adapter: qlora
base_model: deepseek-ai/deepseek-coder-6.7b-instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - a8d492346672a1c8_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/a8d492346672a1c8_train_data.json
  type:
    field_input: source
    field_instruction: prompt
    field_output: instruction
    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: 1
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/bc5a1285-38b6-4c0e-9645-2dfbb7abf009
hub_repo: null
hub_strategy: end
hub_token: null
learning_rate: 0.0001
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
max_steps: 300
micro_batch_size: 1
mlflow_experiment_name: /tmp/a8d492346672a1c8_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 12
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: 1
sequence_len: 512
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: c8942af0-5e05-4c22-897b-cdb519f12f7b
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: c8942af0-5e05-4c22-897b-cdb519f12f7b
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

bc5a1285-38b6-4c0e-9645-2dfbb7abf009

This model is a fine-tuned version of deepseek-ai/deepseek-coder-6.7b-instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.0501

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: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 8
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
3.7335	0.0029	1	3.8303
1.5917	0.0729	25	2.3094
0.9143	0.1458	50	1.4404
1.7505	0.2187	75	1.2449
1.2376	0.2915	100	1.1931
1.2788	0.3644	125	1.1278
1.4158	0.4373	150	1.0998
0.9088	0.5102	175	1.0870
1.1591	0.5831	200	1.0699
1.0273	0.6560	225	1.0602
1.9042	0.7289	250	1.0567
0.9941	0.8017	275	1.0524
1.1544	0.8746	300	1.0501

Framework versions

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