XPK Start: Thu Apr 23 11:24:34 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. 2026-04-23 11:24:59.735883: 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) I0423 11:24:59.945837 135142985008960 max_utils.py:273] Attempting to initialize the jax distributed system... I0423 11:25:08.987545 135142985008960 distributed.py:149] Starting JAX distributed service on [::]:8482 I0423 11:25:08.989896 135142985008960 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-bn72v-slice-job-0-0.mt-06-grad-accum-bn72v:8482 I0423 11:25:10.077167 135142985008960 max_utils.py:284] Jax distributed system initialized! I0423 11:25:16.213277 135142985008960 max_utils.py:800] System Information: Jax Version: 0.9.2 I0423 11:25:16.213381 135142985008960 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0423 11:25:16.213421 135142985008960 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 I0423 11:25:16.213456 135142985008960 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 11:25:16.903559 135142985008960 maxtext_utils.py:1732] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 11:25:17.014910 135142985008960 checkpointing.py:688] Setting up checkpoint logger... I0423 11:25:17.015030 135142985008960 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 11:25:17.015077 135142985008960 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0423 11:25:17.015274 135142985008960 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 0x7ae8b6038230>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 11:25:19.921010 135142985008960 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 11:25:19.921201 135142985008960 checkpoint_manager.py:708] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad4681cbef0>}, handler_registry=None I0423 11:25:19.921440 135142985008960 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad4681cbef0>` 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`. I0423 11:25:19.921490 135142985008960 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad46819f260>` 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`. I0423 11:25:19.921525 135142985008960 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 0x7ad4681cbef0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ad4681cbef0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad46819f260>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ad46819f260>}). I0423 11:25:19.921860 135142985008960 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0423 11:25:19.921929 135142985008960 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ad3903cb4c0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0423 11:25:21.426927 135142985008960 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints I0423 11:25:21.739292 135142985008960 checkpoint_manager.py:929] [process=5][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ad4687edee0> I0423 11:25:21.739466 135142985008960 checkpointing.py:302] Checkpoint manager created! I0423 11:25:23.505034 135142985008960 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 11:25:23.505154 135142985008960 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 11:25:25.248626 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown. I0423 11:25:25.248730 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,). I0423 11:25:25.264110 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown. I0423 11:25:25.264172 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None). I0423 11:25:25.294760 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown. I0423 11:25:25.294830 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp'). I0423 11:25:25.310171 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown. I0423 11:25:25.310227 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None). I0423 11:25:25.340967 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown. I0423 11:25:25.341035 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None). I0423 11:25:25.356243 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown. I0423 11:25:25.356303 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp'). I0423 11:25:25.401988 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown. I0423 11:25:25.402063 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None). I0423 11:25:25.417331 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown. I0423 11:25:25.417387 135142985008960 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp'). I0423 11:25:25.650466 135142985008960 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 11:25:25.650621 135142985008960 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 11:25:25.657021 135142985008960 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 11:25:25.657134 135142985008960 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 11:25:25.674636 135142985008960 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 11:25:25.674788 135142985008960 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 11:25:25.691530 135142985008960 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 11:25:25.691685 135142985008960 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 11:25:25.728834 135142985008960 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 11:25:25.729037 135142985008960 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 11:25:25.745866 135142985008960 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 11:25:25.745998 135142985008960 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 11:25:25.762528 135142985008960 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 11:25:25.762674 135142985008960 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 11:25:25.796285 135142985008960 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 11:25:25.796471 135142985008960 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 11:25:25.818709 135142985008960 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 11:25:25.818789 135142985008960 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 11:25:36.982399 135142985008960 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0423 11:25:36.984809 135142985008960 metric_logger.py:301] number parameters: 1.104 billion I0423 11:25:38.654903 135142985008960 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 11:25:50.532458 135142985008960 checkpointing.py:798] Waited 11.87754201889038 seconds for step 0 to finish before starting checkpointing. I0423 11:25:50.535008 135142985008960 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 11:25:50.536561 135142985008960 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0 I0423 11:25:50.537966 135142985008960 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/0. I0423 11:25:51.300527 135142985008960 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 11:25:51.301607 135142985008960 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0423 11:25:51.301676 135142985008960 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 I0423 11:25:51.571907 135142985008960 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.272123s I0423 11:25:51.572088 135142985008960 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.487 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28113460540771484 s) (per-host) I0423 11:25:51.572140 135142985008960 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281198s (batch_requests_ready=0.003018s, total_serialization_initiated=0.278109s, others=0.000072s) I0423 11:25:51.572236 135142985008960 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285311s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.285292) I0423 11:25:51.573087 135142985008960 event_tracking.py:125] [process=5] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/0. I0423 11:25:51.573395 135014805120768 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 11:45:51.573361 I0423 11:25:51.586555 135142985008960 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 11:25:51.586847 135014280849152 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 11:25:51.587007 135142985008960 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776943550.53499, 'wait_for_prev_duration_secs': 5.936622619628906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776943550.5366008, 'checkpointer_blocking_duration_secs': 1.0369462966918945, 'get_old_steps_start_time': 1776943551.5735726, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776943550.5328777, 'checkpoint_manager_blocking_duration_secs': 1.0540907382965088} I0423 11:25:51.587116 135142985008960 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 11:25:51.587182 135142985008960 max_utils.py:750] Memstats: After params initialized: I0423 11:25:51.587232 135142985008960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0423 11:25:51.587265 135142985008960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0423 11:25:51.587292 135142985008960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0423 11:25:51.587316 135142985008960 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0423 11:25:51.904951 135142985008960 metric_logger.py:196] completed step: 0, seconds: 1.669, TFLOP/s/device: 32.563, Tokens/s/device: 4908.234, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598 I0423 11:25:52.498357 135142985008960 metric_logger.py:196] completed step: 1, seconds: 13.248, TFLOP/s/device: 4.102, Tokens/s/device: 618.346, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598 I0423 11:25:53.076644 135142985008960 metric_logger.py:196] completed step: 2, seconds: 0.016, TFLOP/s/device: 3359.201, Tokens/s/device: 506335.373, total_weights: 262144, loss: 10.561, lm_loss: 10.561, perplexity: 38601.980 I0423 11:25:53.655188 135142985008960 metric_logger.py:196] completed step: 3, seconds: 0.585, TFLOP/s/device: 92.936, Tokens/s/device: 14008.328, total_weights: 262144, loss: 10.270, lm_loss: 10.270, perplexity: 28865.557 I0423 11:25:54.812238 135142985008960 metric_logger.py:196] completed step: 4, seconds: 0.579, TFLOP/s/device: 93.913, Tokens/s/device: 14155.598, total_weights: 262144, loss: 10.021, lm_loss: 10.021, perplexity: 22497.775 I0423 11:25:54.819995 135142985008960 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.026, Tokens/s/device: 14172.643, total_weights: 262144, loss: 9.819, lm_loss: 9.819, perplexity: 18383.324 I0423 11:25:54.861365 2577 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 11:25:56.885528 135018028586752 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/0/items/array_metadatas/process_5 I0423 11:26:29.643735 135014805120768 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 41.186 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.35274076461792 s) (per-host) I0423 11:26:29.643858 135014805120768 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.070354s. I0423 11:26:37.496614 135014805120768 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 45.923094s. I0423 11:26:37.496939 135014280849152 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 11:26:37.497058 135014280849152 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 11:26:37.497110 135014280849152 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 11:26:38.223392 135014280849152 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 11:26:40.025821 135142985008960 metric_logger.py:196] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.935, Tokens/s/device: 7074.565, total_weights: 262144, loss: 9.666, lm_loss: 9.666, perplexity: 15775.117 I0423 11:26:40.604675 135142985008960 metric_logger.py:196] completed step: 7, seconds: 44.629, TFLOP/s/device: 1.218, Tokens/s/device: 183.559, total_weights: 262144, loss: 9.560, lm_loss: 9.560, perplexity: 14192.143 I0423 11:26:41.183090 135142985008960 metric_logger.py:196] completed step: 8, seconds: 0.584, TFLOP/s/device: 93.043, Tokens/s/device: 14024.492, total_weights: 262144, loss: 9.495, lm_loss: 9.495, perplexity: 13292.714 I0423 11:26:41.762217 135142985008960 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 11:26:41.762880 135142985008960 checkpointing.py:798] Waited 0.0006756782531738281 seconds for step 9 to finish before starting checkpointing. I0423 11:26:41.765123 135142985008960 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 11:26:41.767018 135142985008960 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9 I0423 11:26:41.768751 135142985008960 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/9. I0423 11:26:42.492357 135142985008960 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0423 11:26:42.492448 135142985008960 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 I0423 11:26:42.528905 135142985008960 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038300s I0423 11:26:42.529070 135142985008960 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 34.041 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045314788818359375 s) (per-host) I0423 11:26:42.529116 135142985008960 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045371s (batch_requests_ready=0.002694s, total_serialization_initiated=0.042613s, others=0.000063s) I0423 11:26:42.529216 135142985008960 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049630s (all_items=0.000032s, per_item={'items': '0.00003195'}, temp_paths=0.049598) I0423 11:26:42.529987 135142985008960 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/9. I0423 11:26:42.530246 135014280849152 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 11:46:42.530226 I0423 11:26:42.532742 135142985008960 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 11:26:42.532991 135014301820672 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 11:26:42.533152 135142985008960 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776943601.7650898, 'wait_for_prev_duration_secs': 7.605552673339844e-05, 'time_between_consecutive_saves_sec': 3.5416030883789062, 'checkpointer_blocking_start_time': 1776943601.7671013, 'checkpointer_blocking_duration_secs': 0.7632348537445068, 'get_old_steps_start_time': 1776943602.5303605, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776943601.7631192, 'checkpoint_manager_blocking_duration_secs': 0.7699990272521973} I0423 11:26:42.533260 135142985008960 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 11:26:42.533305 135142985008960 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 11:26:47.635644 135014310213376 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806/nnx_xpk_feat_nnx_trainstate_and_training_loop_20260423_093806_06_grad_accum/checkpoints/9/items/array_metadatas/process_5 I0423 11:27:25.249043 135014280849152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 36.936 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.76524806022644 s) (per-host) I0423 11:27:25.249169 135014280849152 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.718854s. I0423 11:27:32.616058 135014280849152 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 50.085728s. I0423 11:27:32.616338 135014301820672 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 11:27:32.616454 135014301820672 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 11:27:32.616501 135014301820672 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 11:27:32.619178 135014301820672 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 11:27:32.619360 135142985008960 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0423 11:27:32.619526 135142985008960 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 11:27:32.620492 135142985008960 metric_logger.py:196] completed step: 9, seconds: 0.579, TFLOP/s/device: 93.845, Tokens/s/device: 14145.356, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12793.173 Per train step: Total TFLOPs: 54.35 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 11:27:44 UTC 2026 EXIT_CODE=0