Upload config.json with huggingface_hub
Browse files- config.json +108 -0
config.json
ADDED
|
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "cadence_punctuation",
|
| 3 |
+
"auto_map": {
|
| 4 |
+
"AutoConfig": "modeling_gemma3_punctuation.Gemma3PunctuationConfig",
|
| 5 |
+
"AutoModelForTokenClassification": "modeling_gemma3_punctuation.Gemma3ForTokenClassification",
|
| 6 |
+
"AutoModel": "modeling_gemma3_punctuation.Gemma3ForTokenClassification"
|
| 7 |
+
},
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Gemma3ForTokenClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_bias": false,
|
| 12 |
+
"attention_dropout": 0.0,
|
| 13 |
+
"attn_logit_softcapping": null,
|
| 14 |
+
"bos_token_id": 2,
|
| 15 |
+
"cache_implementation": "hybrid",
|
| 16 |
+
"classifier_dropout_prob": 0.0,
|
| 17 |
+
"eos_token_id": 1,
|
| 18 |
+
"final_logit_softcapping": null,
|
| 19 |
+
"head_dim": 256,
|
| 20 |
+
"hidden_activation": "gelu_pytorch_tanh",
|
| 21 |
+
"hidden_size": 1152,
|
| 22 |
+
"id2label": {
|
| 23 |
+
"0": "O",
|
| 24 |
+
"1": ".",
|
| 25 |
+
"10": "\"",
|
| 26 |
+
"11": "\u0964",
|
| 27 |
+
"12": "(",
|
| 28 |
+
"13": ")",
|
| 29 |
+
"14": ":",
|
| 30 |
+
"15": "\u066c",
|
| 31 |
+
"16": "\u06d4",
|
| 32 |
+
"17": "\u061f",
|
| 33 |
+
"18": ".\"",
|
| 34 |
+
"19": ").",
|
| 35 |
+
"2": ",",
|
| 36 |
+
"20": "),",
|
| 37 |
+
"21": "\",",
|
| 38 |
+
"22": "\".",
|
| 39 |
+
"23": "?\"",
|
| 40 |
+
"24": "\"?",
|
| 41 |
+
"25": "\u0964\"",
|
| 42 |
+
"26": "\"\u0964",
|
| 43 |
+
"27": "\u060c",
|
| 44 |
+
"28": "\u1c7e",
|
| 45 |
+
"29": "\u0965",
|
| 46 |
+
"3": "?",
|
| 47 |
+
"30": "\u1c7e\u0964",
|
| 48 |
+
"4": "-",
|
| 49 |
+
"5": ";",
|
| 50 |
+
"6": "_",
|
| 51 |
+
"7": "!",
|
| 52 |
+
"8": "'",
|
| 53 |
+
"9": "..."
|
| 54 |
+
},
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 6912,
|
| 57 |
+
"label2id": {
|
| 58 |
+
"!": 7,
|
| 59 |
+
"\"": 10,
|
| 60 |
+
"\",": 21,
|
| 61 |
+
"\".": 22,
|
| 62 |
+
"\"?": 24,
|
| 63 |
+
"\"\u0964": 26,
|
| 64 |
+
"'": 8,
|
| 65 |
+
"(": 12,
|
| 66 |
+
")": 13,
|
| 67 |
+
"),": 20,
|
| 68 |
+
").": 19,
|
| 69 |
+
",": 2,
|
| 70 |
+
"-": 4,
|
| 71 |
+
".": 1,
|
| 72 |
+
".\"": 18,
|
| 73 |
+
"...": 9,
|
| 74 |
+
":": 14,
|
| 75 |
+
";": 5,
|
| 76 |
+
"?": 3,
|
| 77 |
+
"?\"": 23,
|
| 78 |
+
"O": 0,
|
| 79 |
+
"_": 6,
|
| 80 |
+
"\u060c": 27,
|
| 81 |
+
"\u061f": 17,
|
| 82 |
+
"\u066c": 15,
|
| 83 |
+
"\u06d4": 16,
|
| 84 |
+
"\u0964": 11,
|
| 85 |
+
"\u0964\"": 25,
|
| 86 |
+
"\u0965": 29,
|
| 87 |
+
"\u1c7e": 28,
|
| 88 |
+
"\u1c7e\u0964": 30
|
| 89 |
+
},
|
| 90 |
+
"max_position_embeddings": 32768,
|
| 91 |
+
|
| 92 |
+
"num_attention_heads": 4,
|
| 93 |
+
"num_hidden_layers": 26,
|
| 94 |
+
"num_key_value_heads": 1,
|
| 95 |
+
"pad_token_id": 0,
|
| 96 |
+
"query_pre_attn_scalar": 256,
|
| 97 |
+
"rms_norm_eps": 1e-06,
|
| 98 |
+
"rope_local_base_freq": 10000,
|
| 99 |
+
"rope_scaling": null,
|
| 100 |
+
"rope_theta": 1000000,
|
| 101 |
+
"sliding_window": 512,
|
| 102 |
+
"sliding_window_pattern": 6,
|
| 103 |
+
"torch_dtype": "float32",
|
| 104 |
+
"transformers_version": "4.51.3",
|
| 105 |
+
"use_cache": true,
|
| 106 |
+
"use_non_causal_attention": true,
|
| 107 |
+
"vocab_size": 262146
|
| 108 |
+
}
|