Update Model.json (#1005)

* add(mixtral): add model.json for mixtral * archived some models + update the model.json * add(model): add pandora 10.7b * fix(model): update description * fix(model): pump vers and change the featured model to trinity * fix(model): archive neuralchat * fix(model): decapriated all old models * fix(trinity): add cover image and change description * fix(trinity): update cover png * add(pandora): cover image * fix(pandora): cover image * add(mixtral): add model.json for mixtral * archived some models + update the model.json * add(model): add pandora 10.7b * fix(model): update description * fix(model): pump vers and change the featured model to trinity * fix(model): archive neuralchat * fix(model): decapriated all old models * fix(trinity): add cover image and change description * fix(trinity): update cover png * add(pandora): cover image * fix(pandora): cover image * chore: model desc nits * fix(models): adjust the size for solars and pandoras * add(mixtral): description --------- Co-authored-by: 0xSage <n@pragmatic.vc>
janhq · Dec 20, 2023 · a893b59 · a893b59
1 parent 314df59
commit a893b59
Show file tree

Hide file tree

Showing 5 changed files with 110 additions and 88 deletions.
diff --git a/models/mixtral-8x7b-instruct/model.json b/models/mixtral-8x7b-instruct/model.json
@@ -1,22 +1,28 @@
 {
-    "source_url": "https://huggingface.co/TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF/resolve/main/mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf",
-    "id": "mixtral-8x7b-instruct",
-    "object": "model",
-    "name": "Mixtral 8x7B Instruct Q4",
-    "version": "1.0",
-    "description": "The Mixtral-8x7B Large Language Model (LLM) is a pretrained generative Sparse Mixture of Experts. The Mixtral-8x7B outperforms Llama 2 70B on most benchmarks we tested.",
-    "format": "gguf",
-    "settings": {
-      "ctx_len": 4096,
-      "prompt_template": "[INST] {prompt} [/INST]"
-    },
-    "parameters": {
-      "max_tokens": 4096
-    },
-    "metadata": {
-      "author": "MistralAI, TheBloke",
-      "tags": ["MOE", "Foundational Model"],
-      "size": 26440000000
-    },
-    "engine": "nitro"
-  }
+  "source": [
+    {
+      "filename": "mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf",
+      "url": "https://huggingface.co/TheBloke/Mixtral-8x7B-Instruct-v0.1-GGUF/resolve/main/mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf"
+    }
+  ],
+  "id": "mixtral-8x7b-instruct",
+  "object": "model",
+  "name": "Mixtral 8x7B Instruct Q4",
+  "version": "1.0",
+  "description": "The Mixtral-8x7B Large Language Model (LLM) is a pretrained generative Sparse Mixture of Experts. The Mixtral-8x7B outperforms Llama 2 70B on most benchmarks we tested.",
+  "format": "gguf",
+  "settings": {
+    "ctx_len": 4096,
+    "prompt_template": "[INST] {prompt} [/INST]",
+    "llama_model_path": "mixtral-8x7b-instruct-v0.1.Q4_K_M.gguf"
+  },
+  "parameters": {
+    "max_tokens": 4096
+  },
+  "metadata": {
+    "author": "MistralAI, TheBloke",
+    "tags": ["MOE", "Foundational Model"],
+    "size": 26440000000
+  },
+  "engine": "nitro"
+}
diff --git a/models/openhermes-neural-7b/cover.png b/models/openhermes-neural-7b/cover.png
diff --git a/models/pandora-10.7b-v1/model.json b/models/pandora-10.7b-v1/model.json
@@ -1,24 +1,29 @@
 {
-    "source_url": "https://huggingface.co/janhq/pandora-v1-10.7b-GGUF/resolve/main/pandora-v1-10.7b.Q4_K_M.gguf",
-    "id": "pandora-10.7b-v1",
-    "object": "model",
-    "name": "Pandora 11B Q4",
-    "version": "1.0",
-    "description": "Pandora, our research model, employs the Passthrough merging technique to merge 2x7B models into 1.",
-    "format": "gguf",
-    "settings": {
-      "ctx_len": 4096,
-      "prompt_template": "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant"
-    },
-    "parameters": {
-      "max_tokens": 4096
-    },
-    "metadata": {
-      "author": "Jan",
-      "tags": ["11B","Finetuned", "Featured"],
-      "size": 6360000000,
-      "cover": "https://raw.githubusercontent.com/janhq/jan/main/models/pandora-10.7b-v1/cover.png"
-    },
-    "engine": "nitro"
-  }
-
+  "source": [
+    {
+      "filename": "pandora-v1-10.7b.Q4_K_M.gguf",
+      "url": "https://huggingface.co/janhq/pandora-v1-10.7b-GGUF/resolve/main/pandora-v1-10.7b.Q4_K_M.gguf"
+    }
+  ],
+  "id": "pandora-10.7b-v1",
+  "object": "model",
+  "name": "Pandora 11B Q4",
+  "version": "1.0",
+  "description": "Pandora, our research model, employs the Passthrough merging technique to merge 2x7B models into 1.",
+  "format": "gguf",
+  "settings": {
+    "ctx_len": 4096,
+    "prompt_template": "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant",
+    "llama_model_path": "pandora-v1-10.7b.Q4_K_M.gguf"
+  },
+  "parameters": {
+    "max_tokens": 4096
+  },
+  "metadata": {
+    "author": "Jan",
+    "tags": ["11B", "Finetuned", "Featured"],
+    "size": 6360000000,
+    "cover": "https://raw.githubusercontent.com/janhq/jan/main/models/pandora-10.7b-v1/cover.png"
+  },
+  "engine": "nitro"
+}
diff --git a/models/solar-10.7b-instruct/model.json b/models/solar-10.7b-instruct/model.json
@@ -1,23 +1,28 @@
 {
-    "source_url": "https://huggingface.co/janhq/Solar-10.7B-Instruct-v1.0/resolve/main/solar-10.7b-instruct-v1.0.Q4_K_M.gguf",
-    "id": "solar-10.7b-instruct",
-    "object": "model",
-    "name": "Solar Instruct 10.7B Q4",
-    "version": "1.0",
-    "description": "SOLAR-10.7B model built on the Llama2 architecture with Depth Up-Scaling and integrated Mistral 7B weights. Its robustness and adaptability make it ideal for fine-tuning applications, significantly enhancing performance with simple instruction-based techniques.",
-    "format": "gguf",
-    "settings": {
-      "ctx_len": 4096,
-      "prompt_template": "### User: {prompt}\n### Assistant:"
-    },
-    "parameters": {
-      "max_tokens": 4096
-    },
-    "metadata": {
-      "author": "Upstage, Jan",
-      "tags": ["11B","Finetuned"],
-      "size": 6360000000
-    },
-    "engine": "nitro"
-  }
-
+  "source": [
+    {
+      "filename": "solar-10.7b-instruct-v1.0.Q4_K_M.gguf",
+      "url": "https://huggingface.co/janhq/Solar-10.7B-Instruct-v1.0/resolve/main/solar-10.7b-instruct-v1.0.Q4_K_M.gguf"
+    }
+  ],
+  "id": "solar-10.7b-instruct",
+  "object": "model",
+  "name": "Solar Instruct 10.7B Q4",
+  "version": "1.0",
+  "description": "SOLAR-10.7B model built on the Llama2 architecture with Depth Up-Scaling and integrated Mistral 7B weights. Its robustness and adaptability make it ideal for fine-tuning applications, significantly enhancing performance with simple instruction-based techniques.",
+  "format": "gguf",
+  "settings": {
+    "ctx_len": 4096,
+    "prompt_template": "### User: {prompt}\n### Assistant:",
+    "llama_model_path": "solar-10.7b-instruct-v1.0.Q4_K_M.gguf"
+  },
+  "parameters": {
+    "max_tokens": 4096
+  },
+  "metadata": {
+    "author": "Upstage, Jan",
+    "tags": ["11B", "Finetuned"],
+    "size": 6360000000
+  },
+  "engine": "nitro"
+}
diff --git a/models/trinity-v1-7b/model.json b/models/trinity-v1-7b/model.json
@@ -1,23 +1,29 @@
 {
-    "source_url": "https://huggingface.co/janhq/trinity-v1-GGUF/resolve/main/trinity-v1.Q4_K_M.gguf",
-    "id": "trinity-v1-7b",
-    "object": "model",
-    "name": "Trinity 7B Q4",
-    "version": "1.0",
-    "description": "Trinity is an experimental model merge of GreenNodeLM & LeoScorpius using the Slerp method. Recommended for daily assistance purposes.",
-    "format": "gguf",
-    "settings": {
-      "ctx_len": 4096,
-      "prompt_template": "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant"
-    },
-    "parameters": {
-      "max_tokens": 4096
-    },
-    "metadata": {
-      "author": "Jan",
-      "tags": ["7B", "Merged", "Featured"],
-      "size": 4370000000,
-      "cover": "https://raw.githubusercontent.com/janhq/jan/main/models/trinity-v1-7b/cover.png"
-    },
-    "engine": "nitro"
-  }  
+  "source": [
+    {
+      "filename": "trinity-v1.Q4_K_M.gguf",
+      "url": "https://huggingface.co/janhq/trinity-v1-GGUF/resolve/main/trinity-v1.Q4_K_M.gguf"
+    }
+  ],
+  "id": "trinity-v1-7b",
+  "object": "model",
+  "name": "Trinity 7B Q4",
+  "version": "1.0",
+  "description": "Trinity is an experimental model merge of GreenNodeLM & LeoScorpius using the Slerp method. Recommended for daily assistance purposes.",
+  "format": "gguf",
+  "settings": {
+    "ctx_len": 4096,
+    "prompt_template": "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant",
+    "llama_model_path": "trinity-v1.Q4_K_M.gguf"
+  },
+  "parameters": {
+    "max_tokens": 4096
+  },
+  "metadata": {
+    "author": "Jan",
+    "tags": ["7B", "Merged", "Featured"],
+    "size": 4370000000,
+    "cover": "https://raw.githubusercontent.com/janhq/jan/main/models/trinity-v1-7b/cover.png"
+  },
+  "engine": "nitro"
+}