See axolotl config

axolotl version: 0.5.2

adapter: lora
auto_find_batch_size: true
base_model: NousResearch/Nous-Hermes-2-Mistral-7B-DPO
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 039e297ae683b655_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/039e297ae683b655_train_data.json
  type:
    field_input: input
    field_instruction: instruction
    field_output: output
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: /workspace/axolotl/configs/deepspeed_stage2.json
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_sample_packing: false
eval_steps: 10
eval_strategy: steps
eval_table_size: null
flash_attention: true
fp16: false
gpu_memory_limit: 80GiB
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: true
hub_model_id: PhoenixB/b4135751-e0a8-44f9-ab89-14d1a2372fa2
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 2e-4
liger_fused_linear_cross_entropy: true
liger_glu_activation: true
liger_layer_norm: true
liger_rms_norm: true
liger_rope: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 5
lora_alpha: 16
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: 10000
micro_batch_size: 2
mlflow_experiment_name: /tmp/039e297ae683b655_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_torch_fused
output_dir: miner_id_24
pad_to_sequence_len: true
plugins:
- axolotl.integrations.liger.LigerPlugin
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 20
sequence_len: 8196
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: 744374a3-fddf-43c9-b5b6-239201c8a6f3
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 744374a3-fddf-43c9-b5b6-239201c8a6f3
warmup_steps: 20
weight_decay: 0.0

b4135751-e0a8-44f9-ab89-14d1a2372fa2

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: 0.4380

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: 4
• total_train_batch_size: 16
• total_eval_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 20
• training_steps: 1604

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0019	1	0.7127
0.588	0.0187	10	0.6048
0.5482	0.0374	20	0.5107
0.4637	0.0561	30	0.4935
0.4991	0.0748	40	0.4785
0.5014	0.0935	50	0.4741
0.452	0.1123	60	0.4645
0.475	0.1310	70	0.4595
0.4362	0.1497	80	0.4561
0.4559	0.1684	90	0.4512
0.468	0.1871	100	0.4463
0.4543	0.2058	110	0.4432
0.4628	0.2245	120	0.4371
0.4253	0.2432	130	0.4436
0.4366	0.2619	140	0.4373
0.4712	0.2806	150	0.4380

Framework versions

• PEFT 0.13.2
• Transformers 4.46.3
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3