metadata
license: llama2
language:
- en
- ar
metrics:
- accuracy
- f1
library_name: transformers
llama-7b-v2-Receipt-Key-Extraction
llama-7b-v2-Receipt-Key-Extraction is a 7 billion parameter based on LLamA v1
Uses
The model is intended for research-only use in English and Arabic for key information extraction for items in receipts.
How to Get Started with the Model
Use the code below to get started with the model.
# pip install -q transformers
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig
try:
if torch.backends.mps.is_available():
device = "mps"
except:
pass
checkpoint = "abdoelsayed/llama-7b-v2-Receipt-Key-Extraction"
device = "cuda" if torch.cuda.is_available() else "cpu"
tokenizer = AutoTokenizer.from_pretrained(checkpoint, model_max_length=512,
padding_side="right",
use_fast=False,)
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)
def generate_response(instruction, input_text, max_new_tokens=100, temperature=0.1, num_beams=4 , top_p=0.75, top_k=40):
prompt = f"Below is an instruction that describes a task, paired with an input that provides further context.\n\n### Instruction:\n{instruction}\n\n### Input:\n{input_text}\n\n### Response:"
inputs = tokenizer(prompt, return_tensors="pt")
input_ids = inputs["input_ids"].to(device)
generation_config = GenerationConfig(
temperature=temperature,
top_p=top_p,
top_k=top_k,
num_beams=num_beams,
)
with torch.no_grad():
outputs = model.generate(input_ids,generation_config=generation_config, max_new_tokens=max_new_tokens,return_dict_in_generate=True,output_scores=True,)
outputs = tokenizer.decode(outputs.sequences[0])
return outputs.split("### Response:")[-1].strip().replace("</s>","")
instruction = "Extract the class, Brand, Weight, Number of units, Size of units, Price, T.Price, Pack, Unit from the following sentence"
input_text = "Americana Okra zero 400 gm"
response = generate_response(instruction, input_text)
print(response)
How to Cite
Please cite this model using this format.
@misc{abdallah2023amurd,
title={AMuRD: Annotated Multilingual Receipts Dataset for Cross-lingual Key Information Extraction and Classification},
author={Abdelrahman Abdallah and Mahmoud Abdalla and Mohamed Elkasaby and Yasser Elbendary and Adam Jatowt},
year={2023},
eprint={2309.09800},
archivePrefix={arXiv},
primaryClass={cs.CL}
}