Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.spm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,47 @@
+# Gemma AI Models
+
+This directory contains the downloaded Gemma 3N (2B parameters) model files required for the M.AI application's on-device AI features.
+
+## Included Files
+
+1.  **`gemma-1.1-2b-it-gpu-int4.bin`** (~1.35 GB)
+    *   **Description:** The main Large Language Model, instruction-tuned and quantized (INT4) for efficient GPU inference on mobile devices.
+    *   **Source:** Google's Gemma 1.1 2B IT TFLite repository (Hugging Face).
+    *   **SHA-256:** `53f7defdb5554dd517a6863d2d605a144d14361d6c486bbeaac2870497b75747`
+
+2.  **`tokenizer.spm`** (~4.2 MB)
+    *   **Description:** The SentencePiece tokenizer model used to process text input before sending it to the LLM.
+    *   **Source:** Google's Gemma 2B repository (Hugging Face).
+    *   **SHA-256:** `61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2`
+
+## 🚀 Next Steps: Hosting & Configuration
+
+Since these models are gated by Google, you **cannot** use direct Hugging Face links in your production app. You must self-host them.
+
+### 1. Upload to Hosting Service
+Upload both files to a public file host. Recommended options:
+*   **GitHub Releases:** Create a release in your repo (e.g., `v1.0-models`) and attach these files. (Free, fast, public).
+*   **Firebase Storage:** Upload to your project's storage bucket.
+*   **AWS S3 / Cloudflare R2:** Standard object storage.
+
+### 2. Get Public URLs
+Obtain the direct download link for each file.
+*   *Test:* Running `curl -I <LINK>` should return `200 OK` and the correct file size (not a login page).
+
+### 3. Update App Configuration
+Open `apps/ai_apps/maiv2/lib/core/config/environment_config.dart` and update the `production` configuration:
+
+```dart
+static const EnvironmentConfig production = EnvironmentConfig(
+  // ... other config ...
+  gemmaModelUrl: 'https://your-host.com/path/to/gemma-1.1-2b-it-gpu-int4.bin',
+  gemmaTokenizerUrl: 'https://your-host.com/path/to/tokenizer.spm',
+  
+  // Use the SHA-256 checksums listed above
+  gemmaModelChecksum: '53f7defdb5554dd517a6863d2d605a144d14361d6c486bbeaac2870497b75747',
+  gemmaTokenizerChecksum: '61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2',
+);
+```
+
+### 4. Verify
+Run the app in release mode. The "Download AI Model" button in the chat interface will now fetch these specific files and verify their integrity using the checksums.

gemma-1.1-2b-it-gpu-int4.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53f7defdb5554dd517a6863d2d605a144d14361d6c486bbeaac2870497b75747
+size 1346502656

tokenizer.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2
+size 4241003