| seed: 42 | |
| data: | |
| root: data | |
| protocol: remoteclip_rgb_bpe_pairs_v1 | |
| image_size: 224 | |
| channels: 3 | |
| context_length: 77 | |
| vocabulary_size: 49408 | |
| sot_token_id: 49406 | |
| eot_token_id: 49407 | |
| pad_token_id: 0 | |
| train_samples: 8 | |
| test_samples: 4 | |
| num_semantic_groups: 2 | |
| model: | |
| image_size: 224 | |
| patch_size: 32 | |
| embed_dim: 64 | |
| vision_width: 64 | |
| vision_layers: 2 | |
| vision_heads: 4 | |
| text_width: 64 | |
| text_layers: 2 | |
| text_heads: 4 | |
| training: | |
| epochs: 1 | |
| batch_size: 4 | |
| learning_rate: 0.0005 | |
| weight_decay: 0.01 | |
| num_workers: 0 | |
| runtime: | |
| device: auto | |
| amp: true | |
| paths: | |
| checkpoint: result/checkpoints/remoteclip.pt | |
| training_metrics: result/training/metrics.json | |
| inference_dir: result/output | |
| evaluation_dir: result/evaluation | |