Hugging Face's logo

bzhang0426
/
LLaMA3-70B-qlora-bnb-A61K-15950

PEFT
Safetensors
llama
llama-factory
lora
unsloth
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files Community
LLaMA3-70B-qlora-bnb-A61K-15950 / README.md
bzhang0426's picture
bzhang0426
Upload 15 files
162c34b verified
preview code
|
raw
history blame contribute delete
1.77 kB
metadata
license: other
library_name: peft
tags:
  - llama-factory
  - lora
  - unsloth
  - generated_from_trainer
base_model: unsloth/llama-3-70b-Instruct-bnb-4bit
model-index:
  - name: A61K-15950_2
    results: []

A61K-15950_2

This model is a fine-tuned version of /home/user/research/LLaMA-Factory/saves/LLaMA3-70B-qlora-bnb/lora/sft/A61K-15950/checkpoint-200 on the patents_A61K_cla_abs_det_sample_2100_balance_train_llama-factory-15950 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3507

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: 1
  • eval_batch_size: 1
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 1.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.3663 0.4938 100 0.3549
0.3383 0.9877 200 0.3507

Framework versions

  • PEFT 0.10.0
  • Transformers 4.40.1
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1