preliminary embeddinggemma support

Author: BBC-Esq

src/constants.py

@@ -324,7 +324,7 @@
     "timm==1.0.19",
     "tokenizers==0.22.0",
     "tqdm==4.67.1",
-    "transformers==4.56.0",
+    "transformers==4.56.1",
     "typing-inspection==0.4.1", # required by pydantic and pydantic-settings
     "typing_extensions==4.15.0",
     "unstructured-client==0.42.3",
@@ -592,7 +592,7 @@
 VECTOR_MODELS = {
     'BAAI': [
         {
-            'name': 'bge-small-en-v1.5',# transformers 4.30.0
+            'name': 'bge-small-en-v1.5',
             'dimensions': 384,
             'max_sequence': 512,
             'size_mb': 134,
@@ -627,22 +627,24 @@
             'precision': 'float32',
             'license': 'mit',
         },
+    ],
+    # 'Google': [
         # {
-            # 'name': 'bge-code-v1',# transformers 4.49.0
-            # 'dimensions': 1536,
-            # 'max_sequence': 4096,
-            # 'size_mb': 1340,
-            # 'repo_id': 'BAAI/bge-code-v1',
-            # 'cache_dir': 'BAAI--bge-code-v1',
+            # 'name': 'embeddinggemma-300m',
+            # 'dimensions': 768,
+            # 'max_sequence': 2048,
+            # 'size_mb': 1210,
+            # 'repo_id': 'google/embeddinggemma-300m',
+            # 'cache_dir': 'google--embeddinggemma-300m',
             # 'type': 'vector',
-            # 'parameters': '1540m',
+            # 'parameters': '303m',
             # 'precision': 'float32',
-            # 'license': 'apache-2.0',
+            # 'license': 'gemma - commercial ok',
         # },
-    ],
+    # ],
     'IBM': [
         {
-            'name': 'Granite-30m-English',# transformers 4.38.2
+            'name': 'Granite-30m-English',
             'dimensions': 384,
             'max_sequence': 512,
             'size_mb': 61,
@@ -668,7 +670,7 @@
     ],
     'infly': [
         {
-            'name': 'inf-retriever-v1-1.5b',# transformers 4.48.1
+            'name': 'inf-retriever-v1-1.5b',
             'dimensions': 1536,
             'max_sequence': 8192,
             'size_mb': 3090,
@@ -680,7 +682,7 @@
             'license': 'apache-2.0',
         },
         {
-            'name': 'inf-retriever-v1-7b',# transformers 4.44.2
+            'name': 'inf-retriever-v1-7b',
             'dimensions': 3584,
             'max_sequence': 8192,
             'size_mb': 14130,
@@ -694,7 +696,7 @@
     ],
     'intfloat': [
         {
-            'name': 'e5-small-v2',# 4.29.0.dev0
+            'name': 'e5-small-v2',
             'dimensions': 384,
             'max_sequence': 512,
             'size_mb': 134,
@@ -732,7 +734,7 @@
     ],
     'Qwen': [
         {
-            'name': 'Qwen3-Embedding-0.6B',# transformers 4.51.3
+            'name': 'Qwen3-Embedding-0.6B',
             'dimensions': 1024,
             'max_sequence':8192,
             'size_mb': 1190,
@@ -770,7 +772,7 @@
     ],
     'Snowflake': [
         {
-            'name': 'arctic-embed-m-v2.0',# transformers 4.39.3
+            'name': 'arctic-embed-m-v2.0',
             'dimensions': 768,
             'max_sequence':8192,
             'size_mb': 1220,
@@ -1561,11 +1563,6 @@
         "folder": "cuDF",
         "scraper_class": "PydataThemeScraper"
     },
-    # "CuPy": {
-        # "URL": "https://docs.cupy.dev/en/stable/",
-        # "folder": "cupy",
-        # "scraper_class": "PydataThemeScraper"
-    # },
     "cycler": {
         "URL": "https://matplotlib.org/cycler/",
         "folder": "cycler"
@@ -1891,11 +1888,6 @@
         "folder": "platformdirs",
         "scraper_class": "FuroThemeScraper"
     },
-    # "Playwright": {
-        # "URL": "https://playwright.dev/python/",
-        # "folder": "playwright",
-        # "scraper_class": "DivClassThemeDocMarkdownMarkdownScraper"
-    # },
     "pluggy": {
         "URL": "https://pluggy.readthedocs.io/en/stable/",
         "folder": "pluggy",
@@ -3159,7 +3151,6 @@ def _generate_button_style(cls, color_values):
 system_message = "You are a helpful person who clearly and directly answers questions in a succinct fashion based on contexts provided to you. If you cannot find the answer within the contexts simply tell me that the contexts do not provide an answer. However, if the contexts partially address my question I still want you to answer based on what the contexts say and then briefly summarize the parts of my question that the contexts didn't provide an answer."
 rag_string = "Here are the contexts to base your answer on.  However, I need to reiterate that I only want you to base your response on these contexts and do not use outside knowledge that you may have been trained with."
 
-
 r"""
 
 ====================================================================
@@ -3326,23 +3317,20 @@ def _generate_button_style(cls, color_values):
 * Triton 3.2 works with PyTorch >= 2.6 .
 * Triton 3.1 works with PyTorch >= 2.4 . PyTorch 2.3 and older are not supported.
 
-
 ************
 cuDNN & CUDA
 ************
 
 # Nvidia promises that all cuDNN 9+ releases are compatible with all CUDA 12.x releases.
 # However, certain version of the torch library are built/tested with certain versions of cuDNN.  Doesn't always mean incompatibility.
 
-
 ***********************
 LINUX Flash Attention 2
 ***********************
 
 # HIGHLY CUDA SPECIFIC
 # Obtained from https://github.com/Dao-AILab/flash-attention/blob/main/.github/workflows/publish.yml
 
-
 *************************
 WINDOWS Flash Attention 2
 *************************
@@ -3358,7 +3346,6 @@ def _generate_button_style(cls, color_values):
 +--------------+-----------------------------------------+------------+
 # Windows wheels: https://github.com/kingbri1/flash-attention
 
-
 ********
 Xformers
 ********
@@ -3382,7 +3369,6 @@ def _generate_button_style(cls, color_values):
 * FA2 support determined by by https://github.com/facebookresearch/xformers/blob/main/xformers/ops/fmha/flash.py
 * CUDA support determined by https://github.com/facebookresearch/xformers/blob/main/.github/actions/setup-build-cuda/action.yml
 
-
 ***************
 **CTRANSLATE2**
 ***************