See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_resume_from_checkpoints: false
base_model: JackFram/llama-68m
bf16: auto
chat_template: llama3
dataset_prepared_path: null
dataset_processes: 
datasets:
- data_files:
  - 623d9787d7fd7d0e_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/623d9787d7fd7d0e_train_data.json
  type:
    field_instruction: text
    field_output: title
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 200
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: false
hub_model_id: error577/24906023-f4b5-47c9-a7fb-6b92b845bc17
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: null
micro_batch_size: 24
mlflow_experiment_name: /tmp/623d9787d7fd7d0e_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 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: 200
sequence_len: 256
special_tokens:
  pad_token: </s>
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: af6de128-8865-42e8-800b-ff2d2b1acccd
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: af6de128-8865-42e8-800b-ff2d2b1acccd
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

24906023-f4b5-47c9-a7fb-6b92b845bc17

This model is a fine-tuned version of JackFram/llama-68m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0011

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: 24
• eval_batch_size: 24
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 48
• 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: 30
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
2.152	0.0016	1	2.3080
0.0209	0.3108	200	0.0117
0.0028	0.6216	400	0.0062
0.0012	0.9324	600	0.0035
0.0022	1.2432	800	0.0031
0.0013	1.5540	1000	0.0030
0.001	1.8648	1200	0.0022
0.0026	2.1756	1400	0.0018
0.0011	2.4864	1600	0.0018
0.0005	2.7972	1800	0.0017
0.0004	3.1080	2000	0.0012
0.0074	3.4188	2200	0.0011
0.0003	3.7296	2400	0.0012
0.0003	4.0404	2600	0.0011
0.0006	4.3512	2800	0.0011
0.0005	4.6620	3000	0.0011
0.0009	4.9728	3200	0.0011

Framework versions

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