XPK Start: Fri Apr 24 12:15:32 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. 2026-04-24 12:15:57.601756: 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) I0424 12:15:57.813595 140260153067328 max_utils.py:273] Attempting to initialize the jax distributed system... I0424 12:16:06.854602 140260153067328 distributed.py:149] Starting JAX distributed service on [::]:8482 I0424 12:16:06.856884 140260153067328 distributed.py:172] Connecting to JAX distributed service on mt-04-int8-68pmm-slice-job-0-0.mt-04-int8-68pmm:8482 I0424 12:16:08.197973 140260153067328 max_utils.py:284] Jax distributed system initialized! I0424 12:16:14.530256 140260153067328 max_utils.py:800] System Information: Jax Version: 0.9.2 I0424 12:16:14.530368 140260153067328 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0424 12:16:14.530408 140260153067328 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 I0424 12:16:14.530444 140260153067328 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 12:16:15.258490 140260153067328 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 12:16:15.259082 140260153067328 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 12:16:15.259399 140260153067328 checkpointing.py:688] Setting up checkpoint logger... I0424 12:16:15.259451 140260153067328 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0424 12:16:15.259496 140260153067328 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0424 12:16:15.259828 140260153067328 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 0x7f904e7d7470>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 12:16:18.162538 140260153067328 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 12:16:18.162783 140260153067328 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 0x7f7b8c58f3e0>}, handler_registry=None I0424 12:16:18.163020 140260153067328 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f7b8c58f3e0>` 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`. I0424 12:16:18.163070 140260153067328 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f7b8c590f80>` 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`. I0424 12:16:18.163118 140260153067328 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 0x7f7b8c58f3e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f7b8c58f3e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f7b8c590f80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f7b8c590f80>}). I0424 12:16:18.163441 140260153067328 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0424 12:16:18.163511 140260153067328 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f7b8c1d9e40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0424 12:16:19.324965 140260153067328 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints I0424 12:16:19.578793 140260153067328 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=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/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f7b8c590500> I0424 12:16:19.578971 140260153067328 checkpointing.py:302] Checkpoint manager created! I0424 12:16:20.731731 140260153067328 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 12:16:20.731846 140260153067328 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 12:16:21.119330 140260153067328 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 12:16:21.119425 140260153067328 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 12:16:21.137167 140260153067328 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 12:16:21.137228 140260153067328 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 12:16:21.204459 140260153067328 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 12:16:21.204550 140260153067328 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 12:16:21.222045 140260153067328 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 12:16:21.222121 140260153067328 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 12:16:21.239808 140260153067328 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 12:16:21.239877 140260153067328 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 12:16:21.266385 140260153067328 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0424 12:16:21.266458 140260153067328 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 12:16:21.337395 140260153067328 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 12:16:21.337493 140260153067328 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 12:16:21.866428 140260153067328 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 12:16:21.866540 140260153067328 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 12:16:24.401177 140260153067328 maxtext_utils.py:1880] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0424 12:16:24.401304 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 12:16:24.401366 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 12:16:24.401424 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0424 12:16:24.401481 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0424 12:16:24.401525 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0424 12:16:24.401581 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0424 12:16:24.401637 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0424 12:16:24.401677 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0424 12:16:24.401715 140260153067328 maxtext_utils.py:1880] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0424 12:16:24.401764 140260153067328 maxtext_utils.py:1880] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0424 12:16:24.401813 140260153067328 maxtext_utils.py:1880] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0424 12:16:25.808974 140260153067328 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 12:16:25.809073 140260153067328 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 12:16:25.824790 140260153067328 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 12:16:25.824850 140260153067328 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 12:16:39.862234 140260153067328 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0424 12:16:39.863032 140260153067328 metric_logger.py:301] number parameters: 1.104 billion I0424 12:16:54.687680 140260153067328 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 12:16:54.932708 140260153067328 checkpointing.py:798] Waited 0.24500346183776855 seconds for step 0 to finish before starting checkpointing. I0424 12:16:54.935154 140260153067328 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 12:16:54.937952 140260153067328 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0424 12:16:54.939355 140260153067328 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/0. I0424 12:16:55.295334 140260153067328 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 12:16:55.296267 140260153067328 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 12:16:55.296325 140260153067328 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 I0424 12:16:55.586269 140260153067328 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.291007s I0424 12:16:55.586450 140260153067328 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.203 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2964768409729004 s) (per-host) I0424 12:16:55.586508 140260153067328 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.296546s (batch_requests_ready=0.002225s, total_serialization_initiated=0.294243s, others=0.000078s) I0424 12:16:55.586610 140260153067328 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.300906s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.300889) I0424 12:16:55.587432 140260153067328 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.65 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/0. I0424 12:16:55.587775 140126001833728 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 12:36:55.587736 I0424 12:16:55.589900 140260153067328 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 12:16:55.590145 140123854411520 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 12:16:55.590307 140260153067328 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777033014.935136, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777033014.9379914, 'checkpointer_blocking_duration_secs': 0.6499476432800293, 'get_old_steps_start_time': 1777033015.5879645, 'get_old_steps_duration_secs': 3.1948089599609375e-05, 'checkpoint_manager_blocking_start_time': 1777033014.933234, 'checkpoint_manager_blocking_duration_secs': 0.6570351123809814} I0424 12:16:55.590408 140260153067328 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 12:16:55.590456 140260153067328 max_utils.py:750] Memstats: After params initialized: I0424 12:16:55.590502 140260153067328 max_utils.py:756] Using (GB) 0.44 / 31.25 (1.408000%) on TPU_24(process=6,(0,6,0,0)) I0424 12:16:55.590541 140260153067328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 12:16:55.590567 140260153067328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 12:16:55.590591 140260153067328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 12:16:55.915760 140260153067328 metric_logger.py:196] completed step: 0, seconds: 14.825, TFLOP/s/device: 0.917, Tokens/s/device: 138.149, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277 I0424 12:16:56.164795 140260153067328 metric_logger.py:196] completed step: 1, seconds: 1.227, TFLOP/s/device: 11.078, Tokens/s/device: 1669.727, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52796.277 I0424 12:16:56.577242 140260153067328 metric_logger.py:196] completed step: 2, seconds: 0.020, TFLOP/s/device: 675.003, Tokens/s/device: 101743.753, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28634.820 I0424 12:16:56.806620 140260153067328 metric_logger.py:196] completed step: 3, seconds: 0.413, TFLOP/s/device: 32.915, Tokens/s/device: 4961.324, total_weights: 65536, loss: 9.731, lm_loss: 9.731, perplexity: 16827.199 I0424 12:16:57.265661 140260153067328 metric_logger.py:196] completed step: 4, seconds: 0.235, TFLOP/s/device: 57.932, Tokens/s/device: 8732.135, total_weights: 65536, loss: 9.272, lm_loss: 9.272, perplexity: 10638.531 I0424 12:16:57.271808 140260153067328 metric_logger.py:196] completed step: 5, seconds: 0.229, TFLOP/s/device: 59.224, Tokens/s/device: 8926.820, total_weights: 65536, loss: 8.887, lm_loss: 8.887, perplexity: 7233.941 I0424 12:16:58.510128 2998 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 12:17:00.544763 140123912009472 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/0/items/array_metadatas/process_6 I0424 12:17:26.810663 140260153067328 metric_logger.py:196] completed step: 6, seconds: 0.460, TFLOP/s/device: 29.555, Tokens/s/device: 4454.847, total_weights: 65536, loss: 8.588, lm_loss: 8.588, perplexity: 5365.468 I0424 12:17:27.041744 140260153067328 metric_logger.py:196] completed step: 7, seconds: 29.311, TFLOP/s/device: 0.464, Tokens/s/device: 69.871, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4358.779 I0424 12:17:27.271078 140260153067328 metric_logger.py:196] completed step: 8, seconds: 0.233, TFLOP/s/device: 58.198, Tokens/s/device: 8772.193, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.453 I0424 12:17:27.499805 140260153067328 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 12:17:27.500499 140260153067328 checkpointing.py:798] Waited 0.0007135868072509766 seconds for step 9 to finish before starting checkpointing. I0424 12:17:27.502603 140260153067328 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 12:17:27.888607 140126001833728 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 48.456 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 32.59858703613281 s) (per-host) I0424 12:17:27.888730 140126001833728 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.300837s. I0424 12:17:38.361063 140126001833728 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 42.773152s. I0424 12:17:38.361398 140123854411520 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 12:17:38.361535 140123854411520 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 12:17:38.361587 140123854411520 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 12:17:38.364772 140123854411520 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 12:17:38.364932 140260153067328 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. W0424 12:17:38.365077 140260153067328 checkpoint_manager.py:1452] Waiting for previous save to complete took 10.862470 seconds. If this number is high, consider checkpointing less frequently. I0424 12:17:38.367013 140260153067328 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0424 12:17:38.369260 140260153067328 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/9. I0424 12:17:38.667396 140260153067328 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 12:17:38.667489 140260153067328 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 I0424 12:17:38.701505 140260153067328 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036031s I0424 12:17:38.701674 140260153067328 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 39.092 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.039460182189941406 s) (per-host) I0424 12:17:38.701725 140260153067328 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039519s (batch_requests_ready=0.001780s, total_serialization_initiated=0.037673s, others=0.000066s) I0424 12:17:38.701819 140260153067328 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043534s (all_items=0.000015s, per_item={'items': '0.00001526'}, temp_paths=0.043519) I0424 12:17:38.702489 140260153067328 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/9. I0424 12:17:38.702821 140123854411520 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 12:37:38.702783 I0424 12:17:38.709148 140260153067328 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 12:17:38.709436 140123912009472 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 12:17:38.709600 140260153067328 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777033047.502573, 'wait_for_prev_duration_secs': 10.862470388412476, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777033058.3670506, 'checkpointer_blocking_duration_secs': 0.33591437339782715, 'get_old_steps_start_time': 1777033058.7029872, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1777033047.5007346, 'checkpoint_manager_blocking_duration_secs': 11.208828926086426} I0424 12:17:38.709708 140260153067328 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 12:17:38.709754 140260153067328 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 12:17:44.186181 140105675147008 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_04_int8/checkpoints/9/items/array_metadatas/process_6 I0424 12:18:19.859130 140123854411520 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.343 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.19686484336853 s) (per-host) I0424 12:18:19.859239 140123854411520 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.156305s. I0424 12:18:29.908988 140123854411520 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.206038s. I0424 12:18:29.909305 140123912009472 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 12:18:29.909438 140123912009472 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 12:18:29.909486 140123912009472 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 12:18:29.910944 140123912009472 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 12:18:29.911132 140260153067328 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. I0424 12:18:29.911286 140260153067328 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 12:18:29.912347 140260153067328 metric_logger.py:196] completed step: 9, seconds: 0.231, TFLOP/s/device: 58.858, Tokens/s/device: 8871.677, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3552.333 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 24 12:18:42 UTC 2026 EXIT_CODE=0