See axolotl config

axolotl version: 0.4.1

base_model: Qwen/Qwen2.5-14B-Instruct
model_type: Qwen2ForCausalLM
tokenizer_type: Qwen2Tokenizer

trust_remote_code: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: dwikitheduck/genesist-inst-rag-39K
    type: completion
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/lora-out

sequence_len: 4096  
sample_packing: false
pad_to_sequence_len:

adapter: lora
lora_model_dir:
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project: axolotl-soca
wandb_entity: soca-ai
wandb_watch:
wandb_name:
wandb_log_model:

hub_model_id: dwikitheduck/gen-try-1

gradient_accumulation_steps: 8
micro_batch_size: 2
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
s2_attention:

warmup_steps: 10
evals_per_epoch: 2
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:

save_safetensors: true

gen-try-1

This model is a fine-tuned version of Qwen/Qwen2.5-14B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.8327

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• total_eval_batch_size: 4
• optimizer: Use adamw_bnb_8bit 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: 10
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.1832	0.0008	1	1.5919
0.656	0.5003	620	0.8327

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.1+cu124
• Datasets 3.0.1
• Tokenizers 0.20.3