Merge pull request #70 from lmstudio-ai/google-gemma

Google gemma

models/google-gemma-2b.json

@@ -0,0 +1,42 @@
+{
+    "_descriptorVersion": "0.0.1",  
+    "datePublished": "2024-02-21T16:54:57.000Z",
+    "name": "Google's Gemma 2B Instruct",
+    "description": "** Requires LM Studio 0.2.15 or newer ** Gemma is a family of lightweight LLMs built from the same research and technology Google used to create the Gemini models. Gemma models are available in two sizes, 2 billion and 7 billion parameters. These models are trained on up to 6T tokens of primarily English web documents, mathematics, and code, using a transformer architecture with enhancements like Multi-Query Attention, RoPE Embeddings, GeGLU Activations, and advanced normalization techniques.",
+    "author": {
+      "name": "Google DeepMind",
+      "url": "https://deepmind.google",
+      "blurb": "We’re a team of scientists, engineers, ethicists and more, working to build the next generation of AI systems safely and responsibly."
+    },
+    "numParameters": "2B",
+    "resources": {
+      "canonicalUrl": "https://huggingface.co/google/gemma-2b-it",
+      "paperUrl": "https://blog.google/technology/developers/gemma-open-models/",
+      "downloadUrl": "https://huggingface.co/lmstudio-ai/gemma-2b-it-GGUF"
+    },
+    "trainedFor": "chat",
+    "arch": "gemma",
+    "files": {
+      "highlighted": {
+        "economical": {
+          "name": "gemma-2b-it-q8_0.gguf"
+        }
+      },
+      "all": [
+        {
+          "name": "gemma-2b-it-q8_0.gguf",
+          "url": "https://huggingface.co/lmstudio-ai/gemma-2b-it-GGUF/resolve/main/gemma-2b-it-q8_0.gguf",
+          "sizeBytes": 2669351840,
+          "quantization": "Q8_0",
+          "format": "gguf",
+          "sha256checksum": "ec68b50d23469882716782da8b680402246356c3f984e9a3b9bcc5bc15273140",
+          "publisher": {
+            "name": "LM Studio",
+            "socialUrl": "https://twitter.com/LMStudioAI"
+          },
+          "respository": "lmstudio-ai/gemma-2b-it-GGUF",
+          "repositoryUrl": "https://huggingface.co/lmstudio-ai/gemma-2b-it-GGUF"
+        }
+      ]
+    }
+  }

models/qwen-1.5.json

@@ -1,6 +1,6 @@
 {
     "_descriptorVersion": "0.0.1",  
-    "datePublished": "2024-02-03T11:59:54",
+    "datePublished": "2024-02-03T16:59:54.000Z",
     "name": "Qwen 1.5",
     "description": "Qwen1.5 is the large language model series developed by Qwen Team, Alibaba Group. It is a transformer-based decoder-only language model pretrained on large-scale multilingual data covering a wide range of domains and it is aligned with human preferences.",
     "author": {

schema.json

@@ -51,15 +51,15 @@
       },
       "numParameters": {
         "type": "string",
-        "enum": ["1.5B", "3B", "4B", "6.7B", "7B", "13B", "15B", "30B", "65B", "unknown"]
+        "enum": ["1.5B", "2B", "3B", "4B", "6.7B", "7B", "13B", "15B", "30B", "65B", "unknown"]
       },
       "trainedFor": {
         "type": "string",
         "enum": ["chat", "instruct", "code_completion", "other"]
       },
       "arch": {
         "type": "string",
-        "enum": ["llama", "pythia", "gpt-neo-x", "gpt-j", "mpt", "replit", "starcoder", "falcon", "mistral", "stablelm", "phi2", "qwen2"]
+        "enum": ["llama", "pythia", "gpt-neo-x", "gpt-j", "mpt", "replit", "starcoder", "falcon", "mistral", "stablelm", "phi2", "qwen2", "gemma"]
       },
       "description": {
         "type": "string"