Upload model
Browse files- config.json +40 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "./experiments/raw_wiki/rn/model",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DebertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"base_model": "deberta",
|
| 8 |
+
"bos_token_token_id": 2,
|
| 9 |
+
"cls_token_token_id": 4,
|
| 10 |
+
"eos_token_token_id": 3,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 312,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 1200,
|
| 16 |
+
"layer_norm_eps": 1e-07,
|
| 17 |
+
"mask_token_token_id": 6,
|
| 18 |
+
"max_position_embeddings": 512,
|
| 19 |
+
"max_relative_positions": -1,
|
| 20 |
+
"model_type": "deberta",
|
| 21 |
+
"num_attention_heads": 12,
|
| 22 |
+
"num_hidden_layers": 4,
|
| 23 |
+
"pad_token_id": 0,
|
| 24 |
+
"pad_token_token_id": 0,
|
| 25 |
+
"pooler_dropout": 0,
|
| 26 |
+
"pooler_hidden_act": "gelu",
|
| 27 |
+
"pooler_hidden_size": 312,
|
| 28 |
+
"pos_att_type": [
|
| 29 |
+
"c2p",
|
| 30 |
+
"p2c"
|
| 31 |
+
],
|
| 32 |
+
"position_biased_input": true,
|
| 33 |
+
"relative_attention": true,
|
| 34 |
+
"sep_token_token_id": 5,
|
| 35 |
+
"torch_dtype": "float32",
|
| 36 |
+
"transformers_version": "4.41.2",
|
| 37 |
+
"type_vocab_size": 0,
|
| 38 |
+
"unk_token_token_id": 1,
|
| 39 |
+
"vocab_size": 786
|
| 40 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5b6735d02eed4c3adfbade0a5c3f622c99bb657f7914b67b2bdf04ce83b2886
|
| 3 |
+
size 24297896
|