Hugging Face's logo

wangd12
/
sentiment_llama_3_1_8b_8bits_r8_rslora

Safetensors
Model card Files Community
sentiment_llama_3_1_8b_8bits_r8_rslora / README.md
wangd12's picture
wangd12
Add/update LoRA model: sentiment_llama_3_1_8b_8bits_r8_rslora
78367c9 verified
preview code
|
raw
history blame contribute delete
4.04 kB

Built with Axolotl

See axolotl config

axolotl version: 0.9.1.post1

base_model: meta-llama/Llama-3.1-8B-Instruct
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer
gradient_accumulation_steps: 2
micro_batch_size: 8
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0001
load_in_8bit: true
load_in_4bit: false
adapter: lora
lora_model_dir: null
lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
- q_proj
- v_proj
- k_proj
datasets:
- path: /workspace/FinLoRA/data/train/finlora_sentiment_train.jsonl
  type:
    system_prompt: ''
    field_system: system
    field_instruction: context
    field_output: target
    format: '[INST] {instruction} [/INST]'
    no_input_format: '[INST] {instruction} [/INST]'
dataset_prepared_path: null
val_set_size: 0.02
output_dir: /workspace/FinLoRA/lora/axolotl-output/sentiment_llama_3_1_8b_8bits_r8_rslora
peft_use_dora: false
peft_use_rslora: true
sequence_len: 4096
sample_packing: false
pad_to_sequence_len: false
wandb_project: finlora_models
wandb_entity: null
wandb_watch: gradients
wandb_name: sentiment_llama_3_1_8b_8bits_r8_rslora
wandb_log_model: 'false'
bf16: auto
tf32: false
gradient_checkpointing: true
resume_from_checkpoint: null
logging_steps: 500
flash_attention: false
deepspeed: deepspeed_configs/zero1.json
warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
weight_decay: 0.0
special_tokens:
  pad_token: <|end_of_text|>
chat_template: llama3

workspace/FinLoRA/lora/axolotl-output/sentiment_llama_3_1_8b_8bits_r8_rslora

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the /workspace/FinLoRA/data/train/finlora_sentiment_train.jsonl dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2509

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: 0.0001
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 3
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 48
  • total_eval_batch_size: 24
  • 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: 10
  • num_epochs: 4.0

Training results

Training Loss Epoch Step Validation Loss
No log 0.0007 1 3.6225
No log 0.2502 372 0.2253
0.2875 0.5003 744 0.2231
0.1383 0.7505 1116 0.2171
0.1383 1.0007 1488 0.2199
0.1134 1.2508 1860 0.2214
0.0907 1.5010 2232 0.2217
0.0848 1.7512 2604 0.2090
0.0848 2.0013 2976 0.2118
0.0803 2.2515 3348 0.2197
0.0626 2.5017 3720 0.2119
0.0628 2.7518 4092 0.2203
0.0628 3.0020 4464 0.2186
0.0614 3.2522 4836 0.2530
0.0496 3.5024 5208 0.2486
0.0489 3.7525 5580 0.2509

Framework versions

  • PEFT 0.15.2
  • Transformers 4.51.3
  • Pytorch 2.8.0.dev20250319+cu128
  • Datasets 3.5.1
  • Tokenizers 0.21.1