Safetensors
HelixWorld-preview / inference_config.json
lintsinghua's picture
noizai
HelixWorld Preview v1
0f6aa9f
Raw History Blame Contribute Delete
996 Bytes
{
"schema_version": 1,
"model_family": "HelixWorld joint audio-video camera/action",
"deployment": "helixworld_preview_v1",
"dtype": "bfloat16",
"model": {
"repo_id": "NoizAI/HelixWorld-preview",
"filename": "weights/model.safetensors",
"checkpoint_files": 1
},
"text_encoder": {
"repo_id": "google/gemma-3-12b-it-qat-q4_0-unquantized",
"revision": "68f7ee4fbd59087436ada77ed2d62f373fdd4482"
},
"generation": {
"width": 768,
"height": 512,
"fps": 24,
"generated_frames_per_turn": 121,
"audio_video_joint_generation": true
},
"caption": {
"component_order": ["caption_video", "caption_audio", "caption_av"],
"canonical_template": "Video description:\n{caption_video}\n\nAudio description:\n{caption_audio}\n\nJoint audio-visual description:\n{caption_av}"
},
"controls": {
"camera_action_required": true,
"supported_keys": ["W", "A", "S", "D", "left", "right", "up", "down"],
"combination_separator": "+"
}
}