| { | |
| "architecture_name": "tabicl_v2", | |
| "config": { | |
| "max_classes": 10, | |
| "num_quantiles": 999, | |
| "embed_dim": 128, | |
| "col_num_blocks": 3, | |
| "col_nhead": 8, | |
| "col_num_inds": 128, | |
| "col_affine": false, | |
| "col_feature_group": "same", | |
| "col_feature_group_size": 3, | |
| "col_target_aware": true, | |
| "col_ssmax": "qassmax-mlp-elementwise", | |
| "row_num_blocks": 3, | |
| "row_nhead": 8, | |
| "row_num_cls": 4, | |
| "row_rope_base": 100000, | |
| "row_rope_interleaved": false, | |
| "icl_num_blocks": 12, | |
| "icl_nhead": 8, | |
| "icl_ssmax": "qassmax-mlp-elementwise", | |
| "ff_factor": 2, | |
| "dropout": 0.0, | |
| "activation": "gelu", | |
| "norm_first": true, | |
| "bias_free_ln": false, | |
| "recompute": false | |
| } | |
| } | |