XPK Start: Tue Apr 21 13:08:10 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 13:08:34.870165: 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 13:08:35.410933 135048246929216 max_utils.py:273] Attempting to initialize the jax distributed system... I0421 13:08:44.452215 135048246929216 distributed.py:149] Starting JAX distributed service on [::]:8482 I0421 13:08:44.454588 135048246929216 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-uy9w9-slice-job-0-0.mt-16-zarr3-false-sv-uy9w9:8482 I0421 13:08:45.377709 135048246929216 max_utils.py:284] Jax distributed system initialized! I0421 13:08:51.587873 135048246929216 max_utils.py:800] System Information: Jax Version: 0.9.2 I0421 13:08:51.587972 135048246929216 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0421 13:08:51.588011 135048246929216 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 13:08:51.588043 135048246929216 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 13:08:52.287571 135048246929216 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 13:08:52.288192 135048246929216 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 13:08:52.288371 135048246929216 checkpointing.py:688] Setting up checkpoint logger... I0421 13:08:52.288420 135048246929216 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0421 13:08:52.288462 135048246929216 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0421 13:08:52.288800 135048246929216 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7ad2d97ea6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 13:08:55.168511 135048246929216 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0421 13:08:55.168761 135048246929216 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad19b2a9d30>}, handler_registry=None I0421 13:08:55.169009 135048246929216 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad19b2a9d30>` 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 13:08:55.169057 135048246929216 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7abe6460b890>` 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 13:08:55.169090 135048246929216 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 0x7ad19b2a9d30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad19b2a9d30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7abe6460b890>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7abe6460b890>}). I0421 13:08:55.169412 135048246929216 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0421 13:08:55.169482 135048246929216 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7abe643b0180> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0421 13:08:56.207504 135048246929216 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints I0421 13:08:56.615700 135048246929216 checkpoint_manager.py:929] [process=7][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_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7abe64608080> I0421 13:08:56.615916 135048246929216 checkpointing.py:302] Checkpoint manager created! I0421 13:08:57.550238 135048246929216 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 13:08:57.550339 135048246929216 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 13:08:57.927479 135048246929216 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 13:08:57.927573 135048246929216 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 13:08:57.943989 135048246929216 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 13:08:57.944050 135048246929216 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 13:08:57.968485 135048246929216 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 13:08:57.968552 135048246929216 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 13:08:57.984927 135048246929216 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 13:08:57.984986 135048246929216 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 13:08:58.001285 135048246929216 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 13:08:58.001340 135048246929216 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 13:08:58.026025 135048246929216 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 13:08:58.026098 135048246929216 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 13:08:58.046998 135048246929216 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 13:08:58.047063 135048246929216 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 13:08:58.260192 135048246929216 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0421 13:08:58.260297 135048246929216 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 13:08:59.709416 135048246929216 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0421 13:08:59.709541 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 13:08:59.709594 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 13:08:59.709652 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0421 13:08:59.709726 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0421 13:08:59.709821 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0421 13:08:59.709885 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0421 13:08:59.709940 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0421 13:08:59.709980 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0421 13:08:59.710014 135048246929216 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0421 13:08:59.710060 135048246929216 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 13:08:59.710106 135048246929216 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 13:09:00.201361 135048246929216 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 13:09:00.201457 135048246929216 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 13:09:00.217096 135048246929216 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 13:09:00.217160 135048246929216 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 13:09:11.173498 135048246929216 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. I0421 13:09:11.174344 135048246929216 metric_logger.py:301] number parameters: 1.104 billion I0421 13:09:22.651766 135048246929216 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0421 13:09:22.805578 135048246929216 checkpointing.py:798] Waited 0.1538071632385254 seconds for step 0 to finish before starting checkpointing. I0421 13:09:22.808015 135048246929216 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 13:09:22.809713 135048246929216 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 0 I0421 13:09:22.811142 135048246929216 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/0. I0421 13:09:23.146501 135048246929216 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 13:09:23.148661 135048246929216 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 13:09:23.148719 135048246929216 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 13:09:23.423033 135048246929216 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276605s I0421 13:09:23.423207 135048246929216 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.468 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28209686279296875 s) (per-host) I0421 13:09:23.423259 135048246929216 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.282158s (batch_requests_ready=0.002187s, total_serialization_initiated=0.279901s, others=0.000070s) I0421 13:09:23.423354 135048246929216 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.286458s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.286440) I0421 13:09:23.424221 135048246929216 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/0. I0421 13:09:23.424521 134919418279680 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:29:23.424488 I0421 13:09:23.435191 135048246929216 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 13:09:23.435431 134917241431808 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 13:09:23.435550 135048246929216 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776776962.8079965, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776776962.809768, 'checkpointer_blocking_duration_secs': 0.6149168014526367, 'get_old_steps_start_time': 1776776963.4247108, 'get_old_steps_duration_secs': 4.3392181396484375e-05, 'checkpoint_manager_blocking_start_time': 1776776962.8061216, 'checkpoint_manager_blocking_duration_secs': 0.6293909549713135} I0421 13:09:23.435663 135048246929216 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0421 13:09:23.435711 135048246929216 max_utils.py:750] Memstats: After params initialized: I0421 13:09:23.435773 135048246929216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0421 13:09:23.435809 135048246929216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0421 13:09:23.435835 135048246929216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0421 13:09:23.435858 135048246929216 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0421 13:09:23.752316 135048246929216 metric_logger.py:196] completed step: 0, seconds: 11.477, TFLOP/s/device: 1.184, Tokens/s/device: 178.439, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 13:09:23.924874 135048246929216 metric_logger.py:196] completed step: 1, seconds: 1.099, TFLOP/s/device: 12.365, Tokens/s/device: 1863.850, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 13:09:24.346903 135048246929216 metric_logger.py:196] completed step: 2, seconds: 0.026, TFLOP/s/device: 532.182, Tokens/s/device: 80216.208, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0421 13:09:24.494224 135048246929216 metric_logger.py:196] completed step: 3, seconds: 0.422, TFLOP/s/device: 32.187, Tokens/s/device: 4851.620, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0421 13:09:24.498718 135048246929216 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0421 13:09:24.788376 135048246929216 checkpointing.py:798] Waited 0.28962159156799316 seconds for step 5 to finish before starting checkpointing. I0421 13:09:24.791353 135048246929216 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 13:09:26.755810 2824 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 13:09:28.817098 134918302639872 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/0/items/array_metadatas/process_7 I0421 13:10:06.002017 134919418279680 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 36.854 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.860865354537964 s) (per-host) I0421 13:10:06.002128 134919418279680 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.577487s. I0421 13:10:12.023685 134919418279680 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 48.599028s. I0421 13:10:12.023925 134917241431808 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 13:10:12.023998 134917241431808 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 13:10:12.024042 134917241431808 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 13:10:12.027553 134917241431808 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 13:10:12.027777 135048246929216 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 13:10:12.027930 135048246929216 checkpoint_manager.py:1452] Waiting for previous save to complete took 47.236577 seconds. If this number is high, consider checkpointing less frequently. I0421 13:10:12.029705 135048246929216 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 5 I0421 13:10:12.032099 135048246929216 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/5. I0421 13:10:12.346657 135048246929216 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 13:10:12.346773 135048246929216 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 13:10:12.380544 135048246929216 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034845s I0421 13:10:12.380705 135048246929216 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 40.154 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03841686248779297 s) (per-host) I0421 13:10:12.380784 135048246929216 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038502s (batch_requests_ready=0.001794s, total_serialization_initiated=0.036616s, others=0.000092s) I0421 13:10:12.380903 135048246929216 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043563s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.043548) I0421 13:10:12.381561 135048246929216 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/5. I0421 13:10:12.381897 134917241431808 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:30:12.381861 I0421 13:10:12.390249 135048246929216 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 13:10:12.390531 134897897293568 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 13:10:12.390692 135048246929216 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776776964.7913232, 'wait_for_prev_duration_secs': 47.236576557159424, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776777012.0297592, 'checkpointer_blocking_duration_secs': 0.35229063034057617, 'get_old_steps_start_time': 1776777012.3820727, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776776964.788992, 'checkpoint_manager_blocking_duration_secs': 47.60166549682617} I0421 13:10:12.390822 135048246929216 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0421 13:10:12.391661 135048246929216 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.790, Tokens/s/device: 13383.347, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0421 13:10:12.397454 135048246929216 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.195, Tokens/s/device: 13896.617, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0421 13:10:17.497495 134918302639872 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/5/items/array_metadatas/process_7 I0421 13:10:35.631475 135048246929216 metric_logger.py:196] completed step: 6, seconds: 47.897, TFLOP/s/device: 0.284, Tokens/s/device: 42.758, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0421 13:10:35.778836 135048246929216 metric_logger.py:196] completed step: 7, seconds: 23.088, TFLOP/s/device: 0.589, Tokens/s/device: 88.705, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0421 13:10:35.926014 135048246929216 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.824, Tokens/s/device: 13388.509, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0421 13:10:36.072645 135048246929216 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0421 13:10:36.073345 135048246929216 checkpointing.py:798] Waited 0.0007197856903076172 seconds for step 9 to finish before starting checkpointing. I0421 13:10:36.075485 135048246929216 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 13:10:54.434709 134917241431808 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.527 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.09238314628601 s) (per-host) I0421 13:10:54.434846 134917241431808 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.052830s. I0421 13:11:00.260130 134917241431808 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 47.878099s. I0421 13:11:00.260412 134897897293568 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 13:11:00.260542 134897897293568 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 13:11:00.260611 134897897293568 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 13:11:00.262494 134897897293568 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 13:11:00.262677 135048246929216 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0421 13:11:00.262821 135048246929216 checkpoint_manager.py:1452] Waiting for previous save to complete took 24.187345 seconds. If this number is high, consider checkpointing less frequently. I0421 13:11:00.264484 135048246929216 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 9 I0421 13:11:00.266484 135048246929216 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/9. I0421 13:11:00.991014 135048246929216 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 13:11:00.991107 135048246929216 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 13:11:01.025856 135048246929216 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035747s I0421 13:11:01.025993 135048246929216 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 39.346 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.039205312728881836 s) (per-host) I0421 13:11:01.026040 135048246929216 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039262s (batch_requests_ready=0.001624s, total_serialization_initiated=0.037577s, others=0.000061s) I0421 13:11:01.026115 135048246929216 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043318s (all_items=0.000011s, per_item={'items': '0.00001144'}, temp_paths=0.043306) I0421 13:11:01.026793 135048246929216 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/9. I0421 13:11:01.027112 134918302639872 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:31:01.027076 I0421 13:11:01.029007 135048246929216 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 13:11:01.029277 134897897293568 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 13:11:01.029407 135048246929216 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776777036.0754526, 'wait_for_prev_duration_secs': 24.187344789505005, 'time_between_consecutive_saves_sec': 24.04785680770874, 'checkpointer_blocking_start_time': 1776777060.2645235, 'checkpointer_blocking_duration_secs': 0.762688159942627, 'get_old_steps_start_time': 1776777061.027231, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776777036.0735693, 'checkpoint_manager_blocking_duration_secs': 24.95580744743347} I0421 13:11:01.029577 135048246929216 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0421 13:11:01.029619 135048246929216 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 13:11:06.825299 134917270796032 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_16_zarr3_false/checkpoints/9/items/array_metadatas/process_7 I0421 13:11:42.699441 134918302639872 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.869 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.71263337135315 s) (per-host) I0421 13:11:42.699545 134918302639872 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.671850s. I0421 13:11:49.042849 134918302639872 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 48.015138s. I0421 13:11:49.043153 134897897293568 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 13:11:49.043270 134897897293568 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 13:11:49.043311 134897897293568 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 13:11:49.044846 134897897293568 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 13:11:49.045007 135048246929216 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 13:11:49.045153 135048246929216 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 13:11:49.046176 135048246929216 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.564, Tokens/s/device: 13952.189, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 13:12:01 UTC 2026 EXIT_CODE=0