See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Llama-3.2-3B
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - cfe4cfbd460d8124_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/cfe4cfbd460d8124_train_data.json
  type:
    field_input: OriginalAddress1
    field_instruction: PermitTypeDesc
    field_output: Description
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
device_map:
  ? ''
  : 0,1,2,3,4,5,6,7
early_stopping_patience: 2
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: true
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: Alphatao/bfdb40d4-c9b7-45f9-b919-894a52b54660
hub_repo: null
hub_strategy: null
hub_token: null
learning_rate: 0.0002
load_best_model_at_end: true
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
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
- down_proj
- up_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 2070
micro_batch_size: 4
mlflow_experiment_name: /tmp/cfe4cfbd460d8124_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
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: 100
sequence_len: 2048
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.04127830659875009
wandb_entity: null
wandb_mode: online
wandb_name: 1c0ec80c-26dd-4dfe-858e-1b40744d4a00
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 1c0ec80c-26dd-4dfe-858e-1b40744d4a00
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

bfdb40d4-c9b7-45f9-b919-894a52b54660

This model is a fine-tuned version of unsloth/Llama-3.2-3B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.4649

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• 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: 10
• training_steps: 2070

Training results

Training Loss	Epoch	Step	Validation Loss
4.5394	0.0003	1	4.5230
1.9168	0.0276	100	1.8147
1.6452	0.0551	200	1.7164
1.6071	0.0827	300	1.6726
1.7231	0.1102	400	1.6363
1.2157	0.1378	500	1.6138
1.5274	0.1653	600	1.5968
1.5334	0.1929	700	1.5771
1.6462	0.2204	800	1.5659
1.7249	0.2480	900	1.5432
1.6462	0.2755	1000	1.5300
1.4306	0.3031	1100	1.5184
1.673	0.3307	1200	1.5094
1.5232	0.3582	1300	1.4975
1.6796	0.3858	1400	1.4885
1.6199	0.4133	1500	1.4817
1.6107	0.4409	1600	1.4766
1.5798	0.4684	1700	1.4720
1.7002	0.4960	1800	1.4678
1.4047	0.5235	1900	1.4657
1.1164	0.5511	2000	1.4649

Framework versions

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