richarderkhov/mala-lm_-_lucky52-bloom-7b1-no-2-gguf overview

{
  "metadata": {},
  "card_data": {
    "frontmatter": {},
    "hero_image_url": "",
    "summary": "Quantization made by Richard Erkhov. Github Discord Request more models lucky52-bloom-7b1-no-2 - GGUF | Name | Quant method | Size | | ---- | ---- | ---- | | lucky52-bloom-7b1-no-2.Q2_K.gguf | Q2_K | 3.2GB | | lucky52-bloom-7b1-no-2.IQ3_XS.gguf | IQ3_XS | 3.56GB | | lucky52-bloom-7b1-no-2.IQ3_S.gguf | IQ3_S | 3.63GB | | lucky52-bloom-7b1-no-2.Q3_K_S.gguf | Q3_K_S | 3.63GB | | lucky52-bloom-7b1-no-2.IQ3_M.gguf | IQ3_M | 3.9GB | | lucky52-bloom-7b1-no-2.Q3_K.gguf | Q3_K | 4.14GB | | lucky52-bloom-7b1-no-2.Q3_K_M.gguf | Q3_K_M | 4.14GB | | lucky52-bloom-7b1-no-2.Q3_K_L.gguf | Q3_K_L | 4.42GB | | lucky52-bloom-7b1-no-2.IQ4_XS.gguf | IQ4_XS | 4.33GB | | lucky52-bloom-7b1-no-2.Q4_0.gguf | Q4_0 | 4.51GB | | lucky52-bloom-7b1-no-2.IQ4_NL.gguf | IQ4_NL | 4.53GB | | lucky52-bloom-7b1-no-2.Q4_K_S.gguf | Q4_K_S | 4.53GB | | lucky52-bloom-7b1-no-2.Q4_K.gguf | Q4_K | 4.91GB | | lucky52-bloom-7b1-no-2.Q4_K_M.gguf | Q4_K_M | 4.91GB | | lucky52-bloom-7b1-no-2.Q4_1.gguf | Q4_1 | 4.92GB | | lucky52-bloom-7b1-no-2.Q5_0.gguf | Q5_0 | 5.33GB | | lucky52-bloom-7b1-no-2.Q5_K_S.gguf | Q5_K_S | 5.33GB | | lucky52-bloom-7b1-no-2.Q5_K.gguf | Q5_K | 5.63GB | | lucky52-bloom-7b1-no-2.Q5_K_M.gguf | Q5_K_M | 5.63GB | | lucky52-bloom-7b1-no-2.Q5_1.gguf | Q5_1 | 5.74GB | | lucky52-bloom-7b1-no-2.Q6_K.gguf | Q6_K | 6.2GB | | lucky52-bloom-7b1-no-2.Q8_0.gguf | Q8_0 | 8.03GB | Original model description: --- library_name: transformers pipeline_tag: text-generation language: tags: datasets: license: cc-by-nc-4.0 --- ###  Model Description This HF repository hosts instruction fine-tuned multilingual BLOOM model using the parallel instruction dataset called Bactrain-X in 52 languages. We progressively add a language during instruction fine-tuning at each time, and train 52 models in total. Then, we evaluate those models in three multilingual benchmarks. Please refer to our paper for more details. * Base model: BLOOM 7B1 * Instruction languages: English, Chinese * Instruction language codes: en, zh * Training method: full-parameter fine-tuning. ### Usage The model checkpoint should be loaded using transformers library. ``python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained(\"MaLA-LM/lucky52-bloom-7b1-no-2\") model = AutoModelForCausalLM.from_pretrained(\"MaLA-LM/lucky52-bloom-7b1-no-2\") ` ### Citation ` @inproceedings{ji2025lucky52, title={How Many Languages Make Good Multilingual Instruction Tuning? A Case Study on BLOOM}, author={Shaoxiong Ji and Pinzhen Chen}, year={2025}, booktitle={Proceedings of COLING}, url={https://arxiv.org/abs/2404.04850}, } ``",
    "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\nlucky52-bloom-7b1-no-2 - GGUF\n- Model creator: https://huggingface.co/MaLA-LM/\n- Original model: https://huggingface.co/MaLA-LM/lucky52-bloom-7b1-no-2/\n\n\n| Name | Quant method | Size |\n| ---- | ---- | ---- |\n| [lucky52-bloom-7b1-no-2.Q2_K.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q2_K.gguf) | Q2_K | 3.2GB |\n| [lucky52-bloom-7b1-no-2.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.IQ3_XS.gguf) | IQ3_XS | 3.56GB |\n| [lucky52-bloom-7b1-no-2.IQ3_S.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.IQ3_S.gguf) | IQ3_S | 3.63GB |\n| [lucky52-bloom-7b1-no-2.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q3_K_S.gguf) | Q3_K_S | 3.63GB |\n| [lucky52-bloom-7b1-no-2.IQ3_M.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.IQ3_M.gguf) | IQ3_M | 3.9GB |\n| [lucky52-bloom-7b1-no-2.Q3_K.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q3_K.gguf) | Q3_K | 4.14GB |\n| [lucky52-bloom-7b1-no-2.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q3_K_M.gguf) | Q3_K_M | 4.14GB |\n| [lucky52-bloom-7b1-no-2.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q3_K_L.gguf) | Q3_K_L | 4.42GB |\n| [lucky52-bloom-7b1-no-2.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.IQ4_XS.gguf) | IQ4_XS | 4.33GB |\n| [lucky52-bloom-7b1-no-2.Q4_0.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q4_0.gguf) | Q4_0 | 4.51GB |\n| [lucky52-bloom-7b1-no-2.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.IQ4_NL.gguf) | IQ4_NL | 4.53GB |\n| [lucky52-bloom-7b1-no-2.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q4_K_S.gguf) | Q4_K_S | 4.53GB |\n| [lucky52-bloom-7b1-no-2.Q4_K.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q4_K.gguf) | Q4_K | 4.91GB |\n| [lucky52-bloom-7b1-no-2.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q4_K_M.gguf) | Q4_K_M | 4.91GB |\n| [lucky52-bloom-7b1-no-2.Q4_1.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q4_1.gguf) | Q4_1 | 4.92GB |\n| [lucky52-bloom-7b1-no-2.Q5_0.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q5_0.gguf) | Q5_0 | 5.33GB |\n| [lucky52-bloom-7b1-no-2.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q5_K_S.gguf) | Q5_K_S | 5.33GB |\n| [lucky52-bloom-7b1-no-2.Q5_K.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q5_K.gguf) | Q5_K | 5.63GB |\n| [lucky52-bloom-7b1-no-2.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q5_K_M.gguf) | Q5_K_M | 5.63GB |\n| [lucky52-bloom-7b1-no-2.Q5_1.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q5_1.gguf) | Q5_1 | 5.74GB |\n| [lucky52-bloom-7b1-no-2.Q6_K.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q6_K.gguf) | Q6_K | 6.2GB |\n| [lucky52-bloom-7b1-no-2.Q8_0.gguf](https://huggingface.co/RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf/blob/main/lucky52-bloom-7b1-no-2.Q8_0.gguf) | Q8_0 | 8.03GB |\n\n\n\n\nOriginal model description:\n\n---\nlibrary_name: transformers\npipeline_tag: text-generation\nlanguage:\n- multilingual\ntags:\n- generation\n- question answering\n- instruction tuning\ndatasets:\n- MBZUAI/Bactrian-X\nlicense: cc-by-nc-4.0\n---\n\n###  Model Description\n\nThis HF repository hosts instruction fine-tuned multilingual BLOOM model using the parallel instruction dataset called Bactrain-X in 52 languages. \nWe progressively add a language during instruction fine-tuning at each time, and train 52 models in total. Then, we evaluate those models in three multilingual benchmarks. \n\nPlease refer to [our paper](https://arxiv.org/abs/2404.04850) for more details.\n\n* Base model: [BLOOM 7B1](https://huggingface.co/bigscience/bloom-7b1)\n* Instruction languages: English, Chinese\n* Instruction language codes: en, zh\n* Training method: full-parameter fine-tuning.\n\n### Usage\nThe model checkpoint should be loaded using `transformers` library.\n\n```python\nfrom transformers import AutoTokenizer, AutoModelForCausalLM\n\ntokenizer = AutoTokenizer.from_pretrained(\"MaLA-LM/lucky52-bloom-7b1-no-2\")\nmodel = AutoModelForCausalLM.from_pretrained(\"MaLA-LM/lucky52-bloom-7b1-no-2\")\n```\n\n### Citation\n```\n@inproceedings{ji2025lucky52,\n      title={How Many Languages Make Good Multilingual Instruction Tuning? A Case Study on BLOOM}, \n      author={Shaoxiong Ji and Pinzhen Chen},\n      year={2025},\n      booktitle={Proceedings of COLING},\n      url={https://arxiv.org/abs/2404.04850}, \n}\n```\n\n\n\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "arxiv:2404.04850",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 0,
  "downloads": 108,
  "gated": false,
  "private": false,
  "last_modified": "2025-07-26T06:08:06.000Z",
  "created_at": "2025-07-26T05:05:37.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "6884622125c291a5bc73a303",
  "id": "RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf",
  "modelId": "RichardErkhov/MaLA-LM_-_lucky52-bloom-7b1-no-2-gguf",
  "sha": "d6e8fb88c6dd6632e0cf94700e62be9205b77cf1",
  "createdAt": "2025-07-26T05:05:37.000Z",
  "lastModified": "2025-07-26T06:08:06.000Z",
  "author": "RichardErkhov",
  "downloads": 108,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 24
}