metadata

library_name: peft
license: cc
language:
  - en
pipeline_tag: text-generation
base_model:
  - meta-llama/Llama-2-13b-hf

Psychometric LLaMA

This model-repository contains a LLaMA 2 13B LoRA adapter, trained to create psychometric items for psychological testing. I created it as a part of my master's thesis. For further Information about scope and usage visit https://github.com/BjoernGilles/PsychometricLLaMA.

Training procedure

The following bitsandbytes quantization config was used during training:

load_in_8bit: False
load_in_4bit: True
llm_int8_threshold: 6.0
llm_int8_skip_modules: None
llm_int8_enable_fp32_cpu_offload: False
llm_int8_has_fp16_weight: False
bnb_4bit_quant_type: nf4
bnb_4bit_use_double_quant: True
bnb_4bit_compute_dtype: float16

Framework versions

PEFT 0.4.0