jan/models/capybara-34b/model.json
Hoang Ha e6812b1247
chore: pre-populate Jan's /models folder with model.jsons (#775)
* draft model.json

* islm3b update

* capybara 34b update

* deepseek coder update

* dolphin yi update

* fix the maxtokens of islm

* lzlv 70b update

* marx3b update

* mythomax 13b update

* update neural chat 7b

* noromaid 20b update

* update openchat 7b

* openhermes7b update

* openorca 7b

* orca 13b update

* phind 34b update

* rocket 3b update

* starling 7b update

* storytelling 70b update

* tiefighter 13B

* update tiefighter tags

* tinyllama update

* wizard coder 13b

* update wizard coder 13b description

* wizard coder 34b update

* wizard coder minor fix

* xwin 70b update

* yarn 70b

* yi 34b

* zephyr beta 7b

* neuralhermes-7b update

* change path + ctxlen

* update id

* fix startling
2023-12-01 17:20:58 +07:00

24 lines
769 B
JSON

{
"source_url": "https://huggingface.co/TheBloke/Nous-Capybara-34B-GGUF/resolve/main/nous-capybara-34b.Q5_K_M.gguf",
"id": "capybara-34b",
"object": "model",
"name": "Capybara 200k 34B",
"version": 1.0,
"description": "Nous Capybara 34B, a variant of the Yi-34B model, is the first Nous model with a 200K context length, trained for three epochs on the innovative Capybara dataset.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "USER: ",
"ai_prompt": "ASSISTANT: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "NousResearch, The Bloke",
"tags": ["General", "Big Context Length"],
"size": 24320000000
}
}