See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Yarn-Llama-2-13b-128k
bf16: true
chat_template: llama3
datasets:
- data_files:
  - bd2c9b6bc73ab3f1_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/bd2c9b6bc73ab3f1_train_data.json
  type:
    field_input: content
    field_instruction: path
    field_output: license
    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: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: false
hub_model_id: lesso11/422f759a-85ab-42c5-a424-65df0fc79cde
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
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_memory:
  0: 77GiB
max_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/bd2c9b6bc73ab3f1_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: null
sample_packing: false
save_steps: 25
save_strategy: steps
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: null
wandb_mode: online
wandb_name: 422f759a-85ab-42c5-a424-65df0fc79cde
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 422f759a-85ab-42c5-a424-65df0fc79cde
warmup_steps: 10
weight_decay: 0.01
xformers_attention: false

422f759a-85ab-42c5-a424-65df0fc79cde

This model is a fine-tuned version of NousResearch/Yarn-Llama-2-13b-128k on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2631

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss
14.468	0.0002	1	6.9906
12.5367	0.0015	9	5.4372
3.9705	0.0030	18	1.9645
1.1097	0.0045	27	0.4604
0.9314	0.0060	36	0.4227
0.6036	0.0075	45	0.3315
0.5354	0.0090	54	0.3403
0.4412	0.0104	63	0.2953
0.5696	0.0119	72	0.3060
0.464	0.0134	81	0.2763
0.4093	0.0149	90	0.2648
0.5967	0.0164	99	0.2631

Framework versions

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