[2026-01-28 18:42:35,684] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:11475] baseline 0.000GB () [2026-01-28 18:42:35,684] [INFO] [axolotl.cli.config.load_cfg:259] [PID:11475] config: { "activation_offloading": false, "adapter": "lora", "axolotl_config_path": "Guten.yaml", "base_model": "allenai/Olmo-3.1-32B-Instruct", "base_model_config": "allenai/Olmo-3.1-32B-Instruct", "batch_size": 8, "bf16": true, "capabilities": { "bf16": true, "compute_capability": "sm_80", "fp8": false, "n_gpu": 1, "n_node": 1 }, "chat_template": "chatml", "context_parallel_size": 1, "dataloader_num_workers": 1, "dataloader_pin_memory": true, "dataloader_prefetch_factor": 256, "dataset_num_proc": 12, "datasets": [ { "chat_template": "tokenizer_default", "message_property_mappings": { "content": "content", "role": "role" }, "path": "./gutenbeg-2-sft-train.jsonl", "train_on_eos": "all", "trust_remote_code": false, "type": "chat_template" }, { "chat_template": "tokenizer_default", "message_property_mappings": { "content": "content", "role": "role" }, "path": "./condor-sft-train.jsonl", "train_on_eos": "all", "trust_remote_code": false, "type": "chat_template" } ], "ddp": false, "device": "cuda:0", "dion_rank_fraction": 1.0, "dion_rank_multiple_of": 1, "env_capabilities": { "torch_version": "2.8.0" }, "eot_tokens": [ "<|im_end|>" ], "eval_batch_size": 2, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_sample_packing": true, "eval_table_size": 0, "experimental_skip_move_to_device": true, "flash_attention": true, "fp16": false, "gradient_accumulation_steps": 4, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "include_tkps": true, "learning_rate": 2.5e-05, "liger_fused_linear_cross_entropy": true, "liger_glu_activation": true, "liger_layer_norm": true, "liger_rms_norm": true, "liger_rope": true, "lisa_layers_attribute": "model.layers", "load_best_model_at_end": false, "load_in_4bit": false, "load_in_8bit": true, "local_rank": 0, "logging_steps": 1, "lora_alpha": 128, "lora_dropout": 0.0, "lora_mlp_kernel": false, "lora_o_kernel": false, "lora_qkv_kernel": false, "lora_r": 64, "lora_target_linear": true, "loraplus_lr_embedding": 1e-06, "loraplus_lr_ratio": 16.0, "lr_scheduler": "constant_with_warmup", "max_grad_norm": 1.0, "mean_resizing_embeddings": false, "micro_batch_size": 2, "model_config_type": "olmo3", "num_epochs": 2.0, "optimizer": "paged_adamw_8bit", "otel_metrics_host": "localhost", "otel_metrics_port": 8000, "output_dir": "./Olmo-Guten-SFT", "pad_to_sequence_len": true, "plugins": [ "axolotl.integrations.liger.LigerPlugin" ], "pretrain_multipack_attn": true, "profiler_steps_start": 0, "qlora_sharded_model_loading": false, "ray_num_workers": 1, "resources_per_worker": { "GPU": 1 }, "sample_packing": true, "sample_packing_bin_size": 200, "sample_packing_group_size": 100000, "save_only_model": false, "save_safetensors": true, "seed": 42, "sequence_len": 6144, "shuffle_before_merging_datasets": false, "shuffle_merged_datasets": true, "skip_prepare_dataset": false, "streaming_multipack_buffer_size": 10000, "strict": false, "tensor_parallel_size": 1, "tf32": true, "tiled_mlp_use_original_mlp": true, "tokenizer_config": "allenai/Olmo-3.1-32B-Instruct", "tokenizer_save_jinja_files": true, "torch_dtype": "torch.bfloat16", "train_on_inputs": false, "trl": { "log_completions": false, "mask_truncated_completions": false, "ref_model_mixup_alpha": 0.9, "ref_model_sync_steps": 64, "scale_rewards": true, "sync_ref_model": false, "use_vllm": false, "vllm_server_host": "0.0.0.0", "vllm_server_port": 8000 }, "use_otel_metrics": false, "use_ray": false, "use_tensorboard": true, "val_set_size": 0.0, "vllm": { "device": "auto", "dtype": "auto", "gpu_memory_utilization": 0.9, "host": "0.0.0.0", "port": 8000 }, "warmup_ratio": 0.05, "weight_decay": 0.0, "world_size": 1 } [2026-01-28 18:42:35,686] [WARNING] [axolotl.cli.checks.check_user_token:46] [PID:11475] Error verifying HuggingFace token. Remember to log in using `huggingface-cli login` and get your access token from https://huggingface.co/settings/tokens if you want to use gated models or datasets. [2026-01-28 18:42:36,259] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:285] [PID:11475] EOS: 100257 / <|endoftext|> [2026-01-28 18:42:36,259] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:11475] BOS: 100257 / <|endoftext|> [2026-01-28 18:42:36,259] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:11475] PAD: 100277 / <|pad|> [2026-01-28 18:42:36,259] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:11475] UNK: 100257 / <|endoftext|> [2026-01-28 18:42:36,260] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:481] [PID:11475] Unable to find prepared dataset in last_run_prepared/ef6f6bab0786bc209a0a8382fdc0cfca [2026-01-28 18:42:36,261] [INFO] [axolotl.utils.data.sft._load_raw_datasets:320] [PID:11475] Loading raw datasets... [2026-01-28 18:42:36,261] [WARNING] [axolotl.utils.data.sft._load_raw_datasets:322] [PID:11475] Processing datasets during training can lead to VRAM instability. Please pre-process your dataset using `axolotl preprocess path/to/config.yml`. [2026-01-28 18:42:36,382] [INFO] [axolotl.utils.data.wrappers.get_dataset_wrapper:87] [PID:11475] Loading dataset: ./gutenbeg-2-sft-train.jsonl with base_type: chat_template and prompt_style: None [2026-01-28 18:42:36,385] [INFO] [axolotl.prompt_strategies.chat_template.__call__:996] [PID:11475] Using chat template: --- {% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + ' ' + message['content'] + '<|im_end|>' + ' '}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant ' }}{% endif %} --- [2026-01-28 18:42:36,521] [INFO] [axolotl.utils.data.wrappers.get_dataset_wrapper:87] [PID:11475] Loading dataset: ./condor-sft-train.jsonl with base_type: chat_template and prompt_style: None [2026-01-28 18:42:36,521] [INFO] [axolotl.prompt_strategies.chat_template.__call__:996] [PID:11475] Using chat template: --- {% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + ' ' + message['content'] + '<|im_end|>' + ' '}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant ' }}{% endif %} --- [2026-01-28 18:42:36,606] [INFO] [axolotl.utils.data.shared.merge_datasets:556] [PID:11475] Merging datasets... [2026-01-28 18:42:36,612] [DEBUG] [axolotl.utils.data.shared.merge_datasets:560] [PID:11475] Shuffling merged datasets... [2026-01-28 18:42:36,618] [INFO] [axolotl.utils.data.utils.handle_long_seq_in_dataset:224] [PID:11475] min_input_len: 45 [2026-01-28 18:42:36,618] [INFO] [axolotl.utils.data.utils.handle_long_seq_in_dataset:226] [PID:11475] max_input_len: 14764 [2026-01-28 18:42:36,653] [WARNING] [axolotl.utils.data.utils.handle_long_seq_in_dataset:270] [PID:11475] Dropped 30 samples from dataset Saving the dataset (0/2 shards): 0%| | 0/513 [00:00 [2026-01-28 18:42:39,970] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:11475] BOS: 100257 / <|endoftext|> [2026-01-28 18:42:39,970] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:11475] PAD: 100277 / <|pad|> [2026-01-28 18:42:39,970] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:11475] UNK: 100257 / <|endoftext|> [2026-01-28 18:42:39,970] [DEBUG] [axolotl.train.setup_model_and_tokenizer:82] [PID:11475] Loading model [2026-01-28 18:42:40,046] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:11475] Patched Trainer.evaluation_loop with nanmean loss calculation [2026-01-28 18:42:40,048] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:11475] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation [2026-01-28 18:42:40,049] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:345] [PID:11475] Applying multipack dataloader patch for sample packing... [2026-01-28 18:42:40,170] [INFO] [axolotl.integrations.liger.plugin.pre_model_load:98] [PID:11475] Applying LIGER to olmo3 with kwargs: {'rope': True, 'cross_entropy': None, 'fused_linear_cross_entropy': True, 'rms_norm': True, 'swiglu': True} model.safetensors.index.json: 0.00B [00:00, ?B/s] model.safetensors.index.json: 59.1kB [00:00, 98.1MB/s] model-00001-of-00014.safetensors: 0%| | 0.00/4.99G [00:00