See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/Llama-3.2-1B-Instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - da89d6d0d6c47035_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/da89d6d0d6c47035_train_data.json
  type:
    field_input: v1_rejected
    field_instruction: prompt
    field_output: ground_truth_chosen
    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: 4
flash_attention: true
fp16:
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/9bbe02f2-64a5-465f-bf4c-4ea2695fbef3
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: true
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
micro_batch_size: 2
mlflow_experiment_name: /tmp/da89d6d0d6c47035_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 4
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: 768
max_steps: 100
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: online
wandb_name: a9d1d646-8ba4-40a6-8b74-692cf970db39
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: a9d1d646-8ba4-40a6-8b74-692cf970db39
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null
special_tokens:
   pad_token: <|end_of_text|>

9bbe02f2-64a5-465f-bf4c-4ea2695fbef3

This model is a fine-tuned version of unsloth/Llama-3.2-1B-Instruct on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2094

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: 10
• training_steps: 100

Training results

Training Loss	Epoch	Step	Validation Loss
1.3023	0.0005	1	1.6697
1.848	0.0037	7	1.5881
1.3956	0.0075	14	1.3877
1.2391	0.0112	21	1.3272
1.497	0.0149	28	1.2778
1.4533	0.0187	35	1.2552
1.2165	0.0224	42	1.2408
1.1767	0.0262	49	1.2297
0.9731	0.0299	56	1.2223
0.8316	0.0336	63	1.2189
1.3272	0.0374	70	1.2140
1.1467	0.0411	77	1.2112
1.2043	0.0448	84	1.2099
1.3629	0.0486	91	1.2091
1.2862	0.0523	98	1.2094

Framework versions

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