richarderkhov/saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf overview

{
  "metadata": {},
  "card_data": {
    "frontmatter": {},
    "hero_image_url": "https://www.linkbricks.com/wp-content/uploads/2022/03/%E1%84%85%E1%85%B5%E1%86%BC%E1%84%8F%E1%85%B3%E1%84%87%E1%85%B3%E1%84%85%E1%85%B5%E1%86%A8%E1%84%89%E1%85%B3%E1%84%85%E1%85%A9%E1%84%80%E1%85%A9-2-1024x804.png",
    "summary": "AI 와 빅데이터 분석 전문 기업인 Linkbricks의 데이터사이언티스트인 지윤성 박사(Saxo)가 beomi/llama-2-koen-13b 베이스모델을 GCP상의 A100-40G 4개를 통해 SFT 훈련을 한(2048 Tokens) 인스트럭션 모델. Accelerate, Deepspeed Zero-3 라이브러리를 사용했으며 Flash Attention 은 Disable  로 설정 Dr. Yunsung Ji (Saxo), a data scientist at Linkbricks, a company specializing in AI and big data analytics, trained the beomi/llama-2-koen-13b base model on 4 A100-40Gs on GCP for 4 hours of instructional training (2048 Tokens). Accelerate, Deepspeed Zero-3 libraries were used. www.linkbricks.com, www.linkbricks.vc",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "Quantization made by Richard Erkhov.\n\n[Github](https://github.com/RichardErkhov)\n\n[Discord](https://discord.gg/pvy7H8DZMG)\n\n[Request more models](https://github.com/RichardErkhov/quant_request)\n\n\nyunsung-llama-2-koen-13b-linkbricks-sft-basic-v1 - GGUF\n- Model creator: https://huggingface.co/Saxo/\n- Original model: https://huggingface.co/Saxo/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1/\n\n\n| Name | Quant method | Size |\n| ---- | ---- | ---- |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q2_K.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q2_K.gguf) | Q2_K | 4.6GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_XS.gguf) | IQ3_XS | 5.08GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_S.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_S.gguf) | IQ3_S | 5.36GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_S.gguf) | Q3_K_S | 5.36GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_M.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ3_M.gguf) | IQ3_M | 5.66GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K.gguf) | Q3_K | 5.99GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_M.gguf) | Q3_K_M | 5.99GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q3_K_L.gguf) | Q3_K_L | 6.54GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ4_XS.gguf) | IQ4_XS | 6.63GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_0.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_0.gguf) | Q4_0 | 6.95GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.IQ4_NL.gguf) | IQ4_NL | 7.0GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K_S.gguf) | Q4_K_S | 7.01GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K.gguf) | Q4_K | 7.42GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_K_M.gguf) | Q4_K_M | 7.42GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_1.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q4_1.gguf) | Q4_1 | 7.71GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_0.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_0.gguf) | Q5_0 | 8.46GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K_S.gguf) | Q5_K_S | 8.46GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K.gguf) | Q5_K | 8.7GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_K_M.gguf) | Q5_K_M | 8.7GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_1.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q5_1.gguf) | Q5_1 | 9.21GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q6_K.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q6_K.gguf) | Q6_K | 10.06GB |\n| [yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q8_0.gguf](https://huggingface.co/RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf/blob/main/yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1.Q8_0.gguf) | Q8_0 | 13.03GB |\n\n\n\n\nOriginal model description:\n---\nlibrary_name: transformers\nlicense: apache-2.0\nbasemodel: beomi/llama-2-koen-13b\ndatasets:\n- Saxo/total_ko_train_set_small_basic\n- beomi/KoAlpaca-v1.1a\n- kyujinpy/KOR-OpenOrca-Platypus-v2\n- nlpai-lab/databricks-dolly-15k-ko\nlanguage:\n- ko\n- en\npipeline_tag: text-generation\n---\n\n# Model Card for Model ID\n\n<div align=\"center\">\n<img src=\"https://www.linkbricks.com/wp-content/uploads/2022/03/%E1%84%85%E1%85%B5%E1%86%BC%E1%84%8F%E1%85%B3%E1%84%87%E1%85%B3%E1%84%85%E1%85%B5%E1%86%A8%E1%84%89%E1%85%B3%E1%84%85%E1%85%A9%E1%84%80%E1%85%A9-2-1024x804.png\" />\n</div>\n\n\nAI 와 빅데이터 분석 전문 기업인 Linkbricks의 데이터사이언티스트인 지윤성 박사(Saxo)가 beomi/llama-2-koen-13b 베이스모델을 GCP상의 A100-40G 4개를 통해 SFT 훈련을 한(2048 Tokens) 인스트럭션 모델.\n Accelerate, Deepspeed Zero-3 라이브러리를 사용했으며 Flash Attention 은 Disable  로 설정\n\nDr. Yunsung Ji (Saxo), a data scientist at Linkbricks, a company specializing in AI and big data analytics, trained the beomi/llama-2-koen-13b base model on 4 A100-40Gs on GCP for 4 hours of instructional training (2048 Tokens).\nAccelerate, Deepspeed Zero-3 libraries were used. \n\nwww.linkbricks.com, www.linkbricks.vc\n\n## Configuration including BitsandBytes\n---\nlearning_rate = 2e-4\nnum_epochs = 5 \nbatch_size = 4 \nblock_size = 2048 \ntrainer = \"sft\" \nwarmup_ratio = 0.1 \nweight_decay = 0.01 \ngradient_accumulation = 4 \nmixed_precision = \"fp16\" \npeft = True \nquantization = \"int4\" \nlora_r = 64 \nlora_alpha = 16 \nlora_dropout = 0.1 \nmodel_max_length = 2048 \n---\n## Dataset Format\nAlpaca Format Prompt Text\n\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 0,
  "downloads": 128,
  "gated": false,
  "private": false,
  "last_modified": "2024-08-10T04:20:18.000Z",
  "created_at": "2024-08-10T01:17:51.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "66b6bfbf503e23d3bd2ba959",
  "id": "RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf",
  "modelId": "RichardErkhov/Saxo_-_yunsung-llama-2-koen-13b-linkbricks-sft-basic-v1-gguf",
  "sha": "36926e8f3a5293e88daf5b089156b3e4be2b1ce0",
  "createdAt": "2024-08-10T01:17:51.000Z",
  "lastModified": "2024-08-10T04:20:18.000Z",
  "author": "RichardErkhov",
  "downloads": 128,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 24
}