XPK Start: Tue Apr 21 08:43:17 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-21 08:43:41.302773: 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 08:43:41.838665 140194653390656 max_utils.py:273] Attempting to initialize the jax distributed system...
I0421 08:43:50.880166 140194653390656 distributed.py:149] Starting JAX distributed service on [::]:8482
I0421 08:43:50.882564 140194653390656 distributed.py:172] Connecting to JAX distributed service on mt-04-int8-tqphp-slice-job-0-0.mt-04-int8-tqphp:8482
I0421 08:43:52.798838 140194653390656 max_utils.py:284] Jax distributed system initialized!
I0421 08:43:58.838463 140194653390656 max_utils.py:800] System Information: Jax Version: 0.9.2
I0421 08:43:58.838566 140194653390656 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0421 08:43:58.838607 140194653390656 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 08:43:58.838642 140194653390656 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0421 08:43:59.599988 140194653390656 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 08:43:59.600574 140194653390656 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0421 08:43:59.600893 140194653390656 checkpointing.py:688] Setting up checkpoint logger...
I0421 08:43:59.600947 140194653390656 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0421 08:43:59.600993 140194653390656 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0421 08:43:59.601332 140194653390656 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 0x7f810d91b170>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0421 08:44:02.579818 140194653390656 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0421 08:44:02.580046 140194653390656 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 0x7f7fd8eac830>}, handler_registry=None
I0421 08:44:02.580284 140194653390656 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f7fd8eac830>` 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 08:44:02.580332 140194653390656 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f6cf4745c40>` 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 08:44:02.580366 140194653390656 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 0x7f7fd8eac830>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f7fd8eac830>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f6cf4745c40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f6cf4745c40>}).
I0421 08:44:02.580682 140194653390656 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0421 08:44:02.580769 140194653390656 async_checkpointer.py:192] [process=4][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f6cf44bfec0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0421 08:44:04.421340 140194653390656 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints
I0421 08:44:04.423555 140194653390656 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_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f6cf4745dc0>
I0421 08:44:04.423666 140194653390656 checkpointing.py:302] Checkpoint manager created!
I0421 08:44:05.565519 140194653390656 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0421 08:44:05.565627 140194653390656 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 08:44:05.947241 140194653390656 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 08:44:05.947328 140194653390656 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 08:44:05.963542 140194653390656 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0421 08:44:05.963603 140194653390656 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0421 08:44:06.029394 140194653390656 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 08:44:06.029479 140194653390656 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 08:44:06.045778 140194653390656 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 08:44:06.045838 140194653390656 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 08:44:06.062134 140194653390656 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 08:44:06.062198 140194653390656 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 08:44:06.092022 140194653390656 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0421 08:44:06.092115 140194653390656 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0421 08:44:06.158692 140194653390656 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0421 08:44:06.158787 140194653390656 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0421 08:44:06.672729 140194653390656 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0421 08:44:06.672837 140194653390656 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0421 08:44:08.696355 140194653390656 maxtext_utils.py:1821] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0421 08:44:08.696486 140194653390656 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)
I0421 08:44:08.696549 140194653390656 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)
I0421 08:44:08.696611 140194653390656 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')
I0421 08:44:08.696666 140194653390656 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)
I0421 08:44:08.696740 140194653390656 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)
I0421 08:44:08.696803 140194653390656 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)
I0421 08:44:08.696859 140194653390656 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')
I0421 08:44:08.696900 140194653390656 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)
I0421 08:44:08.696935 140194653390656 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)
I0421 08:44:08.696983 140194653390656 maxtext_utils.py:1821] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0421 08:44:08.697029 140194653390656 maxtext_utils.py:1821] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0421 08:44:10.531588 140194653390656 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 08:44:10.531682 140194653390656 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 08:44:10.547323 140194653390656 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0421 08:44:10.547384 140194653390656 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0421 08:44:24.012942 140194653390656 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 08:44:24.013767 140194653390656 metric_logger.py:301] number parameters: 1.104 billion
I0421 08:44:39.244941 140194653390656 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0421 08:44:39.483078 140194653390656 checkpointing.py:798] Waited 0.23812103271484375 seconds for step 0 to finish before starting checkpointing.
I0421 08:44:39.485950 140194653390656 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 08:44:39.487943 140194653390656 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 0
I0421 08:44:39.489385 140194653390656 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/0.
I0421 08:44:40.236649 140194653390656 signaling_client.py:364] Using JaxDistributedSignalingClient
I0421 08:44:40.242853 140194653390656 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 08:44:40.242916 140194653390656 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 08:44:40.525254 140194653390656 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.288678s
I0421 08:44:40.525433 140194653390656 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 5.244 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29414987564086914 s) (per-host)
I0421 08:44:40.525489 140194653390656 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.294218s (batch_requests_ready=0.002217s, total_serialization_initiated=0.291925s, others=0.000076s)
I0421 08:44:40.525584 140194653390656 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.298704s (all_items=0.000025s, per_item={'items': '0.00002503'}, temp_paths=0.298679)
I0421 08:44:40.526434 140194653390656 event_tracking.py:125] [process=4] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/0.
I0421 08:44:40.526737 140064881178368 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 09:04:40.526689
I0421 08:44:40.528716 140194653390656 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 08:44:40.529000 140062718174976 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0421 08:44:40.529164 140194653390656 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776761079.4859319, 'wait_for_prev_duration_secs': 6.318092346191406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776761079.4879806, 'checkpointer_blocking_duration_secs': 1.038907527923584, 'get_old_steps_start_time': 1776761080.5269132, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776761079.4835796, 'checkpoint_manager_blocking_duration_secs': 1.0455377101898193}
I0421 08:44:40.529272 140194653390656 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0421 08:44:40.529323 140194653390656 max_utils.py:750]
Memstats: After params initialized:
I0421 08:44:40.529371 140194653390656 max_utils.py:756] Using (GB) 0.44 / 31.25 (1.408000%) on TPU_16(process=4,(0,4,0,0))
I0421 08:44:40.529402 140194653390656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_17(process=4,(1,4,0,0))
I0421 08:44:40.529428 140194653390656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_20(process=4,(0,5,0,0))
I0421 08:44:40.529453 140194653390656 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_21(process=4,(1,5,0,0))
I0421 08:44:40.855724 140194653390656 metric_logger.py:196] completed step: 0, seconds: 15.231, TFLOP/s/device: 0.892, Tokens/s/device: 134.462, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277
I0421 08:44:41.101957 140194653390656 metric_logger.py:196] completed step: 1, seconds: 1.609, TFLOP/s/device: 8.444, Tokens/s/device: 1272.733, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277
I0421 08:44:41.515956 140194653390656 metric_logger.py:196] completed step: 2, seconds: 0.017, TFLOP/s/device: 776.985, Tokens/s/device: 117115.572, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28634.820
I0421 08:44:41.745611 140194653390656 metric_logger.py:196] completed step: 3, seconds: 0.415, TFLOP/s/device: 32.767, Tokens/s/device: 4938.998, total_weights: 65536, loss: 9.731, lm_loss: 9.731, perplexity: 16827.199
I0421 08:44:42.204616 140194653390656 metric_logger.py:196] completed step: 4, seconds: 0.235, TFLOP/s/device: 57.923, Tokens/s/device: 8730.758, total_weights: 65536, loss: 9.272, lm_loss: 9.272, perplexity: 10638.531
I0421 08:44:42.210807 140194653390656 metric_logger.py:196] completed step: 5, seconds: 0.229, TFLOP/s/device: 59.218, Tokens/s/device: 8926.003, total_weights: 65536, loss: 8.887, lm_loss: 8.887, perplexity: 7233.941
I0421 08:44:43.906573 2932 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0421 08:44:46.257299 140062726567680 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/0/items/array_metadatas/process_4
I0421 08:45:12.328202 140194653390656 metric_logger.py:196] completed step: 6, seconds: 0.460, TFLOP/s/device: 29.563, Tokens/s/device: 4456.078, total_weights: 65536, loss: 8.588, lm_loss: 8.588, perplexity: 5365.468
I0421 08:45:12.557720 140194653390656 metric_logger.py:196] completed step: 7, seconds: 29.889, TFLOP/s/device: 0.455, Tokens/s/device: 68.520, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4358.779
I0421 08:45:12.787056 140194653390656 metric_logger.py:196] completed step: 8, seconds: 0.234, TFLOP/s/device: 58.090, Tokens/s/device: 8755.953, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.453
I0421 08:45:13.015786 140194653390656 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0421 08:45:13.016465 140194653390656 checkpointing.py:798] Waited 0.0006992816925048828 seconds for step 9 to finish before starting checkpointing.
I0421 08:45:13.018553 140194653390656 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 08:45:13.911910 140064881178368 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 46.899 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 33.68059754371643 s) (per-host)
I0421 08:45:13.912019 140064881178368 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.385169s.
I0421 08:45:22.543291 140064881178368 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 42.016425s.
I0421 08:45:22.543567 140062718174976 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 08:45:22.543689 140062718174976 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0421 08:45:22.543756 140062718174976 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0421 08:45:22.545449 140062718174976 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0421 08:45:22.545607 140194653390656 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 08:45:22.545756 140194653390656 checkpoint_manager.py:1452] Waiting for previous save to complete took 9.527203 seconds. If this number is high, consider checkpointing less frequently.
I0421 08:45:22.547738 140194653390656 checkpoint_manager.py:1512] [process=4] Saving checkpoint at step 9
I0421 08:45:22.549820 140194653390656 event_tracking.py:70] [process=4] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/9.
I0421 08:45:23.267897 140194653390656 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0421 08:45:23.267991 140194653390656 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 08:45:23.299868 140194653390656 base_pytree_checkpoint_handler.py:154] [process=4][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.033994s
I0421 08:45:23.300038 140194653390656 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/blocking_gbytes_per_sec: 41.242 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.037403106689453125 s) (per-host)
I0421 08:45:23.300090 140194653390656 base_pytree_checkpoint_handler.py:768] [process=4][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037465s (batch_requests_ready=0.001769s, total_serialization_initiated=0.035626s, others=0.000069s)
I0421 08:45:23.300200 140194653390656 composite_checkpoint_handler.py:715] [process=4][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041747s (all_items=0.000014s, per_item={'items': '0.00001431'}, temp_paths=0.041733)
I0421 08:45:23.300861 140194653390656 event_tracking.py:125] [process=4] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/9.
I0421 08:45:23.301191 140062726567680 async_checkpointer.py:76] [process=4][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 09:05:23.301153
I0421 08:45:23.304111 140194653390656 checkpoint_manager.py:1560] [process=4][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0421 08:45:23.304413 140062718174976 async_checkpointer.py:280] [process=4][thread=save_finalize] Waiting for background save thread=async_save.
I0421 08:45:23.304571 140194653390656 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776761113.0185244, 'wait_for_prev_duration_secs': 9.52720308303833, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776761122.5477757, 'checkpointer_blocking_duration_secs': 0.7535614967346191, 'get_old_steps_start_time': 1776761123.301358, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776761113.0166888, 'checkpoint_manager_blocking_duration_secs': 10.287847995758057}
I0421 08:45:23.304678 140194653390656 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0421 08:45:23.304739 140194653390656 checkpoint_manager.py:2020] [process=4][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0421 08:45:28.874089 140043679160064 array_metadata_store.py:203] [process=4][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238/linen_xpk_feat_nnx_trainstate_and_training_loop_20260421_083238_04_int8/checkpoints/9/items/array_metadatas/process_4
I0421 08:46:05.441080 140062726567680 base_pytree_checkpoint_handler.py:130] [process=4] /jax/orbax/write/gbytes_per_sec: 37.450 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.17840552330017 s) (per-host)
I0421 08:46:05.441205 140062726567680 async_checkpointer.py:90] [process=4][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.139903s.
I0421 08:46:13.358117 140062726567680 async_checkpointer.py:160] [process=4][thread=async_save] Background save thread done. Time taken: 50.056799s.
I0421 08:46:13.358421 140062718174976 async_checkpointer.py:288] [process=4][thread=save_finalize] Done with waiting for background save thread=async_save.
I0421 08:46:13.358542 140062718174976 async_checkpointer.py:298] [process=4][thread=save_finalize] No errors found in background save thread=async_save.
I0421 08:46:13.358591 140062718174976 checkpoint_manager.py:2137] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0421 08:46:13.359971 140062718174976 checkpoint_manager.py:2146] [process=4][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0421 08:46:13.360166 140194653390656 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 08:46:13.360314 140194653390656 checkpoint_manager.py:2009] [process=4][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0421 08:46:13.361280 140194653390656 metric_logger.py:196] completed step: 9, seconds: 0.229, TFLOP/s/device: 59.214, Tokens/s/device: 8925.381, 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 08:46:22 UTC 2026
EXIT_CODE=0