Add gemma2 to available models

[prodrigestivill]

Add gemma2 to available models.

Keep in mind that to run this model is required ollama version 0.1.47 but is preferable version 0.1.48.

Because of the JSON lint used in available_models.json, I detail here the new object added here in case you prefer to do the change in any other way:

{
  "gemma2": {
    "url": "https://ollama.com/library/gemma2",
    "description": "Google Gemma 2 is now available in 2 sizes, 9B and 27B.",
    "tags": [
      ["latest", "5.4GB"],
      ["27b", "16GB"],
      ["9b", "5.4GB"],
      ["27b-instruct-fp16", "54GB"],
      ["27b-instruct-q2_K", "10GB"],
      ["27b-instruct-q3_K_L", "15GB"],
      ["27b-instruct-q3_K_M", "13GB"],
      ["27b-instruct-q3_K_S", "12GB"],
      ["27b-instruct-q4_0", "16GB"],
      ["27b-instruct-q4_1", "17GB"],
      ["27b-instruct-q4_K_M", "17GB"],
      ["27b-instruct-q4_K_S", "16GB"],
      ["27b-instruct-q5_0", "19GB"],
      ["27b-instruct-q5_1", "21GB"],
      ["27b-instruct-q5_K_M", "19GB"],
      ["27b-instruct-q5_K_S", "19GB"],
      ["27b-instruct-q6_K", "22GB"],
      ["27b-instruct-q8_0", "29GB"],
      ["9b-instruct-fp16", "18GB"],
      ["9b-instruct-q2_K", "3.8GB"],
      ["9b-instruct-q3_K_L", "5.1GB"],
      ["9b-instruct-q3_K_M", "4.8GB"],
      ["9b-instruct-q3_K_S", "4.3GB"],
      ["9b-instruct-q4_0", "5.4GB"],
      ["9b-instruct-q4_1", "6.0GB"],
      ["9b-instruct-q4_K_M", "5.8GB"],
      ["9b-instruct-q4_K_S", "5.5GB"],
      ["9b-instruct-q5_0", "6.5GB"],
      ["9b-instruct-q5_1", "7.0GB"],
      ["9b-instruct-q5_K_M", "6.6GB"],
      ["9b-instruct-q5_K_S", "6.5GB"],
      ["9b-instruct-q6_K", "7.6GB"],
      ["9b-instruct-q8_0", "9.8GB"]
    ],
    "image": false,
    "author": "Google DeepMind"
  }
}

[prodrigestivill]

Just saw this issue from yesterday #119. As I understand this PR is of no use since you use an external script to generate the JSON.

[Jeffser]

Hi, Google Gemma 2 will be supported on the next release, like you said I do use an external script to generate the JSON, so I will be using that just for consistency, so I won't be merging this