MaxView

XPK Start: Mon Apr 20 08:21:06 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-20 08:21:30.101116: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0420 08:21:30.279622 138964451231552 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 08:21:39.318940 138964451231552 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 08:21:39.321232 138964451231552 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-uxzvb-slice-job-0-0.mt-03-dropout-uxzvb:8482
I0420 08:21:40.914929 138964451231552 max_utils.py:284] Jax distributed system initialized!
I0420 08:21:47.099283 138964451231552 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 08:21:47.099384 138964451231552 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 08:21:47.099425 138964451231552 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
I0420 08:21:47.099460 138964451231552 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 08:21:48.147407 138964451231552 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 08:21:48.147998 138964451231552 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 08:21:48.148175 138964451231552 checkpointing.py:688] Setting up checkpoint logger...
I0420 08:21:48.148223 138964451231552 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 08:21:48.148266 138964451231552 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 08:21:48.148594 138964451231552 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 0x7e6276e4c470>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 08:21:51.270012 138964451231552 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 08:21:51.270202 138964451231552 checkpoint_manager.py:708] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e616aca6300>}, handler_registry=None
I0420 08:21:51.270442 138964451231552 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e616aca6300>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0420 08:21:51.270490 138964451231552 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e616accc5c0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0420 08:21:51.270525 138964451231552 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 0x7e616aca6300>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e616aca6300>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e616accc5c0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e616accc5c0>}).
I0420 08:21:51.270931 138964451231552 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 08:21:51.271004 138964451231552 async_checkpointer.py:192] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e4e3c593740> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 08:21:51.956747 138964451231552 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints
I0420 08:21:52.417902 138964451231552 checkpoint_manager.py:929] [process=1][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e616aca6c00>
I0420 08:21:52.418078 138964451231552 checkpointing.py:302] Checkpoint manager created!
I0420 08:21:52.766825 138964451231552 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 08:21:52.766926 138964451231552 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 08:21:53.152036 138964451231552 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 08:21:53.152123 138964451231552 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 08:21:53.168223 138964451231552 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 08:21:53.168281 138964451231552 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None).
I0420 08:21:53.191638 138964451231552 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 08:21:53.191721 138964451231552 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 08:21:53.207932 138964451231552 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 08:21:53.207991 138964451231552 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 08:21:53.224046 138964451231552 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 08:21:53.224104 138964451231552 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 08:21:53.249317 138964451231552 attentions.py:1197] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 08:21:53.249388 138964451231552 attentions.py:1197] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None).
I0420 08:21:53.275149 138964451231552 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 08:21:53.275214 138964451231552 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None).
I0420 08:21:53.498013 138964451231552 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 08:21:53.498116 138964451231552 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 08:21:54.935663 138964451231552 maxtext_utils.py:1821]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0420 08:21:54.935814 138964451231552 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0420 08:21:54.935869 138964451231552 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0420 08:21:54.935927 138964451231552 maxtext_utils.py:1821]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0420 08:21:54.935981 138964451231552 maxtext_utils.py:1821]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0420 08:21:54.936022 138964451231552 maxtext_utils.py:1821]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0420 08:21:54.936076 138964451231552 maxtext_utils.py:1821]  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)
I0420 08:21:54.936128 138964451231552 maxtext_utils.py:1821]  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')
I0420 08:21:54.936168 138964451231552 maxtext_utils.py:1821]  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)
I0420 08:21:54.936203 138964451231552 maxtext_utils.py:1821]  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)
I0420 08:21:54.936249 138964451231552 maxtext_utils.py:1821]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0420 08:21:54.936294 138964451231552 maxtext_utils.py:1821]  params/params/token_embedder/embedding

    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0420 08:21:55.458008 138964451231552 train.py:157] train/xent Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0420 08:21:55.458100 138964451231552 train.py:157] train/xent Physical: float32[32,128]............................................. ('fsdp', None).
I0420 08:21:55.473344 138964451231552 train.py:164] train/z_loss Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0420 08:21:55.473405 138964451231552 train.py:164] train/z_loss Physical: float32[32,128]............................................. ('fsdp', None).
I0420 08:21:59.102794 138964451231552 max_utils.py:791] Total memory size: 0.8 GB, Output size: 0.4 GB, Temp size: 0.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0420 08:21:59.103629 138964451231552 metric_logger.py:301] number parameters: 1.104 billion
I0420 08:22:03.241470 138964451231552 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 08:22:03.315098 138964451231552 checkpointing.py:798] Waited 0.07361006736755371 seconds for step 0 to finish before starting checkpointing.
I0420 08:22:03.317620 138964451231552 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 08:22:03.319474 138964451231552 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 0
I0420 08:22:03.320867 138964451231552 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/0.
I0420 08:22:03.617051 138964451231552 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 08:22:03.618044 138964451231552 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0420 08:22:03.618102 138964451231552 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0420 08:22:03.895360 138964451231552 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278379s
I0420 08:22:03.895534 138964451231552 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 5.439 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28361964225769043 s) (per-host)
I0420 08:22:03.895589 138964451231552 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.283685s (batch_requests_ready=0.002131s, total_serialization_initiated=0.281481s, others=0.000072s)
I0420 08:22:03.895680 138964451231552 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.288337s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.288319)
I0420 08:22:03.896508 138964451231552 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.58 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/0.
I0420 08:22:03.896783 138836227127040 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 08:42:03.896760
I0420 08:22:03.940153 138964451231552 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 08:22:03.940469 138835103028992 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 08:22:03.940609 138964451231552 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776673323.3176029, 'wait_for_prev_duration_secs': 6.0558319091796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776673323.3195114, 'checkpointer_blocking_duration_secs': 0.5773763656616211, 'get_old_steps_start_time': 1776673323.896913, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776673323.3155758, 'checkpoint_manager_blocking_duration_secs': 0.6249923706054688}
I0420 08:22:03.940732 138964451231552 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 08:22:03.940785 138964451231552 max_utils.py:750] 
Memstats: After params initialized:
I0420 08:22:03.940838 138964451231552 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_2(process=1,(2,0,0,0))
I0420 08:22:03.940872 138964451231552 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_3(process=1,(3,0,0,0))
I0420 08:22:03.940896 138964451231552 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_6(process=1,(2,1,0,0))
I0420 08:22:03.940922 138964451231552 max_utils.py:756] 	Using (GB) 0.41 / 31.25 (1.312000%) on TPU_7(process=1,(3,1,0,0))
I0420 08:22:04.250671 138964451231552 metric_logger.py:196] completed step: 0, seconds: 4.138, TFLOP/s/device: 0.194, Tokens/s/device: 30.935, total_weights: 4096, loss: 10.889, lm_loss: 10.889, perplexity: 53560.699
I0420 08:22:04.365380 138964451231552 metric_logger.py:196] completed step: 1, seconds: 1.008, TFLOP/s/device: 0.795, Tokens/s/device: 127.028, total_weights: 4096, loss: 10.902, lm_loss: 10.902, perplexity: 54298.523
I0420 08:22:04.791131 138964451231552 metric_logger.py:196] completed step: 2, seconds: 0.044, TFLOP/s/device: 18.026, Tokens/s/device: 2881.001, total_weights: 4096, loss: 9.956, lm_loss: 9.956, perplexity: 21082.051
I0420 08:22:04.861964 138964451231552 metric_logger.py:196] completed step: 3, seconds: 0.426, TFLOP/s/device: 1.879, Tokens/s/device: 300.250, total_weights: 4096, loss: 9.139, lm_loss: 9.139, perplexity: 9314.691
I0420 08:22:05.003605 138964451231552 metric_logger.py:196] completed step: 4, seconds: 0.076, TFLOP/s/device: 10.551, Tokens/s/device: 1686.318, total_weights: 4096, loss: 8.459, lm_loss: 8.459, perplexity: 4715.493
I0420 08:22:05.009853 138964451231552 metric_logger.py:196] completed step: 5, seconds: 0.071, TFLOP/s/device: 11.342, Tokens/s/device: 1812.697, total_weights: 4096, loss: 7.924, lm_loss: 7.924, perplexity: 2762.373
I0420 08:22:06.989414    2789 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 08:22:09.535617 138835111421696 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/0/items/array_metadatas/process_1
I0420 08:22:14.502179 138964451231552 metric_logger.py:196] completed step: 6, seconds: 0.142, TFLOP/s/device: 5.623, Tokens/s/device: 898.643, total_weights: 4096, loss: 7.553, lm_loss: 7.553, perplexity: 1906.617
I0420 08:22:14.573140 138964451231552 metric_logger.py:196] completed step: 7, seconds: 9.423, TFLOP/s/device: 0.085, Tokens/s/device: 13.584, total_weights: 4096, loss: 7.267, lm_loss: 7.267, perplexity: 1431.653
I0420 08:22:14.643757 138964451231552 metric_logger.py:196] completed step: 8, seconds: 0.075, TFLOP/s/device: 10.647, Tokens/s/device: 1701.652, total_weights: 4096, loss: 7.111, lm_loss: 7.111, perplexity: 1225.983
I0420 08:22:14.713956 138964451231552 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 08:22:14.714576 138964451231552 checkpointing.py:798] Waited 0.0006368160247802734 seconds for step 9 to finish before starting checkpointing.
I0420 08:22:14.716775 138964451231552 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 08:22:36.728365 138836227127040 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 47.698 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 33.11641812324524 s) (per-host)
I0420 08:22:36.728470 138836227127040 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.831613s.
I0420 08:22:44.179994 138836227127040 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 40.283120s.
I0420 08:22:44.180212 138835103028992 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 08:22:44.180271 138835103028992 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 08:22:44.180325 138835103028992 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 08:22:44.182046 138835103028992 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 08:22:44.182207 138964451231552 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0420 08:22:44.182338 138964451231552 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.465567 seconds. If this number is high, consider checkpointing less frequently.
I0420 08:22:44.184283 138964451231552 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 9
I0420 08:22:44.186427 138964451231552 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/9.
I0420 08:22:44.475886 138964451231552 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0420 08:22:44.475972 138964451231552 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0420 08:22:44.509675 138964451231552 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034913s
I0420 08:22:44.509827 138964451231552 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 40.362 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03821825981140137 s) (per-host)
I0420 08:22:44.509872 138964451231552 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038273s (batch_requests_ready=0.001707s, total_serialization_initiated=0.036505s, others=0.000061s)
I0420 08:22:44.509954 138964451231552 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042486s (all_items=0.000014s, per_item={'items': '0.00001359'}, temp_paths=0.042473)
I0420 08:22:44.510651 138964451231552 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/9.
I0420 08:22:44.510966 138835103028992 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 08:42:44.510934
I0420 08:22:44.520210 138964451231552 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 08:22:44.520466 138835111421696 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save.
I0420 08:22:44.520602 138964451231552 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776673334.7167416, 'wait_for_prev_duration_secs': 29.465566873550415, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776673364.1843212, 'checkpointer_blocking_duration_secs': 0.3267934322357178, 'get_old_steps_start_time': 1776673364.511137, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776673334.7148204, 'checkpoint_manager_blocking_duration_secs': 29.805745363235474}
I0420 08:22:44.520722 138964451231552 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 08:22:44.520770 138964451231552 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 08:22:49.700387 138829738526464 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533/linen_xpk_feat_nnx_trainstate_and_training_loop_20260420_081533_03_dropout/checkpoints/9/items/array_metadatas/process_1
I0420 08:23:25.032562 138835103028992 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 38.944 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.56091260910034 s) (per-host)
I0420 08:23:25.032685 138835103028992 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.521605s.
I0420 08:23:33.024588 138835103028992 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 48.513489s.
I0420 08:23:33.024895 138835111421696 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 08:23:33.025013 138835111421696 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save.
I0420 08:23:33.025061 138835111421696 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 08:23:33.026824 138835111421696 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 08:23:33.027004 138964451231552 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0420 08:23:33.027157 138964451231552 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 08:23:33.028214 138964451231552 metric_logger.py:196] completed step: 9, seconds: 0.071, TFLOP/s/device: 11.328, Tokens/s/device: 1810.569, total_weights: 4096, loss: 6.981, lm_loss: 6.981, perplexity: 1075.867
Per train step:
 Total TFLOPs: 0.80 
 split as 99.60% learnable weight flops and 0.40% attention flops
XPK End: Mon Apr 20 08:23:41 UTC 2026
EXIT_CODE=0