See axolotl config

axolotl version: 0.4.1

adapter: qlora
base_model: DeepMount00/Llama-3-8b-Ita
bf16: auto
chat_template: llama3
dataloader_num_workers: 6
dataset_prepared_path: null
datasets:
- data_files:
  - 8bb3de2515f68e34_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/8bb3de2515f68e34_train_data.json
  type:
    field_instruction: instruction
    field_output: output
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping:
  metric: eval_loss
  mode: min
  patience: 3
eval_max_new_tokens: 128
eval_steps: 20
eval_table_size: null
evals_per_epoch: null
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: true
group_by_length: true
hub_model_id: error577/a0f51937-c16e-40f7-9eb9-a36807c20372
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0003
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.3
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 300
micro_batch_size: 1
mlflow_experiment_name: /tmp/8bb3de2515f68e34_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 50
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: 20
sequence_len: 512
special_tokens:
  pad_token: <|eot_id|>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.02
wandb_entity: null
wandb_mode: online
wandb_name: 3402e0d8-6057-4b92-a376-02ad538dbab9
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 3402e0d8-6057-4b92-a376-02ad538dbab9
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

a0f51937-c16e-40f7-9eb9-a36807c20372

This model is a fine-tuned version of DeepMount00/Llama-3-8b-Ita on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1541

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• 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: 300

Training results

Training Loss	Epoch	Step	Validation Loss
0.2993	0.0008	1	0.5944
0.1408	0.0165	20	0.2069
0.2487	0.0330	40	0.1814
0.1628	0.0495	60	0.1897
0.1704	0.0660	80	0.1721
0.1819	0.0825	100	0.1865
0.1674	0.0990	120	0.1747
0.1445	0.1154	140	0.1637
0.1094	0.1319	160	0.1691
0.133	0.1484	180	0.1613
0.2585	0.1649	200	0.1585
0.1557	0.1814	220	0.1573
0.1189	0.1979	240	0.1555
0.1831	0.2144	260	0.1547
0.1477	0.2309	280	0.1542
0.1795	0.2474	300	0.1541

Framework versions

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