See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/gemma-2-9b-it
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - b8257c8e0fe17671_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/b8257c8e0fe17671_train_data.json
  type:
    field_input: choices
    field_instruction: question
    field_output: subject
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: false
fp16: true
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: ardaspear/bebbaeef-09b0-445e-9095-26a2b67be0db
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 3
lora_alpha: 128
lora_dropout: 0.1
lora_fan_in_fan_out: true
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
max_memory:
  0: 72GB
max_steps: 150
micro_batch_size: 4
mlflow_experiment_name: /tmp/b8257c8e0fe17671_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: false
sample_packing: false
saves_per_epoch: 4
sequence_len: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: leixa-personal
wandb_mode: online
wandb_name: bebbaeef-09b0-445e-9095-26a2b67be0db
wandb_project: Gradients-On-Two
wandb_run: your_name
wandb_runid: bebbaeef-09b0-445e-9095-26a2b67be0db
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

bebbaeef-09b0-445e-9095-26a2b67be0db

This model is a fine-tuned version of unsloth/gemma-2-9b-it on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0004

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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: 10
• training_steps: 8

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.3636	1	9.7078
No log	0.7273	2	8.5528
10.3937	1.1364	3	5.9608
10.3937	1.5	4	3.2141
10.3937	1.8636	5	0.5463
3.3022	2.2727	6	0.0036
3.3022	2.6364	7	0.0004
3.3022	3.0455	8	0.0004

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1