See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: WhiteRabbitNeo/Llama-3-WhiteRabbitNeo-8B-v2.0
bf16: auto
datasets:
- data_files:
  - a11f8217f40ca211_train_data.json
  ds_type: json
  format: custom
  path: a11f8217f40ca211_train_data.json
  type:
    field: null
    field_input: null
    field_instruction: input
    field_output: response_a
    field_system: null
    format: null
    no_input_format: null
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: null
eval_max_new_tokens: 128
eval_sample_packing: false
eval_table_size: null
evals_per_epoch: 4
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
hub_model_id: FatCat87/taopanda-3_b49985cb-1972-4835-840c-c05792e5f494
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: true
local_rank: null
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_r: 32
lora_target_linear: true
lr_scheduler: cosine
micro_batch_size: 2
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_bnb_8bit
output_dir: ./outputs/out/taopanda-3_b49985cb-1972-4835-840c-c05792e5f494
pad_to_sequence_len: true
resume_from_checkpoint: null
sample_packing: true
saves_per_epoch: 1
seed: 18127
sequence_len: 4096
special_tokens: null
strict: false
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.1
wandb_entity: fatcat87-taopanda
wandb_log_model: null
wandb_mode: online
wandb_name: taopanda-3_b49985cb-1972-4835-840c-c05792e5f494
wandb_project: subnet56
wandb_runid: taopanda-3_b49985cb-1972-4835-840c-c05792e5f494
wandb_watch: null
warmup_ratio: 0.05
weight_decay: 0.0
xformers_attention: null

taopanda-3_b49985cb-1972-4835-840c-c05792e5f494

This model is a fine-tuned version of WhiteRabbitNeo/Llama-3-WhiteRabbitNeo-8B-v2.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6800

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: 18127
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 2
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.222	0.0396	1	1.2185
0.9168	0.2772	7	0.8508
0.7745	0.5545	14	0.7478
0.7129	0.8317	21	0.7133
0.7005	1.0792	28	0.6945
0.6863	1.3564	35	0.6855
0.6688	1.6337	42	0.6808
0.6772	1.9109	49	0.6800

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1