See axolotl config

axolotl version: 0.6.0

base_model: Qwen/Qwen2.5-7B-Instruct
trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: medalpaca/medical_meadow_medqa
    type: alpaca
dataset_prepared_path:
val_set_size: 0.1
output_dir: ./qlora-qwen25-instruct

sequence_len: 2048
sample_packing: true
eval_sample_packing: true
pad_to_sequence_len: true

adapter: qlora
lora_model_dir:
lora_r: 256
lora_alpha: 256
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 2
num_epochs: 3
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.00002

train_on_inputs: false
group_by_length: false
bf16: true
fp16:
tf32: 

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps:
evals_per_epoch: 4
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
  - full_shard
  - auto_wrap
fsdp_config:
  fsdp_limit_all_gathers: true
  fsdp_sync_module_states: true
  fsdp_offload_params: true
  fsdp_use_orig_params: false
  fsdp_cpu_ram_efficient_loading: true
  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
  fsdp_transformer_layer_cls_to_wrap: Qwen2DecoderLayer
  fsdp_state_dict_type: FULL_STATE_DICT
  fsdp_sharding_strategy: FULL_SHARD
special_tokens:

wandb_project: qlora-qwen-25-7b-instruct
wandb_entity: 
wandb_watch:
wandb_name: 
wandb_log_model: 

hub_model_id: neginashz/qlora-qwen-25-7b-instruct-s
hub_strategy: 
early_stopping_patience:

resume_from_checkpoint:
auto_resume_from_checkpoints: true

qlora-qwen-25-7b-instruct-s

This model is a fine-tuned version of Qwen/Qwen2.5-7B-Instruct on the medalpaca/medical_meadow_medqa dataset. It achieves the following results on the evaluation set:

• Loss: 0.1608

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 8
• total_eval_batch_size: 8
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 14
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.1216	0.2530	42	0.1267
0.1366	0.5060	84	0.1142
0.0914	0.7590	126	0.1104
0.0814	1.0060	168	0.1050
0.0763	1.2590	210	0.1113
0.0746	1.5120	252	0.1147
0.0467	1.7651	294	0.1125
0.0176	2.0120	336	0.1154
0.0367	2.2651	378	0.1605
0.0349	2.5181	420	0.1571
0.0173	2.7711	462	0.1608

Framework versions

• PEFT 0.14.0
• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0