VideoPoet

{{Short description|Text-to-video model by Google}}

{{Use mdy dates|date=February 2024}}

{{Infobox software

| screenshot = {{multiple image

| border = infobox

| perrow = 2

| total_width = 220

| caption_align = center

| image1 = Dog popcorn with audio 6781BD0C.webm

| caption1 = "A dog eating popcorn at the cinema"

| image2 = Drums with audio DA80C510.webm

| caption2 = "A teddy bear with a cap, sunglasses, and leather jacket playing drums"

}}

| caption = Example videos generated by the model from texts

| developer = Google

| genre = Large language model

| released = {{start date and age|2024|02|08}}

| platform =

}}

VideoPoet is a large language model developed by Google Research in 2023 for video making.{{Cite web |last=Krithika |first=K. L. |date=2023-12-20 |title=Google Unveils VideoPoet, a New LLM for Video Generation |url=https://analyticsindiamag.com/google-unveils-videopoet-a-new-llm-for-video-generation/ |access-date=2024-04-29 |website=Analytics India Magazine |language=en-US}}{{cite arXiv|title=VideoPoet: A Large Language Model for Zero-Shot Video Generation|first1=Dan|last1=Kondratyuk|first2=Lijun|last2=Yu|first3=Xiuye|last3=Gu|first4=José|last4=Lezama|first5=Jonathan|last5=Huang|first6=Rachel|last6=Hornung|first7=Hartwig|last7=Adam|first8=Hassan|last8=Akbari|first9=Yair|last9=Alon|first10=Vighnesh|last10=Birodkar|first11=Yong|last11=Cheng|first12=Ming-Chang|last12=Chiu|first13=Josh|last13=Dillon|first14=Irfan|last14=Essa|first15=Agrim|last15=Gupta|first16=Meera|last16=Hahn|first17=Anja|last17=Hauth|first18=David|last18=Hendon|first19=Alonso|last19=Martinez|first20=David|last20=Minnen|first21=David|last21=Ross|first22=Grant|last22=Schindler|first23=Mikhail|last23=Sirotenko|first24=Kihyuk|last24=Sohn|first25=Krishna|last25=Somandepalli|first26=Huisheng|last26=Wang|first27=Jimmy|last27=Yan|first28=Ming-Hsuan|last28=Yang|first29=Xuan|last29=Yang|first30=Bryan|last30=Seybold|first31=Lu|last31=Jiang|date=December 21, 2023|class=cs.CV|eprint=2312.14125}}{{Cite news |date=December 21, 2023 |title=Google has introduced VideoPOET breaking new ground in coherent video generation |url=https://www.gizmochina.com/2023/12/21/google-videopoet-10-second-coherent-video-generation/ |work=Gizmochina}}{{Cite web |title=VideoPoet |url=https://sites.research.google/videopoet/ |access-date=2024-04-29 |website=Google Research |language=en}} It can be asked to animate still images.{{Cite news |last=Franzen |first=Carl |date=December 20, 2023 |title=Google's new multimodal AI video generator VideoPoet looks incredible |url=https://venturebeat.com/ai/googles-new-videopoet-multimodal-ai-video-generation-model-looks-incredible/ |work=VentureBeat}} The model accepts text, images, and videos as inputs, with a program to add feature for any input to any format generated content. VideoPoet was publicly announced on December 19, 2023. It uses an autoregressive language model.