Model Intelligence Sheet
richarderkhov/moemoekkung_-_frankenstein-moe-en-10.7bx6-gguf overview
Method To initialize the gate projection weight of the MoE layer, the H6 trainset was sampled and used. We sampled 400 and selected the final 30 with low PPL. trufulqa used gpt4 to generate data. ### Evals in progress
Downloads
116
Likes
0
Pipeline
—
Library
—
Visibility
Public
Access
Open
Repository Files & Downloads
24 files detected
Direct downloads for all repository files
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| Frankenstein-MoE-en-10.7Bx6.IQ3_M.gguf | GGUF | IQ3_M | 21.65 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.IQ3_S.gguf | GGUF | IQ3_S | 21.29 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.IQ3_XS.gguf | GGUF | IQ3_XS | 20.14 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.IQ4_NL.gguf | GGUF | IQ4_NL | 28.08 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.IQ4_XS.gguf | GGUF | IQ4_XS | 26.61 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q2_K.gguf | GGUF | Q2_K | 17.99 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q3_K.gguf | GGUF | Q3_K | 23.61 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q3_K_L.gguf | GGUF | Q3_K_L | 25.57 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q3_K_M.gguf | GGUF | Q3_K_M | 23.61 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q3_K_S.gguf | GGUF | Q3_K_S | 21.27 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q4_0.gguf | GGUF | — | 27.81 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q4_1.gguf | GGUF | — | 30.89 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q4_K.gguf | GGUF | Q4_K | 29.86 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q4_K_M.gguf | GGUF | Q4_K_M | 29.86 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q4_K_S.gguf | GGUF | Q4_K_S | 28.06 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q5_0.gguf | GGUF | — | 33.96 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q5_1.gguf | GGUF | — | 37.04 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q5_K.gguf | GGUF | Q5_K | 35.02 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q5_K_M.gguf | GGUF | Q5_K_M | 35.02 GB | Download |
| Frankenstein-MoE-en-10.7Bx6.Q5_K_S.gguf | GGUF | Q5_K_S | 33.96 GB | Download |
| Frankenstein-MoE-en-10.7Bx6_Q6_K-00001-of-00002.gguf | GGUF | Q6_K | 36.99 GB | Download |
| Frankenstein-MoE-en-10.7Bx6_Q6_K-00002-of-00002.gguf | GGUF | Q6_K | 3.52 GB | Download |
| Frankenstein-MoE-en-10.7Bx6_Q8_0-00001-of-00002.gguf | GGUF | — | 37.22 GB | Download |
| Frankenstein-MoE-en-10.7Bx6_Q8_0-00002-of-00002.gguf | GGUF | — | 15.25 GB | Download |
Model Details Live
Metadata Inspector
Normalized metadata (stored in metadata_json)
{
"metadata": {},
"card_data": {
"frontmatter": {},
"hero_image_url": "",
"summary": "### Method To initialize the gate projection weight of the MoE layer, the H6 trainset was sampled and used. We sampled 400 and selected the final 30 with low PPL. trufulqa used gpt4 to generate data. ### Evals in progress",
"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\nFrankenstein-MoE-en-10.7Bx6 - GGUF\n- Model creator: https://huggingface.co/MoEMoEKKung/\n- Original model: https://huggingface.co/MoEMoEKKung/Frankenstein-MoE-en-10.7Bx6/\n\n\n| Name | Quant method | Size |\n| ---- | ---- | ---- |\n| [Frankenstein-MoE-en-10.7Bx6.Q2_K.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q2_K.gguf) | Q2_K | 17.99GB |\n| [Frankenstein-MoE-en-10.7Bx6.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.IQ3_XS.gguf) | IQ3_XS | 20.14GB |\n| [Frankenstein-MoE-en-10.7Bx6.IQ3_S.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.IQ3_S.gguf) | IQ3_S | 21.29GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q3_K_S.gguf) | Q3_K_S | 21.27GB |\n| [Frankenstein-MoE-en-10.7Bx6.IQ3_M.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.IQ3_M.gguf) | IQ3_M | 21.65GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q3_K.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q3_K.gguf) | Q3_K | 23.61GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q3_K_M.gguf) | Q3_K_M | 23.61GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q3_K_L.gguf) | Q3_K_L | 25.57GB |\n| [Frankenstein-MoE-en-10.7Bx6.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.IQ4_XS.gguf) | IQ4_XS | 26.61GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q4_0.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q4_0.gguf) | Q4_0 | 27.81GB |\n| [Frankenstein-MoE-en-10.7Bx6.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.IQ4_NL.gguf) | IQ4_NL | 28.08GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q4_K_S.gguf) | Q4_K_S | 28.06GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q4_K.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q4_K.gguf) | Q4_K | 29.86GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q4_K_M.gguf) | Q4_K_M | 29.86GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q4_1.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q4_1.gguf) | Q4_1 | 30.89GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q5_0.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q5_0.gguf) | Q5_0 | 33.96GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q5_K_S.gguf) | Q5_K_S | 33.96GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q5_K.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q5_K.gguf) | Q5_K | 35.02GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q5_K_M.gguf) | Q5_K_M | 35.02GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q5_1.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/blob/main/Frankenstein-MoE-en-10.7Bx6.Q5_1.gguf) | Q5_1 | 37.04GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q6_K.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/tree/main/) | Q6_K | 40.5GB |\n| [Frankenstein-MoE-en-10.7Bx6.Q8_0.gguf](https://huggingface.co/RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf/tree/main/) | Q8_0 | 52.46GB |\n\n\n\n\nOriginal model description:\n---\nlanguage:\n- en\nlicense: cc-by-nc-sa-4.0\n---\n\n# Frankenstein-MoE\n\n### Method\nTo initialize the gate projection weight of the MoE layer, the H6 trainset was sampled and used. We sampled 400 and selected the final 30 with low PPL. \n\ntrufulqa used gpt4 to generate data.\n\n### Evals\nin progress\n\n",
"related_quantizations": []
},
"tags": [
"gguf",
"endpoints_compatible",
"region:us",
"conversational"
],
"likes": 0,
"downloads": 116,
"gated": false,
"private": false,
"last_modified": "2024-07-28T18:21:00.000Z",
"created_at": "2024-07-27T13:33:01.000Z",
"pipeline_tag": "",
"library_name": ""
}
Source payload excerpt (from Hugging Face API)
{
"_id": "66a4f70d0b1430428a6c6994",
"id": "RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf",
"modelId": "RichardErkhov/MoEMoEKKung_-_Frankenstein-MoE-en-10.7Bx6-gguf",
"sha": "c70e16bdb51dcb4957777855091715e2dbeb8fbb",
"createdAt": "2024-07-27T13:33:01.000Z",
"lastModified": "2024-07-28T18:21:00.000Z",
"author": "RichardErkhov",
"downloads": 116,
"likes": 0,
"gated": false,
"private": false,
"pipeline_tag": "",
"library_name": "",
"siblings_count": 26
}