# \#fine-tuning **URL:** https://discuss.ai.google.dev/tag/fine-tuning/9.md [Latest](https://discuss.ai.google.dev/latest.md) · [Categories](https://discuss.ai.google.dev/categories.md) · [Tags](https://discuss.ai.google.dev/tags.md) --- ## [ATRIA | EchoTrace: Fine-Tuning MedGemma 1.5 for Polygon-Based Heart Structure Contouring](https://discuss.ai.google.dev/t/atria-echotrace-fine-tuning-medgemma-1-5-for-polygon-based-heart-structure-contouring/172907)
**Author:** [@ShehabAnwer](https://discuss.ai.google.dev/u/ShehabAnwer)\ **Replies:** 4\ **Last updated:** [October 11, 2026, 2:30pm UTC](https://discuss.ai.google.dev/t/atria-echotrace-fine-tuning-medgemma-1-5-for-polygon-based-heart-structure-contouring/172907 "2026-10-11T14:30:42Z")
Esteemed HAI-DEF & MedGemma Team, I’m Shehab Anwer, a cardiologist. I recently shared some of my progress with Rory Pilgrim, who kindly suggested I post here to connect with the right folks for your valuable insights an… --- ## [Google Cloud Agent Platform Gemma 4 fine-tuning ETA](https://discuss.ai.google.dev/t/google-cloud-agent-platform-gemma-4-fine-tuning-eta/170582)
**Author:** [@yogayoga](https://discuss.ai.google.dev/u/yogayoga)\ **Replies:** 0\ **Last updated:** [June 10, 2026, 8:38am UTC](https://discuss.ai.google.dev/t/google-cloud-agent-platform-gemma-4-fine-tuning-eta/170582 "2026-06-10T08:38:14Z")
Hi, Is there an expected ETA on Gemma 4 fine-tuning availability on Google Cloud’s Agent Platform? We are specifically interested in fine-tuning the Gemma 4 12B model Thanks, --- ## [Request: Allowlist Access for Gemini 3.1 Fine-Tuning on Vertex AI (Google for Startups Member)](https://discuss.ai.google.dev/t/request-allowlist-access-for-gemini-3-1-fine-tuning-on-vertex-ai-google-for-startups-member/136120)
**Author:** [@Austin\_Ollar](https://discuss.ai.google.dev/u/Austin_Ollar)\ **Replies:** 5\ **Last updated:** [May 8, 2026, 1:32pm UTC](https://discuss.ai.google.dev/t/request-allowlist-access-for-gemini-3-1-fine-tuning-on-vertex-ai-google-for-startups-member/136120 "2026-05-08T13:32:19Z")
Hi team, I’m the founder of TypeScience, an AI personality-classification platform, and a current Google for Startups Cloud Program member. We’re actively building on Vertex AI and have already completed a supervised f… --- ## [Supervised fine-tuned Gemini 2.5 Pro endpoints — what happens after Oct 17, 2026 shutdown?](https://discuss.ai.google.dev/t/supervised-fine-tuned-gemini-2-5-pro-endpoints-what-happens-after-oct-17-2026-shutdown/137946)
**Author:** [@J2H](https://discuss.ai.google.dev/u/J2H)\ **Replies:** 2\ **Last updated:** [May 8, 2026, 1:30pm UTC](https://discuss.ai.google.dev/t/supervised-fine-tuned-gemini-2-5-pro-endpoints-what-happens-after-oct-17-2026-shutdown/137946 "2026-05-08T13:30:37Z")
I run a production service using supervised fine-tuned Gemini 2.5 Pro endpoints on Vertex AI. The deprecation page shows gemini-2.5-pro shutting down on Oct 17, 2026. I’d like to request clarification on two points: … --- ## [Can Gemini LLM Be Fine-Tuned Locally on Private Servers for Sensitive Data?](https://discuss.ai.google.dev/t/can-gemini-llm-be-fine-tuned-locally-on-private-servers-for-sensitive-data/51896)
**Author:** [@Zahra\_Hosseini](https://discuss.ai.google.dev/u/Zahra_Hosseini)\ **Replies:** 2\ **Last updated:** [March 12, 2026, 1:14pm UTC](https://discuss.ai.google.dev/t/can-gemini-llm-be-fine-tuned-locally-on-private-servers-for-sensitive-data/51896 "2026-03-12T13:14:54Z")
Hello, community! We work with private data that must remain on our local servers due to strict data privacy regulations. We’re exploring the possibility of using Gemini for fine-tuning our models but need to ensure tha… --- ## [Discontinuation applies for fine-tuned models?](https://discuss.ai.google.dev/t/discontinuation-applies-for-fine-tuned-models/115760)
**Author:** [@Manuel\_Cojulun](https://discuss.ai.google.dev/u/Manuel_Cojulun)\ **Replies:** 0\ **Last updated:** [January 16, 2026, 6:58am UTC](https://discuss.ai.google.dev/t/discontinuation-applies-for-fine-tuned-models/115760 "2026-01-16T06:58:35Z")
Gemini 2.5 Flash has a discontinuation date setted to June 17, 2026. If I fine tune Gemini 2.5 Flash, will I stop having access to that fine tuned model on June 17, 2026? --- ## [maxLength not respected by schema](https://discuss.ai.google.dev/t/maxlength-not-respected-by-schema/114526)
**Author:** [@Aaryan\_B](https://discuss.ai.google.dev/u/Aaryan_B)\ **Replies:** 4\ **Last updated:** [January 12, 2026, 11:55am UTC](https://discuss.ai.google.dev/t/maxlength-not-respected-by-schema/114526 "2026-01-12T11:55:48Z")
Hey everyone, Im using a finetuned endpoint of gemin 2.5 flash lite and im facing an issue, In my prompt i ask it to do some thinking to get a better result. Now i have kepty a limit on the max char in the thinking st… --- ## [Video tuning gemini-2.5-flash fails for larger train datasets](https://discuss.ai.google.dev/t/video-tuning-gemini-2-5-flash-fails-for-larger-train-datasets/109633)
**Author:** [@Venkat\_31](https://discuss.ai.google.dev/u/Venkat_31)\ **Replies:** 7\ **Last updated:** [January 6, 2026, 6:26pm UTC](https://discuss.ai.google.dev/t/video-tuning-gemini-2-5-flash-fails-for-larger-train-datasets/109633 "2026-01-06T18:26:45Z")
Hi, I am fine-tuning gemini-2.5-flash with video data. My fine-tuning job say v0 succeeded without any errors for a dataset with over 2800 videos (7235244 tokens). But subsequent runs with the same training and validatio… --- ## [Can Fine tuning with Gemini API be brought back?](https://discuss.ai.google.dev/t/can-fine-tuning-with-gemini-api-be-brought-back/109382)
**Author:** [@Aijaz\_Qureshi](https://discuss.ai.google.dev/u/Aijaz_Qureshi)\ **Replies:** 2\ **Last updated:** [November 19, 2025, 4:12pm UTC](https://discuss.ai.google.dev/t/can-fine-tuning-with-gemini-api-be-brought-back/109382 "2025-11-19T16:12:49Z")
I have been playing with the Kaggle Jupyter Notebooks that refer to training the model/gemini-1.5-flash-001 model, but the model is taken away and is not available for local training. Since training a model is very impor… --- ## [Half the Battle Is A Masterful Scaffold](https://discuss.ai.google.dev/t/half-the-battle-is-a-masterful-scaffold/106655)
**Author:** [@zoid](https://discuss.ai.google.dev/u/zoid)\ **Replies:** 5\ **Last updated:** [October 4, 2025, 9:22am UTC](https://discuss.ai.google.dev/t/half-the-battle-is-a-masterful-scaffold/106655 "2025-10-04T09:22:49Z")
The Emergence of Structured Reasoning: Inducing Metacognitive‑Like Capabilities via Scaffold‑Driven Fine‑Tuning Author: Zachary Robert Bennett Abstract We present Scaffold‑Driven Fine‑Tuning (SDFT), a method that fine‑t… --- ## [VertexAI API does not provide ability to set FPS for videos when fine-tuning models](https://discuss.ai.google.dev/t/vertexai-api-does-not-provide-ability-to-set-fps-for-videos-when-fine-tuning-models/105133)
**Author:** [@Darien\_Ford](https://discuss.ai.google.dev/u/Darien_Ford)\ **Replies:** 1\ **Last updated:** [September 19, 2025, 3:51am UTC](https://discuss.ai.google.dev/t/vertexai-api-does-not-provide-ability-to-set-fps-for-videos-when-fine-tuning-models/105133 "2025-09-19T03:51:38Z")
I’m trying to fine-tune gemini-2.5-flash to do better video analysis. The tuning parameters are very limited and do not include the videoMetadata options that a normal inference would have: This means the model is miss… --- ## [Fine-tuning Gemini models via Gemini API / Hugging Face / Unsloth](https://discuss.ai.google.dev/t/fine-tuning-gemini-models-via-gemini-api-hugging-face-unsloth/105403)
**Author:** [@Charon](https://discuss.ai.google.dev/u/Charon)\ **Replies:** 1\ **Last updated:** [September 19, 2025, 3:43am UTC](https://discuss.ai.google.dev/t/fine-tuning-gemini-models-via-gemini-api-hugging-face-unsloth/105403 "2025-09-19T03:43:45Z")
Hi experts, Currently, the Genuine Gemini API does not clarify when the fine-tuning APIs with theoretical explanations and methodologies for specific models will be provided. In VertexAI, the some HTTP endpoints, which… --- ## [2.5 flash lite - fine tuning](https://discuss.ai.google.dev/t/2-5-flash-lite-fine-tuning/95206)
**Author:** [@Yash\_Shukla](https://discuss.ai.google.dev/u/Yash_Shukla)\ **Replies:** 2\ **Last updated:** [September 15, 2025, 3:27am UTC](https://discuss.ai.google.dev/t/2-5-flash-lite-fine-tuning/95206 "2025-09-15T03:27:42Z")
Today Gemini launched 2.5 Flash Lite General availability - but they have explicitly mentioned that it doesn’t support fine-tuning - Just wanted to know if they do this kind of thing, and they will launch SFT soon? In h… --- ## [How a 4-Hour Overnight Fine-Tune for My Boyfriend Became a £11,550 Google Cloud Bill – and What I Learned](https://discuss.ai.google.dev/t/how-a-4-hour-overnight-fine-tune-for-my-boyfriend-became-a-11-550-google-cloud-bill-and-what-i-learned/88535)
**Author:** [@Lyahn](https://discuss.ai.google.dev/u/Lyahn)\ **Replies:** 10\ **Last updated:** [September 10, 2025, 3:18pm UTC](https://discuss.ai.google.dev/t/how-a-4-hour-overnight-fine-tune-for-my-boyfriend-became-a-11-550-google-cloud-bill-and-what-i-learned/88535 "2025-09-10T15:18:30Z")
Last week I tried something fun and romantic: I wanted to give my boyfriend a special birthday surprise, so I fine-tuned a large-language model with Google Vertex AI using our past chat history. The goal was for the mode… --- ## [Fine-tuning and go-sdk support in docs](https://discuss.ai.google.dev/t/fine-tuning-and-go-sdk-support-in-docs/96930)
**Author:** [@matzyn](https://discuss.ai.google.dev/u/matzyn)\ **Replies:** 1\ **Last updated:** [September 8, 2025, 8:58am UTC](https://discuss.ai.google.dev/t/fine-tuning-and-go-sdk-support-in-docs/96930 "2025-09-08T08:58:29Z")
When will you return fine-tuning for 2.0/2.5 models? Also, are you planning to add Go examples in the documentation? Right now, the main topics are covered, but deeper dives only include Python and JS. Will this change? --- ## [Are we able to fine tune the video understanding on Gemini 2.5 Pro?](https://discuss.ai.google.dev/t/are-we-able-to-fine-tune-the-video-understanding-on-gemini-2-5-pro/99521)
**Author:** [@harrythentrepreneur](https://discuss.ai.google.dev/u/harrythentrepreneur)\ **Replies:** 3\ **Last updated:** [September 8, 2025, 5:36am UTC](https://discuss.ai.google.dev/t/are-we-able-to-fine-tune-the-video-understanding-on-gemini-2-5-pro/99521 "2025-09-08T05:36:12Z")
I have looked around and I am confused whether we can or we can not fine tune Gemini 2.5 Pro vision video understanding capabilities? --- ## [Please enable Fine-Tuning for Gemini 2.0 Flash for Educational Applications](https://discuss.ai.google.dev/t/please-enable-fine-tuning-for-gemini-2-0-flash-for-educational-applications/99064)
**Author:** [@Rohan\_Khan](https://discuss.ai.google.dev/u/Rohan_Khan)\ **Replies:** 3\ **Last updated:** [August 21, 2025, 9:05am UTC](https://discuss.ai.google.dev/t/please-enable-fine-tuning-for-gemini-2-0-flash-for-educational-applications/99064 "2025-08-21T09:05:01Z")
Gemini-2.0-flash is the perfect model for my application considering the cost and speed. Please enable fine-tuning on Gemini-2.0-flash. --- ## [Seeking Tips & Use Cases for Gemma 3 270M in Flutter](https://discuss.ai.google.dev/t/seeking-tips-use-cases-for-gemma-3-270m-in-flutter/98700)
**Author:** [@itjimkr](https://discuss.ai.google.dev/u/itjimkr)\ **Replies:** 1\ **Last updated:** [August 19, 2025, 7:19am UTC](https://discuss.ai.google.dev/t/seeking-tips-use-cases-for-gemma-3-270m-in-flutter/98700 "2025-08-19T07:19:07Z")
Hi Gemma community, I’m a developer who is currently experimenting with integrating the Gemma 3 270M model into a Flutter application. I’ve become very curious about how to bring out the full potential of this model. Fo… --- ## [Slm local offload processing power](https://discuss.ai.google.dev/t/slm-local-offload-processing-power/96880)
**Author:** [@Alfred\_Berrios](https://discuss.ai.google.dev/u/Alfred_Berrios)\ **Replies:** 2\ **Last updated:** [August 7, 2025, 4:22pm UTC](https://discuss.ai.google.dev/t/slm-local-offload-processing-power/96880 "2025-08-07T16:22:55Z")
Proposal: Google-Developed SLM and App for Hybrid Gemini Integration with Wi-Fi Offloading and Enhancement Potential Date: August 5, 2025 Prepared by: Anonymous Developer (with assistance from Grok 3, xAI) Purpose: To … --- ## [Fine Tuning gemini flash-lite 2.0](https://discuss.ai.google.dev/t/fine-tuning-gemini-flash-lite-2-0/86371)
**Author:** [@CraftersMC-Guides](https://discuss.ai.google.dev/u/CraftersMC-Guides)\ **Replies:** 4\ **Last updated:** [August 6, 2025, 11:15pm UTC](https://discuss.ai.google.dev/t/fine-tuning-gemini-flash-lite-2-0/86371 "2025-08-06T23:15:10Z")
Hey, we are a small community team who are trying to create an AI model for our project. We saw that gemini 1.5 is discontinued and went with gemini 2.0 flash-lite. But I’m not sure how to fine tune it as everywhere I go… --- ## [Tuned model creation failed ERROR!](https://discuss.ai.google.dev/t/tuned-model-creation-failed-error/38069)
**Author:** [@Raw\_MC](https://discuss.ai.google.dev/u/Raw_MC)\ **Replies:** 1\ **Last updated:** [July 30, 2025, 6:23am UTC](https://discuss.ai.google.dev/t/tuned-model-creation-failed-error/38069 "2025-07-30T06:23:22Z")
Hi all, I’m trying to fine tune on my dataset. Everytime I click tune button, I keep on getting Error: Tuned model creation failed as shown in the attached image. I’ve read other similiar threads whose solution was maki… --- ## [FPS Processing Capability of Gemini 2.5 Flash for Video Fine-Tuning](https://discuss.ai.google.dev/t/fps-processing-capability-of-gemini-2-5-flash-for-video-fine-tuning/95200)
**Author:** [@harishkumar](https://discuss.ai.google.dev/u/harishkumar)\ **Replies:** 3\ **Last updated:** [August 5, 2025, 11:00am UTC](https://discuss.ai.google.dev/t/fps-processing-capability-of-gemini-2-5-flash-for-video-fine-tuning/95200 "2025-08-05T11:00:06Z")
Hello everyone, I am currently working on something that involves tracking objects in a video footage using the Gemini 2.5 Flash model. I initially tried using the pre-trained model, but it didn’t worked out of the for … --- ## [Fine tuned models returning no content when running using the AI studio interface](https://discuss.ai.google.dev/t/fine-tuned-models-returning-no-content-when-running-using-the-ai-studio-interface/6459)
**Author:** [@fickle\_Fox](https://discuss.ai.google.dev/u/fickle_Fox)\ **Replies:** 1\ **Last updated:** [July 17, 2025, 9:50am UTC](https://discuss.ai.google.dev/t/fine-tuned-models-returning-no-content-when-running-using-the-ai-studio-interface/6459 "2025-07-17T09:50:05Z")
So I need successfully fine-tuned the model using the very easy to use tuned model interface, with a relatively small dataset to test things out. However upon attempting to run the structure of prompt with the fine-tuned… --- ## [Tuned model sometimes fail to generate content](https://discuss.ai.google.dev/t/tuned-model-sometimes-fail-to-generate-content/41765)
**Author:** [@seungduk](https://discuss.ai.google.dev/u/seungduk)\ **Replies:** 1\ **Last updated:** [July 15, 2025, 9:51am UTC](https://discuss.ai.google.dev/t/tuned-model-sometimes-fail-to-generate-content/41765 "2025-07-15T09:51:12Z")
I got the following error when using a fine-tuned model from Vertex. And the weirdest thing was that, sometimes the request succeeded. rpc error: code = Invalid Argument desc = Request contains an invalid argument. It… --- ## [Tuning doesn't work even with Gemini Pro 1.0 for datasets containing outputs with more than 5000 characters](https://discuss.ai.google.dev/t/tuning-doesnt-work-even-with-gemini-pro-1-0-for-datasets-containing-outputs-with-more-than-5000-characters/41495)
**Author:** [@Abrar](https://discuss.ai.google.dev/u/Abrar)\ **Replies:** 1\ **Last updated:** [July 14, 2025, 3:33pm UTC](https://discuss.ai.google.dev/t/tuning-doesnt-work-even-with-gemini-pro-1-0-for-datasets-containing-outputs-with-more-than-5000-characters/41495 "2025-07-14T15:33:42Z")
From the official website, it mentions the output character limit for training dataset as 5000 only for Gemini Flash 1.5 model. But I am unable to tune even with Gemini Pro 1.0 model. Important to note that I was able to… --- ## [I cannot create a tuned model;, and my previous tuned models are no longer workingf. somebody help me](https://discuss.ai.google.dev/t/i-cannot-create-a-tuned-model-and-my-previous-tuned-models-are-no-longer-workingf-somebody-help-me/58482)
**Author:** [@Ezekiel\_Jatto](https://discuss.ai.google.dev/u/Ezekiel_Jatto)\ **Replies:** 5\ **Last updated:** [July 12, 2025, 9:30am UTC](https://discuss.ai.google.dev/t/i-cannot-create-a-tuned-model-and-my-previous-tuned-models-are-no-longer-workingf-somebody-help-me/58482 "2025-07-12T09:30:02Z")
i cannot use gemini in my app anymore. i keep getting the error messge, an internal error has occured --- ## [Gemini API fine-tuning disabled?](https://discuss.ai.google.dev/t/gemini-api-fine-tuning-disabled/92763)
**Author:** [@Cristobal](https://discuss.ai.google.dev/u/Cristobal)\ **Replies:** 1\ **Last updated:** [July 7, 2025, 4:59pm UTC](https://discuss.ai.google.dev/t/gemini-api-fine-tuning-disabled/92763 "2025-07-07T16:59:42Z")
Hi everyone! Has the Gemini API fine-tuning option been disabled? I was trying implement a tuning via python and it seems that there are no available models anymore and the remaining option is VertexAI based, whose billi… --- ## [Suggestion for Google AI Studio](https://discuss.ai.google.dev/t/suggestion-for-google-ai-studio/80630)
**Author:** [@King\_Gaming\_Roblox](https://discuss.ai.google.dev/u/King_Gaming_Roblox)\ **Replies:** 1\ **Last updated:** [July 1, 2025, 7:52am UTC](https://discuss.ai.google.dev/t/suggestion-for-google-ai-studio/80630 "2025-07-01T07:52:36Z")
I have a suggestion. When fine tuning models, I feel like we should be able to provide it with files for a knowledge base. Let’s say we’re making a bot that assists people with programming tasks for our website that use… --- ## [Google disabled createTunedModel action?](https://discuss.ai.google.dev/t/google-disabled-createtunedmodel-action/90052)
**Author:** [@angusma](https://discuss.ai.google.dev/u/angusma)\ **Replies:** 1\ **Last updated:** [June 23, 2025, 6:29am UTC](https://discuss.ai.google.dev/t/google-disabled-createtunedmodel-action/90052 "2025-06-23T06:29:36Z")
I was able to tune gemini 1.5 flash model through API, but it failed a few days ago. I tried to find a model that supports the action createTunedModel and found nothing. Did Google disabled this action? --- ## [Fine tuning / low langage ressource](https://discuss.ai.google.dev/t/fine-tuning-low-langage-ressource/74899)
**Author:** [@Assani\_moufida](https://discuss.ai.google.dev/u/Assani_moufida)\ **Replies:** 1\ **Last updated:** [June 19, 2025, 9:48am UTC](https://discuss.ai.google.dev/t/fine-tuning-low-langage-ressource/74899 "2025-06-19T09:48:19Z")
Hello/Good evening, I have a question regarding the use of Google AI Studio. Is it appropriate to use it to fine-tune a model in order to teach it new languages? I have a high-quality dataset of 500 rows. [Next page](https://discuss.ai.google.dev/tag/fine-tuning/9.md?match_all_tags=true&page=1&tags%5B%5D=fine-tuning)