See axolotl config

axolotl version: 0.4.1

adapter: qlora
auto_resume_from_checkpoints: true
base_model: bigcode/starcoder2-3b
bf16: auto
chat_template: llama3
dataset_prepared_path: null
dataset_processes: 6
datasets:
- data_files:
  - ad8f783c2fd92065_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/ad8f783c2fd92065_train_data.json
  type:
    field_input: span_labels
    field_instruction: source_text
    field_output: target_text
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 5
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: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/a0abd2fc-0e9c-4a53-83bd-5435c4990984
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: true
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: 2
mlflow_experiment_name: /tmp/ad8f783c2fd92065_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
save_steps: 200
sequence_len: 512
special_tokens:
  pad_token: <|endoftext|>
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.005
wandb_entity: null
wandb_mode: online
wandb_name: 2807bbf9-436c-4c90-b92a-3dbfd7f919cd
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 2807bbf9-436c-4c90-b92a-3dbfd7f919cd
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

a0abd2fc-0e9c-4a53-83bd-5435c4990984

This model is a fine-tuned version of bigcode/starcoder2-3b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0052

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• 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: 30
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
18.6994	0.0001	1	0.3745
0.3429	0.0143	200	0.0124
0.1255	0.0287	400	0.0100
0.5354	0.0430	600	0.0093
0.1077	0.0573	800	0.0069
0.1981	0.0717	1000	0.0083
0.0596	0.0860	1200	0.0060
0.0976	0.1003	1400	0.0060
0.0257	0.1147	1600	0.0061
0.0718	0.1290	1800	0.0059
0.1601	0.1433	2000	0.0063
0.0453	0.1576	2200	0.0051
0.1005	0.1720	2400	0.0055
0.1717	0.1863	2600	0.0053
0.0081	0.2006	2800	0.0054
0.0468	0.2150	3000	0.0055
0.0827	0.2293	3200	0.0052

Framework versions

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