{"templateName":"base-page-template54","cssClassNames":"page basicpage summit-page","canonicalLink":"https://www.snowflake.com/en/artificial-intelligence/machine-learning/model-training/fine-tuning/","robotsTags":[],"allowedRenditionsWidth":["320","480","640","768","960","1200","1440","1920"],"description":"Learn how LLM fine-tuning works, when to use LoRA or QLoRA, how it compares with RAG and prompting, and how to build and evaluate training data.","language":"en","title":"LLM Fine-Tuning: Methods, LoRA and QLoRA | Snowflake","analyticsPageType":"homepage","analyticsCategory":"general","analyticsSubCategory":"","excludeFromAnalytics":false,":items":{"root":{"columnCount":12,"columnClassNames":{"markup_editor_928258845":"aem-GridColumn aem-GridColumn--default--12","experiencefragment-banner":"aem-GridColumn aem-GridColumn--default--12","experiencefragment-header":"aem-GridColumn aem-GridColumn--default--12","responsivegrid":"aem-GridColumn aem-GridColumn--default--12","markup_editor_597730182":"aem-GridColumn aem-GridColumn--default--12","experiencefragment-footer":"aem-GridColumn aem-GridColumn--default--12","experiencefragment":"aem-GridColumn aem-GridColumn--default--12","modal_container":"aem-GridColumn aem-GridColumn--default--12","markup_editor":"aem-GridColumn aem-GridColumn--default--12"},"gridClassNames":"aem-Grid aem-Grid--12 aem-Grid--default--12",":items":{"experiencefragment-banner":{"id":"experiencefragment-da663d0872","localizedFragmentVariationPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/pushdown-banner/master/jcr:content","configured":true,":type":"snowflake-site/components/experiencefragment","xfModelPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/pushdown-banner/master.xfmodel.json"},"experiencefragment-header":{"id":"experiencefragment-2b2798fee1","localizedFragmentVariationPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/mega-nav-header/master/jcr:content","configured":true,":type":"snowflake-site/components/experiencefragment","xfModelPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/mega-nav-header/master.xfmodel.json","languageNavPath":"/content/snowflake-site/global/en/artificial-intelligence/machine-learning/model-training/fine-tuning.languagenav.json"},"responsivegrid":{"columnCount":12,"columnClassNames":{"flexible_column_cont_939100716":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_1158003461":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_663228916":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_912630531":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_1398138236":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_1786318617":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_1467213961":"aem-GridColumn aem-GridColumn--default--12","flexible_column_cont_1377146023":"aem-GridColumn aem-GridColumn--default--12"},"gridClassNames":"aem-Grid aem-Grid--12 aem-Grid--default--12",":items":{"flexible_column_cont":{"id":"flexible-column-container-2828cdac82","propertiesId":"hub-hero-breadcrumbs","type":"1-column","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"small","bottomPadding":"none","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"propertiesCSSClasses":"page-section","backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"container-23c22f4322",":items":{"breadcrumb":{"id":"breadcrumb-cb28e4c4c0","items":[{"id":"breadcrumb-cb28e4c4c0-item-f710b50fc9","link":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/"},"active":false,"current":false,"title":"Machine Learning","appliedCssClassNames":"summit-page",":type":"snowflake-site/components/structure/page"},{"id":"breadcrumb-cb28e4c4c0-item-7b414e79ce","link":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/model-training/"},"active":false,"current":false,"title":"ML Model Training","appliedCssClassNames":"summit-page",":type":"snowflake-site/components/structure/page"},{"id":"breadcrumb-cb28e4c4c0-item-1f01b51074","link":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/model-training/fine-tuning/"},"active":true,"current":true,"title":"Fine-Tuning","appliedCssClassNames":"summit-page",":type":"snowflake-site/components/structure/page"}],":type":"snowflake-site/components/breadcrumb"}},":itemsOrder":["breadcrumb"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_939100716":{"id":"flexible-column-container-486265e544","propertiesId":"hub-hero","type":"2-column-even","alignColumns":"center","containerMaxWidth":"extra-large","topPadding":"extra-small","bottomPadding":"extra-small","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"propertiesCSSClasses":"page-section","backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"container-7a34232625",":items":{"title_v2":{"id":"title-v2-58f4b0880e","additionalClasses":"hub-hero__headline","type":"heading1","lines":["Fine-Tuning Has Changed: A Practitioner’s Guide to Modern Model Adaptation"],":type":"snowflake-site/components/title-v2","appliedCssClassNames":"left-alignment"},"text":{"id":"text-be8c66a1a7","additionalClasses":"hub-hero__subheadline","text":"\u003Cp\u003EFrom full fine-tuning to lightweight adapters, today’s teams have more ways to specialize AI models than ever. Here’s how to match the technique to the task.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"},"container":{"additionalClasses":"hub-hero__authors","layout":"RESPONSIVE_GRID","columnCount":12,"columnClassNames":{"content_chip_copy":"aem-GridColumn aem-GridColumn--default--12","content_chip":"aem-GridColumn aem-GridColumn--default--12"},"gridClassNames":"aem-Grid aem-Grid--12 aem-Grid--default--12","id":"container-e64d8ad53c",":items":{"content_chip":{"id":"content-chip-b3f7ca298a","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/blog/authors/Laurie-Macpherson/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Read bio"},"image":{"id":"image","height":"800","src":"https://www.snowflake.com/adobe/dynamicmedia/deliver/dm-aid--a7e9fdff-6f08-4edc-9cd1-e213bb234aaa/laurie-macpherson.jpg?preferwebp=true&quality=85","lazyEnabled":true,"alt":"Laurie MacPherson","isLcpImage":true,"width":"800",":type":"snowflake-site/components/image"},"headline":{"id":"title","type":"heading5","lines":["Laurie MacPherson","Technical Writer, Snowflake"],":type":"snowflake-site/components/title-v2"},":type":"snowflake-site/components/content-chip"},"content_chip_copy":{"id":"content-chip-4f308a9d57","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/blog/authors/david-gaule/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Read bio"},"image":{"id":"image","height":"512","src":"https://www.snowflake.com/adobe/dynamicmedia/deliver/dm-aid--fd9454ea-3b59-4d19-95be-534774cbd226/david.jpg?preferwebp=true&quality=85","lazyEnabled":true,"alt":"David Gaule","isLcpImage":false,"width":"512",":type":"snowflake-site/components/image"},"headline":{"id":"title","type":"heading5","lines":["David Gaule","Technical Editor, Snowflake"],":type":"snowflake-site/components/title-v2"},":type":"snowflake-site/components/content-chip"}},":itemsOrder":["content_chip","content_chip_copy"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-small",":type":"snowflake-site/components/container"}},":itemsOrder":["title_v2","text","container"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-extra-small",":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},"flexible_column_content_container_2":{"additionalClasses":"hub-hero__video-column","layout":"SIMPLE","id":"container-fea68e91ac",":items":{"youtube":{"id":"embed-3a2b140fa9","youtubeVideoId":"pzr1LXHMtFs","layout":"responsive","youtubeAspectRatio":"56.25","youtubeAutoPlay":false,"youtubeLoop":false,"youtubeMute":false,"youtubePlaysInline":false,"youtubeRel":false,"embeddableResourceType":"core/wcm/components/embed/v1/embed/embeddable/youtube","type":"EMBEDDABLE",":type":"snowflake-site/components/youtube"}},":itemsOrder":["youtube"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_1398138236":{"id":"flexible-column-container-c844ee846b","propertiesId":"hub-hero-related-topics","type":"1-column","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"extra-small","bottomPadding":"medium","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"propertiesCSSClasses":"page-section","backgroundImageOption":"none","flexible_column_content_container_1":{"additionalClasses":"related-topics-outer-container border-top","layout":"SIMPLE","id":"container-c83b5c5375",":items":{"text_894059747":{"id":"text-bfc7ee70b4","additionalClasses":"seo-hub-hero__related-topic-label","text":"\u003Cp\u003EML Model Training Topics:\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"},"text":{"id":"text-dbd1ba94e3","additionalClasses":"related-topics ","text":"\u003Cul\u003E\r\n\u003Cli\u003E\u003Ca href=\"https://www.snowflake.com/en/artificial-intelligence/machine-learning/model-training/data-augmentation/\"\u003EData Augmentation\u003C/a\u003E\u003C/li\u003E\r\n\u003Cli\u003E\u003Ca href=\"https://www.snowflake.com/en/artificial-intelligence/machine-learning/model-training/gradient-descent/\"\u003EGradient Descent\u003C/a\u003E\u003C/li\u003E\r\n\u003C/ul\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-small"}},":itemsOrder":["text_894059747","text"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_663228916":{"id":"flexible-column-container-ee105a1e7c","propertiesId":"hub-body","type":"2-column-60-40","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"medium","bottomPadding":"medium","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"propertiesCSSClasses":"page-section","backgroundImageOption":"none","flexible_column_content_container_1":{"additionalClasses":"longform-content","layout":"SIMPLE","id":"hub-body-content",":items":{"callout__0":{"id":"text-23965b8589","additionalClasses":"callout callout--general","text":"\u003Cp\u003E\u003Cstrong\u003EFINE-TUNING DEFINED\u003C/strong\u003E\u003C/p\u003E\n\u003Cp\u003EFine-tuning is the process of continuing to train a pretrained model for a narrower objective, updating all or a selected subset of its parameters to make its outputs better match a specific task, domain or behavior.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"},"text__0":{"id":"text-cda7ff745f","text":"\u003Cp\u003EFine-tuning once implied updating an entire model, a process that demanded substantial compute, memory and operational support. As foundation models grew, that approach meant specialization was beyond the practical reach of many teams — especially when the application required several task-specific versions or frequent experimentation.\u003C/p\u003E\n\u003Cp\u003EThe category has since expanded. Parameter-efficient techniques such as low-rank adaptation (LoRA) can train a comparatively small set of additional parameters while leaving the base model frozen, and QLoRA reduces \u003Ca href=\"https://www.snowflake.com/en/artificial-intelligence/machine-learning/model-training/\"\u003Emodel training\u003C/a\u003E memory requirements further by loading those frozen weights in a quantized representation. Compact adapters can support multiple specializations around the same base model, while distillation workflows allow teams to transfer a bounded capability from a larger model to a smaller one. Managed fine-tuning services have also reduced the infrastructure work required to prepare data, run training jobs and deploy the resulting model.\u003C/p\u003E\n\u003Cp\u003EFine-tuning now describes a collection of training objectives and parameter-update strategies rather than a single procedure. Understanding how those choices fit together helps practitioners decide what behavior to teach, how extensively to modify the model and how to account for the cost, governance and maintenance that follow.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_what-is-fine-tuning":{"id":"title-v2-bc9b4e9932","additionalClasses":"anchor-title anchor-title--what-is-fine-tuning","type":"heading2","lines":["What is fine-tuning?"],":type":"snowflake-site/components/title-v2"},"text_what-is-fine-tuning_0":{"id":"text-0afcb8c2af","text":"\u003Cp\u003EFine-tuning is the process of continuing to train a pretrained model on a narrower data set, task or objective. During training, optimization adjusts either all of the model’s parameters or a selected subset, shifting the probability of its outputs toward patterns represented in the fine-tuning examples.\u003C/p\u003E\n\u003Cp\u003EThe pretrained model supplies the starting point. Its initial training has already captured statistical patterns across a broad collection of text, images, code or other content, so a team can specialize those existing capabilities without repeating the full pretraining process. This reuse of learned representations is one form of transfer learning.\u003C/p\u003E\n\u003Cp\u003EFor \u003Ca href=\"https://www.snowflake.com/en/fundamentals/large-language-model/\"\u003Elarge language models\u003C/a\u003E (LLMs), fine-tuning often happens within a broader post-training pipeline. A base model may undergo supervised instruction tuning, preference optimization and safety training before being released. Another round of fine-tuning might specialize the model for a narrower task, such as extracting fields from contracts, assigning support tickets to a controlled taxonomy or generating output that conforms to an application schema.\u003C/p\u003E\n\u003Cp\u003EFine-tuning differs from continued pretraining, although both update model parameters. Continued pretraining generally uses the original language-modeling objective on a large body of unlabeled domain text, helping the model adapt to the vocabulary and distribution of a field. Task-specific fine-tuning usually works with labeled examples that demonstrate the desired response.\u003C/p\u003E\n\u003Cp\u003E\u003Cem\u003EWatch Snowflake’s Arun Agarwal, Yusuf Ozuysal and James Cha-Earley discuss why AI model flexibility is crucial for real-world applications:\u003C/em\u003E\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"yt_what-is-fine-tuning_0":{"id":"embed-95464cca88","youtubeVideoId":"ZUZsEzFptAA","layout":"responsive","youtubeAspectRatio":"56.25","youtubeAutoPlay":false,"youtubeLoop":false,"youtubeMute":false,"youtubePlaysInline":false,"youtubeRel":false,"embeddableResourceType":"core/wcm/components/embed/v1/embed/embeddable/youtube","type":"EMBEDDABLE",":type":"snowflake-site/components/youtube"},"title_what-fine-tuning-can-change":{"id":"title-v2-67a45a1e11","additionalClasses":"anchor-title anchor-title--what-fine-tuning-can-change","type":"heading2","lines":["What fine-tuning can change"],":type":"snowflake-site/components/title-v2"},"text_what-fine-tuning-can-change_0":{"id":"text-56a5b9f068","text":"\u003Cp\u003EA fine-tuning data set contains repeated mappings between inputs and desired outputs. As training proceeds, gradient updates increase the likelihood that the model will reproduce similar patterns when it receives a related input.\u003C/p\u003E\n\u003Cp\u003EThose patterns may describe several kinds of behavior:\u003C/p\u003E\n\u003Cul\u003E\n\u003Cli\u003ESelecting a label from a defined taxonomy\u003C/li\u003E\n\u003Cli\u003EExtracting fields into a required schema\u003C/li\u003E\n\u003Cli\u003EFollowing an organization’s terminology or response conventions\u003C/li\u003E\n\u003Cli\u003EProducing a particular style or level of detail\u003C/li\u003E\n\u003Cli\u003ECalling tools or generating function arguments in an expected format\u003C/li\u003E\n\u003Cli\u003EHandling recurring domain inputs that differ from the model’s general training distribution\u003C/li\u003E\n\u003C/ul\u003E\n\u003Cp\u003EFine-tuning can also support model compression workflows. A larger teacher model can generate labels, explanations or other training signals, which are then used to fine-tune a smaller student model for a bounded task. \u003Ca href=\"https://arxiv.org/abs/2305.02301\" target=\"_blank\"\u003EResearch has shown\u003C/a\u003E that teacher-generated rationales can help smaller models perform competitively with much larger models on the tasks included in the experiment. Results will still depend on the quality of the teacher output, the student architecture and the evaluation set.\u003C/p\u003E\n\u003Cp\u003EThe resulting model has limits. Training examples may improve its handling of a domain without making every generated claim accurate. Sparse coverage can leave important edge cases unaddressed, while continued fine-tuning can reduce performance on previously learned tasks. \u003Ca href=\"https://arxiv.org/abs/2308.08747\" target=\"_blank\"\u003EIn one empirical study\u003C/a\u003E, researchers observed this form of catastrophic forgetting across tests of domain knowledge, reasoning and reading comprehension. Information encoded through parameter updates is also difficult to inspect, attribute or refresh one record at a time.\u003C/p\u003E\n\u003Cp\u003EFor facts that change frequently or must be traced to an authoritative source, \u003Ca href=\"https://www.snowflake.com/en/fundamentals/rag/\"\u003Eretrieval-augmented generation\u003C/a\u003E (RAG) usually provides a better mechanism. Fine-tuning is generally strongest when the desired behavior remains stable across many model calls.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"card_v2_what-fine-tuning-can-change_0":{"id":"card-v2-568a3c7634","additionalClasses":"seo-customer","configurationStatus":{"configured":true,"message":""},":type":"snowflake-site/components/card-v2","title":{"id":"title","type":"heading4","lines":["Jet2"],":type":"snowflake-site/components/title-v2"},"button":{"id":"button","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/customers/all-customers/case-study/jet2/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Read the full customer story"},"image":{"id":"image","height":"351","src":"https://www.snowflake.com/adobe/dynamicmedia/deliver/dm-aid--668fc1ee-25d2-49b6-a9cc-cf526ad23f4d/jet2-logo.png?preferwebp=true&quality=85","lazyEnabled":true,"alt":"Jet2","isLcpImage":false,"width":"624",":type":"snowflake-site/components/image"},"type":"content-card","text":{"id":"text","text":"\u003Cp\u003EJet2 uses Snowflake’s AI Data Cloud and Cortex AI to analyze customer intent and sentiment across thousands of call transcripts, helping the UK tour operator and airline better understand why customers contact its service centers. With a fine-tuned LLM in Cortex AI, Jet2 went from no prior LLM experience to a production-ready use case in three months, and was analyzing roughly 2,000 calls per day while maintaining a balance of scale, accuracy and cost [as of June 2025].\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text"},"layoutStyle":"horizontal"},"title_fine-tuning-vs-prompting-rag-and-continued-pretraining":{"id":"title-v2-bb4f59a268","additionalClasses":"anchor-title anchor-title--fine-tuning-vs-prompting-rag-and-continued-pretraining","type":"heading2","lines":["Fine-tuning vs. prompting, RAG and continued pretraining"],":type":"snowflake-site/components/title-v2"},"text_fine-tuning-vs-prompting-rag-and-continued-pretraining_0":{"id":"text-17a7661c1e","text":"\u003Cp\u003EPrompting, RAG, fine-tuning and continued pretraining operate at different points in the model lifecycle. Selecting among them starts with the condition the system needs to change.\u003C/p\u003E\n\u003Ctable\u003E\u003Cthead\u003E\u003Ctr\u003E\u003Cth\u003EApproach\u003C/th\u003E\u003Cth\u003EUpdates model parameters?\u003C/th\u003E\u003Cth\u003EAppropriate for\u003C/th\u003E\u003Cth\u003EPrimary constraint\u003C/th\u003E\u003C/tr\u003E\u003C/thead\u003E\u003Ctbody\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cstrong\u003EPrompting\u003C/strong\u003E\u003C/td\u003E\u003Ctd\u003ENo\u003C/td\u003E\u003Ctd\u003ESupplying instructions, examples and contextual constraints for a model call\u003C/td\u003E\u003Ctd\u003EInstructions consume context and may produce variable results across inputs\u003C/td\u003E\u003C/tr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cstrong\u003ERAG\u003C/strong\u003E\u003C/td\u003E\u003Ctd\u003ENo\u003C/td\u003E\u003Ctd\u003EProviding current, private or attributable information during inference\u003C/td\u003E\u003Ctd\u003ERetrieval quality and the model’s use of retrieved context affect the final answer\u003C/td\u003E\u003C/tr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cstrong\u003EFine-tuning\u003C/strong\u003E\u003C/td\u003E\u003Ctd\u003EYes\u003C/td\u003E\u003Ctd\u003ETeaching stable task behavior, response patterns and output conventions\u003C/td\u003E\u003Ctd\u003ERequires representative training data, evaluation and ongoing maintenance\u003C/td\u003E\u003C/tr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cstrong\u003EContinued pretraining\u003C/strong\u003E\u003C/td\u003E\u003Ctd\u003EYes\u003C/td\u003E\u003Ctd\u003EAdapting a model to a broad domain corpus, vocabulary or distribution\u003C/td\u003E\u003Ctd\u003EUsually requires substantially more data and compute than task-specific fine-tuning\u003C/td\u003E\u003C/tr\u003E\u003Ctr\u003E\u003Ctd\u003E\u003Cstrong\u003EHybrid architecture\u003C/strong\u003E\u003C/td\u003E\u003Ctd\u003ESometimes\u003C/td\u003E\u003Ctd\u003ECombining specialized behavior with external information\u003C/td\u003E\u003Ctd\u003EAdds components that must be evaluated together\u003C/td\u003E\u003C/tr\u003E\u003C/tbody\u003E\u003C/table\u003E\n\u003Cp\u003EA practical rule of thumb is to use prompting for instructions that can fit reliably within the model call, RAG for information that changes or requires source attribution, and fine-tuning for repeated behaviors that remain inconsistent after the prompt and surrounding application have been improved.\u003C/p\u003E\n\u003Cp\u003EThe boundaries aren’t absolute, however. Retrieved documents can contain examples and instructions that influence how a model responds, while fine-tuning data can expose a model to domain facts and terminology. The distinction lies in the mechanism: RAG inserts content into the inference context, whereas fine-tuning incorporates patterns through parameter updates.\u003C/p\u003E\n\u003Cp\u003EWhen an application requires both specialized behavior and retrieved information, the techniques can work together. Retrieval-Augmented Fine-Tuning (RAFT), for example, trains a model on questions paired with relevant and irrelevant retrieved documents. The model learns to use supporting material while disregarding distractors, improving its behavior within the retrieval workflow examined by the researchers. RAFT provides a useful training recipe for in-domain, open-book question answering; it doesn’t establish that every RAG system requires fine-tuning.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_choose-the-fine-tuning-objective":{"id":"title-v2-3e75b3eb88","additionalClasses":"anchor-title anchor-title--choose-the-fine-tuning-objective","type":"heading2","lines":["Choose the fine-tuning objective"],":type":"snowflake-site/components/title-v2"},"text_choose-the-fine-tuning-objective_0":{"id":"text-4db48cc2ca","text":"\u003Cp\u003EBefore selecting LoRA, QLoRA or full fine-tuning, a team needs to define what the training examples will teach. The objective determines the signal used to update the model, while the parameter strategy determines where those updates are applied.\u003C/p\u003E\n\u003Ch3\u003ESupervised fine-tuning\u003C/h3\u003E\n\u003Cp\u003ESupervised fine-tuning (SFT) trains a model on labeled input-output pairs. For an LLM, each example may contain a prompt and an approved completion, allowing the model to learn the relationship between an instruction and the response it should generate.\u003C/p\u003E\n\u003Cp\u003EA classification data set might pair each support request with its assigned category. An extraction data set could pair a document passage with validated JSON. For a text-to-SQL application, examples could connect natural-language questions to reviewed SQL statements.\u003C/p\u003E\n\u003Cp\u003EInstruction tuning is a form of supervised fine-tuning in which the data set covers tasks expressed as instructions. Training across a varied instruction collection can improve a model’s ability to follow requests, including requests that weren’t represented exactly in the training set.\u003C/p\u003E\n\u003Cp\u003EThe quality of the target completions sets an upper bound on what the model can learn from them. Inconsistent labels, contradictory instructions and superficially polished answers with factual errors all transfer undesirable patterns into training.\u003C/p\u003E\n\u003Ch3\u003EPreference optimization\u003C/h3\u003E\n\u003Cp\u003ESome behaviors are easier to express through comparisons than through a single approved answer. A preference data set presents two or more candidate responses, along with an indication of which response is preferred.\u003C/p\u003E\n\u003Cp\u003EDirect Preference Optimization (DPO) uses these pairs to increase the likelihood of preferred responses and decrease the likelihood of rejected responses. The \u003Ca href=\"https://arxiv.org/abs/2305.18290\" target=\"_blank\"\u003Eoriginal DPO method\u003C/a\u003E formulates preference training as a classification loss, avoiding the separate reward-model training and reinforcement-learning stages used in a conventional reinforcement learning from human feedback pipeline.\u003C/p\u003E\n\u003Cp\u003EPreference optimization can refine qualities such as relevance, concision, adherence to policy or appropriate refusal behavior. Its results still depend on the criteria behind the rankings. When reviewers apply different standards, the model receives an inconsistent learning signal.\u003C/p\u003E\n\u003Ch3\u003EReinforcement learning from feedback\u003C/h3\u003E\n\u003Cp\u003EReinforcement learning from human feedback (\u003Ca href=\"https://www.snowflake.com/en/artificial-intelligence/machine-learning/reinforcement-learning/rlhf/\"\u003ERLHF\u003C/a\u003E) commonly begins with supervised fine-tuning, followed by the collection of human rankings across model responses. A reward model learns from those rankings, and reinforcement learning then adjusts the language model to produce responses that receive higher predicted rewards.\u003C/p\u003E\n\u003Cp\u003EThe \u003Ca href=\"https://arxiv.org/abs/2203.02155\" target=\"_blank\"\u003EInstructGPT research\u003C/a\u003E used this sequence to improve instruction following and human preference ratings relative to the underlying GPT-3 models. The researchers also documented remaining mistakes, illustrating the difference between improved preference alignment and guaranteed correctness.\u003C/p\u003E\n\u003Cp\u003ERLHF offers considerable control over broad model behavior, though its data collection and training pipeline is more involved than a straightforward supervised fine-tuning job. For a bounded enterprise task with clear approved outputs, SFT may provide a more direct starting point.\u003C/p\u003E\n\u003Ch3\u003EDistillation\u003C/h3\u003E\n\u003Cp\u003EDistillation transfers a capability from a larger teacher model to a smaller student model. The teacher may produce labels, completions, preference pairs or reasoning traces; the student then trains on those signals, often through supervised fine-tuning.\u003C/p\u003E\n\u003Cp\u003EFor workloads with high request volume, a successful distillation workflow can shift inference to a smaller model with a different latency and cost profile. The student should be evaluated independently, since teacher errors can enter the training data and a smaller model may not reproduce every capability used to generate the examples.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_choose-which-parameters-to-update":{"id":"title-v2-2ec20638d2","additionalClasses":"anchor-title anchor-title--choose-which-parameters-to-update","type":"heading2","lines":["Choose which parameters to update"],":type":"snowflake-site/components/title-v2"},"text_choose-which-parameters-to-update_0":{"id":"text-1b812183e8","text":"\u003Cp\u003EOnce the training objective and data set are defined, the next decision concerns the model parameters. Full fine-tuning updates the entire model, while parameter-efficient fine-tuning (PEFT) limits training to a smaller set of parameters or added components.\u003C/p\u003E\n\u003Ch3\u003EFull fine-tuning\u003C/h3\u003E\n\u003Cp\u003EFull fine-tuning allows optimization to update all model parameters. Because the training process can modify the model throughout its layers, it provides substantial capacity for adaptation.\u003C/p\u003E\n\u003Cp\u003EThat flexibility requires memory for the model weights, gradients and optimizer states, along with storage and operational support for the resulting model. Each specialized version may require its own full set of weights. As model size increases, the infrastructure and experimentation costs can make full fine-tuning difficult to justify for a narrowly defined task.\u003C/p\u003E\n\u003Cp\u003EFull fine-tuning may still be appropriate when the target distribution differs substantially from the base model’s training, broad adaptation is required and evaluation shows a meaningful advantage over parameter-efficient methods.\u003C/p\u003E\n\u003Ch3\u003EParameter-efficient fine-tuning\u003C/h3\u003E\n\u003Cp\u003EPEFT methods keep most pretrained parameters frozen and train a smaller collection of parameters. Since the base weights remain unchanged, teams can store and manage compact task-specific components alongside a shared model.\u003C/p\u003E\n\u003Cp\u003EFor many enterprise experiments, PEFT provides a practical starting point. It lowers training memory requirements and allows several specializations to reuse the same base model, although quality remains dependent on the task, model, adapter configuration and data.\u003C/p\u003E\n\u003Ch3\u003ELoRA\u003C/h3\u003E\n\u003Cp\u003ELow-rank adaptation (LoRA) freezes the original model weights and introduces trainable low-rank matrices into selected layers. During training, these matrices learn a compact approximation of the parameter updates required for the task.\u003C/p\u003E\n\u003Cp\u003EIn the \u003Ca href=\"https://arxiv.org/abs/2106.09685\" target=\"_blank\"\u003Eoriginal LoRA experiments\u003C/a\u003E, the method reduced the number of trainable parameters and GPU memory requirements substantially compared with full fine-tuning. Those results established LoRA as an efficient adaptation technique, while later research has examined how \u003Ca href=\"https://arxiv.org/abs/2402.12354\" target=\"_blank\"\u003Erank allocation\u003C/a\u003E and \u003Ca href=\"https://arxiv.org/abs/2303.10512\" target=\"_blank\"\u003Eoptimizer settings\u003C/a\u003E affect LoRA performance under different parameter budgets and model configurations.\u003C/p\u003E\n\u003Cp\u003ELoRA is useful when teams need portable adapters, several task-specific versions of one base model or lower training overhead. Its performance should still be compared with the original model and, for high-value workloads, with a full fine-tuning baseline.\u003C/p\u003E\n\u003Ch3\u003EQLoRA\u003C/h3\u003E\n\u003Cp\u003EQuantized low-rank adaptation (QLoRA) reduces memory requirements further by loading the frozen base model in a 4-bit quantized representation and backpropagating gradients through it into LoRA adapters.\u003C/p\u003E\n\u003Cp\u003E\u003Ca href=\"https://arxiv.org/abs/2305.14314\" target=\"_blank\"\u003EThe QLoRA paper\u003C/a\u003E demonstrated fine-tuning of a 65-billion-parameter model on a single 48-GB GPU in its experimental setup. Quantization made the base model less expensive to hold in memory during training, while the adapters retained higher-precision trainable parameters.\u003C/p\u003E\n\u003Cp\u003EQLoRA is especially relevant when GPU memory limits the feasible model size. Quantization can introduce its own implementation tradeoffs, so teams should test training stability, throughput and task performance under the configuration they plan to deploy.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_build-and-evaluate-the-fine-tuning-data-set":{"id":"title-v2-cdb9e9666d","additionalClasses":"anchor-title anchor-title--build-and-evaluate-the-fine-tuning-data-set","type":"heading2","lines":["Build and evaluate the fine-tuning data set"],":type":"snowflake-site/components/title-v2"},"text_build-and-evaluate-the-fine-tuning-data-set_0":{"id":"text-9a5b1c4d45","text":"\u003Cp\u003EModel selection receives much of the attention in a fine-tuning project, although the training and evaluation data usually determine whether the adaptation achieves the original objective.\u003C/p\u003E\n\u003Cp\u003EBegin with a specific behavior that can be observed and scored. “Improve responses” leaves too much room for interpretation, for example, while “return one of 18 approved categories and a confidence score in valid JSON” defines a result that an evaluation can measure.\u003C/p\u003E\n\u003Cp\u003EBefore training, run the base model against a held-out test set using the strongest reasonable prompt and any retrieval components planned for production. This baseline shows which errors fine-tuning needs to reduce. It also prevents a team from attributing gains to training when a prompt, parser or retrieval adjustment would have produced the same result.\u003C/p\u003E\n\u003Cp\u003ETraining examples should reflect the distribution the model will encounter after deployment. Common cases establish the primary pattern, while edge cases, ambiguous inputs and negative examples define its boundaries. When production inputs include incomplete records, long documents or conflicting terminology, the fine-tuning data should represent those conditions.\u003C/p\u003E\n\u003Cp\u003ESeparate data supports separate decisions:\u003C/p\u003E\n\u003Cul\u003E\n\u003Cli\u003ETraining data supplies the examples used for parameter updates.\u003C/li\u003E\n\u003Cli\u003EValidation data guides checkpoint selection and hyperparameter choices.\u003C/li\u003E\n\u003Cli\u003ETest data measures the selected model against examples that didn’t influence training decisions.\u003C/li\u003E\n\u003C/ul\u003E\n\u003Cp\u003EThe split should also account for near-duplicates and related records. When versions of the same document appear across training and test sets, evaluation results can overstate the model’s ability to generalize.\u003C/p\u003E\n\u003Cp\u003EFor generative tasks, exact-match accuracy rarely tells the whole story. Evaluation may need to measure schema validity, field-level accuracy, label precision and recall, unsupported claims, instruction adherence, latency and cost. Human review remains useful when the desired output depends on domain judgment, although reviewers need a shared rubric and representative samples.\u003C/p\u003E\n\u003Cp\u003EA fine-tuned model should also be tested for regressions. Stronger performance on the target task can accompany weaker behavior on adjacent inputs, especially when the training data is narrow. Comparing the tuned model with the base model across both target and general evaluations surfaces those trade-offs before deployment.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"callout_build-and-evaluate-the-fine-tuning-data-set_0":{"id":"text-8b8245a750","additionalClasses":"callout callout--tip","text":"\u003Cp\u003E\u003Cstrong\u003EQUICK TIP\u003C/strong\u003E\u003C/p\u003E\n\u003Cp\u003EBuild the test set before expanding the training set. A stable, held-out evaluation gives the team a consistent way to decide which new examples are actually improving the model.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"},"title_when-to-fine-tune":{"id":"title-v2-8a497009d5","additionalClasses":"anchor-title anchor-title--when-to-fine-tune","type":"heading2","lines":["When to fine-tune"],":type":"snowflake-site/components/title-v2"},"text_when-to-fine-tune_0":{"id":"text-23afffb68c","text":"\u003Cp\u003EFine-tuning is a strong candidate when the system shows a stable, repeated error pattern and the desired correction can be demonstrated through representative examples.\u003C/p\u003E\n\u003Cp\u003ETypical conditions include:\u003C/p\u003E\n\u003Cul\u003E\n\u003Cli\u003EThe task runs frequently enough to justify training and maintenance.\u003C/li\u003E\n\u003Cli\u003EPrompt revisions have stopped producing meaningful improvement.\u003C/li\u003E\n\u003Cli\u003EOutputs must follow a controlled schema, taxonomy or response convention.\u003C/li\u003E\n\u003Cli\u003EThe organization has reviewed examples that represent production inputs.\u003C/li\u003E\n\u003Cli\u003EThe task is stable enough for today’s examples to remain useful.\u003C/li\u003E\n\u003Cli\u003EA smaller specialized model could reduce inference latency or cost.\u003C/li\u003E\n\u003Cli\u003EThe team can measure target-task gains and broader regressions.\u003C/li\u003E\n\u003C/ul\u003E\n\u003Cp\u003EBut other conditions point elsewhere. When the application lacks current or private information, for example, retrieval addresses the information path directly. When requirements are still changing, application-level instructions are easier to revise than a training data set and model. When no reliable evaluation exists, the team cannot determine whether fine-tuning improved the system or merely changed it.\u003C/p\u003E\n\u003Cp\u003EA small experiment can resolve uncertainty. Establish a baseline, fine-tune on a carefully selected sample and compare the result on held-out data. If the gain is small, the team has avoided committing to a larger training pipeline. If the gain is material, the same evaluation framework can guide data expansion and deployment.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"callout_when-to-fine-tune_0":{"id":"text-3b36bbfbf4","additionalClasses":"callout callout--warning","text":"\u003Cp\u003E\u003Cstrong\u003ECOMMON PITFALL\u003C/strong\u003E\u003C/p\u003E\n\u003Cp\u003EA common mistake is fine-tuning while requirements are still changing. Updating prompts or retrieval logic is usually faster and easier to reverse than rebuilding a training set and producing another model version.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"},"title_account-for-the-full-cost-of-fine-tuning":{"id":"title-v2-c53dc23e19","additionalClasses":"anchor-title anchor-title--account-for-the-full-cost-of-fine-tuning","type":"heading2","lines":["Account for the full cost of fine-tuning"],":type":"snowflake-site/components/title-v2"},"text_account-for-the-full-cost-of-fine-tuning_0":{"id":"text-1c1a5c2031","text":"\u003Cp\u003ETraining compute is only one part of fine-tuning cost. Data preparation often requires domain experts to define labels, review completions, resolve disagreements and add cases that the first training run exposed. Evaluation then repeats across models, checkpoints, prompts and adapter settings.\u003C/p\u003E\n\u003Cp\u003EThe technical cost includes several components:\u003C/p\u003E\n\u003Cul\u003E\n\u003Cli\u003EProcessing the training tokens across one or more epochs\u003C/li\u003E\n\u003Cli\u003EStoring base models, adapters or complete tuned checkpoints\u003C/li\u003E\n\u003Cli\u003ERunning experiments that don’t produce a deployable model\u003C/li\u003E\n\u003Cli\u003EHosting the model for online or batch inference\u003C/li\u003E\n\u003Cli\u003EMonitoring production quality and collecting new examples\u003C/li\u003E\n\u003Cli\u003ERetraining when inputs, policies or task definitions change\u003C/li\u003E\n\u003C/ul\u003E\n\u003Cp\u003EInference economics should be evaluated alongside training cost. A specialized small model may require an upfront fine-tuning investment while reducing the cost of each subsequent request. A large model with a carefully designed prompt may cost more per call while avoiding the operational work of maintaining a tuned model. Request volume, latency requirements and expected model lifetime determine which trade-off is favorable.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_fine-tuning-and-model-training-with-snowflake":{"id":"title-v2-f01f372b90","additionalClasses":"anchor-title anchor-title--fine-tuning-and-model-training-with-snowflake","type":"heading2","lines":["Fine-tuning and model training with Snowflake"],":type":"snowflake-site/components/title-v2"},"text_fine-tuning-and-model-training-with-snowflake_0":{"id":"text-84c0234069","text":"\u003Cp\u003EFor supported LLMs, \u003Ca href=\"https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-finetuning\"\u003ESnowflake Cortex Fine-tuning\u003C/a\u003E provides a managed parameter-efficient fine-tuning workflow using training and validation data stored in Snowflake. Teams supply prompt-completion examples and select a supported base model, then Snowflake runs the training job, stores the resulting adapter as a model object and makes the fine-tuned model available for inference. Because the source data and model are governed as Snowflake objects, organizations can apply their existing role-based access controls throughout the workflow. Cortex Fine-tuning also supports `arctic-extract` models, allowing teams to adapt document extraction to their own document formats and fields.\u003C/p\u003E\r\n\u003Cp\u003ECustom fine-tuning and broader model-training workloads can run through \u003Ca href=\"https://www.snowflake.com/en/product/features/end-to-end-ml-workflows/\"\u003ESnowflake ML\u003C/a\u003E. Snowflake Container Runtime provides preconfigured, customizable environments on Snowpark Container Services for interactive experimentation, hyperparameter tuning, fine-tuning and batch training, using CPU or GPU compute. Teams can work with open source frameworks and libraries in Snowflake Notebooks, then use Snowflake ML Jobs to run the same code as a repeatable, resource-intensive workload from a notebook, local development environment or orchestration pipeline. For models and data sets that exceed a single node, Snowflake ML includes distributed trainers for frameworks such as PyTorch, XGBoost and LightGBM.\u003C/p\u003E\r\n\u003Cp\u003EThe training run sits within a larger model lifecycle. Snowflake Feature Store helps teams define reusable features and produce point-in-time-correct training data, while Snowflake ML Experiments records parameters, metrics and artifacts so practitioners can compare training runs systematically. After selection, Snowflake Model Registry manages model versions, metadata and access, then supports batch or real-time inference through Snowflake compute. ML lineage and observability capabilities extend that record into production by connecting models to their source data and tracking performance or drift over time.\u003C/p\u003E\r\n\u003Cp\u003E\u003Ca href=\"https://www.snowflake.com/en/product/snowflake-coco/\"\u003ESnowflake CoCo\u003C/a\u003E can assist across those stages through a natural-language interface. It can explore training data, engineer features, train and evaluate models, submit workloads to compute pools, record experiment results and prepare selected models for deployment, while practitioners review the objective and approve key decisions.\u003C/p\u003E\r\n\u003Cp\u003ETogether, these capabilities allow teams to select a managed adaptation path for supported models or retain deeper control over custom training, while keeping data preparation, experimentation, model management and production operations within the same governed environment.\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"title_selecting-the-right-adaptation-strategy":{"id":"title-v2-98e175329e","additionalClasses":"anchor-title anchor-title--selecting-the-right-adaptation-strategy","type":"heading2","lines":["Selecting the right adaptation strategy"],":type":"snowflake-site/components/title-v2"},"text_selecting-the-right-adaptation-strategy_0":{"id":"text-2356aa1465","text":"\u003Cp\u003EFine-tuning should begin with an observed system behavior, a representative data set and an evaluation that can distinguish improvement from change. The choice among supervised fine-tuning, preference optimization and other objectives follows from the type of feedback available; the choice among full fine-tuning, LoRA and QLoRA follows from the scope of adaptation and the available compute.\u003C/p\u003E\n\u003Cp\u003EPrompting and retrieval remain part of the design. A tuned model may still need explicit instructions, and a specialized model may still require current records from a retrieval system. By assigning each mechanism a defined responsibility, teams can improve model behavior without expecting parameter updates to solve every source of error.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"},"callout_selecting-the-right-adaptation-strategy_0":{"id":"text-fcc85b3428","additionalClasses":"callout callout--general","text":"\u003Cp\u003E\u003Cstrong\u003EKEY TAKEAWAY\u003C/strong\u003E\u003C/p\u003E\n\u003Cp\u003EA fine-tuned model isn’t automatically a better model. Its value depends on whether it outperforms a strong baseline on representative data without introducing unacceptable regressions, costs or maintenance requirements.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular text-color-text-05"}},":itemsOrder":["callout__0","text__0","title_what-is-fine-tuning","text_what-is-fine-tuning_0","yt_what-is-fine-tuning_0","title_what-fine-tuning-can-change","text_what-fine-tuning-can-change_0","card_v2_what-fine-tuning-can-change_0","title_fine-tuning-vs-prompting-rag-and-continued-pretraining","text_fine-tuning-vs-prompting-rag-and-continued-pretraining_0","title_choose-the-fine-tuning-objective","text_choose-the-fine-tuning-objective_0","title_choose-which-parameters-to-update","text_choose-which-parameters-to-update_0","title_build-and-evaluate-the-fine-tuning-data-set","text_build-and-evaluate-the-fine-tuning-data-set_0","callout_build-and-evaluate-the-fine-tuning-data-set_0","title_when-to-fine-tune","text_when-to-fine-tune_0","callout_when-to-fine-tune_0","title_account-for-the-full-cost-of-fine-tuning","text_account-for-the-full-cost-of-fine-tuning_0","title_fine-tuning-and-model-training-with-snowflake","text_fine-tuning-and-model-training-with-snowflake_0","title_selecting-the-right-adaptation-strategy","text_selecting-the-right-adaptation-strategy_0","callout_selecting-the-right-adaptation-strategy_0"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-medium",":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},"flexible_column_content_container_2":{"additionalClasses":"hub-sidebar","layout":"SIMPLE","id":"hub-body-aside",":items":{"container":{"additionalClasses":"sticky-sidebar","layout":"RESPONSIVE_GRID","columnCount":12,"columnClassNames":{"text_943981956_copy_":"aem-GridColumn aem-GridColumn--default--12","text_copy":"aem-GridColumn aem-GridColumn--default--12"},"gridClassNames":"aem-Grid aem-Grid--12 aem-Grid--default--12","id":"container-767af28adf",":items":{"text_943981956_copy_":{"id":"text-4604e03c0a","additionalClasses":"eyebrow-text","text":"\u003Cp\u003EIn This Guide\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-regular"},"text_copy":{"id":"text-34e341ac27","additionalClasses":"page-toc","text":"\u003Cul\u003E\u003Cli data-anchor=\"what-is-fine-tuning\"\u003EWhat is fine-tuning?\u003C/li\u003E\u003Cli data-anchor=\"what-fine-tuning-can-change\"\u003EWhat fine-tuning can change\u003C/li\u003E\u003Cli data-anchor=\"fine-tuning-vs-prompting-rag-and-continued-pretraining\"\u003EFine-tuning vs. prompting, RAG and continued pretraining\u003C/li\u003E\u003Cli data-anchor=\"choose-the-fine-tuning-objective\"\u003EChoose the fine-tuning objective\u003C/li\u003E\u003Cli data-anchor=\"choose-which-parameters-to-update\"\u003EChoose which parameters to update\u003C/li\u003E\u003Cli data-anchor=\"build-and-evaluate-the-fine-tuning-data-set\"\u003EBuild and evaluate the fine-tuning data set\u003C/li\u003E\u003Cli data-anchor=\"when-to-fine-tune\"\u003EWhen to fine-tune\u003C/li\u003E\u003Cli data-anchor=\"account-for-the-full-cost-of-fine-tuning\"\u003EAccount for the full cost of fine-tuning\u003C/li\u003E\u003Cli data-anchor=\"fine-tuning-and-model-training-with-snowflake\"\u003EFine-tuning and model training with Snowflake\u003C/li\u003E\u003Cli data-anchor=\"selecting-the-right-adaptation-strategy\"\u003ESelecting the right adaptation strategy\u003C/li\u003E\u003C/ul\u003E","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-size-small text-color-text-05"}},":itemsOrder":["text_943981956_copy_","text_copy"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-medium",":type":"snowflake-site/components/container"}},":itemsOrder":["container"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-small",":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false},"flexible_column_cont_1786318617":{"id":"flexible-column-container-626c4d457f","propertiesId":"hub-faq","type":"2-column-40-60","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"large","bottomPadding":"large","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"hub-faq-intro",":items":{"title_v2_copy":{"id":"title-v2-b812d54bbf","additionalClasses":"hub-faq__headline","type":"heading2","lines":["Frequently Asked Questions"],":type":"snowflake-site/components/title-v2","appliedCssClassNames":"left-alignment"},"text_copy":{"id":"text-a45b9b488c","additionalClasses":"hub-faq__subheadline","text":"\u003Cp\u003EYour common questions about fine-tuning, answered by Snowflake experts.\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"}},":itemsOrder":["title_v2_copy","text_copy"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-extra-small",":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},"flexible_column_content_container_2":{"layout":"SIMPLE","id":"hub-faq-accordions",":items":{"simple_snowflake_acc":{"id":"simple-snowflake-accordion-d6e860e6b5","additionalClasses":"seo-hub__faqs","showDivider":false,"accordionItemsList":[{"title":"Does fine-tuning add new knowledge to a model?","richText":"\u003Cp\u003EFine-tuning can teach patterns, terminology and associations contained in the training examples, so a model may perform better on questions related to that data. It doesn’t provide a dependable or easily updated knowledge store. RAG is generally better suited to current, private or attributable information because the source content is supplied during inference.\u003C/p\u003E"},{"title":"How much data is needed to fine-tune an LLM?","richText":"\u003Cp\u003EThe required amount varies with the task, base model, diversity of the inputs and quality of the examples. A narrow classification or formatting task may show gains with hundreds of carefully reviewed examples, while broader behavior changes can require much more data. Teams should begin with a representative sample, evaluate it against a baseline and add examples that address observed errors.\u003C/p\u003E"},{"title":"When should fine-tuning and RAG be used together?","richText":"\u003Cp\u003EThe combination is useful when an application needs consistent task behavior and access to external information. Fine-tuning can teach the model how to structure an answer or use retrieved context, while RAG supplies the records needed for a particular request. Retrieval-aware methods such as RAFT extend this approach by including relevant and distracting documents in the training examples.\u003C/p\u003E"},{"title":"What is the difference between LoRA and QLoRA?","richText":"\u003Cp\u003ELoRA freezes the base model and trains low-rank adapter matrices. QLoRA also uses LoRA adapters, while holding the frozen base model in a quantized representation to reduce memory consumption during training. LoRA may suit environments with sufficient GPU memory and a simpler training stack; QLoRA can make larger models feasible when memory is the primary constraint.\u003C/p\u003E"}],":type":"snowflake-site/components/simple-snowflake-accordion"}},":itemsOrder":["simple_snowflake_acc"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_1467213961":{"id":"flexible-column-container-e9698838f1","propertiesId":"hub-explore-resources-header","type":"1-column","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"large","bottomPadding":"none","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"container-327a424b5c",":items":{"title_v2":{"id":"title-v2-337d9dd283","additionalClasses":"hub-explore-resources-header__headline","type":"heading2","lines":["Explore AI Resources"],":type":"snowflake-site/components/title-v2","appliedCssClassNames":"left-alignment"}},":itemsOrder":["title_v2"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_912630531":{"id":"flexible-column-container-f023f7860c","propertiesId":"hub-explore-resources-grid","type":"1-column","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"small","bottomPadding":"large","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"hub-explore-resources-grid-inner",":items":{"resource_chip_0":{"id":"content-chip-39e7b27caf","tagText":"BLOG","tagColor":"#29B5E8","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"https://www.snowflake.com/en/blog/engineering/arctic-cookbook-series-instruction-tuning-arctic/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_EXTERNAL","text":"Read more"},"headline":{"id":"title","type":"heading5","lines":["Snowflake Arctic Cookbook Series: Instruction-Tuning Arctic"],":type":"snowflake-site/components/title-v2"},"appliedCssClassNames":"snowflake-content-chip-white-bg",":type":"snowflake-site/components/content-chip"},"resource_chip_1":{"id":"content-chip-fb3d3316a6","tagText":"BLOG","tagColor":"#29B5E8","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"https://www.snowflake.com/en/blog/engineering/fine-tune-llms-snowflake-arctictraining-ml-jobs/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_EXTERNAL","text":"Read more"},"headline":{"id":"title","type":"heading5","lines":["From General-Purpose to Domain Expert: Fine-Tune LLMs Directly in Snowflake"],":type":"snowflake-site/components/title-v2"},"appliedCssClassNames":"snowflake-content-chip-white-bg",":type":"snowflake-site/components/content-chip"},"resource_chip_2":{"id":"content-chip-5d5dda8d54","tagText":"GUIDE","tagColor":"#71D3DC","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"https://www.snowflake.com/en/developers/guides/finetuning-llm-using-snowflake-cortex-ai/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_EXTERNAL","text":"Read more"},"headline":{"id":"title-v2-0e2470c51d","type":"heading5","lines":["Serverless LLM Fine-tuning using Snowflake Cortex AI"],":type":"snowflake-site/components/title-v2"},"appliedCssClassNames":"snowflake-content-chip-white-bg",":type":"snowflake-site/components/content-chip"},"resource_chip_3":{"id":"content-chip-6a167e7e28","tagText":"BLOG","tagColor":"#29B5E8","cta":{"id":"cta","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"https://www.snowflake.com/en/blog/engineering/arctic-extract-fine-tuning/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_EXTERNAL","text":"Read more"},"headline":{"id":"title","type":"heading5","lines":["Achieving Human-Level Document Extraction with AI_EXTRACT Fine-Tuning"],":type":"snowflake-site/components/title-v2"},"appliedCssClassNames":"snowflake-content-chip-white-bg",":type":"snowflake-site/components/content-chip"}},":itemsOrder":["resource_chip_0","resource_chip_1","resource_chip_2","resource_chip_3"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-gray-10-bg"},"flexible_column_cont_1158003461":{"id":"flexible-column-container-a4a75f2a02","propertiesId":"hub-explore-topics-header","type":"1-column","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"large","bottomPadding":"none","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"container-a995b2f1fd",":items":{"title_v2_copy_copy":{"id":"title-v2-359b11e240","additionalClasses":"hub-explore-topics-header__headline","type":"heading2","lines":["Explore AI Topics"],":type":"snowflake-site/components/title-v2","appliedCssClassNames":"left-alignment"},"text_copy_copy":{"id":"text-6cb85e42a7","additionalClasses":"hub-explore-topics-header__subheadline","text":"\u003Cp\u003EDeep dives into every aspect of artificial intelligence\u003C/p\u003E\r\n","richText":true,":type":"snowflake-site/components/text","appliedCssClassNames":"text-color-text-05"}},":itemsOrder":["title_v2_copy_copy","text_copy_copy"],"appliedCssClassNames":"snowflake-responsive-container-inner-padding-extra-small",":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-white-bg"},"flexible_column_cont_1377146023":{"id":"flexible-column-container-fef36afb17","propertiesId":"hub-explore-topics-grid","type":"3-column-even","alignColumns":"top","containerMaxWidth":"extra-large","topPadding":"small","bottomPadding":"large","spaceBetween":"small","reverseOnMobile":false,"carouselOnMobile":false,"backgroundImageOption":"none","flexible_column_content_container_1":{"layout":"SIMPLE","id":"container-d3495c04ba",":items":{"topic_card_0":{"id":"card-v2-2008ecb6cb","configurationStatus":{"configured":true,"message":""},":type":"snowflake-site/components/card-v2","title":{"id":"title","type":"heading4","lines":["MLOps"],":type":"snowflake-site/components/title-v2"},"button":{"id":"button","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/mlops/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Learn more"},"type":"content-card","text":{"id":"text","text":"\u003Cp\u003EMLOps is the operating discipline that keeps the data, features, infrastructure, approvals and feedback loops behind production ML working together.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text"},"layoutStyle":"vertical"}},":itemsOrder":["topic_card_0"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},"flexible_column_content_container_2":{"layout":"SIMPLE","id":"container-6dbde133de",":items":{"topic_card_1":{"id":"card-v2-5f16cfa27c","configurationStatus":{"configured":true,"message":""},":type":"snowflake-site/components/card-v2","title":{"id":"title","type":"heading4","lines":["Model Evaluation"],":type":"snowflake-site/components/title-v2"},"button":{"id":"button","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/model-evaluation/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Learn more"},"type":"content-card","text":{"id":"text","text":"\u003Cp\u003EModel evaluation helps teams validate performance on unseen data, analyze errors and determine whether a model is ready for production.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text"},"layoutStyle":"vertical"}},":itemsOrder":["topic_card_1"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},"flexible_column_content_container_3":{"layout":"SIMPLE","id":"container-0286447090",":items":{"topic_card_2":{"id":"card-v2-fffa0031a9","configurationStatus":{"configured":true,"message":""},":type":"snowflake-site/components/card-v2","title":{"id":"title","type":"heading4","lines":["Model Training"],":type":"snowflake-site/components/title-v2"},"button":{"id":"button","showOutboundIcon":false,"buttonLink":{"valid":true,"url":"/en/artificial-intelligence/machine-learning/model-training/"},"linkTargetContentType":"GENERIC",":type":"snowflake-site/components/button","linkType":"SNOWFLAKE_INTERNAL","text":"Learn more"},"type":"content-card","text":{"id":"text","text":"\u003Cp\u003EModel training turns quality data, optimization choices, and production constraints into reliable ML systems that can scale.\u003C/p\u003E","richText":true,":type":"snowflake-site/components/text"},"layoutStyle":"vertical"}},":itemsOrder":["topic_card_2"],":type":"snowflake-site/components/flexible-column-container/flexible-column-content-container"},":type":"snowflake-site/components/flexible-column-container","isBlogPage":false,"isActiveTOC":false,"appliedCssClassNames":"snowflake-flexible-column-container-white-bg"}},":itemsOrder":["flexible_column_cont","flexible_column_cont_939100716","flexible_column_cont_1398138236","flexible_column_cont_663228916","flexible_column_cont_1786318617","flexible_column_cont_1467213961","flexible_column_cont_912630531","flexible_column_cont_1158003461","flexible_column_cont_1377146023"],":type":"wcm/foundation/components/responsivegrid"},"modal_container":{"layout":"SIMPLE","id":"container-85e4e5a37c",":items":{},":itemsOrder":[],":type":"snowflake-site/components/modal/modal-container"},"markup_editor_928258845":{"id":"markup-editor-3de868aeee","title":" ","cssContent":".snowflake-flexible-column-container-gray-10-bg\u003E.snowflake-flexible-column-container{background-color:var(--ui-background-05) !important}.text-size-regular:has(.seo-hub-hero__related-topic-label){display:flex;align-items:center}.hub-hero__headline span{text-transform:none !important}.hub-hero__subheadline p{max-width:70ch;margin-top:8px}.hub-hero__authors \u003E .container \u003E .cmp-container \u003E .aem-container{display:flex;flex-direction:row}.hub-hero__authors \u003E .container \u003E .cmp-container \u003E .aem-container \u003E div{width:auto !important;margin:24px 48px 0 0 !important}.hub-hero__authors .heading-5-v2{gap:var(--spacing-00)}.hub-hero__authors .snowflake-content-chip-button{display:none !important}.hub-hero__authors .snowflake-person-chip-content .body-2,.hub-hero__authors .snowflake-content-chip-content .snowflake-title-v2-line{font-size:16px !important;line-height:20px !important;font-family:\"Lato\",sans-serif !important;color:#000 !important;font-weight:600 !important}.hub-hero__authors .snowflake-person-chip-content .body-3,.hub-hero__authors .snowflake-content-chip-content .snowflake-title-v2-line:not(:first-child){font-weight:400 !important;color:var(--text-05) !important;font-size:16px !important}.hub-hero__authors .snowflake-image-container img{aspect-ratio:1 !important;border-radius:100%;overflow:hidden}.hub-hero__authors .snowflake-person-chip-avatar{width:56px;height:56px}.hub-hero__authors .snowflake-content-chip{align-items:center;display:inline-flex}.hub-hero__authors .snowflake-content-chip-image{max-width:56px;line-height:0;margin-right:var(--spacing-03)}.hub-hero__authors .snowflake-person-chip-inner-horizontal{gap:var(--spacing-03)}@media screen and (min-width:1367px){.hub-hero__headline .heading-1-v2{font-size:48px;line-height:44px}}#hub-explore-resources-grid-inner\u003E.container\u003E.cmp-container\u003E.aem-container{display:flex;flex-direction:row;flex-wrap:wrap;gap:24px}#hub-explore-resources-grid-inner\u003E.container\u003E.cmp-container\u003E.aem-container::before,#hub-explore-resources-grid-inner\u003E.container\u003E.cmp-container\u003E.aem-container::after{display:none}#hub-explore-resources-grid-inner\u003E.container\u003E.cmp-container\u003E.aem-container\u003Ediv{width:calc(25% - 18px)}.text-color-text-05 .snowflake-text h2,.text-color-text-05.cq-Editable-dom h2,.text-color-text-05 .snowflake-text h3,.text-color-text-05.cq-Editable-dom h3,.text-color-text-05 .snowflake-text h4,.text-color-text-05.cq-Editable-dom h4,.text-color-text-05 .snowflake-text h5,.text-color-text-05.cq-Editable-dom h5,.text-color-text-05 .snowflake-text h6,.text-color-text-05.cq-Editable-dom h6{color:#000 !important}",":type":"snowflake-site/components/markup-editor","isGSAPEnabled":false},"markup_editor_597730182":{"id":"markup-editor-7609b55bc5","title":" ","cssContent":".sf-copy-markdown [data-copy-md]{display:inline-flex;align-items:center;gap:6px;padding:8px 16px;font-family:'Texta',sans-serif;text-transform:uppercase;font-weight:800 !important;font-size:14px;font-weight:500;color:#11567f;background:#f0faff;border:1px solid #b8e6f9;border-radius:24px;cursor:pointer;transition:background .2s ease,border-color .2s ease,color .2s ease}.sf-copy-markdown [data-copy-md]:hover{background:#ddf3fc;border-color:#29b5e8}.sf-copy-markdown [data-copy-md][data-copied=\"1\"]{color:#0f7b3e;background:#ecfdf5;border-color:#6ee7a0;pointer-events:none}.sf-copy-markdown [data-copy-md] svg{flex-shrink:0}.longform-conten .snowflake-content-chip-white-bg .snowflake-content-chip{box-shadow:0 0 24px 4px rgba(0,0,0,.02),0 4px 8px 0 rgba(0,0,0,.04);flex-direction:row-reverse;align-items:center}.longform-conten .snowflake-content-chip-button{display:none}.longform-conten .snowflake-content-chip-image__inner{aspect-ratio:5 / 3;display:flex;justify-content:center;align-items:center;background-color:var(--ui-01);border-radius:4px}.longform-conten .snowflake-content-chip-image{margin-right:0;margin-left:48px}.longform-conten .snowflake-content-chip-image img{width:50%;border-radius:0 !important;object-fit:contain}.longform-content .black-blue-text-color .snowflake-title-v2-line:not(:first-child){font-size:14px !important;font-weight:400 !important;color:rgba(0,0,0,.6) !important;margin-top:8px !important}.page-toc ul li:first-child{padding-top:0 !important}.page-toc ul li:last-child{padding-bottom:0 !important}.seo-hub__top-bar \u003E .container \u003E .cmp-container \u003E .aem-container \u003E div:first-child{flex-grow:1}.sf-copy-markdown{margin-top:40px !important}.page-toc ul{margin-top:16px !important}.seo-hub__top-bar \u003E .container \u003E .cmp-container \u003E .aem-container{display:flex;justify-content:space-between}.sf-copy-markdown{}.callout.snowflake-text p:not(:first-child){margin-top:var(--spacing-01)}.callout \u003E span \u003E p:first-child \u003E strong,.callout \u003E span \u003E p:first-child \u003E b{text-transform:uppercase;font-family:'Texta',sans-serif;font-size:16px !important;color:var(--ui-01) !important}.seo-hub-hero__subheadline p{max-width:50ch}.tag-group ul{list-style-type:none;padding:0;margin:0;display:flex;flex-direction:row;row-gap:12px;column-gap:8px;align-items:center;flex-wrap:wrap}.tag-group ul li:first-child{flex-shrink:0}.tag-group ul li a{display:inline-block;padding:2px 12px;border-radius:48px;background-color:#ededed;color:#666;font-size:14px !important}@media screen and (min-width:1367px){.seo-hub-hero__headline span.snowflake-title-v2-line{font-size:56px !important}}.callout.snowflake-text p:not(:first-child){margin-top:var(--spacing-01)}.callout \u003E span \u003E p:first-child \u003E b{text-transform:uppercase;font-family:'Texta',sans-serif;font-size:16px !important;color:var(--ui-01) !important}.seo-hub-hero__subheadline p{max-width:80ch}#hero:has(.snowflake-youtube-lite) .seo-hub-hero__subheadline p{max-width:50ch}.tag-group ul{list-style-type:none;padding:0;margin:0;display:flex;flex-direction:row;row-gap:12px;column-gap:8px;align-items:center;flex-wrap:wrap}.tag-group ul li:first-child{width:100%;flex-shrink:0}.tag-group ul li a{display:inline-block;padding:2px 12px;border-radius:48px;background-color:#ededed;color:#666;font-size:14px !important}@media screen and (min-width:1367px){.seo-hub-hero__headline span.snowflake-title-v2-line{font-size:56px !important}}",":type":"snowflake-site/components/markup-editor","isGSAPEnabled":false},"markup_editor":{"id":"markup-editor-8a2f1b5d97","title":" ","cssContent":"div.snowflake-breadcrumb a.snowflake-breadcrumb-item,.snowflake-breadcrumb div.snowflake-breadcrumb-item{text-transform:none;font-weight:500}.snowflake-breadcrumb svg{display:none !important}.snowflake-breadcrumb a:has(svg)::after{content:'/';margin:0 12px;color:#666}.hub-sidebar{padding:0 40px}.sticky-sidebar{max-width:340px;margin-left:auto}.page-toc ul{list-style-type:none;padding:0}.page-toc li{padding:8px 16px;border-left:4px solid var(--ui-01);cursor:pointer;transition:300ms ease all}.page-toc li:hover{color:var(--ui-01);border-color:#7fd3f1;transition:300ms ease all}.callout,.customer-card{background-color:#eef9fd;border-left:4px solid var(--ui-01);padding:24px 24px 24px 32px;border-radius:4px}.logo-container{max-width:180px}.longform-content li{margin-top:1rem !important}div.longform-content p{max-width:80ch}.bolder .snowflake-title-v2-line{font-weight:900 !important}.border-top\u003Ediv{border-top:1px solid #ccc;padding-top:48px}.related-topics ul{list-style-type:none;padding:0;margin:0;display:flex;gap:8px;flex-wrap:wrap}.related-topics li{display:inline-block;border:1px solid #ccc;padding:4px 12px;border-radius:24px}div.longform-content .snowflake-text h2,div.longform-content .snowflake-text .heading-2-v2,div.longform-content .snowflake-text h3,div.longform-content .snowflake-text .heading-3-v2,div.longform-content .snowflake-title-v2 .heading-3-v2,div.longform-content .snowflake-text h4,div.longform-content .snowflake-text .heading-4-v2,div.longform-content .snowflake-title-v2 .heading-4-v2,div.longform-content .snowflake-text h5,div.longform-content .snowflake-text .heading-5-v2,div.longform-content .snowflake-title-v2 .heading-5-v2,div.longform-content .snowflake-text h6,div.longform-content .snowflake-title-v2 .heading-6-v2,div.longform-content .snowflake-text .heading-6-v2{text-transform:none !important}div.longform-content .snowflake-text h2,div.longform-content .snowflake-text .heading-2-v2,div.longform-content .snowflake-text h3,div.longform-content .snowflake-text .heading-3-v2,div.longform-content .snowflake-text h4,div.longform-content .snowflake-text .heading-4-v2,div.longform-content .snowflake-text h5,div.longform-content .snowflake-text .heading-5-v2,div.longform-content .snowflake-text h6,div.longform-content .snowflake-text .heading-6-v2{margin-top:1.5rem !important;line-height:1.1 !important}div.longform-content .snowflake-text h3,div.longform-content .snowflake-text .heading-3-v2,div.longform-content .snowflake-title-v2 .heading-3-v2,div.longform-content .snowflake-text h4,div.longform-content .snowflake-text .heading-4-v2,div.longform-content .snowflake-title-v2 .heading-4-v2,div.longform-content .snowflake-text h5,div.longform-content .snowflake-text .heading-5-v2,div.longform-content .snowflake-title-v2 .heading-5-v2,div.longform-content .snowflake-text h6,div.longform-content .snowflake-text .heading-6-v2,div.longform-content .snowflake-title-v2 .heading-6-v2{font-family:Lato,sans-serif !important;font-weight:800 !important}div.longform-content .snowflake-text h2,div.longform-content .snowflake-text .heading-2-v2,div.longform-content .snowflake-title-v2 .heading-2-v2{text-transform:none !important;font-size:28px !important}div.longform-content .snowflake-text h3,div.longform-content .snowflake-text .heading-3-v2,div.longform-content .snowflake-title-v2 .heading-3-v2{font-size:22px !important}div.longform-content .snowflake-text h4,div.longform-content .snowflake-text .heading-4-v2,div.longform-content .snowflake-title-v2 .heading-4-v2{font-size:18px !important}div.longform-content .snowflake-text h5,div.longform-content .snowflake-text .heading-5-v2,div.longform-content .snowflake-title-v2 .heading-5-v2{font-size:16px !important}div.longform-content .snowflake-text h6,div.longform-content .snowflake-text .heading-6-v2,div.longform-content .snowflake-title-v2 .heading-6-v2{font-size:14px !important}@media screen and (min-width:992px){div.longform-content .snowflake-text h2,div.longform-content .snowflake-text .heading-2-v2,div.longform-content .snowflake-title-v2 .heading-2-v2{font-size:38px !important}div.longform-content .snowflake-text h3,div.longform-content .snowflake-text .heading-3-v2,div.longform-content .snowflake-title-v2 .heading-3-v2{font-size:26px !important}div.longform-content .snowflake-text h4,div.longform-content .snowflake-text .heading-4-v2,div.longform-content .snowflake-title-v2 .heading-4-v2{font-size:22px !important}div.longform-content .snowflake-text h5,div.longform-content .snowflake-text .heading-5-v2,div.longform-content .snowflake-title-v2 .heading-5-v2{font-size:18px !important}div.longform-content .snowflake-text h6,div.longform-content .snowflake-text .heading-6-v2,div.longform-content .snowflake-title-v2 .heading-6-v2{font-size:16px !important}}.sticky-sidebar .page-toc li.is-active{font-weight:600;color:var(--snow-blue,#29b5e8)}.sticky-sidebar .page-toc li[data-anchor]{cursor:pointer}.longform-content table{margin-top:24px;margin-bottom:24px;width:100%;background-color:var(--ui-background-01);border-collapse:collapse;border:2px solid var(--ui-background-09);font-family:'Lato',sans-serif;color:var(--ui-background-09)}.longform-content table thead{background-color:var(--ui-01)}.longform-content th,.longform-content td{min-width:120px;border:2px solid var(--ui-background-09);padding:var(--spacing-01)}.longform-content ol{margin-top:0 !important}.longform-content ol li{margin-bottom:1rem !important}.longform-content ul li{margin:0;padding:0 0 0 32px;position:relative}.longform-content ul{list-style-type:none}.longform-content ul li::before{content:\"\";display:block;border-radius:100%;background:#29b5e8;width:18px;height:18px;position:absolute;top:4px;left:0;border:5px solid #e5f2f7;box-sizing:border-box}.seo-customer.snowflake-card-v2-advanced-horizontal .snowflake-card-v2-advanced-image-container{max-width:200px}.seo-customer.snowflake-card-v2-advanced-horizontal .snowflake-card-v2-advanced-image-container img{object-fit:contain}.related-topics-outer-container \u003E .container \u003E .cmp-container \u003E .aem-container{display:flex;flex-direction:row}.related-topics-outer-container \u003E .container \u003E .cmp-container \u003E .aem-container \u003E div{width:auto !important;margin:0 !important}.related-topics-outer-container \u003E .container \u003E .cmp-container \u003E .aem-container \u003E div:first-child{margin-right:16px !important;flex-shrink:0}","jsContent":"(function(){var OFFSET=100;if(window.gsap&&window.ScrollTrigger){gsap.registerPlugin(ScrollTrigger);var sidebar=document.querySelector('.sticky-sidebar');var body=document.querySelector('.longform-content');if(sidebar&&body){ScrollTrigger.create({trigger:sidebar,start:'top 100px',endTrigger:body,end:'bottom bottom',pin:sidebar,pinSpacing:false});}}document.addEventListener('click',function(e){var li=e.target.closest('li[data-anchor]');if(!li)return;var slug=li.getAttribute('data-anchor');var heading=document.querySelector('.anchor-title--'+CSS.escape(slug));if(!heading)return;e.preventDefault();var top=heading.getBoundingClientRect().top+window.pageYOffset-OFFSET;window.scrollTo({top:top,behavior:'smooth'});history.replaceState(null,'','#'+slug);},false);var headings=document.querySelectorAll('[class*=\"anchor-title--\"]');if(headings.length&&'IntersectionObserver'in window){var io=new IntersectionObserver(function(entries){entries.forEach(function(entry){if(!entry.isIntersecting)return;var cls=Array.from(entry.target.classList).find(function(c){return c.indexOf('anchor-title--')===0;});if(!cls)return;var slug=cls.replace('anchor-title--','');document.querySelectorAll('li[data-anchor]').forEach(function(li){li.classList.toggle('is-active',li.getAttribute('data-anchor')===slug);});});},{rootMargin:'-20% 0px -70% 0px',threshold:0});headings.forEach(function(h){io.observe(h);});}})();",":type":"snowflake-site/components/markup-editor","isGSAPEnabled":true},"experiencefragment-footer":{"id":"experiencefragment-785aacbada","localizedFragmentVariationPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/footer/master/jcr:content","configured":true,":type":"snowflake-site/components/experiencefragment","xfModelPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/footer/master.xfmodel.json"},"experiencefragment":{"id":"experiencefragment-ce4c31dcb1","localizedFragmentVariationPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/footer-legal-disclaimers/master/jcr:content","configured":true,":type":"snowflake-site/components/experiencefragment","xfModelPath":"/content/experience-fragments/snowflake-site/language-masters/en/site/footer-legal-disclaimers/master.xfmodel.json"}},":itemsOrder":["experiencefragment-banner","experiencefragment-header","responsivegrid","modal_container","markup_editor_928258845","markup_editor_597730182","markup_editor","experiencefragment-footer","experiencefragment"],":type":"wcm/foundation/components/responsivegrid"}},":itemsOrder":["root"],":mappedPath":"/en/artificial-intelligence/machine-learning/model-training/fine-tuning/",":type":"snowflake-site/components/structure/page",":hierarchyType":"page",":path":"/content/snowflake-site/global/en/artificial-intelligence/machine-learning/model-training/fine-tuning","isPasswordProtected":false,"analyticsContentTags":[],"analyticsEnabled":true,"coveoConfig":{"searchHub":"snowflake.com","organizationId":"snowflakecomputingproduction8neljofn","apiKey":"xx335921a6-2a0a-40f2-a167-e390b4766c3d","pipeline":"snowflake.com"},"analyticsDebugMode":false,"analyticsData":{"excludeFromAnalytics":false,"subCategory":"","pageType":"homepage","templateName":"base-page-template54","siteName":"snowflake","pageUrl":"/content/snowflake-site/global/en/artificial-intelligence/machine-learning/model-training/fine-tuning","language":"en","category":"general","pageName":"Fine-Tuning Has Changed: A Practitioner’s Guide to Modern Model Adaptation","contentTags":[]},"locale":"en"}