See axolotl config

axolotl version: 0.4.1

adapter: lora
auto_find_batch_size: false
base_model: bigcode/starcoder2-3b
bf16: auto
chat_template: llama3
dataloader_num_workers: 12
dataset_prepared_path: null
datasets:
- data_files:
  - 693fb74cca31a376_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/693fb74cca31a376_train_data.json
  type:
    field_input: abstract
    field_instruction: prompt
    field_output: y_true
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 14
eval_strategy: null
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: mrferr3t/6c3b0c75-5225-48e1-9542-ff6e2c9e5e22
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0004
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 14
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 
micro_batch_size: 16
mlflow_experiment_name: /tmp/693fb74cca31a376_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 100
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: /workspace/hub_repo/last-checkpoint
s2_attention: null
sample_packing: false
save_steps: 14
saves_per_epoch: 0
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: 0.05
wandb_entity: null
wandb_mode: 
wandb_name: f5f34ed3-c062-4e3e-8192-6302a80934c2
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: f5f34ed3-c062-4e3e-8192-6302a80934c2
warmup_steps: 100
weight_decay: 0.0
xformers_attention: null

6c3b0c75-5225-48e1-9542-ff6e2c9e5e22

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: 0.1165

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: 0.0004
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• 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: 100
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0086	1	0.2813
20.9456	0.1294	15	0.2647
10.9112	0.2589	30	0.1761
4.7789	0.3883	45	0.1488
2.9387	0.5178	60	0.1449
1.916	0.6472	75	0.1359
2.0292	0.7767	90	0.1308
1.5231	0.9061	105	0.1256
1.6669	1.0399	120	0.1252
1.1464	1.1694	135	0.1226
1.0988	1.2988	150	0.1204
0.9338	1.4283	165	0.1221
1.2596	1.5577	180	0.1170
1.1981	1.6872	195	0.1156
1.1855	1.8166	210	0.1156
1.1852	1.9493	225	0.1137
1.1871	2.0787	240	0.1168
0.9239	2.2082	255	0.1160
0.9451	2.3452	270	0.1165
1.0584	2.4746	285	0.1210
1.0776	2.6041	300	0.1165

Framework versions

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