[2026-02-24 10:05:43,624] [DEBUG] [axolotl.utils.config.resolve_dtype:66] [PID:1] bf16 support detected, enabling for this configuration. [2026-02-24 10:05:43,882] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:1] baseline 0.000GB () [2026-02-24 10:05:43,883] [INFO] [axolotl.cli.config.load_cfg:259] [PID:1] config: { "activation_offloading": false, "adapter": "qlora", "axolotl_config_path": "/app/results/7a775f96-133a-44da-975d-d4875774c579/axolotl_config.yml", "base_model": "unsloth/Llama-3.2-1B-Instruct", "base_model_config": "unsloth/Llama-3.2-1B-Instruct", "batch_size": 4, "bf16": true, "capabilities": { "bf16": true, "compute_capability": "sm_86", "fp8": false, "n_gpu": 1, "n_node": 1 }, "context_parallel_size": 1, "dataloader_num_workers": 1, "dataloader_pin_memory": true, "dataloader_prefetch_factor": 256, "dataset_num_proc": 16, "datasets": [ { "message_property_mappings": { "content": "content", "role": "role" }, "path": "yahma/alpaca-cleaned", "trust_remote_code": false, "type": "alpaca" } ], "ddp": false, "device": "cuda:0", "device_map": "auto", "dion_rank_fraction": 1.0, "dion_rank_multiple_of": 1, "eaft_alpha": 1.0, "eaft_k": 20, "env_capabilities": { "torch_version": "2.9.1" }, "eval_batch_size": 1, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_table_size": 0, "experimental_skip_move_to_device": true, "flash_attention": false, "fp16": false, "gradient_accumulation_steps": 4, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "hf_use_auth_token": true, "hub_model_id": "HarrySoteriou/smoke-test-output", "include_tkps": true, "is_falcon_derived_model": false, "is_llama_derived_model": true, "is_mistral_derived_model": false, "learning_rate": 0.0002, "lisa_layers_attribute": "model.layers", "load_best_model_at_end": false, "load_in_4bit": true, "load_in_8bit": false, "local_rank": 0, "logging_steps": 10, "lora_alpha": 32, "lora_dropout": 0.05, "lora_r": 16, "lora_target_linear": true, "loraplus_lr_embedding": 1e-06, "lr_scheduler": "cosine", "max_steps": 10, "mean_resizing_embeddings": false, "micro_batch_size": 1, "model_config_type": "llama", "num_epochs": 1.0, "optimizer": "adamw_bnb_8bit", "otel_metrics_host": "localhost", "otel_metrics_port": 8000, "output_dir": "/app/results/7a775f96-133a-44da-975d-d4875774c579", "pretrain_multipack_attn": true, "profiler_steps_start": 0, "qlora_sharded_model_loading": false, "ray_num_workers": 1, "resources_per_worker": { "GPU": 1 }, "sample_packing_bin_size": 200, "sample_packing_group_size": 100000, "save_only_model": false, "save_safetensors": true, "save_steps": 100, "sequence_len": 2048, "shuffle_before_merging_datasets": false, "shuffle_merged_datasets": true, "skip_prepare_dataset": false, "streaming_multipack_buffer_size": 10000, "strict": false, "tensor_parallel_size": 1, "tiled_mlp_use_original_mlp": true, "tokenizer_config": "unsloth/Llama-3.2-1B-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 }, "type_of_model": "AutoModelForCausalLM", "use_otel_metrics": false, "use_ray": false, "val_set_size": 0.0, "vllm": { "device": "auto", "dtype": "auto", "gpu_memory_utilization": 0.9, "host": "0.0.0.0", "port": 8000 }, "warmup_steps": 10, "weight_decay": 0.0, "world_size": 1 } [2026-02-24 10:05:46,094] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:285] [PID:1] EOS: 128009 / <|eot_id|> [2026-02-24 10:05:46,094] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:1] BOS: 128000 / <|begin_of_text|> [2026-02-24 10:05:46,094] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:1] PAD: 128004 / <|finetune_right_pad_id|> [2026-02-24 10:05:46,094] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:1] UNK: None / None [2026-02-24 10:05:46,096] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:481] [PID:1] Unable to find prepared dataset in last_run_prepared/8f37d688038ca7c67f271223b45b9dc1 [2026-02-24 10:05:46,096] [INFO] [axolotl.utils.data.sft._load_raw_datasets:320] [PID:1] Loading raw datasets... [2026-02-24 10:05:46,096] [WARNING] [axolotl.utils.data.sft._load_raw_datasets:322] [PID:1] Processing datasets during training can lead to VRAM instability. Please pre-process your dataset using `axolotl preprocess path/to/config.yml`. Fetching 0 files: 0it [00:00, ?it/s] Fetching 0 files: 0it [00:00, ?it/s] [2026-02-24 10:05:49,392] [INFO] [axolotl.utils.data.wrappers.get_dataset_wrapper:87] [PID:1] Loading dataset: yahma/alpaca-cleaned with base_type: alpaca and prompt_style: None [2026-02-24 10:05:49,833] [INFO] [axolotl.utils.data.utils.handle_long_seq_in_dataset:224] [PID:1] min_input_len: 33 [2026-02-24 10:05:49,834] [INFO] [axolotl.utils.data.utils.handle_long_seq_in_dataset:226] [PID:1] max_input_len: 1051 Dropping Long Sequences (>2048) (num_proc=16): 0%| | 0/51760 [00:002048) (num_proc=16): 2%|▏ | 1000/51760 [00:00<00:50, 1002.44 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 29%|██▉ | 15000/51760 [00:01<00:02, 18202.31 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 44%|████▍ | 23000/51760 [00:01<00:01, 23976.79 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 60%|█████▉ | 31000/51760 [00:01<00:00, 31009.89 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 71%|███████▏ | 37000/51760 [00:01<00:00, 35982.37 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 91%|█████████ | 47115/51760 [00:01<00:00, 48954.07 examples/s] Dropping Long Sequences (>2048) (num_proc=16): 100%|██████████| 51760/51760 [00:01<00:00, 27495.53 examples/s] Saving the dataset (0/16 shards): 0%| | 0/51760 [00:00 [2026-02-24 10:05:55,171] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:1] BOS: 128000 / <|begin_of_text|> [2026-02-24 10:05:55,171] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:1] PAD: 128004 / <|finetune_right_pad_id|> [2026-02-24 10:05:55,171] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:1] UNK: None / None [2026-02-24 10:05:55,171] [DEBUG] [axolotl.train.setup_model_and_tokenizer:82] [PID:1] Loading model [2026-02-24 10:05:55,436] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:1] Patched Trainer.evaluation_loop with nanmean loss calculation [2026-02-24 10:05:55,438] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:1] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation [2026-02-24 10:06:03,554] [INFO] [axolotl.loaders.model._prepare_model_for_quantization:859] [PID:1] converting PEFT model w/ prepare_model_for_kbit_training [2026-02-24 10:06:03,557] [INFO] [axolotl.loaders.model._configure_embedding_dtypes:351] [PID:1] Converting modules to torch.bfloat16 [2026-02-24 10:06:03,560] [DEBUG] [axolotl.loaders.model.log_gpu_memory_usage:127] [PID:1] Memory usage after model load 1.936GB (+1.936GB allocated, +2.162GB reserved) [2026-02-24 10:06:03,561] [INFO] [axolotl.loaders.adapter.load_lora:81] [PID:1] found linear modules: ['down_proj', 'gate_proj', 'k_proj', 'o_proj', 'q_proj', 'up_proj', 'v_proj'] trainable params: 11,272,192 || all params: 1,247,086,592 || trainable%: 0.9039 [2026-02-24 10:06:03,949] [DEBUG] [axolotl.loaders.model.log_gpu_memory_usage:127] [PID:1] after adapters 0.999GB (+0.999GB allocated, +2.203GB reserved) [2026-02-24 10:06:11,856] [INFO] [axolotl.train.save_initial_configs:402] [PID:1] Pre-saving adapter config to /app/results/7a775f96-133a-44da-975d-d4875774c579... [2026-02-24 10:06:11,857] [INFO] [axolotl.train.save_initial_configs:406] [PID:1] Pre-saving tokenizer to /app/results/7a775f96-133a-44da-975d-d4875774c579... [2026-02-24 10:06:12,115] [INFO] [axolotl.train.save_initial_configs:411] [PID:1] Pre-saving model config to /app/results/7a775f96-133a-44da-975d-d4875774c579... [2026-02-24 10:06:12,121] [INFO] [axolotl.train.execute_training:207] [PID:1] Starting trainer... 0%| | 0/10 [00:00