See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/llama-3-8b-Instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 4068f8101ab29bee_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/4068f8101ab29bee_train_data.json
  type:
    field_instruction: infobox
    field_output: summary
    format: '{instruction}'
    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: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: false
group_by_length: false
hub_model_id: dimasik87/6813e4a9-59b5-485f-b2d9-38de43e7887d
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: 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: 70GiB
max_steps: 50
micro_batch_size: 1
mlflow_experiment_name: /tmp/4068f8101ab29bee_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 4
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 4
sequence_len: 2028
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 6813e4a9-59b5-485f-b2d9-38de43e7887d
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 6813e4a9-59b5-485f-b2d9-38de43e7887d
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

6813e4a9-59b5-485f-b2d9-38de43e7887d

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

• Loss: 0.2675

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss
0.7376	0.0027	1	0.7193
0.6835	0.0108	4	0.6272
0.3742	0.0215	8	0.3806
0.2788	0.0323	12	0.3253
0.5915	0.0431	16	0.2901
0.2583	0.0538	20	0.2826
0.2147	0.0646	24	0.2890
0.2077	0.0754	28	0.2844
0.4248	0.0861	32	0.2731
0.3099	0.0969	36	0.2734
0.2465	0.1077	40	0.2704
0.2693	0.1184	44	0.2685
0.231	0.1292	48	0.2675

Framework versions

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