\n<\/p>\n
We labored with the studio to curate a high-quality dataset of Ross\u2019 private sketches, utilizing it to fine-tune our text-to-image mannequin, Imagen. By coaching the mannequin on the studio\u2019s chosen work, we have been in a position to incorporate the core parts of Ross\u2019 design language \u2014 the particular curves, structural logic and natural patterns, which allowed us to generate new ideas that have been rooted in Ross\u2019 distinctive fashion.<\/p>\n
From the outset, we approached the venture as a human-led inquiry. The studio decided the necessity to prioritize language alongside the visible dataset, working to decode and articulate Ross\u2019 design lexicon to successfully information the mannequin\u2019s output. We targeted on constructing a selected vocabulary that described the studio\u2019s work, realizing that the fitting prompts have been key to getting significant outcomes.<\/p>\n
All through the method, Lovegrove Studio noticed how the mannequin responded to particular phrases and used these insights to align the outputs towards the meant design outcomes. This dialogue between designer and AI was an important a part of the method. We paid shut consideration to how the mannequin interpreted sure phrases, utilizing that suggestions to refine prompts and steer the output nearer to the studio\u2019s imaginative and prescient. We challenged the mannequin to generate a chair with out ever utilizing the phrase \u201cchair,\u201d as a substitute utilizing artistic synonyms to supply extra numerous outputs and a richer exploration of kind and performance.<\/p>\n
We developed many ideas with this specialised mannequin and the Lovegrove Studio staff, then used Gemini to push the artistic exploration additional to ideate on supplies and visualize the chair from totally different varieties and viewpoints.<\/p>\n<\/div>\n\n","protected":false},"excerpt":{"rendered":"
Our strategy: distillation and dialogue We labored with the studio to curate a high-quality dataset of Ross\u2019 private sketches, utilizing it to fine-tune our text-to-image mannequin, Imagen. By coaching the mannequin on the studio\u2019s chosen work, we have been in a position to incorporate the core parts of Ross\u2019 design language \u2014 the particular curves, […]<\/p>\n","protected":false},"author":2,"featured_media":7248,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[55],"tags":[799,1113,81,5663,5664,381],"class_list":["post-7246","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-machine-learning","tag-deepmind","tag-design","tag-google","tag-lovegrove","tag-partner","tag-ross"],"_links":{"self":[{"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/posts\/7246","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7246"}],"version-history":[{"count":1,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/posts\/7246\/revisions"}],"predecessor-version":[{"id":7247,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/posts\/7246\/revisions\/7247"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=\/wp\/v2\/media\/7248"}],"wp:attachment":[{"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7246"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7246"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/techtrendfeed.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7246"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}