ludis/tsukasa-llama-3-70b-qlora-gguf overview

{
  "metadata": {},
  "card_data": {
    "datasets": [
      "PygmalionAI/PIPPA",
      "lemonilia/LimaRP"
    ],
    "frontmatter": {
      "datasets": [
        "PygmalionAI/PIPPA",
        "lemonilia/LimaRP"
      ]
    },
    "hero_image_url": "",
    "summary": "big thanks to lore for the 8xH100 gpus",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\ndatasets:\n  - PygmalionAI/PIPPA\n  - lemonilia/LimaRP\n---\n\nbig thanks to lore for the 8xH100 gpus\n\n## gguf\n\nlittle endian\n\n## training\n\nbase model is meta llama 3 8b instruct\ntrained on pippa then i trained that model on limarp, both at 8k context for 2 epochs each\n\n## gen settings\n\ni would **start with** every sampler off and **temperature at 1 and just make min p 0.05**, i got good prompts from this but u can also try to gen settings from shori which are copy pasted below\n\n- **Main choice** (may have repetition issues)\n   - **Temperature**: 1.0; **Min-P**: 0.05-0.10; **Presence Penalty**: 0.35-0.45 \n- **Alternative 1** (appears to solve repetition issues while being coherent, but reponses might possibly be less truthful)\n   - **Temperature**: 2.40-2.50; **Min-P**: 0.40; **Frequency penalty**: 0.10-0.15; Temperature last.\n- **Alternative 2**\n   - **Mirostat type**: 2, **Mirostat Tau**: 2.80-3.00; **Mirostat Eta**: 0.0175-0.0200; neutralize or disable all other samplers\n\n## prompting\n\nuse the llama 3 instruct format\n\n`<|eot_id|>` as stopping sequence/string/token\n\nST jsons:\n[instruct](https://files.catbox.moe/ocnjb7.json)\n[context](https://files.catbox.moe/hjkawf.json)\n\nagnaistic prompt:\n```\n<|begin_of_text|><|start_header_id|>system<|end_header_id|>{{#if system}}<|begin_of_text|><|start_header_id|>system<|end_header_id|>{{system}}<|eot_id|>{{/if}}Write {{char}}'s next reply in a fictional roleplay chat between {{#each bot}}{{.name}}, {{/each}}{{char}} and {{user}}.\n\n{{char}}'s Persona: {{personality}}\n\n{{#if memory}}\nImportant details:\n{{memory}}\n{{/if}}\n\n{{#if example_dialogue}}This is how {{char}} should talk:\n{{example_dialogue}}{{/if}}\n\nThis scenario of the conversation: {{scenario}}\n\nThen the roleplay chat between {{#each bot}}{{.name}}, {{/each}}{{char}} and {{user}} begins.<|eot_id|>\n\n{{#each msg}}{{#if .isbot}}<|start_header_id|>response<|end_header_id|>{{/if}}{{#if .isuser}}<|start_header_id|>user<|end_header_id|>{{/if}}{{.name}}: {{.msg}}<|eot_id|>\n{{/each}}\n{{#if ujb}}<|begin_of_text|><|start_header_id|>system<|end_header_id|>{{ujb}}<|eot_id|>{{/if}}\n<|start_header_id|>response<|end_header_id|>{{post}}\n```",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "dataset:PygmalionAI/PIPPA",
    "dataset:lemonilia/LimaRP",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 0,
  "downloads": 88,
  "gated": false,
  "private": false,
  "last_modified": "2024-04-23T01:16:28.000Z",
  "created_at": "2024-04-21T13:52:06.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "66251a06473c87d5504b814d",
  "id": "ludis/tsukasa-llama-3-70b-qlora-gguf",
  "modelId": "ludis/tsukasa-llama-3-70b-qlora-gguf",
  "sha": "44f1510ff74bcc613f98626519b251e66b54a60b",
  "createdAt": "2024-04-21T13:52:06.000Z",
  "lastModified": "2024-04-23T01:16:28.000Z",
  "author": "ludis",
  "downloads": 88,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 16
}