XPK Start: Sat Apr 25 14:37:43 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 14:38:08.144622: 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 14:38:08.356883 137486310885184 max_utils.py:273] Attempting to initialize the jax distributed system... I0425 14:38:17.397529 137486310885184 distributed.py:149] Starting JAX distributed service on [::]:8482 I0425 14:38:17.399955 137486310885184 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-n0k3v-slice-job-0-0.mt-15-ocdbt-false-sv-n0k3v:8482 I0425 14:38:18.676448 137486310885184 max_utils.py:284] Jax distributed system initialized! I0425 14:38:24.869347 137486310885184 max_utils.py:800] System Information: Jax Version: 0.9.2 I0425 14:38:24.869451 137486310885184 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0425 14:38:24.869492 137486310885184 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 14:38:24.869528 137486310885184 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 14:38:25.569635 137486310885184 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 14:38:25.684587 137486310885184 checkpointing.py:688] Setting up checkpoint logger... I0425 14:38:25.684706 137486310885184 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0425 14:38:25.684754 137486310885184 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0425 14:38:25.684954 137486310885184 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7d0a4ee4c0b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 14:38:28.544862 137486310885184 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 14:38:28.545109 137486310885184 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 0x7cf4d87efb90>}, handler_registry=None I0425 14:38:28.545357 137486310885184 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cf4d87efb90>` 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 14:38:28.545407 137486310885184 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cf5b81e0440>` 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 14:38:28.545471 137486310885184 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 0x7cf4d87efb90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cf4d87efb90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cf5b81e0440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cf5b81e0440>}). I0425 14:38:28.545806 137486310885184 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0425 14:38:28.545876 137486310885184 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cf4d836f6a0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0425 14:38:29.285801 137486310885184 checkpoint_manager.py:1812] Found 0 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_false/checkpoints I0425 14:38:29.722047 137486310885184 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_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cf4d87da1e0> I0425 14:38:29.722245 137486310885184 checkpointing.py:302] Checkpoint manager created! I0425 14:38:30.708930 137486310885184 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 14:38:30.709050 137486310885184 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 14:38:32.083547 137486310885184 maxtext_utils.py:1889] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 14:38:32.083667 137486310885184 maxtext_utils.py:1889] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 14:38:32.083719 137486310885184 maxtext_utils.py:1889] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 14:38:32.083761 137486310885184 maxtext_utils.py:1889] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 14:38:32.083798 137486310885184 maxtext_utils.py:1889] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 14:38:32.083832 137486310885184 maxtext_utils.py:1889] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 14:38:32.083867 137486310885184 maxtext_utils.py:1889] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:38:32.083902 137486310885184 maxtext_utils.py:1889] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 14:38:32.083935 137486310885184 maxtext_utils.py:1889] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:38:32.083966 137486310885184 maxtext_utils.py:1889] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 14:38:32.083997 137486310885184 maxtext_utils.py:1889] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 14:38:32.084029 137486310885184 maxtext_utils.py:1889] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 14:38:32.488929 137486310885184 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 14:38:32.489022 137486310885184 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:38:32.494342 137486310885184 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 14:38:32.494397 137486310885184 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:38:32.517582 137486310885184 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 14:38:32.517691 137486310885184 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:38:32.533768 137486310885184 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0425 14:38:32.533829 137486310885184 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0425 14:38:32.559494 137486310885184 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 14:38:32.559567 137486310885184 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:38:32.575797 137486310885184 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 14:38:32.575858 137486310885184 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:38:32.591788 137486310885184 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 14:38:32.591849 137486310885184 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:38:32.621962 137486310885184 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0425 14:38:32.622038 137486310885184 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0425 14:38:32.642212 137486310885184 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 14:38:32.642274 137486310885184 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0425 14:38:44.235170 137486310885184 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 14:38:44.237589 137486310885184 metric_logger.py:301] number parameters: 1.104 billion I0425 14:38:55.559623 137486310885184 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 14:38:55.711062 137486310885184 checkpointing.py:798] Waited 0.15142560005187988 seconds for step 0 to finish before starting checkpointing. I0425 14:38:55.713315 137486310885184 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 14:38:55.715080 137486310885184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0425 14:38:55.716559 137486310885184 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_false/checkpoints/0. I0425 14:38:56.004648 137486310885184 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 14:38:56.005647 137486310885184 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:38:56.005707 137486310885184 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 14:38:56.279463 137486310885184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275604s I0425 14:38:56.279649 137486310885184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.420 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28462862968444824 s) (per-host) I0425 14:38:56.279700 137486310885184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.284690s (batch_requests_ready=0.002963s, total_serialization_initiated=0.281658s, others=0.000069s) I0425 14:38:56.279799 137486310885184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.288870s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.288852) I0425 14:38:56.280607 137486310885184 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.57 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_false/checkpoints/0. I0425 14:38:56.280901 137356837852928 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:58:56.280869 I0425 14:38:56.292955 137486310885184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:38:56.293299 137355259930368 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:38:56.293483 137486310885184 standard_logger.py:34] {'step': 0, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777127935.7132971, 'wait_for_prev_duration_secs': 6.008148193359375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777127935.7151322, 'checkpointer_blocking_duration_secs': 0.5659143924713135, 'get_old_steps_start_time': 1777127936.281072, 'get_old_steps_duration_secs': 4.673004150390625e-05, 'checkpoint_manager_blocking_start_time': 1777127935.711485, 'checkpoint_manager_blocking_duration_secs': 0.5819611549377441} I0425 14:38:56.293595 137486310885184 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 14:38:56.293663 137486310885184 max_utils.py:750] Memstats: After params initialized: I0425 14:38:56.293711 137486310885184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0425 14:38:56.293741 137486310885184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0425 14:38:56.293767 137486310885184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0425 14:38:56.293791 137486310885184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0425 14:38:56.613677 137486310885184 metric_logger.py:196] completed step: 0, seconds: 11.321, TFLOP/s/device: 1.200, Tokens/s/device: 180.903, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0425 14:38:56.779920 137486310885184 metric_logger.py:196] completed step: 1, seconds: 1.052, TFLOP/s/device: 12.914, Tokens/s/device: 1946.565, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0425 14:38:57.183509 137486310885184 metric_logger.py:196] completed step: 2, seconds: 0.020, TFLOP/s/device: 672.630, Tokens/s/device: 101386.139, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38565.637 I0425 14:38:57.330966 137486310885184 metric_logger.py:196] completed step: 3, seconds: 0.393, TFLOP/s/device: 34.534, Tokens/s/device: 5205.381, total_weights: 65536, loss: 9.979, lm_loss: 9.979, perplexity: 21564.932 I0425 14:38:57.336699 137486310885184 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0425 14:38:57.625645 137486310885184 checkpointing.py:798] Waited 0.28891801834106445 seconds for step 5 to finish before starting checkpointing. I0425 14:38:57.628348 137486310885184 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 14:38:59.655990 2830 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 14:39:01.662992 137355268323072 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_false/checkpoints/0/items/array_metadatas/process_6 I0425 14:39:06.997588 137356837852928 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 143.567 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.002530097961426 s) (per-host) I0425 14:39:06.997713 137356837852928 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.716703s. I0425 14:39:14.320194 137356837852928 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 18.039167s. I0425 14:39:14.320513 137355259930368 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:39:14.320631 137355259930368 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:39:14.320682 137355259930368 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:39:14.323075 137355259930368 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 14:39:14.323274 137486310885184 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 14:39:14.323415 137486310885184 checkpoint_manager.py:1452] Waiting for previous save to complete took 16.695066 seconds. If this number is high, consider checkpointing less frequently. I0425 14:39:14.325639 137486310885184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0425 14:39:14.328115 137486310885184 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_false/checkpoints/5. I0425 14:39:14.621762 137486310885184 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:39:14.621856 137486310885184 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 14:39:14.655485 137486310885184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.035322s I0425 14:39:14.655664 137486310885184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 36.245 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04255938529968262 s) (per-host) I0425 14:39:14.655716 137486310885184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042622s (batch_requests_ready=0.002538s, total_serialization_initiated=0.040016s, others=0.000069s) I0425 14:39:14.655816 137486310885184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.047537s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.047520) I0425 14:39:14.656575 137486310885184 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.33 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_false/checkpoints/5. I0425 14:39:14.656911 137355259930368 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:59:14.656875 I0425 14:39:14.676356 137486310885184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:39:14.676643 137356334065408 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:39:14.676814 137486310885184 standard_logger.py:34] {'step': 5, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777127937.6283197, 'wait_for_prev_duration_secs': 16.695066452026367, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777127954.3256776, 'checkpointer_blocking_duration_secs': 0.3313753604888916, 'get_old_steps_start_time': 1777127954.657077, 'get_old_steps_duration_secs': 4.4345855712890625e-05, 'checkpoint_manager_blocking_start_time': 1777127937.6259844, 'checkpoint_manager_blocking_duration_secs': 17.050792932510376} I0425 14:39:14.676934 137486310885184 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0425 14:39:14.678083 137486310885184 metric_logger.py:196] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.650, Tokens/s/device: 12457.876, total_weights: 65536, loss: 9.457, lm_loss: 9.457, perplexity: 12802.436 I0425 14:39:14.685997 137486310885184 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.497, Tokens/s/device: 13942.216, total_weights: 65536, loss: 8.955, lm_loss: 8.955, perplexity: 7748.778 I0425 14:39:19.725661 137355268323072 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_false/checkpoints/5/items/array_metadatas/process_6 I0425 14:39:25.601982 137355259930368 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 143.746 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.988839387893677 s) (per-host) I0425 14:39:25.602123 137355259930368 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.945099s. I0425 14:39:31.359018 137355259930368 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.701978s. I0425 14:39:31.359323 137356334065408 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:39:31.359438 137356334065408 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:39:31.359488 137356334065408 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:39:31.361190 137356334065408 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0425 14:39:38.030844 137486310885184 metric_logger.py:196] completed step: 6, seconds: 17.347, TFLOP/s/device: 0.783, Tokens/s/device: 118.059, total_weights: 65536, loss: 8.465, lm_loss: 8.465, perplexity: 4743.984 I0425 14:39:38.178429 137486310885184 metric_logger.py:196] completed step: 7, seconds: 23.199, TFLOP/s/device: 0.586, Tokens/s/device: 88.278, total_weights: 65536, loss: 7.998, lm_loss: 7.998, perplexity: 2976.266 I0425 14:39:38.325733 137486310885184 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.611, Tokens/s/device: 13356.377, total_weights: 65536, loss: 7.567, lm_loss: 7.567, perplexity: 1933.479 I0425 14:39:38.474063 137486310885184 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 14:39:38.474760 137486310885184 checkpointing.py:798] Waited 0.0007092952728271484 seconds for step 9 to finish before starting checkpointing. I0425 14:39:38.476657 137486310885184 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 14:39:38.478718 137486310885184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0425 14:39:38.480057 137486310885184 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_false/checkpoints/9. I0425 14:39:38.782835 137486310885184 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0425 14:39:38.783176 137486310885184 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 14:39:38.818487 137486310885184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037340s I0425 14:39:38.818637 137486310885184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 34.921 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04417276382446289 s) (per-host) I0425 14:39:38.818700 137486310885184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.044246s (batch_requests_ready=0.002627s, total_serialization_initiated=0.041540s, others=0.000079s) I0425 14:39:38.818812 137486310885184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048171s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.048160) I0425 14:39:38.819555 137486310885184 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_false/checkpoints/9. I0425 14:39:38.819890 137350394914560 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-25 14:59:38.819853 I0425 14:39:38.821881 137486310885184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 14:39:38.822134 137332879501056 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 14:39:38.822297 137486310885184 standard_logger.py:34] {'step': 9, '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_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777127978.4766264, 'wait_for_prev_duration_secs': 7.200241088867188e-05, 'time_between_consecutive_saves_sec': 7.115394353866577, 'checkpointer_blocking_start_time': 1777127978.4787562, 'checkpointer_blocking_duration_secs': 0.3412768840789795, 'get_old_steps_start_time': 1777127978.8200517, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1777127978.474995, 'checkpoint_manager_blocking_duration_secs': 0.34726572036743164} I0425 14:39:38.822418 137486310885184 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 14:39:38.822464 137486310885184 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 14:39:43.850077 137334364288768 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_false/checkpoints/9/items/array_metadatas/process_6 I0425 14:39:49.181552 137350394914560 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 151.781 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.407056093215942 s) (per-host) I0425 14:39:49.181680 137350394914560 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.361678s. I0425 14:39:55.499352 137350394914560 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.679327s. I0425 14:39:55.499652 137332879501056 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 14:39:55.499769 137332879501056 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 14:39:55.499810 137332879501056 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 14:39:55.501419 137332879501056 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 14:39:55.501598 137486310885184 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 14:39:55.501763 137486310885184 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 14:39:55.502764 137486310885184 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.821, Tokens/s/device: 13840.269, total_weights: 65536, loss: 7.176, lm_loss: 7.176, perplexity: 1307.366 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 25 14:40:08 UTC 2026 EXIT_CODE=0