XPK Start: Fri Apr 24 12:20:13 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 12:20:38.428193: 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 12:20:38.636295 139277192738624 max_utils.py:273] Attempting to initialize the jax distributed system...
I0424 12:20:47.676945 139277192738624 distributed.py:149] Starting JAX distributed service on [::]:8482
I0424 12:20:47.679400 139277192738624 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-j2zi4-slice-job-0-0.mt-06-grad-accum-j2zi4:8482
I0424 12:20:48.869081 139277192738624 max_utils.py:284] Jax distributed system initialized!
I0424 12:20:55.035182 139277192738624 max_utils.py:800] System Information: Jax Version: 0.9.2
I0424 12:20:55.035302 139277192738624 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0424 12:20:55.035342 139277192738624 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 12:20:55.035378 139277192738624 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0424 12:20:55.734362 139277192738624 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 12:20:55.734953 139277192738624 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 12:20:55.735133 139277192738624 checkpointing.py:688] Setting up checkpoint logger...
I0424 12:20:55.735184 139277192738624 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0424 12:20:55.735232 139277192738624 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0424 12:20:55.735574 139277192738624 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 0x7eab641141d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0424 12:20:58.682293 139277192738624 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0424 12:20:58.682533 139277192738624 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 0x7eabb43cf3e0>}, handler_registry=None
I0424 12:20:58.682797 139277192738624 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eabb43cf3e0>` 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 12:20:58.682848 139277192738624 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e979c334aa0>` 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 12:20:58.682883 139277192738624 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 0x7eabb43cf3e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7eabb43cf3e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e979c334aa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e979c334aa0>}).
I0424 12:20:58.683200 139277192738624 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0424 12:20:58.683270 139277192738624 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e970447dd00> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0424 12:20:59.406863 139277192738624 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_06_grad_accum/checkpoints
I0424 12:20:59.478781 139277192738624 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/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e979c332d50>
I0424 12:20:59.478949 139277192738624 checkpointing.py:302] Checkpoint manager created!
I0424 12:21:01.199070 139277192738624 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0424 12:21:01.199177 139277192738624 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:21:01.584037 139277192738624 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 12:21:01.584134 139277192738624 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:21:01.600565 139277192738624 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 12:21:01.600619 139277192738624 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 12:21:01.624489 139277192738624 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:21:01.624562 139277192738624 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:21:01.641150 139277192738624 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:21:01.641215 139277192738624 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:21:01.657695 139277192738624 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 12:21:01.657756 139277192738624 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:21:01.682557 139277192738624 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0424 12:21:01.682625 139277192738624 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 12:21:01.703517 139277192738624 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0424 12:21:01.703583 139277192738624 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0424 12:21:01.922935 139277192738624 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0424 12:21:01.923046 139277192738624 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0424 12:21:03.360160 139277192738624 maxtext_utils.py:1880] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0424 12:21:03.360290 139277192738624 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 12:21:03.360345 139277192738624 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 12:21:03.360403 139277192738624 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 12:21:03.360457 139277192738624 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 12:21:03.360499 139277192738624 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 12:21:03.360553 139277192738624 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 12:21:03.360606 139277192738624 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 12:21:03.360663 139277192738624 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 12:21:03.360706 139277192738624 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 12:21:03.360756 139277192738624 maxtext_utils.py:1880] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0424 12:21:03.360808 139277192738624 maxtext_utils.py:1880] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0424 12:21:03.384379 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0424 12:21:03.384449 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0424 12:21:03.399391 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0424 12:21:03.399451 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0424 12:21:03.429245 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0424 12:21:03.429309 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0424 12:21:03.444370 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0424 12:21:03.444424 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0424 12:21:03.474076 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0424 12:21:03.474143 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0424 12:21:03.488962 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0424 12:21:03.489019 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0424 12:21:03.533790 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0424 12:21:03.533858 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0424 12:21:03.548772 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0424 12:21:03.548830 139277192738624 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0424 12:21:04.225306 139277192738624 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 12:21:04.225401 139277192738624 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 12:21:04.241128 139277192738624 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0424 12:21:04.241191 139277192738624 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0424 12:21:15.756363 139277192738624 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.
I0424 12:21:15.757240 139277192738624 metric_logger.py:301] number parameters: 1.104 billion
I0424 12:21:17.598383 139277192738624 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0424 12:21:28.321645 139277192738624 checkpointing.py:798] Waited 10.723241329193115 seconds for step 0 to finish before starting checkpointing.
I0424 12:21:28.324143 139277192738624 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 12:21:28.325873 139277192738624 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0
I0424 12:21:28.327576 139277192738624 event_tracking.py:70] [process=5] [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_06_grad_accum/checkpoints/0.
I0424 12:21:28.662591 139277192738624 signaling_client.py:364] Using JaxDistributedSignalingClient
I0424 12:21:28.663506 139277192738624 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 12:21:28.663562 139277192738624 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 12:21:28.936103 139277192738624 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273586s
I0424 12:21:28.936288 139277192738624 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.526 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27916455268859863 s) (per-host)
I0424 12:21:28.936362 139277192738624 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.279249s (batch_requests_ready=0.002206s, total_serialization_initiated=0.276949s, others=0.000095s)
I0424 12:21:28.936480 139277192738624 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.283501s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.283484)
I0424 12:21:28.937323 139277192738624 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.61 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_06_grad_accum/checkpoints/0.
I0424 12:21:28.937668 139151249237760 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 12:41:28.937614
I0424 12:21:28.958594 139277192738624 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 12:21:28.958892 139151840114432 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0424 12:21:28.959052 139277192738624 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_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777033288.3241246, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777033288.3259106, 'checkpointer_blocking_duration_secs': 0.6119189262390137, 'get_old_steps_start_time': 1777033288.9378583, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1777033288.3221703, 'checkpoint_manager_blocking_duration_secs': 0.636843204498291}
I0424 12:21:28.959158 139277192738624 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0424 12:21:28.959208 139277192738624 max_utils.py:750]
Memstats: After params initialized:
I0424 12:21:28.959259 139277192738624 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0))
I0424 12:21:28.959292 139277192738624 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0))
I0424 12:21:28.959318 139277192738624 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0))
I0424 12:21:28.959341 139277192738624 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0))
I0424 12:21:29.277857 139277192738624 metric_logger.py:196] completed step: 0, seconds: 1.841, TFLOP/s/device: 29.521, Tokens/s/device: 4449.673, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0424 12:21:29.883444 139277192738624 metric_logger.py:196] completed step: 1, seconds: 11.678, TFLOP/s/device: 4.654, Tokens/s/device: 701.490, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0424 12:21:30.461685 139277192738624 metric_logger.py:196] completed step: 2, seconds: 0.027, TFLOP/s/device: 1976.669, Tokens/s/device: 297945.081, total_weights: 262144, loss: 10.563, lm_loss: 10.563, perplexity: 38662.707
I0424 12:21:31.039618 139277192738624 metric_logger.py:196] completed step: 3, seconds: 0.584, TFLOP/s/device: 93.074, Tokens/s/device: 14029.055, total_weights: 262144, loss: 10.272, lm_loss: 10.272, perplexity: 28909.668
I0424 12:21:31.905678 2531 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0424 12:21:32.196338 139277192738624 metric_logger.py:196] completed step: 4, seconds: 0.578, TFLOP/s/device: 93.995, Tokens/s/device: 14167.936, total_weights: 262144, loss: 10.022, lm_loss: 10.022, perplexity: 22524.992
I0424 12:21:32.203255 139277192738624 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.021, Tokens/s/device: 14171.809, total_weights: 262144, loss: 9.820, lm_loss: 9.820, perplexity: 18401.865
I0424 12:21:33.936862 139151257630464 array_metadata_store.py:203] [process=5][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_06_grad_accum/checkpoints/0/items/array_metadatas/process_5
I0424 12:22:06.687691 139151249237760 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 41.535 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.03052473068237 s) (per-host)
I0424 12:22:06.687815 139151249237760 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.750029s.
I0424 12:22:15.168926 139151249237760 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 46.231123s.
I0424 12:22:15.169220 139151840114432 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 12:22:15.169346 139151840114432 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0424 12:22:15.169396 139151840114432 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0424 12:22:15.514994 139151840114432 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0424 12:22:18.401577 139277192738624 metric_logger.py:196] completed step: 6, seconds: 1.158, TFLOP/s/device: 46.931, Tokens/s/device: 7073.936, total_weights: 262144, loss: 9.667, lm_loss: 9.667, perplexity: 15787.604
I0424 12:22:18.979558 139277192738624 metric_logger.py:196] completed step: 7, seconds: 45.621, TFLOP/s/device: 1.191, Tokens/s/device: 179.567, total_weights: 262144, loss: 9.561, lm_loss: 9.561, perplexity: 14203.827
I0424 12:22:19.557614 139277192738624 metric_logger.py:196] completed step: 8, seconds: 0.583, TFLOP/s/device: 93.238, Tokens/s/device: 14053.772, total_weights: 262144, loss: 9.496, lm_loss: 9.496, perplexity: 13302.920
I0424 12:22:20.135369 139277192738624 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0424 12:22:20.136093 139277192738624 checkpointing.py:798] Waited 0.000743865966796875 seconds for step 9 to finish before starting checkpointing.
I0424 12:22:20.138256 139277192738624 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 12:22:20.139987 139277192738624 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9
I0424 12:22:20.141474 139277192738624 event_tracking.py:70] [process=5] [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_06_grad_accum/checkpoints/9.
I0424 12:22:20.731457 139277192738624 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0424 12:22:20.731551 139277192738624 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 12:22:20.768366 139277192738624 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037933s
I0424 12:22:20.768505 139277192738624 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 37.333 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04131913185119629 s) (per-host)
I0424 12:22:20.768552 139277192738624 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.041374s (batch_requests_ready=0.001656s, total_serialization_initiated=0.039657s, others=0.000062s)
I0424 12:22:20.768627 139277192738624 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045726s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.045711)
I0424 12:22:20.769270 139277192738624 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.63 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_06_grad_accum/checkpoints/9.
I0424 12:22:20.769608 139128536553216 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 12:42:20.769571
I0424 12:22:20.771596 139277192738624 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 12:22:20.771895 139151813179136 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0424 12:22:20.772028 139277192738624 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_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777033340.1382077, 'wait_for_prev_duration_secs': 9.5367431640625e-05, 'time_between_consecutive_saves_sec': 4.623120069503784, 'checkpointer_blocking_start_time': 1777033340.140027, 'checkpointer_blocking_duration_secs': 0.6297388076782227, 'get_old_steps_start_time': 1777033340.7697897, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1777033340.1363335, 'checkpoint_manager_blocking_duration_secs': 0.6356611251831055}
I0424 12:22:20.772195 139277192738624 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0424 12:22:20.772243 139277192738624 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 12:22:25.922852 139156646774528 array_metadata_store.py:203] [process=5][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_06_grad_accum/checkpoints/9/items/array_metadatas/process_5
I0424 12:23:02.931865 139128536553216 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.427 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.20464587211609 s) (per-host)
I0424 12:23:02.931988 139128536553216 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.162252s.
I0424 12:23:10.864218 139128536553216 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 50.094455s.
I0424 12:23:10.864512 139151813179136 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 12:23:10.864626 139151813179136 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0424 12:23:10.864695 139151813179136 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0424 12:23:10.866008 139151813179136 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0424 12:23:10.866173 139277192738624 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.
I0424 12:23:10.866328 139277192738624 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 12:23:10.867255 139277192738624 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 93.983, Tokens/s/device: 14166.197, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12802.546
Per train step:
Total TFLOPs: 54.35
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 24 12:23:23 UTC 2026
EXIT_CODE=0