XPK Start: Sat Apr 25 13:03:45 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. 2026-04-25 13:04:09.920886: 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) I0425 13:04:10.133887 136401630017344 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 13:04:19.175352 136401630017344 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 13:04:19.177858 136401630017344 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-81teh-slice-job-0-0.mt-15-ocdbt-true-sv-81teh:8482 I0425 13:04:20.133727 136401630017344 max_utils.py:284] Jax distributed system initialized! I0425 13:04:25.268914 136401630017344 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 13:04:25.269018 136401630017344 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 13:04:25.269060 136401630017344 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 I0425 13:04:25.269210 136401630017344 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 13:04:25.966345 136401630017344 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 13:04:25.967006 136401630017344 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 13:04:25.967237 136401630017344 checkpointing.py:688] Setting up checkpoint logger... I0425 13:04:25.967304 136401630017344 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 13:04:25.967370 136401630017344 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 13:04:25.967741 136401630017344 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 0x7c0de1b1ce90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 13:04:30.386779 136401630017344 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 13:04:30.387118 136401630017344 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 0x7bfa28627140>}, handler_registry=None I0425 13:04:30.387362 136401630017344 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bfa28627140>` 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`. I0425 13:04:30.387413 136401630017344 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfa286267b0>` 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`. I0425 13:04:30.387451 136401630017344 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 0x7bfa28627140>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bfa28627140>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfa286267b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfa286267b0>}). I0425 13:04:30.387784 136401630017344 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 13:04:30.387856 136401630017344 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bf9443e1e40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 13:04:31.086543 136401630017344 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints I0425 13:04:31.479614 136401630017344 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_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bfa28626db0> I0425 13:04:31.479786 136401630017344 checkpointing.py:302] Checkpoint manager created! I0425 13:04:32.413323 136401630017344 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 13:04:32.413430 136401630017344 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 13:04:32.792486 136401630017344 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 13:04:32.792582 136401630017344 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 13:04:32.809276 136401630017344 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 13:04:32.809335 136401630017344 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 13:04:32.833750 136401630017344 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:04:32.833821 136401630017344 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 13:04:32.850917 136401630017344 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:04:32.850981 136401630017344 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 13:04:32.867910 136401630017344 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 13:04:32.867971 136401630017344 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 13:04:32.893140 136401630017344 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 13:04:32.893208 136401630017344 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 13:04:32.914629 136401630017344 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 13:04:32.914695 136401630017344 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 13:04:33.131357 136401630017344 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 13:04:33.131462 136401630017344 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 13:04:34.575881 136401630017344 maxtext_utils.py:1880] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0425 13:04:34.576008 136401630017344 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 13:04:34.576061 136401630017344 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0425 13:04:34.576131 136401630017344 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0425 13:04:34.576186 136401630017344 maxtext_utils.py:1880] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0425 13:04:34.576224 136401630017344 maxtext_utils.py:1880] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0425 13:04:34.576275 136401630017344 maxtext_utils.py:1880] 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) I0425 13:04:34.576330 136401630017344 maxtext_utils.py:1880] 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') I0425 13:04:34.576371 136401630017344 maxtext_utils.py:1880] 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) I0425 13:04:34.576406 136401630017344 maxtext_utils.py:1880] 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) I0425 13:04:34.576452 136401630017344 maxtext_utils.py:1880] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0425 13:04:34.576507 136401630017344 maxtext_utils.py:1880] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0425 13:04:35.075914 136401630017344 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 13:04:35.076006 136401630017344 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0425 13:04:35.091569 136401630017344 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0425 13:04:35.091629 136401630017344 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0425 13:04:46.217190 136401630017344 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. I0425 13:04:46.218055 136401630017344 metric_logger.py:301] number parameters: 1.104 billion I0425 13:04:57.638551 136401630017344 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 13:04:57.792505 136401630017344 checkpointing.py:798] Waited 0.15393924713134766 seconds for step 0 to finish before starting checkpointing. I0425 13:04:57.795123 136401630017344 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 13:04:57.796862 136401630017344 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0425 13:04:57.798323 136401630017344 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/0. I0425 13:04:58.574381 136401630017344 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 13:04:58.575427 136401630017344 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 13:04:58.575491 136401630017344 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 I0425 13:04:58.855881 136401630017344 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281579s I0425 13:04:58.856051 136401630017344 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 2.145 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7190768718719482 s) (per-host) I0425 13:04:58.856117 136401630017344 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.719152s (batch_requests_ready=0.434204s, total_serialization_initiated=0.284865s, others=0.000082s) I0425 13:04:58.856213 136401630017344 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.723235s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.723218) I0425 13:04:58.857024 136401630017344 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/0. I0425 13:04:58.857386 136276264900352 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 13:24:58.857349 I0425 13:04:58.859471 136401630017344 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 13:04:58.859793 136273578682112 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 13:04:58.859929 136401630017344 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777122297.7950907, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777122297.7969007, 'checkpointer_blocking_duration_secs': 1.060645580291748, 'get_old_steps_start_time': 1777122298.857572, 'get_old_steps_duration_secs': 3.4809112548828125e-05, 'checkpoint_manager_blocking_start_time': 1777122297.793015, 'checkpoint_manager_blocking_duration_secs': 1.06687593460083} I0425 13:04:58.860090 136401630017344 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 13:04:58.860152 136401630017344 max_utils.py:750] Memstats: After params initialized: I0425 13:04:58.860200 136401630017344 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 13:04:58.860232 136401630017344 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 13:04:58.860259 136401630017344 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 13:04:58.860283 136401630017344 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 13:04:59.173599 136401630017344 metric_logger.py:196] completed step: 0, seconds: 11.420, TFLOP/s/device: 1.190, Tokens/s/device: 179.329, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0425 13:04:59.355611 136401630017344 metric_logger.py:196] completed step: 1, seconds: 1.534, TFLOP/s/device: 8.860, Tokens/s/device: 1335.501, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0425 13:04:59.777494 136401630017344 metric_logger.py:196] completed step: 2, seconds: 0.035, TFLOP/s/device: 391.436, Tokens/s/device: 59001.469, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0425 13:04:59.924842 136401630017344 metric_logger.py:196] completed step: 3, seconds: 0.422, TFLOP/s/device: 32.173, Tokens/s/device: 4849.484, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0425 13:04:59.929988 136401630017344 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 13:05:00.218843 136401630017344 checkpointing.py:798] Waited 0.2888174057006836 seconds for step 5 to finish before starting checkpointing. I0425 13:05:00.222304 136401630017344 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 13:05:01.381801 2746 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 13:05:03.418343 136274670565120 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_6 I0425 13:05:34.529007 136276264900352 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 43.405 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.39199352264404 s) (per-host) I0425 13:05:34.529149 136276264900352 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.671649s. I0425 13:05:42.308567 136276264900352 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 43.451050s. I0425 13:05:42.308837 136273578682112 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 13:05:42.308964 136273578682112 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 13:05:42.309015 136273578682112 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 13:05:42.311377 136273578682112 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 13:05:42.311557 136401630017344 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0425 13:05:42.311693 136401630017344 checkpoint_manager.py:1452] Waiting for previous save to complete took 42.089389 seconds. If this number is high, consider checkpointing less frequently. I0425 13:05:42.313590 136401630017344 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0425 13:05:42.316012 136401630017344 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/5. I0425 13:05:42.855290 136401630017344 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 13:05:42.855386 136401630017344 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 I0425 13:05:42.896413 136401630017344 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.042164s I0425 13:05:42.896564 136401630017344 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.754 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045700788497924805 s) (per-host) I0425 13:05:42.896614 136401630017344 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045758s (batch_requests_ready=0.001896s, total_serialization_initiated=0.043799s, others=0.000064s) I0425 13:05:42.896703 136401630017344 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.050734s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.050718) I0425 13:05:42.897456 136401630017344 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.58 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/5. I0425 13:05:42.897793 136248358328064 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 13:25:42.897757 I0425 13:05:42.900451 136401630017344 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 13:05:42.900797 136273578682112 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 13:05:42.900955 136401630017344 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777122300.2222729, 'wait_for_prev_duration_secs': 42.089388847351074, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777122342.3136296, 'checkpointer_blocking_duration_secs': 0.5843112468719482, 'get_old_steps_start_time': 1777122342.8979642, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1777122300.2196088, 'checkpoint_manager_blocking_duration_secs': 42.681312084198} I0425 13:05:42.901069 136401630017344 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 13:05:42.901899 136401630017344 metric_logger.py:196] completed step: 4, seconds: 0.154, TFLOP/s/device: 88.469, Tokens/s/device: 13335.070, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0425 13:05:42.907879 136401630017344 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.289, Tokens/s/device: 13910.775, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0425 13:05:48.031022 136274670565120 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_6 I0425 13:06:05.133080 136401630017344 metric_logger.py:196] completed step: 6, seconds: 42.976, TFLOP/s/device: 0.316, Tokens/s/device: 47.654, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0425 13:06:05.280444 136401630017344 metric_logger.py:196] completed step: 7, seconds: 22.079, TFLOP/s/device: 0.615, Tokens/s/device: 92.758, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0425 13:06:05.427897 136401630017344 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.263, Tokens/s/device: 13454.653, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0425 13:06:05.574459 136401630017344 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 13:06:05.575155 136401630017344 checkpointing.py:798] Waited 0.0007157325744628906 seconds for step 9 to finish before starting checkpointing. I0425 13:06:05.577192 136401630017344 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 13:06:24.048526 136248358328064 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.342 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.197622537612915 s) (per-host) I0425 13:06:24.048667 136248358328064 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.150756s. I0425 13:06:31.357306 136248358328064 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.459378s. I0425 13:06:31.357609 136273578682112 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 13:06:31.357721 136273578682112 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 13:06:31.357769 136273578682112 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 13:06:31.360106 136273578682112 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 13:06:31.360322 136401630017344 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0425 13:06:31.360499 136401630017344 checkpoint_manager.py:1452] Waiting for previous save to complete took 25.783304 seconds. If this number is high, consider checkpointing less frequently. I0425 13:06:31.361918 136401630017344 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0425 13:06:31.364933 136401630017344 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9. I0425 13:06:31.723785 136401630017344 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0425 13:06:31.723881 136401630017344 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 I0425 13:06:31.764877 136401630017344 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.041933s I0425 13:06:31.765041 136401630017344 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 33.924 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0454709529876709 s) (per-host) I0425 13:06:31.765105 136401630017344 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045534s (batch_requests_ready=0.001760s, total_serialization_initiated=0.043706s, others=0.000068s) I0425 13:06:31.765203 136401630017344 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049552s (all_items=0.000010s, per_item={'items': '0.00001049'}, temp_paths=0.049541) I0425 13:06:31.765958 136401630017344 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.40 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9. I0425 13:06:31.766295 136274670565120 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 13:26:31.766257 I0425 13:06:31.768301 136401630017344 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 13:06:31.768664 136273578682112 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 13:06:31.768798 136401630017344 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777122365.57716, 'wait_for_prev_duration_secs': 25.78330397605896, 'time_between_consecutive_saves_sec': 23.26574182510376, 'checkpointer_blocking_start_time': 1777122391.3619595, 'checkpointer_blocking_duration_secs': 0.40444350242614746, 'get_old_steps_start_time': 1777122391.7664218, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1777122365.5753791, 'checkpoint_manager_blocking_duration_secs': 26.1933856010437} I0425 13:06:31.768958 136401630017344 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 13:06:31.769001 136401630017344 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 13:06:36.818028 136272483948288 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/linen_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_6 I0425 13:07:12.947967 136274670565120 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.313 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.228357553482056 s) (per-host) I0425 13:07:12.948106 136274670565120 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.181182s. I0425 13:07:19.901831 136274670565120 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.134903s. I0425 13:07:19.902142 136273578682112 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 13:07:19.902258 136273578682112 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 13:07:19.902314 136273578682112 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 13:07:19.903944 136273578682112 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 13:07:19.904126 136401630017344 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0425 13:07:19.904281 136401630017344 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 13:07:19.905336 136401630017344 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.197, Tokens/s/device: 13896.900, 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: Sat Apr 25 13:07:30 UTC 2026 EXIT_CODE=0