Add functiongemma model card #65
Conversation
Summary of ChangesHello @ilopezluna, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request expands the model card collection by introducing documentation for the FunctionGemma model in two variants: one for vLLM and another for the GGUF version by Unsloth. These additions provide detailed information on the model's architecture, intended applications, and integration with Docker Model Runner, enhancing the discoverability and usability of this specialized function-calling AI model. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey - I've left some high level feedback:
- The two model card files duplicate the same long-form description; consider extracting shared content or clearly highlighting only the differences (e.g., GGUF vs vLLM) to make maintenance easier and reduce the risk of divergence.
- Only the GGUF variant mentions Unsloth explicitly and links to the Unsloth GGUF docs, while the vLLM card omits any vLLM-specific link or explanation; aligning both cards so each clearly calls out its serving/runtime specifics would make the distinction between them clearer.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- The two model card files duplicate the same long-form description; consider extracting shared content or clearly highlighting only the differences (e.g., GGUF vs vLLM) to make maintenance easier and reduce the risk of divergence.
- Only the GGUF variant mentions Unsloth explicitly and links to the Unsloth GGUF docs, while the vLLM card omits any vLLM-specific link or explanation; aligning both cards so each clearly calls out its serving/runtime specifics would make the distinction between them clearer.Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request adds model cards for two variants of FunctionGemma: a vLLM version and a GGUF version. The cards provide a good overview of the model. However, I've found a few critical and high-severity issues that need to be addressed.
- There's a factual error regarding the model's context window, which is 8K tokens, not 32K as stated.
- The base model is incorrectly cited as 'Gemma 3' instead of 'Gemma 2'.
- There are also some minor typos and opportunities to improve readability by hyphenating compound words and breaking up a very long paragraph.
I've provided specific suggestions to correct these issues in both files. Addressing these will ensure the model cards are accurate and easy to read.
| ## Inputs and outputs | ||
| Input: | ||
| - Text string, such as a question, a prompt, or a document to be summarized | ||
| - Total input context of 32K tokens |
There was a problem hiding this comment.
The context window for FunctionGemma is 8K tokens, not 32K. Please correct this value. The official documentation states an 8,192 token context window.
| - Total input context of 32K tokens | |
| - Total input context of 8K tokens |
|
|
||
| Output: | ||
| - Generated text in response to the input, such as an answer to a question, or a summary of a document | ||
| - Total output context up to 32K tokens per request, subtracting the request input tokens |
There was a problem hiding this comment.
The total output context should also reflect the 8K token limit. The maximum output is 8K minus the input tokens.
| - Total output context up to 32K tokens per request, subtracting the request input tokens | |
| - Total output context up to 8K tokens per request, subtracting the request input tokens |
| ## Inputs and outputs | ||
| Input: | ||
| - Text string, such as a question, a prompt, or a document to be summarized | ||
| - Total input context of 32K tokens |
There was a problem hiding this comment.
The context window for FunctionGemma is 8K tokens, not 32K. Please correct this value to align with the official model specifications, which state an 8,192 token context window.
| - Total input context of 32K tokens | |
| - Total input context of 8K tokens |
|
|
||
| Output: | ||
| - Generated text in response to the input, such as an answer to a question, or a summary of a document | ||
| - Total output context up to 32K tokens per request, subtracting the request input tokens |
There was a problem hiding this comment.
The total output context should also reflect the 8K token limit. The maximum output is 8K minus the number of input tokens.
| - Total output context up to 32K tokens per request, subtracting the request input tokens | |
| - Total output context up to 8K tokens per request, subtracting the request input tokens |
|
|
||
| ## Intended uses | ||
|
|
||
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 3 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 3, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be finetuned on single turn or multiturn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. |
There was a problem hiding this comment.
This paragraph has some opportunities for improvement in accuracy, consistency, and readability.
- Accuracy: FunctionGemma is based on Gemma 2, not Gemma 3. This is mentioned twice and should be corrected.
- Typos: 'state of a art' should be 'state of the art'.
- Consistency: For better readability, terms like 'finetuned', 'single turn', and 'multiturn' should be hyphenated ('fine-tuned', 'single-turn', 'multi-turn').
- Readability: The paragraph is very long. Consider splitting it into smaller paragraphs to make it easier to digest.
I've provided a code suggestion that fixes the accuracy and typo issues. You may want to further break up the paragraph for readability.
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 3 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 3, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be finetuned on single turn or multiturn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. | |
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 2 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 2, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be fine-tuned on single-turn or multi-turn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. |
|
|
||
| ## Intended uses | ||
|
|
||
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 3 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 3, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be finetuned on single turn or multiturn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. |
There was a problem hiding this comment.
This paragraph has some opportunities for improvement in accuracy, consistency, and readability.
- Accuracy: FunctionGemma is based on Gemma 2, not Gemma 3. This is mentioned twice and should be corrected.
- Typos: 'state of a art' should be 'state of the art'.
- Consistency: For better readability, terms like 'finetuned', 'single turn', and 'multurn' should be hyphenated ('fine-tuned', 'single-turn', 'multi-turn').
- Readability: The paragraph is very long. Consider splitting it into smaller paragraphs to make it easier to digest.
I've provided a code suggestion that fixes the accuracy and typo issues. You may want to further break up the paragraph for readability.
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 3 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 3, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be finetuned on single turn or multiturn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. | |
| FunctionGemma is a lightweight, open model from Google, built as a foundation for creating your own specialized function calling models. FunctionGemma is not intended for use as a direct dialogue model, and is designed to be highly performant after further fine-tuning, as is typical of models this size. Built on the Gemma 2 270M model and with the same research and technology used to create the Gemini models, FunctionGemma has been trained specifically for function calling. The model has the same architecture as Gemma 2, but uses a different chat format. The model is well suited for text-only function calling. The uniquely small size makes it possible to deploy in environments with limited resources such as laptops, desktops or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. Furthermore, akin to the base Gemma 270M, the model has been optimized to be extremely versatile, performant on a variety of hardware in single turn scenarios, but should be fine-tuned on single-turn or multi-turn task specific data to achieve best accuracy in specific domains. To demonstrate how specializing the 270M parameter model can achieve high performance on specific agentic workflows, we have highlighted two use cases in the Google AI Edge Gallery app. |
No description provided.