[2026-03-27 09:52:09,822] [DEBUG] [axolotl.utils.config.resolve_dtype:74] [PID:64102] bf16 support detected, enabling for this configuration.
[2026-03-27 09:52:09,985] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:64102] baseline 0.000GB ()
[2026-03-27 09:52:09,985] [INFO] [axolotl.cli.config.load_cfg:341] [PID:64102] config:
{
"activation_offloading": false,
"axolotl_config_path": "ministral3-3b-qlora.yaml",
"base_model": "mistralai/Ministral-3-3B-Instruct-2512-BF16",
"base_model_config": "mistralai/Ministral-3-3B-Instruct-2512-BF16",
"batch_size": 1,
"bf16": true,
"capabilities": {
"bf16": true,
"compute_capability": "sm_89",
"fp8": false,
"n_gpu": 1,
"n_node": 1,
"tf32": true
},
"context_parallel_size": 1,
"cut_cross_entropy": true,
"dataloader_num_workers": 0,
"dataset_num_proc": 16,
"dataset_prepared_path": "last_run_prepared",
"datasets": [
{
"message_property_mappings": {
"content": "content",
"role": "role"
},
"path": "AlexHung29629/test_data_123",
"trust_remote_code": false,
"type": {
"field_instruction": "input",
"field_output": "output",
"field_system": "system",
"format": "{instruction}",
"no_input_format": "{instruction}",
"system_prompt": ""
}
}
],
"ddp": false,
"device": "cuda:0",
"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_sample_packing": true,
"eval_table_size": 0,
"evals_per_epoch": 1,
"experimental_skip_move_to_device": true,
"flex_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": 1,
"gradient_checkpointing": true,
"include_tkps": true,
"is_multimodal": true,
"layer_offloading": false,
"learning_rate": 2e-05,
"lisa_layers_attribute": "model.layers",
"load_best_model_at_end": false,
"load_in_4bit": false,
"load_in_8bit": false,
"local_rank": 0,
"logging_steps": 1,
"lora_dropout": 0.0,
"loraplus_lr_embedding": 1e-06,
"lr_scheduler": "constant",
"max_grad_norm": 1.0,
"mean_resizing_embeddings": false,
"merge_method": "memory_efficient",
"micro_batch_size": 1,
"model_config_type": "mistral3",
"model_config_type_text": "ministral3",
"num_epochs": 2.0,
"num_generation_samples": 3,
"optimizer": "adamw_bnb_8bit",
"otel_metrics_host": "localhost",
"otel_metrics_port": 8000,
"output_dir": "./outputs/out",
"pad_to_sequence_len": true,
"plugins": [
"axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin"
],
"pretrain_multipack_attn": true,
"processor_config": "mistralai/Ministral-3-3B-Instruct-2512-BF16",
"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": 0.5,
"saves_per_epoch": 1,
"scaling_softmax": true,
"sequence_len": 32768,
"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": false,
"tiled_mlp_use_original_mlp": true,
"tokenizer_config": "mistralai/Ministral-3-3B-Instruct-2512-BF16",
"tokenizer_save_jinja_files": true,
"tokenizer_use_mistral_common": false,
"torch_dtype": "torch.bfloat16",
"train_on_inputs": false,
"trl": {
"async_prefetch": false,
"log_completions": false,
"mask_truncated_completions": false,
"ref_model_mixup_alpha": 0.9,
"ref_model_sync_steps": 64,
"replay_buffer_size": 0,
"replay_recompute_logps": true,
"reroll_max_groups": 1,
"reroll_start_fraction": 1.0,
"reward_num_workers": 1,
"scale_rewards": true,
"skip_zero_advantage_batches": true,
"sync_ref_model": false,
"use_data_producer": false,
"use_vllm": false,
"vllm_lora_sync": false,
"vllm_server_host": "0.0.0.0",
"vllm_server_port": 8000
},
"unfrozen_parameters": [
"^model.language_model.norm.weight$",
"^model.language_model.layers.2[0-5].[.a-z_]+$"
],
"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_ratio": 0.0,
"weight_decay": 0.0,
"world_size": 1
}
[2026-03-27 09:52:10,190] [DEBUG] [axolotl.loaders.utils.check_model_config:88] [PID:64102] Loaded image size: 1540 from model config
[2026-03-27 09:52:11,872] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:307] [PID:64102] EOS: 2 /
[2026-03-27 09:52:11,872] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:308] [PID:64102] BOS: 1 /
[2026-03-27 09:52:11,873] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:309] [PID:64102] PAD: 11 /
[2026-03-27 09:52:11,873] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:310] [PID:64102] UNK: 0 /
[2026-03-27 09:52:11,874] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:475] [PID:64102] Loading prepared dataset from disk at last_run_prepared/0a6d77d9f0fbd2dd6692eaf810500a77...
[2026-03-27 09:52:11,880] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:420] [PID:64102] total_num_tokens: 336_571
[2026-03-27 09:52:11,881] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:438] [PID:64102] `total_supervised_tokens: 3_275`
[2026-03-27 09:52:14,284] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:64102] generate_batches time: 0.9476313591003418
[2026-03-27 09:52:15,197] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:64102] generate_batches time: 0.9123454093933105
[2026-03-27 09:52:16,158] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:64102] generate_batches time: 0.9608397483825684
[2026-03-27 09:52:17,093] [DEBUG] [axolotl.utils.samplers.multipack.__len__:462] [PID:64102] generate_batches time: 0.9345409870147705
[2026-03-27 09:52:17,119] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:438] [PID:64102] gather_len_batches: [12]
[2026-03-27 09:52:17,119] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:495] [PID:64102] data_loader_len: 12
[2026-03-27 09:52:17,119] [INFO] [axolotl.utils.trainer.calc_sample_packing_eff_est:504] [PID:64102] sample_packing_eff_est across ranks: [0.9337574351917614]
[2026-03-27 09:52:17,119] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:516] [PID:64102] sample_packing_eff_est: 0.94
[2026-03-27 09:52:17,119] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:521] [PID:64102] total_num_steps: 24
[2026-03-27 09:52:17,119] [INFO] [axolotl.utils.data.sft._prepare_standard_dataset:121] [PID:64102] Maximum number of steps set at 24
[2026-03-27 09:52:17,148] [DEBUG] [axolotl.train.setup_model_and_tokenizer:70] [PID:64102] loading tokenizer... mistralai/Ministral-3-3B-Instruct-2512-BF16
[2026-03-27 09:52:19,018] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:307] [PID:64102] EOS: 2 /
[2026-03-27 09:52:19,019] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:308] [PID:64102] BOS: 1 /
[2026-03-27 09:52:19,019] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:309] [PID:64102] PAD: 11 /
[2026-03-27 09:52:19,019] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:310] [PID:64102] UNK: 0 /
[2026-03-27 09:52:23,670] [DEBUG] [axolotl.train.setup_model_and_tokenizer:81] [PID:64102] Loading model
[2026-03-27 09:52:23,794] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:94] [PID:64102] Patched Trainer.evaluation_loop with nanmean loss calculation
[2026-03-27 09:52:23,796] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:148] [PID:64102] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation
[2026-03-27 09:52:23,797] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:402] [PID:64102] Applying multipack dataloader patch for sample packing...
[2026-03-27 09:52:23,820] [INFO] [axolotl.integrations.cut_cross_entropy.pre_model_load:94] [PID:64102] Applying Cut Cross Entropy to model type: mistral3
Loading weights: 0%| | 0/458 [00:00