daebakgazua/250526_OhLoRA_LLM_SBERT

1. Overview

S-BERT for Oh-LoRA 👱‍♀️ (오로라) LLM memory, for Oh-LoRA v3 Project.

• This S-BERT model is a Fine-tuned version of klue/roberta-base.
• Detailed info (in Korean)
• Memory (RAG-like concept) S-BERT model
• Ethics (detect bad user behavior) S-BERT model

2. Save Path

Save downloaded files in directory 2025_05_26_OhLoRA_v3/llm/models/ethics_sbert/trained_sbert_model & 2025_05_26_OhLoRA_v3/llm/models/memory_sbert/trained_sbert_model as below:

- llm
  - models
    - ethics_sbert
      - trained_sbert_model
        - 1_Pooling
          - config.json
        - eval
          - similarity_evaluation_valid_evaluator_results.csv
        - config.json
        - config_sentence_transformers.json
        - model.safetensors
        - modules.json
        - README.md
        - sentence_bert_config.json
        - special_tokens_map.json
        - tokenizer.json
        - tokenizer_config.json
        - vocab.txt
    - memory_sbert
      - trained_sbert_model
        - 1_Pooling
          - config.json
        - eval
          - similarity_evaluation_valid_evaluator_results.csv
        - config.json
        - config_sentence_transformers.json
        - model.safetensors
        - modules.json
        - README.md
        - sentence_bert_config.json
        - special_tokens_map.json
        - tokenizer.json
        - tokenizer_config.json
        - vocab.txt