duyntnet/mn-12b-mag-mell-r1-imatrix-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "other",
    "language": [
      "en"
    ],
    "pipeline_tag": "text-generation",
    "inference": false,
    "tags": [
      "transformers",
      "gguf",
      "imatrix",
      "MN-12B-Mag-Mell-R1"
    ],
    "frontmatter": {
      "license": "other",
      "language": [
        "en"
      ],
      "pipeline_tag": "text-generation",
      "inference": "false",
      "tags": [
        "transformers",
        "gguf",
        "imatrix",
        "MN-12B-Mag-Mell-R1"
      ]
    },
    "hero_image_url": "",
    "summary": "NOTE for newer users: \"R1\" here means \"Revision 1\". This model predates DeepSeek's R1; DeepSeek inadvertently made using this versioning scheme very annoying! This is a merge of pre-trained language models created using mergekit.",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: other\nlanguage:\n- en\npipeline_tag: text-generation\ninference: false\ntags:\n- transformers\n- gguf\n- imatrix\n- MN-12B-Mag-Mell-R1\n---\nQuantizations of https://huggingface.co/inflatebot/MN-12B-Mag-Mell-R1\n\n\n### Open source inference clients/UIs\n* [llama.cpp](https://github.com/ggerganov/llama.cpp)\n* [KoboldCPP](https://github.com/LostRuins/koboldcpp)\n* [text-generation-webui](https://github.com/oobabooga/text-generation-webui)\n* [ollama](https://github.com/ollama/ollama)\n* [jan](https://github.com/janhq/jan)\n* [GPT4All](https://github.com/nomic-ai/gpt4all)\n\n### Closed source inference clients/UIs\n* [LM Studio](https://lmstudio.ai/)\n* [Backyard AI](https://backyard.ai/)\n* More will be added...\n---\n\n# From original readme\n\nNOTE for newer users: \"R1\" here means \"Revision 1\". This model predates DeepSeek's R1; DeepSeek inadvertently made using this versioning scheme very annoying!\n\nThis is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).\n\n## Usage Details\n\n### Sampler Settings\nMag Mell R1 was tested with Temp 1.25 and MinP 0.2. This was fairly stable up to 10K, but this might be too \"hot\".\nIf issues with coherency occur, try *in*creasing MinP or *de*creasing Temperature.\n\nOther samplers shouldn't be necessary. XTC was shown to break outputs. DRY should be okay if used sparingly. Other penalty-type samplers should probably be avoided.\n\n### Formatting\nThe base model for Mag Mell is [Mistral-Nemo-Base-2407-chatml](https://huggingface.co/IntervitensInc/Mistral-Nemo-Base-2407-chatml), and as such ChatML formatting is recommended.\n\nEarly testing versions had a tendency to leak tokens, but this should be more or less hammered out. It recently (12-18-2024) came to attention that Cache Quantization may either cause or exacerbate this issue.\n\n## Merge Details\nMag Mell is a multi-stage merge, Inspired by hyper-merges like [Tiefighter](https://huggingface.co/KoboldAI/LLaMA2-13B-Tiefighter) and [Umbral Mind.](https://huggingface.co/Casual-Autopsy/L3-Umbral-Mind-RP-v2.0-8B)\nIntended to be a general purpose \"Best of Nemo\" model for any fictional, creative use case.\n\n6 models were chosen based on 3 categories; they were then paired up and merged via layer-weighted SLERP to create intermediate \"specialists\" which are then evaluated in their domain.\nThe specialists were then merged into the base via DARE-TIES, with hyperparameters chosen to reduce interference caused by the overlap of the three domains. \nThe idea with this approach is to extract the best qualities of each component part, and produce models whose task vectors represent more than the sum of their parts.\n\nThe three specialists are as follows:\n\n- Hero (RP, kink/trope coverage): [Chronos Gold](https://huggingface.co/elinas/Chronos-Gold-12B-1.0), [Sunrose](https://huggingface.co/Fizzarolli/MN-12b-Sunrose).\n\n- Monk (Intelligence, groundedness): [Bophades](https://huggingface.co/nbeerbower/mistral-nemo-bophades-12B), [Wissenschaft](https://huggingface.co/nbeerbower/mistral-nemo-wissenschaft-12B).\n\n- Deity (Prose, flair): [Gutenberg v4](https://huggingface.co/nbeerbower/mistral-nemo-gutenberg-12B-v4), [Magnum 2.5 KTO](https://huggingface.co/anthracite-org/magnum-v2.5-12b-kto).\n\nI've been dreaming about this merge since Nemo tunes started coming out in earnest. From our testing, Mag Mell demonstrates worldbuilding capabilities unlike any model in its class, comparable to old adventuring models like Tiefighter, and prose that exhibits minimal \"slop\" (not bad for no finetuning,) frequently devising electrifying metaphors that left us consistently astonished.\n\nI don't want to toot my own bugle though; I'm really proud of how this came out, but please leave your feedback, good or bad.\n\nSpecial thanks as usual to Toaster for his feedback and Fizz for helping fund compute, as well as the KoboldAI Discord for their resources.\n\n### Merge Method\n\nThis model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using [IntervitensInc/Mistral-Nemo-Base-2407-chatml](https://huggingface.co/IntervitensInc/Mistral-Nemo-Base-2407-chatml) as a base.\n\n### Models Merged\n\nThe following models were included in the merge:\n* IntervitensInc/Mistral-Nemo-Base-2407-chatml\n* nbeerbower/mistral-nemo-bophades-12B\n* nbeerbower/mistral-nemo-wissenschaft-12B\n* elinas/Chronos-Gold-12B-1.0\n* Fizzarolli/MN-12b-Sunrose\n* nbeerbower/mistral-nemo-gutenberg-12B-v4\n* anthracite-org/magnum-12b-v2.5-kto\n\n### Configuration\n\nThe following YAML configurations were used to produce this model:\n\n#### Monk:\n```yaml\nmodels:\n  - model: nbeerbower/mistral-nemo-bophades-12B\n  - model: nbeerbower/mistral-nemo-wissenschaft-12B\nmerge_method: slerp\nbase_model: nbeerbower/mistral-nemo-bophades-12B\nparameters:\n  t: [0.1, 0.2, 0.4, 0.6, 0.6, 0.4, 0.2, 0.1]\ndtype: bfloat16\ntokenizer_source: base\n```\n\n#### Hero:\n```yaml\nmodels:\n  - model: elinas/Chronos-Gold-12B-1.0\n  - model: Fizzarolli/MN-12b-Sunrose\nmerge_method: slerp\nbase_model: elinas/Chronos-Gold-12B-1.0\nparameters:\n  t: [0.1, 0.2, 0.4, 0.6, 0.6, 0.4, 0.2, 0.1]\ndtype: bfloat16\ntokenizer_source: base\n```\n\n#### Deity:\n```yaml\nmodels:\n  - model: nbeerbower/mistral-nemo-gutenberg-12B-v4\n  - model: anthracite-org/magnum-12b-v2.5-kto\nmerge_method: slerp\nbase_model: nbeerbower/mistral-nemo-gutenberg-12B-v4\nparameters:\n  t: [0, 0.1, 0.2, 0.25, 0.25, 0.2, 0.1, 0]\ndtype: bfloat16\ntokenizer_source: base\n```\n\n#### Mag Mell:\n```yaml\nmodels:\n - model: monk\n   parameters:\n     density: 0.7\n     weight: 0.5\n - model: hero\n   parameters:\n     density: 0.9\n     weight: 1\n - model: deity\n   parameters:\n     density: 0.5\n     weight: 0.7\nmerge_method: dare_ties\nbase_model: IntervitensInc/Mistral-Nemo-Base-2407-chatml\ntokenizer_source: base\n```\n\n\n`In Irish mythology, Mag Mell (modern spelling: Magh Meall, meaning 'delightful plain') is one of the names for the Celtic Otherworld, a mythical realm achievable through death and/or glory... Never explicitly stated in any surviving mythological account to be an afterlife; rather, it is usually portrayed as a paradise populated by deities, which is occasionally visited by some adventurous mortals. In its island guise, it was visited by various legendary Irish heroes and monks, forming the basis of the adventure myth or echtrae...`",
    "related_quantizations": []
  },
  "tags": [
    "transformers",
    "gguf",
    "imatrix",
    "MN-12B-Mag-Mell-R1",
    "text-generation",
    "en",
    "arxiv:2311.03099",
    "arxiv:2306.01708",
    "license:other",
    "region:us",
    "conversational"
  ],
  "likes": 0,
  "downloads": 236,
  "gated": false,
  "private": false,
  "last_modified": "2025-04-23T12:55:23.000Z",
  "created_at": "2025-04-23T11:27:45.000Z",
  "pipeline_tag": "text-generation",
  "library_name": "transformers"
}

{
  "_id": "6808ceb1ccabba3e07582a7d",
  "id": "duyntnet/MN-12B-Mag-Mell-R1-imatrix-GGUF",
  "modelId": "duyntnet/MN-12B-Mag-Mell-R1-imatrix-GGUF",
  "sha": "9ef62b6bec4d0d55062b3d74af3dd86c0246c351",
  "createdAt": "2025-04-23T11:27:45.000Z",
  "lastModified": "2025-04-23T12:55:23.000Z",
  "author": "duyntnet",
  "downloads": 236,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "text-generation",
  "library_name": "transformers",
  "siblings_count": 29
}