XPK Start: Sat Apr 25 14:32:36 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 14:33:01.736345: 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 14:33:01.946666 139783692478272 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 14:33:10.988387 139783692478272 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 14:33:10.990757 139783692478272 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-rs-p1k1q-slice-job-0-0.mt-15-ocdbt-true-rs-p1k1q:8482 I0425 14:33:11.985277 139783692478272 max_utils.py:284] Jax distributed system initialized! I0425 14:33:18.119931 139783692478272 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 14:33:18.120040 139783692478272 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 14:33:18.120080 139783692478272 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 14:33:18.120134 139783692478272 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 14:33:18.817883 139783692478272 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 14:33:18.933050 139783692478272 checkpointing.py:688] Setting up checkpoint logger... I0425 14:33:18.933183 139783692478272 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 14:33:18.933230 139783692478272 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 14:33:18.933438 139783692478272 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 0x7f2135b1e6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 14:33:20.206709 139783692478272 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 14:33:20.206940 139783692478272 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 0x7f0cd842ff20>}, handler_registry=None I0425 14:33:20.207201 139783692478272 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f0cd842ff20>` 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 14:33:20.207258 139783692478272 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f0cd837aab0>` 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 14:33:20.207300 139783692478272 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 0x7f0cd842ff20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f0cd842ff20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f0cd837aab0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f0cd837aab0>}). I0425 14:33:20.207643 139783692478272 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 14:33:20.207715 139783692478272 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f0cd8437560> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 14:33:22.056721 139661716477696 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1777127344244358454, 'commit_timestamp_nsecs': 1777127383982248580, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0425 14:33:22.304404 139661716477696 checkpoint_manager.py:1871] Missing metrics for step 0 I0425 14:33:22.331855 139783692478272 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints I0425 14:33:22.737951 139783692478272 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/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f0bf06d86e0> I0425 14:33:22.738149 139783692478272 checkpointing.py:302] Checkpoint manager created! I0425 14:33:23.703644 139783692478272 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 14:33:23.703761 139783692478272 checkpointing.py:582] restoring from this run's directory step 9 I0425 14:33:23.959138 139783692478272 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9. I0425 14:33:24.206481 139783692478272 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9. I0425 14:33:24.206911 139783692478272 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9. I0425 14:33:25.620727 1461 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 14:33:31.704478 139783692478272 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 259.481 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.090133190155029 s) (per-host) I0425 14:33:31.713326 139783692478272 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 243.522 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.498822450637817 s) (per-host) I0425 14:33:34.168519 139783692478272 event_tracking.py:138] [process=6] [sync] Finished load in 10.21 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/9 I0425 14:33:34.168674 139783692478272 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints', 'checkpointer_start_time': 1777127603.9590213, 'checkpointer_duration_secs': 10.209622144699097, 'checkpoint_manager_start_time': 1777127603.7044077, 'checkpoint_manager_duration_secs': 10.464237928390503} fsdp: 32 I0425 14:33:34.244259 139783692478272 maxtext_utils.py:1889] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 14:33:34.244375 139783692478272 maxtext_utils.py:1889] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 14:33:34.244426 139783692478272 maxtext_utils.py:1889] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 14:33:34.244467 139783692478272 maxtext_utils.py:1889] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 14:33:34.244505 139783692478272 maxtext_utils.py:1889] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 14:33:34.244541 139783692478272 maxtext_utils.py:1889] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 14:33:34.244577 139783692478272 maxtext_utils.py:1889] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:33:34.244611 139783692478272 maxtext_utils.py:1889] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 14:33:34.244645 139783692478272 maxtext_utils.py:1889] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:33:34.244678 139783692478272 maxtext_utils.py:1889] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:33:34.244711 139783692478272 maxtext_utils.py:1889] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 14:33:34.244743 139783692478272 maxtext_utils.py:1889] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 14:33:34.648603 139783692478272 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 14:33:34.648705 139783692478272 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:33:34.654178 139783692478272 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 14:33:34.654235 139783692478272 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:33:34.670818 139783692478272 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 14:33:34.670878 139783692478272 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:33:34.686767 139783692478272 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 14:33:34.686827 139783692478272 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:33:34.712083 139783692478272 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 14:33:34.712171 139783692478272 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:33:34.728132 139783692478272 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 14:33:34.728193 139783692478272 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:33:34.744042 139783692478272 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 14:33:34.744112 139783692478272 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:33:34.773463 139783692478272 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 14:33:34.773530 139783692478272 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:33:34.793513 139783692478272 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 14:33:34.793573 139783692478272 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 14:33:46.411252 139783692478272 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 14:33:46.413763 139783692478272 metric_logger.py:301] number parameters: 1.104 billion I0425 14:33:58.031814 139783692478272 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0425 14:33:58.188251 139783692478272 checkpointing.py:798] Waited 0.15642261505126953 seconds for step 10 to finish before starting checkpointing. I0425 14:33:58.190801 139783692478272 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 14:33:58.192718 139783692478272 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0425 14:33:58.194054 139783692478272 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/10. I0425 14:33:58.876273 139783692478272 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 14:33:58.877279 139783692478272 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:33:58.877336 139783692478272 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 14:33:59.157056 139783692478272 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281555s I0425 14:33:59.157244 139783692478272 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.350 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28835058212280273 s) (per-host) I0425 14:33:59.157296 139783692478272 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.288412s (batch_requests_ready=0.002559s, total_serialization_initiated=0.285784s, others=0.000070s) I0425 14:33:59.157394 139783692478272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.292402s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.292385) I0425 14:33:59.158068 139783692478272 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.97 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/10. I0425 14:33:59.158427 139660021970688 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:53:59.158386 I0425 14:33:59.164326 139783692478272 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:33:59.164559 139661104105216 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:33:59.164673 139783692478272 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_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': 1777127638.1907823, 'wait_for_prev_duration_secs': 6.341934204101562e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777127638.192761, 'checkpointer_blocking_duration_secs': 0.9658269882202148, 'get_old_steps_start_time': 1777127639.158611, 'get_old_steps_duration_secs': 3.337860107421875e-05, 'checkpoint_manager_blocking_start_time': 1777127638.1886454, 'checkpoint_manager_blocking_duration_secs': 0.97599196434021} I0425 14:33:59.164780 139783692478272 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0425 14:33:59.164848 139783692478272 max_utils.py:750] Memstats: After params initialized: I0425 14:33:59.164898 139783692478272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 14:33:59.164930 139783692478272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 14:33:59.164957 139783692478272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 14:33:59.164981 139783692478272 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 14:33:59.488957 139783692478272 metric_logger.py:196] completed step: 10, seconds: 11.617, TFLOP/s/device: 1.170, Tokens/s/device: 176.294, total_weights: 65536, loss: 6.827, lm_loss: 6.827, perplexity: 922.537 I0425 14:33:59.671975 139783692478272 metric_logger.py:196] completed step: 11, seconds: 1.455, TFLOP/s/device: 9.336, Tokens/s/device: 1407.232, total_weights: 65536, loss: 6.522, lm_loss: 6.522, perplexity: 679.821 I0425 14:33:59.819428 139783692478272 metric_logger.py:196] completed step: 12, seconds: 0.026, TFLOP/s/device: 520.620, Tokens/s/device: 78473.446, total_weights: 65536, loss: 6.260, lm_loss: 6.260, perplexity: 523.013 I0425 14:33:59.966989 139783692478272 metric_logger.py:196] completed step: 13, seconds: 0.166, TFLOP/s/device: 81.953, Tokens/s/device: 12352.828, total_weights: 65536, loss: 6.039, lm_loss: 6.039, perplexity: 419.520 I0425 14:34:00.104393 139783692478272 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0425 14:34:00.397950 139783692478272 checkpointing.py:798] Waited 0.29354166984558105 seconds for step 15 to finish before starting checkpointing. I0425 14:34:00.400706 139783692478272 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 14:34:04.608786 139660055541504 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/10/items/array_metadatas/process_6 I0425 14:34:40.806727 139660021970688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.665 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.93779635429382 s) (per-host) I0425 14:34:40.806853 139660021970688 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.648309s. I0425 14:34:48.931665 139660021970688 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.773105s. I0425 14:34:48.931904 139661104105216 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:34:48.931971 139661104105216 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:34:48.932020 139661104105216 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:34:48.934287 139661104105216 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0425 14:34:48.934447 139783692478272 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0425 14:34:48.934586 139783692478272 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.533880 seconds. If this number is high, consider checkpointing less frequently. I0425 14:34:48.936434 139783692478272 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0425 14:34:48.939009 139783692478272 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/15. I0425 14:34:49.247769 139783692478272 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:34:49.247867 139783692478272 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 14:34:49.274047 139783692478272 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027874s I0425 14:34:49.274194 139783692478272 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 44.551 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03462481498718262 s) (per-host) I0425 14:34:49.274242 139783692478272 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.034681s (batch_requests_ready=0.002562s, total_serialization_initiated=0.032057s, others=0.000062s) I0425 14:34:49.274317 139783692478272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.039644s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.039631) I0425 14:34:49.275036 139783692478272 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/15. I0425 14:34:49.275299 139661104105216 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:54:49.275276 I0425 14:34:49.302026 139783692478272 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:34:49.302351 139660055541504 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:34:49.302510 139783692478272 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_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': 1777127640.400676, 'wait_for_prev_duration_secs': 48.53388023376465, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777127688.9364736, 'checkpointer_blocking_duration_secs': 0.33892369270324707, 'get_old_steps_start_time': 1777127689.2754216, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1777127640.3983498, 'checkpoint_manager_blocking_duration_secs': 48.90412664413452} I0425 14:34:49.302631 139783692478272 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0425 14:34:49.303789 139783692478272 metric_logger.py:196] completed step: 14, seconds: 0.146, TFLOP/s/device: 92.998, Tokens/s/device: 14017.700, total_weights: 65536, loss: 5.858, lm_loss: 5.858, perplexity: 350.067 I0425 14:34:49.312456 139783692478272 metric_logger.py:196] completed step: 15, seconds: 0.279, TFLOP/s/device: 48.732, Tokens/s/device: 7345.478, total_weights: 65536, loss: 5.713, lm_loss: 5.713, perplexity: 302.780 I0425 14:34:54.724036 139655714006784 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/15/items/array_metadatas/process_6 I0425 14:35:10.750844 139783692478272 metric_logger.py:196] completed step: 16, seconds: 49.206, TFLOP/s/device: 0.276, Tokens/s/device: 41.621, total_weights: 65536, loss: 5.599, lm_loss: 5.599, perplexity: 270.201 I0425 14:35:10.898238 139783692478272 metric_logger.py:196] completed step: 17, seconds: 21.293, TFLOP/s/device: 0.638, Tokens/s/device: 96.180, total_weights: 65536, loss: 5.511, lm_loss: 5.511, perplexity: 247.443 I0425 14:35:11.045905 139783692478272 metric_logger.py:196] completed step: 18, seconds: 0.154, TFLOP/s/device: 88.474, Tokens/s/device: 13335.764, total_weights: 65536, loss: 5.442, lm_loss: 5.442, perplexity: 230.909 I0425 14:35:11.193999 139783692478272 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0425 14:35:11.194720 139783692478272 checkpointing.py:798] Waited 0.0007305145263671875 seconds for step 19 to finish before starting checkpointing. I0425 14:35:11.196840 139783692478272 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 14:35:31.694810 139661104105216 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.206 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.455199241638184 s) (per-host) I0425 14:35:31.694936 139661104105216 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.419561s. I0425 14:35:39.473309 139661104105216 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.197918s. I0425 14:35:39.473598 139660055541504 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:35:39.473761 139660055541504 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:35:39.473815 139660055541504 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:35:39.475288 139660055541504 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0425 14:35:39.475471 139783692478272 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0425 14:35:39.475602 139783692478272 checkpoint_manager.py:1452] Waiting for previous save to complete took 28.278769 seconds. If this number is high, consider checkpointing less frequently. I0425 14:35:39.477274 139783692478272 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0425 14:35:39.479399 139783692478272 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/19. I0425 14:35:39.770114 139783692478272 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:35:39.770209 139783692478272 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 14:35:39.796313 139783692478272 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027747s I0425 14:35:39.796508 139783692478272 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 44.585 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.034598588943481445 s) (per-host) I0425 14:35:39.796564 139783692478272 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.034665s (batch_requests_ready=0.002655s, total_serialization_initiated=0.031932s, others=0.000078s) I0425 14:35:39.796670 139783692478272 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.038760s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.038749) I0425 14:35:39.797268 139783692478272 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/19. I0425 14:35:39.797538 139655714006784 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:55:39.797497 I0425 14:35:39.802165 139783692478272 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:35:39.802449 139660055541504 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:35:39.802603 139783692478272 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_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': 1777127711.1968105, 'wait_for_prev_duration_secs': 28.27876853942871, 'time_between_consecutive_saves_sec': 22.26248288154602, 'checkpointer_blocking_start_time': 1777127739.4773133, 'checkpointer_blocking_duration_secs': 0.32036685943603516, 'get_old_steps_start_time': 1777127739.7976933, 'get_old_steps_duration_secs': 2.2411346435546875e-05, 'checkpoint_manager_blocking_start_time': 1777127711.194962, 'checkpoint_manager_blocking_duration_secs': 28.607608318328857} I0425 14:35:39.802714 139783692478272 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0425 14:35:39.802758 139783692478272 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 14:35:45.133588 139656792352512 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260425_121355/nnx_xpk_feat_nnx_post_train_fixes_20260425_121355_15_ocdbt_true/checkpoints/19/items/array_metadatas/process_6 I0425 14:36:21.074784 139655714006784 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.235 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.312844038009644 s) (per-host) I0425 14:36:21.074888 139655714006784 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.277236s. I0425 14:36:28.899910 139655714006784 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.102241s. I0425 14:36:28.900227 139660055541504 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:36:28.900346 139660055541504 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:36:28.900386 139660055541504 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:36:28.901984 139660055541504 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0425 14:36:28.902167 139783692478272 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0425 14:36:28.902347 139783692478272 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 14:36:28.903263 139783692478272 metric_logger.py:196] completed step: 19, seconds: 0.148, TFLOP/s/device: 91.990, Tokens/s/device: 13865.663, total_weights: 65536, loss: 5.385, lm_loss: 5.385, perplexity: 218.045 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 14:36:41 UTC 2026 EXIT_CODE=0