| { | |
| "model_type": "open-ogma", | |
| "architecture": { | |
| "variant": "transformer", | |
| "d_embed": 64, | |
| "d_model": 128, | |
| "n_layers": 2, | |
| "n_heads": 2, | |
| "vocab_size": 30007, | |
| "max_seq_len": 1024, | |
| "matryoshka_dims": [ | |
| 32, | |
| 64, | |
| 128 | |
| ], | |
| "pooling": "mean", | |
| "d_output": 128, | |
| "ffn_mult": 2.6666666666666665, | |
| "conv_kernel_size": 7, | |
| "spatial_rank": 32, | |
| "n_random_features": 128, | |
| "dropout": 0.0, | |
| "pad_id": 0, | |
| "unk_id": 1, | |
| "bos_id": 2, | |
| "eos_id": 3, | |
| "qry_id": 4, | |
| "doc_id": 5, | |
| "sym_id": 6, | |
| "n_special_tokens": 7 | |
| }, | |
| "heads": { | |
| "proj_small": 384, | |
| "proj_large": 1024 | |
| }, | |
| "default_head": "small384", | |
| "matryoshka_dims": [ | |
| 32, | |
| 64, | |
| 128 | |
| ], | |
| "train_step": 686610, | |
| "small_mteb_at_save": 0.5270759827049664 | |
| } |