This public repository stores a large model artifact from the ko-law-retriever workspace. It is too large for GitHub and is uploaded with Hugging Face's resumable upload-large-folder path.

Scope

Artifact type: full fine-tuned model
This is not a standalone legal-advice system.
Prefer the LoRA adapter releases for lightweight reproduction when possible.

Download

hf download gyung/lfm25-ko-legal-source-bar-full-trial-20260621-v1-final-model --local-dir lfm25-ko-legal-source-bar-full-trial-20260621-v1-final-model

Basic Transformers Loading

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "gyung/lfm25-ko-legal-source-bar-full-trial-20260621-v1-final-model"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype="auto",
    trust_remote_code=True,
)

Related Repositories

Main code: gyung/ko-law-retriever
Large data/eval artifacts: gyung/ko-law-retriever-artifacts-20260622