Commit
•
14a6760
1
Parent(s):
7fe003a
Training in progress, epoch 1
Browse files
config.json
ADDED
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "answerdotai/ModernBERT-base",
|
3 |
+
"architectures": [
|
4 |
+
"ModernBertForSequenceClassification"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 50281,
|
9 |
+
"classifier_activation": "gelu",
|
10 |
+
"classifier_bias": false,
|
11 |
+
"classifier_dropout": 0.0,
|
12 |
+
"classifier_pooling": "mean",
|
13 |
+
"cls_token_id": 50281,
|
14 |
+
"decoder_bias": true,
|
15 |
+
"deterministic_flash_attn": false,
|
16 |
+
"embedding_dropout": 0.0,
|
17 |
+
"eos_token_id": 50282,
|
18 |
+
"global_attn_every_n_layers": 3,
|
19 |
+
"global_rope_theta": 160000.0,
|
20 |
+
"gradient_checkpointing": false,
|
21 |
+
"hidden_activation": "gelu",
|
22 |
+
"hidden_size": 768,
|
23 |
+
"id2label": {
|
24 |
+
"0": "business-and-industrial",
|
25 |
+
"1": "books-and-literature",
|
26 |
+
"10": "home-and-garden",
|
27 |
+
"11": "adult",
|
28 |
+
"12": "science",
|
29 |
+
"13": "food-and-drink",
|
30 |
+
"14": "real-estate",
|
31 |
+
"15": "news",
|
32 |
+
"16": "jobs-and-education",
|
33 |
+
"17": "health",
|
34 |
+
"18": "hobbies-and-leisure",
|
35 |
+
"19": "games",
|
36 |
+
"2": "beauty-and-fitness",
|
37 |
+
"20": "computers-and-electronics",
|
38 |
+
"21": "arts-and-entertainment",
|
39 |
+
"22": "travel-and-transportation",
|
40 |
+
"23": "finance",
|
41 |
+
"24": "law-and-government",
|
42 |
+
"25": "sensitive-subjects",
|
43 |
+
"3": "autos-and-vehicles",
|
44 |
+
"4": "people-and-society",
|
45 |
+
"5": "sports",
|
46 |
+
"6": "shopping",
|
47 |
+
"7": "online-communities",
|
48 |
+
"8": "pets-and-animals",
|
49 |
+
"9": "internet-and-telecom"
|
50 |
+
},
|
51 |
+
"initializer_cutoff_factor": 2.0,
|
52 |
+
"initializer_range": 0.02,
|
53 |
+
"intermediate_size": 1152,
|
54 |
+
"label2id": {
|
55 |
+
"adult": "11",
|
56 |
+
"arts-and-entertainment": "21",
|
57 |
+
"autos-and-vehicles": "3",
|
58 |
+
"beauty-and-fitness": "2",
|
59 |
+
"books-and-literature": "1",
|
60 |
+
"business-and-industrial": "0",
|
61 |
+
"computers-and-electronics": "20",
|
62 |
+
"finance": "23",
|
63 |
+
"food-and-drink": "13",
|
64 |
+
"games": "19",
|
65 |
+
"health": "17",
|
66 |
+
"hobbies-and-leisure": "18",
|
67 |
+
"home-and-garden": "10",
|
68 |
+
"internet-and-telecom": "9",
|
69 |
+
"jobs-and-education": "16",
|
70 |
+
"law-and-government": "24",
|
71 |
+
"news": "15",
|
72 |
+
"online-communities": "7",
|
73 |
+
"people-and-society": "4",
|
74 |
+
"pets-and-animals": "8",
|
75 |
+
"real-estate": "14",
|
76 |
+
"science": "12",
|
77 |
+
"sensitive-subjects": "25",
|
78 |
+
"shopping": "6",
|
79 |
+
"sports": "5",
|
80 |
+
"travel-and-transportation": "22"
|
81 |
+
},
|
82 |
+
"layer_norm_eps": 1e-05,
|
83 |
+
"local_attention": 128,
|
84 |
+
"local_rope_theta": 10000.0,
|
85 |
+
"max_position_embeddings": 8192,
|
86 |
+
"mlp_bias": false,
|
87 |
+
"mlp_dropout": 0.0,
|
88 |
+
"model_type": "modernbert",
|
89 |
+
"norm_bias": false,
|
90 |
+
"norm_eps": 1e-05,
|
91 |
+
"num_attention_heads": 12,
|
92 |
+
"num_hidden_layers": 22,
|
93 |
+
"pad_token_id": 50283,
|
94 |
+
"position_embedding_type": "absolute",
|
95 |
+
"problem_type": "single_label_classification",
|
96 |
+
"reference_compile": false,
|
97 |
+
"sep_token_id": 50282,
|
98 |
+
"sparse_pred_ignore_index": -100,
|
99 |
+
"sparse_prediction": false,
|
100 |
+
"torch_dtype": "float32",
|
101 |
+
"transformers_version": "4.48.0.dev0",
|
102 |
+
"vocab_size": 50368
|
103 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9b783fe937f587916cf9da959c946a336557a78ca7b3cf17e3b14d1d4d22a95f
|
3 |
+
size 598513616
|
runs/Dec30_11-50-32_Davids-MacBook-Pro.local/events.out.tfevents.1735555878.Davids-MacBook-Pro.local.23438.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d263a00eeddf756efcd326aba635f7c291d0b4a4e35c322911b18654119d040a
|
3 |
+
size 7507
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:593b4c0e73ab5ec7b6b2049e8532fe4e79c644312bce0b9f4534f5e257a5174c
|
3 |
+
size 5368
|