Upload model
Browse files- config.json +159 -0
- generation_config.json +7 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,159 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/xmod-base",
|
| 3 |
+
"adapter_layer_norm": false,
|
| 4 |
+
"adapter_reduction_factor": 2,
|
| 5 |
+
"adapter_reuse_layer_norm": true,
|
| 6 |
+
"adapters": {
|
| 7 |
+
"adapters": {
|
| 8 |
+
"en_XX": "9321733f16973c74"
|
| 9 |
+
},
|
| 10 |
+
"config_map": {
|
| 11 |
+
"9321733f16973c74": {
|
| 12 |
+
"adapter_residual_before_ln": false,
|
| 13 |
+
"cross_adapter": false,
|
| 14 |
+
"factorized_phm_W": true,
|
| 15 |
+
"factorized_phm_rule": false,
|
| 16 |
+
"hypercomplex_nonlinearity": "glorot-uniform",
|
| 17 |
+
"init_weights": "bert",
|
| 18 |
+
"inv_adapter": null,
|
| 19 |
+
"inv_adapter_reduction_factor": null,
|
| 20 |
+
"is_parallel": false,
|
| 21 |
+
"learn_phm": true,
|
| 22 |
+
"leave_out": [],
|
| 23 |
+
"ln_after": false,
|
| 24 |
+
"ln_before": false,
|
| 25 |
+
"mh_adapter": false,
|
| 26 |
+
"non_linearity": "gelu",
|
| 27 |
+
"original_ln_after": false,
|
| 28 |
+
"original_ln_before": true,
|
| 29 |
+
"output_adapter": true,
|
| 30 |
+
"phm_bias": true,
|
| 31 |
+
"phm_c_init": "normal",
|
| 32 |
+
"phm_dim": 4,
|
| 33 |
+
"phm_init_range": 0.0001,
|
| 34 |
+
"phm_layer": false,
|
| 35 |
+
"phm_rank": 1,
|
| 36 |
+
"reduction_factor": 2,
|
| 37 |
+
"residual_before_ln": false,
|
| 38 |
+
"scaling": 1.0,
|
| 39 |
+
"shared_W_phm": false,
|
| 40 |
+
"shared_phm_rule": true,
|
| 41 |
+
"use_gating": false
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"fusion_config_map": {},
|
| 45 |
+
"fusions": {}
|
| 46 |
+
},
|
| 47 |
+
"architectures": [
|
| 48 |
+
"XmodAdapterModel"
|
| 49 |
+
],
|
| 50 |
+
"attention_probs_dropout_prob": 0.1,
|
| 51 |
+
"bos_token_id": 0,
|
| 52 |
+
"classifier_dropout": null,
|
| 53 |
+
"custom_heads": {},
|
| 54 |
+
"default_language": null,
|
| 55 |
+
"eos_token_id": 2,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_dropout_prob": 0.1,
|
| 58 |
+
"hidden_size": 768,
|
| 59 |
+
"initializer_range": 0.02,
|
| 60 |
+
"intermediate_size": 3072,
|
| 61 |
+
"languages": [
|
| 62 |
+
"en_XX",
|
| 63 |
+
"id_ID",
|
| 64 |
+
"vi_VN",
|
| 65 |
+
"ru_RU",
|
| 66 |
+
"fa_IR",
|
| 67 |
+
"sv_SE",
|
| 68 |
+
"ja_XX",
|
| 69 |
+
"fr_XX",
|
| 70 |
+
"de_DE",
|
| 71 |
+
"ro_RO",
|
| 72 |
+
"ko_KR",
|
| 73 |
+
"hu_HU",
|
| 74 |
+
"es_XX",
|
| 75 |
+
"fi_FI",
|
| 76 |
+
"uk_UA",
|
| 77 |
+
"da_DK",
|
| 78 |
+
"pt_XX",
|
| 79 |
+
"no_XX",
|
| 80 |
+
"th_TH",
|
| 81 |
+
"pl_PL",
|
| 82 |
+
"bg_BG",
|
| 83 |
+
"nl_XX",
|
| 84 |
+
"zh_CN",
|
| 85 |
+
"he_IL",
|
| 86 |
+
"el_GR",
|
| 87 |
+
"it_IT",
|
| 88 |
+
"sk_SK",
|
| 89 |
+
"hr_HR",
|
| 90 |
+
"tr_TR",
|
| 91 |
+
"ar_AR",
|
| 92 |
+
"cs_CZ",
|
| 93 |
+
"lt_LT",
|
| 94 |
+
"hi_IN",
|
| 95 |
+
"zh_TW",
|
| 96 |
+
"ca_ES",
|
| 97 |
+
"ms_MY",
|
| 98 |
+
"sl_SI",
|
| 99 |
+
"lv_LV",
|
| 100 |
+
"ta_IN",
|
| 101 |
+
"bn_IN",
|
| 102 |
+
"et_EE",
|
| 103 |
+
"az_AZ",
|
| 104 |
+
"sq_AL",
|
| 105 |
+
"sr_RS",
|
| 106 |
+
"kk_KZ",
|
| 107 |
+
"ka_GE",
|
| 108 |
+
"tl_XX",
|
| 109 |
+
"ur_PK",
|
| 110 |
+
"is_IS",
|
| 111 |
+
"hy_AM",
|
| 112 |
+
"ml_IN",
|
| 113 |
+
"mk_MK",
|
| 114 |
+
"be_BY",
|
| 115 |
+
"la_VA",
|
| 116 |
+
"te_IN",
|
| 117 |
+
"eu_ES",
|
| 118 |
+
"gl_ES",
|
| 119 |
+
"mn_MN",
|
| 120 |
+
"kn_IN",
|
| 121 |
+
"ne_NP",
|
| 122 |
+
"sw_KE",
|
| 123 |
+
"si_LK",
|
| 124 |
+
"mr_IN",
|
| 125 |
+
"af_ZA",
|
| 126 |
+
"gu_IN",
|
| 127 |
+
"cy_GB",
|
| 128 |
+
"eo_EO",
|
| 129 |
+
"km_KH",
|
| 130 |
+
"ky_KG",
|
| 131 |
+
"uz_UZ",
|
| 132 |
+
"ps_AF",
|
| 133 |
+
"pa_IN",
|
| 134 |
+
"ga_IE",
|
| 135 |
+
"ha_NG",
|
| 136 |
+
"am_ET",
|
| 137 |
+
"lo_LA",
|
| 138 |
+
"ku_TR",
|
| 139 |
+
"so_SO",
|
| 140 |
+
"my_MM",
|
| 141 |
+
"or_IN",
|
| 142 |
+
"sa_IN"
|
| 143 |
+
],
|
| 144 |
+
"layer_norm_eps": 1e-05,
|
| 145 |
+
"ln_before_adapter": true,
|
| 146 |
+
"max_position_embeddings": 514,
|
| 147 |
+
"model_type": "xmod",
|
| 148 |
+
"num_attention_heads": 12,
|
| 149 |
+
"num_hidden_layers": 12,
|
| 150 |
+
"pad_token_id": 1,
|
| 151 |
+
"position_embedding_type": "absolute",
|
| 152 |
+
"pre_norm": false,
|
| 153 |
+
"prediction_heads": {},
|
| 154 |
+
"torch_dtype": "float32",
|
| 155 |
+
"transformers_version": "4.31.0",
|
| 156 |
+
"type_vocab_size": 1,
|
| 157 |
+
"use_cache": true,
|
| 158 |
+
"vocab_size": 250002
|
| 159 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 1,
|
| 6 |
+
"transformers_version": "4.31.0"
|
| 7 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a1d1e4f3568e32f1f2a9a0094b16b2f2179c6ecbf71147e4ab2f335829c5700
|
| 3 |
+
size 1140623721
|