See axolotl config

axolotl version: 0.10.0.dev0

adapter: lora
base_model: tiiuae/falcon-7b
bf16: auto
chat_template: llama3
datasets:
- data_files:
  - ba46d26616a00d4d_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/
  type:
    field_input: None
    field_instruction: instruct
    field_output: output
    field_system: None
    format: None
    no_input_format: None
    system_format: '{system}'
    system_prompt: None
eval_max_new_tokens: 256
evals_per_epoch: 2
flash_attention: false
fp16: false
gradient_accumulation_steps: 1
gradient_checkpointing: true
group_by_length: true
hub_model_id: apriasmoro/1e44d36c-d8d9-41f4-9e74-ce5a46f731d8
learning_rate: 0.0002
logging_steps: 10
lora_alpha: 16
lora_dropout: 0.05
lora_fan_in_fan_out: false
lora_r: 8
lora_target_linear: true
lr_scheduler: cosine
max_steps: 57
micro_batch_size: 4
mixed_precision: bf16
mlflow_experiment_name: /tmp/ba46d26616a00d4d_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
sample_packing: false
save_steps: 18
sequence_len: 2048
special_tokens:
  pad_token: <|endoftext|>
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.05
wandb_entity: null
wandb_mode: online
wandb_name: 68211432-3aeb-4a3e-8679-cdad4b1d8067
wandb_project: Gradients-On-Demand
wandb_run: apriasmoro
wandb_runid: 68211432-3aeb-4a3e-8679-cdad4b1d8067
warmup_steps: 100
weight_decay: 0.01

1e44d36c-d8d9-41f4-9e74-ce5a46f731d8

This model is a fine-tuned version of tiiuae/falcon-7b on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.5030

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: 4
• eval_batch_size: 4
• seed: 42
• 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: 100
• training_steps: 57

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0005	1	2.9748
2.3031	0.0053	10	2.9873
2.0132	0.0106	20	2.9482
1.8501	0.0159	30	2.8389
3.2085	0.0212	40	2.5878
3.2721	0.0265	50	2.5030

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.5.1+cu124
• Datasets 3.5.1
• Tokenizers 0.21.1