See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: true
base_model: NousResearch/Nous-Hermes-llama-2-7b
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - c1b617ce82c7310e_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/c1b617ce82c7310e_train_data.json
  type:
    field_instruction: prompt
    field_output: chosen
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
early_stopping_threshold: 0.001
eval_max_new_tokens: 128
eval_steps: 20
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: false
group_by_length: false
hub_model_id: mrferr3t/24b7ffa1-83e7-48cf-b96e-84d19d9f7ba9
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0003
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 100
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
micro_batch_size: 32
mlflow_experiment_name: /tmp/c1b617ce82c7310e_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 5
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
s2_attention: null
sample_packing: false
save_steps: 20
saves_per_epoch: 0
sequence_len: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 7f85a073-7b5c-430c-9a22-9fdc7c748e1c
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 7f85a073-7b5c-430c-9a22-9fdc7c748e1c
warmup_ratio: 0.05
weight_decay: 0.0
xformers_attention: null

24b7ffa1-83e7-48cf-b96e-84d19d9f7ba9

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

• Loss: 1.0472

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: 0.0003
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use adamw_bnb_8bit 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: 69
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0009	1	1.1513
No log	0.0180	20	1.1314
No log	0.0360	40	1.0877
No log	0.0540	60	1.0745
No log	0.0720	80	1.0684
1.0992	0.0900	100	1.0650
1.0992	0.1080	120	1.0614
1.0992	0.1260	140	1.0593
1.0992	0.1439	160	1.0582
1.0992	0.1619	180	1.0569
1.0811	0.1799	200	1.0547
1.0811	0.1979	220	1.0529
1.0811	0.2159	240	1.0531
1.0811	0.2339	260	1.0509
1.0811	0.2519	280	1.0513
1.0675	0.2699	300	1.0513
1.0675	0.2879	320	1.0504
1.0675	0.3059	340	1.0485
1.0675	0.3239	360	1.0485
1.0675	0.3419	380	1.0487
1.0808	0.3599	400	1.0465
1.0808	0.3779	420	1.0466
1.0808	0.3959	440	1.0472
1.0808	0.4139	460	1.0472

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.3.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1