commaVQ - GPT2M
A GPT2M model trained on a larger version of the commaVQ dataset.
This model is able to generate driving video unconditionally.
Below is an example of 5 seconds of imagined video using GPT2M.
A GPT2M model trained on a larger version of the commaVQ dataset.
This model is able to generate driving video unconditionally.
Below is an example of 5 seconds of imagined video using GPT2M.
pytorch_model.bin 1.4 GBgpt2m.onnx 614 MBgpt2m_share_buffer.onnx 614 MBtemporal_decoder.onnx 310 MBdecoder.onnx 171 MBdecoder_pytorch_model.bin 171 MBencoder.onnx 120 MBencoder_pytorch_model.bin 119 MB