adapter: qlora
base_model: VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - aff0bb4998a5c186_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/aff0bb4998a5c186_train_data.json
  type:
    field_input: ''
    field_instruction: input_text
    field_output: target_text
    format: '{instruction}'
    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: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 16
gradient_checkpointing: false
group_by_length: false
hub_model_id: error577/4d263682-5479-402b-86f4-c3d0c7fb15de
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: 32
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_steps: 10
micro_batch_size: 1
mlflow_experiment_name: /tmp/aff0bb4998a5c186_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 4
optimizer: adamw_torch_4bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
saves_per_epoch: 4
sequence_len: 128
special_tokens:
  pad_token: <|eot_id|>
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: 6bc59e91-007b-44b2-ab79-c8605e7e3fd3
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 6bc59e91-007b-44b2-ab79-c8605e7e3fd3
warmup_steps: 5
weight_decay: 0.0
xformers_attention: false

4d263682-5479-402b-86f4-c3d0c7fb15de

This model is a fine-tuned version of VAGOsolutions/Llama-3.1-SauerkrautLM-8b-Instruct on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.3580

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: 1
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 16
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: 5
training_steps: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.7653	0.0006	1	0.8461
0.8482	0.0011	2	0.8418
0.8872	0.0017	3	0.8183
0.7429	0.0023	4	0.7554
0.8967	0.0028	5	0.6476
0.5598	0.0034	6	0.5238
0.5344	0.0040	7	0.4375
0.4827	0.0045	8	0.3868
0.3745	0.0051	9	0.3641
0.2627	0.0057	10	0.3580

Framework versions

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