You can deploy this model and start earning money today!

ElenaSenger/
DiSTER-Llama-3-8B-Instruct

Name: DiSTER-Llama-3-8B-Instruct
Rating: 2.5 (13 reviews)

safetensors

llama

term-extraction

named-entity-recognition

fine-tuned

domain-adaptation

meta-llama/Meta-Llama-3-8B-Instruct

finetune

llama3

Views

No views yet

DiSTER-Llama-3-8B-Instruct

This is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on our dataset SynTerm for cross-domain technical / scientific term extraction. The model was optimized to output lists of domain-relevant terms given an input text.

Model Details

Base Model: meta-llama/Meta-Llama-3-8B-Instruct
Fine-tuned on: SynTerm (ElenaSenger/SynTerm-fine-tuning)
Tasks: Term Extraction
License: llama3-license

Example Usage

The model was trained with conversation-style prompts, so the same format should be used for inference.

python

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "ElenaSenger/DiSTER-Llama-3-8B-Instruct"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype="auto"
)


# Use the chat format as in SynTerm-fine-tuning
prompt = (
    '{"id": "test_0", "conversations": [\n'
    '    {"from": "human", "value": "Text: We used dropout regularization and AdamW optimizer to train a CNN on MRI images."},\n'
    '    {"from": "gpt", "value": "I\'ve read this text."},\n'
    '    {"from": "human", "value": "What describes (technical or scientific) terms in the text, that are relevant to the domain medical-imaging?"},\n'
    '    {"from": "gpt", "value": ""}\n'
    ']}\n'
)



inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

  with torch.no_grad():
      outputs = model.generate(
            **inputs,
            max_new_tokens=128,
            do_sample=True,  # Enable sampling for temperature to take effect
            temperature=0.1,
            top_p=1.0
        )
  decoded = tokenizer.decode(outputs[0], skip_special_tokens=True)
  print("=== Decoded output ===")
  print(decoded)
  # Print only the model's reply
  reply = decoded[len(prompt):].strip()
  print("\n=== Model reply only ===")
  print(reply)