richarderkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf overview

{
  "metadata": {},
  "card_data": {
    "frontmatter": {},
    "hero_image_url": "",
    "summary": "",
    "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\nllama_with_eeve_new_03_150m - GGUF\n- Model creator: https://huggingface.co/kikikara/\n- Original model: https://huggingface.co/kikikara/llama_with_eeve_new_03_150m/\n\n\n| Name | Quant method | Size |\n| ---- | ---- | ---- |\n| [llama_with_eeve_new_03_150m.Q2_K.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q2_K.gguf) | Q2_K | 0.06GB |\n| [llama_with_eeve_new_03_150m.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.IQ3_XS.gguf) | IQ3_XS | 0.07GB |\n| [llama_with_eeve_new_03_150m.IQ3_S.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.IQ3_S.gguf) | IQ3_S | 0.07GB |\n| [llama_with_eeve_new_03_150m.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q3_K_S.gguf) | Q3_K_S | 0.07GB |\n| [llama_with_eeve_new_03_150m.IQ3_M.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.IQ3_M.gguf) | IQ3_M | 0.07GB |\n| [llama_with_eeve_new_03_150m.Q3_K.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q3_K.gguf) | Q3_K | 0.08GB |\n| [llama_with_eeve_new_03_150m.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q3_K_M.gguf) | Q3_K_M | 0.08GB |\n| [llama_with_eeve_new_03_150m.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q3_K_L.gguf) | Q3_K_L | 0.08GB |\n| [llama_with_eeve_new_03_150m.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.IQ4_XS.gguf) | IQ4_XS | 0.08GB |\n| [llama_with_eeve_new_03_150m.Q4_0.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q4_0.gguf) | Q4_0 | 0.09GB |\n| [llama_with_eeve_new_03_150m.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.IQ4_NL.gguf) | IQ4_NL | 0.09GB |\n| [llama_with_eeve_new_03_150m.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q4_K_S.gguf) | Q4_K_S | 0.09GB |\n| [llama_with_eeve_new_03_150m.Q4_K.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q4_K.gguf) | Q4_K | 0.09GB |\n| [llama_with_eeve_new_03_150m.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q4_K_M.gguf) | Q4_K_M | 0.09GB |\n| [llama_with_eeve_new_03_150m.Q4_1.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q4_1.gguf) | Q4_1 | 0.09GB |\n| [llama_with_eeve_new_03_150m.Q5_0.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q5_0.gguf) | Q5_0 | 0.1GB |\n| [llama_with_eeve_new_03_150m.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q5_K_S.gguf) | Q5_K_S | 0.1GB |\n| [llama_with_eeve_new_03_150m.Q5_K.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q5_K.gguf) | Q5_K | 0.1GB |\n| [llama_with_eeve_new_03_150m.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q5_K_M.gguf) | Q5_K_M | 0.1GB |\n| [llama_with_eeve_new_03_150m.Q5_1.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q5_1.gguf) | Q5_1 | 0.11GB |\n| [llama_with_eeve_new_03_150m.Q6_K.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q6_K.gguf) | Q6_K | 0.12GB |\n| [llama_with_eeve_new_03_150m.Q8_0.gguf](https://huggingface.co/RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf/blob/main/llama_with_eeve_new_03_150m.Q8_0.gguf) | Q8_0 | 0.15GB |\n\n\n\n\nOriginal model description:\n---\ndatasets:\n- maywell/ko_wikidata_QA\n- nlpai-lab/kullm-v2\n- heegyu/kowikitext\n- MarkrAI/KoCommercial-Dataset\n- heegyu/CoT-collection-ko\n- HAERAE-HUB/Korean-Human-Judgements\n- instructkr/ko_elo_arena_0207\n- HAERAE-HUB/K2-Feedback\n- heegyu/open-korean-instructions\n- heegyu/aulm-0809\nlanguage:\n- ko\n---\n\n# llama_with_eeve_new_03_150m\n\n\n## Model Info\n llama 아키텍처와 eeve 토크나이저를 사용해 랜덤 가중치에서 시작해 사전학습된 모델입니다\n\n\n다음 시스템 프롬프트가 주어진 상태로 학습하였습니다(모델 사용 시 프롬프트를 포함해야 합니다).\n\n'''### System:\\n당신은 비도덕적이거나, 성적이거나, 불법적이거나 또는 사회 통념적으로 허용되지 않는 발언은 하지 않습니다.\n사용자와 즐겁게 대화하며, 사용자의 응답에 가능한 정확하고 친절하게 응답함으로써 최대한 도와주려고 노력합니다.\n\n\\n\\n### User:\\n {question}'''\n\n\n### Evaluation results\n\nllm as a judge 방식으로 평가를 진행했습니다.\n자세한 내용은 \" \"를 참고해주세요\n\n| Model                                                                                                   | params | Fluency | Coherence | Accuracy | Completeness |\n|---------------------------------------------------------------------------------------------------------|--------|---------|-----------|----------|--------------|\n| **[kikikara/llama_with_eeve_new_03_150m](https://huggingface.co/kikikara/llama_with_eeve_new_03_150m)(this)** | **0.15B**  | **63.12%**  |   **37.18%**  |  **23.75%**  |    **23.75%**  |\n| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b)                       | 1.3B   | 51.25%  |   40.31%  |  34.68%  |    32.5%     |\n| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b)                       | 5.8B   | 54.37%  |   40.62%  |  41.25%  |     35%      |\n\n\n\n### How to use\n\n```python\nfrom transformers import AutoTokenizer, AutoModelForCausalLM, pipeline\n\ntokenizer = AutoTokenizer.from_pretrained(\"kikikara/llama_with_eeve_new_03_150m\")\nmodel = AutoModelForCausalLM.from_pretrained(\"kikikara/llama_with_eeve_new_03_150m\")\n\nquestion = \"너는 누구야?\"\n\nprompt = f\"### System:\\n당신은 비도덕적이거나, 성적이거나, 불법적이거나 또는 사회 통념적으로 허용되지 않는 발언은 하지 않습니다.\\n사용자와 즐겁게 대화하며, 사용자의 응답에 가능한 정확하고 친절하게 응답함으로써 최대한 도와주려고 노력합니다.\\n\\n\\n### User:\\n {question}\"\npipe = pipeline(task=\"text-generation\", model=model, tokenizer=tokenizer, max_length=400, repetition_penalty=1.12)\nresult = pipe(prompt)\n\nprint(result[0]['generated_text'])```\n\n\n\n\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 0,
  "downloads": 243,
  "gated": false,
  "private": false,
  "last_modified": "2024-07-05T05:04:39.000Z",
  "created_at": "2024-07-05T05:02:37.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "66877e6dabca3bbcec629924",
  "id": "RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf",
  "modelId": "RichardErkhov/kikikara_-_llama_with_eeve_new_03_150m-gguf",
  "sha": "ce6568bc1505f73e05d6481a0221d3445b87c0dd",
  "createdAt": "2024-07-05T05:02:37.000Z",
  "lastModified": "2024-07-05T05:04:39.000Z",
  "author": "RichardErkhov",
  "downloads": 243,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 24
}