adapter: lora
base_model: Orenguteng/Llama-3-8B-Lexi-Uncensored
bf16: auto
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 4404b6e6064c8d37_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/4404b6e6064c8d37_train_data.json
  type:
    field_input: story_id
    field_instruction: input_sentence_1
    field_output: sentence_quiz1
    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: 1
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: error577/634fe6e7-ba15-40a0-84cd-c93ce43b7688
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: 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
max_steps: 1000
micro_batch_size: 1
mlflow_experiment_name: /tmp/4404b6e6064c8d37_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 10
optimizer: paged_adamw_32bit
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: 512
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: 279e5cfb-d198-4bf0-8895-5af873459233
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 279e5cfb-d198-4bf0-8895-5af873459233
warmup_steps: 20
weight_decay: 0.0
xformers_attention: null

634fe6e7-ba15-40a0-84cd-c93ce43b7688

This model is a fine-tuned version of Orenguteng/Llama-3-8B-Lexi-Uncensored on the None dataset. It achieves the following results on the evaluation set:

Loss: 3.2154

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: 8
total_train_batch_size: 8
optimizer: Use OptimizerNames.PAGED_ADAMW 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: 20
training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss
4.3457	0.0050	1	4.1774
2.5186	0.5041	100	2.3878
2.131	1.0082	200	2.3347
1.6391	1.5123	300	2.4039
1.1056	2.0164	400	2.4127
1.2669	2.5205	500	2.6209
0.5887	3.0246	600	2.6697
0.5802	3.5287	700	2.9570
0.2365	4.0328	800	3.0128
0.3664	4.5369	900	3.2092
0.1177	5.0410	1000	3.2154

Framework versions

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