Hugging Face's logo

zeromodels
/
xlm_roberta_large

Fill-Mask
KerasFormers
Keras
PyTorch
JAX
TensorFlow
xlm-roberta
multilingual
text-encoder
Model card Files
xet
 Community
IMvision12 commited on
Commit
d6ceddd
·
0 Parent(s):

Super-squash branch 'main' using huggingface_hub

Browse files
Files changed (6) hide show
  1. .gitattributes +35 -0
  2. README.md +95 -0
  3. kf_config.json +23 -0
  4. model.weights.json +601 -0
  5. model_00000.weights.h5 +3 -0
  6. tokenizer.json +0 -0
.gitattributes ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,95 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pipeline_tag: fill-mask
3
+ license: mit
4
+ base_model: FacebookAI/xlm-roberta-large
5
+ library_name: kerasformers
6
+ tags:
7
+ - keras
8
+ - kerasformers
9
+ - xlm-roberta
10
+ - fill-mask
11
+ - multilingual
12
+ - text-encoder
13
+ - arxiv:1911.02116
14
+ - pytorch
15
+ - jax
16
+ - tf
17
+ ---
18
+
19
+ ## ***See [our collection](https://huggingface.co/collections/kerasformers/xlm-roberta-6a6e8fd0a258b1a8991cf608) for all versions of XLM-RoBERTa.***
20
+
21
+ # Run XLM-RoBERTa with Keras 3: JAX, PyTorch, or TensorFlow
22
+
23
+ [![GitHub](https://img.shields.io/badge/GitHub-KerasFormers-black?logo=github)](https://github.com/IMvision12/KerasFormers) [![Docs](https://img.shields.io/badge/Docs-XLM--RoBERTa-blue)](https://imvision12.github.io/KerasFormers/xlm_roberta/) [![Collection](https://img.shields.io/badge/HF-XLM--RoBERTa%20collection-yellow)](https://huggingface.co/collections/kerasformers/xlm-roberta-6a6e8fd0a258b1a8991cf608)
24
+
25
+ # kerasformers/xlm_roberta_large
26
+
27
+ Paper: [Unsupervised Cross-lingual Representation Learning at Scale (arXiv:1911.02116)](https://arxiv.org/abs/1911.02116) · [HF Papers](https://huggingface.co/papers/1911.02116)
28
+
29
+ XLM-RoBERTa is the **multilingual** RoBERTa: same encoder architecture, pretrained on 2.5TB CommonCrawl across **100 languages**, with a 250k SentencePiece vocabulary (mask token `<mask>`).
30
+
31
+ For more details on the model, please go to the upstream [model card](https://huggingface.co/FacebookAI/xlm-roberta-large).
32
+
33
+ Pure-**Keras 3** conversion of [`FacebookAI/xlm-roberta-large`](https://huggingface.co/FacebookAI/xlm-roberta-large) for [kerasformers](https://github.com/IMvision12/KerasFormers). One implementation runs unmodified on **TensorFlow / Torch / JAX**.
34
+
35
+ This is a **fill-mask / encoder** checkpoint (`XLMRobertaMaskedLM`, large). Task heads load via `hf:` fine-tunes.
36
+
37
+ ## ✨ Quick start (multilingual fill-mask)
38
+
39
+ ```python
40
+ import os
41
+ os.environ["KERAS_BACKEND"] = "torch" # or "jax" / "tensorflow"
42
+
43
+ from kerasformers.models.xlm_roberta import (
44
+ XLMRobertaMaskedLM,
45
+ XLMRobertaTokenizer,
46
+ )
47
+
48
+ mlm = XLMRobertaMaskedLM.from_weights("kerasformers/xlm_roberta_large")
49
+ tokenizer = XLMRobertaTokenizer.from_weights("kerasformers/xlm_roberta_large")
50
+
51
+ # Multilingual: same <mask> API as RoBERTa, 100-language SentencePiece vocab.
52
+ inputs = tokenizer("La capitale de la France est <mask>.")
53
+ logits = mlm(inputs) # (1, L, vocab_size)
54
+ mask = int((inputs["input_ids"][0] == tokenizer.mask_token_id).argmax())
55
+ print(tokenizer.decode([int(logits[0, mask].argmax())]))
56
+ ```
57
+
58
+ Load any XLM-RoBERTa variant the same way with `from_weights("kerasformers/<variant>")`:
59
+
60
+ | Variant | Hub |
61
+ |---|---|
62
+ | `xlm_roberta_base` | [`kerasformers/xlm_roberta_base`](https://huggingface.co/kerasformers/xlm_roberta_base) |
63
+ | `xlm_roberta_large` | [`kerasformers/xlm_roberta_large`](https://huggingface.co/kerasformers/xlm_roberta_large) |
64
+
65
+ ## Available classes
66
+
67
+ Load any of these from this repo with `from_weights("kerasformers/xlm_roberta_large")` (or on the fly via the `hf:` prefix). The pretrained backbone is shared; task heads not stored in this checkpoint start randomly initialized, ready for fine-tuning (or load a `hf:` fine-tune).
68
+
69
+ | Class | Task |
70
+ |---|---|
71
+ | `XLMRobertaModel` | Encoder backbone |
72
+ | `XLMRobertaMaskedLM` | Masked language modeling (fill-mask) |
73
+ | `XLMRobertaSequenceClassify` | Sequence classification |
74
+ | `XLMRobertaTokenClassify` | Token classification (NER / POS) |
75
+ | `XLMRobertaQnA` | Extractive question answering |
76
+ | `XLMRobertaMultipleChoice` | Multiple choice |
77
+
78
+ ```python
79
+ from kerasformers.models.xlm_roberta import XLMRobertaSequenceClassify
80
+ model = XLMRobertaSequenceClassify.from_weights("kerasformers/xlm_roberta_large")
81
+ ```
82
+
83
+ ## Tips
84
+
85
+ - Set `KERAS_BACKEND` **before** importing Keras / kerasformers.
86
+ - Prefer `XLMRobertaTokenizer.from_weights(...)` so the SentencePiece vocab matches.
87
+ - Use `<mask>` (not `[MASK]`).
88
+ - See [XLM-RoBERTa docs](https://imvision12.github.io/KerasFormers/xlm_roberta/) and [Loading Weights](https://imvision12.github.io/KerasFormers/loading_weights/).
89
+ - Community / upstream safetensors still work via the `hf:` prefix, e.g. `XLMRobertaMaskedLM.from_weights("hf:FacebookAI/xlm-roberta-large")`.
90
+
91
+ ## Special Thanks
92
+
93
+ A huge thank you to the Facebook AI XLM-RoBERTa authors for creating and releasing these models.
94
+
95
+ License: MIT.
kf_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "library_name": "kerasformers",
3
+ "kerasformers_version": "1.2.1",
4
+ "model_module": "kerasformers.models.xlm_roberta",
5
+ "model_class": "XLMRobertaModel",
6
+ "variant": "xlm_roberta_large",
7
+ "weights": "model.weights.json",
8
+ "schema_version": 2,
9
+ "weight_dtype": "float32",
10
+ "model_type": "xlm_roberta",
11
+ "text_config": {
12
+ "vocab_size": 250002,
13
+ "embed_dim": 1024,
14
+ "num_layers": 24,
15
+ "num_heads": 16,
16
+ "mlp_dim": 4096,
17
+ "max_position_embeddings": 514,
18
+ "type_vocab_size": 1,
19
+ "hidden_act": "gelu",
20
+ "layer_norm_eps": 1e-05,
21
+ "pad_token_id": 1
22
+ }
23
+ }
model.weights.json ADDED
@@ -0,0 +1,601 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 3268776520.0
4
+ },
5
+ "weight_map": {
6
+ "/layers/roberta_embeddings/layer_norm/vars": [
7
+ "model_00000.weights.h5"
8
+ ],
9
+ "/layers/roberta_embeddings/position_embeddings/vars": "model_00000.weights.h5",
10
+ "/layers/roberta_embeddings/token_type_embeddings/vars": "model_00000.weights.h5",
11
+ "/layers/roberta_embeddings/word_embeddings/vars": "model_00000.weights.h5",
12
+ "/layers/roberta_self_attention/key/vars": [
13
+ "model_00000.weights.h5"
14
+ ],
15
+ "/layers/roberta_self_attention/query/vars": [
16
+ "model_00000.weights.h5"
17
+ ],
18
+ "/layers/roberta_self_attention/value/vars": [
19
+ "model_00000.weights.h5"
20
+ ],
21
+ "/layers/dense/vars": [
22
+ "model_00000.weights.h5"
23
+ ],
24
+ "/layers/layer_normalization/vars": [
25
+ "model_00000.weights.h5"
26
+ ],
27
+ "/layers/dense_1/vars": [
28
+ "model_00000.weights.h5"
29
+ ],
30
+ "/layers/dense_2/vars": [
31
+ "model_00000.weights.h5"
32
+ ],
33
+ "/layers/layer_normalization_1/vars": [
34
+ "model_00000.weights.h5"
35
+ ],
36
+ "/layers/roberta_self_attention_1/key/vars": [
37
+ "model_00000.weights.h5"
38
+ ],
39
+ "/layers/roberta_self_attention_1/query/vars": [
40
+ "model_00000.weights.h5"
41
+ ],
42
+ "/layers/roberta_self_attention_1/value/vars": [
43
+ "model_00000.weights.h5"
44
+ ],
45
+ "/layers/dense_3/vars": [
46
+ "model_00000.weights.h5"
47
+ ],
48
+ "/layers/layer_normalization_2/vars": [
49
+ "model_00000.weights.h5"
50
+ ],
51
+ "/layers/dense_4/vars": [
52
+ "model_00000.weights.h5"
53
+ ],
54
+ "/layers/dense_5/vars": [
55
+ "model_00000.weights.h5"
56
+ ],
57
+ "/layers/layer_normalization_3/vars": [
58
+ "model_00000.weights.h5"
59
+ ],
60
+ "/layers/roberta_self_attention_2/key/vars": [
61
+ "model_00000.weights.h5"
62
+ ],
63
+ "/layers/roberta_self_attention_2/query/vars": [
64
+ "model_00000.weights.h5"
65
+ ],
66
+ "/layers/roberta_self_attention_2/value/vars": [
67
+ "model_00000.weights.h5"
68
+ ],
69
+ "/layers/dense_6/vars": [
70
+ "model_00000.weights.h5"
71
+ ],
72
+ "/layers/layer_normalization_4/vars": [
73
+ "model_00000.weights.h5"
74
+ ],
75
+ "/layers/dense_7/vars": [
76
+ "model_00000.weights.h5"
77
+ ],
78
+ "/layers/dense_8/vars": [
79
+ "model_00000.weights.h5"
80
+ ],
81
+ "/layers/layer_normalization_5/vars": [
82
+ "model_00000.weights.h5"
83
+ ],
84
+ "/layers/roberta_self_attention_3/key/vars": [
85
+ "model_00000.weights.h5"
86
+ ],
87
+ "/layers/roberta_self_attention_3/query/vars": [
88
+ "model_00000.weights.h5"
89
+ ],
90
+ "/layers/roberta_self_attention_3/value/vars": [
91
+ "model_00000.weights.h5"
92
+ ],
93
+ "/layers/dense_9/vars": [
94
+ "model_00000.weights.h5"
95
+ ],
96
+ "/layers/layer_normalization_6/vars": [
97
+ "model_00000.weights.h5"
98
+ ],
99
+ "/layers/dense_10/vars": [
100
+ "model_00000.weights.h5"
101
+ ],
102
+ "/layers/dense_11/vars": [
103
+ "model_00000.weights.h5"
104
+ ],
105
+ "/layers/layer_normalization_7/vars": [
106
+ "model_00000.weights.h5"
107
+ ],
108
+ "/layers/roberta_self_attention_4/key/vars": [
109
+ "model_00000.weights.h5"
110
+ ],
111
+ "/layers/roberta_self_attention_4/query/vars": [
112
+ "model_00000.weights.h5"
113
+ ],
114
+ "/layers/roberta_self_attention_4/value/vars": [
115
+ "model_00000.weights.h5"
116
+ ],
117
+ "/layers/dense_12/vars": [
118
+ "model_00000.weights.h5"
119
+ ],
120
+ "/layers/layer_normalization_8/vars": [
121
+ "model_00000.weights.h5"
122
+ ],
123
+ "/layers/dense_13/vars": [
124
+ "model_00000.weights.h5"
125
+ ],
126
+ "/layers/dense_14/vars": [
127
+ "model_00000.weights.h5"
128
+ ],
129
+ "/layers/layer_normalization_9/vars": [
130
+ "model_00000.weights.h5"
131
+ ],
132
+ "/layers/roberta_self_attention_5/key/vars": [
133
+ "model_00000.weights.h5"
134
+ ],
135
+ "/layers/roberta_self_attention_5/query/vars": [
136
+ "model_00000.weights.h5"
137
+ ],
138
+ "/layers/roberta_self_attention_5/value/vars": [
139
+ "model_00000.weights.h5"
140
+ ],
141
+ "/layers/dense_15/vars": [
142
+ "model_00000.weights.h5"
143
+ ],
144
+ "/layers/layer_normalization_10/vars": [
145
+ "model_00000.weights.h5"
146
+ ],
147
+ "/layers/dense_16/vars": [
148
+ "model_00000.weights.h5"
149
+ ],
150
+ "/layers/dense_17/vars": [
151
+ "model_00000.weights.h5"
152
+ ],
153
+ "/layers/layer_normalization_11/vars": [
154
+ "model_00000.weights.h5"
155
+ ],
156
+ "/layers/roberta_self_attention_6/key/vars": [
157
+ "model_00000.weights.h5"
158
+ ],
159
+ "/layers/roberta_self_attention_6/query/vars": [
160
+ "model_00000.weights.h5"
161
+ ],
162
+ "/layers/roberta_self_attention_6/value/vars": [
163
+ "model_00000.weights.h5"
164
+ ],
165
+ "/layers/dense_18/vars": [
166
+ "model_00000.weights.h5"
167
+ ],
168
+ "/layers/layer_normalization_12/vars": [
169
+ "model_00000.weights.h5"
170
+ ],
171
+ "/layers/dense_19/vars": [
172
+ "model_00000.weights.h5"
173
+ ],
174
+ "/layers/dense_20/vars": [
175
+ "model_00000.weights.h5"
176
+ ],
177
+ "/layers/layer_normalization_13/vars": [
178
+ "model_00000.weights.h5"
179
+ ],
180
+ "/layers/roberta_self_attention_7/key/vars": [
181
+ "model_00000.weights.h5"
182
+ ],
183
+ "/layers/roberta_self_attention_7/query/vars": [
184
+ "model_00000.weights.h5"
185
+ ],
186
+ "/layers/roberta_self_attention_7/value/vars": [
187
+ "model_00000.weights.h5"
188
+ ],
189
+ "/layers/dense_21/vars": [
190
+ "model_00000.weights.h5"
191
+ ],
192
+ "/layers/layer_normalization_14/vars": [
193
+ "model_00000.weights.h5"
194
+ ],
195
+ "/layers/dense_22/vars": [
196
+ "model_00000.weights.h5"
197
+ ],
198
+ "/layers/dense_23/vars": [
199
+ "model_00000.weights.h5"
200
+ ],
201
+ "/layers/layer_normalization_15/vars": [
202
+ "model_00000.weights.h5"
203
+ ],
204
+ "/layers/roberta_self_attention_8/key/vars": [
205
+ "model_00000.weights.h5"
206
+ ],
207
+ "/layers/roberta_self_attention_8/query/vars": [
208
+ "model_00000.weights.h5"
209
+ ],
210
+ "/layers/roberta_self_attention_8/value/vars": [
211
+ "model_00000.weights.h5"
212
+ ],
213
+ "/layers/dense_24/vars": [
214
+ "model_00000.weights.h5"
215
+ ],
216
+ "/layers/layer_normalization_16/vars": [
217
+ "model_00000.weights.h5"
218
+ ],
219
+ "/layers/dense_25/vars": [
220
+ "model_00000.weights.h5"
221
+ ],
222
+ "/layers/dense_26/vars": [
223
+ "model_00000.weights.h5"
224
+ ],
225
+ "/layers/layer_normalization_17/vars": [
226
+ "model_00000.weights.h5"
227
+ ],
228
+ "/layers/roberta_self_attention_9/key/vars": [
229
+ "model_00000.weights.h5"
230
+ ],
231
+ "/layers/roberta_self_attention_9/query/vars": [
232
+ "model_00000.weights.h5"
233
+ ],
234
+ "/layers/roberta_self_attention_9/value/vars": [
235
+ "model_00000.weights.h5"
236
+ ],
237
+ "/layers/dense_27/vars": [
238
+ "model_00000.weights.h5"
239
+ ],
240
+ "/layers/layer_normalization_18/vars": [
241
+ "model_00000.weights.h5"
242
+ ],
243
+ "/layers/dense_28/vars": [
244
+ "model_00000.weights.h5"
245
+ ],
246
+ "/layers/dense_29/vars": [
247
+ "model_00000.weights.h5"
248
+ ],
249
+ "/layers/layer_normalization_19/vars": [
250
+ "model_00000.weights.h5"
251
+ ],
252
+ "/layers/roberta_self_attention_10/key/vars": [
253
+ "model_00000.weights.h5"
254
+ ],
255
+ "/layers/roberta_self_attention_10/query/vars": [
256
+ "model_00000.weights.h5"
257
+ ],
258
+ "/layers/roberta_self_attention_10/value/vars": [
259
+ "model_00000.weights.h5"
260
+ ],
261
+ "/layers/dense_30/vars": [
262
+ "model_00000.weights.h5"
263
+ ],
264
+ "/layers/layer_normalization_20/vars": [
265
+ "model_00000.weights.h5"
266
+ ],
267
+ "/layers/dense_31/vars": [
268
+ "model_00000.weights.h5"
269
+ ],
270
+ "/layers/dense_32/vars": [
271
+ "model_00000.weights.h5"
272
+ ],
273
+ "/layers/layer_normalization_21/vars": [
274
+ "model_00000.weights.h5"
275
+ ],
276
+ "/layers/roberta_self_attention_11/key/vars": [
277
+ "model_00000.weights.h5"
278
+ ],
279
+ "/layers/roberta_self_attention_11/query/vars": [
280
+ "model_00000.weights.h5"
281
+ ],
282
+ "/layers/roberta_self_attention_11/value/vars": [
283
+ "model_00000.weights.h5"
284
+ ],
285
+ "/layers/dense_33/vars": [
286
+ "model_00000.weights.h5"
287
+ ],
288
+ "/layers/layer_normalization_22/vars": [
289
+ "model_00000.weights.h5"
290
+ ],
291
+ "/layers/dense_34/vars": [
292
+ "model_00000.weights.h5"
293
+ ],
294
+ "/layers/dense_35/vars": [
295
+ "model_00000.weights.h5"
296
+ ],
297
+ "/layers/layer_normalization_23/vars": [
298
+ "model_00000.weights.h5"
299
+ ],
300
+ "/layers/roberta_self_attention_12/key/vars": [
301
+ "model_00000.weights.h5"
302
+ ],
303
+ "/layers/roberta_self_attention_12/query/vars": [
304
+ "model_00000.weights.h5"
305
+ ],
306
+ "/layers/roberta_self_attention_12/value/vars": [
307
+ "model_00000.weights.h5"
308
+ ],
309
+ "/layers/dense_36/vars": [
310
+ "model_00000.weights.h5"
311
+ ],
312
+ "/layers/layer_normalization_24/vars": [
313
+ "model_00000.weights.h5"
314
+ ],
315
+ "/layers/dense_37/vars": [
316
+ "model_00000.weights.h5"
317
+ ],
318
+ "/layers/dense_38/vars": [
319
+ "model_00000.weights.h5"
320
+ ],
321
+ "/layers/layer_normalization_25/vars": [
322
+ "model_00000.weights.h5"
323
+ ],
324
+ "/layers/roberta_self_attention_13/key/vars": [
325
+ "model_00000.weights.h5"
326
+ ],
327
+ "/layers/roberta_self_attention_13/query/vars": [
328
+ "model_00000.weights.h5"
329
+ ],
330
+ "/layers/roberta_self_attention_13/value/vars": [
331
+ "model_00000.weights.h5"
332
+ ],
333
+ "/layers/dense_39/vars": [
334
+ "model_00000.weights.h5"
335
+ ],
336
+ "/layers/layer_normalization_26/vars": [
337
+ "model_00000.weights.h5"
338
+ ],
339
+ "/layers/dense_40/vars": [
340
+ "model_00000.weights.h5"
341
+ ],
342
+ "/layers/dense_41/vars": [
343
+ "model_00000.weights.h5"
344
+ ],
345
+ "/layers/layer_normalization_27/vars": [
346
+ "model_00000.weights.h5"
347
+ ],
348
+ "/layers/roberta_self_attention_14/key/vars": [
349
+ "model_00000.weights.h5"
350
+ ],
351
+ "/layers/roberta_self_attention_14/query/vars": [
352
+ "model_00000.weights.h5"
353
+ ],
354
+ "/layers/roberta_self_attention_14/value/vars": [
355
+ "model_00000.weights.h5"
356
+ ],
357
+ "/layers/dense_42/vars": [
358
+ "model_00000.weights.h5"
359
+ ],
360
+ "/layers/layer_normalization_28/vars": [
361
+ "model_00000.weights.h5"
362
+ ],
363
+ "/layers/dense_43/vars": [
364
+ "model_00000.weights.h5"
365
+ ],
366
+ "/layers/dense_44/vars": [
367
+ "model_00000.weights.h5"
368
+ ],
369
+ "/layers/layer_normalization_29/vars": [
370
+ "model_00000.weights.h5"
371
+ ],
372
+ "/layers/roberta_self_attention_15/key/vars": [
373
+ "model_00000.weights.h5"
374
+ ],
375
+ "/layers/roberta_self_attention_15/query/vars": [
376
+ "model_00000.weights.h5"
377
+ ],
378
+ "/layers/roberta_self_attention_15/value/vars": [
379
+ "model_00000.weights.h5"
380
+ ],
381
+ "/layers/dense_45/vars": [
382
+ "model_00000.weights.h5"
383
+ ],
384
+ "/layers/layer_normalization_30/vars": [
385
+ "model_00000.weights.h5"
386
+ ],
387
+ "/layers/dense_46/vars": [
388
+ "model_00000.weights.h5"
389
+ ],
390
+ "/layers/dense_47/vars": [
391
+ "model_00000.weights.h5"
392
+ ],
393
+ "/layers/layer_normalization_31/vars": [
394
+ "model_00000.weights.h5"
395
+ ],
396
+ "/layers/roberta_self_attention_16/key/vars": [
397
+ "model_00000.weights.h5"
398
+ ],
399
+ "/layers/roberta_self_attention_16/query/vars": [
400
+ "model_00000.weights.h5"
401
+ ],
402
+ "/layers/roberta_self_attention_16/value/vars": [
403
+ "model_00000.weights.h5"
404
+ ],
405
+ "/layers/dense_48/vars": [
406
+ "model_00000.weights.h5"
407
+ ],
408
+ "/layers/layer_normalization_32/vars": [
409
+ "model_00000.weights.h5"
410
+ ],
411
+ "/layers/dense_49/vars": [
412
+ "model_00000.weights.h5"
413
+ ],
414
+ "/layers/dense_50/vars": [
415
+ "model_00000.weights.h5"
416
+ ],
417
+ "/layers/layer_normalization_33/vars": [
418
+ "model_00000.weights.h5"
419
+ ],
420
+ "/layers/roberta_self_attention_17/key/vars": [
421
+ "model_00000.weights.h5"
422
+ ],
423
+ "/layers/roberta_self_attention_17/query/vars": [
424
+ "model_00000.weights.h5"
425
+ ],
426
+ "/layers/roberta_self_attention_17/value/vars": [
427
+ "model_00000.weights.h5"
428
+ ],
429
+ "/layers/dense_51/vars": [
430
+ "model_00000.weights.h5"
431
+ ],
432
+ "/layers/layer_normalization_34/vars": [
433
+ "model_00000.weights.h5"
434
+ ],
435
+ "/layers/dense_52/vars": [
436
+ "model_00000.weights.h5"
437
+ ],
438
+ "/layers/dense_53/vars": [
439
+ "model_00000.weights.h5"
440
+ ],
441
+ "/layers/layer_normalization_35/vars": [
442
+ "model_00000.weights.h5"
443
+ ],
444
+ "/layers/roberta_self_attention_18/key/vars": [
445
+ "model_00000.weights.h5"
446
+ ],
447
+ "/layers/roberta_self_attention_18/query/vars": [
448
+ "model_00000.weights.h5"
449
+ ],
450
+ "/layers/roberta_self_attention_18/value/vars": [
451
+ "model_00000.weights.h5"
452
+ ],
453
+ "/layers/dense_54/vars": [
454
+ "model_00000.weights.h5"
455
+ ],
456
+ "/layers/layer_normalization_36/vars": [
457
+ "model_00000.weights.h5"
458
+ ],
459
+ "/layers/dense_55/vars": [
460
+ "model_00000.weights.h5"
461
+ ],
462
+ "/layers/dense_56/vars": [
463
+ "model_00000.weights.h5"
464
+ ],
465
+ "/layers/layer_normalization_37/vars": [
466
+ "model_00000.weights.h5"
467
+ ],
468
+ "/layers/roberta_self_attention_19/key/vars": [
469
+ "model_00000.weights.h5"
470
+ ],
471
+ "/layers/roberta_self_attention_19/query/vars": [
472
+ "model_00000.weights.h5"
473
+ ],
474
+ "/layers/roberta_self_attention_19/value/vars": [
475
+ "model_00000.weights.h5"
476
+ ],
477
+ "/layers/dense_57/vars": [
478
+ "model_00000.weights.h5"
479
+ ],
480
+ "/layers/layer_normalization_38/vars": [
481
+ "model_00000.weights.h5"
482
+ ],
483
+ "/layers/dense_58/vars": [
484
+ "model_00000.weights.h5"
485
+ ],
486
+ "/layers/dense_59/vars": [
487
+ "model_00000.weights.h5"
488
+ ],
489
+ "/layers/layer_normalization_39/vars": [
490
+ "model_00000.weights.h5"
491
+ ],
492
+ "/layers/roberta_self_attention_20/key/vars": [
493
+ "model_00000.weights.h5"
494
+ ],
495
+ "/layers/roberta_self_attention_20/query/vars": [
496
+ "model_00000.weights.h5"
497
+ ],
498
+ "/layers/roberta_self_attention_20/value/vars": [
499
+ "model_00000.weights.h5"
500
+ ],
501
+ "/layers/dense_60/vars": [
502
+ "model_00000.weights.h5"
503
+ ],
504
+ "/layers/layer_normalization_40/vars": [
505
+ "model_00000.weights.h5"
506
+ ],
507
+ "/layers/dense_61/vars": [
508
+ "model_00000.weights.h5"
509
+ ],
510
+ "/layers/dense_62/vars": [
511
+ "model_00000.weights.h5"
512
+ ],
513
+ "/layers/layer_normalization_41/vars": [
514
+ "model_00000.weights.h5"
515
+ ],
516
+ "/layers/roberta_self_attention_21/key/vars": [
517
+ "model_00000.weights.h5"
518
+ ],
519
+ "/layers/roberta_self_attention_21/query/vars": [
520
+ "model_00000.weights.h5"
521
+ ],
522
+ "/layers/roberta_self_attention_21/value/vars": [
523
+ "model_00000.weights.h5"
524
+ ],
525
+ "/layers/dense_63/vars": [
526
+ "model_00000.weights.h5"
527
+ ],
528
+ "/layers/layer_normalization_42/vars": [
529
+ "model_00000.weights.h5"
530
+ ],
531
+ "/layers/dense_64/vars": [
532
+ "model_00000.weights.h5"
533
+ ],
534
+ "/layers/dense_65/vars": [
535
+ "model_00000.weights.h5"
536
+ ],
537
+ "/layers/layer_normalization_43/vars": [
538
+ "model_00000.weights.h5"
539
+ ],
540
+ "/layers/roberta_self_attention_22/key/vars": [
541
+ "model_00000.weights.h5"
542
+ ],
543
+ "/layers/roberta_self_attention_22/query/vars": [
544
+ "model_00000.weights.h5"
545
+ ],
546
+ "/layers/roberta_self_attention_22/value/vars": [
547
+ "model_00000.weights.h5"
548
+ ],
549
+ "/layers/dense_66/vars": [
550
+ "model_00000.weights.h5"
551
+ ],
552
+ "/layers/layer_normalization_44/vars": [
553
+ "model_00000.weights.h5"
554
+ ],
555
+ "/layers/dense_67/vars": [
556
+ "model_00000.weights.h5"
557
+ ],
558
+ "/layers/dense_68/vars": [
559
+ "model_00000.weights.h5"
560
+ ],
561
+ "/layers/layer_normalization_45/vars": [
562
+ "model_00000.weights.h5"
563
+ ],
564
+ "/layers/roberta_self_attention_23/key/vars": [
565
+ "model_00000.weights.h5"
566
+ ],
567
+ "/layers/roberta_self_attention_23/query/vars": [
568
+ "model_00000.weights.h5"
569
+ ],
570
+ "/layers/roberta_self_attention_23/value/vars": [
571
+ "model_00000.weights.h5"
572
+ ],
573
+ "/layers/dense_69/vars": [
574
+ "model_00000.weights.h5"
575
+ ],
576
+ "/layers/layer_normalization_46/vars": [
577
+ "model_00000.weights.h5"
578
+ ],
579
+ "/layers/dense_70/vars": [
580
+ "model_00000.weights.h5"
581
+ ],
582
+ "/layers/dense_71/vars": [
583
+ "model_00000.weights.h5"
584
+ ],
585
+ "/layers/layer_normalization_47/vars": [
586
+ "model_00000.weights.h5"
587
+ ],
588
+ "/layers/dense_72/vars": [
589
+ "model_00000.weights.h5"
590
+ ],
591
+ "/layers/layer_normalization_48/vars": [
592
+ "model_00000.weights.h5"
593
+ ],
594
+ "/layers/dense_73/vars": [
595
+ "model_00000.weights.h5"
596
+ ],
597
+ "/layers/dense_74/vars": [
598
+ "model_00000.weights.h5"
599
+ ]
600
+ }
601
+ }
model_00000.weights.h5 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc557161ee4ed489de3a2dfa9e2a9d9a92dc768bf793021149689034ace64be2
3
+ size 3269684728
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff