adapter: qlora
base_model: unsloth/Qwen2.5-Coder-1.5B-Instruct
bf16: auto
chat_template: llama3
dataloader_num_workers: 6
dataset_prepared_path: null
datasets:
- data_files:
  - 000dac3a8cb81c80_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/000dac3a8cb81c80_train_data.json
  type:
    field_input: ''
    field_instruction: instruction
    field_output: output
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 3
eval_max_new_tokens: 128
eval_steps: 50
eval_table_size: null
evals_per_epoch: null
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: true
group_by_length: true
hub_model_id: error577/e4d743c3-e9f1-423a-a19a-e0d6ed3f5f22
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0003
load_in_4bit: true
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.3
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 128
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 600
auto_resume_from_checkpoints: true
micro_batch_size: 1
mlflow_experiment_name: /tmp/000dac3a8cb81c80_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 50
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: 50
sequence_len: 512
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.002
wandb_entity: null
wandb_mode: online
wandb_name: c30b7643-9ea3-489c-b95b-6fdd775d8f75
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: c30b7643-9ea3-489c-b95b-6fdd775d8f75
warmup_steps: 10
weight_decay: 0.01
xformers_attention: null

e4d743c3-e9f1-423a-a19a-e0d6ed3f5f22

This model is a fine-tuned version of unsloth/Qwen2.5-Coder-1.5B-Instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5509

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: 0.0003
train_batch_size: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
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: 10
training_steps: 600

Training results

Training Loss	Epoch	Step	Validation Loss
0.6802	0.0001	1	0.6829
0.6973	0.0056	50	0.5870
0.5528	0.0112	100	0.5732
0.5774	0.0168	150	0.5664
0.507	0.0224	200	0.5608
0.5901	0.0280	250	0.5569
0.5583	0.0336	300	0.5562
0.6019	0.0392	350	0.5567
0.5378	0.0448	400	0.5553
0.4796	0.0504	450	0.5526
0.5788	0.0560	500	0.5517
0.5497	0.0616	550	0.5511
0.5998	0.0672	600	0.5509

Framework versions

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