| { |
| "image_processor": { |
| "discrete_image_end_token": "<|discrete_image_end|>", |
| "discrete_image_ratios": [ |
| [ |
| 1, |
| 1 |
| ], |
| [ |
| 1, |
| 2 |
| ], |
| [ |
| 2, |
| 1 |
| ], |
| [ |
| 3, |
| 4 |
| ], |
| [ |
| 4, |
| 3 |
| ], |
| [ |
| 9, |
| 16 |
| ], |
| [ |
| 16, |
| 9 |
| ], |
| [ |
| 1, |
| 3 |
| ], |
| [ |
| 3, |
| 1 |
| ] |
| ], |
| "discrete_image_start_token": "<|discrete_image_start|>", |
| "discrete_image_token": "<|DISCRETE_IMAGE_PAD|>", |
| "discrete_token_size": 27, |
| "do_center_crop": false, |
| "do_convert_rgb": true, |
| "do_normalize": true, |
| "do_rescale": true, |
| "do_resize": true, |
| "image_end_token": "<|image_end|>", |
| "image_mean": [ |
| 0.48145466, |
| 0.4578275, |
| 0.40821073 |
| ], |
| "image_processor_type": "Qwen2VLImageProcessor", |
| "image_start_token": "<|image_start|>", |
| "image_std": [ |
| 0.26862954, |
| 0.26130258, |
| 0.27577711 |
| ], |
| "image_token": "<|IMAGE_PAD|>", |
| "merge_size": 2, |
| "patch_size": 14, |
| "resample": 3, |
| "rescale_factor": 0.00392156862745098, |
| "resize_mode": "smart_resize", |
| "size": { |
| "longest_edge": 2073600, |
| "shortest_edge": 3136 |
| }, |
| "temporal_patch_size": 2, |
| "use_discrete_token": false, |
| "vision_eof_token": "<|vision_eof|>", |
| "vision_eol_token": "<|vision_eol|>" |
| }, |
| "processor_class": "HyperCLOVAXVisionV2Processor", |
| "video_processor": { |
| "do_convert_rgb": true, |
| "do_normalize": true, |
| "do_rescale": true, |
| "do_resize": true, |
| "do_sample_frames": false, |
| "image_mean": [ |
| 0.48145466, |
| 0.4578275, |
| 0.40821073 |
| ], |
| "image_std": [ |
| 0.26862954, |
| 0.26130258, |
| 0.27577711 |
| ], |
| "max_frames": 120, |
| "max_video_tokens": 128000, |
| "merge_size": 2, |
| "min_frames": 4, |
| "patch_size": 14, |
| "resample": 3, |
| "rescale_factor": 0.00392156862745098, |
| "return_metadata": false, |
| "size": { |
| "longest_edge": 285768, |
| "shortest_edge": 3136 |
| }, |
| "temporal_patch_size": 2, |
| "use_audio_in_video": false, |
| "use_discrete_token": false, |
| "video_audio_token": "<|VIDEO_AUDIO_PAD|>", |
| "video_end_token": "<|video_end|>", |
| "video_processor_type": "Qwen2VLVideoProcessor", |
| "video_start_token": "<|video_start|>", |
| "video_token": "<|VIDEO_PAD|>", |
| "vision_eof_token": "<|vision_eof|>", |
| "vision_eol_token": "<|vision_eol|>" |
| } |
| } |
|
|