Rethinking Negative Instances for Generative Named Entity Recognition

Model Card for GNER-LLaMA-7B

We introduce GNER, a Generative Named Entity Recognition framework, which demonstrates enhanced zero-shot capabilities across unseen entity domains. Experiments on two representative generative models, i.e., LLaMA and Flan-T5, show that the integration of negative instances into the training process yields substantial performance enhancements. The resulting models, GNER-LLaMA and GNER-T5, outperform state-of-the-art (SoTA) approaches by a large margin, achieving improvements of 8 and 11 points in $F_1$ score, respectively. Code and models are publicly available.

• 💻 Code: https://github.com/yyDing1/GNER/
• 📖 Paper: Rethinking Negative Instances for Generative Named Entity Recognition
• 💾 Models in the 🤗 HuggingFace Hub: GNER-Models
• 🧪 Reproduction Materials: Reproduction Materials
• 🎨 Example Jupyter Notebooks: GNER Notebook

PreTrained Models

We release five GNER models based on LLaMA (7B) and Flan-T5 (base, large, xl and xxl).

Model	# Params	Zero-shot Average $F_1$	Supervised Average $F_1$	🤗 HuggingFace Download Link
GNER-LLaMA	7B	66.1	86.09	link
GNER-T5-base	248M	59.5	83.21	link
GNER-T5-large	783M	63.5	85.45	link
GNER-T5-xl	3B	66.1	85.94	link
GNER-T5-xxl	11B	69.1	86.15	link

Demo usage

You should install the dependencies:

pip install torch datasets deepspeed accelerate transformers protobuf

Please check out Example Jupyter Notebooks for guidance on utilizing GNER models.

A simple inference example is as follows:

Below is an example using GNER-LLaMA

>>> import torch
>>> from transformers import AutoTokenizer, AutoModelForCausalLM
>>> tokenizer = AutoTokenizer.from_pretrained("dyyyyyyyy/GNER-LLaMA-7B")
>>> model = AutoModelForCausalLM.from_pretrained("dyyyyyyyy/GNER-LLaMA-7B", torch_dtype=torch.bfloat16).cuda()
>>> model = model.eval()
>>> instruction_template = "Please analyze the sentence provided, identifying the type of entity for each word on a token-by-token basis.\nOutput format is: word_1(label_1), word_2(label_2), ...\nWe'll use the BIO-format to label the entities, where:\n1. B- (Begin) indicates the start of a named entity.\n2. I- (Inside) is used for words within a named entity but are not the first word.\n3. O (Outside) denotes words that are not part of a named entity.\n"
>>> sentence = "did george clooney make a musical in the 1980s"
>>> entity_labels = ["genre", "rating", "review", "plot", "song", "average ratings", "director", "character", "trailer", "year", "actor", "title"]
>>> instruction = f"{instruction_template}\nUse the specific entity tags: {', '.join(entity_labels)} and O.\nSentence: {sentence}"
>>> instruction = f"[INST] {instruction} [/INST]"
>>> inputs = tokenizer(instruction, return_tensors="pt").to("cuda")
>>> outputs = model.generate(**inputs, max_new_tokens=640)
>>> response = tokenizer.decode(outputs[0], skip_special_tokens=True)
>>> response = response[response.find("[/INST]") + len("[/INST]"):].strip()
>>> print(response)
"did(O) george(B-actor) clooney(I-actor) make(O) a(O) musical(B-genre) in(O) the(O) 1980s(B-year)"

Citation

@misc{ding2024rethinking,
      title={Rethinking Negative Instances for Generative Named Entity Recognition}, 
      author={Yuyang Ding and Juntao Li and Pinzheng Wang and Zecheng Tang and Bowen Yan and Min Zhang},
      year={2024},
      eprint={2402.16602},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}