.815986 138973401782080 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e6380b962a0>}, handler_registry=None
I0419 00:02:30.816224 138973401782080 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e6380b962a0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0419 00:02:30.816292 138973401782080 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6380b87380>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0419 00:02:30.816330 138973401782080 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 0x7e6380b962a0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e6380b962a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6380b87380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e6380b87380>}).
I0419 00:02:30.816643 138973401782080 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 00:02:30.816714 138973401782080 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e505c33ec00> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 00:02:31.930809 138973401782080 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints
I0419 00:02:31.939435 138973401782080 checkpoint_manager.py:929] [process=3][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_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e6380b94a40>
I0419 00:02:31.939552 138973401782080 checkpointing.py:302] Checkpoint manager created!
I0419 00:02:33.644999 138973401782080 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 00:02:33.645106 138973401782080 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 00:02:34.022196 138973401782080 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 00:02:34.022296 138973401782080 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 00:02:34.038138 138973401782080 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 00:02:34.038195 138973401782080 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 00:02:34.061550 138973401782080 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 00:02:34.061618 138973401782080 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 00:02:34.077689 138973401782080 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 00:02:34.077748 138973401782080 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 00:02:34.093717 138973401782080 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 00:02:34.093777 138973401782080 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 00:02:34.117819 138973401782080 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 00:02:34.117894 138973401782080 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 00:02:34.139288 138973401782080 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 00:02:34.139357 138973401782080 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 00:02:34.350201 138973401782080 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 00:02:34.350320 138973401782080 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 00:02:35.769005 138973401782080 maxtext_utils.py:1827]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0419 00:02:35.769130 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 00:02:35.769184 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0419 00:02:35.769289 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0419 00:02:35.769357 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 00:02:35.769399 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0419 00:02:35.769453 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0419 00:02:35.769507 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   P('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0419 00:02:35.769549 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0419 00:02:35.769586 138973401782080 maxtext_utils.py:1827]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   P('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0419 00:02:35.769636 138973401782080 maxtext_utils.py:1827]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)

I0419 00:02:35.769685 138973401782080 maxtext_utils.py:1827]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0419 00:02:35.791800 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0419 00:02:35.791866 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0419 00:02:35.806520 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0419 00:02:35.806577 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0419 00:02:35.835661 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0419 00:02:35.835731 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0419 00:02:35.850268 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0419 00:02:35.850322 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0419 00:02:35.879178 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0419 00:02:35.879256 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0419 00:02:35.893806 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0419 00:02:35.893863 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0419 00:02:35.937309 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0419 00:02:35.937381 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0419 00:02:35.951873 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0419 00:02:35.951932 138973401782080 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0419 00:02:36.610028 138973401782080 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 00:02:36.610122 138973401782080 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 00:02:36.625501 138973401782080 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0419 00:02:36.625558 138973401782080 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0419 00:02:47.900347 138973401782080 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0419 00:02:47.901172 138973401782080 metric_logger.py:301] number parameters: 1.104 billion
I0419 00:02:49.756930 138973401782080 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0419 00:03:00.603784 138973401782080 checkpointing.py:798] Waited 10.846832275390625 seconds for step 0 to finish before starting checkpointing.
I0419 00:03:00.606337 138973401782080 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 00:03:00.608291 138973401782080 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0
I0419 00:03:00.609624 138973401782080 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/0.
I0419 00:03:01.807428 138973401782080 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 00:03:01.808474 138973401782080 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 00:03:01.808530 138973401782080 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0419 00:03:02.084492 138973401782080 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277134s
I0419 00:03:02.084672 138973401782080 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 5.460 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2825033664703369 s) (per-host)
I0419 00:03:02.084747 138973401782080 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282589s (batch_requests_ready=0.002196s, total_serialization_initiated=0.280298s, others=0.000096s)
I0419 00:03:02.084858 138973401782080 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286797s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.286780)
I0419 00:03:02.085679 138973401782080 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.48 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/0.
I0419 00:03:02.086007 138844679788288 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 00:23:02.085971
I0419 00:03:02.092408 138973401782080 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 00:03:02.092696 138843018700544 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0419 00:03:02.092827 138973401782080 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776556980.6063178, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776556980.6083293, 'checkpointer_blocking_duration_secs': 1.4778258800506592, 'get_old_steps_start_time': 1776556982.0861788, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776556980.6043, 'checkpoint_manager_blocking_duration_secs': 1.4884917736053467}
I0419 00:03:02.092981 138973401782080 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0419 00:03:02.093031 138973401782080 max_utils.py:750] 
Memstats: After params initialized:
I0419 00:03:02.093078 138973401782080 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0))
I0419 00:03:02.093110 138973401782080 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0))
I0419 00:03:02.093137 138973401782080 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0))
I0419 00:03:02.093163 138973401782080 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0))
I0419 00:03:02.404185 138973401782080 metric_logger.py:196] completed step: 0, seconds: 1.856, TFLOP/s/device: 29.288, Tokens/s/device: 4414.673, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0419 00:03:03.019652 138973401782080 metric_logger.py:196] completed step: 1, seconds: 12.646, TFLOP/s/device: 4.298, Tokens/s/device: 647.807, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0419 00:03:03.598176 138973401782080 metric_logger.py:196] completed step: 2, seconds: 0.037, TFLOP/s/device: 1470.588, Tokens/s/device: 221663.014, total_weights: 262144, loss: 10.563, lm_loss: 10.563, perplexity: 38662.707
I0419 00:03:04.177025 138973401782080 metric_logger.py:196] completed step: 3, seconds: 0.584, TFLOP/s/device: 92.987, Tokens/s/device: 14016.021, total_weights: 262144, loss: 10.272, lm_loss: 10.272, perplexity: 28909.668
I0419 00:03:05.334419 138973401782080 metric_logger.py:196] completed step: 4, seconds: 0.578, TFLOP/s/device: 93.950, Tokens/s/device: 14161.103, total_weights: 262144, loss: 10.022, lm_loss: 10.022, perplexity: 22524.992
I0419 00:03:05.340361 138973401782080 metric_logger.py:196] completed step: 5, seconds: 0.579, TFLOP/s/device: 93.911, Tokens/s/device: 14155.328, total_weights: 262144, loss: 9.820, lm_loss: 9.820, perplexity: 18401.865
I0419 00:03:05.384285    2614 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 00:03:08.188517 138843027093248 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/0/items/array_metadatas/process_3
I0419 00:03:39.763264 138844679788288 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 41.611 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.9610481262207 s) (per-host)
I0419 00:03:39.763394 138844679788288 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.677275s.
I0419 00:03:48.475168 138844679788288 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 46.389034s.
I0419 00:03:48.475461 138843018700544 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 00:03:48.475584 138843018700544 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0419 00:03:48.475632 138843018700544 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 00:03:48.477462 138843018700544 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 00:03:52.381025 138973401782080 metric_logger.py:196] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.932, Tokens/s/device: 7074.077, total_weights: 262144, loss: 9.667, lm_loss: 9.667, perplexity: 15787.604
I0419 00:03:52.959814 138973401782080 metric_logger.py:196] completed step: 7, seconds: 46.463, TFLOP/s/device: 1.170, Tokens/s/device: 176.314, total_weights: 262144, loss: 9.561, lm_loss: 9.561, perplexity: 14203.827
I0419 00:03:53.538440 138973401782080 metric_logger.py:196] completed step: 8, seconds: 0.584, TFLOP/s/device: 93.117, Tokens/s/device: 14035.665, total_weights: 262144, loss: 9.496, lm_loss: 9.496, perplexity: 13302.920
I0419 00:03:54.116375 138973401782080 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0419 00:03:54.117001 138973401782080 checkpointing.py:798] Waited 0.0006475448608398438 seconds for step 9 to finish before starting checkpointing.
I0419 00:03:54.119397 138973401782080 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 00:03:54.121052 138973401782080 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 9
I0419 00:03:54.122503 138973401782080 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/9.
I0419 00:03:54.837769 138973401782080 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0419 00:03:54.837864 138973401782080 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0419 00:03:54.873785 138973401782080 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037038s
I0419 00:03:54.873957 138973401782080 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 38.184 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.040398597717285156 s) (per-host)
I0419 00:03:54.874010 138973401782080 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040461s (batch_requests_ready=0.001604s, total_serialization_initiated=0.038787s, others=0.000070s)
I0419 00:03:54.874106 138973401782080 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.044900s (all_items=0.000014s, per_item={'items': '0.00001407'}, temp_paths=0.044886)
I0419 00:03:54.874801 138973401782080 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/9.
I0419 00:03:54.875197 138844673054464 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 00:23:54.875151
I0419 00:03:54.877103 138973401782080 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 00:03:54.877376 138838743693056 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save.
I0419 00:03:54.877540 138973401782080 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776557034.119364, 'wait_for_prev_duration_secs': 7.390975952148438e-05, 'time_between_consecutive_saves_sec': 5.6418633460998535, 'checkpointer_blocking_start_time': 1776557034.1210904, 'checkpointer_blocking_duration_secs': 0.7542710304260254, 'get_old_steps_start_time': 1776557034.8753817, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776557034.1172278, 'checkpoint_manager_blocking_duration_secs': 0.760277271270752}
I0419 00:03:54.877650 138973401782080 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0419 00:03:54.877693 138973401782080 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 00:04:02.026447 138843018700544 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_06_grad_accum/checkpoints/9/items/array_metadatas/process_3
I0419 00:04:38.221592 138844673054464 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.406 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.38799500465393 s) (per-host)
I0419 00:04:38.221719 138844673054464 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.346388s.
I0419 00:04:47.054514 138844673054464 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 52.179168s.
I0419 00:04:47.054799 138838743693056 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 00:04:47.054936 138838743693056 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save.
I0419 00:04:47.054989 138838743693056 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 00:04:47.056917 138838743693056 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 00:04:47.057099 138973401782080 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0419 00:04:47.057265 138973401782080 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 00:04:47.058156 138973401782080 metric_logger.py:196] completed step: 9, seconds: 0.579, TFLOP/s/device: 93.932, Tokens/s/device: 14158.460, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12802.546
Per train step:
 Total TFLOPs: 54.35 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 00:04:59 UTC 2026
EXIT_CODE=0