Training in progress, step 5000
Browse files
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"DebertaForMaskedLM"
|
4 |
+
],
|
5 |
+
"attention_probs_dropout_prob": 0.1,
|
6 |
+
"hidden_act": "gelu",
|
7 |
+
"hidden_dropout_prob": 0.1,
|
8 |
+
"hidden_size": 768,
|
9 |
+
"initializer_range": 0.02,
|
10 |
+
"intermediate_size": 3072,
|
11 |
+
"layer_norm_eps": 1e-07,
|
12 |
+
"max_position_embeddings": 512,
|
13 |
+
"max_relative_positions": -1,
|
14 |
+
"model_type": "deberta",
|
15 |
+
"num_attention_heads": 12,
|
16 |
+
"num_hidden_layers": 6,
|
17 |
+
"pad_token_id": 0,
|
18 |
+
"pooler_dropout": 0,
|
19 |
+
"pooler_hidden_act": "gelu",
|
20 |
+
"pooler_hidden_size": 768,
|
21 |
+
"pos_att_type": null,
|
22 |
+
"position_biased_input": true,
|
23 |
+
"relative_attention": false,
|
24 |
+
"torch_dtype": "float32",
|
25 |
+
"transformers_version": "4.20.1",
|
26 |
+
"type_vocab_size": 0,
|
27 |
+
"vocab_size": 10000
|
28 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:abf89cb0f8214723b198a1e1cad1e899cefa2b0b5487127cfa2572da677e850b
|
3 |
+
size 204831069
|
runs/Jul14_08-45-47_903978113fc8/1657788364.9445837/events.out.tfevents.1657788364.903978113fc8.188.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:67c860792a04b3d6cba229da9f8e9e36a0b334f21d4fa789c4c13442c00215a8
|
3 |
+
size 5327
|
runs/Jul14_08-45-47_903978113fc8/events.out.tfevents.1657788364.903978113fc8.188.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:eb3a12fe94c7fd9854f12f6ed19522f2d96881419324020edf1c67660abad6de
|
3 |
+
size 4020
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:43d6070e12ca8f682255f534010aef717d5cde93a5cee5b39a22fbbaf4d2b736
|
3 |
+
size 3311
|