[2025-10-18 19:02:01,879] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:42363] baseline 0.000GB () [2025-10-18 19:02:01,880] [INFO] [axolotl.cli.config.load_cfg:248] [PID:42363] config: { "activation_offloading": false, "adapter": "qlora", "axolotl_config_path": "mrCuddle-stream.yaml", "base_model": "google/gemma-2-2b-it", "base_model_config": "google/gemma-2-2b-it", "batch_size": 8, "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": 12, "dataset_prepared_path": "last_run_prepared", "datasets": [ { "chat_template": "jinja", "chat_template_jinja": "{{ bos_token }}\n{% for m in messages %}\n {% set role = 'model' if m['role']=='assistant' else 'user' %}\n {{ '' + role + '\\n' + m['content'] | trim + '\\n' }}\n{% endfor %}\n{% if add_generation_prompt %}\n{{ 'model\\n' }}\n{% endif %}\n", "data_files": "conversations_V3.jsonl", "field_messages": "conversations", "message_property_mappings": { "content": "value", "role": "from" }, "path": "AiAF/conversations", "roles_to_train": [ "assistant" ], "split": "train", "train_on_eos": "turn", "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": "2.7.1" }, "eot_tokens": [ "" ], "eval_batch_size": 2, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_sample_packing": true, "eval_steps": 50, "eval_strategy": "steps", "eval_table_size": 0, "evaluation_strategy": "steps", "experimental_skip_move_to_device": true, "flash_attention": true, "fp16": false, "gradient_accumulation_steps": 4, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "hub_model_id": "AiAF/rp-2b", "hub_strategy": "checkpoint", "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": 1, "lora_alpha": 128, "lora_dropout": 0.05, "lora_r": 64, "lora_target_linear": true, "loraplus_lr_embedding": 1e-06, "lr_scheduler": "cosine", "max_steps": 1000, "mean_resizing_embeddings": false, "micro_batch_size": 2, "model_config_type": "gemma2", "num_epochs": 1.0, "optimizer": "adamw_bnb_8bit", "output_dir": "./outputs/sft/gemma-2-2b-it-rp-sft-qlora", "pad_to_sequence_len": true, "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, "save_steps": 50, "save_strategy": "steps", "save_total_limit": 100, "sequence_len": 2048, "shuffle_before_merging_datasets": false, "shuffle_merged_datasets": true, "skip_prepare_dataset": false, "special_tokens": { "bos_token": "", "eos_token": "", "pad_token": "" }, "streaming": true, "streaming_multipack_buffer_size": 10000, "strict": false, "tensor_parallel_size": 1, "test_datasets": [ { "chat_template": "jinja", "chat_template_jinja": "{{ bos_token }}\n{% for m in messages %}\n {% set role = 'model' if m['role']=='assistant' else 'user' %}\n {{ '' + role + '\\n' + m['content'] | trim + '\\n' }}\n{% endfor %}\n{% if add_generation_prompt %}\n{{ 'model\\n' }}\n{% endif %}\n", "data_files": "eval-datasets/shuf-1000_conversations_V2.jsonl", "field_messages": "conversations", "message_property_mappings": { "content": "value", "role": "from" }, "name": "json", "path": ".", "roles_to_train": [ "assistant" ], "split": "train", "trust_remote_code": false, "type": "chat_template" } ], "tf32": true, "tiled_mlp_use_original_mlp": true, "tokenizer_config": "google/gemma-2-2b-it", "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 }, "type_of_model": "AutoModelForCausalLM", "use_ray": false, "use_wandb": true, "val_set_size": 0.0, "vllm": { "device": "auto", "dtype": "auto", "gpu_memory_utilization": 0.9, "host": "0.0.0.0", "port": 8000 }, "wandb_log_model": "false", "wandb_name": "gemma-2-2b-it-rp-sft-qlora", "wandb_project": "rp-sft", "wandb_run_id": "gemma-2-2b-it-rp-sft-qlora", "weight_decay": 0.0, "world_size": 1 } [2025-10-18 19:02:03,610] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:42363] EOS: 1 / [2025-10-18 19:02:03,610] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:42363] BOS: 2 / [2025-10-18 19:02:03,610] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:42363] PAD: 0 / [2025-10-18 19:02:03,610] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:42363] UNK: 3 / [2025-10-18 19:02:17,503] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:470] [PID:42363] Loading prepared dataset from disk at last_run_prepared/323978649404d0f4da7e1f3e2dc7b3de... [2025-10-18 19:02:17,508] [DEBUG] [axolotl.train.setup_model_and_tokenizer:65] [PID:42363] Loading tokenizer... google/gemma-2-2b-it [2025-10-18 19:02:18,978] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:42363] EOS: 1 / [2025-10-18 19:02:18,978] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:42363] BOS: 2 / [2025-10-18 19:02:18,978] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:42363] PAD: 0 / [2025-10-18 19:02:18,979] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:42363] UNK: 3 / [2025-10-18 19:02:18,979] [DEBUG] [axolotl.train.setup_model_and_tokenizer:74] [PID:42363] Loading model [2025-10-18 19:02:19,156] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:42363] Patched Trainer.evaluation_loop with nanmean loss calculation [2025-10-18 19:02:19,159] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:42363] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation [2025-10-18 19:02:19,160] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:301] [PID:42363] Applying multipack dataloader patch for sample packing... Loading checkpoint shards: 0%| | 0/2 [00:00' + role + '\n' + m['content'] | trim + '\n' }} {% endfor %} {% if add_generation_prompt %} {{ 'model\n' }} {% endif %} --- Tokenizing Prompts (num_proc=12): 0%| | 0/10000 [00:00' + role + '\n' + m['content'] | trim + '\n' }} {% endfor %} {% if add_generation_prompt %} {{ 'model\n' }} {% endif %} --- Tokenizing Prompts (num_proc=12): 0%| | 0/10000 [00:00' + role + '\n' + m['content'] | trim + '\n' }} {% endfor %} {% if add_generation_prompt %} {{ 'model\n' }} {% endif %} --- Tokenizing Prompts (num_proc=12): 0%| | 0/10000 [00:00' + role + '\n' + m['content'] | trim + '\n' }} {% endfor %} {% if add_generation_prompt %} {{ 'model\n' }} {% endif %} --- Tokenizing Prompts (num_proc=12): 0%| | 0/10000 [00:00' + role + '\n' + m['content'] | trim + '\n' }} {% endfor %} {% if add_generation_prompt %} {{ 'model\n' }} {% endif %} --- Tokenizing Prompts (num_proc=12): 0%| | 0/10000 [00:00