Nicholai/jan
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #829 from janhq/update-model-hub

Browse Source 
Update model hub
This commit is contained in:
Hoang Ha 2023-12-04 21:57:28 +07:00 committed by GitHub
commit b475a6f551
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
33 changed files with 236 additions and 352 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
models/capybara-34b/model.json
View File

@ -2,22 +2,22 @@
"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,
"name": "Capybara 200k 34B Q5",
"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,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "USER: ",
"ai_prompt": "ASSISTANT: "
"user_prompt": "USER:\n",
"ai_prompt": "ASSISTANT:\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "NousResearch, The Bloke",
"tags": ["General", "Big Context Length"],
"tags": ["Recommended", "Medium", "Finetuned"],
"size": 24320000000
}
}

12
models/deepseek-coder-1.3b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/deepseek-coder-1.3b-base-GGUF/resolve/main/deepseek-coder-1.3b-base.Q4_K_M.gguf",
"id": "deepseek-coder-1.3b",
"object": "model",
"name": "Deepseek Coder 1.3B",
"name": "Deepseek Coder 1.3B Q4",
"version": "1.0",
"description": "",
"description": "Deepseek Coder trained on 2T tokens (87% code, 13% English/Chinese), excelling in project-level code completion with advanced capabilities across multiple programming languages.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "",
"ai_prompt": ""
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "deepseek, The Bloke",
"tags": ["Code"],
"author": "Deepseek, The Bloke",
"tags": ["Tiny", "Foundational Model"],
"size": 870000000
}
}

24
models/deepseek-coder-34b/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/deepseek-coder-33B-instruct-GGUF/resolve/main/deepseek-coder-33b-instruct.Q5_K_M.gguf",
"id": "deepseek-coder-34b",
"object": "model",
"name": "Deepseek Coder 33B Q5",
"version": "1.0",
"description": "Deepseek Coder trained on 2T tokens (87% code, 13% English/Chinese), excelling in project-level code completion with advanced capabilities across multiple programming languages.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "### Instruction:\n",
"ai_prompt": "### Response:\n"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "Deepseek, The Bloke",
"tags": ["Recommended", "Medium", "Foundational Model"],
"size": 26040000000
}
}

24
models/dolphin-yi-34b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/dolphin-2_2-yi-34b-GGUF/resolve/main/dolphin-2_2-yi-34b.Q5_K_M.gguf",
"id": "dolphin-yi-34b",
"object": "model",
"name": "Dolphin Yi 34B",
"version": "1.0",
"description": "Dolphin, based on the Yi-34B model and enhanced with features like conversation and empathy, is trained on a unique dataset for advanced multi-turn conversations. Notably uncensored, it requires careful implementation of an alignment layer for ethical use.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "ehartford, The Bloke",
"tags": ["General Use", "Role-playing"],
"size": 24320000000
}
}

24
models/islm-3b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/UmbrellaCorp/IS-LM-3B_GGUF/resolve/main/IS-LM-Q4_K_M.gguf",
"id": "islm-3b",
"object": "model",
"name": "IS LM 3B",
"version": "1.0",
"description": "IS LM 3B, based on the StableLM 3B model is specifically finetuned for economic analysis using DataForge Economics and QLoRA over three epochs, enhancing its proficiency in economic forecasting and analysis.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "USER: ",
"ai_prompt": "ASSISTANT: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "UmbrellaCorp, The Bloke",
"tags": ["General Use", "Economics"],
"size": 1710000000
}
}

24
models/llama2-chat-70b-q4/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/Llama-2-70B-Chat-GGUF/resolve/main/llama-2-70b-chat.Q4_K_M.gguf",
"id": "llama2-chat-70b-q4",
"object": "model",
"name": "Llama 2 Chat 70B Q4",
"version": "1.0",
"description": "This is a 4-bit quantized version of Meta AI's Llama 2 Chat 70b model.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "[INST] <<SYS>>\n",
"user_prompt": "<</SYS>>\n",
"ai_prompt": "[/INST]"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "MetaAI, The Bloke",
"tags": ["Large", "Foundational Model"],
"size": 43920000000
}
}

24
models/llama2-chat-7b-q4/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGUF/resolve/main/llama-2-7b-chat.Q4_K_M.gguf",
"id": "llama2-chat-7b-q4",
"object": "model",
"name": "Llama 2 Chat 7B Q4",
"version": "1.0",
"description": "This is a 4-bit quantized iteration of Meta AI's Llama 2 Chat 7b model, specifically designed for a comprehensive understanding through training on extensive internet data.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "[INST] <<SYS>>\n",
"user_prompt": "<</SYS>>\n",
"ai_prompt": "[/INST]"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "MetaAI, The Bloke",
"tags": ["Small", "Foundational Model"],
"size": 4080000000
}
}

24
models/llama2-chat-7b-q5/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGUF/resolve/main/llama-2-7b-chat.Q5_K_M.gguf",
"id": "llama2-chat-7b-q5",
"object": "model",
"name": "Llama 2 Chat 7B Q5",
"version": "1.0",
"description": "This is a 5-bit quantized iteration of Meta AI's Llama 2 Chat 7b model, specifically designed for a comprehensive understanding through training on extensive internet data.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "[INST] <<SYS>>\n",
"user_prompt": "<</SYS>>\n",
"ai_prompt": "[/INST]"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "MetaAI, The Bloke",
"tags": ["Small", "Foundational Model"],
"size": 4780000000
}
}

14
models/lzlv-70b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/lzlv_70B-GGUF/resolve/main/lzlv_70b_fp16_hf.Q5_K_M.gguf",
"id": "lzlv-70b",
"object": "model",
"name": "Lzlv 70B",
"name": "Lzlv 70B Q4",
"version": "1.0",
"description": "lzlv_70B is a sophisticated AI model designed for roleplaying and creative tasks. This merge aims to combine intelligence with creativity, seemingly outperforming its individual components in complex scenarios and creative outputs.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "USER: ",
"ai_prompt": "ASSISTANT: "
"user_prompt": "USER:\n",
"ai_prompt": "ASSISTANT:\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "lizpreciatior, The Bloke",
"tags": ["General Use", "Role-playing"],
"author": "Lizpreciatior, The Bloke",
"tags": ["Recommended", "Large", "Finetuned"],
"size": 48750000000
}
}

23
models/marx-3b/model.json
View File

@ -1,23 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/Marx-3B-v3-GGUF/resolve/main/marx-3b-v3.Q4_K_M.gguf",
"id": "marx-3b",
"object": "model",
"name": "Marx 3B",
"version": "1.0",
"description": "Marx 3B, based on the StableLM 3B model is specifically finetuned for chating using EverythingLM data and QLoRA over two epochs, enhancing its proficiency in understand general knowledege.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "### HUMAN: ",
"ai_prompt": "### RESPONSE: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "Bohan Du, The Bloke",
"tags": ["General Use"],
"size": 1620000000
}
}

24
models/mistral-ins-7b-q4/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF/resolve/main/mistral-7b-instruct-v0.1.Q4_K_M.gguf",
"id": "mistral-ins-7b-q4",
"object": "model",
"name": "Mistral Instruct 7B Q4",
"version": "1.0",
"description": "This is a 4-bit quantized iteration of MistralAI's Mistral Instruct 7b model, specifically designed for a comprehensive understanding through training on extensive internet data.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "<s>[INST]",
"ai_prompt": "[/INST]"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "MistralAI, The Bloke",
"tags": ["Small", "Foundational Model"],
"size": 4370000000
}
}

24
models/mistral-ins-7b-q5/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF/resolve/main/mistral-7b-instruct-v0.1.Q5_K_M.gguf",
"id": "mistral-ins-7b-q5",
"object": "model",
"name": "Mistral Instruct 7B Q5",
"version": "1.0",
"description": "This is a 5-bit quantized iteration of MistralAI's Mistral Instruct 7b model, specifically designed for a comprehensive understanding through training on extensive internet data.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "<s>[INST]",
"ai_prompt": "[/INST]"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "MistralAI, The Bloke",
"tags": ["Small", "Foundational Model"],
"size": 5130000000
}
}

24
models/mythomax-13b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/MythoMax-L2-13B-GGUF/resolve/main/mythomax-l2-13b.Q5_K_M.gguf",
"id": "mythomax-13b",
"object": "model",
"name": "Mythomax L2 13B",
"version": "1.0",
"description": "Mythomax L2 13b, an advanced AI model derived from MythoMix, merges MythoLogic-L2's deep comprehension with Huginn's writing skills through a unique tensor merge technique, excelling in roleplaying and storytelling.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "### Instruction: ",
"ai_prompt": "### Response: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "Gryphe, The Bloke",
"tags": ["Role-playing"],
"size": 9230000000
}
}

14
models/neural-chat-7b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/neural-chat-7B-v3-1-GGUF/resolve/main/neural-chat-7b-v3-1.Q4_K_M.gguf",
"id": "neural-chat-7b",
"object": "model",
"name": "Neural Chat 7B",
"name": "Neural Chat 7B Q4",
"version": "1.0",
"description": "The Neural Chat 7B model, developed on the foundation of mistralai/Mistral-7B-v0.1, has been fine-tuned using the Open-Orca/SlimOrca dataset and aligned with the Direct Preference Optimization (DPO) algorithm. It has demonstrated substantial improvements in various AI tasks and performance well on the open_llm_leaderboard.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "### System: ",
"user_prompt": "### User: ",
"ai_prompt": "### Assistant: "
"ctx_len": 2048,
"system_prompt": "### System:\n",
"user_prompt": "### User:\n",
"ai_prompt": "### Assistant:\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "Intel, The Bloke",
"tags": ["General Use", "Role-playing", "Big Context Length"],
"tags": ["Recommended", "Small", "Finetuned"],
"size": 4370000000
}
}

24
models/neuralhermes-7b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/NeuralHermes-2.5-Mistral-7B-GGUF/resolve/main/neuralhermes-2.5-mistral-7b.Q4_K_M.gguf",
"id": "neuralhermes-7b",
"object": "model",
"name": "NeuralHermes 7B",
"version": "1.0",
"description": "NeuralHermes 2.5 has been enhanced using Direct Preference Optimization. This fine-tuning, inspired by the RLHF process of Neural-chat-7b and OpenHermes-2.5-Mistral-7B, has led to improved performance across several benchmarks.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "Intel, The Bloke",
"tags": ["General Use", "Code", "Big Context Length"],
"size": 4370000000
}
}

12
models/noromaid-20b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/Noromaid-20B-v0.1.1-GGUF/resolve/main/noromaid-20b-v0.1.1.Q4_K_M.gguf",
"id": "noromaid-20b",
"object": "model",
"name": "Noromaid 20B",
"name": "Noromaid 20B Q4",
"version": "1.0",
"description": "The Noromaid 20b model is designed for role-playing and general use, featuring a unique touch with the no_robots dataset that enhances human-like behavior.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "### Instruction: ",
"ai_prompt": "### Response: "
"user_prompt": "### Instruction:\n",
"ai_prompt": "### Response:\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "NeverSleep, The Bloke",
"tags": ["Role-playing"],
"tags": ["Medium", "Finetuned"],
"size": 12040000000
}
}

24
models/openchat-7b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/openchat_3.5-GGUF/resolve/main/openchat_3.5.Q4_K_M.gguf",
"id": "openchat-7b",
"object": "model",
"name": "Open Chat 3.5 7B",
"version": "1.0",
"description": "OpenChat represents a breakthrough in the realm of open-source language models. By implementing the C-RLFT fine-tuning strategy, inspired by offline reinforcement learning, this 7B model achieves results on par with ChatGPT (March).",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "GPT4 User: ",
"ai_prompt": "<|end_of_turn|>\nGPT4 Assistant: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "OpenChat, The Bloke",
"tags": ["General", "Code"],
"size": 4370000000
}
}

24
models/openhermes-mistral-7b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/OpenHermes-2.5-Mistral-7B-GGUF/resolve/main/openhermes-2.5-mistral-7b.Q4_K_M.gguf",
"id": "openhermes-mistral-7b",
"object": "model",
"name": "Openhermes 2.5 Mistral 7B",
"version": "1.0",
"description": "The OpenHermes 2.5 Mistral 7B incorporates additional code datasets, more than a million GPT-4 generated data examples, and other high-quality open datasets. This enhancement led to significant improvement in benchmarks, highlighting its improved skill in handling code-centric tasks.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "Teknium, The Bloke",
"tags": ["General", "Roleplay"],
"size": 4370000000
}
}

24
models/openhermes-neural-7b/model.json Normal file
View File

@ -0,0 +1,24 @@
{
"source_url": "https://huggingface.co/TheBloke/OpenHermes-2.5-neural-chat-7B-v3-2-7B-GGUF/resolve/main/openhermes-2.5-neural-chat-7b-v3-2-7b.Q4_K_M.gguf",
"id": "openhermes-neural-7b",
"object": "model",
"name": "OpenHermes Neural 7B Q4",
"version": "1.0",
"description": "OpenHermes Neural is a merged model from OpenHermes-2.5-Mistral-7B and neural-chat-7b-v3-2 with the TIES method.",
"format": "gguf",
"settings": {
"ctx_len": 2048,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 2048
},
"metadata": {
"author": "Intel, The Bloke",
"tags": ["Recommended", "Small", "Merged"],
"size": 4370000000
}
}

8
models/openorca-13b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/Orca-2-13B-GGUF/resolve/main/orca-2-13b.Q5_K_M.gguf",
"id": "openorca-13b",
"object": "model",
"name": "Orca 2 13B",
"name": "Orca 2 13B Q5",
"version": "1.0",
"description": "Orca 2 is a finetuned version of LLAMA-2, designed primarily for single-turn responses in reasoning, reading comprehension, math problem solving, and text summarization.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "Microsoft, The Bloke",
"tags": ["General Use"],
"tags": ["Medium", "Finetuned"],
"size": 9230000000
}
}

24
models/openorca-7b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/Mistral-7B-OpenOrca-GGUF/resolve/main/mistral-7b-openorca.Q4_K_M.gguf",
"id": "openorca-7b",
"object": "model",
"name": "OpenOrca 7B",
"version": "1.0",
"description": "OpenOrca 8k 7B is a model based on Mistral 7B, fine-tuned using the OpenOrca dataset. Notably ranked first on the HF Leaderboard for models under 30B, it excels in efficiency and accessibility.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "OpenOrca, The Bloke",
"tags": ["General", "Code"],
"size": 4370000000
}
}

8
models/phind-34b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/Phind-CodeLlama-34B-v2-GGUF/resolve/main/phind-codellama-34b-v2.Q5_K_M.gguf",
"id": "phind-34b",
"object": "model",
"name": "Phind 34B",
"name": "Phind 34B Q5",
"version": "1.0",
"description": "Phind-CodeLlama-34B-v2 is an AI model fine-tuned on 1.5B tokens of high-quality programming data. It's a SOTA open-source model in coding. This multi-lingual model excels in various programming languages, including Python, C/C++, TypeScript, Java, and is designed to be steerable and user-friendly.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "### System Prompt\n",
"user_prompt": "### User Message\n",
"ai_prompt": "### Assistant\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "Phind, The Bloke",
"tags": ["Code", "Big Context Length"],
"tags": ["Recommended", "Medium", "Finetuned"],
"size": 24320000000
}
}

8
models/rocket-3b/model.json
View File

@ -2,7 +2,7 @@
"source_url": "https://huggingface.co/TheBloke/rocket-3B-GGUF/resolve/main/rocket-3b.Q4_K_M.gguf",
"id": "rocket-3b",
"object": "model",
"name": "Rocket 3B",
"name": "Rocket 3B Q4",
"version": "1.0",
"description": "Rocket-3B is a GPT-like model, primarily English, fine-tuned on diverse public datasets. It outperforms larger models in benchmarks, showcasing superior understanding and text generation, making it an effective chat model for its size.",
"format": "gguf",
@ -13,11 +13,11 @@
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "pansophic, The Bloke",
"tags": ["General Use"],
"author": "Pansophic, The Bloke",
"tags": ["Tiny", "Finetuned"],
"size": 1710000000
}
}

8
models/starling-7b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/Starling-LM-7B-alpha-GGUF/resolve/main/starling-lm-7b-alpha.Q4_K_M.gguf",
"id": "starling-7b",
"object": "model",
"name": "Strarling alpha 7B",
"name": "Strarling alpha 7B Q4",
"version": "1.0",
"description": "Starling-RM-7B-alpha is a language model finetuned with Reinforcement Learning from AI Feedback from Openchat 3.5. It stands out for its impressive performance using GPT-4 as a judge, making it one of the top-performing models in its category.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "GPT4 User: ",
"ai_prompt": "<|end_of_turn|>\nGPT4 Assistant: "
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "Berkeley-nest, The Bloke",
"tags": ["General", "Code"],
"tags": ["Recommended", "Small","Finetuned"],
"size": 4370000000
}
}

24
models/storytelling-70b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/GOAT-70B-Storytelling-GGUF/resolve/main/goat-70b-storytelling.Q5_K_M.gguf",
"id": "storytelling-70b",
"object": "model",
"name": "Storytelling 70B",
"version": "1.0",
"description": "The GOAT-70B-Storytelling model is designed for autonomous story-writing, including crafting books and movie scripts. Based on the LLaMA 2 70B architecture, this model excels in generating cohesive and engaging narratives using inputs like plot outlines and character profiles.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "### USER: ",
"ai_prompt": "\n### ASSISTANT: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "GOAT-AI, The Bloke",
"tags": ["General Use", "Writing"],
"size": 48750000000
}
}

8
models/tiefighter-13b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/LLaMA2-13B-Tiefighter-GGUF/resolve/main/llama2-13b-tiefighter.Q5_K_M.gguf",
"id": "tiefighter-13b",
"object": "model",
"name": "Tiefighter 13B",
"name": "Tiefighter 13B Q5",
"version": "1.0",
"description": "Tiefighter-13B is a highly creative, merged AI model achieved by combining various 'LORAs' on top of an existing merge, particularly focusing on storytelling and improvisation. This model excels in story writing, chatbots, and adventuring, and is designed to perform better with less detailed inputs, leveraging its inherent creativity.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "### Instruction: ",
"ai_prompt": "\n### Response: "
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "KoboldAI, The Bloke",
"tags": ["General Use", "Role-playing", "Writing"],
"tags": ["Medium", "Finetuned"],
"size": 9230000000
}
}

4
models/tinyllama-1.1b/model.json
View File

@ -2,7 +2,7 @@
"source_url": "https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v0.6/resolve/main/ggml-model-q4_0.gguf",
"id": "tinyllama-1.1b",
"object": "model",
"name": "TinyLlama Chat 1.1B",
"name": "TinyLlama Chat 1.1B Q4",
"version": "1.0",
"description": "The TinyLlama project, featuring a 1.1B parameter Llama model, is pretrained on an expansive 3 trillion token dataset. Its design ensures easy integration with various Llama-based open-source projects. Despite its smaller size, it efficiently utilizes lower computational and memory resources, drawing on GPT-4's analytical prowess to enhance its conversational abilities and versatility.",
"format": "gguf",
@ -17,7 +17,7 @@
},
"metadata": {
"author": "TinyLlama",
"tags": ["General Use"],
"tags": ["Tiny", "Foundation Model"],
"size": 637000000
}
}

8
models/wizardcoder-13b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/WizardCoder-Python-13B-V1.0-GGUF/resolve/main/wizardcoder-python-13b-v1.0.Q5_K_M.gguf",
"id": "wizardcoder-13b",
"object": "model",
"name": "Wizard Coder Python 13B",
"name": "Wizard Coder Python 13B Q5",
"version": "1.0",
"description": "WizardCoder-Python-13B is a Python coding model major models like ChatGPT-3.5. This model based on the Llama2 architecture, demonstrate high proficiency in specific domains like coding and mathematics.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "",
"user_prompt": "### Instruction:\n",
"ai_prompt": "### Response:\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "WizardLM, The Bloke",
"tags": ["Code", "Big Context Length"],
"tags": ["Recommended", "Medium", "Finetuned"],
"size": 9230000000
}
}

24
models/wizardcoder-34b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/WizardCoder-Python-34B-V1.0-GGUF/resolve/main/wizardcoder-python-34b-v1.0.Q5_K_M.gguf",
"id": "wizardcoder-34b",
"object": "model",
"name": "Wizard Coder Python 34B",
"version": "1.0",
"description": "WizardCoder-Python-34B is a Python coding model major models like ChatGPT-3.5. This model based on the Llama2 architecture, demonstrate high proficiency in specific domains like coding and mathematics.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "### Instruction:\n",
"ai_prompt": "### Response:\n"
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "WizardLM, The Bloke",
"tags": ["Code", "Big Context Length"],
"size": 24320000000
}
}

24
models/xwin-70b/model.json
View File

@ -1,24 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/Xwin-LM-70B-V0.1-GGUF/resolve/main/xwin-lm-70b-v0.1.Q5_K_M.gguf",
"id": "xwin-70b",
"object": "model",
"name": "Xwin LM 70B",
"version": "1.0",
"description": "Xwin-LM, based on Llama2 models, emphasizes alignment and exhibits advanced language understanding, text generation, and role-playing abilities.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"system_prompt": "",
"user_prompt": "USER: ",
"ai_prompt": "ASSISTANT: "
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "Xwin-LM, The Bloke",
"tags": ["General Use", "Role-playing"],
"size": 48750000000
}
}

21
models/yarn-70b/model.json
View File

@ -1,21 +0,0 @@
{
"source_url": "https://huggingface.co/TheBloke/Yarn-Llama-2-70B-32k-GGUF/resolve/main/yarn-llama-2-70b-32k.Q5_K_M.gguf",
"id": "yarn-70b",
"object": "model",
"name": "Yarn 32k 70B",
"version": "1,0",
"description": "Yarn-Llama-2-70b-32k is designed specifically for handling long contexts. It represents an extension of the Llama-2-70b-hf model, now supporting a 32k token context window.",
"format": "gguf",
"settings": {
"ctx_len": 4096
},
"parameters": {
"max_tokens": 4096
},
"metadata": {
"author": "NousResearch, The Bloke",
"tags": ["General Use", "Big Context Length"],
"size": 48750000000
}
}

8
models/yi-34b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/Yi-34B-Chat-GGUF/resolve/main/yi-34b-chat.Q5_K_M.gguf",
"id": "yi-34b",
"object": "model",
"name": "Yi 34B",
"name": "Yi 34B Q5",
"version": "1.0",
"description": "Yi-34B, a specialized chat model, is known for its diverse and creative responses and excels across various NLP tasks and benchmarks.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "<|im_start|>system\n",
"user_prompt": "<|im_end|>\n<|im_start|>user\n",
"ai_prompt": "<|im_end|>\n<|im_start|>assistant\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "01-ai, The Bloke",
"tags": ["General", "Role-playing", "Writing"],
"tags": ["Medium", "Foundational Model"],
"size": 24320000000
}
}

10
models/zephyr-beta-7b/model.json
View File

@ -2,22 +2,22 @@
"source_url": "https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/resolve/main/zephyr-7b-beta.Q4_K_M.gguf",
"id": "zephyr-beta-7b",
"object": "model",
"name": "Zephyr Beta 7B",
"name": "Zephyr Beta 7B Q4",
"version": "1.0",
"description": "The Zephyr-7B-β model marks the second iteration in the Zephyr series, designed to function as an effective assistant. It has been fine-tuned from the mistralai/Mistral-7B-v0.1 base model, utilizing a combination of public and synthetic datasets with the application of Direct Preference Optimization.",
"description": "The Zephyr-7B-β model is trained by HuggingFace, designed to function as a practical assistant. It has been fine-tuned from the mistralai/Mistral-7B-v0.1 base model, utilizing a combination of public and synthetic datasets with the application of Direct Preference Optimization.",
"format": "gguf",
"settings": {
"ctx_len": 4096,
"ctx_len": 2048,
"system_prompt": "<|system|>\n",
"user_prompt": "</s>\n<|user|>\n",
"ai_prompt": "</s>\n<|assistant|>\n"
},
"parameters": {
"max_tokens": 4096
"max_tokens": 2048
},
"metadata": {
"author": "HuggingFaceH4, The Bloke",
"tags": ["General Use", "Big Context Length"],
"tags": ["Recommended", "Small", "Finetuned"],
"size": 4370000000
}
}