HACAModel
HACAModel copied to clipboard
chitwansaharia
Implementation of "Watch, Listen, and Describe: Globally and Locally Aligned Cross-Modal Attentions for Video Captioning" (https://arxiv.org/abs/1804.05448)
Hello, i wish to reproduce your results. Can you tell me how you extracted the features from the dataset? Regards, Apostolos.
when i run it, it doesn't utilize multiple gpus, though torch sees 2 gpus. how to parallelize the model to use multiple gpus?
Metadata
Owner
Metadata
Implementation of "Watch, Listen, and Describe: Globally and Locally Aligned Cross-Modal Attentions for Video Captioning" (https://arxiv.org/abs/1804.05448)