shuoxing/Llama3-8b-Review-lora-lr-1e-5-a-256-alpha-512

See axolotl config

axolotl version: 0.7.0

base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
# optionally might have model_type or tokenizer_type
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

load_in_8bit: true
load_in_4bit: false
strict: false

chat_template: llama3
datasets:
  - path: data/training_data.json
    type: chat_template
    field_messages: messages
    message_property_mappings:
      role: role
      content: content
    roles:
      user:
        - user
      assistant:
        - assistant

dataset_prepared_path:
val_set_size: 0.1
output_dir: ./output/llama_1e-5_r_256_alpha_512

sequence_len: 8000
sample_packing: false
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 256
lora_alpha: 512
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: llama-review
wandb_entity:
wandb_watch:
wandb_name: llama_1e-5_r_256_alpha_512
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 1e-5

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
s2_attention:

warmup_steps: 100
evals_per_epoch: 
eval_table_size:
eval_max_new_tokens:
saves_per_epoch: 1
debug:
deepspeed: deepspeed_3.json
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
   pad_token: <|end_of_text|>

output/llama_1e-5_r_256_alpha_512

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the data/training_data.json dataset. It achieves the following results on the evaluation set:

• Loss: 13.8053

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• total_eval_batch_size: 4
• optimizer: Use adamw_bnb_8bit 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
• num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
13.8163	0.9995	491	13.8053

Framework versions

• PEFT 0.14.0
• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0