models.json

[
    {
        "id": "Internal",
        "model_title": "AI Assistant",
        "model_file": "ggml-model-Q8_0.gguf",
        "model_url": "https://",
        "model_info_url": "https://huggingface.co/princeton-nlp/Sheared-LLaMA-1.3B",
        "model_avatar": "ava0",
        "model_intention": "It's good for talking and casual writing.",
        "model_license": "license_llama2.txt",
        "model_license_info": "Meta Llama 2 Community License Agreement",
        "model_license_url": "https://ai.meta.com/llama/license/",
        "model_description": "It is an AI assistant who can talk with you and help solve simple problems. It's based on a lite LLAMA2 model developed by Meta Inc.",
        "developer": "Meta",
        "developer_url": "https://ai.meta.com/llama/",
        "file_size": 1430,
        "context" : 2048,
        "temp" : 0.6,
        "prompt_format" : "<human>: {{prompt}}\n<bot>: ",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "llama",
        "n_batch" : 10,
        "template_name" : "HumanBot",
        "is_ready": true,
        "is_internal": true,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
    },
    {
         "id": "LiteLlama-460M-1T-Q8",
         "model_title": "LiteLlama",
         "model_file": "LiteLlama-460M-1T-Q8_0.gguf",
         "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/LiteLlama-460M-1T-Q8_0.gguf?download=true",
         "model_info_url": "https://huggingface.co/ahxt/LiteLlama-460M-1T",
         "model_avatar": "logo_litellama",
         "model_intention": "This is a 460 parameters' very small model for test purpose only",
         "model_license": "license_llama2.txt",
         "model_license_info": "Meta Llama 2 Community License Agreement",
         "model_license_url": "https://ai.meta.com/llama/license/",
         "model_description": "It's a very small LLAMA2 model with only 460M parameters trained with 1T tokens. It's best for testing.",
         "developer": "Xiaotian Han from Texas A&M University",
         "developer_url": "https://huggingface.co/ahxt/LiteLlama-460M-1T",
         "file_size": 493,
         "context" : 1024,
         "temp" : 0.6,
         "prompt_format" : "<human>: {{prompt}}\n<bot>:",
         "top_k" : 5,
         "top_p" : 0.9,
         "model_inference" : "llama",
         "n_batch" : 10,
         "template_name" : "TinyLlama",
         "is_ready": true,
         "is_internal": false,
         "use_metal": true,
         "mlock": false,
         "mmap": true,
         "repeat_last_n": 64,
         "repeat_penalty": 1.2,
         "add_bos_token": true,
         "add_eos_token": false,
         "parse_special_tokens": true
    },
    {
         "id": "tinyllama-1.1B-chat-Q8",
         "model_title": "TinyLlama-1.1B-chat",
         "model_file": "tinyllama-1.1B-chat-v1.0-Q8_0.gguf",
         "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/tinyllama-1.1B-chat-v1.0-Q8_0.gguf?download=true",
         "model_info_url": "https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v1.0",
         "model_avatar": "logo_tinyllama",
         "model_intention": "It's good for question & answer.",
         "model_license": "license_llama2.txt",
         "model_license_info": "Meta Llama 2 Community License Agreement",
         "model_license_url": "https://ai.meta.com/llama/license/",
         "model_description": "The TinyLlama project aims to pretrain a 1.1B Llama model on 3 trillion tokens. With some proper optimization, we can achieve this within a span of just 90 days using 16 A100-40G GPUs. The training has started on 2023-09-01.",
         "developer": "Zhang Peiyuan",
         "developer_url": "https://github.com/jzhang38/TinyLlama",
         "file_size": 1170,
         "context" : 4096,
         "temp" : 0.6,
         "prompt_format" : "<|system|>You are a friendly chatbot who always responds in the style of a pirate.</s><|user|>{{prompt}}</s><|assistant|>",
         "top_k" : 5,
         "top_p" : 0.9,
         "model_inference" : "llama",
         "n_batch" : 10,
         "template_name" : "TinyLlama",
         "is_ready": true,
         "is_internal": false,
         "use_metal": true,
         "mlock": false,
         "mmap": true,
         "repeat_last_n": 64,
         "repeat_penalty": 1.2,
         "add_bos_token": true,
         "add_eos_token": false,
         "parse_special_tokens": true
    },
    {
        "id": "mistral-7b-instruct-v0.2-Q8",
        "model_title": "Mistral 7B v0.2",
        "model_file": "mistral-7b-instruct-v0.2.Q8_0.gguf",
        "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/mistral-7b-instruct-v0.2.Q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2",
        "model_avatar": "logo_mistralai",
        "model_intention": "It's a 7B large model for Q&A purpose. But it requires a high-end device to run.",
        "model_license": "license_apache2.txt",
        "model_license_info": "APACHE LICENSE, VERSION 2.0",
        "model_license_url": "https://www.apache.org/licenses/LICENSE-2.0",
        "model_description": "The Mistral-7B-v0.2 Large Language Model (LLM) is a pretrained generative text model with 7 billion parameters. Mistral-7B-v0.2 outperforms Llama 2 13B on all benchmarks we tested.",
        "developer": "Mistral AI",
        "developer_url": "https://mistral.ai/",
        "file_size": 7695,
        "context" : 4096,
        "temp" : 0.6,
        "prompt_format" : "<s>[INST]{{prompt}}[/INST]</s>",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "llama",
        "n_batch" : 10,
        "template_name" : "Mistral",
        "is_ready": true,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   },
   {
        "id": "openchat-3.5-1210-Q8",
        "model_title": "OpenChat 3.5",
        "model_file": "mistral-7b-instruct-v0.2.Q8.gguf",
        "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/openchat-3.5-1210.Q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/openchat/openchat_3.5",
        "model_avatar": "logo_openchat",
        "model_intention": "It's a 7B large model and performs really good for Q&A. But it requires a high-end device to run.",
        "model_license": "license_apache2.txt",
        "model_license_info": "APACHE LICENSE, VERSION 2.0",
        "model_license_url": "https://www.apache.org/licenses/LICENSE-2.0",
        "model_description": "OpenChat is an innovative library of open-source language models, fine-tuned with C-RLFT - a strategy inspired by offline reinforcement learning. Our models learn from mixed-quality data without preference labels, delivering exceptional performance on par with ChatGPT, even with a 7B model. Despite our simple approach, we are committed to developing a high-performance, commercially viable, open-source large language model, and we continue to make significant strides toward this vision.",
        "developer": "OpenChat Team",
        "developer_url": "https://openchat.team/",
        "file_size": 7695,
        "context" : 4096,
        "temp" : 0.6,
        "prompt_format" : "<s>[INST]{{prompt}}[/INST]</s>",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "llama",
        "n_batch" : 10,
        "template_name" : "Mistral",
        "is_ready": true,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   },
   {
        "id": "phi-2",
        "model_title": "Phi-2",
        "model_file": "phi-2.Q8_0.gguf",
        "model_url": "https://huggingface.co/ggml-org/models/resolve/main/phi-2/ggml-model-q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/microsoft/phi-2",
        "model_avatar": "logo_phi",
        "model_intention": "It's a 2.7B model and is intended for QA, chat, and code purposes",
        "model_license": "license_mit.txt",
        "model_license_info": "The MIT License",
        "model_license_url": "https://opensource.org/license/mit",
        "model_description": "Phi-2 is a Transformer with 2.7 billion parameters. It was trained using the same data sources as Phi-1.5, augmented with a new data source that consists of various NLP synthetic texts and filtered websites (for safety and educational value). When assessed against benchmarks testing common sense, language understanding, and logical reasoning, Phi-2 showcased a nearly state-of-the-art performance among models with less than 13 billion parameters.",
        "developer": "Microsoft",
        "developer_url": "https://huggingface.co/microsoft/phi-2",
        "file_size": 2960,
        "context" : 4096,
        "temp" : 0.6,
        "prompt_format" : "Instruct: {{prompt}}\nOutput:",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "llama",
        "n_batch" : 10,
        "template_name" : "PHI",
        "is_ready": true,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   },
   {
        "id": "yi-6b",
        "model_title": "Yi 6B Chat",
        "model_file": "yi-6b-chat-Q8_0.gguf",
        "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/yi-6b-chat-Q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/01-ai/Yi-6B-Chat",
        "model_avatar": "logo_yi",
        "model_intention": "It's a 6B model and can understand English and Chinese. It's good for QA and Chat",
        "model_license": "license_apache2.txt",
        "model_license_info": "APACHE LICENSE, VERSION 2.0",
        "model_license_url": "https://www.apache.org/licenses/LICENSE-2.0",
        "model_description": "The Yi series models are the next generation of open-source large language models trained from scratch by 01.AI. Targeted as a bilingual language model and trained on 3T multilingual corpus, the Yi series models become one of the strongest LLM worldwide, showing promise in language understanding, commonsense reasoning, reading comprehension, and more. For example, For English language capability, the Yi series models ranked 2nd (just behind GPT-4), outperforming other LLMs (such as LLaMA2-chat-70B, Claude 2, and ChatGPT) on the AlpacaEval Leaderboard in Dec 2023. For Chinese language capability, the Yi series models landed in 2nd place (following GPT-4), surpassing other LLMs (such as Baidu ERNIE, Qwen, and Baichuan) on the SuperCLUE in Oct 2023.",
        "developer": "01.AI",
        "developer_url": "https://01.ai/",
        "file_size": 6440,
        "context" : 200000,
        "temp" : 0.6,
        "prompt_format" : "<|im_start|>user\n<|im_end|>\n{{prompt}}\n<|im_start|>assistant\n",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "llama",
        "n_batch" : 10,
        "template_name" : "yi",
        "is_ready": true,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   },
   {
        "id": "gemma-2b",
        "model_title": "Google Gemma 2B",
        "model_file": "gemma-2b-it-q8_0.gguf",
        "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/gemma-2b-it-q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/google/gemma-2b",
        "model_avatar": "logo_google",
        "model_intention": "It's a 2B large model for Q&A purpose. But it requires a high-end device to run.",
        "model_license": "license_gemma.txt",
        "model_license_info": "APACHE LICENSE, VERSION 2.0",
        "model_license_url": "https://www.apache.org/licenses/LICENSE-2.0",
        "model_description": "Gemma is a family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models. Developed by Google DeepMind and other teams across Google, Gemma is named after the Latin gemma, meaning 'precious stone.' The Gemma model weights are supported by developer tools that promote innovation, collaboration, and the responsible use of artificial intelligence (AI).",
        "developer": "Google",
        "developer_url": "https://huggingface.co/google",
        "file_size": 2669,
        "context" : 8192,
        "temp" : 0.6,
        "prompt_format" : "<bos><start_of_turn>user\n{{prompt}}<end_of_turn>\n<start_of_turn>model\n",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "gemma",
        "n_batch" : 10,
        "template_name" : "gemma",
        "is_ready": false,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   },
   {
        "id": "starcoder2-3b",
        "model_title": "StarCoder2 3B",
        "model_file": "starcoder2-3b-instruct-gguf_Q8_0.gguf",
        "model_url": "https://huggingface.co/flyingfishinwater/goodmodels/resolve/main/starcoder2-3b-instruct-gguf_Q8_0.gguf?download=true",
        "model_info_url": "https://huggingface.co/bigcode/starcoder2-3b",
        "model_avatar": "logo_starcoder",
        "model_intention": "The model is good at 17 programming languages. It can help you resolve programming requirements",
        "model_license": "license_bigcode.txt",
        "model_license_info": "APACHE LICENSE, VERSION 2.0",
        "model_license_url": "https://www.apache.org/licenses/LICENSE-2.0",
        "model_description": "StarCoder2-3B model is a 3B parameter model trained on 17 programming languages from The Stack v2, with opt-out requests excluded. The model uses Grouped Query Attention, a context window of 16,384 tokens with a sliding window attention of 4,096 tokens, and was trained using the Fill-in-the-Middle objective on 3+ trillion tokens",
        "developer": "Bigcode",
        "developer_url": "https://www.bigcode-project.org/",
        "file_size": 3220,
        "context" : 8192,
        "temp" : 0.6,
        "prompt_format" : "### Instruction\n{{prompt}}### Response\n",
        "top_k" : 5,
        "top_p" : 0.9,
        "model_inference" : "starcoder",
        "n_batch" : 10,
        "template_name" : "starcoder",
        "is_ready": false,
        "is_internal": false,
        "use_metal": true,
        "mlock": false,
        "mmap": true,
        "repeat_last_n": 64,
        "repeat_penalty": 1.2,
        "add_bos_token": true,
        "add_eos_token": false,
        "parse_special_tokens": true
   }
]