base_model: Qwen/Qwen2.5-Coder-7B-Instruct
model_type: Qwen2ForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false

datasets:
  - path: felixwangg/codenet-c-cpp-stage1
    type: chat_template
    split: train
test_datasets:
  - path: felixwangg/codenet-c-cpp-stage1
    type: chat_template
    split: validation
dataset_prepared_path: /home/tkwang/scratch/SecSteer-v2/axolotl-datasets/lora/Qwen2.5-Coder-7B/func-stage1
val_set_size: 0
output_dir: /home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-func-stage1
sequence_len: 4096
sample_packing: false
eval_sample_packing: false
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 16
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
merge_lora: true

wandb_project: sft-primevul-sweep-ctx-0
wandb_entity: wtkuan
wandb_watch: "false"
wandb_name: Qwen2.5-Coder-7B-func-stage1
wandb_log_model: "false"


gradient_accumulation_steps: 8
micro_batch_size: 4
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 4e-5

bf16: true
tf32: false

train_on_inputs: false
roles_to_train: ['assistant']

gradient_checkpointing: true
resume_from_checkpoint:
logging_steps: 1
flash_attention: true

num_epochs: 1
warmup_ratio: 0.1
early_stopping_patience: 1000
eval_steps: 5
save_steps: 5
save_total_limit: 1000
load_best_model_at_end: true

weight_decay: 0.02
special_tokens:

plugins:

home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-func-stage1

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-7B-Instruct on the felixwangg/codenet-c-cpp-stage1 dataset. It achieves the following results on the evaluation set:

Loss: 0.6311
Ppl: 1.8797
Memory/max Active (gib): 38.19
Memory/max Allocated (gib): 38.19
Memory/device Reserved (gib): 52.47

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: 4e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 2
gradient_accumulation_steps: 8
total_train_batch_size: 64
total_eval_batch_size: 8
optimizer: Use OptimizerNames.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: 2
training_steps: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Ppl	Active (gib)	Allocated (gib)	Reserved (gib)
No log	0	0	0.6554	1.9259	37.85	37.85	41.82
0.5850	0.3540	5	0.6517	1.9189	38.19	38.19	51.31
0.6071	0.7080	10	0.6357	1.8884	38.19	38.19	52.47
0.6697	1.0	15	0.6311	1.8797	38.19	38.19	52.47

Framework versions

PEFT 0.19.1
Transformers 5.5.4
Pytorch 2.11.0+cu130
Datasets 4.5.0
Tokenizers 0.22.2

felixwangg/Qwen2.5-Coder-7B-func-stage1

home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-func-stage1

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

felixwangg/
Qwen2.5-Coder-7B-func-stage1