anuragshas
commited on
Commit
•
7faa6b8
1
Parent(s):
7a43c3b
End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +427 -0
all_results.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 60.0,
|
3 |
+
"eval_loss": 0.2578442692756653,
|
4 |
+
"eval_runtime": 33.4828,
|
5 |
+
"eval_samples": 1193,
|
6 |
+
"eval_samples_per_second": 35.63,
|
7 |
+
"eval_steps_per_second": 2.24,
|
8 |
+
"eval_wer": 0.22734592299809692,
|
9 |
+
"train_loss": 1.5886853283982936,
|
10 |
+
"train_runtime": 6665.2028,
|
11 |
+
"train_samples": 2606,
|
12 |
+
"train_samples_per_second": 23.459,
|
13 |
+
"train_steps_per_second": 0.738
|
14 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 60.0,
|
3 |
+
"eval_loss": 0.2578442692756653,
|
4 |
+
"eval_runtime": 33.4828,
|
5 |
+
"eval_samples": 1193,
|
6 |
+
"eval_samples_per_second": 35.63,
|
7 |
+
"eval_steps_per_second": 2.24,
|
8 |
+
"eval_wer": 0.22734592299809692
|
9 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 60.0,
|
3 |
+
"train_loss": 1.5886853283982936,
|
4 |
+
"train_runtime": 6665.2028,
|
5 |
+
"train_samples": 2606,
|
6 |
+
"train_samples_per_second": 23.459,
|
7 |
+
"train_steps_per_second": 0.738
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,427 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 60.0,
|
5 |
+
"global_step": 4920,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 1.22,
|
12 |
+
"learning_rate": 7.35e-06,
|
13 |
+
"loss": 12.2749,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 2.44,
|
18 |
+
"learning_rate": 1.485e-05,
|
19 |
+
"loss": 4.8071,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 3.66,
|
24 |
+
"learning_rate": 2.2349999999999998e-05,
|
25 |
+
"loss": 3.5841,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 4.88,
|
30 |
+
"learning_rate": 2.985e-05,
|
31 |
+
"loss": 3.1829,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 4.88,
|
36 |
+
"eval_loss": 3.1227903366088867,
|
37 |
+
"eval_runtime": 33.8472,
|
38 |
+
"eval_samples_per_second": 35.247,
|
39 |
+
"eval_steps_per_second": 2.216,
|
40 |
+
"eval_wer": 1.0,
|
41 |
+
"step": 400
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 6.1,
|
45 |
+
"learning_rate": 3.735e-05,
|
46 |
+
"loss": 3.0323,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 7.32,
|
51 |
+
"learning_rate": 4.484999999999999e-05,
|
52 |
+
"loss": 2.9603,
|
53 |
+
"step": 600
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 8.54,
|
57 |
+
"learning_rate": 5.234999999999999e-05,
|
58 |
+
"loss": 2.8982,
|
59 |
+
"step": 700
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 9.76,
|
63 |
+
"learning_rate": 5.985e-05,
|
64 |
+
"loss": 2.8675,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 9.76,
|
69 |
+
"eval_loss": 2.8616490364074707,
|
70 |
+
"eval_runtime": 33.8201,
|
71 |
+
"eval_samples_per_second": 35.275,
|
72 |
+
"eval_steps_per_second": 2.218,
|
73 |
+
"eval_wer": 0.9992680427463037,
|
74 |
+
"step": 800
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 10.98,
|
78 |
+
"learning_rate": 6.735e-05,
|
79 |
+
"loss": 2.7893,
|
80 |
+
"step": 900
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 12.2,
|
84 |
+
"learning_rate": 7.484999999999999e-05,
|
85 |
+
"loss": 2.2812,
|
86 |
+
"step": 1000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 13.41,
|
90 |
+
"learning_rate": 7.312499999999999e-05,
|
91 |
+
"loss": 1.7369,
|
92 |
+
"step": 1100
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 14.63,
|
96 |
+
"learning_rate": 7.121173469387755e-05,
|
97 |
+
"loss": 1.583,
|
98 |
+
"step": 1200
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 14.63,
|
102 |
+
"eval_loss": 0.6392253637313843,
|
103 |
+
"eval_runtime": 33.9553,
|
104 |
+
"eval_samples_per_second": 35.134,
|
105 |
+
"eval_steps_per_second": 2.209,
|
106 |
+
"eval_wer": 0.6239203630507978,
|
107 |
+
"step": 1200
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 15.85,
|
111 |
+
"learning_rate": 6.92984693877551e-05,
|
112 |
+
"loss": 1.4572,
|
113 |
+
"step": 1300
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 17.07,
|
117 |
+
"learning_rate": 6.738520408163265e-05,
|
118 |
+
"loss": 1.3366,
|
119 |
+
"step": 1400
|
120 |
+
},
|
121 |
+
{
|
122 |
+
"epoch": 18.29,
|
123 |
+
"learning_rate": 6.547193877551019e-05,
|
124 |
+
"loss": 1.2668,
|
125 |
+
"step": 1500
|
126 |
+
},
|
127 |
+
{
|
128 |
+
"epoch": 19.51,
|
129 |
+
"learning_rate": 6.355867346938775e-05,
|
130 |
+
"loss": 1.1959,
|
131 |
+
"step": 1600
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 19.51,
|
135 |
+
"eval_loss": 0.36018967628479004,
|
136 |
+
"eval_runtime": 33.3624,
|
137 |
+
"eval_samples_per_second": 35.759,
|
138 |
+
"eval_steps_per_second": 2.248,
|
139 |
+
"eval_wer": 0.3651002781437564,
|
140 |
+
"step": 1600
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 20.73,
|
144 |
+
"learning_rate": 6.164540816326531e-05,
|
145 |
+
"loss": 1.1694,
|
146 |
+
"step": 1700
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 21.95,
|
150 |
+
"learning_rate": 5.9732142857142846e-05,
|
151 |
+
"loss": 1.1001,
|
152 |
+
"step": 1800
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 23.17,
|
156 |
+
"learning_rate": 5.7818877551020404e-05,
|
157 |
+
"loss": 1.0737,
|
158 |
+
"step": 1900
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 24.39,
|
162 |
+
"learning_rate": 5.5905612244897955e-05,
|
163 |
+
"loss": 1.0276,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 24.39,
|
168 |
+
"eval_loss": 0.3020930588245392,
|
169 |
+
"eval_runtime": 33.6823,
|
170 |
+
"eval_samples_per_second": 35.419,
|
171 |
+
"eval_steps_per_second": 2.227,
|
172 |
+
"eval_wer": 0.29805299370516763,
|
173 |
+
"step": 2000
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 25.61,
|
177 |
+
"learning_rate": 5.3992346938775506e-05,
|
178 |
+
"loss": 1.0293,
|
179 |
+
"step": 2100
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 26.83,
|
183 |
+
"learning_rate": 5.207908163265306e-05,
|
184 |
+
"loss": 0.9966,
|
185 |
+
"step": 2200
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 28.05,
|
189 |
+
"learning_rate": 5.016581632653061e-05,
|
190 |
+
"loss": 0.9721,
|
191 |
+
"step": 2300
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 29.27,
|
195 |
+
"learning_rate": 4.8271683673469385e-05,
|
196 |
+
"loss": 0.9671,
|
197 |
+
"step": 2400
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 29.27,
|
201 |
+
"eval_loss": 0.28717485070228577,
|
202 |
+
"eval_runtime": 33.2455,
|
203 |
+
"eval_samples_per_second": 35.885,
|
204 |
+
"eval_steps_per_second": 2.256,
|
205 |
+
"eval_wer": 0.27389840433318696,
|
206 |
+
"step": 2400
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 30.49,
|
210 |
+
"learning_rate": 4.635841836734693e-05,
|
211 |
+
"loss": 0.9536,
|
212 |
+
"step": 2500
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 31.71,
|
216 |
+
"learning_rate": 4.444515306122448e-05,
|
217 |
+
"loss": 0.934,
|
218 |
+
"step": 2600
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 32.93,
|
222 |
+
"learning_rate": 4.253188775510204e-05,
|
223 |
+
"loss": 0.9136,
|
224 |
+
"step": 2700
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 34.15,
|
228 |
+
"learning_rate": 4.061862244897959e-05,
|
229 |
+
"loss": 0.873,
|
230 |
+
"step": 2800
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 34.15,
|
234 |
+
"eval_loss": 0.259307324886322,
|
235 |
+
"eval_runtime": 33.4272,
|
236 |
+
"eval_samples_per_second": 35.689,
|
237 |
+
"eval_steps_per_second": 2.244,
|
238 |
+
"eval_wer": 0.24593763724198506,
|
239 |
+
"step": 2800
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 35.37,
|
243 |
+
"learning_rate": 3.870535714285714e-05,
|
244 |
+
"loss": 0.8649,
|
245 |
+
"step": 2900
|
246 |
+
},
|
247 |
+
{
|
248 |
+
"epoch": 36.59,
|
249 |
+
"learning_rate": 3.679209183673469e-05,
|
250 |
+
"loss": 0.8464,
|
251 |
+
"step": 3000
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 37.8,
|
255 |
+
"learning_rate": 3.487882653061224e-05,
|
256 |
+
"loss": 0.8545,
|
257 |
+
"step": 3100
|
258 |
+
},
|
259 |
+
{
|
260 |
+
"epoch": 39.02,
|
261 |
+
"learning_rate": 3.2965561224489794e-05,
|
262 |
+
"loss": 0.8513,
|
263 |
+
"step": 3200
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 39.02,
|
267 |
+
"eval_loss": 0.2617344856262207,
|
268 |
+
"eval_runtime": 33.3367,
|
269 |
+
"eval_samples_per_second": 35.786,
|
270 |
+
"eval_steps_per_second": 2.25,
|
271 |
+
"eval_wer": 0.24725516029863856,
|
272 |
+
"step": 3200
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 40.24,
|
276 |
+
"learning_rate": 3.1052295918367345e-05,
|
277 |
+
"loss": 0.8424,
|
278 |
+
"step": 3300
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 41.46,
|
282 |
+
"learning_rate": 2.9139030612244896e-05,
|
283 |
+
"loss": 0.8244,
|
284 |
+
"step": 3400
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 42.68,
|
288 |
+
"learning_rate": 2.7225765306122448e-05,
|
289 |
+
"loss": 0.8157,
|
290 |
+
"step": 3500
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 43.9,
|
294 |
+
"learning_rate": 2.53125e-05,
|
295 |
+
"loss": 0.8132,
|
296 |
+
"step": 3600
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 43.9,
|
300 |
+
"eval_loss": 0.25475597381591797,
|
301 |
+
"eval_runtime": 33.5269,
|
302 |
+
"eval_samples_per_second": 35.583,
|
303 |
+
"eval_steps_per_second": 2.237,
|
304 |
+
"eval_wer": 0.2425706338749817,
|
305 |
+
"step": 3600
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 45.12,
|
309 |
+
"learning_rate": 2.339923469387755e-05,
|
310 |
+
"loss": 0.8008,
|
311 |
+
"step": 3700
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 46.34,
|
315 |
+
"learning_rate": 2.14859693877551e-05,
|
316 |
+
"loss": 0.7863,
|
317 |
+
"step": 3800
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 47.56,
|
321 |
+
"learning_rate": 1.9572704081632652e-05,
|
322 |
+
"loss": 0.7965,
|
323 |
+
"step": 3900
|
324 |
+
},
|
325 |
+
{
|
326 |
+
"epoch": 48.78,
|
327 |
+
"learning_rate": 1.7659438775510203e-05,
|
328 |
+
"loss": 0.7935,
|
329 |
+
"step": 4000
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 48.78,
|
333 |
+
"eval_loss": 0.2637428939342499,
|
334 |
+
"eval_runtime": 33.4727,
|
335 |
+
"eval_samples_per_second": 35.641,
|
336 |
+
"eval_steps_per_second": 2.241,
|
337 |
+
"eval_wer": 0.23525106133801785,
|
338 |
+
"step": 4000
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 50.0,
|
342 |
+
"learning_rate": 1.5746173469387754e-05,
|
343 |
+
"loss": 0.7857,
|
344 |
+
"step": 4100
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 51.22,
|
348 |
+
"learning_rate": 1.3832908163265304e-05,
|
349 |
+
"loss": 0.7834,
|
350 |
+
"step": 4200
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 52.44,
|
354 |
+
"learning_rate": 1.1919642857142855e-05,
|
355 |
+
"loss": 0.7732,
|
356 |
+
"step": 4300
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 53.66,
|
360 |
+
"learning_rate": 1.0006377551020406e-05,
|
361 |
+
"loss": 0.7565,
|
362 |
+
"step": 4400
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 53.66,
|
366 |
+
"eval_loss": 0.2629481256008148,
|
367 |
+
"eval_runtime": 33.9009,
|
368 |
+
"eval_samples_per_second": 35.191,
|
369 |
+
"eval_steps_per_second": 2.212,
|
370 |
+
"eval_wer": 0.23217684087249305,
|
371 |
+
"step": 4400
|
372 |
+
},
|
373 |
+
{
|
374 |
+
"epoch": 54.88,
|
375 |
+
"learning_rate": 8.093112244897959e-06,
|
376 |
+
"loss": 0.7525,
|
377 |
+
"step": 4500
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 56.1,
|
381 |
+
"learning_rate": 6.198979591836734e-06,
|
382 |
+
"loss": 0.7716,
|
383 |
+
"step": 4600
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 57.32,
|
387 |
+
"learning_rate": 4.2857142857142855e-06,
|
388 |
+
"loss": 0.7521,
|
389 |
+
"step": 4700
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 58.54,
|
393 |
+
"learning_rate": 2.372448979591836e-06,
|
394 |
+
"loss": 0.7359,
|
395 |
+
"step": 4800
|
396 |
+
},
|
397 |
+
{
|
398 |
+
"epoch": 58.54,
|
399 |
+
"eval_loss": 0.2578625977039337,
|
400 |
+
"eval_runtime": 33.5888,
|
401 |
+
"eval_samples_per_second": 35.518,
|
402 |
+
"eval_steps_per_second": 2.233,
|
403 |
+
"eval_wer": 0.22529644268774704,
|
404 |
+
"step": 4800
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 59.76,
|
408 |
+
"learning_rate": 4.591836734693877e-07,
|
409 |
+
"loss": 0.7476,
|
410 |
+
"step": 4900
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 60.0,
|
414 |
+
"step": 4920,
|
415 |
+
"total_flos": 1.5656877251307375e+19,
|
416 |
+
"train_loss": 1.5886853283982936,
|
417 |
+
"train_runtime": 6665.2028,
|
418 |
+
"train_samples_per_second": 23.459,
|
419 |
+
"train_steps_per_second": 0.738
|
420 |
+
}
|
421 |
+
],
|
422 |
+
"max_steps": 4920,
|
423 |
+
"num_train_epochs": 60,
|
424 |
+
"total_flos": 1.5656877251307375e+19,
|
425 |
+
"trial_name": null,
|
426 |
+
"trial_params": null
|
427 |
+
}
|