MaxView

XPK Start: Thu Apr 23 14:32:22 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 14:32:47.021956: 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 14:32:47.232500 138022031345472 max_utils.py:273] Attempting to initialize the jax distributed system...
I0423 14:32:56.272145 138022031345472 distributed.py:149] Starting JAX distributed service on [::]:8482
I0423 14:32:56.275042 138022031345472 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-ccfqt-slice-job-0-0.mt-06-grad-accum-ccfqt:8482
I0423 14:32:57.089503 138022031345472 max_utils.py:284] Jax distributed system initialized!
I0423 14:33:03.325280 138022031345472 max_utils.py:800] System Information: Jax Version: 0.9.2
I0423 14:33:03.325383 138022031345472 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0423 14:33:03.325424 138022031345472 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 14:33:03.325459 138022031345472 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 14:33:04.018037 138022031345472 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 14:33:04.133449 138022031345472 checkpointing.py:688] Setting up checkpoint logger...
I0423 14:33:04.133567 138022031345472 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0423 14:33:04.133612 138022031345472 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0423 14:33:04.133831 138022031345472 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 0x7d8733314aa0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 14:33:07.065498 138022031345472 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0423 14:33:07.065755 138022031345472 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 0x7d72d06eb260>}, handler_registry=None
I0423 14:33:07.066000 138022031345472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d72d06eb260>` 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 14:33:07.066048 138022031345472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72f442a690>` 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 14:33:07.066084 138022031345472 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 0x7d72d06eb260>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d72d06eb260>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72f442a690>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72f442a690>}).
I0423 14:33:07.066405 138022031345472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0423 14:33:07.066476 138022031345472 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d72f441b7e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0423 14:33:07.982497 138022031345472 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints
I0423 14:33:08.426962 138022031345472 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_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d72f444a240>
I0423 14:33:08.427149 138022031345472 checkpointing.py:302] Checkpoint manager created!
I0423 14:33:10.194217 138022031345472 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 14:33:10.194335 138022031345472 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0423 14:33:11.590281 138022031345472 maxtext_utils.py:1889]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0423 14:33:11.590400 138022031345472 maxtext_utils.py:1889]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0423 14:33:11.590452 138022031345472 maxtext_utils.py:1889]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0423 14:33:11.590496 138022031345472 maxtext_utils.py:1889]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0423 14:33:11.590534 138022031345472 maxtext_utils.py:1889]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0423 14:33:11.590569 138022031345472 maxtext_utils.py:1889]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0423 14:33:11.590605 138022031345472 maxtext_utils.py:1889]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:33:11.590640 138022031345472 maxtext_utils.py:1889]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0423 14:33:11.590689 138022031345472 maxtext_utils.py:1889]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:33:11.590726 138022031345472 maxtext_utils.py:1889]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0423 14:33:11.590759 138022031345472 maxtext_utils.py:1889]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0423 14:33:11.590793 138022031345472 maxtext_utils.py:1889]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0423 14:33:11.984199 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0423 14:33:11.984292 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0423 14:33:11.999339 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0423 14:33:11.999401 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0423 14:33:12.029238 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0423 14:33:12.029310 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0423 14:33:12.044183 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0423 14:33:12.044238 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0423 14:33:12.073900 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0423 14:33:12.073968 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0423 14:33:12.088999 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0423 14:33:12.089053 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0423 14:33:12.133452 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0423 14:33:12.133529 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0423 14:33:12.148917 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0423 14:33:12.148979 138022031345472 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0423 14:33:12.376171 138022031345472 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 14:33:12.376333 138022031345472 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:33:12.383133 138022031345472 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 14:33:12.383236 138022031345472 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:33:12.400260 138022031345472 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 14:33:12.400393 138022031345472 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:33:12.416842 138022031345472 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 14:33:12.416969 138022031345472 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0423 14:33:12.447861 138022031345472 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 14:33:12.448006 138022031345472 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:33:12.464469 138022031345472 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 14:33:12.464601 138022031345472 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:33:12.480919 138022031345472 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 14:33:12.481054 138022031345472 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:33:12.509224 138022031345472 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 14:33:12.509359 138022031345472 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0423 14:33:12.536847 138022031345472 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 14:33:12.537009 138022031345472 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0423 14:33:24.229689 138022031345472 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 14:33:24.232064 138022031345472 metric_logger.py:301] number parameters: 1.104 billion
I0423 14:33:26.372499 138022031345472 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0423 14:33:36.946247 138022031345472 checkpointing.py:798] Waited 10.573730707168579 seconds for step 0 to finish before starting checkpointing.
I0423 14:33:36.948595 138022031345472 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 14:33:36.950453 138022031345472 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0
I0423 14:33:36.951847 138022031345472 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/0.
I0423 14:33:37.332148 138022031345472 signaling_client.py:364] Using JaxDistributedSignalingClient
I0423 14:33:37.333174 138022031345472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0423 14:33:37.333236 138022031345472 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 14:33:37.608798 138022031345472 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.277463s
I0423 14:33:37.608983 138022031345472 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.372 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28714919090270996 s) (per-host)
I0423 14:33:37.609040 138022031345472 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287215s (batch_requests_ready=0.003400s, total_serialization_initiated=0.283741s, others=0.000074s)
I0423 14:33:37.609140 138022031345472 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291318s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.291301)
I0423 14:33:37.609861 138022031345472 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.66 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/0.
I0423 14:33:37.610126 137897374119680 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 14:53:37.610104
I0423 14:33:37.616963 138022031345472 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 14:33:37.617221 137897320396544 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0423 14:33:37.617338 138022031345472 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776954816.948576, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776954816.9504912, 'checkpointer_blocking_duration_secs': 0.6597442626953125, 'get_old_steps_start_time': 1776954817.6102579, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776954816.9466364, 'checkpoint_manager_blocking_duration_secs': 0.6706650257110596}
I0423 14:33:37.617439 138022031345472 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0423 14:33:37.617500 138022031345472 max_utils.py:750] 
Memstats: After params initialized:
I0423 14:33:37.617547 138022031345472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0))
I0423 14:33:37.617578 138022031345472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0))
I0423 14:33:37.617605 138022031345472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0))
I0423 14:33:37.617629 138022031345472 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0))
I0423 14:33:37.930166 138022031345472 metric_logger.py:196] completed step: 0, seconds: 2.139, TFLOP/s/device: 25.405, Tokens/s/device: 3829.249, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598
I0423 14:33:38.529093 138022031345472 metric_logger.py:196] completed step: 1, seconds: 11.556, TFLOP/s/device: 4.703, Tokens/s/device: 708.916, total_weights: 262144, loss: 10.876, lm_loss: 10.876, perplexity: 52883.598
I0423 14:33:39.108014 138022031345472 metric_logger.py:196] completed step: 2, seconds: 0.022, TFLOP/s/device: 2462.217, Tokens/s/device: 371132.152, total_weights: 262144, loss: 10.561, lm_loss: 10.561, perplexity: 38601.980
I0423 14:33:39.686605 138022031345472 metric_logger.py:196] completed step: 3, seconds: 0.585, TFLOP/s/device: 92.947, Tokens/s/device: 14009.909, total_weights: 262144, loss: 10.270, lm_loss: 10.270, perplexity: 28865.557
I0423 14:33:40.464035    2512 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0423 14:33:40.843995 138022031345472 metric_logger.py:196] completed step: 4, seconds: 0.579, TFLOP/s/device: 93.809, Tokens/s/device: 14139.936, total_weights: 262144, loss: 10.021, lm_loss: 10.021, perplexity: 22497.775
I0423 14:33:40.851823 138022031345472 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.012, Tokens/s/device: 14170.534, total_weights: 262144, loss: 9.819, lm_loss: 9.819, perplexity: 18383.324
I0423 14:33:42.808550 137897328789248 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_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/0/items/array_metadatas/process_5
I0423 14:34:14.671261 137897374119680 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 42.292 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 37.34940552711487 s) (per-host)
I0423 14:34:14.671366 137897374119680 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.061170s.
I0423 14:34:23.833613 137897374119680 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 46.223401s.
I0423 14:34:23.833840 137897320396544 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 14:34:23.833900 137897320396544 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0423 14:34:23.833965 137897320396544 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0423 14:34:24.752481 137897320396544 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0423 14:34:26.457250 138022031345472 metric_logger.py:196] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.926, Tokens/s/device: 7073.191, total_weights: 262144, loss: 9.666, lm_loss: 9.666, perplexity: 15775.117
I0423 14:34:27.035821 138022031345472 metric_logger.py:196] completed step: 7, seconds: 45.029, TFLOP/s/device: 1.207, Tokens/s/device: 181.929, total_weights: 262144, loss: 9.560, lm_loss: 9.560, perplexity: 14192.143
I0423 14:34:27.614204 138022031345472 metric_logger.py:196] completed step: 8, seconds: 0.584, TFLOP/s/device: 93.009, Tokens/s/device: 14019.307, total_weights: 262144, loss: 9.495, lm_loss: 9.495, perplexity: 13292.714
I0423 14:34:28.193452 138022031345472 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0423 14:34:28.194134 138022031345472 checkpointing.py:798] Waited 0.0006928443908691406 seconds for step 9 to finish before starting checkpointing.
I0423 14:34:28.196238 138022031345472 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 14:34:28.197885 138022031345472 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9
I0423 14:34:28.199375 138022031345472 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/9.
I0423 14:34:28.497408 138022031345472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0423 14:34:28.497503 138022031345472 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 14:34:28.535361 138022031345472 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.039604s
I0423 14:34:28.535537 138022031345472 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 32.842 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04696965217590332 s) (per-host)
I0423 14:34:28.535604 138022031345472 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.047048s (batch_requests_ready=0.002779s, total_serialization_initiated=0.044182s, others=0.000087s)
I0423 14:34:28.535738 138022031345472 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.051144s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=0.051130)
I0423 14:34:28.536485 138022031345472 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.34 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/9.
I0423 14:34:28.536769 137897320396544 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-23 14:54:28.536737
I0423 14:34:28.538675 138022031345472 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 14:34:28.538944 137896828856064 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0423 14:34:28.539082 138022031345472 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776954868.196203, 'wait_for_prev_duration_secs': 7.748603820800781e-05, 'time_between_consecutive_saves_sec': 3.4436278343200684, 'checkpointer_blocking_start_time': 1776954868.1979227, 'checkpointer_blocking_duration_secs': 0.3389873504638672, 'get_old_steps_start_time': 1776954868.5369325, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776954868.1943908, 'checkpoint_manager_blocking_duration_secs': 0.34465885162353516}
I0423 14:34:28.539239 138022031345472 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0423 14:34:28.539284 138022031345472 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 14:34:33.664586 137896837248768 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_post_train_fixes_20260423_124541/nnx_xpk_feat_nnx_post_train_fixes_20260423_124541_06_grad_accum/checkpoints/9/items/array_metadatas/process_5
I0423 14:35:10.309117 137897320396544 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.771 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.820510149002075 s) (per-host)
I0423 14:35:10.309246 137897320396544 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.772368s.
I0423 14:35:19.373609 137897320396544 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 50.836717s.
I0423 14:35:19.373913 137896828856064 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 14:35:19.374035 137896828856064 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0423 14:35:19.374083 137896828856064 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0423 14:35:19.375512 137896828856064 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0423 14:35:19.375689 138022031345472 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 14:35:19.375860 138022031345472 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 14:35:19.376828 138022031345472 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 93.962, Tokens/s/device: 14163.037, 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 14:35:32 UTC 2026
EXIT_CODE=0