See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Yarn-Llama-2-7b-64k
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 396327bda852888c_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/396327bda852888c_train_data.json
  type:
    field_instruction: title
    field_output: abstract
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 5
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: false
group_by_length: false
hub_model_id: tuantmdev/ef4bc250-47fb-42e2-b3cb-78e33e14928a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 2e-05
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 40
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
max_steps: 200
micro_batch_size: 2
mlflow_experiment_name: /tmp/396327bda852888c_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_strategy: best
saves_per_epoch: 5
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: a16fc738-6a5a-4a95-a014-1684ce5ab17a
wandb_project: Gradients-On-Demand
wandb_run: unknown
wandb_runid: a16fc738-6a5a-4a95-a014-1684ce5ab17a
warmup_steps: 80
weight_decay: 0.01
xformers_attention: null

ef4bc250-47fb-42e2-b3cb-78e33e14928a

This model is a fine-tuned version of NousResearch/Yarn-Llama-2-7b-64k on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.7235

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: 2e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB 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: 80
• training_steps: 200

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0001	1	2.1100
33.3493	0.0044	40	2.0666
30.9777	0.0088	80	1.8364
28.7871	0.0132	120	1.7560
28.0053	0.0175	160	1.7281
27.6745	0.0219	200	1.7235

Framework versions

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