{ "model_type": "parvulus", "architectures": [ "ParvulusModel" ], "base_channels": 48, "channel_mults": [ 1, 2, 2, 4 ], "num_res_blocks": 1, "cond_dim": 256, "action_dim": 7, "chunk_size": 8, "tokens_per_layer": 4, "resolution": 256, "num_ode_steps": 10, "text_encoder": { "name": "t5", "model_name": "google-t5/t5-small", "revision": "df1b051c49625cf57a3d0d8d3863ed4d13564fe4", "cond_dim": 256, "t5_d_model": 512 }, "auxiliary_decoders": { "video_decoder": { "out_channels": 6, "num_res_blocks": 1 }, "self_decoder": { "out_channels": 6, "num_res_blocks": 1 }, "change_decoder": { "out_channels": 2, "num_res_blocks": 1 } }, "action_scale_file": "action_scale.json", "torch_dtype": "float32" }