See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Nous-Hermes-2-Mistral-7B-DPO
bf16: true
chat_template: llama3
data_processes: 16
dataset_prepared_path: null
datasets:
- data_files:
  - 1029f694c22a0116_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/1029f694c22a0116_train_data.json
  type:
    field_instruction: instructions
    field_output: content
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
do_eval: true
early_stopping_patience: 5
eval_batch_size: 4
eval_max_new_tokens: 128
eval_steps: 50
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: true
hub_model_id: lesso/0a04dd1a-7337-44bc-85f9-d780e7c92e21
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001001
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 128
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: linear
max_grad_norm: 1.0
max_steps: 200
micro_batch_size: 4
mlflow_experiment_name: /tmp/god01/1029f694c22a0116_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optim_args:
  adam_beta1: 0.9
  adam_beta2: 0.95
  adam_epsilon: 1e-5
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: 50
saves_per_epoch: null
sequence_len: 1024
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 7aeff1cf-86b0-475f-8cec-ec31521214cb
wandb_project: ab-god01
wandb_run: your_name
wandb_runid: 7aeff1cf-86b0-475f-8cec-ec31521214cb
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

0a04dd1a-7337-44bc-85f9-d780e7c92e21

This model is a fine-tuned version of NousResearch/Nous-Hermes-2-Mistral-7B-DPO on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3480

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: 0.0001001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 10
• training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss
2.3012	0.0012	1	1.1002
1.0936	0.0613	50	0.4213
0.9928	0.1226	100	0.3831
0.945	0.1839	150	0.3597
0.8969	0.2452	200	0.3480

Framework versions

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