[2026-03-15 11:07:02,974] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:38782] baseline 0.000GB () [2026-03-15 11:07:02,974] [INFO] [axolotl.cli.config.load_cfg:340] [PID:38782] config: { "activation_offloading": false, "adapter": "qlora", "axolotl_config_path": "/workspace/kgpt_config.yaml", "base_model": "Qwen/Qwen2.5-7B", "base_model_config": "Qwen/Qwen2.5-7B", "batch_size": 16, "bf16": true, "capabilities": { "bf16": true, "compute_capability": "sm_89", "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": 32, "datasets": [ { "ds_type": "json", "message_property_mappings": { "content": "content", "role": "role" }, "path": "/workspace/kgpt_combined.jsonl", "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": "2.8.0" }, "eval_batch_size": 1, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_sample_packing": true, "eval_steps": 100, "eval_table_size": 0, "experimental_skip_move_to_device": true, "flash_attention": true, "fp16": false, "generate_samples": false, "generation_do_sample": true, "generation_max_new_tokens": 50, "generation_prompt_ratio": 0.5, "generation_temperature": 0.7, "gradient_accumulation_steps": 16, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "include_tkps": true, "is_falcon_derived_model": false, "is_llama_derived_model": false, "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": 128, "lora_dropout": 0.05, "lora_r": 64, "lora_target_modules": [ "q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj" ], "loraplus_lr_embedding": 1e-06, "lr_scheduler": "cosine", "mean_resizing_embeddings": false, "micro_batch_size": 1, "model_config_type": "qwen2", "num_epochs": 3.0, "num_generation_samples": 3, "optimizer": "adamw_torch", "otel_metrics_host": "localhost", "otel_metrics_port": 8000, "output_dir": "/workspace/outputs/kgpt-v1", "pad_to_sequence_len": true, "pretrain_multipack_attn": true, "profiler_steps_start": 0, "qlora_sharded_model_loading": false, "quantize_moe_experts": 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, "save_steps": 200, "save_total_limit": 3, "sequence_len": 1024, "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": "Qwen/Qwen2.5-7B", "tokenizer_save_jinja_files": true, "tokenizer_type": "AutoTokenizer", "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 }, "trust_remote_code": true, "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": 100, "weight_decay": 0.01, "world_size": 1 } [2026-03-15 11:07:02,975] [WARNING] [axolotl.cli.checks.check_user_token:46] [PID:38782] Error verifying HuggingFace token. Remember to log in using `hf auth login` and get your access token from https://huggingface.co/settings/tokens if you want to use gated models or datasets. [2026-03-15 11:07:04,067] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:299] [PID:38782] EOS: 151643 / <|endoftext|> [2026-03-15 11:07:04,068] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:300] [PID:38782] BOS: None / None [2026-03-15 11:07:04,068] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:301] [PID:38782] PAD: 151643 / <|endoftext|> [2026-03-15 11:07:04,068] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:302] [PID:38782] UNK: None / None [2026-03-15 11:07:04,068] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:481] [PID:38782] Unable to find prepared dataset in last_run_prepared/9f3f78e8a3cb814f704820380de29656 [2026-03-15 11:07:04,068] [INFO] [axolotl.utils.data.sft._load_raw_datasets:320] [PID:38782] Loading raw datasets... [2026-03-15 11:07:04,068] [WARNING] [axolotl.utils.data.sft._load_raw_datasets:322] [PID:38782] Processing datasets during training can lead to VRAM instability. Please pre-process your dataset using `axolotl preprocess path/to/config.yml`. [2026-03-15 11:07:04,452] [INFO] [axolotl.utils.data.wrappers.get_dataset_wrapper:87] [PID:38782] Loading dataset: /workspace/kgpt_combined.jsonl with base_type: alpaca and prompt_style: None Tokenizing Prompts (num_proc=32): 0%| | 0/48697 [00:001024) (num_proc=32): 0%| | 0/48697 [00:001024) (num_proc=32): 3%|█▊ | 1522/48697 [00:00<00:14, 3168.25 examples/s] Dropping Invalid Sequences (1024) (num_proc=32): 100%|██████████████████████████████████████████████████████| 48697/48697 [00:00<00:00, 70617.69 examples/s] Drop Samples with Zero Trainable Tokens (num_proc=32): 0%| | 0/48697 [00:00 [2026-03-15 11:07:41,646] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:300] [PID:38782] BOS: None / None [2026-03-15 11:07:41,646] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:301] [PID:38782] PAD: 151643 / <|endoftext|> [2026-03-15 11:07:41,646] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:302] [PID:38782] UNK: None / None [2026-03-15 11:07:41,646] [DEBUG] [axolotl.train.setup_model_and_tokenizer:82] [PID:38782] Loading model [2026-03-15 11:07:41,791] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:91] [PID:38782] Patched Trainer.evaluation_loop with nanmean loss calculation [2026-03-15 11:07:41,792] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:142] [PID:38782] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation [2026-03-15 11:07:41,792] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:389] [PID:38782] Applying multipack dataloader patch for sample packing... Fetching 22 files: 0%| | 0/22 [00:00