See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: unsloth/llama-2-7b-chat
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 5419de7e3f52d124_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/5419de7e3f52d124_train_data.json
  type:
    field_instruction: short_questions
    field_output: multiple_questions
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
device: cuda
early_stopping_patience: null
eval_max_new_tokens: 128
eval_table_size: null
evals_per_epoch: 4
flash_attention: false
fp16: null
fsdp: null
fsdp_config:
  max_steps: 50
  weight_decay: 0.01
gradient_accumulation_steps: 4
gradient_checkpointing: false
group_by_length: false
hub_model_id: nadejdatarabukina/c5765cf0-ec39-4ba5-b0b2-77f0aa40129a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 3
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_memory:
  0: 70GiB
max_steps: 48
micro_batch_size: 2
mlflow_experiment_name: /tmp/5419de7e3f52d124_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 5
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: 70
sequence_len: 1024
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: c5765cf0-ec39-4ba5-b0b2-77f0aa40129a
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: c5765cf0-ec39-4ba5-b0b2-77f0aa40129a
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

c5765cf0-ec39-4ba5-b0b2-77f0aa40129a

This model is a fine-tuned version of unsloth/llama-2-7b-chat on the None dataset. It achieves the following results on the evaluation set:

• Loss: nan

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 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: 48

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0022	1	nan
0.0	0.0066	3	nan
0.0	0.0132	6	nan
0.0	0.0198	9	nan
0.0	0.0264	12	nan
0.0	0.0330	15	nan
0.0	0.0396	18	nan
0.0	0.0463	21	nan
0.0	0.0529	24	nan
0.0	0.0595	27	nan
0.0	0.0661	30	nan
0.0	0.0727	33	nan
0.0	0.0793	36	nan
0.0	0.0859	39	nan
0.0	0.0925	42	nan
0.0	0.0991	45	nan
0.0	0.1057	48	nan

Framework versions

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