Hi everyone, I need help finding a dataset of images annotated with human actions<\/strong> [such as sitting+in-chair, working+on-laptop<\/strong>, etc.]. I found a model capable of generating such tags on Huggingface here<\/a>, however I was unable to locate its source dataset.<\/p>\n Just for context, I am trying to create a fine-tuned ViT model<\/a>, that incorporates as broad a set of visual tags as possible. My plan is to optimize this model for edge devices [using Quantization aware training + TFLite model<\/a> conversion] and open-source the weights<\/strong>. Eventually, I am hoping this can be used for a broad range of visual search\/tagging\/QnA tasks. Currently, I am training the model on top 2500 Danbooru tags<\/a> + MIT SUN<\/a> indoor location tags.<\/p>\n An online demo<\/a> of the model can be found here. If anyone has any suggestions regarding what other dataset\/tags to add, or would like to help with the training efforts, please drop a line. I would really appreciate it.<\/p>\n [Disclosures: I am not affiliated in any way with any of the HuggingFace \/Arxiv\/Mit.edu links I posted here. The link to the online-demo is maintained by me, but there are no ads or anything else that procures me financial gain on it.]<\/p>\n<\/div>\n submitted by \/u\/DisintegratingBo <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
