alvanli
commited on
Commit
•
97aa366
1
Parent(s):
7e8eb91
Added rest of the files for 10.1 CER
Browse files- README.md +13 -19
- pytorch_model.bin +1 -1
- rng_state.pth +2 -2
- runs/{Dec12_00-47-25_b23159305be2/1670806047.5609205/events.out.tfevents.1670806047.b23159305be2.526.1 → Jan23_04-20-07_441e7d2b355d/1674447614.1245308/events.out.tfevents.1674447614.441e7d2b355d.71996.1} +2 -2
- runs/{Dec12_00-47-25_b23159305be2/events.out.tfevents.1670806047.b23159305be2.526.0 → Jan23_04-20-07_441e7d2b355d/events.out.tfevents.1674447614.441e7d2b355d.71996.0} +2 -2
- scaler.pt +1 -1
- scheduler.pt +1 -1
- tokenizer_config.json +1 -1
- trainer_state.json +2746 -637
- training_args.bin +2 -2
README.md
CHANGED
@@ -22,48 +22,42 @@ model-index:
|
|
22 |
metrics:
|
23 |
- name: Cer
|
24 |
type: cer
|
25 |
-
value: 11
|
26 |
---
|
27 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
28 |
should probably proofread and complete it, then remove this comment. -->
|
29 |
|
30 |
# Whisper Small zh-HK - Alvin
|
31 |
|
32 |
-
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
|
33 |
|
34 |
-
## Model description
|
35 |
-
More information needed
|
36 |
-
## Intended uses & limitations
|
37 |
-
More information needed
|
38 |
## Training and evaluation data
|
39 |
For training, three datasets were used:
|
40 |
- Common Voice 11 Canto Train Set
|
41 |
- CantoMap: Winterstein, Grégoire, Tang, Carmen and Lai, Regine (2020) "CantoMap: a Hong Kong Cantonese MapTask Corpus", in Proceedings of The 12th Language Resources and Evaluation Conference, Marseille: European Language Resources Association, p. 2899-2906.
|
42 |
- Cantonse-ASR: Yu, Tiezheng, Frieske, Rita, Xu, Peng, Cahyawijaya, Samuel, Yiu, Cheuk Tung, Lovenia, Holy, Dai, Wenliang, Barezi, Elham, Chen, Qifeng, Ma, Xiaojuan, Shi, Bertram, Fung, Pascale (2022) "Automatic Speech Recognition Datasets in Cantonese: A Survey and New Dataset", 2022. Link: https://arxiv.org/pdf/2201.02419.pdf
|
43 |
|
44 |
-
## Training procedure
|
45 |
-
|
46 |
## Training Hyperparameters
|
47 |
-
- learning_rate:
|
48 |
-
- train_batch_size:
|
49 |
- eval_batch_size: 8
|
50 |
- gradient_accumulation_steps: 2
|
51 |
- total_train_batch_size: 16x2x2=64
|
52 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
53 |
- lr_scheduler_type: linear
|
54 |
- lr_scheduler_warmup_steps: 500
|
55 |
-
- training_steps:
|
56 |
- mixed_precision_training: Native AMP
|
|
|
57 |
|
58 |
## Training Results
|
59 |
|
60 |
| Training Loss | Epoch | Step | Validation Loss | Cer |
|
61 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
62 |
-
| 0.
|
63 |
-
| 0.
|
64 |
-
| 0.
|
65 |
-
| 0.
|
66 |
-
| 0.
|
67 |
-
|
68 |
-
|
69 |
-
### Framework versions
|
|
|
22 |
metrics:
|
23 |
- name: Cer
|
24 |
type: cer
|
25 |
+
value: 10.11
|
26 |
---
|
27 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
28 |
should probably proofread and complete it, then remove this comment. -->
|
29 |
|
30 |
# Whisper Small zh-HK - Alvin
|
31 |
|
32 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset. This version has a lower CER (by 1%) compared to the previous one.
|
33 |
|
|
|
|
|
|
|
|
|
34 |
## Training and evaluation data
|
35 |
For training, three datasets were used:
|
36 |
- Common Voice 11 Canto Train Set
|
37 |
- CantoMap: Winterstein, Grégoire, Tang, Carmen and Lai, Regine (2020) "CantoMap: a Hong Kong Cantonese MapTask Corpus", in Proceedings of The 12th Language Resources and Evaluation Conference, Marseille: European Language Resources Association, p. 2899-2906.
|
38 |
- Cantonse-ASR: Yu, Tiezheng, Frieske, Rita, Xu, Peng, Cahyawijaya, Samuel, Yiu, Cheuk Tung, Lovenia, Holy, Dai, Wenliang, Barezi, Elham, Chen, Qifeng, Ma, Xiaojuan, Shi, Bertram, Fung, Pascale (2022) "Automatic Speech Recognition Datasets in Cantonese: A Survey and New Dataset", 2022. Link: https://arxiv.org/pdf/2201.02419.pdf
|
39 |
|
|
|
|
|
40 |
## Training Hyperparameters
|
41 |
+
- learning_rate: 5e-5
|
42 |
+
- train_batch_size: 25 (on 2 GPUs)
|
43 |
- eval_batch_size: 8
|
44 |
- gradient_accumulation_steps: 2
|
45 |
- total_train_batch_size: 16x2x2=64
|
46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
47 |
- lr_scheduler_type: linear
|
48 |
- lr_scheduler_warmup_steps: 500
|
49 |
+
- training_steps: 14000
|
50 |
- mixed_precision_training: Native AMP
|
51 |
+
- augmentation: SpecAugment
|
52 |
|
53 |
## Training Results
|
54 |
|
55 |
| Training Loss | Epoch | Step | Validation Loss | Cer |
|
56 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
57 |
+
| 0.4610 | 0.55 | 2000 | 0.3106 | 13.08 |
|
58 |
+
| 0.3441 | 1.11 | 4000 | 0.2875 | 11.79 |
|
59 |
+
| 0.3466 | 1.66 | 6000 | 0.2820 | 11.44 |
|
60 |
+
| 0.2539 | 2.22 | 8000 | 0.2777 | 10.59 |
|
61 |
+
| 0.2312 | 2.77 | 10000 | 0.2822 | 10.60 |
|
62 |
+
| 0.1639 | 3.32 | 12000 | 0.2859 | 10.17 |
|
63 |
+
| 0.1569 | 3.88 | 14000 | 0.2866 | 10.11 |
|
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 967102601
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9563aaf2c8c73c9368146c47fb8767f7b2fc4ec19fa3d2f43de391939240e744
|
3 |
size 967102601
|
rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f84ff25589d8196892ac54f0030211cdc8c9f81de14e6da66a3c7ef0031b4dc0
|
3 |
+
size 15597
|
runs/{Dec12_00-47-25_b23159305be2/1670806047.5609205/events.out.tfevents.1670806047.b23159305be2.526.1 → Jan23_04-20-07_441e7d2b355d/1674447614.1245308/events.out.tfevents.1674447614.441e7d2b355d.71996.1}
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1429d2ad0089c4a83f7b1b1c7a802c58bd80ca29c7972990dc86b0daf336801c
|
3 |
+
size 5941
|
runs/{Dec12_00-47-25_b23159305be2/events.out.tfevents.1670806047.b23159305be2.526.0 → Jan23_04-20-07_441e7d2b355d/events.out.tfevents.1674447614.441e7d2b355d.71996.0}
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f5f45a3b0a6cfa3824cd25ac989119717eac3154fc17ac5584776b58c7386df3
|
3 |
+
size 103538
|
scaler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 557
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b691e5fdc050ffec60fabc6f70ec25a5c9c5a415de9ceda0c71eee2c5823fe64
|
3 |
size 557
|
scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 627
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:931a8817a8317861e6b08e3c4cd7129610781f77531c4af236b65ab04753f525
|
3 |
size 627
|
tokenizer_config.json
CHANGED
@@ -19,7 +19,7 @@
|
|
19 |
},
|
20 |
"errors": "replace",
|
21 |
"model_max_length": 1024,
|
22 |
-
"name_or_path": "
|
23 |
"pad_token": null,
|
24 |
"processor_class": "WhisperProcessor",
|
25 |
"return_attention_mask": false,
|
|
|
19 |
},
|
20 |
"errors": "replace",
|
21 |
"model_max_length": 1024,
|
22 |
+
"name_or_path": "openai/whisper-small",
|
23 |
"pad_token": null,
|
24 |
"processor_class": "WhisperProcessor",
|
25 |
"return_attention_mask": false,
|
trainer_state.json
CHANGED
@@ -1,1261 +1,3370 @@
|
|
1 |
{
|
2 |
-
"best_metric":
|
3 |
-
"best_model_checkpoint": "./
|
4 |
-
"epoch": 3.
|
5 |
-
"global_step":
|
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.
|
12 |
-
"learning_rate":
|
13 |
-
"loss":
|
14 |
"step": 25
|
15 |
},
|
16 |
{
|
17 |
-
"epoch": 0.
|
18 |
-
"learning_rate":
|
19 |
-
"loss":
|
20 |
"step": 50
|
21 |
},
|
22 |
{
|
23 |
-
"epoch": 0.
|
24 |
-
"learning_rate":
|
25 |
-
"loss":
|
26 |
"step": 75
|
27 |
},
|
28 |
{
|
29 |
-
"epoch": 0.
|
30 |
-
"learning_rate":
|
31 |
-
"loss": 0.
|
32 |
"step": 100
|
33 |
},
|
34 |
{
|
35 |
-
"epoch": 0.
|
36 |
-
"learning_rate":
|
37 |
-
"loss": 0.
|
38 |
"step": 125
|
39 |
},
|
40 |
{
|
41 |
-
"epoch": 0.
|
42 |
-
"learning_rate":
|
43 |
-
"loss": 0.
|
44 |
"step": 150
|
45 |
},
|
46 |
{
|
47 |
-
"epoch": 0.
|
48 |
-
"learning_rate":
|
49 |
-
"loss": 0.
|
50 |
"step": 175
|
51 |
},
|
52 |
{
|
53 |
-
"epoch": 0.
|
54 |
-
"learning_rate":
|
55 |
-
"loss": 0.
|
56 |
"step": 200
|
57 |
},
|
58 |
{
|
59 |
-
"epoch": 0.
|
60 |
-
"learning_rate":
|
61 |
-
"loss": 0.
|
62 |
"step": 225
|
63 |
},
|
64 |
{
|
65 |
-
"epoch": 0.
|
66 |
-
"learning_rate":
|
67 |
-
"loss": 0.
|
68 |
"step": 250
|
69 |
},
|
70 |
{
|
71 |
-
"epoch": 0.
|
72 |
-
"learning_rate":
|
73 |
-
"loss": 0.
|
74 |
"step": 275
|
75 |
},
|
76 |
{
|
77 |
-
"epoch": 0.
|
78 |
-
"learning_rate":
|
79 |
-
"loss": 0.
|
80 |
"step": 300
|
81 |
},
|
82 |
{
|
83 |
-
"epoch": 0.
|
84 |
-
"learning_rate":
|
85 |
-
"loss": 0.
|
86 |
"step": 325
|
87 |
},
|
88 |
{
|
89 |
-
"epoch": 0.
|
90 |
-
"learning_rate":
|
91 |
-
"loss": 0.
|
92 |
"step": 350
|
93 |
},
|
94 |
{
|
95 |
-
"epoch": 0.
|
96 |
-
"learning_rate":
|
97 |
-
"loss": 0.
|
98 |
"step": 375
|
99 |
},
|
100 |
{
|
101 |
-
"epoch": 0.
|
102 |
-
"learning_rate":
|
103 |
-
"loss": 0.
|
104 |
"step": 400
|
105 |
},
|
106 |
{
|
107 |
-
"epoch": 0.
|
108 |
-
"learning_rate":
|
109 |
-
"loss": 0.
|
110 |
"step": 425
|
111 |
},
|
112 |
{
|
113 |
-
"epoch": 0.
|
114 |
-
"learning_rate":
|
115 |
-
"loss": 0.
|
116 |
"step": 450
|
117 |
},
|
118 |
{
|
119 |
-
"epoch": 0.
|
120 |
-
"learning_rate":
|
121 |
-
"loss": 0.
|
122 |
"step": 475
|
123 |
},
|
124 |
{
|
125 |
-
"epoch": 0.
|
126 |
-
"learning_rate":
|
127 |
-
"loss": 0.
|
128 |
"step": 500
|
129 |
},
|
130 |
{
|
131 |
-
"epoch": 0.
|
132 |
-
"
|
133 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
134 |
"step": 525
|
135 |
},
|
136 |
{
|
137 |
-
"epoch": 0.
|
138 |
-
"learning_rate":
|
139 |
-
"loss": 0.
|
140 |
"step": 550
|
141 |
},
|
142 |
{
|
143 |
-
"epoch": 0.
|
144 |
-
"learning_rate":
|
145 |
-
"loss": 0.
|
146 |
"step": 575
|
147 |
},
|
148 |
{
|
149 |
-
"epoch": 0.
|
150 |
-
"learning_rate":
|
151 |
-
"loss": 0.
|
152 |
"step": 600
|
153 |
},
|
154 |
{
|
155 |
-
"epoch": 0.
|
156 |
-
"learning_rate":
|
157 |
-
"loss": 0.
|
158 |
"step": 625
|
159 |
},
|
160 |
{
|
161 |
-
"epoch": 0.
|
162 |
-
"learning_rate":
|
163 |
-
"loss": 0.
|
164 |
"step": 650
|
165 |
},
|
166 |
{
|
167 |
-
"epoch": 0.
|
168 |
-
"learning_rate":
|
169 |
-
"loss": 0.
|
170 |
"step": 675
|
171 |
},
|
172 |
{
|
173 |
-
"epoch": 0.
|
174 |
-
"learning_rate":
|
175 |
-
"loss": 0.
|
176 |
"step": 700
|
177 |
},
|
178 |
{
|
179 |
-
"epoch": 0.
|
180 |
-
"learning_rate":
|
181 |
-
"loss": 0.
|
182 |
"step": 725
|
183 |
},
|
184 |
{
|
185 |
-
"epoch": 0.
|
186 |
-
"learning_rate":
|
187 |
-
"loss": 0.
|
188 |
"step": 750
|
189 |
},
|
190 |
{
|
191 |
-
"epoch": 0.
|
192 |
-
"learning_rate":
|
193 |
-
"loss": 0.
|
194 |
"step": 775
|
195 |
},
|
196 |
{
|
197 |
-
"epoch": 0.
|
198 |
-
"learning_rate":
|
199 |
-
"loss": 0.
|
200 |
"step": 800
|
201 |
},
|
202 |
{
|
203 |
-
"epoch": 0.
|
204 |
-
"learning_rate":
|
205 |
-
"loss": 0.
|
206 |
"step": 825
|
207 |
},
|
208 |
{
|
209 |
-
"epoch": 0.
|
210 |
-
"learning_rate":
|
211 |
-
"loss": 0.
|
212 |
"step": 850
|
213 |
},
|
214 |
{
|
215 |
-
"epoch": 0.
|
216 |
-
"learning_rate":
|
217 |
-
"loss": 0.
|
218 |
"step": 875
|
219 |
},
|
220 |
{
|
221 |
-
"epoch": 0.
|
222 |
-
"learning_rate":
|
223 |
-
"loss": 0.
|
224 |
"step": 900
|
225 |
},
|
226 |
{
|
227 |
-
"epoch": 0.
|
228 |
-
"learning_rate":
|
229 |
-
"loss": 0.
|
230 |
"step": 925
|
231 |
},
|
232 |
{
|
233 |
-
"epoch": 0.
|
234 |
-
"learning_rate":
|
235 |
-
"loss": 0.
|
236 |
"step": 950
|
237 |
},
|
238 |
{
|
239 |
-
"epoch": 0.
|
240 |
-
"learning_rate":
|
241 |
-
"loss": 0.
|
242 |
"step": 975
|
243 |
},
|
244 |
{
|
245 |
-
"epoch": 0.
|
246 |
-
"learning_rate":
|
247 |
-
"loss": 0.
|
248 |
"step": 1000
|
249 |
},
|
250 |
{
|
251 |
-
"epoch": 0.
|
252 |
-
"eval_cer": 14.
|
253 |
-
"eval_loss": 0.
|
254 |
-
"eval_runtime":
|
255 |
-
"eval_samples_per_second": 4.
|
256 |
-
"eval_steps_per_second": 0.
|
257 |
"step": 1000
|
258 |
},
|
259 |
{
|
260 |
-
"epoch": 0.
|
261 |
-
"learning_rate":
|
262 |
-
"loss": 0.
|
263 |
"step": 1025
|
264 |
},
|
265 |
{
|
266 |
-
"epoch": 0.
|
267 |
-
"learning_rate":
|
268 |
-
"loss": 0.
|
269 |
"step": 1050
|
270 |
},
|
271 |
{
|
272 |
-
"epoch": 0.
|
273 |
-
"learning_rate":
|
274 |
-
"loss": 0.
|
275 |
"step": 1075
|
276 |
},
|
277 |
{
|
278 |
-
"epoch": 0.
|
279 |
-
"learning_rate":
|
280 |
-
"loss": 0.
|
281 |
"step": 1100
|
282 |
},
|
283 |
{
|
284 |
-
"epoch": 0.
|
285 |
-
"learning_rate":
|
286 |
-
"loss": 0.
|
287 |
"step": 1125
|
288 |
},
|
289 |
{
|
290 |
-
"epoch": 0.
|
291 |
-
"learning_rate":
|
292 |
-
"loss": 0.
|
293 |
"step": 1150
|
294 |
},
|
295 |
{
|
296 |
-
"epoch": 0.
|
297 |
-
"learning_rate":
|
298 |
-
"loss": 0.
|
299 |
"step": 1175
|
300 |
},
|
301 |
{
|
302 |
-
"epoch": 0.
|
303 |
-
"learning_rate":
|
304 |
-
"loss": 0.
|
305 |
"step": 1200
|
306 |
},
|
307 |
{
|
308 |
-
"epoch": 0.
|
309 |
-
"learning_rate":
|
310 |
-
"loss": 0.
|
311 |
"step": 1225
|
312 |
},
|
313 |
{
|
314 |
-
"epoch": 0.
|
315 |
-
"learning_rate":
|
316 |
-
"loss": 0.
|
317 |
"step": 1250
|
318 |
},
|
319 |
{
|
320 |
-
"epoch": 0.
|
321 |
-
"learning_rate":
|
322 |
-
"loss": 0.
|
323 |
"step": 1275
|
324 |
},
|
325 |
{
|
326 |
-
"epoch": 0.
|
327 |
-
"learning_rate":
|
328 |
-
"loss": 0.
|
329 |
"step": 1300
|
330 |
},
|
331 |
{
|
332 |
-
"epoch": 0.
|
333 |
-
"learning_rate":
|
334 |
-
"loss": 0.
|
335 |
"step": 1325
|
336 |
},
|
337 |
{
|
338 |
-
"epoch": 0.
|
339 |
-
"learning_rate":
|
340 |
-
"loss": 0.
|
341 |
"step": 1350
|
342 |
},
|
343 |
{
|
344 |
-
"epoch": 0.
|
345 |
-
"learning_rate":
|
346 |
-
"loss": 0.
|
347 |
"step": 1375
|
348 |
},
|
349 |
{
|
350 |
-
"epoch": 0.
|
351 |
-
"learning_rate":
|
352 |
-
"loss": 0.
|
353 |
"step": 1400
|
354 |
},
|
355 |
{
|
356 |
-
"epoch": 0.
|
357 |
-
"learning_rate":
|
358 |
-
"loss": 0.
|
359 |
"step": 1425
|
360 |
},
|
361 |
{
|
362 |
-
"epoch": 0.
|
363 |
-
"learning_rate":
|
364 |
-
"loss": 0.
|
365 |
"step": 1450
|
366 |
},
|
367 |
{
|
368 |
-
"epoch": 0.
|
369 |
-
"learning_rate":
|
370 |
-
"loss": 0.
|
371 |
"step": 1475
|
372 |
},
|
373 |
{
|
374 |
-
"epoch":
|
375 |
-
"learning_rate":
|
376 |
-
"loss": 0.
|
377 |
"step": 1500
|
378 |
},
|
379 |
{
|
380 |
-
"epoch":
|
381 |
-
"
|
382 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
383 |
"step": 1525
|
384 |
},
|
385 |
{
|
386 |
-
"epoch":
|
387 |
-
"learning_rate":
|
388 |
-
"loss": 0.
|
389 |
"step": 1550
|
390 |
},
|
391 |
{
|
392 |
-
"epoch":
|
393 |
-
"learning_rate":
|
394 |
-
"loss": 0.
|
395 |
"step": 1575
|
396 |
},
|
397 |
{
|
398 |
-
"epoch":
|
399 |
-
"learning_rate":
|
400 |
-
"loss": 0.
|
401 |
"step": 1600
|
402 |
},
|
403 |
{
|
404 |
-
"epoch":
|
405 |
-
"learning_rate":
|
406 |
-
"loss": 0.
|
407 |
"step": 1625
|
408 |
},
|
409 |
{
|
410 |
-
"epoch":
|
411 |
-
"learning_rate":
|
412 |
-
"loss": 0.
|
413 |
"step": 1650
|
414 |
},
|
415 |
{
|
416 |
-
"epoch":
|
417 |
-
"learning_rate":
|
418 |
-
"loss": 0.
|
419 |
"step": 1675
|
420 |
},
|
421 |
{
|
422 |
-
"epoch":
|
423 |
-
"learning_rate":
|
424 |
-
"loss": 0.
|
425 |
"step": 1700
|
426 |
},
|
427 |
{
|
428 |
-
"epoch":
|
429 |
-
"learning_rate":
|
430 |
-
"loss": 0.
|
431 |
"step": 1725
|
432 |
},
|
433 |
{
|
434 |
-
"epoch":
|
435 |
-
"learning_rate":
|
436 |
-
"loss": 0.
|
437 |
"step": 1750
|
438 |
},
|
439 |
{
|
440 |
-
"epoch":
|
441 |
-
"learning_rate":
|
442 |
-
"loss": 0.
|
443 |
"step": 1775
|
444 |
},
|
445 |
{
|
446 |
-
"epoch":
|
447 |
-
"learning_rate":
|
448 |
-
"loss": 0.
|
449 |
"step": 1800
|
450 |
},
|
451 |
{
|
452 |
-
"epoch":
|
453 |
-
"learning_rate":
|
454 |
-
"loss": 0.
|
455 |
"step": 1825
|
456 |
},
|
457 |
{
|
458 |
-
"epoch":
|
459 |
-
"learning_rate":
|
460 |
-
"loss": 0.
|
461 |
"step": 1850
|
462 |
},
|
463 |
{
|
464 |
-
"epoch":
|
465 |
-
"learning_rate":
|
466 |
-
"loss": 0.
|
467 |
"step": 1875
|
468 |
},
|
469 |
{
|
470 |
-
"epoch":
|
471 |
-
"learning_rate":
|
472 |
-
"loss": 0.
|
473 |
"step": 1900
|
474 |
},
|
475 |
{
|
476 |
-
"epoch":
|
477 |
-
"learning_rate":
|
478 |
-
"loss": 0.
|
479 |
"step": 1925
|
480 |
},
|
481 |
{
|
482 |
-
"epoch":
|
483 |
-
"learning_rate":
|
484 |
-
"loss": 0.
|
485 |
"step": 1950
|
486 |
},
|
487 |
{
|
488 |
-
"epoch":
|
489 |
-
"learning_rate":
|
490 |
-
"loss": 0.
|
491 |
"step": 1975
|
492 |
},
|
493 |
{
|
494 |
-
"epoch":
|
495 |
-
"learning_rate":
|
496 |
-
"loss": 0.
|
497 |
"step": 2000
|
498 |
},
|
499 |
{
|
500 |
-
"epoch":
|
501 |
-
"eval_cer":
|
502 |
-
"eval_loss": 0.
|
503 |
-
"eval_runtime":
|
504 |
-
"eval_samples_per_second": 4.
|
505 |
-
"eval_steps_per_second": 0.
|
506 |
"step": 2000
|
507 |
},
|
508 |
{
|
509 |
-
"epoch":
|
510 |
-
"learning_rate":
|
511 |
-
"loss": 0.
|
512 |
"step": 2025
|
513 |
},
|
514 |
{
|
515 |
-
"epoch":
|
516 |
-
"learning_rate":
|
517 |
-
"loss": 0.
|
518 |
"step": 2050
|
519 |
},
|
520 |
{
|
521 |
-
"epoch":
|
522 |
-
"learning_rate":
|
523 |
-
"loss": 0.
|
524 |
"step": 2075
|
525 |
},
|
526 |
{
|
527 |
-
"epoch":
|
528 |
-
"learning_rate":
|
529 |
-
"loss": 0.
|
530 |
"step": 2100
|
531 |
},
|
532 |
{
|
533 |
-
"epoch":
|
534 |
-
"learning_rate":
|
535 |
-
"loss": 0.
|
536 |
"step": 2125
|
537 |
},
|
538 |
{
|
539 |
-
"epoch":
|
540 |
-
"learning_rate":
|
541 |
-
"loss": 0.
|
542 |
"step": 2150
|
543 |
},
|
544 |
{
|
545 |
-
"epoch":
|
546 |
-
"learning_rate":
|
547 |
-
"loss": 0.
|
548 |
"step": 2175
|
549 |
},
|
550 |
{
|
551 |
-
"epoch":
|
552 |
-
"learning_rate":
|
553 |
-
"loss": 0.
|
554 |
"step": 2200
|
555 |
},
|
556 |
{
|
557 |
-
"epoch":
|
558 |
-
"learning_rate":
|
559 |
-
"loss": 0.
|
560 |
"step": 2225
|
561 |
},
|
562 |
{
|
563 |
-
"epoch":
|
564 |
-
"learning_rate":
|
565 |
-
"loss": 0.
|
566 |
"step": 2250
|
567 |
},
|
568 |
{
|
569 |
-
"epoch":
|
570 |
-
"learning_rate":
|
571 |
-
"loss": 0.
|
572 |
"step": 2275
|
573 |
},
|
574 |
{
|
575 |
-
"epoch":
|
576 |
-
"learning_rate":
|
577 |
-
"loss": 0.
|
578 |
"step": 2300
|
579 |
},
|
580 |
{
|
581 |
-
"epoch":
|
582 |
-
"learning_rate":
|
583 |
-
"loss": 0.
|
584 |
"step": 2325
|
585 |
},
|
586 |
{
|
587 |
-
"epoch":
|
588 |
-
"learning_rate":
|
589 |
-
"loss": 0.
|
590 |
"step": 2350
|
591 |
},
|
592 |
{
|
593 |
-
"epoch":
|
594 |
-
"learning_rate":
|
595 |
-
"loss": 0.
|
596 |
"step": 2375
|
597 |
},
|
598 |
{
|
599 |
-
"epoch":
|
600 |
-
"learning_rate":
|
601 |
-
"loss": 0.
|
602 |
"step": 2400
|
603 |
},
|
604 |
{
|
605 |
-
"epoch":
|
606 |
-
"learning_rate":
|
607 |
-
"loss": 0.
|
608 |
"step": 2425
|
609 |
},
|
610 |
{
|
611 |
-
"epoch":
|
612 |
-
"learning_rate":
|
613 |
-
"loss": 0.
|
614 |
"step": 2450
|
615 |
},
|
616 |
{
|
617 |
-
"epoch":
|
618 |
-
"learning_rate":
|
619 |
-
"loss": 0.
|
620 |
"step": 2475
|
621 |
},
|
622 |
{
|
623 |
-
"epoch":
|
624 |
-
"learning_rate":
|
625 |
-
"loss": 0.
|
626 |
"step": 2500
|
627 |
},
|
628 |
{
|
629 |
-
"epoch":
|
630 |
-
"
|
631 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
632 |
"step": 2525
|
633 |
},
|
634 |
{
|
635 |
-
"epoch":
|
636 |
-
"learning_rate":
|
637 |
-
"loss": 0.
|
638 |
"step": 2550
|
639 |
},
|
640 |
{
|
641 |
-
"epoch":
|
642 |
-
"learning_rate":
|
643 |
-
"loss": 0.
|
644 |
"step": 2575
|
645 |
},
|
646 |
{
|
647 |
-
"epoch":
|
648 |
-
"learning_rate":
|
649 |
-
"loss": 0.
|
650 |
"step": 2600
|
651 |
},
|
652 |
{
|
653 |
-
"epoch":
|
654 |
-
"learning_rate":
|
655 |
-
"loss": 0.
|
656 |
"step": 2625
|
657 |
},
|
658 |
{
|
659 |
-
"epoch":
|
660 |
-
"learning_rate":
|
661 |
-
"loss": 0.
|
662 |
"step": 2650
|
663 |
},
|
664 |
{
|
665 |
-
"epoch":
|
666 |
-
"learning_rate":
|
667 |
-
"loss": 0.
|
668 |
"step": 2675
|
669 |
},
|
670 |
{
|
671 |
-
"epoch":
|
672 |
-
"learning_rate":
|
673 |
-
"loss": 0.
|
674 |
"step": 2700
|
675 |
},
|
676 |
{
|
677 |
-
"epoch":
|
678 |
-
"learning_rate":
|
679 |
-
"loss": 0.
|
680 |
"step": 2725
|
681 |
},
|
682 |
{
|
683 |
-
"epoch":
|
684 |
-
"learning_rate":
|
685 |
-
"loss": 0.
|
686 |
"step": 2750
|
687 |
},
|
688 |
{
|
689 |
-
"epoch":
|
690 |
-
"learning_rate": 4.
|
691 |
-
"loss": 0.
|
692 |
"step": 2775
|
693 |
},
|
694 |
{
|
695 |
-
"epoch":
|
696 |
-
"learning_rate": 4.
|
697 |
-
"loss": 0.
|
698 |
"step": 2800
|
699 |
},
|
700 |
{
|
701 |
-
"epoch":
|
702 |
-
"learning_rate": 4.
|
703 |
-
"loss": 0.
|
704 |
"step": 2825
|
705 |
},
|
706 |
{
|
707 |
-
"epoch":
|
708 |
-
"learning_rate": 4.
|
709 |
-
"loss": 0.
|
710 |
"step": 2850
|
711 |
},
|
712 |
{
|
713 |
-
"epoch":
|
714 |
-
"learning_rate": 4.
|
715 |
-
"loss": 0.
|
716 |
"step": 2875
|
717 |
},
|
718 |
{
|
719 |
-
"epoch":
|
720 |
-
"learning_rate": 4.
|
721 |
-
"loss": 0.
|
722 |
"step": 2900
|
723 |
},
|
724 |
{
|
725 |
-
"epoch":
|
726 |
-
"learning_rate": 4.
|
727 |
-
"loss": 0.
|
728 |
"step": 2925
|
729 |
},
|
730 |
{
|
731 |
-
"epoch":
|
732 |
-
"learning_rate": 4.
|
733 |
-
"loss": 0.
|
734 |
"step": 2950
|
735 |
},
|
736 |
{
|
737 |
-
"epoch":
|
738 |
-
"learning_rate": 4.
|
739 |
-
"loss": 0.
|
740 |
"step": 2975
|
741 |
},
|
742 |
{
|
743 |
-
"epoch":
|
744 |
-
"learning_rate": 4.
|
745 |
-
"loss": 0.
|
746 |
"step": 3000
|
747 |
},
|
748 |
{
|
749 |
-
"epoch":
|
750 |
-
"eval_cer":
|
751 |
-
"eval_loss": 0.
|
752 |
-
"eval_runtime":
|
753 |
-
"eval_samples_per_second": 4.
|
754 |
-
"eval_steps_per_second": 0.
|
755 |
"step": 3000
|
756 |
},
|
757 |
{
|
758 |
-
"epoch":
|
759 |
-
"learning_rate": 4.
|
760 |
-
"loss": 0.
|
761 |
"step": 3025
|
762 |
},
|
763 |
{
|
764 |
-
"epoch":
|
765 |
-
"learning_rate": 4.
|
766 |
-
"loss": 0.
|
767 |
"step": 3050
|
768 |
},
|
769 |
{
|
770 |
-
"epoch":
|
771 |
-
"learning_rate": 4.
|
772 |
-
"loss": 0.
|
773 |
"step": 3075
|
774 |
},
|
775 |
{
|
776 |
-
"epoch":
|
777 |
-
"learning_rate": 4.
|
778 |
-
"loss": 0.
|
779 |
"step": 3100
|
780 |
},
|
781 |
{
|
782 |
-
"epoch":
|
783 |
-
"learning_rate": 4.
|
784 |
-
"loss": 0.
|
785 |
"step": 3125
|
786 |
},
|
787 |
{
|
788 |
-
"epoch":
|
789 |
-
"learning_rate": 4.
|
790 |
-
"loss": 0.
|
791 |
"step": 3150
|
792 |
},
|
793 |
{
|
794 |
-
"epoch":
|
795 |
-
"learning_rate": 4.
|
796 |
-
"loss": 0.
|
797 |
"step": 3175
|
798 |
},
|
799 |
{
|
800 |
-
"epoch":
|
801 |
-
"learning_rate": 4.
|
802 |
-
"loss": 0.
|
803 |
"step": 3200
|
804 |
},
|
805 |
{
|
806 |
-
"epoch":
|
807 |
-
"learning_rate":
|
808 |
-
"loss": 0.
|
809 |
"step": 3225
|
810 |
},
|
811 |
{
|
812 |
-
"epoch":
|
813 |
-
"learning_rate":
|
814 |
-
"loss": 0.
|
815 |
"step": 3250
|
816 |
},
|
817 |
{
|
818 |
-
"epoch":
|
819 |
-
"learning_rate":
|
820 |
-
"loss": 0.
|
821 |
"step": 3275
|
822 |
},
|
823 |
{
|
824 |
-
"epoch":
|
825 |
-
"learning_rate":
|
826 |
-
"loss": 0.
|
827 |
"step": 3300
|
828 |
},
|
829 |
{
|
830 |
-
"epoch":
|
831 |
-
"learning_rate":
|
832 |
-
"loss": 0.
|
833 |
"step": 3325
|
834 |
},
|
835 |
{
|
836 |
-
"epoch":
|
837 |
-
"learning_rate":
|
838 |
-
"loss": 0.
|
839 |
"step": 3350
|
840 |
},
|
841 |
{
|
842 |
-
"epoch":
|
843 |
-
"learning_rate":
|
844 |
-
"loss": 0.
|
845 |
"step": 3375
|
846 |
},
|
847 |
{
|
848 |
-
"epoch":
|
849 |
-
"learning_rate":
|
850 |
-
"loss": 0.
|
851 |
"step": 3400
|
852 |
},
|
853 |
{
|
854 |
-
"epoch":
|
855 |
-
"learning_rate": 3.
|
856 |
-
"loss": 0.
|
857 |
"step": 3425
|
858 |
},
|
859 |
{
|
860 |
-
"epoch":
|
861 |
-
"learning_rate": 3.
|
862 |
-
"loss": 0.
|
863 |
"step": 3450
|
864 |
},
|
865 |
{
|
866 |
-
"epoch":
|
867 |
-
"learning_rate": 3.
|
868 |
-
"loss": 0.
|
869 |
"step": 3475
|
870 |
},
|
871 |
{
|
872 |
-
"epoch":
|
873 |
-
"learning_rate": 3.
|
874 |
-
"loss": 0.
|
875 |
"step": 3500
|
876 |
},
|
877 |
{
|
878 |
-
"epoch":
|
879 |
-
"
|
880 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
881 |
"step": 3525
|
882 |
},
|
883 |
{
|
884 |
-
"epoch":
|
885 |
-
"learning_rate": 3.
|
886 |
-
"loss": 0.
|
887 |
"step": 3550
|
888 |
},
|
889 |
{
|
890 |
-
"epoch":
|
891 |
-
"learning_rate": 3.
|
892 |
-
"loss": 0.
|
893 |
"step": 3575
|
894 |
},
|
895 |
{
|
896 |
-
"epoch":
|
897 |
-
"learning_rate": 3.
|
898 |
-
"loss": 0.
|
899 |
"step": 3600
|
900 |
},
|
901 |
{
|
902 |
-
"epoch":
|
903 |
-
"learning_rate": 3.
|
904 |
-
"loss": 0.
|
905 |
"step": 3625
|
906 |
},
|
907 |
{
|
908 |
-
"epoch":
|
909 |
-
"learning_rate": 3.
|
910 |
-
"loss": 0.
|
911 |
"step": 3650
|
912 |
},
|
913 |
{
|
914 |
-
"epoch":
|
915 |
-
"learning_rate":
|
916 |
-
"loss": 0.
|
917 |
"step": 3675
|
918 |
},
|
919 |
{
|
920 |
-
"epoch":
|
921 |
-
"learning_rate":
|
922 |
-
"loss": 0.
|
923 |
"step": 3700
|
924 |
},
|
925 |
{
|
926 |
-
"epoch":
|
927 |
-
"learning_rate":
|
928 |
-
"loss": 0.
|
929 |
"step": 3725
|
930 |
},
|
931 |
{
|
932 |
-
"epoch":
|
933 |
-
"learning_rate":
|
934 |
-
"loss": 0.
|
935 |
"step": 3750
|
936 |
},
|
937 |
{
|
938 |
-
"epoch":
|
939 |
-
"learning_rate":
|
940 |
-
"loss": 0.
|
941 |
"step": 3775
|
942 |
},
|
943 |
{
|
944 |
-
"epoch":
|
945 |
-
"learning_rate":
|
946 |
-
"loss": 0.
|
947 |
"step": 3800
|
948 |
},
|
949 |
{
|
950 |
-
"epoch":
|
951 |
-
"learning_rate":
|
952 |
-
"loss": 0.
|
953 |
"step": 3825
|
954 |
},
|
955 |
{
|
956 |
-
"epoch":
|
957 |
-
"learning_rate":
|
958 |
-
"loss": 0.
|
959 |
"step": 3850
|
960 |
},
|
961 |
{
|
962 |
-
"epoch":
|
963 |
-
"learning_rate":
|
964 |
-
"loss": 0.
|
965 |
"step": 3875
|
966 |
},
|
967 |
{
|
968 |
-
"epoch":
|
969 |
-
"learning_rate":
|
970 |
-
"loss": 0.
|
971 |
"step": 3900
|
972 |
},
|
973 |
{
|
974 |
-
"epoch":
|
975 |
-
"learning_rate":
|
976 |
-
"loss": 0.
|
977 |
"step": 3925
|
978 |
},
|
979 |
{
|
980 |
-
"epoch":
|
981 |
-
"learning_rate":
|
982 |
-
"loss": 0.
|
983 |
"step": 3950
|
984 |
},
|
985 |
{
|
986 |
-
"epoch":
|
987 |
-
"learning_rate":
|
988 |
-
"loss": 0.
|
989 |
"step": 3975
|
990 |
},
|
991 |
{
|
992 |
-
"epoch":
|
993 |
-
"learning_rate":
|
994 |
-
"loss": 0.
|
995 |
"step": 4000
|
996 |
},
|
997 |
{
|
998 |
-
"epoch":
|
999 |
-
"eval_cer": 11.
|
1000 |
-
"eval_loss": 0.
|
1001 |
-
"eval_runtime":
|
1002 |
-
"eval_samples_per_second": 4.
|
1003 |
-
"eval_steps_per_second": 0.
|
1004 |
"step": 4000
|
1005 |
},
|
1006 |
{
|
1007 |
-
"epoch":
|
1008 |
-
"learning_rate":
|
1009 |
-
"loss": 0.
|
1010 |
"step": 4025
|
1011 |
},
|
1012 |
{
|
1013 |
-
"epoch":
|
1014 |
-
"learning_rate":
|
1015 |
-
"loss": 0.
|
1016 |
"step": 4050
|
1017 |
},
|
1018 |
{
|
1019 |
-
"epoch":
|
1020 |
-
"learning_rate":
|
1021 |
-
"loss": 0.
|
1022 |
"step": 4075
|
1023 |
},
|
1024 |
{
|
1025 |
-
"epoch":
|
1026 |
-
"learning_rate":
|
1027 |
-
"loss": 0.
|
1028 |
"step": 4100
|
1029 |
},
|
1030 |
{
|
1031 |
-
"epoch":
|
1032 |
-
"learning_rate":
|
1033 |
-
"loss": 0.
|
1034 |
"step": 4125
|
1035 |
},
|
1036 |
{
|
1037 |
-
"epoch":
|
1038 |
-
"learning_rate":
|
1039 |
-
"loss": 0.
|
1040 |
"step": 4150
|
1041 |
},
|
1042 |
{
|
1043 |
-
"epoch":
|
1044 |
-
"learning_rate":
|
1045 |
-
"loss": 0.
|
1046 |
"step": 4175
|
1047 |
},
|
1048 |
{
|
1049 |
-
"epoch":
|
1050 |
-
"learning_rate":
|
1051 |
-
"loss": 0.
|
1052 |
"step": 4200
|
1053 |
},
|
1054 |
{
|
1055 |
-
"epoch":
|
1056 |
-
"learning_rate":
|
1057 |
-
"loss": 0.
|
1058 |
"step": 4225
|
1059 |
},
|
1060 |
{
|
1061 |
-
"epoch":
|
1062 |
-
"learning_rate":
|
1063 |
-
"loss": 0.
|
1064 |
"step": 4250
|
1065 |
},
|
1066 |
{
|
1067 |
-
"epoch":
|
1068 |
-
"learning_rate":
|
1069 |
-
"loss": 0.
|
1070 |
"step": 4275
|
1071 |
},
|
1072 |
{
|
1073 |
-
"epoch":
|
1074 |
-
"learning_rate":
|
1075 |
-
"loss": 0.
|
1076 |
"step": 4300
|
1077 |
},
|
1078 |
{
|
1079 |
-
"epoch": 2
|
1080 |
-
"learning_rate":
|
1081 |
-
"loss": 0.
|
1082 |
"step": 4325
|
1083 |
},
|
1084 |
{
|
1085 |
-
"epoch":
|
1086 |
-
"learning_rate":
|
1087 |
-
"loss": 0.
|
1088 |
"step": 4350
|
1089 |
},
|
1090 |
{
|
1091 |
-
"epoch":
|
1092 |
-
"learning_rate":
|
1093 |
-
"loss": 0.
|
1094 |
"step": 4375
|
1095 |
},
|
1096 |
{
|
1097 |
-
"epoch":
|
1098 |
-
"learning_rate":
|
1099 |
-
"loss": 0.
|
1100 |
"step": 4400
|
1101 |
},
|
1102 |
{
|
1103 |
-
"epoch":
|
1104 |
-
"learning_rate":
|
1105 |
-
"loss": 0.
|
1106 |
"step": 4425
|
1107 |
},
|
1108 |
{
|
1109 |
-
"epoch":
|
1110 |
-
"learning_rate":
|
1111 |
-
"loss": 0.
|
1112 |
"step": 4450
|
1113 |
},
|
1114 |
{
|
1115 |
-
"epoch":
|
1116 |
-
"learning_rate":
|
1117 |
-
"loss": 0.
|
1118 |
"step": 4475
|
1119 |
},
|
1120 |
{
|
1121 |
-
"epoch":
|
1122 |
-
"learning_rate":
|
1123 |
-
"loss": 0.
|
1124 |
"step": 4500
|
1125 |
},
|
1126 |
{
|
1127 |
-
"epoch":
|
1128 |
-
"
|
1129 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1130 |
"step": 4525
|
1131 |
},
|
1132 |
{
|
1133 |
-
"epoch":
|
1134 |
-
"learning_rate":
|
1135 |
-
"loss": 0.
|
1136 |
"step": 4550
|
1137 |
},
|
1138 |
{
|
1139 |
-
"epoch":
|
1140 |
-
"learning_rate":
|
1141 |
-
"loss": 0.
|
1142 |
"step": 4575
|
1143 |
},
|
1144 |
{
|
1145 |
-
"epoch":
|
1146 |
-
"learning_rate":
|
1147 |
-
"loss": 0.
|
1148 |
"step": 4600
|
1149 |
},
|
1150 |
{
|
1151 |
-
"epoch":
|
1152 |
-
"learning_rate":
|
1153 |
-
"loss": 0.
|
1154 |
"step": 4625
|
1155 |
},
|
1156 |
{
|
1157 |
-
"epoch":
|
1158 |
-
"learning_rate":
|
1159 |
-
"loss": 0.
|
1160 |
"step": 4650
|
1161 |
},
|
1162 |
{
|
1163 |
-
"epoch": 3
|
1164 |
-
"learning_rate":
|
1165 |
-
"loss": 0.
|
1166 |
"step": 4675
|
1167 |
},
|
1168 |
{
|
1169 |
-
"epoch": 3
|
1170 |
-
"learning_rate":
|
1171 |
-
"loss": 0.
|
1172 |
"step": 4700
|
1173 |
},
|
1174 |
{
|
1175 |
-
"epoch":
|
1176 |
-
"learning_rate":
|
1177 |
-
"loss": 0.
|
1178 |
"step": 4725
|
1179 |
},
|
1180 |
{
|
1181 |
-
"epoch":
|
1182 |
-
"learning_rate":
|
1183 |
-
"loss": 0.
|
1184 |
"step": 4750
|
1185 |
},
|
1186 |
{
|
1187 |
-
"epoch":
|
1188 |
-
"learning_rate":
|
1189 |
-
"loss": 0.
|
1190 |
"step": 4775
|
1191 |
},
|
1192 |
{
|
1193 |
-
"epoch":
|
1194 |
-
"learning_rate":
|
1195 |
-
"loss": 0.
|
1196 |
"step": 4800
|
1197 |
},
|
1198 |
{
|
1199 |
-
"epoch":
|
1200 |
-
"learning_rate": 3.
|
1201 |
-
"loss": 0.
|
1202 |
"step": 4825
|
1203 |
},
|
1204 |
{
|
1205 |
-
"epoch":
|
1206 |
-
"learning_rate": 3.
|
1207 |
-
"loss": 0.
|
1208 |
"step": 4850
|
1209 |
},
|
1210 |
{
|
1211 |
-
"epoch":
|
1212 |
-
"learning_rate":
|
1213 |
-
"loss": 0.
|
1214 |
"step": 4875
|
1215 |
},
|
1216 |
{
|
1217 |
-
"epoch":
|
1218 |
-
"learning_rate":
|
1219 |
-
"loss": 0.
|
1220 |
"step": 4900
|
1221 |
},
|
1222 |
{
|
1223 |
-
"epoch":
|
1224 |
-
"learning_rate":
|
1225 |
-
"loss": 0.
|
1226 |
"step": 4925
|
1227 |
},
|
1228 |
{
|
1229 |
-
"epoch":
|
1230 |
-
"learning_rate":
|
1231 |
-
"loss": 0.
|
1232 |
"step": 4950
|
1233 |
},
|
1234 |
{
|
1235 |
-
"epoch":
|
1236 |
-
"learning_rate":
|
1237 |
-
"loss": 0.
|
1238 |
"step": 4975
|
1239 |
},
|
1240 |
{
|
1241 |
-
"epoch":
|
1242 |
-
"learning_rate":
|
1243 |
-
"loss": 0.
|
1244 |
"step": 5000
|
1245 |
},
|
1246 |
{
|
1247 |
-
"epoch":
|
1248 |
-
"eval_cer": 11.
|
1249 |
-
"eval_loss": 0.
|
1250 |
-
"eval_runtime":
|
1251 |
-
"eval_samples_per_second": 4.
|
1252 |
-
"eval_steps_per_second": 0.
|
1253 |
"step": 5000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1254 |
}
|
1255 |
],
|
1256 |
-
"max_steps":
|
1257 |
-
"num_train_epochs":
|
1258 |
-
"total_flos":
|
1259 |
"trial_name": null,
|
1260 |
"trial_params": null
|
1261 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 10.058774675781,
|
3 |
+
"best_model_checkpoint": "./model_out_trpro_more_more/checkpoint-13000",
|
4 |
+
"epoch": 3.6020224407812718,
|
5 |
+
"global_step": 13000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
+
"epoch": 0.01,
|
12 |
+
"learning_rate": 2.5e-06,
|
13 |
+
"loss": 0.487,
|
14 |
"step": 25
|
15 |
},
|
16 |
{
|
17 |
+
"epoch": 0.01,
|
18 |
+
"learning_rate": 5e-06,
|
19 |
+
"loss": 0.4472,
|
20 |
"step": 50
|
21 |
},
|
22 |
{
|
23 |
+
"epoch": 0.02,
|
24 |
+
"learning_rate": 7.5e-06,
|
25 |
+
"loss": 0.4173,
|
26 |
"step": 75
|
27 |
},
|
28 |
{
|
29 |
+
"epoch": 0.03,
|
30 |
+
"learning_rate": 1e-05,
|
31 |
+
"loss": 0.4623,
|
32 |
"step": 100
|
33 |
},
|
34 |
{
|
35 |
+
"epoch": 0.03,
|
36 |
+
"learning_rate": 1.25e-05,
|
37 |
+
"loss": 0.457,
|
38 |
"step": 125
|
39 |
},
|
40 |
{
|
41 |
+
"epoch": 0.04,
|
42 |
+
"learning_rate": 1.5e-05,
|
43 |
+
"loss": 0.4775,
|
44 |
"step": 150
|
45 |
},
|
46 |
{
|
47 |
+
"epoch": 0.05,
|
48 |
+
"learning_rate": 1.75e-05,
|
49 |
+
"loss": 0.4445,
|
50 |
"step": 175
|
51 |
},
|
52 |
{
|
53 |
+
"epoch": 0.06,
|
54 |
+
"learning_rate": 2e-05,
|
55 |
+
"loss": 0.4751,
|
56 |
"step": 200
|
57 |
},
|
58 |
{
|
59 |
+
"epoch": 0.06,
|
60 |
+
"learning_rate": 2.25e-05,
|
61 |
+
"loss": 0.4633,
|
62 |
"step": 225
|
63 |
},
|
64 |
{
|
65 |
+
"epoch": 0.07,
|
66 |
+
"learning_rate": 2.5e-05,
|
67 |
+
"loss": 0.4682,
|
68 |
"step": 250
|
69 |
},
|
70 |
{
|
71 |
+
"epoch": 0.08,
|
72 |
+
"learning_rate": 2.7500000000000004e-05,
|
73 |
+
"loss": 0.4807,
|
74 |
"step": 275
|
75 |
},
|
76 |
{
|
77 |
+
"epoch": 0.08,
|
78 |
+
"learning_rate": 3e-05,
|
79 |
+
"loss": 0.4882,
|
80 |
"step": 300
|
81 |
},
|
82 |
{
|
83 |
+
"epoch": 0.09,
|
84 |
+
"learning_rate": 3.2500000000000004e-05,
|
85 |
+
"loss": 0.4681,
|
86 |
"step": 325
|
87 |
},
|
88 |
{
|
89 |
+
"epoch": 0.1,
|
90 |
+
"learning_rate": 3.5e-05,
|
91 |
+
"loss": 0.4918,
|
92 |
"step": 350
|
93 |
},
|
94 |
{
|
95 |
+
"epoch": 0.1,
|
96 |
+
"learning_rate": 3.7500000000000003e-05,
|
97 |
+
"loss": 0.4871,
|
98 |
"step": 375
|
99 |
},
|
100 |
{
|
101 |
+
"epoch": 0.11,
|
102 |
+
"learning_rate": 4e-05,
|
103 |
+
"loss": 0.4986,
|
104 |
"step": 400
|
105 |
},
|
106 |
{
|
107 |
+
"epoch": 0.12,
|
108 |
+
"learning_rate": 4.25e-05,
|
109 |
+
"loss": 0.4859,
|
110 |
"step": 425
|
111 |
},
|
112 |
{
|
113 |
+
"epoch": 0.12,
|
114 |
+
"learning_rate": 4.5e-05,
|
115 |
+
"loss": 0.4761,
|
116 |
"step": 450
|
117 |
},
|
118 |
{
|
119 |
+
"epoch": 0.13,
|
120 |
+
"learning_rate": 4.75e-05,
|
121 |
+
"loss": 0.4884,
|
122 |
"step": 475
|
123 |
},
|
124 |
{
|
125 |
+
"epoch": 0.14,
|
126 |
+
"learning_rate": 5e-05,
|
127 |
+
"loss": 0.5453,
|
128 |
"step": 500
|
129 |
},
|
130 |
{
|
131 |
+
"epoch": 0.14,
|
132 |
+
"eval_cer": 14.794288634766497,
|
133 |
+
"eval_loss": 0.3284708261489868,
|
134 |
+
"eval_runtime": 1225.6472,
|
135 |
+
"eval_samples_per_second": 4.562,
|
136 |
+
"eval_steps_per_second": 0.286,
|
137 |
+
"step": 500
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 0.15,
|
141 |
+
"learning_rate": 4.991379310344828e-05,
|
142 |
+
"loss": 0.5147,
|
143 |
"step": 525
|
144 |
},
|
145 |
{
|
146 |
+
"epoch": 0.15,
|
147 |
+
"learning_rate": 4.982758620689655e-05,
|
148 |
+
"loss": 0.514,
|
149 |
"step": 550
|
150 |
},
|
151 |
{
|
152 |
+
"epoch": 0.16,
|
153 |
+
"learning_rate": 4.9741379310344836e-05,
|
154 |
+
"loss": 0.5397,
|
155 |
"step": 575
|
156 |
},
|
157 |
{
|
158 |
+
"epoch": 0.17,
|
159 |
+
"learning_rate": 4.9655172413793107e-05,
|
160 |
+
"loss": 0.5269,
|
161 |
"step": 600
|
162 |
},
|
163 |
{
|
164 |
+
"epoch": 0.17,
|
165 |
+
"learning_rate": 4.9568965517241384e-05,
|
166 |
+
"loss": 0.4992,
|
167 |
"step": 625
|
168 |
},
|
169 |
{
|
170 |
+
"epoch": 0.18,
|
171 |
+
"learning_rate": 4.9482758620689655e-05,
|
172 |
+
"loss": 0.538,
|
173 |
"step": 650
|
174 |
},
|
175 |
{
|
176 |
+
"epoch": 0.19,
|
177 |
+
"learning_rate": 4.939655172413793e-05,
|
178 |
+
"loss": 0.5123,
|
179 |
"step": 675
|
180 |
},
|
181 |
{
|
182 |
+
"epoch": 0.19,
|
183 |
+
"learning_rate": 4.931034482758621e-05,
|
184 |
+
"loss": 0.5051,
|
185 |
"step": 700
|
186 |
},
|
187 |
{
|
188 |
+
"epoch": 0.2,
|
189 |
+
"learning_rate": 4.922413793103449e-05,
|
190 |
+
"loss": 0.5342,
|
191 |
"step": 725
|
192 |
},
|
193 |
{
|
194 |
+
"epoch": 0.21,
|
195 |
+
"learning_rate": 4.913793103448276e-05,
|
196 |
+
"loss": 0.5037,
|
197 |
"step": 750
|
198 |
},
|
199 |
{
|
200 |
+
"epoch": 0.21,
|
201 |
+
"learning_rate": 4.905172413793104e-05,
|
202 |
+
"loss": 0.5492,
|
203 |
"step": 775
|
204 |
},
|
205 |
{
|
206 |
+
"epoch": 0.22,
|
207 |
+
"learning_rate": 4.896551724137931e-05,
|
208 |
+
"loss": 0.5256,
|
209 |
"step": 800
|
210 |
},
|
211 |
{
|
212 |
+
"epoch": 0.23,
|
213 |
+
"learning_rate": 4.8879310344827586e-05,
|
214 |
+
"loss": 0.531,
|
215 |
"step": 825
|
216 |
},
|
217 |
{
|
218 |
+
"epoch": 0.24,
|
219 |
+
"learning_rate": 4.8793103448275864e-05,
|
220 |
+
"loss": 0.5245,
|
221 |
"step": 850
|
222 |
},
|
223 |
{
|
224 |
+
"epoch": 0.24,
|
225 |
+
"learning_rate": 4.870689655172414e-05,
|
226 |
+
"loss": 0.4702,
|
227 |
"step": 875
|
228 |
},
|
229 |
{
|
230 |
+
"epoch": 0.25,
|
231 |
+
"learning_rate": 4.862068965517241e-05,
|
232 |
+
"loss": 0.5404,
|
233 |
"step": 900
|
234 |
},
|
235 |
{
|
236 |
+
"epoch": 0.26,
|
237 |
+
"learning_rate": 4.853793103448276e-05,
|
238 |
+
"loss": 0.4968,
|
239 |
"step": 925
|
240 |
},
|
241 |
{
|
242 |
+
"epoch": 0.26,
|
243 |
+
"learning_rate": 4.8451724137931036e-05,
|
244 |
+
"loss": 0.4905,
|
245 |
"step": 950
|
246 |
},
|
247 |
{
|
248 |
+
"epoch": 0.27,
|
249 |
+
"learning_rate": 4.836551724137931e-05,
|
250 |
+
"loss": 0.5281,
|
251 |
"step": 975
|
252 |
},
|
253 |
{
|
254 |
+
"epoch": 0.28,
|
255 |
+
"learning_rate": 4.827931034482759e-05,
|
256 |
+
"loss": 0.5144,
|
257 |
"step": 1000
|
258 |
},
|
259 |
{
|
260 |
+
"epoch": 0.28,
|
261 |
+
"eval_cer": 14.356672842266658,
|
262 |
+
"eval_loss": 0.3405072093009949,
|
263 |
+
"eval_runtime": 1158.9803,
|
264 |
+
"eval_samples_per_second": 4.824,
|
265 |
+
"eval_steps_per_second": 0.302,
|
266 |
"step": 1000
|
267 |
},
|
268 |
{
|
269 |
+
"epoch": 0.28,
|
270 |
+
"learning_rate": 4.819310344827587e-05,
|
271 |
+
"loss": 0.5357,
|
272 |
"step": 1025
|
273 |
},
|
274 |
{
|
275 |
+
"epoch": 0.29,
|
276 |
+
"learning_rate": 4.810689655172414e-05,
|
277 |
+
"loss": 0.5521,
|
278 |
"step": 1050
|
279 |
},
|
280 |
{
|
281 |
+
"epoch": 0.3,
|
282 |
+
"learning_rate": 4.802068965517242e-05,
|
283 |
+
"loss": 0.5413,
|
284 |
"step": 1075
|
285 |
},
|
286 |
{
|
287 |
+
"epoch": 0.3,
|
288 |
+
"learning_rate": 4.793448275862069e-05,
|
289 |
+
"loss": 0.4707,
|
290 |
"step": 1100
|
291 |
},
|
292 |
{
|
293 |
+
"epoch": 0.31,
|
294 |
+
"learning_rate": 4.784827586206897e-05,
|
295 |
+
"loss": 0.4771,
|
296 |
"step": 1125
|
297 |
},
|
298 |
{
|
299 |
+
"epoch": 0.32,
|
300 |
+
"learning_rate": 4.7762068965517245e-05,
|
301 |
+
"loss": 0.5055,
|
302 |
"step": 1150
|
303 |
},
|
304 |
{
|
305 |
+
"epoch": 0.33,
|
306 |
+
"learning_rate": 4.767586206896552e-05,
|
307 |
+
"loss": 0.5464,
|
308 |
"step": 1175
|
309 |
},
|
310 |
{
|
311 |
+
"epoch": 0.33,
|
312 |
+
"learning_rate": 4.758965517241379e-05,
|
313 |
+
"loss": 0.4839,
|
314 |
"step": 1200
|
315 |
},
|
316 |
{
|
317 |
+
"epoch": 0.34,
|
318 |
+
"learning_rate": 4.750344827586207e-05,
|
319 |
+
"loss": 0.5087,
|
320 |
"step": 1225
|
321 |
},
|
322 |
{
|
323 |
+
"epoch": 0.35,
|
324 |
+
"learning_rate": 4.741724137931035e-05,
|
325 |
+
"loss": 0.4728,
|
326 |
"step": 1250
|
327 |
},
|
328 |
{
|
329 |
+
"epoch": 0.35,
|
330 |
+
"learning_rate": 4.7331034482758626e-05,
|
331 |
+
"loss": 0.4457,
|
332 |
"step": 1275
|
333 |
},
|
334 |
{
|
335 |
+
"epoch": 0.36,
|
336 |
+
"learning_rate": 4.72448275862069e-05,
|
337 |
+
"loss": 0.5246,
|
338 |
"step": 1300
|
339 |
},
|
340 |
{
|
341 |
+
"epoch": 0.37,
|
342 |
+
"learning_rate": 4.7158620689655175e-05,
|
343 |
+
"loss": 0.4953,
|
344 |
"step": 1325
|
345 |
},
|
346 |
{
|
347 |
+
"epoch": 0.37,
|
348 |
+
"learning_rate": 4.7072413793103446e-05,
|
349 |
+
"loss": 0.5318,
|
350 |
"step": 1350
|
351 |
},
|
352 |
{
|
353 |
+
"epoch": 0.38,
|
354 |
+
"learning_rate": 4.6986206896551724e-05,
|
355 |
+
"loss": 0.5062,
|
356 |
"step": 1375
|
357 |
},
|
358 |
{
|
359 |
+
"epoch": 0.39,
|
360 |
+
"learning_rate": 4.69e-05,
|
361 |
+
"loss": 0.4979,
|
362 |
"step": 1400
|
363 |
},
|
364 |
{
|
365 |
+
"epoch": 0.39,
|
366 |
+
"learning_rate": 4.681379310344828e-05,
|
367 |
+
"loss": 0.5202,
|
368 |
"step": 1425
|
369 |
},
|
370 |
{
|
371 |
+
"epoch": 0.4,
|
372 |
+
"learning_rate": 4.672758620689656e-05,
|
373 |
+
"loss": 0.5267,
|
374 |
"step": 1450
|
375 |
},
|
376 |
{
|
377 |
+
"epoch": 0.41,
|
378 |
+
"learning_rate": 4.664137931034483e-05,
|
379 |
+
"loss": 0.4971,
|
380 |
"step": 1475
|
381 |
},
|
382 |
{
|
383 |
+
"epoch": 0.42,
|
384 |
+
"learning_rate": 4.6555172413793106e-05,
|
385 |
+
"loss": 0.49,
|
386 |
"step": 1500
|
387 |
},
|
388 |
{
|
389 |
+
"epoch": 0.42,
|
390 |
+
"eval_cer": 14.058008049575161,
|
391 |
+
"eval_loss": 0.3323359787464142,
|
392 |
+
"eval_runtime": 1178.6215,
|
393 |
+
"eval_samples_per_second": 4.744,
|
394 |
+
"eval_steps_per_second": 0.297,
|
395 |
+
"step": 1500
|
396 |
+
},
|
397 |
+
{
|
398 |
+
"epoch": 0.42,
|
399 |
+
"learning_rate": 4.6468965517241384e-05,
|
400 |
+
"loss": 0.4821,
|
401 |
"step": 1525
|
402 |
},
|
403 |
{
|
404 |
+
"epoch": 0.43,
|
405 |
+
"learning_rate": 4.638275862068966e-05,
|
406 |
+
"loss": 0.4603,
|
407 |
"step": 1550
|
408 |
},
|
409 |
{
|
410 |
+
"epoch": 0.44,
|
411 |
+
"learning_rate": 4.629655172413793e-05,
|
412 |
+
"loss": 0.5213,
|
413 |
"step": 1575
|
414 |
},
|
415 |
{
|
416 |
+
"epoch": 0.44,
|
417 |
+
"learning_rate": 4.621034482758621e-05,
|
418 |
+
"loss": 0.4922,
|
419 |
"step": 1600
|
420 |
},
|
421 |
{
|
422 |
+
"epoch": 0.45,
|
423 |
+
"learning_rate": 4.612413793103448e-05,
|
424 |
+
"loss": 0.5202,
|
425 |
"step": 1625
|
426 |
},
|
427 |
{
|
428 |
+
"epoch": 0.46,
|
429 |
+
"learning_rate": 4.603793103448276e-05,
|
430 |
+
"loss": 0.5046,
|
431 |
"step": 1650
|
432 |
},
|
433 |
{
|
434 |
+
"epoch": 0.46,
|
435 |
+
"learning_rate": 4.5951724137931036e-05,
|
436 |
+
"loss": 0.4795,
|
437 |
"step": 1675
|
438 |
},
|
439 |
{
|
440 |
+
"epoch": 0.47,
|
441 |
+
"learning_rate": 4.5865517241379314e-05,
|
442 |
+
"loss": 0.447,
|
443 |
"step": 1700
|
444 |
},
|
445 |
{
|
446 |
+
"epoch": 0.48,
|
447 |
+
"learning_rate": 4.5779310344827585e-05,
|
448 |
+
"loss": 0.4752,
|
449 |
"step": 1725
|
450 |
},
|
451 |
{
|
452 |
+
"epoch": 0.48,
|
453 |
+
"learning_rate": 4.569310344827586e-05,
|
454 |
+
"loss": 0.515,
|
455 |
"step": 1750
|
456 |
},
|
457 |
{
|
458 |
+
"epoch": 0.49,
|
459 |
+
"learning_rate": 4.560689655172414e-05,
|
460 |
+
"loss": 0.4882,
|
461 |
"step": 1775
|
462 |
},
|
463 |
{
|
464 |
+
"epoch": 0.5,
|
465 |
+
"learning_rate": 4.552068965517242e-05,
|
466 |
+
"loss": 0.4721,
|
467 |
"step": 1800
|
468 |
},
|
469 |
{
|
470 |
+
"epoch": 0.51,
|
471 |
+
"learning_rate": 4.5434482758620696e-05,
|
472 |
+
"loss": 0.4542,
|
473 |
"step": 1825
|
474 |
},
|
475 |
{
|
476 |
+
"epoch": 0.51,
|
477 |
+
"learning_rate": 4.534827586206897e-05,
|
478 |
+
"loss": 0.5197,
|
479 |
"step": 1850
|
480 |
},
|
481 |
{
|
482 |
+
"epoch": 0.52,
|
483 |
+
"learning_rate": 4.5262068965517245e-05,
|
484 |
+
"loss": 0.4589,
|
485 |
"step": 1875
|
486 |
},
|
487 |
{
|
488 |
+
"epoch": 0.53,
|
489 |
+
"learning_rate": 4.5175862068965516e-05,
|
490 |
+
"loss": 0.4668,
|
491 |
"step": 1900
|
492 |
},
|
493 |
{
|
494 |
+
"epoch": 0.53,
|
495 |
+
"learning_rate": 4.50896551724138e-05,
|
496 |
+
"loss": 0.4824,
|
497 |
"step": 1925
|
498 |
},
|
499 |
{
|
500 |
+
"epoch": 0.54,
|
501 |
+
"learning_rate": 4.500344827586207e-05,
|
502 |
+
"loss": 0.4967,
|
503 |
"step": 1950
|
504 |
},
|
505 |
{
|
506 |
+
"epoch": 0.55,
|
507 |
+
"learning_rate": 4.491724137931035e-05,
|
508 |
+
"loss": 0.4913,
|
509 |
"step": 1975
|
510 |
},
|
511 |
{
|
512 |
+
"epoch": 0.55,
|
513 |
+
"learning_rate": 4.483103448275862e-05,
|
514 |
+
"loss": 0.461,
|
515 |
"step": 2000
|
516 |
},
|
517 |
{
|
518 |
+
"epoch": 0.55,
|
519 |
+
"eval_cer": 13.056602568197789,
|
520 |
+
"eval_loss": 0.31062883138656616,
|
521 |
+
"eval_runtime": 1163.7705,
|
522 |
+
"eval_samples_per_second": 4.804,
|
523 |
+
"eval_steps_per_second": 0.301,
|
524 |
"step": 2000
|
525 |
},
|
526 |
{
|
527 |
+
"epoch": 0.56,
|
528 |
+
"learning_rate": 4.47448275862069e-05,
|
529 |
+
"loss": 0.4987,
|
530 |
"step": 2025
|
531 |
},
|
532 |
{
|
533 |
+
"epoch": 0.57,
|
534 |
+
"learning_rate": 4.4658620689655175e-05,
|
535 |
+
"loss": 0.5091,
|
536 |
"step": 2050
|
537 |
},
|
538 |
{
|
539 |
+
"epoch": 0.57,
|
540 |
+
"learning_rate": 4.457241379310345e-05,
|
541 |
+
"loss": 0.5096,
|
542 |
"step": 2075
|
543 |
},
|
544 |
{
|
545 |
+
"epoch": 0.58,
|
546 |
+
"learning_rate": 4.4486206896551724e-05,
|
547 |
+
"loss": 0.4716,
|
548 |
"step": 2100
|
549 |
},
|
550 |
{
|
551 |
+
"epoch": 0.59,
|
552 |
+
"learning_rate": 4.44e-05,
|
553 |
+
"loss": 0.4618,
|
554 |
"step": 2125
|
555 |
},
|
556 |
{
|
557 |
+
"epoch": 0.6,
|
558 |
+
"learning_rate": 4.431379310344827e-05,
|
559 |
+
"loss": 0.4588,
|
560 |
"step": 2150
|
561 |
},
|
562 |
{
|
563 |
+
"epoch": 0.6,
|
564 |
+
"learning_rate": 4.422758620689656e-05,
|
565 |
+
"loss": 0.5037,
|
566 |
"step": 2175
|
567 |
},
|
568 |
{
|
569 |
+
"epoch": 0.61,
|
570 |
+
"learning_rate": 4.414137931034483e-05,
|
571 |
+
"loss": 0.4818,
|
572 |
"step": 2200
|
573 |
},
|
574 |
{
|
575 |
+
"epoch": 0.62,
|
576 |
+
"learning_rate": 4.4055172413793106e-05,
|
577 |
+
"loss": 0.4802,
|
578 |
"step": 2225
|
579 |
},
|
580 |
{
|
581 |
+
"epoch": 0.62,
|
582 |
+
"learning_rate": 4.3968965517241384e-05,
|
583 |
+
"loss": 0.4889,
|
584 |
"step": 2250
|
585 |
},
|
586 |
{
|
587 |
+
"epoch": 0.63,
|
588 |
+
"learning_rate": 4.3882758620689655e-05,
|
589 |
+
"loss": 0.4914,
|
590 |
"step": 2275
|
591 |
},
|
592 |
{
|
593 |
+
"epoch": 0.64,
|
594 |
+
"learning_rate": 4.379655172413793e-05,
|
595 |
+
"loss": 0.4867,
|
596 |
"step": 2300
|
597 |
},
|
598 |
{
|
599 |
+
"epoch": 0.64,
|
600 |
+
"learning_rate": 4.371034482758621e-05,
|
601 |
+
"loss": 0.5006,
|
602 |
"step": 2325
|
603 |
},
|
604 |
{
|
605 |
+
"epoch": 0.65,
|
606 |
+
"learning_rate": 4.362413793103449e-05,
|
607 |
+
"loss": 0.4705,
|
608 |
"step": 2350
|
609 |
},
|
610 |
{
|
611 |
+
"epoch": 0.66,
|
612 |
+
"learning_rate": 4.353793103448276e-05,
|
613 |
+
"loss": 0.482,
|
614 |
"step": 2375
|
615 |
},
|
616 |
{
|
617 |
+
"epoch": 0.66,
|
618 |
+
"learning_rate": 4.3451724137931037e-05,
|
619 |
+
"loss": 0.4818,
|
620 |
"step": 2400
|
621 |
},
|
622 |
{
|
623 |
+
"epoch": 0.67,
|
624 |
+
"learning_rate": 4.336551724137931e-05,
|
625 |
+
"loss": 0.4478,
|
626 |
"step": 2425
|
627 |
},
|
628 |
{
|
629 |
+
"epoch": 0.68,
|
630 |
+
"learning_rate": 4.327931034482759e-05,
|
631 |
+
"loss": 0.4759,
|
632 |
"step": 2450
|
633 |
},
|
634 |
{
|
635 |
+
"epoch": 0.69,
|
636 |
+
"learning_rate": 4.319310344827586e-05,
|
637 |
+
"loss": 0.4518,
|
638 |
"step": 2475
|
639 |
},
|
640 |
{
|
641 |
+
"epoch": 0.69,
|
642 |
+
"learning_rate": 4.310689655172414e-05,
|
643 |
+
"loss": 0.4935,
|
644 |
"step": 2500
|
645 |
},
|
646 |
{
|
647 |
+
"epoch": 0.69,
|
648 |
+
"eval_cer": 13.829617325752253,
|
649 |
+
"eval_loss": 0.3001407980918884,
|
650 |
+
"eval_runtime": 1213.7136,
|
651 |
+
"eval_samples_per_second": 4.607,
|
652 |
+
"eval_steps_per_second": 0.288,
|
653 |
+
"step": 2500
|
654 |
+
},
|
655 |
+
{
|
656 |
+
"epoch": 0.7,
|
657 |
+
"learning_rate": 4.302068965517241e-05,
|
658 |
+
"loss": 0.4943,
|
659 |
"step": 2525
|
660 |
},
|
661 |
{
|
662 |
+
"epoch": 0.71,
|
663 |
+
"learning_rate": 4.293448275862069e-05,
|
664 |
+
"loss": 0.4451,
|
665 |
"step": 2550
|
666 |
},
|
667 |
{
|
668 |
+
"epoch": 0.71,
|
669 |
+
"learning_rate": 4.284827586206897e-05,
|
670 |
+
"loss": 0.456,
|
671 |
"step": 2575
|
672 |
},
|
673 |
{
|
674 |
+
"epoch": 0.72,
|
675 |
+
"learning_rate": 4.2762068965517245e-05,
|
676 |
+
"loss": 0.4944,
|
677 |
"step": 2600
|
678 |
},
|
679 |
{
|
680 |
+
"epoch": 0.73,
|
681 |
+
"learning_rate": 4.267586206896552e-05,
|
682 |
+
"loss": 0.4523,
|
683 |
"step": 2625
|
684 |
},
|
685 |
{
|
686 |
+
"epoch": 0.73,
|
687 |
+
"learning_rate": 4.2589655172413794e-05,
|
688 |
+
"loss": 0.4865,
|
689 |
"step": 2650
|
690 |
},
|
691 |
{
|
692 |
+
"epoch": 0.74,
|
693 |
+
"learning_rate": 4.250344827586207e-05,
|
694 |
+
"loss": 0.4429,
|
695 |
"step": 2675
|
696 |
},
|
697 |
{
|
698 |
+
"epoch": 0.75,
|
699 |
+
"learning_rate": 4.241724137931035e-05,
|
700 |
+
"loss": 0.4705,
|
701 |
"step": 2700
|
702 |
},
|
703 |
{
|
704 |
+
"epoch": 0.75,
|
705 |
+
"learning_rate": 4.233103448275863e-05,
|
706 |
+
"loss": 0.437,
|
707 |
"step": 2725
|
708 |
},
|
709 |
{
|
710 |
+
"epoch": 0.76,
|
711 |
+
"learning_rate": 4.22448275862069e-05,
|
712 |
+
"loss": 0.4218,
|
713 |
"step": 2750
|
714 |
},
|
715 |
{
|
716 |
+
"epoch": 0.77,
|
717 |
+
"learning_rate": 4.2158620689655176e-05,
|
718 |
+
"loss": 0.5038,
|
719 |
"step": 2775
|
720 |
},
|
721 |
{
|
722 |
+
"epoch": 0.78,
|
723 |
+
"learning_rate": 4.2072413793103447e-05,
|
724 |
+
"loss": 0.4579,
|
725 |
"step": 2800
|
726 |
},
|
727 |
{
|
728 |
+
"epoch": 0.78,
|
729 |
+
"learning_rate": 4.1986206896551724e-05,
|
730 |
+
"loss": 0.4738,
|
731 |
"step": 2825
|
732 |
},
|
733 |
{
|
734 |
+
"epoch": 0.79,
|
735 |
+
"learning_rate": 4.19e-05,
|
736 |
+
"loss": 0.4642,
|
737 |
"step": 2850
|
738 |
},
|
739 |
{
|
740 |
+
"epoch": 0.8,
|
741 |
+
"learning_rate": 4.181379310344828e-05,
|
742 |
+
"loss": 0.4422,
|
743 |
"step": 2875
|
744 |
},
|
745 |
{
|
746 |
+
"epoch": 0.8,
|
747 |
+
"learning_rate": 4.172758620689655e-05,
|
748 |
+
"loss": 0.4212,
|
749 |
"step": 2900
|
750 |
},
|
751 |
{
|
752 |
+
"epoch": 0.81,
|
753 |
+
"learning_rate": 4.164137931034483e-05,
|
754 |
+
"loss": 0.4803,
|
755 |
"step": 2925
|
756 |
},
|
757 |
{
|
758 |
+
"epoch": 0.82,
|
759 |
+
"learning_rate": 4.1555172413793106e-05,
|
760 |
+
"loss": 0.4333,
|
761 |
"step": 2950
|
762 |
},
|
763 |
{
|
764 |
+
"epoch": 0.82,
|
765 |
+
"learning_rate": 4.1468965517241384e-05,
|
766 |
+
"loss": 0.4663,
|
767 |
"step": 2975
|
768 |
},
|
769 |
{
|
770 |
+
"epoch": 0.83,
|
771 |
+
"learning_rate": 4.1382758620689655e-05,
|
772 |
+
"loss": 0.4836,
|
773 |
"step": 3000
|
774 |
},
|
775 |
{
|
776 |
+
"epoch": 0.83,
|
777 |
+
"eval_cer": 13.110905257778061,
|
778 |
+
"eval_loss": 0.29782944917678833,
|
779 |
+
"eval_runtime": 1227.0769,
|
780 |
+
"eval_samples_per_second": 4.556,
|
781 |
+
"eval_steps_per_second": 0.285,
|
782 |
"step": 3000
|
783 |
},
|
784 |
{
|
785 |
+
"epoch": 0.84,
|
786 |
+
"learning_rate": 4.129655172413793e-05,
|
787 |
+
"loss": 0.482,
|
788 |
"step": 3025
|
789 |
},
|
790 |
{
|
791 |
+
"epoch": 0.84,
|
792 |
+
"learning_rate": 4.121034482758621e-05,
|
793 |
+
"loss": 0.4706,
|
794 |
"step": 3050
|
795 |
},
|
796 |
{
|
797 |
+
"epoch": 0.85,
|
798 |
+
"learning_rate": 4.112413793103448e-05,
|
799 |
+
"loss": 0.4658,
|
800 |
"step": 3075
|
801 |
},
|
802 |
{
|
803 |
+
"epoch": 0.86,
|
804 |
+
"learning_rate": 4.1037931034482766e-05,
|
805 |
+
"loss": 0.4624,
|
806 |
"step": 3100
|
807 |
},
|
808 |
{
|
809 |
+
"epoch": 0.87,
|
810 |
+
"learning_rate": 4.095172413793104e-05,
|
811 |
+
"loss": 0.4754,
|
812 |
"step": 3125
|
813 |
},
|
814 |
{
|
815 |
+
"epoch": 0.87,
|
816 |
+
"learning_rate": 4.0865517241379315e-05,
|
817 |
+
"loss": 0.4866,
|
818 |
"step": 3150
|
819 |
},
|
820 |
{
|
821 |
+
"epoch": 0.88,
|
822 |
+
"learning_rate": 4.0779310344827586e-05,
|
823 |
+
"loss": 0.4723,
|
824 |
"step": 3175
|
825 |
},
|
826 |
{
|
827 |
+
"epoch": 0.89,
|
828 |
+
"learning_rate": 4.069310344827586e-05,
|
829 |
+
"loss": 0.4707,
|
830 |
"step": 3200
|
831 |
},
|
832 |
{
|
833 |
+
"epoch": 0.89,
|
834 |
+
"learning_rate": 4.060689655172414e-05,
|
835 |
+
"loss": 0.3989,
|
836 |
"step": 3225
|
837 |
},
|
838 |
{
|
839 |
+
"epoch": 0.9,
|
840 |
+
"learning_rate": 4.052068965517242e-05,
|
841 |
+
"loss": 0.4714,
|
842 |
"step": 3250
|
843 |
},
|
844 |
{
|
845 |
+
"epoch": 0.91,
|
846 |
+
"learning_rate": 4.043448275862069e-05,
|
847 |
+
"loss": 0.4451,
|
848 |
"step": 3275
|
849 |
},
|
850 |
{
|
851 |
+
"epoch": 0.91,
|
852 |
+
"learning_rate": 4.034827586206897e-05,
|
853 |
+
"loss": 0.4785,
|
854 |
"step": 3300
|
855 |
},
|
856 |
{
|
857 |
+
"epoch": 0.92,
|
858 |
+
"learning_rate": 4.026206896551724e-05,
|
859 |
+
"loss": 0.4509,
|
860 |
"step": 3325
|
861 |
},
|
862 |
{
|
863 |
+
"epoch": 0.93,
|
864 |
+
"learning_rate": 4.017586206896552e-05,
|
865 |
+
"loss": 0.4523,
|
866 |
"step": 3350
|
867 |
},
|
868 |
{
|
869 |
+
"epoch": 0.94,
|
870 |
+
"learning_rate": 4.0089655172413794e-05,
|
871 |
+
"loss": 0.4264,
|
872 |
"step": 3375
|
873 |
},
|
874 |
{
|
875 |
+
"epoch": 0.94,
|
876 |
+
"learning_rate": 4.000344827586207e-05,
|
877 |
+
"loss": 0.4685,
|
878 |
"step": 3400
|
879 |
},
|
880 |
{
|
881 |
+
"epoch": 0.95,
|
882 |
+
"learning_rate": 3.991724137931035e-05,
|
883 |
+
"loss": 0.4632,
|
884 |
"step": 3425
|
885 |
},
|
886 |
{
|
887 |
+
"epoch": 0.96,
|
888 |
+
"learning_rate": 3.983103448275862e-05,
|
889 |
+
"loss": 0.4548,
|
890 |
"step": 3450
|
891 |
},
|
892 |
{
|
893 |
+
"epoch": 0.96,
|
894 |
+
"learning_rate": 3.97448275862069e-05,
|
895 |
+
"loss": 0.4306,
|
896 |
"step": 3475
|
897 |
},
|
898 |
{
|
899 |
+
"epoch": 0.97,
|
900 |
+
"learning_rate": 3.9658620689655176e-05,
|
901 |
+
"loss": 0.4782,
|
902 |
"step": 3500
|
903 |
},
|
904 |
{
|
905 |
+
"epoch": 0.97,
|
906 |
+
"eval_cer": 12.711620775570179,
|
907 |
+
"eval_loss": 0.2923573851585388,
|
908 |
+
"eval_runtime": 1186.1965,
|
909 |
+
"eval_samples_per_second": 4.713,
|
910 |
+
"eval_steps_per_second": 0.295,
|
911 |
+
"step": 3500
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 0.98,
|
915 |
+
"learning_rate": 3.9572413793103454e-05,
|
916 |
+
"loss": 0.4219,
|
917 |
"step": 3525
|
918 |
},
|
919 |
{
|
920 |
+
"epoch": 0.98,
|
921 |
+
"learning_rate": 3.9486206896551725e-05,
|
922 |
+
"loss": 0.471,
|
923 |
"step": 3550
|
924 |
},
|
925 |
{
|
926 |
+
"epoch": 0.99,
|
927 |
+
"learning_rate": 3.94e-05,
|
928 |
+
"loss": 0.4687,
|
929 |
"step": 3575
|
930 |
},
|
931 |
{
|
932 |
+
"epoch": 1.0,
|
933 |
+
"learning_rate": 3.931379310344827e-05,
|
934 |
+
"loss": 0.4356,
|
935 |
"step": 3600
|
936 |
},
|
937 |
{
|
938 |
+
"epoch": 1.0,
|
939 |
+
"learning_rate": 3.922758620689656e-05,
|
940 |
+
"loss": 0.4144,
|
941 |
"step": 3625
|
942 |
},
|
943 |
{
|
944 |
+
"epoch": 1.01,
|
945 |
+
"learning_rate": 3.914137931034483e-05,
|
946 |
+
"loss": 0.3605,
|
947 |
"step": 3650
|
948 |
},
|
949 |
{
|
950 |
+
"epoch": 1.02,
|
951 |
+
"learning_rate": 3.9055172413793106e-05,
|
952 |
+
"loss": 0.3652,
|
953 |
"step": 3675
|
954 |
},
|
955 |
{
|
956 |
+
"epoch": 1.03,
|
957 |
+
"learning_rate": 3.896896551724138e-05,
|
958 |
+
"loss": 0.3725,
|
959 |
"step": 3700
|
960 |
},
|
961 |
{
|
962 |
+
"epoch": 1.03,
|
963 |
+
"learning_rate": 3.8882758620689655e-05,
|
964 |
+
"loss": 0.3481,
|
965 |
"step": 3725
|
966 |
},
|
967 |
{
|
968 |
+
"epoch": 1.04,
|
969 |
+
"learning_rate": 3.879655172413793e-05,
|
970 |
+
"loss": 0.3662,
|
971 |
"step": 3750
|
972 |
},
|
973 |
{
|
974 |
+
"epoch": 1.05,
|
975 |
+
"learning_rate": 3.871034482758621e-05,
|
976 |
+
"loss": 0.3835,
|
977 |
"step": 3775
|
978 |
},
|
979 |
{
|
980 |
+
"epoch": 1.05,
|
981 |
+
"learning_rate": 3.862413793103448e-05,
|
982 |
+
"loss": 0.3638,
|
983 |
"step": 3800
|
984 |
},
|
985 |
{
|
986 |
+
"epoch": 1.06,
|
987 |
+
"learning_rate": 3.853793103448276e-05,
|
988 |
+
"loss": 0.3615,
|
989 |
"step": 3825
|
990 |
},
|
991 |
{
|
992 |
+
"epoch": 1.07,
|
993 |
+
"learning_rate": 3.845172413793104e-05,
|
994 |
+
"loss": 0.3327,
|
995 |
"step": 3850
|
996 |
},
|
997 |
{
|
998 |
+
"epoch": 1.07,
|
999 |
+
"learning_rate": 3.8365517241379315e-05,
|
1000 |
+
"loss": 0.3679,
|
1001 |
"step": 3875
|
1002 |
},
|
1003 |
{
|
1004 |
+
"epoch": 1.08,
|
1005 |
+
"learning_rate": 3.827931034482759e-05,
|
1006 |
+
"loss": 0.3549,
|
1007 |
"step": 3900
|
1008 |
},
|
1009 |
{
|
1010 |
+
"epoch": 1.09,
|
1011 |
+
"learning_rate": 3.8193103448275863e-05,
|
1012 |
+
"loss": 0.3728,
|
1013 |
"step": 3925
|
1014 |
},
|
1015 |
{
|
1016 |
+
"epoch": 1.09,
|
1017 |
+
"learning_rate": 3.810689655172414e-05,
|
1018 |
+
"loss": 0.3493,
|
1019 |
"step": 3950
|
1020 |
},
|
1021 |
{
|
1022 |
+
"epoch": 1.1,
|
1023 |
+
"learning_rate": 3.802068965517241e-05,
|
1024 |
+
"loss": 0.3509,
|
1025 |
"step": 3975
|
1026 |
},
|
1027 |
{
|
1028 |
+
"epoch": 1.11,
|
1029 |
+
"learning_rate": 3.793448275862069e-05,
|
1030 |
+
"loss": 0.3441,
|
1031 |
"step": 4000
|
1032 |
},
|
1033 |
{
|
1034 |
+
"epoch": 1.11,
|
1035 |
+
"eval_cer": 11.794863604420877,
|
1036 |
+
"eval_loss": 0.2874628007411957,
|
1037 |
+
"eval_runtime": 1164.5329,
|
1038 |
+
"eval_samples_per_second": 4.801,
|
1039 |
+
"eval_steps_per_second": 0.301,
|
1040 |
"step": 4000
|
1041 |
},
|
1042 |
{
|
1043 |
+
"epoch": 1.12,
|
1044 |
+
"learning_rate": 3.784827586206897e-05,
|
1045 |
+
"loss": 0.3513,
|
1046 |
"step": 4025
|
1047 |
},
|
1048 |
{
|
1049 |
+
"epoch": 1.12,
|
1050 |
+
"learning_rate": 3.7762068965517245e-05,
|
1051 |
+
"loss": 0.3442,
|
1052 |
"step": 4050
|
1053 |
},
|
1054 |
{
|
1055 |
+
"epoch": 1.13,
|
1056 |
+
"learning_rate": 3.7675862068965516e-05,
|
1057 |
+
"loss": 0.3183,
|
1058 |
"step": 4075
|
1059 |
},
|
1060 |
{
|
1061 |
+
"epoch": 1.14,
|
1062 |
+
"learning_rate": 3.7589655172413794e-05,
|
1063 |
+
"loss": 0.3674,
|
1064 |
"step": 4100
|
1065 |
},
|
1066 |
{
|
1067 |
+
"epoch": 1.14,
|
1068 |
+
"learning_rate": 3.750344827586207e-05,
|
1069 |
+
"loss": 0.3491,
|
1070 |
"step": 4125
|
1071 |
},
|
1072 |
{
|
1073 |
+
"epoch": 1.15,
|
1074 |
+
"learning_rate": 3.741724137931035e-05,
|
1075 |
+
"loss": 0.337,
|
1076 |
"step": 4150
|
1077 |
},
|
1078 |
{
|
1079 |
+
"epoch": 1.16,
|
1080 |
+
"learning_rate": 3.733103448275862e-05,
|
1081 |
+
"loss": 0.3711,
|
1082 |
"step": 4175
|
1083 |
},
|
1084 |
{
|
1085 |
+
"epoch": 1.16,
|
1086 |
+
"learning_rate": 3.72448275862069e-05,
|
1087 |
+
"loss": 0.3839,
|
1088 |
"step": 4200
|
1089 |
},
|
1090 |
{
|
1091 |
+
"epoch": 1.17,
|
1092 |
+
"learning_rate": 3.7158620689655176e-05,
|
1093 |
+
"loss": 0.3724,
|
1094 |
"step": 4225
|
1095 |
},
|
1096 |
{
|
1097 |
+
"epoch": 1.18,
|
1098 |
+
"learning_rate": 3.707241379310345e-05,
|
1099 |
+
"loss": 0.3444,
|
1100 |
"step": 4250
|
1101 |
},
|
1102 |
{
|
1103 |
+
"epoch": 1.18,
|
1104 |
+
"learning_rate": 3.698620689655173e-05,
|
1105 |
+
"loss": 0.3997,
|
1106 |
"step": 4275
|
1107 |
},
|
1108 |
{
|
1109 |
+
"epoch": 1.19,
|
1110 |
+
"learning_rate": 3.69e-05,
|
1111 |
+
"loss": 0.3742,
|
1112 |
"step": 4300
|
1113 |
},
|
1114 |
{
|
1115 |
+
"epoch": 1.2,
|
1116 |
+
"learning_rate": 3.681379310344828e-05,
|
1117 |
+
"loss": 0.3657,
|
1118 |
"step": 4325
|
1119 |
},
|
1120 |
{
|
1121 |
+
"epoch": 1.21,
|
1122 |
+
"learning_rate": 3.672758620689655e-05,
|
1123 |
+
"loss": 0.3762,
|
1124 |
"step": 4350
|
1125 |
},
|
1126 |
{
|
1127 |
+
"epoch": 1.21,
|
1128 |
+
"learning_rate": 3.664137931034483e-05,
|
1129 |
+
"loss": 0.3815,
|
1130 |
"step": 4375
|
1131 |
},
|
1132 |
{
|
1133 |
+
"epoch": 1.22,
|
1134 |
+
"learning_rate": 3.655517241379311e-05,
|
1135 |
+
"loss": 0.3628,
|
1136 |
"step": 4400
|
1137 |
},
|
1138 |
{
|
1139 |
+
"epoch": 1.23,
|
1140 |
+
"learning_rate": 3.6468965517241384e-05,
|
1141 |
+
"loss": 0.3533,
|
1142 |
"step": 4425
|
1143 |
},
|
1144 |
{
|
1145 |
+
"epoch": 1.23,
|
1146 |
+
"learning_rate": 3.6382758620689655e-05,
|
1147 |
+
"loss": 0.3857,
|
1148 |
"step": 4450
|
1149 |
},
|
1150 |
{
|
1151 |
+
"epoch": 1.24,
|
1152 |
+
"learning_rate": 3.629655172413793e-05,
|
1153 |
+
"loss": 0.3381,
|
1154 |
"step": 4475
|
1155 |
},
|
1156 |
{
|
1157 |
+
"epoch": 1.25,
|
1158 |
+
"learning_rate": 3.6210344827586204e-05,
|
1159 |
+
"loss": 0.3647,
|
1160 |
"step": 4500
|
1161 |
},
|
1162 |
{
|
1163 |
+
"epoch": 1.25,
|
1164 |
+
"eval_cer": 11.617581294320578,
|
1165 |
+
"eval_loss": 0.28388434648513794,
|
1166 |
+
"eval_runtime": 1169.2298,
|
1167 |
+
"eval_samples_per_second": 4.782,
|
1168 |
+
"eval_steps_per_second": 0.299,
|
1169 |
+
"step": 4500
|
1170 |
+
},
|
1171 |
+
{
|
1172 |
+
"epoch": 1.25,
|
1173 |
+
"learning_rate": 3.612413793103449e-05,
|
1174 |
+
"loss": 0.3935,
|
1175 |
"step": 4525
|
1176 |
},
|
1177 |
{
|
1178 |
+
"epoch": 1.26,
|
1179 |
+
"learning_rate": 3.603793103448276e-05,
|
1180 |
+
"loss": 0.3514,
|
1181 |
"step": 4550
|
1182 |
},
|
1183 |
{
|
1184 |
+
"epoch": 1.27,
|
1185 |
+
"learning_rate": 3.595172413793104e-05,
|
1186 |
+
"loss": 0.3925,
|
1187 |
"step": 4575
|
1188 |
},
|
1189 |
{
|
1190 |
+
"epoch": 1.27,
|
1191 |
+
"learning_rate": 3.586551724137931e-05,
|
1192 |
+
"loss": 0.355,
|
1193 |
"step": 4600
|
1194 |
},
|
1195 |
{
|
1196 |
+
"epoch": 1.28,
|
1197 |
+
"learning_rate": 3.5779310344827586e-05,
|
1198 |
+
"loss": 0.3539,
|
1199 |
"step": 4625
|
1200 |
},
|
1201 |
{
|
1202 |
+
"epoch": 1.29,
|
1203 |
+
"learning_rate": 3.5693103448275864e-05,
|
1204 |
+
"loss": 0.3583,
|
1205 |
"step": 4650
|
1206 |
},
|
1207 |
{
|
1208 |
+
"epoch": 1.3,
|
1209 |
+
"learning_rate": 3.560689655172414e-05,
|
1210 |
+
"loss": 0.3312,
|
1211 |
"step": 4675
|
1212 |
},
|
1213 |
{
|
1214 |
+
"epoch": 1.3,
|
1215 |
+
"learning_rate": 3.552068965517242e-05,
|
1216 |
+
"loss": 0.331,
|
1217 |
"step": 4700
|
1218 |
},
|
1219 |
{
|
1220 |
+
"epoch": 1.31,
|
1221 |
+
"learning_rate": 3.543448275862069e-05,
|
1222 |
+
"loss": 0.3703,
|
1223 |
"step": 4725
|
1224 |
},
|
1225 |
{
|
1226 |
+
"epoch": 1.32,
|
1227 |
+
"learning_rate": 3.534827586206897e-05,
|
1228 |
+
"loss": 0.3599,
|
1229 |
"step": 4750
|
1230 |
},
|
1231 |
{
|
1232 |
+
"epoch": 1.32,
|
1233 |
+
"learning_rate": 3.526206896551724e-05,
|
1234 |
+
"loss": 0.3779,
|
1235 |
"step": 4775
|
1236 |
},
|
1237 |
{
|
1238 |
+
"epoch": 1.33,
|
1239 |
+
"learning_rate": 3.517586206896552e-05,
|
1240 |
+
"loss": 0.3338,
|
1241 |
"step": 4800
|
1242 |
},
|
1243 |
{
|
1244 |
+
"epoch": 1.34,
|
1245 |
+
"learning_rate": 3.5089655172413794e-05,
|
1246 |
+
"loss": 0.3249,
|
1247 |
"step": 4825
|
1248 |
},
|
1249 |
{
|
1250 |
+
"epoch": 1.34,
|
1251 |
+
"learning_rate": 3.500344827586207e-05,
|
1252 |
+
"loss": 0.364,
|
1253 |
"step": 4850
|
1254 |
},
|
1255 |
{
|
1256 |
+
"epoch": 1.35,
|
1257 |
+
"learning_rate": 3.491724137931034e-05,
|
1258 |
+
"loss": 0.3755,
|
1259 |
"step": 4875
|
1260 |
},
|
1261 |
{
|
1262 |
+
"epoch": 1.36,
|
1263 |
+
"learning_rate": 3.483103448275862e-05,
|
1264 |
+
"loss": 0.353,
|
1265 |
"step": 4900
|
1266 |
},
|
1267 |
{
|
1268 |
+
"epoch": 1.36,
|
1269 |
+
"learning_rate": 3.47448275862069e-05,
|
1270 |
+
"loss": 0.3483,
|
1271 |
"step": 4925
|
1272 |
},
|
1273 |
{
|
1274 |
+
"epoch": 1.37,
|
1275 |
+
"learning_rate": 3.4658620689655176e-05,
|
1276 |
+
"loss": 0.3675,
|
1277 |
"step": 4950
|
1278 |
},
|
1279 |
{
|
1280 |
+
"epoch": 1.38,
|
1281 |
+
"learning_rate": 3.457241379310345e-05,
|
1282 |
+
"loss": 0.3779,
|
1283 |
"step": 4975
|
1284 |
},
|
1285 |
{
|
1286 |
+
"epoch": 1.39,
|
1287 |
+
"learning_rate": 3.4486206896551725e-05,
|
1288 |
+
"loss": 0.3642,
|
1289 |
"step": 5000
|
1290 |
},
|
1291 |
{
|
1292 |
+
"epoch": 1.39,
|
1293 |
+
"eval_cer": 11.572861432313294,
|
1294 |
+
"eval_loss": 0.28437113761901855,
|
1295 |
+
"eval_runtime": 1173.2699,
|
1296 |
+
"eval_samples_per_second": 4.765,
|
1297 |
+
"eval_steps_per_second": 0.298,
|
1298 |
"step": 5000
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 1.39,
|
1302 |
+
"learning_rate": 3.4399999999999996e-05,
|
1303 |
+
"loss": 0.3872,
|
1304 |
+
"step": 5025
|
1305 |
+
},
|
1306 |
+
{
|
1307 |
+
"epoch": 1.4,
|
1308 |
+
"learning_rate": 3.431379310344828e-05,
|
1309 |
+
"loss": 0.392,
|
1310 |
+
"step": 5050
|
1311 |
+
},
|
1312 |
+
{
|
1313 |
+
"epoch": 1.41,
|
1314 |
+
"learning_rate": 3.422758620689656e-05,
|
1315 |
+
"loss": 0.3545,
|
1316 |
+
"step": 5075
|
1317 |
+
},
|
1318 |
+
{
|
1319 |
+
"epoch": 1.41,
|
1320 |
+
"learning_rate": 3.41448275862069e-05,
|
1321 |
+
"loss": 0.3627,
|
1322 |
+
"step": 5100
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 1.42,
|
1326 |
+
"learning_rate": 3.4058620689655175e-05,
|
1327 |
+
"loss": 0.3362,
|
1328 |
+
"step": 5125
|
1329 |
+
},
|
1330 |
+
{
|
1331 |
+
"epoch": 1.43,
|
1332 |
+
"learning_rate": 3.397241379310345e-05,
|
1333 |
+
"loss": 0.3628,
|
1334 |
+
"step": 5150
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 1.43,
|
1338 |
+
"learning_rate": 3.3886206896551724e-05,
|
1339 |
+
"loss": 0.3157,
|
1340 |
+
"step": 5175
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 1.44,
|
1344 |
+
"learning_rate": 3.38e-05,
|
1345 |
+
"loss": 0.3649,
|
1346 |
+
"step": 5200
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 1.45,
|
1350 |
+
"learning_rate": 3.371379310344828e-05,
|
1351 |
+
"loss": 0.3462,
|
1352 |
+
"step": 5225
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 1.45,
|
1356 |
+
"learning_rate": 3.362758620689656e-05,
|
1357 |
+
"loss": 0.3833,
|
1358 |
+
"step": 5250
|
1359 |
+
},
|
1360 |
+
{
|
1361 |
+
"epoch": 1.46,
|
1362 |
+
"learning_rate": 3.354137931034483e-05,
|
1363 |
+
"loss": 0.3643,
|
1364 |
+
"step": 5275
|
1365 |
+
},
|
1366 |
+
{
|
1367 |
+
"epoch": 1.47,
|
1368 |
+
"learning_rate": 3.3455172413793106e-05,
|
1369 |
+
"loss": 0.3788,
|
1370 |
+
"step": 5300
|
1371 |
+
},
|
1372 |
+
{
|
1373 |
+
"epoch": 1.48,
|
1374 |
+
"learning_rate": 3.336896551724138e-05,
|
1375 |
+
"loss": 0.3625,
|
1376 |
+
"step": 5325
|
1377 |
+
},
|
1378 |
+
{
|
1379 |
+
"epoch": 1.48,
|
1380 |
+
"learning_rate": 3.328275862068966e-05,
|
1381 |
+
"loss": 0.363,
|
1382 |
+
"step": 5350
|
1383 |
+
},
|
1384 |
+
{
|
1385 |
+
"epoch": 1.49,
|
1386 |
+
"learning_rate": 3.319655172413793e-05,
|
1387 |
+
"loss": 0.3714,
|
1388 |
+
"step": 5375
|
1389 |
+
},
|
1390 |
+
{
|
1391 |
+
"epoch": 1.5,
|
1392 |
+
"learning_rate": 3.311034482758621e-05,
|
1393 |
+
"loss": 0.3636,
|
1394 |
+
"step": 5400
|
1395 |
+
},
|
1396 |
+
{
|
1397 |
+
"epoch": 1.5,
|
1398 |
+
"learning_rate": 3.302413793103448e-05,
|
1399 |
+
"loss": 0.3349,
|
1400 |
+
"step": 5425
|
1401 |
+
},
|
1402 |
+
{
|
1403 |
+
"epoch": 1.51,
|
1404 |
+
"learning_rate": 3.293793103448276e-05,
|
1405 |
+
"loss": 0.3516,
|
1406 |
+
"step": 5450
|
1407 |
+
},
|
1408 |
+
{
|
1409 |
+
"epoch": 1.52,
|
1410 |
+
"learning_rate": 3.2851724137931036e-05,
|
1411 |
+
"loss": 0.3327,
|
1412 |
+
"step": 5475
|
1413 |
+
},
|
1414 |
+
{
|
1415 |
+
"epoch": 1.52,
|
1416 |
+
"learning_rate": 3.2765517241379314e-05,
|
1417 |
+
"loss": 0.3493,
|
1418 |
+
"step": 5500
|
1419 |
+
},
|
1420 |
+
{
|
1421 |
+
"epoch": 1.52,
|
1422 |
+
"eval_cer": 11.528141570306012,
|
1423 |
+
"eval_loss": 0.2879265248775482,
|
1424 |
+
"eval_runtime": 1178.4372,
|
1425 |
+
"eval_samples_per_second": 4.744,
|
1426 |
+
"eval_steps_per_second": 0.297,
|
1427 |
+
"step": 5500
|
1428 |
+
},
|
1429 |
+
{
|
1430 |
+
"epoch": 1.53,
|
1431 |
+
"learning_rate": 3.2679310344827585e-05,
|
1432 |
+
"loss": 0.3419,
|
1433 |
+
"step": 5525
|
1434 |
+
},
|
1435 |
+
{
|
1436 |
+
"epoch": 1.54,
|
1437 |
+
"learning_rate": 3.259310344827586e-05,
|
1438 |
+
"loss": 0.3876,
|
1439 |
+
"step": 5550
|
1440 |
+
},
|
1441 |
+
{
|
1442 |
+
"epoch": 1.54,
|
1443 |
+
"learning_rate": 3.250689655172414e-05,
|
1444 |
+
"loss": 0.3392,
|
1445 |
+
"step": 5575
|
1446 |
+
},
|
1447 |
+
{
|
1448 |
+
"epoch": 1.55,
|
1449 |
+
"learning_rate": 3.242068965517241e-05,
|
1450 |
+
"loss": 0.3781,
|
1451 |
+
"step": 5600
|
1452 |
+
},
|
1453 |
+
{
|
1454 |
+
"epoch": 1.56,
|
1455 |
+
"learning_rate": 3.2334482758620696e-05,
|
1456 |
+
"loss": 0.3616,
|
1457 |
+
"step": 5625
|
1458 |
+
},
|
1459 |
+
{
|
1460 |
+
"epoch": 1.57,
|
1461 |
+
"learning_rate": 3.224827586206897e-05,
|
1462 |
+
"loss": 0.3596,
|
1463 |
+
"step": 5650
|
1464 |
+
},
|
1465 |
+
{
|
1466 |
+
"epoch": 1.57,
|
1467 |
+
"learning_rate": 3.2162068965517245e-05,
|
1468 |
+
"loss": 0.3681,
|
1469 |
+
"step": 5675
|
1470 |
+
},
|
1471 |
+
{
|
1472 |
+
"epoch": 1.58,
|
1473 |
+
"learning_rate": 3.2075862068965516e-05,
|
1474 |
+
"loss": 0.3373,
|
1475 |
+
"step": 5700
|
1476 |
+
},
|
1477 |
+
{
|
1478 |
+
"epoch": 1.59,
|
1479 |
+
"learning_rate": 3.1989655172413794e-05,
|
1480 |
+
"loss": 0.3311,
|
1481 |
+
"step": 5725
|
1482 |
+
},
|
1483 |
+
{
|
1484 |
+
"epoch": 1.59,
|
1485 |
+
"learning_rate": 3.190344827586207e-05,
|
1486 |
+
"loss": 0.4004,
|
1487 |
+
"step": 5750
|
1488 |
+
},
|
1489 |
+
{
|
1490 |
+
"epoch": 1.6,
|
1491 |
+
"learning_rate": 3.181724137931035e-05,
|
1492 |
+
"loss": 0.3731,
|
1493 |
+
"step": 5775
|
1494 |
+
},
|
1495 |
+
{
|
1496 |
+
"epoch": 1.61,
|
1497 |
+
"learning_rate": 3.173103448275862e-05,
|
1498 |
+
"loss": 0.3443,
|
1499 |
+
"step": 5800
|
1500 |
+
},
|
1501 |
+
{
|
1502 |
+
"epoch": 1.61,
|
1503 |
+
"learning_rate": 3.16448275862069e-05,
|
1504 |
+
"loss": 0.345,
|
1505 |
+
"step": 5825
|
1506 |
+
},
|
1507 |
+
{
|
1508 |
+
"epoch": 1.62,
|
1509 |
+
"learning_rate": 3.155862068965517e-05,
|
1510 |
+
"loss": 0.312,
|
1511 |
+
"step": 5850
|
1512 |
+
},
|
1513 |
+
{
|
1514 |
+
"epoch": 1.63,
|
1515 |
+
"learning_rate": 3.147241379310345e-05,
|
1516 |
+
"loss": 0.3474,
|
1517 |
+
"step": 5875
|
1518 |
+
},
|
1519 |
+
{
|
1520 |
+
"epoch": 1.63,
|
1521 |
+
"learning_rate": 3.1386206896551724e-05,
|
1522 |
+
"loss": 0.3162,
|
1523 |
+
"step": 5900
|
1524 |
+
},
|
1525 |
+
{
|
1526 |
+
"epoch": 1.64,
|
1527 |
+
"learning_rate": 3.13e-05,
|
1528 |
+
"loss": 0.3227,
|
1529 |
+
"step": 5925
|
1530 |
+
},
|
1531 |
+
{
|
1532 |
+
"epoch": 1.65,
|
1533 |
+
"learning_rate": 3.121379310344828e-05,
|
1534 |
+
"loss": 0.3482,
|
1535 |
+
"step": 5950
|
1536 |
+
},
|
1537 |
+
{
|
1538 |
+
"epoch": 1.66,
|
1539 |
+
"learning_rate": 3.112758620689655e-05,
|
1540 |
+
"loss": 0.3398,
|
1541 |
+
"step": 5975
|
1542 |
+
},
|
1543 |
+
{
|
1544 |
+
"epoch": 1.66,
|
1545 |
+
"learning_rate": 3.104137931034483e-05,
|
1546 |
+
"loss": 0.3466,
|
1547 |
+
"step": 6000
|
1548 |
+
},
|
1549 |
+
{
|
1550 |
+
"epoch": 1.66,
|
1551 |
+
"eval_cer": 11.44349326007794,
|
1552 |
+
"eval_loss": 0.28201985359191895,
|
1553 |
+
"eval_runtime": 1169.509,
|
1554 |
+
"eval_samples_per_second": 4.781,
|
1555 |
+
"eval_steps_per_second": 0.299,
|
1556 |
+
"step": 6000
|
1557 |
+
},
|
1558 |
+
{
|
1559 |
+
"epoch": 1.67,
|
1560 |
+
"learning_rate": 3.0955172413793106e-05,
|
1561 |
+
"loss": 0.3256,
|
1562 |
+
"step": 6025
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 1.68,
|
1566 |
+
"learning_rate": 3.0868965517241384e-05,
|
1567 |
+
"loss": 0.3208,
|
1568 |
+
"step": 6050
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 1.68,
|
1572 |
+
"learning_rate": 3.0782758620689655e-05,
|
1573 |
+
"loss": 0.3267,
|
1574 |
+
"step": 6075
|
1575 |
+
},
|
1576 |
+
{
|
1577 |
+
"epoch": 1.69,
|
1578 |
+
"learning_rate": 3.069655172413793e-05,
|
1579 |
+
"loss": 0.3675,
|
1580 |
+
"step": 6100
|
1581 |
+
},
|
1582 |
+
{
|
1583 |
+
"epoch": 1.7,
|
1584 |
+
"learning_rate": 3.061034482758621e-05,
|
1585 |
+
"loss": 0.3502,
|
1586 |
+
"step": 6125
|
1587 |
+
},
|
1588 |
+
{
|
1589 |
+
"epoch": 1.7,
|
1590 |
+
"learning_rate": 3.052413793103449e-05,
|
1591 |
+
"loss": 0.3586,
|
1592 |
+
"step": 6150
|
1593 |
+
},
|
1594 |
+
{
|
1595 |
+
"epoch": 1.71,
|
1596 |
+
"learning_rate": 3.043793103448276e-05,
|
1597 |
+
"loss": 0.3211,
|
1598 |
+
"step": 6175
|
1599 |
+
},
|
1600 |
+
{
|
1601 |
+
"epoch": 1.72,
|
1602 |
+
"learning_rate": 3.0351724137931037e-05,
|
1603 |
+
"loss": 0.3631,
|
1604 |
+
"step": 6200
|
1605 |
+
},
|
1606 |
+
{
|
1607 |
+
"epoch": 1.72,
|
1608 |
+
"learning_rate": 3.026551724137931e-05,
|
1609 |
+
"loss": 0.3764,
|
1610 |
+
"step": 6225
|
1611 |
+
},
|
1612 |
+
{
|
1613 |
+
"epoch": 1.73,
|
1614 |
+
"learning_rate": 3.017931034482759e-05,
|
1615 |
+
"loss": 0.3636,
|
1616 |
+
"step": 6250
|
1617 |
+
},
|
1618 |
+
{
|
1619 |
+
"epoch": 1.74,
|
1620 |
+
"learning_rate": 3.009310344827586e-05,
|
1621 |
+
"loss": 0.3682,
|
1622 |
+
"step": 6275
|
1623 |
+
},
|
1624 |
+
{
|
1625 |
+
"epoch": 1.75,
|
1626 |
+
"learning_rate": 3.000689655172414e-05,
|
1627 |
+
"loss": 0.3484,
|
1628 |
+
"step": 6300
|
1629 |
+
},
|
1630 |
+
{
|
1631 |
+
"epoch": 1.75,
|
1632 |
+
"learning_rate": 2.9920689655172412e-05,
|
1633 |
+
"loss": 0.336,
|
1634 |
+
"step": 6325
|
1635 |
+
},
|
1636 |
+
{
|
1637 |
+
"epoch": 1.76,
|
1638 |
+
"learning_rate": 2.983448275862069e-05,
|
1639 |
+
"loss": 0.3457,
|
1640 |
+
"step": 6350
|
1641 |
+
},
|
1642 |
+
{
|
1643 |
+
"epoch": 1.77,
|
1644 |
+
"learning_rate": 2.974827586206897e-05,
|
1645 |
+
"loss": 0.3548,
|
1646 |
+
"step": 6375
|
1647 |
+
},
|
1648 |
+
{
|
1649 |
+
"epoch": 1.77,
|
1650 |
+
"learning_rate": 2.966206896551724e-05,
|
1651 |
+
"loss": 0.3455,
|
1652 |
+
"step": 6400
|
1653 |
+
},
|
1654 |
+
{
|
1655 |
+
"epoch": 1.78,
|
1656 |
+
"learning_rate": 2.957586206896552e-05,
|
1657 |
+
"loss": 0.3303,
|
1658 |
+
"step": 6425
|
1659 |
+
},
|
1660 |
+
{
|
1661 |
+
"epoch": 1.79,
|
1662 |
+
"learning_rate": 2.9489655172413794e-05,
|
1663 |
+
"loss": 0.3344,
|
1664 |
+
"step": 6450
|
1665 |
+
},
|
1666 |
+
{
|
1667 |
+
"epoch": 1.79,
|
1668 |
+
"learning_rate": 2.940344827586207e-05,
|
1669 |
+
"loss": 0.3517,
|
1670 |
+
"step": 6475
|
1671 |
+
},
|
1672 |
+
{
|
1673 |
+
"epoch": 1.8,
|
1674 |
+
"learning_rate": 2.9317241379310346e-05,
|
1675 |
+
"loss": 0.3507,
|
1676 |
+
"step": 6500
|
1677 |
+
},
|
1678 |
+
{
|
1679 |
+
"epoch": 1.8,
|
1680 |
+
"eval_cer": 11.012266019293426,
|
1681 |
+
"eval_loss": 0.27717769145965576,
|
1682 |
+
"eval_runtime": 1163.8916,
|
1683 |
+
"eval_samples_per_second": 4.804,
|
1684 |
+
"eval_steps_per_second": 0.301,
|
1685 |
+
"step": 6500
|
1686 |
+
},
|
1687 |
+
{
|
1688 |
+
"epoch": 1.81,
|
1689 |
+
"learning_rate": 2.9231034482758624e-05,
|
1690 |
+
"loss": 0.3283,
|
1691 |
+
"step": 6525
|
1692 |
+
},
|
1693 |
+
{
|
1694 |
+
"epoch": 1.81,
|
1695 |
+
"learning_rate": 2.9144827586206898e-05,
|
1696 |
+
"loss": 0.3808,
|
1697 |
+
"step": 6550
|
1698 |
+
},
|
1699 |
+
{
|
1700 |
+
"epoch": 1.82,
|
1701 |
+
"learning_rate": 2.9058620689655176e-05,
|
1702 |
+
"loss": 0.3473,
|
1703 |
+
"step": 6575
|
1704 |
+
},
|
1705 |
+
{
|
1706 |
+
"epoch": 1.83,
|
1707 |
+
"learning_rate": 2.8972413793103447e-05,
|
1708 |
+
"loss": 0.3502,
|
1709 |
+
"step": 6600
|
1710 |
+
},
|
1711 |
+
{
|
1712 |
+
"epoch": 1.84,
|
1713 |
+
"learning_rate": 2.8886206896551728e-05,
|
1714 |
+
"loss": 0.3507,
|
1715 |
+
"step": 6625
|
1716 |
+
},
|
1717 |
+
{
|
1718 |
+
"epoch": 1.84,
|
1719 |
+
"learning_rate": 2.88e-05,
|
1720 |
+
"loss": 0.3086,
|
1721 |
+
"step": 6650
|
1722 |
+
},
|
1723 |
+
{
|
1724 |
+
"epoch": 1.85,
|
1725 |
+
"learning_rate": 2.8713793103448276e-05,
|
1726 |
+
"loss": 0.3266,
|
1727 |
+
"step": 6675
|
1728 |
+
},
|
1729 |
+
{
|
1730 |
+
"epoch": 1.86,
|
1731 |
+
"learning_rate": 2.862758620689655e-05,
|
1732 |
+
"loss": 0.3236,
|
1733 |
+
"step": 6700
|
1734 |
+
},
|
1735 |
+
{
|
1736 |
+
"epoch": 1.86,
|
1737 |
+
"learning_rate": 2.854137931034483e-05,
|
1738 |
+
"loss": 0.3063,
|
1739 |
+
"step": 6725
|
1740 |
+
},
|
1741 |
+
{
|
1742 |
+
"epoch": 1.87,
|
1743 |
+
"learning_rate": 2.8455172413793106e-05,
|
1744 |
+
"loss": 0.3357,
|
1745 |
+
"step": 6750
|
1746 |
+
},
|
1747 |
+
{
|
1748 |
+
"epoch": 1.88,
|
1749 |
+
"learning_rate": 2.836896551724138e-05,
|
1750 |
+
"loss": 0.3366,
|
1751 |
+
"step": 6775
|
1752 |
+
},
|
1753 |
+
{
|
1754 |
+
"epoch": 1.88,
|
1755 |
+
"learning_rate": 2.828275862068966e-05,
|
1756 |
+
"loss": 0.3459,
|
1757 |
+
"step": 6800
|
1758 |
+
},
|
1759 |
+
{
|
1760 |
+
"epoch": 1.89,
|
1761 |
+
"learning_rate": 2.8196551724137933e-05,
|
1762 |
+
"loss": 0.3396,
|
1763 |
+
"step": 6825
|
1764 |
+
},
|
1765 |
+
{
|
1766 |
+
"epoch": 1.9,
|
1767 |
+
"learning_rate": 2.811034482758621e-05,
|
1768 |
+
"loss": 0.3388,
|
1769 |
+
"step": 6850
|
1770 |
+
},
|
1771 |
+
{
|
1772 |
+
"epoch": 1.9,
|
1773 |
+
"learning_rate": 2.8024137931034485e-05,
|
1774 |
+
"loss": 0.3657,
|
1775 |
+
"step": 6875
|
1776 |
+
},
|
1777 |
+
{
|
1778 |
+
"epoch": 1.91,
|
1779 |
+
"learning_rate": 2.7937931034482763e-05,
|
1780 |
+
"loss": 0.3604,
|
1781 |
+
"step": 6900
|
1782 |
+
},
|
1783 |
+
{
|
1784 |
+
"epoch": 1.92,
|
1785 |
+
"learning_rate": 2.7851724137931033e-05,
|
1786 |
+
"loss": 0.3467,
|
1787 |
+
"step": 6925
|
1788 |
+
},
|
1789 |
+
{
|
1790 |
+
"epoch": 1.93,
|
1791 |
+
"learning_rate": 2.7765517241379315e-05,
|
1792 |
+
"loss": 0.3434,
|
1793 |
+
"step": 6950
|
1794 |
+
},
|
1795 |
+
{
|
1796 |
+
"epoch": 1.93,
|
1797 |
+
"learning_rate": 2.7679310344827586e-05,
|
1798 |
+
"loss": 0.3416,
|
1799 |
+
"step": 6975
|
1800 |
+
},
|
1801 |
+
{
|
1802 |
+
"epoch": 1.94,
|
1803 |
+
"learning_rate": 2.7593103448275863e-05,
|
1804 |
+
"loss": 0.3482,
|
1805 |
+
"step": 7000
|
1806 |
+
},
|
1807 |
+
{
|
1808 |
+
"epoch": 1.94,
|
1809 |
+
"eval_cer": 11.178368363891906,
|
1810 |
+
"eval_loss": 0.2766138017177582,
|
1811 |
+
"eval_runtime": 1182.8471,
|
1812 |
+
"eval_samples_per_second": 4.727,
|
1813 |
+
"eval_steps_per_second": 0.296,
|
1814 |
+
"step": 7000
|
1815 |
+
},
|
1816 |
+
{
|
1817 |
+
"epoch": 1.95,
|
1818 |
+
"learning_rate": 2.7506896551724138e-05,
|
1819 |
+
"loss": 0.3213,
|
1820 |
+
"step": 7025
|
1821 |
+
},
|
1822 |
+
{
|
1823 |
+
"epoch": 1.95,
|
1824 |
+
"learning_rate": 2.7420689655172415e-05,
|
1825 |
+
"loss": 0.359,
|
1826 |
+
"step": 7050
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 1.96,
|
1830 |
+
"learning_rate": 2.733448275862069e-05,
|
1831 |
+
"loss": 0.3506,
|
1832 |
+
"step": 7075
|
1833 |
+
},
|
1834 |
+
{
|
1835 |
+
"epoch": 1.97,
|
1836 |
+
"learning_rate": 2.7248275862068968e-05,
|
1837 |
+
"loss": 0.3178,
|
1838 |
+
"step": 7100
|
1839 |
+
},
|
1840 |
+
{
|
1841 |
+
"epoch": 1.97,
|
1842 |
+
"learning_rate": 2.716206896551724e-05,
|
1843 |
+
"loss": 0.3371,
|
1844 |
+
"step": 7125
|
1845 |
+
},
|
1846 |
+
{
|
1847 |
+
"epoch": 1.98,
|
1848 |
+
"learning_rate": 2.707586206896552e-05,
|
1849 |
+
"loss": 0.3236,
|
1850 |
+
"step": 7150
|
1851 |
+
},
|
1852 |
+
{
|
1853 |
+
"epoch": 1.99,
|
1854 |
+
"learning_rate": 2.6989655172413797e-05,
|
1855 |
+
"loss": 0.3231,
|
1856 |
+
"step": 7175
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 1.99,
|
1860 |
+
"learning_rate": 2.6903448275862068e-05,
|
1861 |
+
"loss": 0.3007,
|
1862 |
+
"step": 7200
|
1863 |
+
},
|
1864 |
+
{
|
1865 |
+
"epoch": 2.0,
|
1866 |
+
"learning_rate": 2.6820689655172414e-05,
|
1867 |
+
"loss": 0.324,
|
1868 |
+
"step": 7225
|
1869 |
+
},
|
1870 |
+
{
|
1871 |
+
"epoch": 2.01,
|
1872 |
+
"learning_rate": 2.6734482758620692e-05,
|
1873 |
+
"loss": 0.2574,
|
1874 |
+
"step": 7250
|
1875 |
+
},
|
1876 |
+
{
|
1877 |
+
"epoch": 2.02,
|
1878 |
+
"learning_rate": 2.6648275862068966e-05,
|
1879 |
+
"loss": 0.2346,
|
1880 |
+
"step": 7275
|
1881 |
+
},
|
1882 |
+
{
|
1883 |
+
"epoch": 2.02,
|
1884 |
+
"learning_rate": 2.6562068965517244e-05,
|
1885 |
+
"loss": 0.2627,
|
1886 |
+
"step": 7300
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 2.03,
|
1890 |
+
"learning_rate": 2.647586206896552e-05,
|
1891 |
+
"loss": 0.2614,
|
1892 |
+
"step": 7325
|
1893 |
+
},
|
1894 |
+
{
|
1895 |
+
"epoch": 2.04,
|
1896 |
+
"learning_rate": 2.6389655172413796e-05,
|
1897 |
+
"loss": 0.2279,
|
1898 |
+
"step": 7350
|
1899 |
+
},
|
1900 |
+
{
|
1901 |
+
"epoch": 2.04,
|
1902 |
+
"learning_rate": 2.630344827586207e-05,
|
1903 |
+
"loss": 0.271,
|
1904 |
+
"step": 7375
|
1905 |
+
},
|
1906 |
+
{
|
1907 |
+
"epoch": 2.05,
|
1908 |
+
"learning_rate": 2.621724137931035e-05,
|
1909 |
+
"loss": 0.2566,
|
1910 |
+
"step": 7400
|
1911 |
+
},
|
1912 |
+
{
|
1913 |
+
"epoch": 2.06,
|
1914 |
+
"learning_rate": 2.613103448275862e-05,
|
1915 |
+
"loss": 0.2471,
|
1916 |
+
"step": 7425
|
1917 |
+
},
|
1918 |
+
{
|
1919 |
+
"epoch": 2.06,
|
1920 |
+
"learning_rate": 2.60448275862069e-05,
|
1921 |
+
"loss": 0.2438,
|
1922 |
+
"step": 7450
|
1923 |
+
},
|
1924 |
+
{
|
1925 |
+
"epoch": 2.07,
|
1926 |
+
"learning_rate": 2.595862068965517e-05,
|
1927 |
+
"loss": 0.2457,
|
1928 |
+
"step": 7475
|
1929 |
+
},
|
1930 |
+
{
|
1931 |
+
"epoch": 2.08,
|
1932 |
+
"learning_rate": 2.587241379310345e-05,
|
1933 |
+
"loss": 0.232,
|
1934 |
+
"step": 7500
|
1935 |
+
},
|
1936 |
+
{
|
1937 |
+
"epoch": 2.08,
|
1938 |
+
"eval_cer": 10.926020571136522,
|
1939 |
+
"eval_loss": 0.2784821093082428,
|
1940 |
+
"eval_runtime": 1168.6896,
|
1941 |
+
"eval_samples_per_second": 4.784,
|
1942 |
+
"eval_steps_per_second": 0.299,
|
1943 |
+
"step": 7500
|
1944 |
+
},
|
1945 |
+
{
|
1946 |
+
"epoch": 2.09,
|
1947 |
+
"learning_rate": 2.5786206896551724e-05,
|
1948 |
+
"loss": 0.256,
|
1949 |
+
"step": 7525
|
1950 |
+
},
|
1951 |
+
{
|
1952 |
+
"epoch": 2.09,
|
1953 |
+
"learning_rate": 2.57e-05,
|
1954 |
+
"loss": 0.2383,
|
1955 |
+
"step": 7550
|
1956 |
+
},
|
1957 |
+
{
|
1958 |
+
"epoch": 2.1,
|
1959 |
+
"learning_rate": 2.5613793103448276e-05,
|
1960 |
+
"loss": 0.2516,
|
1961 |
+
"step": 7575
|
1962 |
+
},
|
1963 |
+
{
|
1964 |
+
"epoch": 2.11,
|
1965 |
+
"learning_rate": 2.5527586206896553e-05,
|
1966 |
+
"loss": 0.229,
|
1967 |
+
"step": 7600
|
1968 |
+
},
|
1969 |
+
{
|
1970 |
+
"epoch": 2.11,
|
1971 |
+
"learning_rate": 2.5441379310344828e-05,
|
1972 |
+
"loss": 0.242,
|
1973 |
+
"step": 7625
|
1974 |
+
},
|
1975 |
+
{
|
1976 |
+
"epoch": 2.12,
|
1977 |
+
"learning_rate": 2.5355172413793105e-05,
|
1978 |
+
"loss": 0.243,
|
1979 |
+
"step": 7650
|
1980 |
+
},
|
1981 |
+
{
|
1982 |
+
"epoch": 2.13,
|
1983 |
+
"learning_rate": 2.5268965517241383e-05,
|
1984 |
+
"loss": 0.2476,
|
1985 |
+
"step": 7675
|
1986 |
+
},
|
1987 |
+
{
|
1988 |
+
"epoch": 2.13,
|
1989 |
+
"learning_rate": 2.5182758620689658e-05,
|
1990 |
+
"loss": 0.2351,
|
1991 |
+
"step": 7700
|
1992 |
+
},
|
1993 |
+
{
|
1994 |
+
"epoch": 2.14,
|
1995 |
+
"learning_rate": 2.5096551724137935e-05,
|
1996 |
+
"loss": 0.2526,
|
1997 |
+
"step": 7725
|
1998 |
+
},
|
1999 |
+
{
|
2000 |
+
"epoch": 2.15,
|
2001 |
+
"learning_rate": 2.5010344827586206e-05,
|
2002 |
+
"loss": 0.2424,
|
2003 |
+
"step": 7750
|
2004 |
+
},
|
2005 |
+
{
|
2006 |
+
"epoch": 2.15,
|
2007 |
+
"learning_rate": 2.4924137931034484e-05,
|
2008 |
+
"loss": 0.264,
|
2009 |
+
"step": 7775
|
2010 |
+
},
|
2011 |
+
{
|
2012 |
+
"epoch": 2.16,
|
2013 |
+
"learning_rate": 2.483793103448276e-05,
|
2014 |
+
"loss": 0.244,
|
2015 |
+
"step": 7800
|
2016 |
+
},
|
2017 |
+
{
|
2018 |
+
"epoch": 2.17,
|
2019 |
+
"learning_rate": 2.4751724137931036e-05,
|
2020 |
+
"loss": 0.2588,
|
2021 |
+
"step": 7825
|
2022 |
+
},
|
2023 |
+
{
|
2024 |
+
"epoch": 2.18,
|
2025 |
+
"learning_rate": 2.4665517241379314e-05,
|
2026 |
+
"loss": 0.2418,
|
2027 |
+
"step": 7850
|
2028 |
+
},
|
2029 |
+
{
|
2030 |
+
"epoch": 2.18,
|
2031 |
+
"learning_rate": 2.4579310344827588e-05,
|
2032 |
+
"loss": 0.2556,
|
2033 |
+
"step": 7875
|
2034 |
+
},
|
2035 |
+
{
|
2036 |
+
"epoch": 2.19,
|
2037 |
+
"learning_rate": 2.4493103448275866e-05,
|
2038 |
+
"loss": 0.2253,
|
2039 |
+
"step": 7900
|
2040 |
+
},
|
2041 |
+
{
|
2042 |
+
"epoch": 2.2,
|
2043 |
+
"learning_rate": 2.440689655172414e-05,
|
2044 |
+
"loss": 0.2629,
|
2045 |
+
"step": 7925
|
2046 |
+
},
|
2047 |
+
{
|
2048 |
+
"epoch": 2.2,
|
2049 |
+
"learning_rate": 2.4320689655172415e-05,
|
2050 |
+
"loss": 0.2512,
|
2051 |
+
"step": 7950
|
2052 |
+
},
|
2053 |
+
{
|
2054 |
+
"epoch": 2.21,
|
2055 |
+
"learning_rate": 2.4234482758620692e-05,
|
2056 |
+
"loss": 0.2186,
|
2057 |
+
"step": 7975
|
2058 |
+
},
|
2059 |
+
{
|
2060 |
+
"epoch": 2.22,
|
2061 |
+
"learning_rate": 2.4148275862068967e-05,
|
2062 |
+
"loss": 0.2539,
|
2063 |
+
"step": 8000
|
2064 |
+
},
|
2065 |
+
{
|
2066 |
+
"epoch": 2.22,
|
2067 |
+
"eval_cer": 10.593815881939564,
|
2068 |
+
"eval_loss": 0.2776859700679779,
|
2069 |
+
"eval_runtime": 1165.9559,
|
2070 |
+
"eval_samples_per_second": 4.795,
|
2071 |
+
"eval_steps_per_second": 0.3,
|
2072 |
+
"step": 8000
|
2073 |
+
},
|
2074 |
+
{
|
2075 |
+
"epoch": 2.22,
|
2076 |
+
"learning_rate": 2.406206896551724e-05,
|
2077 |
+
"loss": 0.2433,
|
2078 |
+
"step": 8025
|
2079 |
+
},
|
2080 |
+
{
|
2081 |
+
"epoch": 2.23,
|
2082 |
+
"learning_rate": 2.397586206896552e-05,
|
2083 |
+
"loss": 0.2383,
|
2084 |
+
"step": 8050
|
2085 |
+
},
|
2086 |
+
{
|
2087 |
+
"epoch": 2.24,
|
2088 |
+
"learning_rate": 2.3889655172413793e-05,
|
2089 |
+
"loss": 0.2454,
|
2090 |
+
"step": 8075
|
2091 |
+
},
|
2092 |
+
{
|
2093 |
+
"epoch": 2.24,
|
2094 |
+
"learning_rate": 2.380344827586207e-05,
|
2095 |
+
"loss": 0.2596,
|
2096 |
+
"step": 8100
|
2097 |
+
},
|
2098 |
+
{
|
2099 |
+
"epoch": 2.25,
|
2100 |
+
"learning_rate": 2.3717241379310345e-05,
|
2101 |
+
"loss": 0.2479,
|
2102 |
+
"step": 8125
|
2103 |
+
},
|
2104 |
+
{
|
2105 |
+
"epoch": 2.26,
|
2106 |
+
"learning_rate": 2.363103448275862e-05,
|
2107 |
+
"loss": 0.2519,
|
2108 |
+
"step": 8150
|
2109 |
+
},
|
2110 |
+
{
|
2111 |
+
"epoch": 2.27,
|
2112 |
+
"learning_rate": 2.3544827586206897e-05,
|
2113 |
+
"loss": 0.2394,
|
2114 |
+
"step": 8175
|
2115 |
+
},
|
2116 |
+
{
|
2117 |
+
"epoch": 2.27,
|
2118 |
+
"learning_rate": 2.345862068965517e-05,
|
2119 |
+
"loss": 0.2336,
|
2120 |
+
"step": 8200
|
2121 |
+
},
|
2122 |
+
{
|
2123 |
+
"epoch": 2.28,
|
2124 |
+
"learning_rate": 2.337241379310345e-05,
|
2125 |
+
"loss": 0.2549,
|
2126 |
+
"step": 8225
|
2127 |
+
},
|
2128 |
+
{
|
2129 |
+
"epoch": 2.29,
|
2130 |
+
"learning_rate": 2.3286206896551727e-05,
|
2131 |
+
"loss": 0.2784,
|
2132 |
+
"step": 8250
|
2133 |
+
},
|
2134 |
+
{
|
2135 |
+
"epoch": 2.29,
|
2136 |
+
"learning_rate": 2.32e-05,
|
2137 |
+
"loss": 0.2365,
|
2138 |
+
"step": 8275
|
2139 |
+
},
|
2140 |
+
{
|
2141 |
+
"epoch": 2.3,
|
2142 |
+
"learning_rate": 2.311379310344828e-05,
|
2143 |
+
"loss": 0.2445,
|
2144 |
+
"step": 8300
|
2145 |
+
},
|
2146 |
+
{
|
2147 |
+
"epoch": 2.31,
|
2148 |
+
"learning_rate": 2.3027586206896554e-05,
|
2149 |
+
"loss": 0.2411,
|
2150 |
+
"step": 8325
|
2151 |
+
},
|
2152 |
+
{
|
2153 |
+
"epoch": 2.31,
|
2154 |
+
"learning_rate": 2.2941379310344828e-05,
|
2155 |
+
"loss": 0.257,
|
2156 |
+
"step": 8350
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 2.32,
|
2160 |
+
"learning_rate": 2.2855172413793106e-05,
|
2161 |
+
"loss": 0.2196,
|
2162 |
+
"step": 8375
|
2163 |
+
},
|
2164 |
+
{
|
2165 |
+
"epoch": 2.33,
|
2166 |
+
"learning_rate": 2.276896551724138e-05,
|
2167 |
+
"loss": 0.2583,
|
2168 |
+
"step": 8400
|
2169 |
+
},
|
2170 |
+
{
|
2171 |
+
"epoch": 2.33,
|
2172 |
+
"learning_rate": 2.2682758620689658e-05,
|
2173 |
+
"loss": 0.2522,
|
2174 |
+
"step": 8425
|
2175 |
+
},
|
2176 |
+
{
|
2177 |
+
"epoch": 2.34,
|
2178 |
+
"learning_rate": 2.2596551724137932e-05,
|
2179 |
+
"loss": 0.2712,
|
2180 |
+
"step": 8450
|
2181 |
+
},
|
2182 |
+
{
|
2183 |
+
"epoch": 2.35,
|
2184 |
+
"learning_rate": 2.2510344827586206e-05,
|
2185 |
+
"loss": 0.2624,
|
2186 |
+
"step": 8475
|
2187 |
+
},
|
2188 |
+
{
|
2189 |
+
"epoch": 2.36,
|
2190 |
+
"learning_rate": 2.2424137931034484e-05,
|
2191 |
+
"loss": 0.2377,
|
2192 |
+
"step": 8500
|
2193 |
+
},
|
2194 |
+
{
|
2195 |
+
"epoch": 2.36,
|
2196 |
+
"eval_cer": 10.659298537021657,
|
2197 |
+
"eval_loss": 0.2783927917480469,
|
2198 |
+
"eval_runtime": 1180.6736,
|
2199 |
+
"eval_samples_per_second": 4.735,
|
2200 |
+
"eval_steps_per_second": 0.296,
|
2201 |
+
"step": 8500
|
2202 |
+
},
|
2203 |
+
{
|
2204 |
+
"epoch": 2.36,
|
2205 |
+
"learning_rate": 2.233793103448276e-05,
|
2206 |
+
"loss": 0.2526,
|
2207 |
+
"step": 8525
|
2208 |
+
},
|
2209 |
+
{
|
2210 |
+
"epoch": 2.37,
|
2211 |
+
"learning_rate": 2.2251724137931036e-05,
|
2212 |
+
"loss": 0.2408,
|
2213 |
+
"step": 8550
|
2214 |
+
},
|
2215 |
+
{
|
2216 |
+
"epoch": 2.38,
|
2217 |
+
"learning_rate": 2.216551724137931e-05,
|
2218 |
+
"loss": 0.2778,
|
2219 |
+
"step": 8575
|
2220 |
+
},
|
2221 |
+
{
|
2222 |
+
"epoch": 2.38,
|
2223 |
+
"learning_rate": 2.2079310344827585e-05,
|
2224 |
+
"loss": 0.253,
|
2225 |
+
"step": 8600
|
2226 |
+
},
|
2227 |
+
{
|
2228 |
+
"epoch": 2.39,
|
2229 |
+
"learning_rate": 2.1993103448275863e-05,
|
2230 |
+
"loss": 0.2416,
|
2231 |
+
"step": 8625
|
2232 |
+
},
|
2233 |
+
{
|
2234 |
+
"epoch": 2.4,
|
2235 |
+
"learning_rate": 2.190689655172414e-05,
|
2236 |
+
"loss": 0.2477,
|
2237 |
+
"step": 8650
|
2238 |
+
},
|
2239 |
+
{
|
2240 |
+
"epoch": 2.4,
|
2241 |
+
"learning_rate": 2.1820689655172415e-05,
|
2242 |
+
"loss": 0.2453,
|
2243 |
+
"step": 8675
|
2244 |
+
},
|
2245 |
+
{
|
2246 |
+
"epoch": 2.41,
|
2247 |
+
"learning_rate": 2.1734482758620693e-05,
|
2248 |
+
"loss": 0.2463,
|
2249 |
+
"step": 8700
|
2250 |
+
},
|
2251 |
+
{
|
2252 |
+
"epoch": 2.42,
|
2253 |
+
"learning_rate": 2.1648275862068967e-05,
|
2254 |
+
"loss": 0.2409,
|
2255 |
+
"step": 8725
|
2256 |
+
},
|
2257 |
+
{
|
2258 |
+
"epoch": 2.42,
|
2259 |
+
"learning_rate": 2.1562068965517245e-05,
|
2260 |
+
"loss": 0.2305,
|
2261 |
+
"step": 8750
|
2262 |
+
},
|
2263 |
+
{
|
2264 |
+
"epoch": 2.43,
|
2265 |
+
"learning_rate": 2.147586206896552e-05,
|
2266 |
+
"loss": 0.2342,
|
2267 |
+
"step": 8775
|
2268 |
+
},
|
2269 |
+
{
|
2270 |
+
"epoch": 2.44,
|
2271 |
+
"learning_rate": 2.1389655172413793e-05,
|
2272 |
+
"loss": 0.2551,
|
2273 |
+
"step": 8800
|
2274 |
+
},
|
2275 |
+
{
|
2276 |
+
"epoch": 2.45,
|
2277 |
+
"learning_rate": 2.130344827586207e-05,
|
2278 |
+
"loss": 0.2396,
|
2279 |
+
"step": 8825
|
2280 |
+
},
|
2281 |
+
{
|
2282 |
+
"epoch": 2.45,
|
2283 |
+
"learning_rate": 2.1217241379310345e-05,
|
2284 |
+
"loss": 0.268,
|
2285 |
+
"step": 8850
|
2286 |
+
},
|
2287 |
+
{
|
2288 |
+
"epoch": 2.46,
|
2289 |
+
"learning_rate": 2.1131034482758623e-05,
|
2290 |
+
"loss": 0.2384,
|
2291 |
+
"step": 8875
|
2292 |
+
},
|
2293 |
+
{
|
2294 |
+
"epoch": 2.47,
|
2295 |
+
"learning_rate": 2.1044827586206898e-05,
|
2296 |
+
"loss": 0.2418,
|
2297 |
+
"step": 8900
|
2298 |
+
},
|
2299 |
+
{
|
2300 |
+
"epoch": 2.47,
|
2301 |
+
"learning_rate": 2.0958620689655172e-05,
|
2302 |
+
"loss": 0.2414,
|
2303 |
+
"step": 8925
|
2304 |
+
},
|
2305 |
+
{
|
2306 |
+
"epoch": 2.48,
|
2307 |
+
"learning_rate": 2.087241379310345e-05,
|
2308 |
+
"loss": 0.2232,
|
2309 |
+
"step": 8950
|
2310 |
+
},
|
2311 |
+
{
|
2312 |
+
"epoch": 2.49,
|
2313 |
+
"learning_rate": 2.0786206896551724e-05,
|
2314 |
+
"loss": 0.248,
|
2315 |
+
"step": 8975
|
2316 |
+
},
|
2317 |
+
{
|
2318 |
+
"epoch": 2.49,
|
2319 |
+
"learning_rate": 2.07e-05,
|
2320 |
+
"loss": 0.2384,
|
2321 |
+
"step": 9000
|
2322 |
+
},
|
2323 |
+
{
|
2324 |
+
"epoch": 2.49,
|
2325 |
+
"eval_cer": 10.659298537021657,
|
2326 |
+
"eval_loss": 0.2808603048324585,
|
2327 |
+
"eval_runtime": 1165.5072,
|
2328 |
+
"eval_samples_per_second": 4.797,
|
2329 |
+
"eval_steps_per_second": 0.3,
|
2330 |
+
"step": 9000
|
2331 |
+
},
|
2332 |
+
{
|
2333 |
+
"epoch": 2.5,
|
2334 |
+
"learning_rate": 2.0613793103448276e-05,
|
2335 |
+
"loss": 0.2448,
|
2336 |
+
"step": 9025
|
2337 |
+
},
|
2338 |
+
{
|
2339 |
+
"epoch": 2.51,
|
2340 |
+
"learning_rate": 2.0527586206896554e-05,
|
2341 |
+
"loss": 0.2268,
|
2342 |
+
"step": 9050
|
2343 |
+
},
|
2344 |
+
{
|
2345 |
+
"epoch": 2.51,
|
2346 |
+
"learning_rate": 2.0441379310344828e-05,
|
2347 |
+
"loss": 0.247,
|
2348 |
+
"step": 9075
|
2349 |
+
},
|
2350 |
+
{
|
2351 |
+
"epoch": 2.52,
|
2352 |
+
"learning_rate": 2.0355172413793106e-05,
|
2353 |
+
"loss": 0.2345,
|
2354 |
+
"step": 9100
|
2355 |
+
},
|
2356 |
+
{
|
2357 |
+
"epoch": 2.53,
|
2358 |
+
"learning_rate": 2.026896551724138e-05,
|
2359 |
+
"loss": 0.2117,
|
2360 |
+
"step": 9125
|
2361 |
+
},
|
2362 |
+
{
|
2363 |
+
"epoch": 2.54,
|
2364 |
+
"learning_rate": 2.0182758620689658e-05,
|
2365 |
+
"loss": 0.2265,
|
2366 |
+
"step": 9150
|
2367 |
+
},
|
2368 |
+
{
|
2369 |
+
"epoch": 2.54,
|
2370 |
+
"learning_rate": 2.0096551724137932e-05,
|
2371 |
+
"loss": 0.2464,
|
2372 |
+
"step": 9175
|
2373 |
+
},
|
2374 |
+
{
|
2375 |
+
"epoch": 2.55,
|
2376 |
+
"learning_rate": 2.0010344827586207e-05,
|
2377 |
+
"loss": 0.2741,
|
2378 |
+
"step": 9200
|
2379 |
+
},
|
2380 |
+
{
|
2381 |
+
"epoch": 2.56,
|
2382 |
+
"learning_rate": 1.9924137931034484e-05,
|
2383 |
+
"loss": 0.2397,
|
2384 |
+
"step": 9225
|
2385 |
+
},
|
2386 |
+
{
|
2387 |
+
"epoch": 2.56,
|
2388 |
+
"learning_rate": 1.983793103448276e-05,
|
2389 |
+
"loss": 0.2357,
|
2390 |
+
"step": 9250
|
2391 |
+
},
|
2392 |
+
{
|
2393 |
+
"epoch": 2.57,
|
2394 |
+
"learning_rate": 1.9751724137931037e-05,
|
2395 |
+
"loss": 0.2435,
|
2396 |
+
"step": 9275
|
2397 |
+
},
|
2398 |
+
{
|
2399 |
+
"epoch": 2.58,
|
2400 |
+
"learning_rate": 1.966551724137931e-05,
|
2401 |
+
"loss": 0.2571,
|
2402 |
+
"step": 9300
|
2403 |
+
},
|
2404 |
+
{
|
2405 |
+
"epoch": 2.58,
|
2406 |
+
"learning_rate": 1.9579310344827585e-05,
|
2407 |
+
"loss": 0.2675,
|
2408 |
+
"step": 9325
|
2409 |
+
},
|
2410 |
+
{
|
2411 |
+
"epoch": 2.59,
|
2412 |
+
"learning_rate": 1.9493103448275863e-05,
|
2413 |
+
"loss": 0.2476,
|
2414 |
+
"step": 9350
|
2415 |
+
},
|
2416 |
+
{
|
2417 |
+
"epoch": 2.6,
|
2418 |
+
"learning_rate": 1.9406896551724137e-05,
|
2419 |
+
"loss": 0.2412,
|
2420 |
+
"step": 9375
|
2421 |
+
},
|
2422 |
+
{
|
2423 |
+
"epoch": 2.6,
|
2424 |
+
"learning_rate": 1.9320689655172415e-05,
|
2425 |
+
"loss": 0.2502,
|
2426 |
+
"step": 9400
|
2427 |
+
},
|
2428 |
+
{
|
2429 |
+
"epoch": 2.61,
|
2430 |
+
"learning_rate": 1.923448275862069e-05,
|
2431 |
+
"loss": 0.2327,
|
2432 |
+
"step": 9425
|
2433 |
+
},
|
2434 |
+
{
|
2435 |
+
"epoch": 2.62,
|
2436 |
+
"learning_rate": 1.9148275862068964e-05,
|
2437 |
+
"loss": 0.2198,
|
2438 |
+
"step": 9450
|
2439 |
+
},
|
2440 |
+
{
|
2441 |
+
"epoch": 2.63,
|
2442 |
+
"learning_rate": 1.9062068965517245e-05,
|
2443 |
+
"loss": 0.2614,
|
2444 |
+
"step": 9475
|
2445 |
+
},
|
2446 |
+
{
|
2447 |
+
"epoch": 2.63,
|
2448 |
+
"learning_rate": 1.897586206896552e-05,
|
2449 |
+
"loss": 0.2495,
|
2450 |
+
"step": 9500
|
2451 |
+
},
|
2452 |
+
{
|
2453 |
+
"epoch": 2.63,
|
2454 |
+
"eval_cer": 10.632147192231521,
|
2455 |
+
"eval_loss": 0.281376451253891,
|
2456 |
+
"eval_runtime": 1164.8408,
|
2457 |
+
"eval_samples_per_second": 4.8,
|
2458 |
+
"eval_steps_per_second": 0.3,
|
2459 |
+
"step": 9500
|
2460 |
+
},
|
2461 |
+
{
|
2462 |
+
"epoch": 2.64,
|
2463 |
+
"learning_rate": 1.8889655172413794e-05,
|
2464 |
+
"loss": 0.2442,
|
2465 |
+
"step": 9525
|
2466 |
+
},
|
2467 |
+
{
|
2468 |
+
"epoch": 2.65,
|
2469 |
+
"learning_rate": 1.880344827586207e-05,
|
2470 |
+
"loss": 0.2369,
|
2471 |
+
"step": 9550
|
2472 |
+
},
|
2473 |
+
{
|
2474 |
+
"epoch": 2.65,
|
2475 |
+
"learning_rate": 1.8717241379310346e-05,
|
2476 |
+
"loss": 0.2186,
|
2477 |
+
"step": 9575
|
2478 |
+
},
|
2479 |
+
{
|
2480 |
+
"epoch": 2.66,
|
2481 |
+
"learning_rate": 1.8631034482758623e-05,
|
2482 |
+
"loss": 0.234,
|
2483 |
+
"step": 9600
|
2484 |
+
},
|
2485 |
+
{
|
2486 |
+
"epoch": 2.67,
|
2487 |
+
"learning_rate": 1.8544827586206898e-05,
|
2488 |
+
"loss": 0.2448,
|
2489 |
+
"step": 9625
|
2490 |
+
},
|
2491 |
+
{
|
2492 |
+
"epoch": 2.67,
|
2493 |
+
"learning_rate": 1.8458620689655172e-05,
|
2494 |
+
"loss": 0.2449,
|
2495 |
+
"step": 9650
|
2496 |
+
},
|
2497 |
+
{
|
2498 |
+
"epoch": 2.68,
|
2499 |
+
"learning_rate": 1.837241379310345e-05,
|
2500 |
+
"loss": 0.2322,
|
2501 |
+
"step": 9675
|
2502 |
+
},
|
2503 |
+
{
|
2504 |
+
"epoch": 2.69,
|
2505 |
+
"learning_rate": 1.8286206896551724e-05,
|
2506 |
+
"loss": 0.2489,
|
2507 |
+
"step": 9700
|
2508 |
+
},
|
2509 |
+
{
|
2510 |
+
"epoch": 2.69,
|
2511 |
+
"learning_rate": 1.8200000000000002e-05,
|
2512 |
+
"loss": 0.2569,
|
2513 |
+
"step": 9725
|
2514 |
+
},
|
2515 |
+
{
|
2516 |
+
"epoch": 2.7,
|
2517 |
+
"learning_rate": 1.8113793103448276e-05,
|
2518 |
+
"loss": 0.2331,
|
2519 |
+
"step": 9750
|
2520 |
+
},
|
2521 |
+
{
|
2522 |
+
"epoch": 2.71,
|
2523 |
+
"learning_rate": 1.802758620689655e-05,
|
2524 |
+
"loss": 0.2647,
|
2525 |
+
"step": 9775
|
2526 |
+
},
|
2527 |
+
{
|
2528 |
+
"epoch": 2.72,
|
2529 |
+
"learning_rate": 1.794137931034483e-05,
|
2530 |
+
"loss": 0.2426,
|
2531 |
+
"step": 9800
|
2532 |
+
},
|
2533 |
+
{
|
2534 |
+
"epoch": 2.72,
|
2535 |
+
"learning_rate": 1.7855172413793103e-05,
|
2536 |
+
"loss": 0.2341,
|
2537 |
+
"step": 9825
|
2538 |
+
},
|
2539 |
+
{
|
2540 |
+
"epoch": 2.73,
|
2541 |
+
"learning_rate": 1.776896551724138e-05,
|
2542 |
+
"loss": 0.24,
|
2543 |
+
"step": 9850
|
2544 |
+
},
|
2545 |
+
{
|
2546 |
+
"epoch": 2.74,
|
2547 |
+
"learning_rate": 1.7682758620689658e-05,
|
2548 |
+
"loss": 0.249,
|
2549 |
+
"step": 9875
|
2550 |
+
},
|
2551 |
+
{
|
2552 |
+
"epoch": 2.74,
|
2553 |
+
"learning_rate": 1.7596551724137933e-05,
|
2554 |
+
"loss": 0.2217,
|
2555 |
+
"step": 9900
|
2556 |
+
},
|
2557 |
+
{
|
2558 |
+
"epoch": 2.75,
|
2559 |
+
"learning_rate": 1.751034482758621e-05,
|
2560 |
+
"loss": 0.2467,
|
2561 |
+
"step": 9925
|
2562 |
+
},
|
2563 |
+
{
|
2564 |
+
"epoch": 2.76,
|
2565 |
+
"learning_rate": 1.7424137931034485e-05,
|
2566 |
+
"loss": 0.2529,
|
2567 |
+
"step": 9950
|
2568 |
+
},
|
2569 |
+
{
|
2570 |
+
"epoch": 2.76,
|
2571 |
+
"learning_rate": 1.733793103448276e-05,
|
2572 |
+
"loss": 0.2578,
|
2573 |
+
"step": 9975
|
2574 |
+
},
|
2575 |
+
{
|
2576 |
+
"epoch": 2.77,
|
2577 |
+
"learning_rate": 1.7251724137931037e-05,
|
2578 |
+
"loss": 0.2312,
|
2579 |
+
"step": 10000
|
2580 |
+
},
|
2581 |
+
{
|
2582 |
+
"epoch": 2.77,
|
2583 |
+
"eval_cer": 10.60020443365489,
|
2584 |
+
"eval_loss": 0.28221753239631653,
|
2585 |
+
"eval_runtime": 1168.5035,
|
2586 |
+
"eval_samples_per_second": 4.785,
|
2587 |
+
"eval_steps_per_second": 0.3,
|
2588 |
+
"step": 10000
|
2589 |
+
},
|
2590 |
+
{
|
2591 |
+
"epoch": 2.78,
|
2592 |
+
"learning_rate": 1.716551724137931e-05,
|
2593 |
+
"loss": 0.2284,
|
2594 |
+
"step": 10025
|
2595 |
+
},
|
2596 |
+
{
|
2597 |
+
"epoch": 2.78,
|
2598 |
+
"learning_rate": 1.7079310344827585e-05,
|
2599 |
+
"loss": 0.2348,
|
2600 |
+
"step": 10050
|
2601 |
+
},
|
2602 |
+
{
|
2603 |
+
"epoch": 2.79,
|
2604 |
+
"learning_rate": 1.6993103448275863e-05,
|
2605 |
+
"loss": 0.2244,
|
2606 |
+
"step": 10075
|
2607 |
+
},
|
2608 |
+
{
|
2609 |
+
"epoch": 2.8,
|
2610 |
+
"learning_rate": 1.6906896551724138e-05,
|
2611 |
+
"loss": 0.2522,
|
2612 |
+
"step": 10100
|
2613 |
+
},
|
2614 |
+
{
|
2615 |
+
"epoch": 2.81,
|
2616 |
+
"learning_rate": 1.6820689655172415e-05,
|
2617 |
+
"loss": 0.2476,
|
2618 |
+
"step": 10125
|
2619 |
+
},
|
2620 |
+
{
|
2621 |
+
"epoch": 2.81,
|
2622 |
+
"learning_rate": 1.673448275862069e-05,
|
2623 |
+
"loss": 0.2279,
|
2624 |
+
"step": 10150
|
2625 |
+
},
|
2626 |
+
{
|
2627 |
+
"epoch": 2.82,
|
2628 |
+
"learning_rate": 1.6648275862068964e-05,
|
2629 |
+
"loss": 0.2317,
|
2630 |
+
"step": 10175
|
2631 |
+
},
|
2632 |
+
{
|
2633 |
+
"epoch": 2.83,
|
2634 |
+
"learning_rate": 1.6562068965517242e-05,
|
2635 |
+
"loss": 0.2305,
|
2636 |
+
"step": 10200
|
2637 |
+
},
|
2638 |
+
{
|
2639 |
+
"epoch": 2.83,
|
2640 |
+
"learning_rate": 1.6475862068965516e-05,
|
2641 |
+
"loss": 0.2364,
|
2642 |
+
"step": 10225
|
2643 |
+
},
|
2644 |
+
{
|
2645 |
+
"epoch": 2.84,
|
2646 |
+
"learning_rate": 1.6389655172413794e-05,
|
2647 |
+
"loss": 0.2572,
|
2648 |
+
"step": 10250
|
2649 |
+
},
|
2650 |
+
{
|
2651 |
+
"epoch": 2.85,
|
2652 |
+
"learning_rate": 1.630344827586207e-05,
|
2653 |
+
"loss": 0.2065,
|
2654 |
+
"step": 10275
|
2655 |
+
},
|
2656 |
+
{
|
2657 |
+
"epoch": 2.85,
|
2658 |
+
"learning_rate": 1.6217241379310346e-05,
|
2659 |
+
"loss": 0.2398,
|
2660 |
+
"step": 10300
|
2661 |
+
},
|
2662 |
+
{
|
2663 |
+
"epoch": 2.86,
|
2664 |
+
"learning_rate": 1.6131034482758624e-05,
|
2665 |
+
"loss": 0.2526,
|
2666 |
+
"step": 10325
|
2667 |
+
},
|
2668 |
+
{
|
2669 |
+
"epoch": 2.87,
|
2670 |
+
"learning_rate": 1.6044827586206898e-05,
|
2671 |
+
"loss": 0.2263,
|
2672 |
+
"step": 10350
|
2673 |
+
},
|
2674 |
+
{
|
2675 |
+
"epoch": 2.87,
|
2676 |
+
"learning_rate": 1.5958620689655172e-05,
|
2677 |
+
"loss": 0.2527,
|
2678 |
+
"step": 10375
|
2679 |
+
},
|
2680 |
+
{
|
2681 |
+
"epoch": 2.88,
|
2682 |
+
"learning_rate": 1.587241379310345e-05,
|
2683 |
+
"loss": 0.2401,
|
2684 |
+
"step": 10400
|
2685 |
+
},
|
2686 |
+
{
|
2687 |
+
"epoch": 2.89,
|
2688 |
+
"learning_rate": 1.5786206896551724e-05,
|
2689 |
+
"loss": 0.2221,
|
2690 |
+
"step": 10425
|
2691 |
+
},
|
2692 |
+
{
|
2693 |
+
"epoch": 2.9,
|
2694 |
+
"learning_rate": 1.5700000000000002e-05,
|
2695 |
+
"loss": 0.2251,
|
2696 |
+
"step": 10450
|
2697 |
+
},
|
2698 |
+
{
|
2699 |
+
"epoch": 2.9,
|
2700 |
+
"learning_rate": 1.5613793103448276e-05,
|
2701 |
+
"loss": 0.247,
|
2702 |
+
"step": 10475
|
2703 |
+
},
|
2704 |
+
{
|
2705 |
+
"epoch": 2.91,
|
2706 |
+
"learning_rate": 1.552758620689655e-05,
|
2707 |
+
"loss": 0.2264,
|
2708 |
+
"step": 10500
|
2709 |
+
},
|
2710 |
+
{
|
2711 |
+
"epoch": 2.91,
|
2712 |
+
"eval_cer": 10.73436401967674,
|
2713 |
+
"eval_loss": 0.2812344431877136,
|
2714 |
+
"eval_runtime": 1186.9835,
|
2715 |
+
"eval_samples_per_second": 4.71,
|
2716 |
+
"eval_steps_per_second": 0.295,
|
2717 |
+
"step": 10500
|
2718 |
+
},
|
2719 |
+
{
|
2720 |
+
"epoch": 2.92,
|
2721 |
+
"learning_rate": 1.544137931034483e-05,
|
2722 |
+
"loss": 0.2459,
|
2723 |
+
"step": 10525
|
2724 |
+
},
|
2725 |
+
{
|
2726 |
+
"epoch": 2.92,
|
2727 |
+
"learning_rate": 1.5355172413793103e-05,
|
2728 |
+
"loss": 0.2541,
|
2729 |
+
"step": 10550
|
2730 |
+
},
|
2731 |
+
{
|
2732 |
+
"epoch": 2.93,
|
2733 |
+
"learning_rate": 1.526896551724138e-05,
|
2734 |
+
"loss": 0.2265,
|
2735 |
+
"step": 10575
|
2736 |
+
},
|
2737 |
+
{
|
2738 |
+
"epoch": 2.94,
|
2739 |
+
"learning_rate": 1.5182758620689655e-05,
|
2740 |
+
"loss": 0.2329,
|
2741 |
+
"step": 10600
|
2742 |
+
},
|
2743 |
+
{
|
2744 |
+
"epoch": 2.94,
|
2745 |
+
"learning_rate": 1.5096551724137931e-05,
|
2746 |
+
"loss": 0.2245,
|
2747 |
+
"step": 10625
|
2748 |
+
},
|
2749 |
+
{
|
2750 |
+
"epoch": 2.95,
|
2751 |
+
"learning_rate": 1.5010344827586207e-05,
|
2752 |
+
"loss": 0.265,
|
2753 |
+
"step": 10650
|
2754 |
+
},
|
2755 |
+
{
|
2756 |
+
"epoch": 2.96,
|
2757 |
+
"learning_rate": 1.4924137931034485e-05,
|
2758 |
+
"loss": 0.2423,
|
2759 |
+
"step": 10675
|
2760 |
+
},
|
2761 |
+
{
|
2762 |
+
"epoch": 2.96,
|
2763 |
+
"learning_rate": 1.4837931034482761e-05,
|
2764 |
+
"loss": 0.2319,
|
2765 |
+
"step": 10700
|
2766 |
+
},
|
2767 |
+
{
|
2768 |
+
"epoch": 2.97,
|
2769 |
+
"learning_rate": 1.4751724137931037e-05,
|
2770 |
+
"loss": 0.2286,
|
2771 |
+
"step": 10725
|
2772 |
+
},
|
2773 |
+
{
|
2774 |
+
"epoch": 2.98,
|
2775 |
+
"learning_rate": 1.4665517241379311e-05,
|
2776 |
+
"loss": 0.2352,
|
2777 |
+
"step": 10750
|
2778 |
+
},
|
2779 |
+
{
|
2780 |
+
"epoch": 2.99,
|
2781 |
+
"learning_rate": 1.4579310344827587e-05,
|
2782 |
+
"loss": 0.2635,
|
2783 |
+
"step": 10775
|
2784 |
+
},
|
2785 |
+
{
|
2786 |
+
"epoch": 2.99,
|
2787 |
+
"learning_rate": 1.4493103448275863e-05,
|
2788 |
+
"loss": 0.2269,
|
2789 |
+
"step": 10800
|
2790 |
+
},
|
2791 |
+
{
|
2792 |
+
"epoch": 3.0,
|
2793 |
+
"learning_rate": 1.440689655172414e-05,
|
2794 |
+
"loss": 0.2296,
|
2795 |
+
"step": 10825
|
2796 |
+
},
|
2797 |
+
{
|
2798 |
+
"epoch": 3.01,
|
2799 |
+
"learning_rate": 1.4324137931034484e-05,
|
2800 |
+
"loss": 0.1896,
|
2801 |
+
"step": 10850
|
2802 |
+
},
|
2803 |
+
{
|
2804 |
+
"epoch": 3.01,
|
2805 |
+
"learning_rate": 1.423793103448276e-05,
|
2806 |
+
"loss": 0.1824,
|
2807 |
+
"step": 10875
|
2808 |
+
},
|
2809 |
+
{
|
2810 |
+
"epoch": 3.02,
|
2811 |
+
"learning_rate": 1.4151724137931036e-05,
|
2812 |
+
"loss": 0.1621,
|
2813 |
+
"step": 10900
|
2814 |
+
},
|
2815 |
+
{
|
2816 |
+
"epoch": 3.03,
|
2817 |
+
"learning_rate": 1.4065517241379312e-05,
|
2818 |
+
"loss": 0.1497,
|
2819 |
+
"step": 10925
|
2820 |
+
},
|
2821 |
+
{
|
2822 |
+
"epoch": 3.03,
|
2823 |
+
"learning_rate": 1.3979310344827586e-05,
|
2824 |
+
"loss": 0.1597,
|
2825 |
+
"step": 10950
|
2826 |
+
},
|
2827 |
+
{
|
2828 |
+
"epoch": 3.04,
|
2829 |
+
"learning_rate": 1.3893103448275862e-05,
|
2830 |
+
"loss": 0.1642,
|
2831 |
+
"step": 10975
|
2832 |
+
},
|
2833 |
+
{
|
2834 |
+
"epoch": 3.05,
|
2835 |
+
"learning_rate": 1.3806896551724138e-05,
|
2836 |
+
"loss": 0.1468,
|
2837 |
+
"step": 11000
|
2838 |
+
},
|
2839 |
+
{
|
2840 |
+
"epoch": 3.05,
|
2841 |
+
"eval_cer": 10.796652398901168,
|
2842 |
+
"eval_loss": 0.2832615077495575,
|
2843 |
+
"eval_runtime": 1183.0471,
|
2844 |
+
"eval_samples_per_second": 4.726,
|
2845 |
+
"eval_steps_per_second": 0.296,
|
2846 |
+
"step": 11000
|
2847 |
+
},
|
2848 |
+
{
|
2849 |
+
"epoch": 3.05,
|
2850 |
+
"learning_rate": 1.3720689655172414e-05,
|
2851 |
+
"loss": 0.1427,
|
2852 |
+
"step": 11025
|
2853 |
+
},
|
2854 |
+
{
|
2855 |
+
"epoch": 3.06,
|
2856 |
+
"learning_rate": 1.363448275862069e-05,
|
2857 |
+
"loss": 0.17,
|
2858 |
+
"step": 11050
|
2859 |
+
},
|
2860 |
+
{
|
2861 |
+
"epoch": 3.07,
|
2862 |
+
"learning_rate": 1.3548275862068965e-05,
|
2863 |
+
"loss": 0.1585,
|
2864 |
+
"step": 11075
|
2865 |
+
},
|
2866 |
+
{
|
2867 |
+
"epoch": 3.08,
|
2868 |
+
"learning_rate": 1.3462068965517241e-05,
|
2869 |
+
"loss": 0.1727,
|
2870 |
+
"step": 11100
|
2871 |
+
},
|
2872 |
+
{
|
2873 |
+
"epoch": 3.08,
|
2874 |
+
"learning_rate": 1.3375862068965517e-05,
|
2875 |
+
"loss": 0.1654,
|
2876 |
+
"step": 11125
|
2877 |
+
},
|
2878 |
+
{
|
2879 |
+
"epoch": 3.09,
|
2880 |
+
"learning_rate": 1.3289655172413793e-05,
|
2881 |
+
"loss": 0.1517,
|
2882 |
+
"step": 11150
|
2883 |
+
},
|
2884 |
+
{
|
2885 |
+
"epoch": 3.1,
|
2886 |
+
"learning_rate": 1.320344827586207e-05,
|
2887 |
+
"loss": 0.1536,
|
2888 |
+
"step": 11175
|
2889 |
+
},
|
2890 |
+
{
|
2891 |
+
"epoch": 3.1,
|
2892 |
+
"learning_rate": 1.3117241379310347e-05,
|
2893 |
+
"loss": 0.1692,
|
2894 |
+
"step": 11200
|
2895 |
+
},
|
2896 |
+
{
|
2897 |
+
"epoch": 3.11,
|
2898 |
+
"learning_rate": 1.3031034482758623e-05,
|
2899 |
+
"loss": 0.1673,
|
2900 |
+
"step": 11225
|
2901 |
+
},
|
2902 |
+
{
|
2903 |
+
"epoch": 3.12,
|
2904 |
+
"learning_rate": 1.2944827586206897e-05,
|
2905 |
+
"loss": 0.1554,
|
2906 |
+
"step": 11250
|
2907 |
+
},
|
2908 |
+
{
|
2909 |
+
"epoch": 3.12,
|
2910 |
+
"learning_rate": 1.2858620689655173e-05,
|
2911 |
+
"loss": 0.1616,
|
2912 |
+
"step": 11275
|
2913 |
+
},
|
2914 |
+
{
|
2915 |
+
"epoch": 3.13,
|
2916 |
+
"learning_rate": 1.277241379310345e-05,
|
2917 |
+
"loss": 0.1553,
|
2918 |
+
"step": 11300
|
2919 |
+
},
|
2920 |
+
{
|
2921 |
+
"epoch": 3.14,
|
2922 |
+
"learning_rate": 1.2686206896551725e-05,
|
2923 |
+
"loss": 0.1692,
|
2924 |
+
"step": 11325
|
2925 |
+
},
|
2926 |
+
{
|
2927 |
+
"epoch": 3.14,
|
2928 |
+
"learning_rate": 1.2600000000000001e-05,
|
2929 |
+
"loss": 0.1582,
|
2930 |
+
"step": 11350
|
2931 |
+
},
|
2932 |
+
{
|
2933 |
+
"epoch": 3.15,
|
2934 |
+
"learning_rate": 1.2513793103448276e-05,
|
2935 |
+
"loss": 0.1634,
|
2936 |
+
"step": 11375
|
2937 |
+
},
|
2938 |
+
{
|
2939 |
+
"epoch": 3.16,
|
2940 |
+
"learning_rate": 1.2427586206896552e-05,
|
2941 |
+
"loss": 0.1584,
|
2942 |
+
"step": 11400
|
2943 |
+
},
|
2944 |
+
{
|
2945 |
+
"epoch": 3.17,
|
2946 |
+
"learning_rate": 1.2341379310344828e-05,
|
2947 |
+
"loss": 0.1678,
|
2948 |
+
"step": 11425
|
2949 |
+
},
|
2950 |
+
{
|
2951 |
+
"epoch": 3.17,
|
2952 |
+
"learning_rate": 1.2255172413793104e-05,
|
2953 |
+
"loss": 0.1767,
|
2954 |
+
"step": 11450
|
2955 |
+
},
|
2956 |
+
{
|
2957 |
+
"epoch": 3.18,
|
2958 |
+
"learning_rate": 1.216896551724138e-05,
|
2959 |
+
"loss": 0.1561,
|
2960 |
+
"step": 11475
|
2961 |
+
},
|
2962 |
+
{
|
2963 |
+
"epoch": 3.19,
|
2964 |
+
"learning_rate": 1.2082758620689656e-05,
|
2965 |
+
"loss": 0.1557,
|
2966 |
+
"step": 11500
|
2967 |
+
},
|
2968 |
+
{
|
2969 |
+
"epoch": 3.19,
|
2970 |
+
"eval_cer": 10.533124640643967,
|
2971 |
+
"eval_loss": 0.28545621037483215,
|
2972 |
+
"eval_runtime": 1176.7108,
|
2973 |
+
"eval_samples_per_second": 4.751,
|
2974 |
+
"eval_steps_per_second": 0.297,
|
2975 |
+
"step": 11500
|
2976 |
+
},
|
2977 |
+
{
|
2978 |
+
"epoch": 3.19,
|
2979 |
+
"learning_rate": 1.1996551724137932e-05,
|
2980 |
+
"loss": 0.172,
|
2981 |
+
"step": 11525
|
2982 |
+
},
|
2983 |
+
{
|
2984 |
+
"epoch": 3.2,
|
2985 |
+
"learning_rate": 1.1910344827586208e-05,
|
2986 |
+
"loss": 0.1571,
|
2987 |
+
"step": 11550
|
2988 |
+
},
|
2989 |
+
{
|
2990 |
+
"epoch": 3.21,
|
2991 |
+
"learning_rate": 1.1824137931034484e-05,
|
2992 |
+
"loss": 0.1605,
|
2993 |
+
"step": 11575
|
2994 |
+
},
|
2995 |
+
{
|
2996 |
+
"epoch": 3.21,
|
2997 |
+
"learning_rate": 1.1737931034482758e-05,
|
2998 |
+
"loss": 0.1612,
|
2999 |
+
"step": 11600
|
3000 |
+
},
|
3001 |
+
{
|
3002 |
+
"epoch": 3.22,
|
3003 |
+
"learning_rate": 1.1651724137931034e-05,
|
3004 |
+
"loss": 0.1615,
|
3005 |
+
"step": 11625
|
3006 |
+
},
|
3007 |
+
{
|
3008 |
+
"epoch": 3.23,
|
3009 |
+
"learning_rate": 1.156551724137931e-05,
|
3010 |
+
"loss": 0.1627,
|
3011 |
+
"step": 11650
|
3012 |
+
},
|
3013 |
+
{
|
3014 |
+
"epoch": 3.23,
|
3015 |
+
"learning_rate": 1.1479310344827588e-05,
|
3016 |
+
"loss": 0.1481,
|
3017 |
+
"step": 11675
|
3018 |
+
},
|
3019 |
+
{
|
3020 |
+
"epoch": 3.24,
|
3021 |
+
"learning_rate": 1.1393103448275863e-05,
|
3022 |
+
"loss": 0.1532,
|
3023 |
+
"step": 11700
|
3024 |
+
},
|
3025 |
+
{
|
3026 |
+
"epoch": 3.25,
|
3027 |
+
"learning_rate": 1.1306896551724139e-05,
|
3028 |
+
"loss": 0.1358,
|
3029 |
+
"step": 11725
|
3030 |
+
},
|
3031 |
+
{
|
3032 |
+
"epoch": 3.26,
|
3033 |
+
"learning_rate": 1.1220689655172415e-05,
|
3034 |
+
"loss": 0.1619,
|
3035 |
+
"step": 11750
|
3036 |
+
},
|
3037 |
+
{
|
3038 |
+
"epoch": 3.26,
|
3039 |
+
"learning_rate": 1.113448275862069e-05,
|
3040 |
+
"loss": 0.1721,
|
3041 |
+
"step": 11775
|
3042 |
+
},
|
3043 |
+
{
|
3044 |
+
"epoch": 3.27,
|
3045 |
+
"learning_rate": 1.1048275862068965e-05,
|
3046 |
+
"loss": 0.1662,
|
3047 |
+
"step": 11800
|
3048 |
+
},
|
3049 |
+
{
|
3050 |
+
"epoch": 3.28,
|
3051 |
+
"learning_rate": 1.0962068965517241e-05,
|
3052 |
+
"loss": 0.1511,
|
3053 |
+
"step": 11825
|
3054 |
+
},
|
3055 |
+
{
|
3056 |
+
"epoch": 3.28,
|
3057 |
+
"learning_rate": 1.0875862068965517e-05,
|
3058 |
+
"loss": 0.1554,
|
3059 |
+
"step": 11850
|
3060 |
+
},
|
3061 |
+
{
|
3062 |
+
"epoch": 3.29,
|
3063 |
+
"learning_rate": 1.0789655172413795e-05,
|
3064 |
+
"loss": 0.1767,
|
3065 |
+
"step": 11875
|
3066 |
+
},
|
3067 |
+
{
|
3068 |
+
"epoch": 3.3,
|
3069 |
+
"learning_rate": 1.070344827586207e-05,
|
3070 |
+
"loss": 0.1613,
|
3071 |
+
"step": 11900
|
3072 |
+
},
|
3073 |
+
{
|
3074 |
+
"epoch": 3.3,
|
3075 |
+
"learning_rate": 1.0617241379310345e-05,
|
3076 |
+
"loss": 0.1752,
|
3077 |
+
"step": 11925
|
3078 |
+
},
|
3079 |
+
{
|
3080 |
+
"epoch": 3.31,
|
3081 |
+
"learning_rate": 1.0531034482758621e-05,
|
3082 |
+
"loss": 0.1493,
|
3083 |
+
"step": 11950
|
3084 |
+
},
|
3085 |
+
{
|
3086 |
+
"epoch": 3.32,
|
3087 |
+
"learning_rate": 1.0444827586206897e-05,
|
3088 |
+
"loss": 0.1927,
|
3089 |
+
"step": 11975
|
3090 |
+
},
|
3091 |
+
{
|
3092 |
+
"epoch": 3.32,
|
3093 |
+
"learning_rate": 1.0358620689655173e-05,
|
3094 |
+
"loss": 0.1639,
|
3095 |
+
"step": 12000
|
3096 |
+
},
|
3097 |
+
{
|
3098 |
+
"epoch": 3.32,
|
3099 |
+
"eval_cer": 10.167380054941546,
|
3100 |
+
"eval_loss": 0.28585749864578247,
|
3101 |
+
"eval_runtime": 1168.3755,
|
3102 |
+
"eval_samples_per_second": 4.785,
|
3103 |
+
"eval_steps_per_second": 0.3,
|
3104 |
+
"step": 12000
|
3105 |
+
},
|
3106 |
+
{
|
3107 |
+
"epoch": 3.33,
|
3108 |
+
"learning_rate": 1.0272413793103448e-05,
|
3109 |
+
"loss": 0.1708,
|
3110 |
+
"step": 12025
|
3111 |
+
},
|
3112 |
+
{
|
3113 |
+
"epoch": 3.34,
|
3114 |
+
"learning_rate": 1.0186206896551724e-05,
|
3115 |
+
"loss": 0.1692,
|
3116 |
+
"step": 12050
|
3117 |
+
},
|
3118 |
+
{
|
3119 |
+
"epoch": 3.35,
|
3120 |
+
"learning_rate": 1.0100000000000002e-05,
|
3121 |
+
"loss": 0.1754,
|
3122 |
+
"step": 12075
|
3123 |
+
},
|
3124 |
+
{
|
3125 |
+
"epoch": 3.35,
|
3126 |
+
"learning_rate": 1.0013793103448278e-05,
|
3127 |
+
"loss": 0.1695,
|
3128 |
+
"step": 12100
|
3129 |
+
},
|
3130 |
+
{
|
3131 |
+
"epoch": 3.36,
|
3132 |
+
"learning_rate": 9.927586206896552e-06,
|
3133 |
+
"loss": 0.1695,
|
3134 |
+
"step": 12125
|
3135 |
+
},
|
3136 |
+
{
|
3137 |
+
"epoch": 3.37,
|
3138 |
+
"learning_rate": 9.841379310344828e-06,
|
3139 |
+
"loss": 0.1718,
|
3140 |
+
"step": 12150
|
3141 |
+
},
|
3142 |
+
{
|
3143 |
+
"epoch": 3.37,
|
3144 |
+
"learning_rate": 9.755172413793104e-06,
|
3145 |
+
"loss": 0.151,
|
3146 |
+
"step": 12175
|
3147 |
+
},
|
3148 |
+
{
|
3149 |
+
"epoch": 3.38,
|
3150 |
+
"learning_rate": 9.66896551724138e-06,
|
3151 |
+
"loss": 0.1672,
|
3152 |
+
"step": 12200
|
3153 |
+
},
|
3154 |
+
{
|
3155 |
+
"epoch": 3.39,
|
3156 |
+
"learning_rate": 9.582758620689654e-06,
|
3157 |
+
"loss": 0.1665,
|
3158 |
+
"step": 12225
|
3159 |
+
},
|
3160 |
+
{
|
3161 |
+
"epoch": 3.39,
|
3162 |
+
"learning_rate": 9.49655172413793e-06,
|
3163 |
+
"loss": 0.1567,
|
3164 |
+
"step": 12250
|
3165 |
+
},
|
3166 |
+
{
|
3167 |
+
"epoch": 3.4,
|
3168 |
+
"learning_rate": 9.410344827586208e-06,
|
3169 |
+
"loss": 0.1691,
|
3170 |
+
"step": 12275
|
3171 |
+
},
|
3172 |
+
{
|
3173 |
+
"epoch": 3.41,
|
3174 |
+
"learning_rate": 9.324137931034484e-06,
|
3175 |
+
"loss": 0.1648,
|
3176 |
+
"step": 12300
|
3177 |
+
},
|
3178 |
+
{
|
3179 |
+
"epoch": 3.42,
|
3180 |
+
"learning_rate": 9.237931034482759e-06,
|
3181 |
+
"loss": 0.1576,
|
3182 |
+
"step": 12325
|
3183 |
+
},
|
3184 |
+
{
|
3185 |
+
"epoch": 3.42,
|
3186 |
+
"learning_rate": 9.151724137931035e-06,
|
3187 |
+
"loss": 0.1689,
|
3188 |
+
"step": 12350
|
3189 |
+
},
|
3190 |
+
{
|
3191 |
+
"epoch": 3.43,
|
3192 |
+
"learning_rate": 9.06551724137931e-06,
|
3193 |
+
"loss": 0.159,
|
3194 |
+
"step": 12375
|
3195 |
+
},
|
3196 |
+
{
|
3197 |
+
"epoch": 3.44,
|
3198 |
+
"learning_rate": 8.979310344827587e-06,
|
3199 |
+
"loss": 0.1682,
|
3200 |
+
"step": 12400
|
3201 |
+
},
|
3202 |
+
{
|
3203 |
+
"epoch": 3.44,
|
3204 |
+
"learning_rate": 8.893103448275863e-06,
|
3205 |
+
"loss": 0.1729,
|
3206 |
+
"step": 12425
|
3207 |
+
},
|
3208 |
+
{
|
3209 |
+
"epoch": 3.45,
|
3210 |
+
"learning_rate": 8.806896551724137e-06,
|
3211 |
+
"loss": 0.168,
|
3212 |
+
"step": 12450
|
3213 |
+
},
|
3214 |
+
{
|
3215 |
+
"epoch": 3.46,
|
3216 |
+
"learning_rate": 8.720689655172415e-06,
|
3217 |
+
"loss": 0.1704,
|
3218 |
+
"step": 12475
|
3219 |
+
},
|
3220 |
+
{
|
3221 |
+
"epoch": 3.46,
|
3222 |
+
"learning_rate": 8.634482758620691e-06,
|
3223 |
+
"loss": 0.1603,
|
3224 |
+
"step": 12500
|
3225 |
+
},
|
3226 |
+
{
|
3227 |
+
"epoch": 3.46,
|
3228 |
+
"eval_cer": 10.212099916948828,
|
3229 |
+
"eval_loss": 0.2885717749595642,
|
3230 |
+
"eval_runtime": 1168.0562,
|
3231 |
+
"eval_samples_per_second": 4.787,
|
3232 |
+
"eval_steps_per_second": 0.3,
|
3233 |
+
"step": 12500
|
3234 |
+
},
|
3235 |
+
{
|
3236 |
+
"epoch": 3.47,
|
3237 |
+
"learning_rate": 8.548275862068967e-06,
|
3238 |
+
"loss": 0.1564,
|
3239 |
+
"step": 12525
|
3240 |
+
},
|
3241 |
+
{
|
3242 |
+
"epoch": 3.48,
|
3243 |
+
"learning_rate": 8.462068965517241e-06,
|
3244 |
+
"loss": 0.1511,
|
3245 |
+
"step": 12550
|
3246 |
+
},
|
3247 |
+
{
|
3248 |
+
"epoch": 3.48,
|
3249 |
+
"learning_rate": 8.375862068965517e-06,
|
3250 |
+
"loss": 0.1575,
|
3251 |
+
"step": 12575
|
3252 |
+
},
|
3253 |
+
{
|
3254 |
+
"epoch": 3.49,
|
3255 |
+
"learning_rate": 8.289655172413793e-06,
|
3256 |
+
"loss": 0.1574,
|
3257 |
+
"step": 12600
|
3258 |
+
},
|
3259 |
+
{
|
3260 |
+
"epoch": 3.5,
|
3261 |
+
"learning_rate": 8.20344827586207e-06,
|
3262 |
+
"loss": 0.1673,
|
3263 |
+
"step": 12625
|
3264 |
+
},
|
3265 |
+
{
|
3266 |
+
"epoch": 3.51,
|
3267 |
+
"learning_rate": 8.117241379310346e-06,
|
3268 |
+
"loss": 0.1605,
|
3269 |
+
"step": 12650
|
3270 |
+
},
|
3271 |
+
{
|
3272 |
+
"epoch": 3.51,
|
3273 |
+
"learning_rate": 8.031034482758622e-06,
|
3274 |
+
"loss": 0.1562,
|
3275 |
+
"step": 12675
|
3276 |
+
},
|
3277 |
+
{
|
3278 |
+
"epoch": 3.52,
|
3279 |
+
"learning_rate": 7.944827586206898e-06,
|
3280 |
+
"loss": 0.1629,
|
3281 |
+
"step": 12700
|
3282 |
+
},
|
3283 |
+
{
|
3284 |
+
"epoch": 3.53,
|
3285 |
+
"learning_rate": 7.858620689655174e-06,
|
3286 |
+
"loss": 0.1629,
|
3287 |
+
"step": 12725
|
3288 |
+
},
|
3289 |
+
{
|
3290 |
+
"epoch": 3.53,
|
3291 |
+
"learning_rate": 7.772413793103448e-06,
|
3292 |
+
"loss": 0.1849,
|
3293 |
+
"step": 12750
|
3294 |
+
},
|
3295 |
+
{
|
3296 |
+
"epoch": 3.54,
|
3297 |
+
"learning_rate": 7.686206896551724e-06,
|
3298 |
+
"loss": 0.1501,
|
3299 |
+
"step": 12775
|
3300 |
+
},
|
3301 |
+
{
|
3302 |
+
"epoch": 3.55,
|
3303 |
+
"learning_rate": 7.6e-06,
|
3304 |
+
"loss": 0.1563,
|
3305 |
+
"step": 12800
|
3306 |
+
},
|
3307 |
+
{
|
3308 |
+
"epoch": 3.55,
|
3309 |
+
"learning_rate": 7.513793103448276e-06,
|
3310 |
+
"loss": 0.1596,
|
3311 |
+
"step": 12825
|
3312 |
+
},
|
3313 |
+
{
|
3314 |
+
"epoch": 3.56,
|
3315 |
+
"learning_rate": 7.427586206896551e-06,
|
3316 |
+
"loss": 0.156,
|
3317 |
+
"step": 12850
|
3318 |
+
},
|
3319 |
+
{
|
3320 |
+
"epoch": 3.57,
|
3321 |
+
"learning_rate": 7.341379310344828e-06,
|
3322 |
+
"loss": 0.1655,
|
3323 |
+
"step": 12875
|
3324 |
+
},
|
3325 |
+
{
|
3326 |
+
"epoch": 3.57,
|
3327 |
+
"learning_rate": 7.255172413793104e-06,
|
3328 |
+
"loss": 0.1642,
|
3329 |
+
"step": 12900
|
3330 |
+
},
|
3331 |
+
{
|
3332 |
+
"epoch": 3.58,
|
3333 |
+
"learning_rate": 7.16896551724138e-06,
|
3334 |
+
"loss": 0.1676,
|
3335 |
+
"step": 12925
|
3336 |
+
},
|
3337 |
+
{
|
3338 |
+
"epoch": 3.59,
|
3339 |
+
"learning_rate": 7.0827586206896555e-06,
|
3340 |
+
"loss": 0.149,
|
3341 |
+
"step": 12950
|
3342 |
+
},
|
3343 |
+
{
|
3344 |
+
"epoch": 3.6,
|
3345 |
+
"learning_rate": 6.9965517241379315e-06,
|
3346 |
+
"loss": 0.1734,
|
3347 |
+
"step": 12975
|
3348 |
+
},
|
3349 |
+
{
|
3350 |
+
"epoch": 3.6,
|
3351 |
+
"learning_rate": 6.910344827586207e-06,
|
3352 |
+
"loss": 0.1599,
|
3353 |
+
"step": 13000
|
3354 |
+
},
|
3355 |
+
{
|
3356 |
+
"epoch": 3.6,
|
3357 |
+
"eval_cer": 10.058774675781,
|
3358 |
+
"eval_loss": 0.2872118651866913,
|
3359 |
+
"eval_runtime": 1162.7195,
|
3360 |
+
"eval_samples_per_second": 4.809,
|
3361 |
+
"eval_steps_per_second": 0.301,
|
3362 |
+
"step": 13000
|
3363 |
}
|
3364 |
],
|
3365 |
+
"max_steps": 15000,
|
3366 |
+
"num_train_epochs": 5,
|
3367 |
+
"total_flos": 3.7516448344080384e+20,
|
3368 |
"trial_name": null,
|
3369 |
"trial_params": null
|
3370 |
}
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a475f6ee95be6b057d14b9fb168421fcdb786f15d44fa584423c26f08dc470e
|
3 |
+
size 3643
|