Template fix: Preserve anyOf/$ref/$defs in Gemma tool declarations

.eval_results/mmmu_pro.yaml

@@ -1,8 +0,0 @@
-- dataset:
-    id: MMMU/MMMU_Pro
-    task_id: mmmu_pro_vision
-  value: 76.9
-  date: '2026-05-12'
-  source:
-    url: https://huggingface.co/google/gemma-4-31B-it
-    name: Model Card

README.md

@@ -3,8 +3,6 @@ library_name: transformers
 license: apache-2.0
 license_link: https://ai.google.dev/gemma/docs/gemma_4_license
 pipeline_tag: image-text-to-text
-base_model:
-- google/gemma-4-31B
 ---
 
 <div align="center">
@@ -200,13 +198,13 @@ Once the model is loaded, you can start generating output by directly referencin
 
 
 ```python
-# Prompt - add audio after text
+# Prompt - add audio before text
 messages = [
     {
         "role": "user",
         "content": [
+            {"type": "audio", "audio": "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/Demos/sample-data/journal1.wav"},
             {"type": "text", "text": "Transcribe the following speech segment in its original language. Follow these specific instructions for formatting the answer:\n* Only output the transcription, with no newlines.\n* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three."},
-            {"type": "audio", "audio": "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/apps/sample-data/journal1.wav"},
         ]
     }
 ]
@@ -263,7 +261,7 @@ Once the model is loaded, you can start generating output by directly referencin
 messages = [
     {
         "role": "user", "content": [
-            {"type": "image", "url": "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/apps/sample-data/GoldenGate.png"},
+            {"type": "image", "url": "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/Demos/sample-data/GoldenGate.png"},
             {"type": "text", "text": "What is shown in this image?"}
         ]
     }
@@ -380,10 +378,7 @@ Compared to Gemma 3, the models use standard `system`, `assistant`, and `user` r
 
 ### 4. Modality order
 
-For optimal performance with multimodal inputs, place:
-
-* Image content **before** the text in your prompt.
-* Audio content **after** the text in your prompt.
+* For optimal performance with multimodal inputs, place image and/or audio content **before** the text in your prompt. 
 
 ### 5. Variable Image Resolution
 
@@ -515,4 +510,4 @@ The development of vision-language models (VLMs) raises several ethical concerns
 
 ### **Benefits**
 
-At the time of release, this family of models provides high-performance open vision-language model implementations designed from the ground up for responsible AI development compared to similarly sized models.
+At the time of release, this family of models provides high-performance open vision-language model implementations designed from the ground up for responsible AI development compared to similarly sized models.

chat_template.jinja

@@ -295,15 +295,6 @@
                                 {%- endif -%}
                             {%- endfor -%}
                             {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
-                            {%- for part in tool_body -%}
-                                {%- if part.get('type') == 'image' -%}
-                                    {{- '<|image|>' -}}
-                                {%- elif part.get('type') == 'audio' -%}
-                                    {{- '<|audio|>' -}}
-                                {%- elif part.get('type') == 'video' -%}
-                                    {{- '<|video|>' -}}
-                                {%- endif -%}
-                            {%- endfor -%}
                         {%- else -%}
                             {{- format_tool_response_block(ns_tname.name, tool_body) -}}
                         {%- endif -%}