adapter: lora
base_model: unsloth/Llama-3.2-1B-Instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - MATH-Hard_train_data.json
  ds_type: json
  path: /workspace/input_data/MATH-Hard_train_data.json
  type:
    field_input: solution
    field_instruction: type
    field_output: problem
    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: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: ncbateman/60c4f9de-24ff-4eeb-a041-3c142f4f4888
hub_repo: ncbateman
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: true
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: cosine
max_steps: 1000
micro_batch_size: 2
mlflow_experiment_name: /tmp/MATH-Hard_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 1
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
save_strategy: steps
sequence_len: 4096
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
val_set_size: 0.05
wandb_entity: breakfasthut
wandb_mode: online
wandb_project: tuning-miner
wandb_run: miner
wandb_runid: 60c4f9de-24ff-4eeb-a041-3c142f4f4888
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

60c4f9de-24ff-4eeb-a041-3c142f4f4888

This model is a fine-tuned version of unsloth/Llama-3.2-1B-Instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.9063

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: 2
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 4
total_train_batch_size: 32
total_eval_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 50
training_steps: 105

Training results

Training Loss	Epoch	Step	Validation Loss
1.4365	0.0095	1	1.4671
1.1781	0.2578	27	1.0924
0.9832	0.5155	54	0.9416
0.7667	0.7733	81	0.9063

Framework versions

PEFT 0.13.2
Transformers 4.45.2
Pytorch 2.4.1+cu124
Datasets 3.0.1
Tokenizers 0.20.1