See axolotl config

axolotl version: 0.4.1

adapter: qlora
auto_resume_from_checkpoints: true
base_model: unsloth/Qwen2.5-0.5B-Instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - d8b0e80d874b8916_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/d8b0e80d874b8916_train_data.json
  type:
    field_instruction: instruction
    field_output: generation
    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: 100
eval_table_size: null
flash_attention: true
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/46c6853f-c5c7-4cf4-8b2f-14c96891228a
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: 128
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: null
micro_batch_size: 2
mlflow_experiment_name: /tmp/d8b0e80d874b8916_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_torch_4bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
sequence_len: 512
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: d7203ae8-4753-4218-8461-cee5d4eb3802
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: d7203ae8-4753-4218-8461-cee5d4eb3802
warmup_steps: 30
weight_decay: 0.0
xformers_attention: null

46c6853f-c5c7-4cf4-8b2f-14c96891228a

This model is a fine-tuned version of unsloth/Qwen2.5-0.5B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9390

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: 4
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH_4BIT 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
1.2794	0.0008	1	1.4075
0.9889	0.0827	100	0.9601
1.1141	0.1653	200	0.9499
0.9383	0.2480	300	0.9463
0.8717	0.3306	400	0.9463
1.0189	0.4133	500	0.9367
1.0395	0.4960	600	0.9364
1.0817	0.5786	700	0.9325
1.0038	0.6613	800	0.9276
0.922	0.7440	900	0.9246
1.0791	0.8266	1000	0.9254
0.9752	0.9093	1100	0.9164
0.8466	0.9919	1200	0.9137
0.6698	1.0748	1300	0.9461
0.8372	1.1575	1400	0.9421
0.7766	1.2401	1500	0.9390

Framework versions

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