# axolotl preprocess config.yaml

adapter: lora
base_model: google/gemma-2-9b-it
bf16: auto
dataset_processes: 32
datasets:
- path: frjonah/training_data5
  type:
    system_prompt: ""
    field_system: system
    field_instruction: prompt
    field_output: completion
    format: "[INST] {instruction} [/INST]"
    no_input_format: "[INST] {instruction} [/INST]"

resize_token_embeddings_to_32x: false
add_special_tokens: false
special_tokens:
  pad_token: null
  eos_token: null
  bos_token: null
  unk_token: null

gradient_accumulation_steps: 1
gradient_checkpointing: true
learning_rate: 0.00002
lisa_layers_attribute: model.layers
load_best_model_at_end: false
load_in_4bit: false
load_in_8bit: true
lora_alpha: 256
lora_dropout: 0.1
lora_r: 128
lora_target_modules:
- q_proj
- v_proj
- k_proj
- o_proj
- gate_proj
- down_proj
- up_proj
loraplus_lr_embedding: 1.0e-06
lr_scheduler: cosine
max_prompt_len: 512
mean_resizing_embeddings: false
micro_batch_size: 16
num_epochs: 30.0
optimizer: adamw_bnb_8bit
output_dir: ./outputs/test3
pretrain_multipack_attn: true
pretrain_multipack_buffer_size: 10000
qlora_sharded_model_loading: false
ray_num_workers: 1
resources_per_worker:
  GPU: 1
sample_packing_bin_size: 200
sample_packing_group_size: 100000
save_only_model: false
save_safetensors: true
sequence_len: 2048
shuffle_merged_datasets: true
skip_prepare_dataset: false
strict: false
train_on_inputs: false
trl:
  log_completions: false
  ref_model_mixup_alpha: 0.9
  ref_model_sync_steps: 64
  sync_ref_model: false
  use_vllm: false
  vllm_device: auto
  vllm_dtype: auto
  vllm_gpu_memory_utilization: 0.9
use_ray: false
val_set_size: 0.0
weight_decay: 0.01

outputs/test3

This model is a fine-tuned version of google/gemma-2-9b-it on the frjonah/training_data5 dataset.

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: 16
eval_batch_size: 16
seed: 42
optimizer: Use OptimizerNames.ADAMW_BNB 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: 12
training_steps: 402

Training results

Framework versions

PEFT 0.15.2
Transformers 4.52.3
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.1

frjonah/test7

outputs/test3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

frjonah/
test7