Upload 4 files
Browse files- config.json +80 -0
- model.safetensors +3 -0
- training_args.bin +3 -0
- training_metadata.json +7 -0
config.json
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "xlm-roberta-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 1024,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LY",
|
| 15 |
+
"1": "SP",
|
| 16 |
+
"2": "ID",
|
| 17 |
+
"3": "NA",
|
| 18 |
+
"4": "HI",
|
| 19 |
+
"5": "IN",
|
| 20 |
+
"6": "OP",
|
| 21 |
+
"7": "IP",
|
| 22 |
+
"8": "it",
|
| 23 |
+
"9": "ne",
|
| 24 |
+
"10": "sr",
|
| 25 |
+
"11": "nb",
|
| 26 |
+
"12": "re",
|
| 27 |
+
"13": "en",
|
| 28 |
+
"14": "ra",
|
| 29 |
+
"15": "dtp",
|
| 30 |
+
"16": "fi",
|
| 31 |
+
"17": "lt",
|
| 32 |
+
"18": "rv",
|
| 33 |
+
"19": "ob",
|
| 34 |
+
"20": "rs",
|
| 35 |
+
"21": "av",
|
| 36 |
+
"22": "ds",
|
| 37 |
+
"23": "ed"
|
| 38 |
+
},
|
| 39 |
+
"initializer_range": 0.02,
|
| 40 |
+
"intermediate_size": 4096,
|
| 41 |
+
"label2id": {
|
| 42 |
+
"HI": 4,
|
| 43 |
+
"ID": 2,
|
| 44 |
+
"IN": 5,
|
| 45 |
+
"IP": 7,
|
| 46 |
+
"LY": 0,
|
| 47 |
+
"NA": 3,
|
| 48 |
+
"OP": 6,
|
| 49 |
+
"SP": 1,
|
| 50 |
+
"av": 21,
|
| 51 |
+
"ds": 22,
|
| 52 |
+
"dtp": 15,
|
| 53 |
+
"ed": 23,
|
| 54 |
+
"en": 13,
|
| 55 |
+
"fi": 16,
|
| 56 |
+
"it": 8,
|
| 57 |
+
"lt": 17,
|
| 58 |
+
"nb": 11,
|
| 59 |
+
"ne": 9,
|
| 60 |
+
"ob": 19,
|
| 61 |
+
"ra": 14,
|
| 62 |
+
"re": 12,
|
| 63 |
+
"rs": 20,
|
| 64 |
+
"rv": 18,
|
| 65 |
+
"sr": 10
|
| 66 |
+
},
|
| 67 |
+
"layer_norm_eps": 1e-05,
|
| 68 |
+
"max_position_embeddings": 514,
|
| 69 |
+
"model_type": "xlm-roberta",
|
| 70 |
+
"num_attention_heads": 16,
|
| 71 |
+
"num_hidden_layers": 24,
|
| 72 |
+
"output_past": true,
|
| 73 |
+
"pad_token_id": 1,
|
| 74 |
+
"position_embedding_type": "absolute",
|
| 75 |
+
"torch_dtype": "bfloat16",
|
| 76 |
+
"transformers_version": "4.39.3",
|
| 77 |
+
"type_vocab_size": 1,
|
| 78 |
+
"use_cache": true,
|
| 79 |
+
"vocab_size": 250002
|
| 80 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0df64c2d30c7ac1313e8705213f39eb50cad2dd975074f4ae299a8ff111396e2
|
| 3 |
+
size 1119878576
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d27f6ad61a07f61d10473c33c79662539547ff90f50593a9dfc4acd64b736176
|
| 3 |
+
size 4984
|
training_metadata.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"batch_size": 8,
|
| 3 |
+
"learning_rate": 3e-05,
|
| 4 |
+
"warmup_ratio": 0.05,
|
| 5 |
+
"stopped_epoch": 9,
|
| 6 |
+
"total_epochs": 30,
|
| 7 |
+
}
|