MaxView

XPK Start: Tue Apr 21 06:25:10 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-21 06:25:36.193028: 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)
I0421 06:25:36.403877 137758269183808 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 06:25:45.445963 137758269183808 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 06:25:45.448569 137758269183808 distributed.py:172] Connecting to JAX distributed service on mt-04-int8-eix5b-slice-job-0-0.mt-04-int8-eix5b:8482
I0421 06:25:46.641819 137758269183808 max_utils.py:284] Jax distributed system initialized!
I0421 06:25:51.910862 137758269183808 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 06:25:51.910972 137758269183808 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 06:25:51.911014 137758269183808 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
I0421 06:25:51.911052 137758269183808 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 06:25:52.612745 137758269183808 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 06:25:52.613173 137758269183808 checkpointing.py:677] Setting up checkpoint logger...
I0421 06:25:52.613234 137758269183808 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0421 06:25:52.613293 137758269183808 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 06:25:52.613667 137758269183808 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 0x7d49a0f50080>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 06:25:55.642197 137758269183808 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0421 06:25:55.642436 137758269183808 checkpoint_manager.py:708] [process=4][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d3624321a90>}, handler_registry=None
I0421 06:25:55.642696 137758269183808 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d3624321a90>` 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`.
I0421 06:25:55.642747 137758269183808 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d36243254f0>` 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`.
I0421 06:25:55.642784 137758269183808 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 0x7d3624321a90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d3624321a90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d36243254f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d36243254f0>}).
I0421 06:25:55.643102 137758269183808 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 06:25:55.643180 137758269183808 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d3624155760> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 06:25:57.832879 137758269183808 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints
I0421 06:25:57.843868 137758269183808 checkpoint_manager.py:929] [process=4][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_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d3624287740>
I0421 06:25:57.843989 137758269183808 checkpointing.py:301] Checkpoint manager created!
I0421 06:25:58.971506 137758269183808 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 06:25:58.971631 137758269183808 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 06:25:59.359594 137758269183808 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 06:25:59.359702 137758269183808 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 06:25:59.376151 137758269183808 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 06:25:59.376212 137758269183808 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 06:25:59.442471 137758269183808 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 06:25:59.442564 137758269183808 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 06:25:59.459172 137758269183808 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 06:25:59.459241 137758269183808 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 06:25:59.475664 137758269183808 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0421 06:25:59.475728 137758269183808 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 06:25:59.500971 137758269183808 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 06:25:59.501043 137758269183808 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 06:25:59.567715 137758269183808 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 06:25:59.567801 137758269183808 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 06:26:00.077659 137758269183808 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 06:26:00.077773 137758269183808 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0421 06:26:02.528729 137758269183808 maxtext_utils.py:1668]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   P('norm',)
    Physical:  (None,)
I0421 06:26:02.528861 137758269183808 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 06:26:02.528916 137758269183808 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   P('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0421 06:26:02.528976 137758269183808 maxtext_utils.py:1668]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   P('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0421 06:26:02.529031 137758269183808 maxtext_utils.py:1668]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 06:26:02.529071 137758269183808 maxtext_utils.py:1668]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   P('norm', 'layers')
    Physical:  (None, None)
I0421 06:26:02.529125 137758269183808 maxtext_utils.py:1668]  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)
I0421 06:26:02.529180 137758269183808 maxtext_utils.py:1668]  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')
I0421 06:26:02.529222 137758269183808 maxtext_utils.py:1668]  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)
I0421 06:26:02.529265 137758269183808 maxtext_utils.py:1668]  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)
I0421 06:26:02.529312 137758269183808 maxtext_utils.py:1668]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   P('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)

I0421 06:26:02.529361 137758269183808 maxtext_utils.py:1668]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   P('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')
I0421 06:26:03.922946 137758269183808 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 06:26:03.923045 137758269183808 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 06:26:03.938593 137758269183808 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 06:26:03.938665 137758269183808 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 06:26:17.770292 137758269183808 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0421 06:26:17.771198 137758269183808 metric_logger.py:301] number parameters: 1.104 billion
I0421 06:26:32.038755 137758269183808 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0421 06:26:32.278318 137758269183808 checkpointing.py:776] Waited 0.23954272270202637 seconds for step 0 to finish before starting checkpointing.
I0421 06:26:32.281109 137758269183808 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 06:26:32.283139 137758269183808 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 0
I0421 06:26:32.284572 137758269183808 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/0.
I0421 06:26:33.039305 137758269183808 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 06:26:33.040186 137758269183808 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 06:26:33.040250 137758269183808 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
I0421 06:26:33.323596 137758269183808 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.284371s
I0421 06:26:33.323792 137758269183808 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 5.318 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29009151458740234 s) (per-host)
I0421 06:26:33.323854 137758269183808 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.290164s (batch_requests_ready=0.002409s, total_serialization_initiated=0.287673s, others=0.000081s)
I0421 06:26:33.323949 137758269183808 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.294487s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.294470)
I0421 06:26:33.324752 137758269183808 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/0.
I0421 06:26:33.325029 137628090095360 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 06:46:33.325004
I0421 06:26:33.334073 137758269183808 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 06:26:33.334369 137627556763392 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0421 06:26:33.334530 137758269183808 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776752792.2810907, 'wait_for_prev_duration_secs': 6.198883056640625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776752792.2831786, 'checkpointer_blocking_duration_secs': 1.0419728755950928, 'get_old_steps_start_time': 1776752793.3251767, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776752792.2788918, 'checkpoint_manager_blocking_duration_secs': 1.0555975437164307}
I0421 06:26:33.334653 137758269183808 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0421 06:26:33.334707 137758269183808 max_utils.py:750] 
Memstats: After params initialized:
I0421 06:26:33.334757 137758269183808 max_utils.py:756] 	Using (GB) 0.44 / 31.25 (1.408000%) on TPU_16(process=4,(0,4,0,0))
I0421 06:26:33.334789 137758269183808 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0))
I0421 06:26:33.334818 137758269183808 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0))
I0421 06:26:33.334844 137758269183808 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0))
I0421 06:26:33.655003 137758269183808 metric_logger.py:196] completed step: 0, seconds: 14.267, TFLOP/s/device: 0.952, Tokens/s/device: 143.544, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277
I0421 06:26:33.912134 137758269183808 metric_logger.py:196] completed step: 1, seconds: 1.615, TFLOP/s/device: 8.415, Tokens/s/device: 1268.388, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277
I0421 06:26:34.331900 137758269183808 metric_logger.py:196] completed step: 2, seconds: 0.028, TFLOP/s/device: 480.773, Tokens/s/device: 72467.358, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28634.820
I0421 06:26:34.561433 137758269183808 metric_logger.py:196] completed step: 3, seconds: 0.421, TFLOP/s/device: 32.293, Tokens/s/device: 4867.614, total_weights: 65536, loss: 9.731, lm_loss: 9.731, perplexity: 16827.199
I0421 06:26:35.021633 137758269183808 metric_logger.py:196] completed step: 4, seconds: 0.234, TFLOP/s/device: 58.050, Tokens/s/device: 8749.968, total_weights: 65536, loss: 9.272, lm_loss: 9.272, perplexity: 10638.531
I0421 06:26:35.027407 137758269183808 metric_logger.py:196] completed step: 5, seconds: 0.230, TFLOP/s/device: 59.179, Tokens/s/device: 8920.055, total_weights: 65536, loss: 8.887, lm_loss: 8.887, perplexity: 7233.941
I0421 06:26:37.967666    2889 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 06:26:40.381908 137627594381056 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/0/items/array_metadatas/process_4
I0421 06:27:04.327629 137758269183808 metric_logger.py:196] completed step: 6, seconds: 0.460, TFLOP/s/device: 29.520, Tokens/s/device: 4449.620, total_weights: 65536, loss: 8.588, lm_loss: 8.588, perplexity: 5365.468
I0421 06:27:04.557541 137758269183808 metric_logger.py:196] completed step: 7, seconds: 29.072, TFLOP/s/device: 0.467, Tokens/s/device: 70.446, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4358.779
I0421 06:27:04.786894 137758269183808 metric_logger.py:196] completed step: 8, seconds: 0.235, TFLOP/s/device: 57.930, Tokens/s/device: 8731.874, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.453
I0421 06:27:05.015499 137758269183808 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0421 06:27:05.016167 137758269183808 checkpointing.py:776] Waited 0.0006937980651855469 seconds for step 9 to finish before starting checkpointing.
I0421 06:27:05.018238 137758269183808 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 06:27:08.324749 137628090095360 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 44.759 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 35.29100704193115 s) (per-host)
I0421 06:27:08.324875 137628090095360 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.999762s.
I0421 06:27:16.114088 137628090095360 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 42.788961s.
I0421 06:27:16.114406 137627556763392 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 06:27:16.114531 137627556763392 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0421 06:27:16.114579 137627556763392 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0421 06:27:16.116031 137627556763392 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0421 06:27:16.116179 137758269183808 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0421 06:27:16.116265 137758269183808 checkpoint_manager.py:1452] Waiting for previous save to complete took 11.098031 seconds. If this number is high, consider checkpointing less frequently.
I0421 06:27:16.118151 137758269183808 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 9
I0421 06:27:16.120252 137758269183808 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/9.
I0421 06:27:16.826102 137758269183808 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 06:27:16.826195 137758269183808 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
I0421 06:27:16.857904 137758269183808 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.032917s
I0421 06:27:16.858078 137758269183808 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 42.332 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.036440134048461914 s) (per-host)
I0421 06:27:16.858130 137758269183808 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.036502s (batch_requests_ready=0.001819s, total_serialization_initiated=0.034614s, others=0.000069s)
I0421 06:27:16.858228 137758269183808 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.040779s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.040764)
I0421 06:27:16.858919 137758269183808 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/9.
I0421 06:27:16.859203 137627556763392 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 06:47:16.859178
I0421 06:27:16.863032 137758269183808 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 06:27:16.863370 137627594381056 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0421 06:27:16.863538 137758269183808 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776752825.0182085, 'wait_for_prev_duration_secs': 11.098030805587769, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776752836.1181884, 'checkpointer_blocking_duration_secs': 0.7411203384399414, 'get_old_steps_start_time': 1776752836.859332, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776752825.016395, 'checkpoint_manager_blocking_duration_secs': 11.8471040725708}
I0421 06:27:16.863662 137758269183808 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0421 06:27:16.863709 137758269183808 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 06:27:23.425943 137603990914816 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260421_061359/linen_xpk_main_20260421_061359_04_int8/checkpoints/9/items/array_metadatas/process_4
I0421 06:28:00.092837 137627556763392 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 36.505 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.27115869522095 s) (per-host)
I0421 06:28:00.092964 137627556763392 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.233680s.
I0421 06:28:10.766931 137627556763392 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 53.907632s.
I0421 06:28:10.767215 137627594381056 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 06:28:10.767334 137627594381056 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0421 06:28:10.767388 137627594381056 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0421 06:28:10.769170 137627594381056 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0421 06:28:10.769340 137758269183808 checkpoint_manager.py:2032] [process=4][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0421 06:28:10.769518 137758269183808 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 06:28:10.770504 137758269183808 metric_logger.py:196] completed step: 9, seconds: 0.229, TFLOP/s/device: 59.281, Tokens/s/device: 8935.467, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3552.333
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Tue Apr 21 06:28:21 UTC 2026
EXIT_CODE=0