richarderkhov/sherlockassistant_-_mistral-7b-instruct-ukrainian-gguf overview

{
  "metadata": {},
  "card_data": {
    "frontmatter": {},
    "hero_image_url": "",
    "summary": "Mistral-7B-UK is a Large Language Model finetuned for the Ukrainian language. Mistral-7B-UK is trained using the following formula: 1. Initial finetuning of Mistral-7B-v0.2 using structured and unstructured datasets. 2. SLERP merge of the finetuned model with a model that performs better than Mistral-7B-v0.2 on OpenLLM benchmark: NeuralTrix-7B 3. DPO of the final model.",
    "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\nMistral-7B-Instruct-Ukrainian - GGUF\n- Model creator: https://huggingface.co/SherlockAssistant/\n- Original model: https://huggingface.co/SherlockAssistant/Mistral-7B-Instruct-Ukrainian/\n\n\n| Name | Quant method | Size |\n| ---- | ---- | ---- |\n| [Mistral-7B-Instruct-Ukrainian.Q2_K.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q2_K.gguf) | Q2_K | 2.53GB |\n| [Mistral-7B-Instruct-Ukrainian.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.IQ3_XS.gguf) | IQ3_XS | 2.81GB |\n| [Mistral-7B-Instruct-Ukrainian.IQ3_S.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.IQ3_S.gguf) | IQ3_S | 2.96GB |\n| [Mistral-7B-Instruct-Ukrainian.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q3_K_S.gguf) | Q3_K_S | 2.95GB |\n| [Mistral-7B-Instruct-Ukrainian.IQ3_M.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.IQ3_M.gguf) | IQ3_M | 3.06GB |\n| [Mistral-7B-Instruct-Ukrainian.Q3_K.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q3_K.gguf) | Q3_K | 3.28GB |\n| [Mistral-7B-Instruct-Ukrainian.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q3_K_M.gguf) | Q3_K_M | 3.28GB |\n| [Mistral-7B-Instruct-Ukrainian.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q3_K_L.gguf) | Q3_K_L | 3.56GB |\n| [Mistral-7B-Instruct-Ukrainian.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.IQ4_XS.gguf) | IQ4_XS | 3.67GB |\n| [Mistral-7B-Instruct-Ukrainian.Q4_0.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q4_0.gguf) | Q4_0 | 3.83GB |\n| [Mistral-7B-Instruct-Ukrainian.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.IQ4_NL.gguf) | IQ4_NL | 3.87GB |\n| [Mistral-7B-Instruct-Ukrainian.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q4_K_S.gguf) | Q4_K_S | 3.86GB |\n| [Mistral-7B-Instruct-Ukrainian.Q4_K.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q4_K.gguf) | Q4_K | 4.07GB |\n| [Mistral-7B-Instruct-Ukrainian.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q4_K_M.gguf) | Q4_K_M | 4.07GB |\n| [Mistral-7B-Instruct-Ukrainian.Q4_1.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q4_1.gguf) | Q4_1 | 4.24GB |\n| [Mistral-7B-Instruct-Ukrainian.Q5_0.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q5_0.gguf) | Q5_0 | 4.65GB |\n| [Mistral-7B-Instruct-Ukrainian.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q5_K_S.gguf) | Q5_K_S | 4.65GB |\n| [Mistral-7B-Instruct-Ukrainian.Q5_K.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q5_K.gguf) | Q5_K | 4.78GB |\n| [Mistral-7B-Instruct-Ukrainian.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q5_K_M.gguf) | Q5_K_M | 4.78GB |\n| [Mistral-7B-Instruct-Ukrainian.Q5_1.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q5_1.gguf) | Q5_1 | 5.07GB |\n| [Mistral-7B-Instruct-Ukrainian.Q6_K.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q6_K.gguf) | Q6_K | 5.53GB |\n| [Mistral-7B-Instruct-Ukrainian.Q8_0.gguf](https://huggingface.co/RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf/blob/main/Mistral-7B-Instruct-Ukrainian.Q8_0.gguf) | Q8_0 | 7.17GB |\n\n\n\n\nOriginal model description:\n---\nlibrary_name: transformers\nlicense: apache-2.0\n---\n\n# Model card for Mistral-7B-Instruct-Ukrainian\n\nMistral-7B-UK is a Large Language Model finetuned for the Ukrainian language.\n\nMistral-7B-UK is trained using the following formula:\n1. Initial finetuning of [Mistral-7B-v0.2](mistralai/Mistral-7B-Instruct-v0.2) using structured and unstructured datasets.\n2. SLERP merge of the finetuned model with a model that performs better than `Mistral-7B-v0.2` on `OpenLLM` benchmark: [NeuralTrix-7B](https://huggingface.co/CultriX/NeuralTrix-7B-v1)\n3. DPO of the final model.\n\n\n\n\n\n## Instruction format\n\nIn order to leverage instruction fine-tuning, your prompt should be surrounded by `[INST]` and `[/INST]` tokens.\n\nE.g.\n```\ntext = \"[INST]Відповідайте лише буквою правильної відповіді: Елементи експресіонізму наявні у творі: A. «Камінний хрест», B. «Інститутка», C. «Маруся», D. «Людина»[/INST]\"\n```\n\nThis format is available as a [chat template](https://huggingface.co/docs/transformers/main/chat_templating) via the `apply_chat_template()` method:\n\n## Model Architecture\nThis instruction model is based on Mistral-7B-v0.2, a transformer model with the following architecture choices:\n- Grouped-Query Attention\n- Sliding-Window Attention\n- Byte-fallback BPE tokenizer\n\n## Datasets - Structured\n- [UA-SQUAD](https://huggingface.co/datasets/FIdo-AI/ua-squad/resolve/main/ua_squad_dataset.json)\n- [Ukrainian StackExchange](https://huggingface.co/datasets/zeusfsx/ukrainian-stackexchange)\n- [UAlpaca Dataset](https://github.com/robinhad/kruk/blob/main/data/cc-by-nc/alpaca_data_translated.json)\n- [Ukrainian Subset from Belebele Dataset](https://github.com/facebookresearch/belebele)\n- [Ukrainian Subset from XQA](https://github.com/thunlp/XQA)\n- [ZNO Dataset provided in UNLP 2024 shared task](https://github.com/unlp-workshop/unlp-2024-shared-task/blob/main/data/zno.train.jsonl)\n\n## Datasets - Unstructured\n- Ukrainian Wiki\n\n## Datasets - DPO\n- Ukrainian translation of [distilabel-indel-orca-dpo-pairs](https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs)\n  \n## 💻 Usage\n\n```python\n!pip install -qU transformers accelerate\n\nfrom transformers import AutoTokenizer\nimport transformers\nimport torch\n\nmodel = \"SherlockAssistant/Mistral-7B-Instruct-Ukrainian\"\nmessages = [{\"role\": \"user\", \"content\": \"What is a large language model?\"}]\n\ntokenizer = AutoTokenizer.from_pretrained(model)\nprompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)\npipeline = transformers.pipeline(\n    \"text-generation\",\n    model=model,\n    torch_dtype=torch.bfloat16,\n    device_map=\"auto\",\n)\n\noutputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)\nprint(outputs[0][\"generated_text\"])\n```\n\n## Citation\n\nIf you are using this model in your research and publishing a paper, please help by citing our paper:\n\n**BIB**\n\n```bib\n@inproceedings{boros-chivereanu-dumitrescu-purcaru-2024-llm-uk,\n    title = \"Fine-tuning and Retrieval Augmented Generation for Question Answering using affordable Large Language Models\",\n    author = \"Boros, Tiberiu and Chivereanu, Radu and Dumitrescu, Stefan Daniel and Purcaru, Octavian\",\n    booktitle = \"Proceedings of the Third Ukrainian Natural Language Processing Workshop, LREC-COLING\",\n    month = may,\n    year = \"2024\",\n    address = \"Torino, Italy\",\n    publisher = \"European Language Resources Association\",\n}\n```\n\n**APA**\n\nBoros, T., Chivereanu, R., Dumitrescu, S., & Purcaru, O. (2024). Fine-tuning and Retrieval Augmented Generation for Question Answering using affordable Large Language Models. In Proceedings of the Third Ukrainian Natural Language Processing Workshop, LREC-COLING. European Language Resources Association.\n\n**MLA**\n\nBoros, Tiberiu, Radu, Chivereanu, Stefan Daniel, Dumitrescu, Octavian, Purcaru. \"Fine-tuning and Retrieval Augmented Generation for Question Answering using affordable Large Language Models.\" Proceedings of the Third Ukrainian Natural Language Processing Workshop, LREC-COLING. European Language Resources Association, 2024.\n\n**Chicago**\n\nBoros, Tiberiu, Radu, Chivereanu, Stefan Daniel, Dumitrescu, and Octavian, Purcaru. \"Fine-tuning and Retrieval Augmented Generation for Question Answering using affordable Large Language Models.\" . In Proceedings of the Third Ukrainian Natural Language Processing Workshop, LREC-COLING. European Language Resources Association, 2024.\n\n\n\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 2,
  "downloads": 149,
  "gated": false,
  "private": false,
  "last_modified": "2024-09-09T09:55:17.000Z",
  "created_at": "2024-09-09T04:43:24.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "66de7cecdd44163a31da0303",
  "id": "RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf",
  "modelId": "RichardErkhov/SherlockAssistant_-_Mistral-7B-Instruct-Ukrainian-gguf",
  "sha": "bfa1f793a4076d7d2988b6cf294651eb82275fe9",
  "createdAt": "2024-09-09T04:43:24.000Z",
  "lastModified": "2024-09-09T09:55:17.000Z",
  "author": "RichardErkhov",
  "downloads": 149,
  "likes": 2,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 24
}