pos-polish-gpt2-small / config.json
nie3e's picture
pos-polish-gpt2-small
6b31691 verified
{
"_name_or_path": "polish-gpt2-small",
"activation_function": "gelu_fast",
"architectures": [
"GPT2ForTokenClassification"
],
"attn_pdrop": 0.1,
"bos_token_id": 0,
"embd_pdrop": 0.1,
"eos_token_id": 2,
"id2label": {
"0": "numcol",
"1": "subst",
"2": "num",
"3": "comp",
"4": "xxx",
"5": "bedzie",
"6": "interp",
"7": "winien",
"8": "interj",
"9": "siebie",
"10": "conj",
"11": "pcon",
"12": "ppas",
"13": "burk",
"14": "imps",
"15": "ppron12",
"16": "pant",
"17": "adj",
"18": "prep",
"19": "adjp",
"20": "fin",
"21": "pred",
"22": "ger",
"23": "praet",
"24": "adjc",
"25": "ppron3",
"26": "adv",
"27": "aglt",
"28": "brev",
"29": "depr",
"30": "inf",
"31": "adja",
"32": "pact",
"33": "qub",
"34": "impt"
},
"initializer_range": 0.02,
"label2id": {
"adj": 17,
"adja": 31,
"adjc": 24,
"adjp": 19,
"adv": 26,
"aglt": 27,
"bedzie": 5,
"brev": 28,
"burk": 13,
"comp": 3,
"conj": 10,
"depr": 29,
"fin": 20,
"ger": 22,
"imps": 14,
"impt": 34,
"inf": 30,
"interj": 8,
"interp": 6,
"num": 2,
"numcol": 0,
"pact": 32,
"pant": 16,
"pcon": 11,
"ppas": 12,
"ppron12": 15,
"ppron3": 25,
"praet": 23,
"pred": 21,
"prep": 18,
"qub": 33,
"siebie": 9,
"subst": 1,
"winien": 7,
"xxx": 4
},
"layer_norm_epsilon": 1e-05,
"model_type": "gpt2",
"n_embd": 768,
"n_head": 12,
"n_inner": 3072,
"n_layer": 12,
"n_positions": 2048,
"reorder_and_upcast_attn": false,
"resid_pdrop": 0.1,
"scale_attn_by_inverse_layer_idx": false,
"scale_attn_weights": true,
"summary_activation": null,
"summary_first_dropout": 0.1,
"summary_proj_to_labels": true,
"summary_type": "cls_index",
"summary_use_proj": true,
"tokenizer_class": "GPT2TokenizerFast",
"torch_dtype": "float32",
"transformers_version": "4.36.2",
"use_cache": true,
"vocab_size": 51200
}