XPK Start: Wed Apr 22 07:40:41 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.
2026-04-22 07:41:06.076248: 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)
I0422 07:41:06.287773 139616549066560 max_utils.py:273] Attempting to initialize the jax distributed system...
I0422 07:41:15.326722 139616549066560 distributed.py:149] Starting JAX distributed service on [::]:8482
I0422 07:41:15.329150 139616549066560 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--35fdy-slice-job-0-0.mt-08-checkpoint-async--35fdy:8482
I0422 07:41:16.336783 139616549066560 max_utils.py:284] Jax distributed system initialized!
I0422 07:41:21.956959 139616549066560 max_utils.py:800] System Information: Jax Version: 0.9.2
I0422 07:41:21.957070 139616549066560 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0422 07:41:21.957109 139616549066560 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
I0422 07:41:21.957144 139616549066560 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 07:41:22.646421 139616549066560 maxtext_utils.py:1604] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 07:41:22.646697 139616549066560 checkpointing.py:677] Setting up checkpoint logger...
I0422 07:41:22.646751 139616549066560 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 07:41:22.646814 139616549066560 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0422 07:41:22.647168 139616549066560 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 0x7efa4b0507a0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 07:41:24.023084 139616549066560 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0422 07:41:24.023339 139616549066560 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ee644549b20>}, handler_registry=None
I0422 07:41:24.023580 139616549066560 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ee644549b20>` 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`.
I0422 07:41:24.023630 139616549066560 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ee64454b500>` 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`.
I0422 07:41:24.023665 139616549066560 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 0x7ee644549b20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ee644549b20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ee64454b500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ee64454b500>}).
I0422 07:41:24.023980 139616549066560 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0422 07:41:24.024051 139616549066560 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ee62c69d800> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0422 07:41:26.312232 139495588468480 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776843517065682358, 'commit_timestamp_nsecs': 1776843564748507861, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/9/_CHECKPOINT_METADATA
I0422 07:41:26.953819 139495596861184 checkpoint_manager.py:1871] Missing metrics for step 5
I0422 07:41:27.335180 139616549066560 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints
I0422 07:41:28.906712 139616549066560 checkpoint_manager.py:929] [process=6][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_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ee6444ad5b0>
I0422 07:41:28.906890 139616549066560 checkpointing.py:301] Checkpoint manager created!
I0422 07:41:29.826474 139616549066560 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 07:41:29.826592 139616549066560 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 07:41:30.206795 139616549066560 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 07:41:30.206888 139616549066560 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 07:41:30.223389 139616549066560 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 07:41:30.223458 139616549066560 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 07:41:30.247158 139616549066560 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 07:41:30.247226 139616549066560 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 07:41:30.263629 139616549066560 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 07:41:30.263697 139616549066560 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 07:41:30.280131 139616549066560 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 07:41:30.280192 139616549066560 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 07:41:30.304741 139616549066560 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 07:41:30.304813 139616549066560 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 07:41:30.326939 139616549066560 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 07:41:30.327013 139616549066560 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 07:41:30.535577 139616549066560 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 07:41:30.535682 139616549066560 checkpointing.py:581] restoring from this run's directory step 9
I0422 07:41:30.797424 139616549066560 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/9.
I0422 07:41:31.740720 139616549066560 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/9.
I0422 07:41:31.741164 139616549066560 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.94 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/9.
I0422 07:41:33.580522 1450 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 07:41:39.720473 139616549066560 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 257.121 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.146040201187134 s) (per-host)
I0422 07:41:39.733550 139616549066560 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 240.972 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.567544937133789 s) (per-host)
I0422 07:41:39.807863 139616549066560 event_tracking.py:138] [process=6] [sync] Finished load in 9.01 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/9
I0422 07:41:39.808013 139616549066560 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776843690.7973256, 'checkpointer_duration_secs': 9.010656118392944, 'checkpoint_manager_start_time': 1776843690.5359213, 'checkpoint_manager_duration_secs': 9.272062540054321}
fsdp: 32
I0422 07:41:40.225772 139616549066560 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0422 07:41:40.225899 139616549066560 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0422 07:41:40.225953 139616549066560 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0422 07:41:40.226011 139616549066560 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0422 07:41:40.226063 139616549066560 maxtext_utils.py:1707] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0422 07:41:40.226106 139616549066560 maxtext_utils.py:1707] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0422 07:41:40.226159 139616549066560 maxtext_utils.py:1707] 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)
I0422 07:41:40.226211 139616549066560 maxtext_utils.py:1707] 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')
I0422 07:41:40.226252 139616549066560 maxtext_utils.py:1707] 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)
I0422 07:41:40.226309 139616549066560 maxtext_utils.py:1707] 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)
I0422 07:41:40.226363 139616549066560 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0422 07:41:40.226411 139616549066560 maxtext_utils.py:1707] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0422 07:41:40.712405 139616549066560 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 07:41:40.712505 139616549066560 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 07:41:40.728039 139616549066560 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0422 07:41:40.728107 139616549066560 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0422 07:41:51.708550 139616549066560 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.
I0422 07:41:51.709419 139616549066560 metric_logger.py:301] number parameters: 1.104 billion
I0422 07:42:03.232543 139616549066560 checkpointing.py:772] Waiting for step 10 to finish before checkpoint...
I0422 07:42:03.384758 139616549066560 checkpointing.py:776] Waited 0.15219354629516602 seconds for step 10 to finish before starting checkpointing.
I0422 07:42:03.387252 139616549066560 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 07:42:03.389184 139616549066560 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10
I0422 07:42:03.390664 139616549066560 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/10.
I0422 07:42:04.151986 139616549066560 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 07:42:04.152978 139616549066560 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 07:42:04.153036 139616549066560 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
I0422 07:42:04.429638 139616549066560 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277730s
I0422 07:42:04.429805 139616549066560 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.474 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28181886672973633 s) (per-host)
I0422 07:42:04.429855 139616549066560 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281878s (batch_requests_ready=0.002381s, total_serialization_initiated=0.279430s, others=0.000067s)
I0422 07:42:04.429944 139616549066560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286011s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.285994)
I0422 07:42:04.430688 139616549066560 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/10.
I0422 07:42:04.430965 139495009666816 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 08:02:04.430942
I0422 07:42:04.439367 139616549066560 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 07:42:04.439670 139491780048640 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 07:42:04.439826 139616549066560 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776843723.3872335, 'wait_for_prev_duration_secs': 7.033348083496094e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776843723.3892243, 'checkpointer_blocking_duration_secs': 1.0418426990509033, 'get_old_steps_start_time': 1776843724.4310904, 'get_old_steps_duration_secs': 3.337860107421875e-05, 'checkpoint_manager_blocking_start_time': 1776843723.3853173, 'checkpoint_manager_blocking_duration_secs': 1.054471731185913}
I0422 07:42:04.439929 139616549066560 checkpointing.py:408] Started an asynchronous checkpoint save for step 10
I0422 07:42:04.439978 139616549066560 max_utils.py:750]
Memstats: After params initialized:
I0422 07:42:04.440026 139616549066560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0422 07:42:04.440059 139616549066560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0422 07:42:04.440087 139616549066560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0422 07:42:04.440112 139616549066560 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0422 07:42:04.751654 139616549066560 metric_logger.py:196] completed step: 10, seconds: 11.523, TFLOP/s/device: 1.179, Tokens/s/device: 177.731, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869
I0422 07:42:05.330632 139616549066560 metric_logger.py:196] completed step: 11, seconds: 1.518, TFLOP/s/device: 8.954, Tokens/s/device: 1349.579, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659
I0422 07:42:05.478152 139616549066560 metric_logger.py:196] completed step: 12, seconds: 0.433, TFLOP/s/device: 31.398, Tokens/s/device: 4732.623, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068
I0422 07:42:05.625694 139616549066560 metric_logger.py:196] completed step: 13, seconds: 0.152, TFLOP/s/device: 89.242, Tokens/s/device: 13451.560, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001
I0422 07:42:05.630198 139616549066560 checkpointing.py:772] Waiting for step 15 to finish before checkpoint...
I0422 07:42:05.919657 139616549066560 checkpointing.py:776] Waited 0.2894248962402344 seconds for step 15 to finish before starting checkpointing.
I0422 07:42:05.922745 139616549066560 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 07:42:09.788993 139491788441344 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_6
I0422 07:42:45.578767 139495009666816 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.126 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.4307427406311 s) (per-host)
I0422 07:42:45.578892 139495009666816 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.147855s.
I0422 07:42:54.023459 139495009666816 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.592405s.
I0422 07:42:54.023756 139491780048640 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 07:42:54.023889 139491780048640 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 07:42:54.023940 139491780048640 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts...
I0422 07:42:54.025418 139491780048640 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts.
I0422 07:42:54.025618 139616549066560 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10.
W0422 07:42:54.025755 139616549066560 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.103007 seconds. If this number is high, consider checkpointing less frequently.
I0422 07:42:54.027796 139616549066560 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15
I0422 07:42:54.030324 139616549066560 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/15.
I0422 07:42:55.172387 139616549066560 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 07:42:55.172483 139616549066560 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
I0422 07:42:55.197824 139616549066560 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027284s
I0422 07:42:55.197990 139616549066560 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 49.962 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03087472915649414 s) (per-host)
I0422 07:42:55.198039 139616549066560 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030932s (batch_requests_ready=0.001913s, total_serialization_initiated=0.028955s, others=0.000065s)
I0422 07:42:55.198128 139616549066560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035903s (all_items=0.000013s, per_item={'items': '0.00001335'}, temp_paths=0.035889)
I0422 07:42:55.198872 139616549066560 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/15.
I0422 07:42:55.199161 139491780048640 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 08:02:55.199132
I0422 07:42:55.209821 139616549066560 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 07:42:55.210094 139491788441344 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 07:42:55.210236 139616549066560 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776843725.9227164, 'wait_for_prev_duration_secs': 48.1030068397522, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776843774.0278378, 'checkpointer_blocking_duration_secs': 1.1714556217193604, 'get_old_steps_start_time': 1776843775.1993184, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776843725.920319, 'checkpoint_manager_blocking_duration_secs': 49.289883852005005}
I0422 07:42:55.210396 139616549066560 checkpointing.py:408] Started an asynchronous checkpoint save for step 15
I0422 07:42:55.211186 139616549066560 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 92.101, Tokens/s/device: 13882.487, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193
I0422 07:42:55.217397 139616549066560 metric_logger.py:196] completed step: 15, seconds: 0.148, TFLOP/s/device: 92.102, Tokens/s/device: 13882.675, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661
I0422 07:43:00.687299 139493923346176 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_6
I0422 07:43:16.235404 139616549066560 metric_logger.py:196] completed step: 16, seconds: 49.586, TFLOP/s/device: 0.274, Tokens/s/device: 41.302, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872
I0422 07:43:16.382825 139616549066560 metric_logger.py:196] completed step: 17, seconds: 20.872, TFLOP/s/device: 0.651, Tokens/s/device: 98.123, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975
I0422 07:43:16.530313 139616549066560 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.116, Tokens/s/device: 13432.503, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333
I0422 07:43:16.677011 139616549066560 checkpointing.py:772] Waiting for step 19 to finish before checkpoint...
I0422 07:43:16.677732 139616549066560 checkpointing.py:776] Waited 0.0007410049438476562 seconds for step 19 to finish before starting checkpointing.
I0422 07:43:16.679827 139616549066560 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 07:43:37.501880 139491780048640 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.312 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.334723472595215 s) (per-host)
I0422 07:43:37.502007 139491780048640 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.302749s.
I0422 07:43:47.619998 139491780048640 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 52.420726s.
I0422 07:43:47.620336 139491788441344 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 07:43:47.620456 139491788441344 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 07:43:47.620503 139491788441344 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts...
I0422 07:43:47.622092 139491788441344 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts.
I0422 07:43:47.622254 139616549066560 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15.
W0422 07:43:47.622392 139616549066560 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.942570 seconds. If this number is high, consider checkpointing less frequently.
I0422 07:43:47.623963 139616549066560 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19
I0422 07:43:47.626148 139616549066560 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/19.
I0422 07:43:48.368330 139616549066560 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0422 07:43:48.368423 139616549066560 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
I0422 07:43:48.393801 139616549066560 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026417s
I0422 07:43:48.393999 139616549066560 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 51.290 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030075788497924805 s) (per-host)
I0422 07:43:48.394056 139616549066560 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030151s (batch_requests_ready=0.001805s, total_serialization_initiated=0.028265s, others=0.000081s)
I0422 07:43:48.394164 139616549066560 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034325s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.034313)
I0422 07:43:48.394904 139616549066560 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/19.
I0422 07:43:48.395200 139491788441344 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 08:03:48.395178
I0422 07:43:48.398521 139616549066560 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 07:43:48.398799 139493923346176 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 07:43:48.398961 139616549066560 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776843796.6797972, 'wait_for_prev_duration_secs': 30.942570447921753, 'time_between_consecutive_saves_sec': 22.65433979034424, 'checkpointer_blocking_start_time': 1776843827.6240458, 'checkpointer_blocking_duration_secs': 0.7712681293487549, 'get_old_steps_start_time': 1776843828.3953328, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776843796.677968, 'checkpoint_manager_blocking_duration_secs': 31.72096014022827}
I0422 07:43:48.399077 139616549066560 checkpointing.py:408] Started an asynchronous checkpoint save for step 19
I0422 07:43:48.399120 139616549066560 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 07:43:53.542334 139490169435904 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260422_071422/linen_xpk_main_20260422_071422_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_6
I0422 07:44:31.779970 139491788441344 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 36.383 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.416014194488525 s) (per-host)
I0422 07:44:31.780103 139491788441344 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.384828s.
I0422 07:44:40.018925 139491788441344 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.623635s.
I0422 07:44:40.019217 139493923346176 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 07:44:40.019357 139493923346176 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 07:44:40.019417 139493923346176 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts...
I0422 07:44:40.021139 139493923346176 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts.
I0422 07:44:40.021343 139616549066560 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19.
I0422 07:44:40.021497 139616549066560 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 07:44:40.022621 139616549066560 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.175, Tokens/s/device: 13893.600, 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: Wed Apr 22 07:44:51 UTC 2026
EXIT_CODE=0