Initial revision.
Browse files- config.json +24 -0
- eval_results.json +1 -0
- pytorch_model.bin +3 -0
- run.sh +15 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- trainer_state.json +345 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "bert-large-uncased",
|
3 |
+
"architectures": [
|
4 |
+
"BertForQuestionAnswering"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"gradient_checkpointing": false,
|
8 |
+
"hidden_act": "gelu",
|
9 |
+
"hidden_dropout_prob": 0.1,
|
10 |
+
"hidden_size": 1024,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"intermediate_size": 4096,
|
13 |
+
"layer_norm_eps": 1e-12,
|
14 |
+
"max_position_embeddings": 512,
|
15 |
+
"model_type": "bert",
|
16 |
+
"num_attention_heads": 16,
|
17 |
+
"num_hidden_layers": 24,
|
18 |
+
"pad_token_id": 0,
|
19 |
+
"position_embedding_type": "absolute",
|
20 |
+
"transformers_version": "4.5.0.dev0",
|
21 |
+
"type_vocab_size": 2,
|
22 |
+
"use_cache": true,
|
23 |
+
"vocab_size": 30522
|
24 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{'exact': 78.6321906847469, 'f1': 81.5816656803201, 'total': 11873, 'HasAns_exact': 73.73481781376518, 'HasAns_f1': 79.64222615088413, 'HasAns_total': 5928, 'NoAns_exact': 83.51555929352396, 'NoAns_f1': 83.51555929352396, 'NoAns_total': 5945, 'best_exact': 78.6321906847469, 'best_exact_thresh': 0.0, 'best_f1': 81.58166568032006, 'best_f1_thresh': 0.0, 'epoch': 1.59}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2403aa6b558c922393a8a0d0f58920bfdd9036cb631d992e9d3c2191200aa117
|
3 |
+
size 1336547639
|
run.sh
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
python run_qa.py \
|
2 |
+
--model_name_or_path bert-large-uncased \
|
3 |
+
--dataset_name squad_v2 \
|
4 |
+
--do_train \
|
5 |
+
--do_eval \
|
6 |
+
--save_steps 2500 \
|
7 |
+
--eval_steps 2500 \
|
8 |
+
--evaluation_strategy steps \
|
9 |
+
--per_device_train_batch_size 12 \
|
10 |
+
--learning_rate 3e-5 \
|
11 |
+
--num_train_epochs 2 \
|
12 |
+
--max_seq_length 384 \
|
13 |
+
--doc_stride 128 \
|
14 |
+
--output_dir bert-large-uncased-squadv2 \
|
15 |
+
--version_2_with_negative 1
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5848985f52ec5922c5fc3e341345b65668fd7e6496fab9e090deab4a4bc50567
|
3 |
+
size 623
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-large-uncased"}
|
trainer_state.json
ADDED
@@ -0,0 +1,345 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.5938069216757742,
|
5 |
+
"global_step": 17500,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.05,
|
12 |
+
"learning_rate": 2.9316939890710385e-05,
|
13 |
+
"loss": 2.0652,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.09,
|
18 |
+
"learning_rate": 2.8633879781420765e-05,
|
19 |
+
"loss": 1.4682,
|
20 |
+
"step": 1000
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.14,
|
24 |
+
"learning_rate": 2.795081967213115e-05,
|
25 |
+
"loss": 1.2878,
|
26 |
+
"step": 1500
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.18,
|
30 |
+
"learning_rate": 2.7267759562841533e-05,
|
31 |
+
"loss": 1.2088,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.23,
|
36 |
+
"learning_rate": 2.6584699453551913e-05,
|
37 |
+
"loss": 1.1539,
|
38 |
+
"step": 2500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"HasAns_exact": 59.78407557354926,
|
42 |
+
"HasAns_f1": 65.49896652234426,
|
43 |
+
"HasAns_total": 5928,
|
44 |
+
"NoAns_exact": 80.87468460891506,
|
45 |
+
"NoAns_f1": 80.87468460891506,
|
46 |
+
"NoAns_total": 5945,
|
47 |
+
"best_exact": 70.34447907015918,
|
48 |
+
"best_exact_thresh": 0.0,
|
49 |
+
"best_f1": 73.19783319670329,
|
50 |
+
"best_f1_thresh": 0.0,
|
51 |
+
"epoch": 0.23,
|
52 |
+
"exact": 70.34447907015918,
|
53 |
+
"f1": 73.19783319670331,
|
54 |
+
"step": 2500,
|
55 |
+
"total": 11873
|
56 |
+
},
|
57 |
+
{
|
58 |
+
"epoch": 0.27,
|
59 |
+
"learning_rate": 2.5901639344262294e-05,
|
60 |
+
"loss": 1.1802,
|
61 |
+
"step": 3000
|
62 |
+
},
|
63 |
+
{
|
64 |
+
"epoch": 0.32,
|
65 |
+
"learning_rate": 2.5218579234972678e-05,
|
66 |
+
"loss": 1.0921,
|
67 |
+
"step": 3500
|
68 |
+
},
|
69 |
+
{
|
70 |
+
"epoch": 0.36,
|
71 |
+
"learning_rate": 2.453551912568306e-05,
|
72 |
+
"loss": 1.047,
|
73 |
+
"step": 4000
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 0.41,
|
77 |
+
"learning_rate": 2.3852459016393442e-05,
|
78 |
+
"loss": 1.0404,
|
79 |
+
"step": 4500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.46,
|
83 |
+
"learning_rate": 2.3169398907103826e-05,
|
84 |
+
"loss": 1.0074,
|
85 |
+
"step": 5000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"HasAns_exact": 65.9412955465587,
|
89 |
+
"HasAns_f1": 71.69124855898053,
|
90 |
+
"HasAns_total": 5928,
|
91 |
+
"NoAns_exact": 83.11185870479395,
|
92 |
+
"NoAns_f1": 83.11185870479395,
|
93 |
+
"NoAns_total": 5945,
|
94 |
+
"best_exact": 74.53886970437127,
|
95 |
+
"best_exact_thresh": 0.0,
|
96 |
+
"best_f1": 77.40972976144505,
|
97 |
+
"best_f1_thresh": 0.0,
|
98 |
+
"epoch": 0.46,
|
99 |
+
"exact": 74.53886970437127,
|
100 |
+
"f1": 77.40972976144509,
|
101 |
+
"step": 5000,
|
102 |
+
"total": 11873
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 0.5,
|
106 |
+
"learning_rate": 2.248633879781421e-05,
|
107 |
+
"loss": 1.017,
|
108 |
+
"step": 5500
|
109 |
+
},
|
110 |
+
{
|
111 |
+
"epoch": 0.55,
|
112 |
+
"learning_rate": 2.180327868852459e-05,
|
113 |
+
"loss": 0.9725,
|
114 |
+
"step": 6000
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.59,
|
118 |
+
"learning_rate": 2.1120218579234974e-05,
|
119 |
+
"loss": 0.9779,
|
120 |
+
"step": 6500
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"epoch": 0.64,
|
124 |
+
"learning_rate": 2.0437158469945358e-05,
|
125 |
+
"loss": 0.9617,
|
126 |
+
"step": 7000
|
127 |
+
},
|
128 |
+
{
|
129 |
+
"epoch": 0.68,
|
130 |
+
"learning_rate": 1.975409836065574e-05,
|
131 |
+
"loss": 0.929,
|
132 |
+
"step": 7500
|
133 |
+
},
|
134 |
+
{
|
135 |
+
"HasAns_exact": 72.33468286099865,
|
136 |
+
"HasAns_f1": 78.59956513847068,
|
137 |
+
"HasAns_total": 5928,
|
138 |
+
"NoAns_exact": 79.22624053826745,
|
139 |
+
"NoAns_f1": 79.22624053826745,
|
140 |
+
"NoAns_total": 5945,
|
141 |
+
"best_exact": 75.78539543502063,
|
142 |
+
"best_exact_thresh": 0.0,
|
143 |
+
"best_f1": 78.9133514815846,
|
144 |
+
"best_f1_thresh": 0.0,
|
145 |
+
"epoch": 0.68,
|
146 |
+
"exact": 75.78539543502063,
|
147 |
+
"f1": 78.9133514815847,
|
148 |
+
"step": 7500,
|
149 |
+
"total": 11873
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.73,
|
153 |
+
"learning_rate": 1.907103825136612e-05,
|
154 |
+
"loss": 0.9407,
|
155 |
+
"step": 8000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.77,
|
159 |
+
"learning_rate": 1.8387978142076503e-05,
|
160 |
+
"loss": 0.921,
|
161 |
+
"step": 8500
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 0.82,
|
165 |
+
"learning_rate": 1.7704918032786887e-05,
|
166 |
+
"loss": 0.9025,
|
167 |
+
"step": 9000
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 0.87,
|
171 |
+
"learning_rate": 1.7021857923497267e-05,
|
172 |
+
"loss": 0.8724,
|
173 |
+
"step": 9500
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 0.91,
|
177 |
+
"learning_rate": 1.633879781420765e-05,
|
178 |
+
"loss": 0.8794,
|
179 |
+
"step": 10000
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"HasAns_exact": 75.55668016194332,
|
183 |
+
"HasAns_f1": 81.90355530149,
|
184 |
+
"HasAns_total": 5928,
|
185 |
+
"NoAns_exact": 75.02102607232969,
|
186 |
+
"NoAns_f1": 75.02102607232969,
|
187 |
+
"NoAns_total": 5945,
|
188 |
+
"best_exact": 75.2884696369915,
|
189 |
+
"best_exact_thresh": 0.0,
|
190 |
+
"best_f1": 78.45736341507896,
|
191 |
+
"best_f1_thresh": 0.0,
|
192 |
+
"epoch": 0.91,
|
193 |
+
"exact": 75.2884696369915,
|
194 |
+
"f1": 78.45736341507894,
|
195 |
+
"step": 10000,
|
196 |
+
"total": 11873
|
197 |
+
},
|
198 |
+
{
|
199 |
+
"epoch": 0.96,
|
200 |
+
"learning_rate": 1.5655737704918035e-05,
|
201 |
+
"loss": 0.8788,
|
202 |
+
"step": 10500
|
203 |
+
},
|
204 |
+
{
|
205 |
+
"epoch": 1.0,
|
206 |
+
"learning_rate": 1.4972677595628415e-05,
|
207 |
+
"loss": 0.8614,
|
208 |
+
"step": 11000
|
209 |
+
},
|
210 |
+
{
|
211 |
+
"epoch": 1.05,
|
212 |
+
"learning_rate": 1.4289617486338798e-05,
|
213 |
+
"loss": 0.5678,
|
214 |
+
"step": 11500
|
215 |
+
},
|
216 |
+
{
|
217 |
+
"epoch": 1.09,
|
218 |
+
"learning_rate": 1.3606557377049181e-05,
|
219 |
+
"loss": 0.5483,
|
220 |
+
"step": 12000
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 1.14,
|
224 |
+
"learning_rate": 1.2923497267759564e-05,
|
225 |
+
"loss": 0.5724,
|
226 |
+
"step": 12500
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"HasAns_exact": 73.582995951417,
|
230 |
+
"HasAns_f1": 80.21421656448703,
|
231 |
+
"HasAns_total": 5928,
|
232 |
+
"NoAns_exact": 80.31959629941127,
|
233 |
+
"NoAns_f1": 80.31959629941127,
|
234 |
+
"NoAns_total": 5945,
|
235 |
+
"best_exact": 76.95611892529269,
|
236 |
+
"best_exact_thresh": 0.0,
|
237 |
+
"best_f1": 80.26698187436004,
|
238 |
+
"best_f1_thresh": 0.0,
|
239 |
+
"epoch": 1.14,
|
240 |
+
"exact": 76.95611892529269,
|
241 |
+
"f1": 80.26698187436024,
|
242 |
+
"step": 12500,
|
243 |
+
"total": 11873
|
244 |
+
},
|
245 |
+
{
|
246 |
+
"epoch": 1.18,
|
247 |
+
"learning_rate": 1.2240437158469946e-05,
|
248 |
+
"loss": 0.5641,
|
249 |
+
"step": 13000
|
250 |
+
},
|
251 |
+
{
|
252 |
+
"epoch": 1.23,
|
253 |
+
"learning_rate": 1.1557377049180328e-05,
|
254 |
+
"loss": 0.5712,
|
255 |
+
"step": 13500
|
256 |
+
},
|
257 |
+
{
|
258 |
+
"epoch": 1.28,
|
259 |
+
"learning_rate": 1.087431693989071e-05,
|
260 |
+
"loss": 0.5751,
|
261 |
+
"step": 14000
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 1.32,
|
265 |
+
"learning_rate": 1.0191256830601094e-05,
|
266 |
+
"loss": 0.5325,
|
267 |
+
"step": 14500
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 1.37,
|
271 |
+
"learning_rate": 9.508196721311476e-06,
|
272 |
+
"loss": 0.5416,
|
273 |
+
"step": 15000
|
274 |
+
},
|
275 |
+
{
|
276 |
+
"HasAns_exact": 72.94197031039137,
|
277 |
+
"HasAns_f1": 79.45948359952195,
|
278 |
+
"HasAns_total": 5928,
|
279 |
+
"NoAns_exact": 83.06139613120268,
|
280 |
+
"NoAns_f1": 83.06139613120268,
|
281 |
+
"NoAns_total": 5945,
|
282 |
+
"best_exact": 78.00892781942221,
|
283 |
+
"best_exact_thresh": 0.0,
|
284 |
+
"best_f1": 81.26301851073573,
|
285 |
+
"best_f1_thresh": 0.0,
|
286 |
+
"epoch": 1.37,
|
287 |
+
"exact": 78.00892781942221,
|
288 |
+
"f1": 81.2630185107358,
|
289 |
+
"step": 15000,
|
290 |
+
"total": 11873
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 1.41,
|
294 |
+
"learning_rate": 8.825136612021857e-06,
|
295 |
+
"loss": 0.5361,
|
296 |
+
"step": 15500
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 1.46,
|
300 |
+
"learning_rate": 8.14207650273224e-06,
|
301 |
+
"loss": 0.5417,
|
302 |
+
"step": 16000
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 1.5,
|
306 |
+
"learning_rate": 7.459016393442623e-06,
|
307 |
+
"loss": 0.5478,
|
308 |
+
"step": 16500
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 1.55,
|
312 |
+
"learning_rate": 6.775956284153006e-06,
|
313 |
+
"loss": 0.5157,
|
314 |
+
"step": 17000
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 1.59,
|
318 |
+
"learning_rate": 6.092896174863388e-06,
|
319 |
+
"loss": 0.5517,
|
320 |
+
"step": 17500
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"HasAns_exact": 73.73481781376518,
|
324 |
+
"HasAns_f1": 79.64222615088413,
|
325 |
+
"HasAns_total": 5928,
|
326 |
+
"NoAns_exact": 83.51555929352396,
|
327 |
+
"NoAns_f1": 83.51555929352396,
|
328 |
+
"NoAns_total": 5945,
|
329 |
+
"best_exact": 78.6321906847469,
|
330 |
+
"best_exact_thresh": 0.0,
|
331 |
+
"best_f1": 81.58166568032006,
|
332 |
+
"best_f1_thresh": 0.0,
|
333 |
+
"epoch": 1.59,
|
334 |
+
"exact": 78.6321906847469,
|
335 |
+
"f1": 81.5816656803201,
|
336 |
+
"step": 17500,
|
337 |
+
"total": 11873
|
338 |
+
}
|
339 |
+
],
|
340 |
+
"max_steps": 21960,
|
341 |
+
"num_train_epochs": 2,
|
342 |
+
"total_flos": 1.616435859777915e+17,
|
343 |
+
"trial_name": null,
|
344 |
+
"trial_params": null
|
345 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:11c9922bff70152d6882db2c0e1bcf88b1d47aa6153da091a6a0230a18497005
|
3 |
+
size 2351
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|