See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/llama-3-8b
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - ed7237070a48a937_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/ed7237070a48a937_train_data.json
  type:
    field_input: tokens
    field_instruction: text
    field_output: ner_tags
    format: '{instruction} {input}'
    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: true
gradient_clipping: 1.0
group_by_length: false
hub_model_id: leixa/2f269ebb-4917-40e0-bd09-8d9b35f735b1
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0001
load_in_4bit: false
load_in_8bit: false
local_rank: 0
logging_steps: 3
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_steps: 100
micro_batch_size: 8
mlflow_experiment_name: /tmp/ed7237070a48a937_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_bnb_8bit
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: 1024
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: techspear-hub
wandb_mode: online
wandb_name: 306028e5-d6af-4d1b-bab1-63a2e73aa431
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 306028e5-d6af-4d1b-bab1-63a2e73aa431
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

2f269ebb-4917-40e0-bd09-8d9b35f735b1

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

• Loss: 0.0423

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0014	1	0.7183
0.4302	0.0127	9	0.3306
0.2097	0.0254	18	0.1815
0.1096	0.0380	27	0.0960
0.0689	0.0507	36	0.0704
0.0632	0.0634	45	0.0626
0.0564	0.0761	54	0.0527
0.0494	0.0888	63	0.0470
0.0413	0.1014	72	0.0443
0.0417	0.1141	81	0.0443
0.0417	0.1268	90	0.0427
0.0397	0.1395	99	0.0423

Framework versions

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