See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: NousResearch/Llama-3.2-1B
bf16: true
chat_template: llama3
datasets:
- data_files:
  - 504e27221bbf4070_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/504e27221bbf4070_train_data.json
  type:
    field_input: base
    field_instruction: src
    field_output: 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: false
fp16: false
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: false
hub_model_id: lesso03/9368a66a-7812-42d5-82b5-91eddf3aa1d4
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 1.0e-05
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 32
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 16
lora_target_linear: true
lr_scheduler: cosine
max_memory:
  0: 70GiB
max_steps: 30
micro_batch_size: 4
mlflow_experiment_name: /tmp/504e27221bbf4070_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_torch
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 20
save_strategy: steps
sequence_len: 1024
special_tokens:
  pad_token: <|end_of_text|>
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: 9368a66a-7812-42d5-82b5-91eddf3aa1d4
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 9368a66a-7812-42d5-82b5-91eddf3aa1d4
warmup_steps: 5
weight_decay: 0.01
xformers_attention: false

9368a66a-7812-42d5-82b5-91eddf3aa1d4

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

• Loss: 1.0527

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 5
• training_steps: 30

Training results

Training Loss	Epoch	Step	Validation Loss
1.1438	0.0004	1	1.1001
1.0895	0.0016	4	1.0994
1.1612	0.0033	8	1.0932
0.9957	0.0049	12	1.0818
1.1647	0.0066	16	1.0695
0.8755	0.0082	20	1.0596
1.0094	0.0099	24	1.0547
0.9408	0.0115	28	1.0527

Framework versions

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