| { | |
| "_name_or_path": "ckpt-last", | |
| "architectures": [ | |
| "SparseBertMiniLM" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "classifier_dropout": null, | |
| "gradient_checkpointing": false, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 768, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "layer_norm_eps": 1e-12, | |
| "layer_skip": 1, | |
| "max_position_embeddings": 512, | |
| "model_type": "bert", | |
| "num_attention_heads": 12, | |
| "num_hidden_layers": 12, | |
| "num_relation_heads": 32, | |
| "pad_token_id": 0, | |
| "position_embedding_type": "absolute", | |
| "sparsity": "90", | |
| "sparsity_map": { | |
| "0": { | |
| "head": {}, | |
| "hidden": { | |
| "-1": 0, | |
| "0": 0, | |
| "1": 0, | |
| "10": 0, | |
| "11": 0, | |
| "2": 0, | |
| "3": 0, | |
| "4": 0, | |
| "5": 0, | |
| "6": 0, | |
| "7": 0, | |
| "8": 0, | |
| "9": 0 | |
| }, | |
| "neuron": {} | |
| }, | |
| "10": { | |
| "head": { | |
| "1": 5, | |
| "11": 1, | |
| "3": 1 | |
| }, | |
| "hidden": { | |
| "-1": 38, | |
| "0": 38, | |
| "1": 38, | |
| "10": 38, | |
| "11": 38, | |
| "2": 38, | |
| "3": 38, | |
| "4": 38, | |
| "5": 38, | |
| "6": 38, | |
| "7": 38, | |
| "8": 38, | |
| "9": 38 | |
| }, | |
| "neuron": { | |
| "0": 190, | |
| "1": 298, | |
| "11": 1, | |
| "2": 1323, | |
| "3": 4, | |
| "4": 5, | |
| "5": 8, | |
| "6": 4, | |
| "7": 7, | |
| "8": 2, | |
| "9": 1 | |
| } | |
| }, | |
| "20": { | |
| "head": { | |
| "0": 1, | |
| "1": 7, | |
| "10": 1, | |
| "11": 4, | |
| "3": 2, | |
| "9": 1 | |
| }, | |
| "hidden": { | |
| "-1": 84, | |
| "0": 84, | |
| "1": 84, | |
| "10": 84, | |
| "11": 84, | |
| "2": 84, | |
| "3": 84, | |
| "4": 84, | |
| "5": 84, | |
| "6": 84, | |
| "7": 84, | |
| "8": 84, | |
| "9": 84 | |
| }, | |
| "neuron": { | |
| "0": 597, | |
| "1": 802, | |
| "11": 16, | |
| "2": 2294, | |
| "3": 48, | |
| "4": 78, | |
| "5": 123, | |
| "6": 22, | |
| "7": 47, | |
| "8": 23, | |
| "9": 5 | |
| } | |
| }, | |
| "30": { | |
| "head": { | |
| "0": 1, | |
| "1": 8, | |
| "10": 2, | |
| "11": 4, | |
| "2": 1, | |
| "3": 2, | |
| "6": 1, | |
| "7": 1, | |
| "8": 1, | |
| "9": 2 | |
| }, | |
| "hidden": { | |
| "-1": 123, | |
| "0": 123, | |
| "1": 123, | |
| "10": 123, | |
| "11": 123, | |
| "2": 123, | |
| "3": 123, | |
| "4": 123, | |
| "5": 123, | |
| "6": 123, | |
| "7": 123, | |
| "8": 123, | |
| "9": 123 | |
| }, | |
| "neuron": { | |
| "0": 980, | |
| "1": 1233, | |
| "10": 2, | |
| "11": 116, | |
| "2": 2512, | |
| "3": 132, | |
| "4": 233, | |
| "5": 386, | |
| "6": 61, | |
| "7": 146, | |
| "8": 73, | |
| "9": 24 | |
| } | |
| }, | |
| "40": { | |
| "head": { | |
| "0": 3, | |
| "1": 8, | |
| "10": 2, | |
| "11": 5, | |
| "2": 1, | |
| "3": 3, | |
| "4": 2, | |
| "5": 1, | |
| "6": 2, | |
| "7": 1, | |
| "8": 1, | |
| "9": 4 | |
| }, | |
| "hidden": { | |
| "-1": 177, | |
| "0": 177, | |
| "1": 177, | |
| "10": 177, | |
| "11": 177, | |
| "2": 177, | |
| "3": 177, | |
| "4": 177, | |
| "5": 177, | |
| "6": 177, | |
| "7": 177, | |
| "8": 177, | |
| "9": 177 | |
| }, | |
| "neuron": { | |
| "0": 1436, | |
| "1": 1664, | |
| "10": 6, | |
| "11": 518, | |
| "2": 2634, | |
| "3": 306, | |
| "4": 449, | |
| "5": 732, | |
| "6": 138, | |
| "7": 321, | |
| "8": 184, | |
| "9": 91 | |
| } | |
| }, | |
| "50": { | |
| "head": { | |
| "0": 5, | |
| "1": 8, | |
| "10": 3, | |
| "11": 5, | |
| "2": 2, | |
| "3": 3, | |
| "4": 3, | |
| "5": 2, | |
| "6": 3, | |
| "7": 2, | |
| "8": 2, | |
| "9": 4 | |
| }, | |
| "hidden": { | |
| "-1": 223, | |
| "0": 223, | |
| "1": 223, | |
| "10": 223, | |
| "11": 223, | |
| "2": 223, | |
| "3": 223, | |
| "4": 223, | |
| "5": 223, | |
| "6": 223, | |
| "7": 223, | |
| "8": 223, | |
| "9": 223 | |
| }, | |
| "neuron": { | |
| "0": 1749, | |
| "1": 1933, | |
| "10": 21, | |
| "11": 915, | |
| "2": 2699, | |
| "3": 519, | |
| "4": 664, | |
| "5": 976, | |
| "6": 252, | |
| "7": 469, | |
| "8": 307, | |
| "9": 187 | |
| } | |
| }, | |
| "60": { | |
| "head": { | |
| "0": 6, | |
| "1": 8, | |
| "10": 5, | |
| "11": 7, | |
| "2": 3, | |
| "3": 4, | |
| "4": 4, | |
| "5": 3, | |
| "6": 5, | |
| "7": 2, | |
| "8": 2, | |
| "9": 4 | |
| }, | |
| "hidden": { | |
| "-1": 284, | |
| "0": 284, | |
| "1": 284, | |
| "10": 284, | |
| "11": 284, | |
| "2": 284, | |
| "3": 284, | |
| "4": 284, | |
| "5": 284, | |
| "6": 284, | |
| "7": 284, | |
| "8": 284, | |
| "9": 284 | |
| }, | |
| "neuron": { | |
| "0": 2065, | |
| "1": 2200, | |
| "10": 67, | |
| "11": 1392, | |
| "2": 2762, | |
| "3": 817, | |
| "4": 991, | |
| "5": 1279, | |
| "6": 454, | |
| "7": 695, | |
| "8": 521, | |
| "9": 397 | |
| } | |
| }, | |
| "70": { | |
| "head": { | |
| "0": 6, | |
| "1": 8, | |
| "10": 6, | |
| "11": 7, | |
| "2": 4, | |
| "3": 6, | |
| "4": 4, | |
| "5": 5, | |
| "6": 6, | |
| "7": 3, | |
| "8": 5, | |
| "9": 5 | |
| }, | |
| "hidden": { | |
| "-1": 346, | |
| "0": 346, | |
| "1": 346, | |
| "10": 346, | |
| "11": 346, | |
| "2": 346, | |
| "3": 346, | |
| "4": 346, | |
| "5": 346, | |
| "6": 346, | |
| "7": 346, | |
| "8": 346, | |
| "9": 346 | |
| }, | |
| "neuron": { | |
| "0": 2266, | |
| "1": 2390, | |
| "10": 206, | |
| "11": 1745, | |
| "2": 2820, | |
| "3": 1173, | |
| "4": 1308, | |
| "5": 1537, | |
| "6": 694, | |
| "7": 951, | |
| "8": 791, | |
| "9": 708 | |
| } | |
| }, | |
| "80": { | |
| "head": { | |
| "0": 6, | |
| "1": 9, | |
| "10": 8, | |
| "11": 8, | |
| "2": 6, | |
| "3": 6, | |
| "4": 6, | |
| "5": 9, | |
| "6": 7, | |
| "7": 4, | |
| "8": 5, | |
| "9": 5 | |
| }, | |
| "hidden": { | |
| "-1": 422, | |
| "0": 422, | |
| "1": 422, | |
| "10": 422, | |
| "11": 422, | |
| "2": 422, | |
| "3": 422, | |
| "4": 422, | |
| "5": 422, | |
| "6": 422, | |
| "7": 422, | |
| "8": 422, | |
| "9": 422 | |
| }, | |
| "neuron": { | |
| "0": 2450, | |
| "1": 2541, | |
| "10": 574, | |
| "11": 2105, | |
| "2": 2863, | |
| "3": 1559, | |
| "4": 1676, | |
| "5": 1862, | |
| "6": 1071, | |
| "7": 1294, | |
| "8": 1154, | |
| "9": 1126 | |
| } | |
| }, | |
| "85": { | |
| "head": { | |
| "0": 6, | |
| "1": 9, | |
| "10": 10, | |
| "11": 8, | |
| "2": 6, | |
| "3": 6, | |
| "4": 7, | |
| "5": 9, | |
| "6": 8, | |
| "7": 7, | |
| "8": 5, | |
| "9": 7 | |
| }, | |
| "hidden": { | |
| "-1": 468, | |
| "0": 468, | |
| "1": 468, | |
| "10": 468, | |
| "11": 468, | |
| "2": 468, | |
| "3": 468, | |
| "4": 468, | |
| "5": 468, | |
| "6": 468, | |
| "7": 468, | |
| "8": 468, | |
| "9": 468 | |
| }, | |
| "neuron": { | |
| "0": 2554, | |
| "1": 2635, | |
| "10": 936, | |
| "11": 2278, | |
| "2": 2889, | |
| "3": 1780, | |
| "4": 1836, | |
| "5": 2037, | |
| "6": 1292, | |
| "7": 1489, | |
| "8": 1360, | |
| "9": 1401 | |
| } | |
| }, | |
| "90": { | |
| "head": { | |
| "0": 6, | |
| "1": 9, | |
| "10": 11, | |
| "11": 8, | |
| "2": 7, | |
| "3": 7, | |
| "4": 9, | |
| "5": 9, | |
| "6": 9, | |
| "7": 9, | |
| "8": 6, | |
| "9": 8 | |
| }, | |
| "hidden": { | |
| "-1": 522, | |
| "0": 522, | |
| "1": 522, | |
| "10": 522, | |
| "11": 522, | |
| "2": 522, | |
| "3": 522, | |
| "4": 522, | |
| "5": 522, | |
| "6": 522, | |
| "7": 522, | |
| "8": 522, | |
| "9": 522 | |
| }, | |
| "neuron": { | |
| "0": 2653, | |
| "1": 2713, | |
| "10": 1390, | |
| "11": 2433, | |
| "2": 2913, | |
| "3": 2013, | |
| "4": 2064, | |
| "5": 2243, | |
| "6": 1569, | |
| "7": 1741, | |
| "8": 1648, | |
| "9": 1688 | |
| } | |
| }, | |
| "95": { | |
| "head": { | |
| "0": 8, | |
| "1": 10, | |
| "10": 11, | |
| "11": 9, | |
| "2": 9, | |
| "3": 9, | |
| "4": 9, | |
| "5": 9, | |
| "6": 11, | |
| "7": 9, | |
| "8": 9, | |
| "9": 9 | |
| }, | |
| "hidden": { | |
| "-1": 599, | |
| "0": 599, | |
| "1": 599, | |
| "10": 599, | |
| "11": 599, | |
| "2": 599, | |
| "3": 599, | |
| "4": 599, | |
| "5": 599, | |
| "6": 599, | |
| "7": 599, | |
| "8": 599, | |
| "9": 599 | |
| }, | |
| "neuron": { | |
| "0": 2753, | |
| "1": 2799, | |
| "10": 2001, | |
| "11": 2654, | |
| "2": 2948, | |
| "3": 2337, | |
| "4": 2381, | |
| "5": 2519, | |
| "6": 1992, | |
| "7": 2148, | |
| "8": 2061, | |
| "9": 2161 | |
| } | |
| }, | |
| "96": { | |
| "head": { | |
| "0": 9, | |
| "1": 10, | |
| "10": 11, | |
| "11": 10, | |
| "2": 9, | |
| "3": 9, | |
| "4": 9, | |
| "5": 10, | |
| "6": 11, | |
| "7": 9, | |
| "8": 9, | |
| "9": 9 | |
| }, | |
| "hidden": { | |
| "-1": 614, | |
| "0": 614, | |
| "1": 614, | |
| "10": 614, | |
| "11": 614, | |
| "2": 614, | |
| "3": 614, | |
| "4": 614, | |
| "5": 614, | |
| "6": 614, | |
| "7": 614, | |
| "8": 614, | |
| "9": 614 | |
| }, | |
| "neuron": { | |
| "0": 2768, | |
| "1": 2822, | |
| "10": 2133, | |
| "11": 2689, | |
| "2": 2955, | |
| "3": 2392, | |
| "4": 2439, | |
| "5": 2579, | |
| "6": 2077, | |
| "7": 2223, | |
| "8": 2148, | |
| "9": 2266 | |
| } | |
| }, | |
| "97": { | |
| "head": { | |
| "0": 10, | |
| "1": 10, | |
| "10": 11, | |
| "11": 10, | |
| "2": 9, | |
| "3": 9, | |
| "4": 9, | |
| "5": 10, | |
| "6": 11, | |
| "7": 10, | |
| "8": 11, | |
| "9": 10 | |
| }, | |
| "hidden": { | |
| "-1": 637, | |
| "0": 637, | |
| "1": 637, | |
| "10": 637, | |
| "11": 637, | |
| "2": 637, | |
| "3": 637, | |
| "4": 637, | |
| "5": 637, | |
| "6": 637, | |
| "7": 637, | |
| "8": 637, | |
| "9": 637 | |
| }, | |
| "neuron": { | |
| "0": 2799, | |
| "1": 2846, | |
| "10": 2294, | |
| "11": 2746, | |
| "2": 2970, | |
| "3": 2487, | |
| "4": 2537, | |
| "5": 2656, | |
| "6": 2207, | |
| "7": 2355, | |
| "8": 2291, | |
| "9": 2409 | |
| } | |
| }, | |
| "98": { | |
| "head": { | |
| "0": 11, | |
| "1": 10, | |
| "10": 11, | |
| "11": 10, | |
| "2": 10, | |
| "3": 10, | |
| "4": 10, | |
| "5": 10, | |
| "6": 11, | |
| "7": 10, | |
| "8": 11, | |
| "9": 10 | |
| }, | |
| "hidden": { | |
| "-1": 660, | |
| "0": 660, | |
| "1": 660, | |
| "10": 660, | |
| "11": 660, | |
| "2": 660, | |
| "3": 660, | |
| "4": 660, | |
| "5": 660, | |
| "6": 660, | |
| "7": 660, | |
| "8": 660, | |
| "9": 660 | |
| }, | |
| "neuron": { | |
| "0": 2826, | |
| "1": 2870, | |
| "10": 2455, | |
| "11": 2794, | |
| "2": 2984, | |
| "3": 2588, | |
| "4": 2634, | |
| "5": 2726, | |
| "6": 2350, | |
| "7": 2501, | |
| "8": 2440, | |
| "9": 2535 | |
| } | |
| }, | |
| "99": { | |
| "head": { | |
| "0": 11, | |
| "1": 10, | |
| "10": 11, | |
| "11": 11, | |
| "2": 11, | |
| "3": 11, | |
| "4": 11, | |
| "5": 11, | |
| "6": 11, | |
| "7": 11, | |
| "8": 11, | |
| "9": 10 | |
| }, | |
| "hidden": { | |
| "-1": 691, | |
| "0": 691, | |
| "1": 691, | |
| "10": 691, | |
| "11": 691, | |
| "2": 691, | |
| "3": 691, | |
| "4": 691, | |
| "5": 691, | |
| "6": 691, | |
| "7": 691, | |
| "8": 691, | |
| "9": 691 | |
| }, | |
| "neuron": { | |
| "0": 2861, | |
| "1": 2915, | |
| "10": 2683, | |
| "11": 2862, | |
| "2": 2998, | |
| "3": 2709, | |
| "4": 2746, | |
| "5": 2845, | |
| "6": 2551, | |
| "7": 2678, | |
| "8": 2635, | |
| "9": 2695 | |
| } | |
| } | |
| }, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.9.2", | |
| "type_vocab_size": 2, | |
| "use_cache": true, | |
| "vocab_size": 30522 | |
| } | |