Modelpublic

kkk0123k/gremlin_gwen_14B_lora_model

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes

Model Card

Gremlin Qwen2.5-Coder 14B LoRA Adapter

This is a LoRA adapter trained on Gremlin graph query tasks using the base model `unsloth/qwen2.5-coder-14b-instruct-bnb-4bit`. It uses PEFT (LoRA) for parameter-efficient fine-tuning.

🧠 Model Details

—Base Model: unsloth/qwen2.5-coder-14b-instruct-bnb-4bit
—LoRA Adapter Size: Lightweight adapter (~100MB)
—Trained for: Graph query generation using the Gremlin query language
—Format: PEFT / LoRA compatible with transformers + peft

🔧 How to Use

python

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

base_model = "unsloth/qwen2.5-coder-14b-instruct-bnb-4bit"
adapter_model = "kkk0123k/gremlin_gwen_14B_lora_model"

tokenizer = AutoTokenizer.from_pretrained(base_model)
model = AutoModelForCausalLM.from_pretrained(base_model, device_map="auto", torch_dtype=torch.float16)
model = PeftModel.from_pretrained(model, adapter_model)

inputs = tokenizer("Convert natural language to a Gremlin query: Get all people who know someone named Alice.", return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))