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/code-evol-instruct-33k
    type: chat_template
    split: train
test_datasets:
  - path: felixwangg/code-evol-instruct-33k
    type: chat_template
    split: validation
dataset_prepared_path: /home/tkwang/scratch/SecSteer-v2/axolotl-datasets/lora/Qwen2.5-Coder-7B/evol-33k-stage1
val_set_size: 0
output_dir: /home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-evol-33k-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: false

wandb_project: sft-primevul-sweep-ctx-0
wandb_entity: wtkuan
wandb_watch: "false"
wandb_name: Qwen2.5-Coder-7B-evol-33k-stage1
wandb_log_model: "false"


gradient_accumulation_steps: 4
micro_batch_size: 4
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 5e-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: 150
save_steps: 150
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-evol-33k-stage1

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-7B-Instruct on the felixwangg/code-evol-instruct-33k dataset. It achieves the following results on the evaluation set:

Loss: 0.7663
Ppl: 2.1518
Memory/max Active (gib): 38.19
Memory/max Allocated (gib): 38.19
Memory/device Reserved (gib): 51.89

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: 5e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 4
total_train_batch_size: 64
total_eval_batch_size: 16
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: 42
training_steps: 424

Training results

Training Loss	Epoch	Step	Validation Loss	Ppl	Active (gib)	Allocated (gib)	Reserved (gib)
No log	0	0	0.7993	2.2240	37.85	37.85	41.82
0.8092	0.3544	150	0.7686	2.1567	38.19	38.19	51.31
0.7755	0.7088	300	0.7666	2.1525	38.19	38.19	51.89
0.7282	1.0	424	0.7663	2.1518	38.19	38.19	51.89

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-evol-33k-stage1

home/tkwang/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-evol-33k-stage1

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

felixwangg/
Qwen2.5-Coder-7B-evol-33k-stage1