Delete trainer_state.json
Browse files- trainer_state.json +0 -1785
trainer_state.json
DELETED
@@ -1,1785 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.3333613374779467,
|
5 |
-
"eval_steps": 248,
|
6 |
-
"global_step": 248,
|
7 |
-
"is_hyper_param_search": false,
|
8 |
-
"is_local_process_zero": true,
|
9 |
-
"is_world_process_zero": true,
|
10 |
-
"log_history": [
|
11 |
-
{
|
12 |
-
"epoch": 0.0013441989414433337,
|
13 |
-
"grad_norm": 9.1875,
|
14 |
-
"learning_rate": 1.0000000000000002e-06,
|
15 |
-
"loss": 1.0378,
|
16 |
-
"step": 1
|
17 |
-
},
|
18 |
-
{
|
19 |
-
"epoch": 0.0013441989414433337,
|
20 |
-
"eval_loss": 3.0436718463897705,
|
21 |
-
"eval_runtime": 5734.1832,
|
22 |
-
"eval_samples_per_second": 3.861,
|
23 |
-
"eval_steps_per_second": 0.483,
|
24 |
-
"step": 1
|
25 |
-
},
|
26 |
-
{
|
27 |
-
"epoch": 0.0026883978828866673,
|
28 |
-
"grad_norm": 10.3125,
|
29 |
-
"learning_rate": 2.0000000000000003e-06,
|
30 |
-
"loss": 1.1143,
|
31 |
-
"step": 2
|
32 |
-
},
|
33 |
-
{
|
34 |
-
"epoch": 0.004032596824330001,
|
35 |
-
"grad_norm": 8.75,
|
36 |
-
"learning_rate": 3e-06,
|
37 |
-
"loss": 1.0933,
|
38 |
-
"step": 3
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 0.005376795765773335,
|
42 |
-
"grad_norm": 6.40625,
|
43 |
-
"learning_rate": 4.000000000000001e-06,
|
44 |
-
"loss": 1.1014,
|
45 |
-
"step": 4
|
46 |
-
},
|
47 |
-
{
|
48 |
-
"epoch": 0.006720994707216668,
|
49 |
-
"grad_norm": 5.34375,
|
50 |
-
"learning_rate": 5e-06,
|
51 |
-
"loss": 1.0504,
|
52 |
-
"step": 5
|
53 |
-
},
|
54 |
-
{
|
55 |
-
"epoch": 0.008065193648660002,
|
56 |
-
"grad_norm": 3.453125,
|
57 |
-
"learning_rate": 6e-06,
|
58 |
-
"loss": 1.0031,
|
59 |
-
"step": 6
|
60 |
-
},
|
61 |
-
{
|
62 |
-
"epoch": 0.009409392590103335,
|
63 |
-
"grad_norm": 3.0,
|
64 |
-
"learning_rate": 7e-06,
|
65 |
-
"loss": 1.0001,
|
66 |
-
"step": 7
|
67 |
-
},
|
68 |
-
{
|
69 |
-
"epoch": 0.01075359153154667,
|
70 |
-
"grad_norm": 2.640625,
|
71 |
-
"learning_rate": 8.000000000000001e-06,
|
72 |
-
"loss": 1.0196,
|
73 |
-
"step": 8
|
74 |
-
},
|
75 |
-
{
|
76 |
-
"epoch": 0.012097790472990002,
|
77 |
-
"grad_norm": 2.21875,
|
78 |
-
"learning_rate": 9e-06,
|
79 |
-
"loss": 0.9207,
|
80 |
-
"step": 9
|
81 |
-
},
|
82 |
-
{
|
83 |
-
"epoch": 0.013441989414433336,
|
84 |
-
"grad_norm": 1.859375,
|
85 |
-
"learning_rate": 1e-05,
|
86 |
-
"loss": 0.9502,
|
87 |
-
"step": 10
|
88 |
-
},
|
89 |
-
{
|
90 |
-
"epoch": 0.01478618835587667,
|
91 |
-
"grad_norm": 2.34375,
|
92 |
-
"learning_rate": 9.999954076906038e-06,
|
93 |
-
"loss": 0.9674,
|
94 |
-
"step": 11
|
95 |
-
},
|
96 |
-
{
|
97 |
-
"epoch": 0.016130387297320005,
|
98 |
-
"grad_norm": 1.6953125,
|
99 |
-
"learning_rate": 9.99981630846772e-06,
|
100 |
-
"loss": 0.8617,
|
101 |
-
"step": 12
|
102 |
-
},
|
103 |
-
{
|
104 |
-
"epoch": 0.017474586238763336,
|
105 |
-
"grad_norm": 1.8046875,
|
106 |
-
"learning_rate": 9.999586697215748e-06,
|
107 |
-
"loss": 0.9222,
|
108 |
-
"step": 13
|
109 |
-
},
|
110 |
-
{
|
111 |
-
"epoch": 0.01881878518020667,
|
112 |
-
"grad_norm": 1.6484375,
|
113 |
-
"learning_rate": 9.999265247367909e-06,
|
114 |
-
"loss": 0.884,
|
115 |
-
"step": 14
|
116 |
-
},
|
117 |
-
{
|
118 |
-
"epoch": 0.020162984121650004,
|
119 |
-
"grad_norm": 1.7578125,
|
120 |
-
"learning_rate": 9.998851964828987e-06,
|
121 |
-
"loss": 0.9086,
|
122 |
-
"step": 15
|
123 |
-
},
|
124 |
-
{
|
125 |
-
"epoch": 0.02150718306309334,
|
126 |
-
"grad_norm": 1.6015625,
|
127 |
-
"learning_rate": 9.99834685719067e-06,
|
128 |
-
"loss": 0.8723,
|
129 |
-
"step": 16
|
130 |
-
},
|
131 |
-
{
|
132 |
-
"epoch": 0.022851382004536673,
|
133 |
-
"grad_norm": 1.7109375,
|
134 |
-
"learning_rate": 9.997749933731397e-06,
|
135 |
-
"loss": 0.8678,
|
136 |
-
"step": 17
|
137 |
-
},
|
138 |
-
{
|
139 |
-
"epoch": 0.024195580945980004,
|
140 |
-
"grad_norm": 1.5625,
|
141 |
-
"learning_rate": 9.997061205416203e-06,
|
142 |
-
"loss": 0.8497,
|
143 |
-
"step": 18
|
144 |
-
},
|
145 |
-
{
|
146 |
-
"epoch": 0.025539779887423338,
|
147 |
-
"grad_norm": 1.515625,
|
148 |
-
"learning_rate": 9.996280684896496e-06,
|
149 |
-
"loss": 0.8475,
|
150 |
-
"step": 19
|
151 |
-
},
|
152 |
-
{
|
153 |
-
"epoch": 0.026883978828866673,
|
154 |
-
"grad_norm": 1.625,
|
155 |
-
"learning_rate": 9.995408386509846e-06,
|
156 |
-
"loss": 0.8673,
|
157 |
-
"step": 20
|
158 |
-
},
|
159 |
-
{
|
160 |
-
"epoch": 0.028228177770310007,
|
161 |
-
"grad_norm": 1.5625,
|
162 |
-
"learning_rate": 9.99444432627971e-06,
|
163 |
-
"loss": 0.8704,
|
164 |
-
"step": 21
|
165 |
-
},
|
166 |
-
{
|
167 |
-
"epoch": 0.02957237671175334,
|
168 |
-
"grad_norm": 1.53125,
|
169 |
-
"learning_rate": 9.993388521915134e-06,
|
170 |
-
"loss": 0.813,
|
171 |
-
"step": 22
|
172 |
-
},
|
173 |
-
{
|
174 |
-
"epoch": 0.030916575653196672,
|
175 |
-
"grad_norm": 1.453125,
|
176 |
-
"learning_rate": 9.992240992810445e-06,
|
177 |
-
"loss": 0.8497,
|
178 |
-
"step": 23
|
179 |
-
},
|
180 |
-
{
|
181 |
-
"epoch": 0.03226077459464001,
|
182 |
-
"grad_norm": 1.453125,
|
183 |
-
"learning_rate": 9.991001760044877e-06,
|
184 |
-
"loss": 0.8027,
|
185 |
-
"step": 24
|
186 |
-
},
|
187 |
-
{
|
188 |
-
"epoch": 0.03360497353608334,
|
189 |
-
"grad_norm": 1.375,
|
190 |
-
"learning_rate": 9.989670846382189e-06,
|
191 |
-
"loss": 0.8099,
|
192 |
-
"step": 25
|
193 |
-
},
|
194 |
-
{
|
195 |
-
"epoch": 0.03494917247752667,
|
196 |
-
"grad_norm": 1.328125,
|
197 |
-
"learning_rate": 9.98824827627025e-06,
|
198 |
-
"loss": 0.7818,
|
199 |
-
"step": 26
|
200 |
-
},
|
201 |
-
{
|
202 |
-
"epoch": 0.03629337141897001,
|
203 |
-
"grad_norm": 1.5,
|
204 |
-
"learning_rate": 9.986734075840591e-06,
|
205 |
-
"loss": 0.8078,
|
206 |
-
"step": 27
|
207 |
-
},
|
208 |
-
{
|
209 |
-
"epoch": 0.03763757036041334,
|
210 |
-
"grad_norm": 1.5625,
|
211 |
-
"learning_rate": 9.985128272907917e-06,
|
212 |
-
"loss": 0.819,
|
213 |
-
"step": 28
|
214 |
-
},
|
215 |
-
{
|
216 |
-
"epoch": 0.03898176930185668,
|
217 |
-
"grad_norm": 1.34375,
|
218 |
-
"learning_rate": 9.983430896969606e-06,
|
219 |
-
"loss": 0.7949,
|
220 |
-
"step": 29
|
221 |
-
},
|
222 |
-
{
|
223 |
-
"epoch": 0.04032596824330001,
|
224 |
-
"grad_norm": 1.46875,
|
225 |
-
"learning_rate": 9.981641979205158e-06,
|
226 |
-
"loss": 0.8044,
|
227 |
-
"step": 30
|
228 |
-
},
|
229 |
-
{
|
230 |
-
"epoch": 0.04167016718474334,
|
231 |
-
"grad_norm": 1.4921875,
|
232 |
-
"learning_rate": 9.97976155247563e-06,
|
233 |
-
"loss": 0.8025,
|
234 |
-
"step": 31
|
235 |
-
},
|
236 |
-
{
|
237 |
-
"epoch": 0.04301436612618668,
|
238 |
-
"grad_norm": 1.3828125,
|
239 |
-
"learning_rate": 9.977789651323025e-06,
|
240 |
-
"loss": 0.7461,
|
241 |
-
"step": 32
|
242 |
-
},
|
243 |
-
{
|
244 |
-
"epoch": 0.04435856506763001,
|
245 |
-
"grad_norm": 1.40625,
|
246 |
-
"learning_rate": 9.975726311969664e-06,
|
247 |
-
"loss": 0.787,
|
248 |
-
"step": 33
|
249 |
-
},
|
250 |
-
{
|
251 |
-
"epoch": 0.045702764009073346,
|
252 |
-
"grad_norm": 1.4609375,
|
253 |
-
"learning_rate": 9.973571572317519e-06,
|
254 |
-
"loss": 0.7837,
|
255 |
-
"step": 34
|
256 |
-
},
|
257 |
-
{
|
258 |
-
"epoch": 0.04704696295051668,
|
259 |
-
"grad_norm": 1.4453125,
|
260 |
-
"learning_rate": 9.971325471947518e-06,
|
261 |
-
"loss": 0.8101,
|
262 |
-
"step": 35
|
263 |
-
},
|
264 |
-
{
|
265 |
-
"epoch": 0.04839116189196001,
|
266 |
-
"grad_norm": 1.4921875,
|
267 |
-
"learning_rate": 9.968988052118804e-06,
|
268 |
-
"loss": 0.783,
|
269 |
-
"step": 36
|
270 |
-
},
|
271 |
-
{
|
272 |
-
"epoch": 0.049735360833403346,
|
273 |
-
"grad_norm": 1.4609375,
|
274 |
-
"learning_rate": 9.966559355768005e-06,
|
275 |
-
"loss": 0.8141,
|
276 |
-
"step": 37
|
277 |
-
},
|
278 |
-
{
|
279 |
-
"epoch": 0.051079559774846676,
|
280 |
-
"grad_norm": 1.359375,
|
281 |
-
"learning_rate": 9.964039427508418e-06,
|
282 |
-
"loss": 0.7753,
|
283 |
-
"step": 38
|
284 |
-
},
|
285 |
-
{
|
286 |
-
"epoch": 0.052423758716290014,
|
287 |
-
"grad_norm": 1.375,
|
288 |
-
"learning_rate": 9.961428313629203e-06,
|
289 |
-
"loss": 0.773,
|
290 |
-
"step": 39
|
291 |
-
},
|
292 |
-
{
|
293 |
-
"epoch": 0.053767957657733345,
|
294 |
-
"grad_norm": 1.3515625,
|
295 |
-
"learning_rate": 9.958726062094533e-06,
|
296 |
-
"loss": 0.7557,
|
297 |
-
"step": 40
|
298 |
-
},
|
299 |
-
{
|
300 |
-
"epoch": 0.055112156599176676,
|
301 |
-
"grad_norm": 1.3203125,
|
302 |
-
"learning_rate": 9.955932722542709e-06,
|
303 |
-
"loss": 0.7241,
|
304 |
-
"step": 41
|
305 |
-
},
|
306 |
-
{
|
307 |
-
"epoch": 0.056456355540620014,
|
308 |
-
"grad_norm": 1.4296875,
|
309 |
-
"learning_rate": 9.953048346285245e-06,
|
310 |
-
"loss": 0.781,
|
311 |
-
"step": 42
|
312 |
-
},
|
313 |
-
{
|
314 |
-
"epoch": 0.057800554482063345,
|
315 |
-
"grad_norm": 1.3984375,
|
316 |
-
"learning_rate": 9.950072986305938e-06,
|
317 |
-
"loss": 0.7644,
|
318 |
-
"step": 43
|
319 |
-
},
|
320 |
-
{
|
321 |
-
"epoch": 0.05914475342350668,
|
322 |
-
"grad_norm": 1.3515625,
|
323 |
-
"learning_rate": 9.947006697259881e-06,
|
324 |
-
"loss": 0.7406,
|
325 |
-
"step": 44
|
326 |
-
},
|
327 |
-
{
|
328 |
-
"epoch": 0.06048895236495001,
|
329 |
-
"grad_norm": 1.3828125,
|
330 |
-
"learning_rate": 9.943849535472468e-06,
|
331 |
-
"loss": 0.762,
|
332 |
-
"step": 45
|
333 |
-
},
|
334 |
-
{
|
335 |
-
"epoch": 0.061833151306393344,
|
336 |
-
"grad_norm": 1.3046875,
|
337 |
-
"learning_rate": 9.940601558938348e-06,
|
338 |
-
"loss": 0.7448,
|
339 |
-
"step": 46
|
340 |
-
},
|
341 |
-
{
|
342 |
-
"epoch": 0.06317735024783667,
|
343 |
-
"grad_norm": 1.390625,
|
344 |
-
"learning_rate": 9.93726282732038e-06,
|
345 |
-
"loss": 0.7658,
|
346 |
-
"step": 47
|
347 |
-
},
|
348 |
-
{
|
349 |
-
"epoch": 0.06452154918928002,
|
350 |
-
"grad_norm": 1.4765625,
|
351 |
-
"learning_rate": 9.933833401948514e-06,
|
352 |
-
"loss": 0.7876,
|
353 |
-
"step": 48
|
354 |
-
},
|
355 |
-
{
|
356 |
-
"epoch": 0.06586574813072335,
|
357 |
-
"grad_norm": 1.3984375,
|
358 |
-
"learning_rate": 9.930313345818683e-06,
|
359 |
-
"loss": 0.8166,
|
360 |
-
"step": 49
|
361 |
-
},
|
362 |
-
{
|
363 |
-
"epoch": 0.06720994707216668,
|
364 |
-
"grad_norm": 1.421875,
|
365 |
-
"learning_rate": 9.92670272359163e-06,
|
366 |
-
"loss": 0.7927,
|
367 |
-
"step": 50
|
368 |
-
},
|
369 |
-
{
|
370 |
-
"epoch": 0.06855414601361001,
|
371 |
-
"grad_norm": 1.359375,
|
372 |
-
"learning_rate": 9.923001601591738e-06,
|
373 |
-
"loss": 0.7906,
|
374 |
-
"step": 51
|
375 |
-
},
|
376 |
-
{
|
377 |
-
"epoch": 0.06989834495505334,
|
378 |
-
"grad_norm": 1.296875,
|
379 |
-
"learning_rate": 9.919210047805791e-06,
|
380 |
-
"loss": 0.7576,
|
381 |
-
"step": 52
|
382 |
-
},
|
383 |
-
{
|
384 |
-
"epoch": 0.07124254389649669,
|
385 |
-
"grad_norm": 1.4453125,
|
386 |
-
"learning_rate": 9.915328131881745e-06,
|
387 |
-
"loss": 0.7847,
|
388 |
-
"step": 53
|
389 |
-
},
|
390 |
-
{
|
391 |
-
"epoch": 0.07258674283794002,
|
392 |
-
"grad_norm": 1.4375,
|
393 |
-
"learning_rate": 9.911355925127433e-06,
|
394 |
-
"loss": 0.7728,
|
395 |
-
"step": 54
|
396 |
-
},
|
397 |
-
{
|
398 |
-
"epoch": 0.07393094177938335,
|
399 |
-
"grad_norm": 1.4375,
|
400 |
-
"learning_rate": 9.907293500509268e-06,
|
401 |
-
"loss": 0.7947,
|
402 |
-
"step": 55
|
403 |
-
},
|
404 |
-
{
|
405 |
-
"epoch": 0.07527514072082668,
|
406 |
-
"grad_norm": 1.3515625,
|
407 |
-
"learning_rate": 9.903140932650891e-06,
|
408 |
-
"loss": 0.7503,
|
409 |
-
"step": 56
|
410 |
-
},
|
411 |
-
{
|
412 |
-
"epoch": 0.07661933966227001,
|
413 |
-
"grad_norm": 1.3671875,
|
414 |
-
"learning_rate": 9.898898297831808e-06,
|
415 |
-
"loss": 0.7519,
|
416 |
-
"step": 57
|
417 |
-
},
|
418 |
-
{
|
419 |
-
"epoch": 0.07796353860371336,
|
420 |
-
"grad_norm": 1.4296875,
|
421 |
-
"learning_rate": 9.894565673985986e-06,
|
422 |
-
"loss": 0.7438,
|
423 |
-
"step": 58
|
424 |
-
},
|
425 |
-
{
|
426 |
-
"epoch": 0.07930773754515669,
|
427 |
-
"grad_norm": 1.4375,
|
428 |
-
"learning_rate": 9.890143140700419e-06,
|
429 |
-
"loss": 0.7635,
|
430 |
-
"step": 59
|
431 |
-
},
|
432 |
-
{
|
433 |
-
"epoch": 0.08065193648660002,
|
434 |
-
"grad_norm": 1.375,
|
435 |
-
"learning_rate": 9.885630779213678e-06,
|
436 |
-
"loss": 0.7176,
|
437 |
-
"step": 60
|
438 |
-
},
|
439 |
-
{
|
440 |
-
"epoch": 0.08199613542804335,
|
441 |
-
"grad_norm": 1.4296875,
|
442 |
-
"learning_rate": 9.881028672414397e-06,
|
443 |
-
"loss": 0.7536,
|
444 |
-
"step": 61
|
445 |
-
},
|
446 |
-
{
|
447 |
-
"epoch": 0.08334033436948668,
|
448 |
-
"grad_norm": 1.4375,
|
449 |
-
"learning_rate": 9.876336904839772e-06,
|
450 |
-
"loss": 0.7572,
|
451 |
-
"step": 62
|
452 |
-
},
|
453 |
-
{
|
454 |
-
"epoch": 0.08468453331093002,
|
455 |
-
"grad_norm": 1.4609375,
|
456 |
-
"learning_rate": 9.871555562673996e-06,
|
457 |
-
"loss": 0.7767,
|
458 |
-
"step": 63
|
459 |
-
},
|
460 |
-
{
|
461 |
-
"epoch": 0.08602873225237335,
|
462 |
-
"grad_norm": 1.4921875,
|
463 |
-
"learning_rate": 9.86668473374668e-06,
|
464 |
-
"loss": 0.763,
|
465 |
-
"step": 64
|
466 |
-
},
|
467 |
-
{
|
468 |
-
"epoch": 0.08737293119381669,
|
469 |
-
"grad_norm": 1.3984375,
|
470 |
-
"learning_rate": 9.861724507531234e-06,
|
471 |
-
"loss": 0.7576,
|
472 |
-
"step": 65
|
473 |
-
},
|
474 |
-
{
|
475 |
-
"epoch": 0.08871713013526002,
|
476 |
-
"grad_norm": 1.421875,
|
477 |
-
"learning_rate": 9.856674975143237e-06,
|
478 |
-
"loss": 0.7342,
|
479 |
-
"step": 66
|
480 |
-
},
|
481 |
-
{
|
482 |
-
"epoch": 0.09006132907670335,
|
483 |
-
"grad_norm": 1.4453125,
|
484 |
-
"learning_rate": 9.851536229338747e-06,
|
485 |
-
"loss": 0.7487,
|
486 |
-
"step": 67
|
487 |
-
},
|
488 |
-
{
|
489 |
-
"epoch": 0.09140552801814669,
|
490 |
-
"grad_norm": 1.390625,
|
491 |
-
"learning_rate": 9.846308364512607e-06,
|
492 |
-
"loss": 0.7604,
|
493 |
-
"step": 68
|
494 |
-
},
|
495 |
-
{
|
496 |
-
"epoch": 0.09274972695959002,
|
497 |
-
"grad_norm": 1.5078125,
|
498 |
-
"learning_rate": 9.840991476696707e-06,
|
499 |
-
"loss": 0.7537,
|
500 |
-
"step": 69
|
501 |
-
},
|
502 |
-
{
|
503 |
-
"epoch": 0.09409392590103335,
|
504 |
-
"grad_norm": 1.4296875,
|
505 |
-
"learning_rate": 9.835585663558221e-06,
|
506 |
-
"loss": 0.7608,
|
507 |
-
"step": 70
|
508 |
-
},
|
509 |
-
{
|
510 |
-
"epoch": 0.09543812484247668,
|
511 |
-
"grad_norm": 1.4140625,
|
512 |
-
"learning_rate": 9.830091024397818e-06,
|
513 |
-
"loss": 0.7298,
|
514 |
-
"step": 71
|
515 |
-
},
|
516 |
-
{
|
517 |
-
"epoch": 0.09678232378392002,
|
518 |
-
"grad_norm": 1.4140625,
|
519 |
-
"learning_rate": 9.824507660147831e-06,
|
520 |
-
"loss": 0.7646,
|
521 |
-
"step": 72
|
522 |
-
},
|
523 |
-
{
|
524 |
-
"epoch": 0.09812652272536336,
|
525 |
-
"grad_norm": 1.3984375,
|
526 |
-
"learning_rate": 9.8188356733704e-06,
|
527 |
-
"loss": 0.7679,
|
528 |
-
"step": 73
|
529 |
-
},
|
530 |
-
{
|
531 |
-
"epoch": 0.09947072166680669,
|
532 |
-
"grad_norm": 1.4765625,
|
533 |
-
"learning_rate": 9.813075168255601e-06,
|
534 |
-
"loss": 0.7263,
|
535 |
-
"step": 74
|
536 |
-
},
|
537 |
-
{
|
538 |
-
"epoch": 0.10081492060825002,
|
539 |
-
"grad_norm": 1.484375,
|
540 |
-
"learning_rate": 9.807226250619522e-06,
|
541 |
-
"loss": 0.7589,
|
542 |
-
"step": 75
|
543 |
-
},
|
544 |
-
{
|
545 |
-
"epoch": 0.10215911954969335,
|
546 |
-
"grad_norm": 1.515625,
|
547 |
-
"learning_rate": 9.801289027902316e-06,
|
548 |
-
"loss": 0.7216,
|
549 |
-
"step": 76
|
550 |
-
},
|
551 |
-
{
|
552 |
-
"epoch": 0.10350331849113668,
|
553 |
-
"grad_norm": 1.4765625,
|
554 |
-
"learning_rate": 9.795263609166243e-06,
|
555 |
-
"loss": 0.7779,
|
556 |
-
"step": 77
|
557 |
-
},
|
558 |
-
{
|
559 |
-
"epoch": 0.10484751743258003,
|
560 |
-
"grad_norm": 1.4609375,
|
561 |
-
"learning_rate": 9.789150105093647e-06,
|
562 |
-
"loss": 0.6941,
|
563 |
-
"step": 78
|
564 |
-
},
|
565 |
-
{
|
566 |
-
"epoch": 0.10619171637402336,
|
567 |
-
"grad_norm": 1.46875,
|
568 |
-
"learning_rate": 9.78294862798494e-06,
|
569 |
-
"loss": 0.7362,
|
570 |
-
"step": 79
|
571 |
-
},
|
572 |
-
{
|
573 |
-
"epoch": 0.10753591531546669,
|
574 |
-
"grad_norm": 1.4296875,
|
575 |
-
"learning_rate": 9.776659291756528e-06,
|
576 |
-
"loss": 0.7103,
|
577 |
-
"step": 80
|
578 |
-
},
|
579 |
-
{
|
580 |
-
"epoch": 0.10888011425691002,
|
581 |
-
"grad_norm": 1.359375,
|
582 |
-
"learning_rate": 9.77028221193872e-06,
|
583 |
-
"loss": 0.6866,
|
584 |
-
"step": 81
|
585 |
-
},
|
586 |
-
{
|
587 |
-
"epoch": 0.11022431319835335,
|
588 |
-
"grad_norm": 1.53125,
|
589 |
-
"learning_rate": 9.763817505673614e-06,
|
590 |
-
"loss": 0.7517,
|
591 |
-
"step": 82
|
592 |
-
},
|
593 |
-
{
|
594 |
-
"epoch": 0.1115685121397967,
|
595 |
-
"grad_norm": 1.453125,
|
596 |
-
"learning_rate": 9.75726529171293e-06,
|
597 |
-
"loss": 0.7508,
|
598 |
-
"step": 83
|
599 |
-
},
|
600 |
-
{
|
601 |
-
"epoch": 0.11291271108124003,
|
602 |
-
"grad_norm": 1.5,
|
603 |
-
"learning_rate": 9.750625690415848e-06,
|
604 |
-
"loss": 0.7275,
|
605 |
-
"step": 84
|
606 |
-
},
|
607 |
-
{
|
608 |
-
"epoch": 0.11425691002268336,
|
609 |
-
"grad_norm": 1.484375,
|
610 |
-
"learning_rate": 9.74389882374678e-06,
|
611 |
-
"loss": 0.6994,
|
612 |
-
"step": 85
|
613 |
-
},
|
614 |
-
{
|
615 |
-
"epoch": 0.11560110896412669,
|
616 |
-
"grad_norm": 1.4453125,
|
617 |
-
"learning_rate": 9.737084815273137e-06,
|
618 |
-
"loss": 0.7365,
|
619 |
-
"step": 86
|
620 |
-
},
|
621 |
-
{
|
622 |
-
"epoch": 0.11694530790557002,
|
623 |
-
"grad_norm": 1.5078125,
|
624 |
-
"learning_rate": 9.730183790163061e-06,
|
625 |
-
"loss": 0.7294,
|
626 |
-
"step": 87
|
627 |
-
},
|
628 |
-
{
|
629 |
-
"epoch": 0.11828950684701336,
|
630 |
-
"grad_norm": 1.375,
|
631 |
-
"learning_rate": 9.72319587518312e-06,
|
632 |
-
"loss": 0.7226,
|
633 |
-
"step": 88
|
634 |
-
},
|
635 |
-
{
|
636 |
-
"epoch": 0.1196337057884567,
|
637 |
-
"grad_norm": 1.390625,
|
638 |
-
"learning_rate": 9.716121198695987e-06,
|
639 |
-
"loss": 0.7126,
|
640 |
-
"step": 89
|
641 |
-
},
|
642 |
-
{
|
643 |
-
"epoch": 0.12097790472990003,
|
644 |
-
"grad_norm": 1.484375,
|
645 |
-
"learning_rate": 9.708959890658074e-06,
|
646 |
-
"loss": 0.7118,
|
647 |
-
"step": 90
|
648 |
-
},
|
649 |
-
{
|
650 |
-
"epoch": 0.12232210367134336,
|
651 |
-
"grad_norm": 1.3515625,
|
652 |
-
"learning_rate": 9.70171208261715e-06,
|
653 |
-
"loss": 0.7164,
|
654 |
-
"step": 91
|
655 |
-
},
|
656 |
-
{
|
657 |
-
"epoch": 0.12366630261278669,
|
658 |
-
"grad_norm": 1.4765625,
|
659 |
-
"learning_rate": 9.69437790770992e-06,
|
660 |
-
"loss": 0.7015,
|
661 |
-
"step": 92
|
662 |
-
},
|
663 |
-
{
|
664 |
-
"epoch": 0.12501050155423002,
|
665 |
-
"grad_norm": 1.3828125,
|
666 |
-
"learning_rate": 9.68695750065959e-06,
|
667 |
-
"loss": 0.7212,
|
668 |
-
"step": 93
|
669 |
-
},
|
670 |
-
{
|
671 |
-
"epoch": 0.12635470049567335,
|
672 |
-
"grad_norm": 1.4140625,
|
673 |
-
"learning_rate": 9.679450997773378e-06,
|
674 |
-
"loss": 0.7301,
|
675 |
-
"step": 94
|
676 |
-
},
|
677 |
-
{
|
678 |
-
"epoch": 0.12769889943711668,
|
679 |
-
"grad_norm": 1.53125,
|
680 |
-
"learning_rate": 9.67185853694002e-06,
|
681 |
-
"loss": 0.7434,
|
682 |
-
"step": 95
|
683 |
-
},
|
684 |
-
{
|
685 |
-
"epoch": 0.12904309837856004,
|
686 |
-
"grad_norm": 1.3828125,
|
687 |
-
"learning_rate": 9.664180257627231e-06,
|
688 |
-
"loss": 0.7503,
|
689 |
-
"step": 96
|
690 |
-
},
|
691 |
-
{
|
692 |
-
"epoch": 0.13038729732000337,
|
693 |
-
"grad_norm": 1.3984375,
|
694 |
-
"learning_rate": 9.656416300879147e-06,
|
695 |
-
"loss": 0.704,
|
696 |
-
"step": 97
|
697 |
-
},
|
698 |
-
{
|
699 |
-
"epoch": 0.1317314962614467,
|
700 |
-
"grad_norm": 1.3515625,
|
701 |
-
"learning_rate": 9.648566809313738e-06,
|
702 |
-
"loss": 0.7091,
|
703 |
-
"step": 98
|
704 |
-
},
|
705 |
-
{
|
706 |
-
"epoch": 0.13307569520289003,
|
707 |
-
"grad_norm": 1.25,
|
708 |
-
"learning_rate": 9.640631927120177e-06,
|
709 |
-
"loss": 0.6939,
|
710 |
-
"step": 99
|
711 |
-
},
|
712 |
-
{
|
713 |
-
"epoch": 0.13441989414433336,
|
714 |
-
"grad_norm": 1.25,
|
715 |
-
"learning_rate": 9.632611800056202e-06,
|
716 |
-
"loss": 0.6645,
|
717 |
-
"step": 100
|
718 |
-
},
|
719 |
-
{
|
720 |
-
"epoch": 0.1357640930857767,
|
721 |
-
"grad_norm": 1.3671875,
|
722 |
-
"learning_rate": 9.62450657544543e-06,
|
723 |
-
"loss": 0.7314,
|
724 |
-
"step": 101
|
725 |
-
},
|
726 |
-
{
|
727 |
-
"epoch": 0.13710829202722002,
|
728 |
-
"grad_norm": 1.359375,
|
729 |
-
"learning_rate": 9.616316402174657e-06,
|
730 |
-
"loss": 0.7022,
|
731 |
-
"step": 102
|
732 |
-
},
|
733 |
-
{
|
734 |
-
"epoch": 0.13845249096866336,
|
735 |
-
"grad_norm": 1.390625,
|
736 |
-
"learning_rate": 9.608041430691126e-06,
|
737 |
-
"loss": 0.7015,
|
738 |
-
"step": 103
|
739 |
-
},
|
740 |
-
{
|
741 |
-
"epoch": 0.1397966899101067,
|
742 |
-
"grad_norm": 1.3046875,
|
743 |
-
"learning_rate": 9.59968181299975e-06,
|
744 |
-
"loss": 0.6831,
|
745 |
-
"step": 104
|
746 |
-
},
|
747 |
-
{
|
748 |
-
"epoch": 0.14114088885155002,
|
749 |
-
"grad_norm": 1.28125,
|
750 |
-
"learning_rate": 9.591237702660335e-06,
|
751 |
-
"loss": 0.6903,
|
752 |
-
"step": 105
|
753 |
-
},
|
754 |
-
{
|
755 |
-
"epoch": 0.14248508779299338,
|
756 |
-
"grad_norm": 1.2421875,
|
757 |
-
"learning_rate": 9.58270925478475e-06,
|
758 |
-
"loss": 0.7236,
|
759 |
-
"step": 106
|
760 |
-
},
|
761 |
-
{
|
762 |
-
"epoch": 0.1438292867344367,
|
763 |
-
"grad_norm": 1.3125,
|
764 |
-
"learning_rate": 9.574096626034077e-06,
|
765 |
-
"loss": 0.7375,
|
766 |
-
"step": 107
|
767 |
-
},
|
768 |
-
{
|
769 |
-
"epoch": 0.14517348567588004,
|
770 |
-
"grad_norm": 1.2421875,
|
771 |
-
"learning_rate": 9.565399974615744e-06,
|
772 |
-
"loss": 0.7051,
|
773 |
-
"step": 108
|
774 |
-
},
|
775 |
-
{
|
776 |
-
"epoch": 0.14651768461732337,
|
777 |
-
"grad_norm": 1.3515625,
|
778 |
-
"learning_rate": 9.556619460280605e-06,
|
779 |
-
"loss": 0.6961,
|
780 |
-
"step": 109
|
781 |
-
},
|
782 |
-
{
|
783 |
-
"epoch": 0.1478618835587667,
|
784 |
-
"grad_norm": 1.2890625,
|
785 |
-
"learning_rate": 9.547755244320013e-06,
|
786 |
-
"loss": 0.731,
|
787 |
-
"step": 110
|
788 |
-
},
|
789 |
-
{
|
790 |
-
"epoch": 0.14920608250021003,
|
791 |
-
"grad_norm": 1.3125,
|
792 |
-
"learning_rate": 9.53880748956286e-06,
|
793 |
-
"loss": 0.6992,
|
794 |
-
"step": 111
|
795 |
-
},
|
796 |
-
{
|
797 |
-
"epoch": 0.15055028144165336,
|
798 |
-
"grad_norm": 1.2578125,
|
799 |
-
"learning_rate": 9.529776360372576e-06,
|
800 |
-
"loss": 0.6954,
|
801 |
-
"step": 112
|
802 |
-
},
|
803 |
-
{
|
804 |
-
"epoch": 0.1518944803830967,
|
805 |
-
"grad_norm": 1.140625,
|
806 |
-
"learning_rate": 9.52066202264412e-06,
|
807 |
-
"loss": 0.6932,
|
808 |
-
"step": 113
|
809 |
-
},
|
810 |
-
{
|
811 |
-
"epoch": 0.15323867932454002,
|
812 |
-
"grad_norm": 1.1875,
|
813 |
-
"learning_rate": 9.511464643800926e-06,
|
814 |
-
"loss": 0.7117,
|
815 |
-
"step": 114
|
816 |
-
},
|
817 |
-
{
|
818 |
-
"epoch": 0.15458287826598335,
|
819 |
-
"grad_norm": 1.1875,
|
820 |
-
"learning_rate": 9.502184392791834e-06,
|
821 |
-
"loss": 0.6992,
|
822 |
-
"step": 115
|
823 |
-
},
|
824 |
-
{
|
825 |
-
"epoch": 0.1559270772074267,
|
826 |
-
"grad_norm": 1.09375,
|
827 |
-
"learning_rate": 9.492821440087978e-06,
|
828 |
-
"loss": 0.6744,
|
829 |
-
"step": 116
|
830 |
-
},
|
831 |
-
{
|
832 |
-
"epoch": 0.15727127614887004,
|
833 |
-
"grad_norm": 1.1796875,
|
834 |
-
"learning_rate": 9.48337595767966e-06,
|
835 |
-
"loss": 0.6553,
|
836 |
-
"step": 117
|
837 |
-
},
|
838 |
-
{
|
839 |
-
"epoch": 0.15861547509031337,
|
840 |
-
"grad_norm": 1.1484375,
|
841 |
-
"learning_rate": 9.473848119073188e-06,
|
842 |
-
"loss": 0.7206,
|
843 |
-
"step": 118
|
844 |
-
},
|
845 |
-
{
|
846 |
-
"epoch": 0.1599596740317567,
|
847 |
-
"grad_norm": 1.125,
|
848 |
-
"learning_rate": 9.4642380992877e-06,
|
849 |
-
"loss": 0.7014,
|
850 |
-
"step": 119
|
851 |
-
},
|
852 |
-
{
|
853 |
-
"epoch": 0.16130387297320004,
|
854 |
-
"grad_norm": 1.1171875,
|
855 |
-
"learning_rate": 9.454546074851927e-06,
|
856 |
-
"loss": 0.7249,
|
857 |
-
"step": 120
|
858 |
-
},
|
859 |
-
{
|
860 |
-
"epoch": 0.16264807191464337,
|
861 |
-
"grad_norm": 1.140625,
|
862 |
-
"learning_rate": 9.444772223800972e-06,
|
863 |
-
"loss": 0.7142,
|
864 |
-
"step": 121
|
865 |
-
},
|
866 |
-
{
|
867 |
-
"epoch": 0.1639922708560867,
|
868 |
-
"grad_norm": 1.0390625,
|
869 |
-
"learning_rate": 9.434916725673023e-06,
|
870 |
-
"loss": 0.6778,
|
871 |
-
"step": 122
|
872 |
-
},
|
873 |
-
{
|
874 |
-
"epoch": 0.16533646979753003,
|
875 |
-
"grad_norm": 1.0390625,
|
876 |
-
"learning_rate": 9.42497976150607e-06,
|
877 |
-
"loss": 0.6831,
|
878 |
-
"step": 123
|
879 |
-
},
|
880 |
-
{
|
881 |
-
"epoch": 0.16668066873897336,
|
882 |
-
"grad_norm": 1.0546875,
|
883 |
-
"learning_rate": 9.414961513834569e-06,
|
884 |
-
"loss": 0.6744,
|
885 |
-
"step": 124
|
886 |
-
},
|
887 |
-
{
|
888 |
-
"epoch": 0.1680248676804167,
|
889 |
-
"grad_norm": 1.1015625,
|
890 |
-
"learning_rate": 9.404862166686089e-06,
|
891 |
-
"loss": 0.7059,
|
892 |
-
"step": 125
|
893 |
-
},
|
894 |
-
{
|
895 |
-
"epoch": 0.16936906662186005,
|
896 |
-
"grad_norm": 1.0390625,
|
897 |
-
"learning_rate": 9.394681905577938e-06,
|
898 |
-
"loss": 0.7245,
|
899 |
-
"step": 126
|
900 |
-
},
|
901 |
-
{
|
902 |
-
"epoch": 0.17071326556330338,
|
903 |
-
"grad_norm": 1.0390625,
|
904 |
-
"learning_rate": 9.384420917513752e-06,
|
905 |
-
"loss": 0.7064,
|
906 |
-
"step": 127
|
907 |
-
},
|
908 |
-
{
|
909 |
-
"epoch": 0.1720574645047467,
|
910 |
-
"grad_norm": 1.0,
|
911 |
-
"learning_rate": 9.374079390980058e-06,
|
912 |
-
"loss": 0.7177,
|
913 |
-
"step": 128
|
914 |
-
},
|
915 |
-
{
|
916 |
-
"epoch": 0.17340166344619004,
|
917 |
-
"grad_norm": 1.0078125,
|
918 |
-
"learning_rate": 9.363657515942814e-06,
|
919 |
-
"loss": 0.7031,
|
920 |
-
"step": 129
|
921 |
-
},
|
922 |
-
{
|
923 |
-
"epoch": 0.17474586238763337,
|
924 |
-
"grad_norm": 0.96875,
|
925 |
-
"learning_rate": 9.35315548384392e-06,
|
926 |
-
"loss": 0.6752,
|
927 |
-
"step": 130
|
928 |
-
},
|
929 |
-
{
|
930 |
-
"epoch": 0.1760900613290767,
|
931 |
-
"grad_norm": 1.0,
|
932 |
-
"learning_rate": 9.342573487597696e-06,
|
933 |
-
"loss": 0.7189,
|
934 |
-
"step": 131
|
935 |
-
},
|
936 |
-
{
|
937 |
-
"epoch": 0.17743426027052003,
|
938 |
-
"grad_norm": 1.0234375,
|
939 |
-
"learning_rate": 9.331911721587345e-06,
|
940 |
-
"loss": 0.6873,
|
941 |
-
"step": 132
|
942 |
-
},
|
943 |
-
{
|
944 |
-
"epoch": 0.17877845921196336,
|
945 |
-
"grad_norm": 1.0,
|
946 |
-
"learning_rate": 9.321170381661383e-06,
|
947 |
-
"loss": 0.7056,
|
948 |
-
"step": 133
|
949 |
-
},
|
950 |
-
{
|
951 |
-
"epoch": 0.1801226581534067,
|
952 |
-
"grad_norm": 0.984375,
|
953 |
-
"learning_rate": 9.310349665130035e-06,
|
954 |
-
"loss": 0.7024,
|
955 |
-
"step": 134
|
956 |
-
},
|
957 |
-
{
|
958 |
-
"epoch": 0.18146685709485003,
|
959 |
-
"grad_norm": 1.015625,
|
960 |
-
"learning_rate": 9.299449770761612e-06,
|
961 |
-
"loss": 0.708,
|
962 |
-
"step": 135
|
963 |
-
},
|
964 |
-
{
|
965 |
-
"epoch": 0.18281105603629338,
|
966 |
-
"grad_norm": 0.8984375,
|
967 |
-
"learning_rate": 9.288470898778863e-06,
|
968 |
-
"loss": 0.6624,
|
969 |
-
"step": 136
|
970 |
-
},
|
971 |
-
{
|
972 |
-
"epoch": 0.18415525497773672,
|
973 |
-
"grad_norm": 0.984375,
|
974 |
-
"learning_rate": 9.277413250855296e-06,
|
975 |
-
"loss": 0.6622,
|
976 |
-
"step": 137
|
977 |
-
},
|
978 |
-
{
|
979 |
-
"epoch": 0.18549945391918005,
|
980 |
-
"grad_norm": 0.9609375,
|
981 |
-
"learning_rate": 9.266277030111474e-06,
|
982 |
-
"loss": 0.7074,
|
983 |
-
"step": 138
|
984 |
-
},
|
985 |
-
{
|
986 |
-
"epoch": 0.18684365286062338,
|
987 |
-
"grad_norm": 0.97265625,
|
988 |
-
"learning_rate": 9.25506244111128e-06,
|
989 |
-
"loss": 0.6745,
|
990 |
-
"step": 139
|
991 |
-
},
|
992 |
-
{
|
993 |
-
"epoch": 0.1881878518020667,
|
994 |
-
"grad_norm": 0.93359375,
|
995 |
-
"learning_rate": 9.243769689858167e-06,
|
996 |
-
"loss": 0.6627,
|
997 |
-
"step": 140
|
998 |
-
},
|
999 |
-
{
|
1000 |
-
"epoch": 0.18953205074351004,
|
1001 |
-
"grad_norm": 0.9609375,
|
1002 |
-
"learning_rate": 9.232398983791363e-06,
|
1003 |
-
"loss": 0.6563,
|
1004 |
-
"step": 141
|
1005 |
-
},
|
1006 |
-
{
|
1007 |
-
"epoch": 0.19087624968495337,
|
1008 |
-
"grad_norm": 1.0234375,
|
1009 |
-
"learning_rate": 9.220950531782069e-06,
|
1010 |
-
"loss": 0.688,
|
1011 |
-
"step": 142
|
1012 |
-
},
|
1013 |
-
{
|
1014 |
-
"epoch": 0.1922204486263967,
|
1015 |
-
"grad_norm": 1.046875,
|
1016 |
-
"learning_rate": 9.209424544129621e-06,
|
1017 |
-
"loss": 0.7021,
|
1018 |
-
"step": 143
|
1019 |
-
},
|
1020 |
-
{
|
1021 |
-
"epoch": 0.19356464756784003,
|
1022 |
-
"grad_norm": 0.87890625,
|
1023 |
-
"learning_rate": 9.197821232557625e-06,
|
1024 |
-
"loss": 0.6826,
|
1025 |
-
"step": 144
|
1026 |
-
},
|
1027 |
-
{
|
1028 |
-
"epoch": 0.19490884650928336,
|
1029 |
-
"grad_norm": 0.9140625,
|
1030 |
-
"learning_rate": 9.186140810210066e-06,
|
1031 |
-
"loss": 0.6904,
|
1032 |
-
"step": 145
|
1033 |
-
},
|
1034 |
-
{
|
1035 |
-
"epoch": 0.19625304545072672,
|
1036 |
-
"grad_norm": 1.0,
|
1037 |
-
"learning_rate": 9.1743834916474e-06,
|
1038 |
-
"loss": 0.6747,
|
1039 |
-
"step": 146
|
1040 |
-
},
|
1041 |
-
{
|
1042 |
-
"epoch": 0.19759724439217005,
|
1043 |
-
"grad_norm": 0.9609375,
|
1044 |
-
"learning_rate": 9.162549492842603e-06,
|
1045 |
-
"loss": 0.7073,
|
1046 |
-
"step": 147
|
1047 |
-
},
|
1048 |
-
{
|
1049 |
-
"epoch": 0.19894144333361338,
|
1050 |
-
"grad_norm": 0.9765625,
|
1051 |
-
"learning_rate": 9.150639031177211e-06,
|
1052 |
-
"loss": 0.6773,
|
1053 |
-
"step": 148
|
1054 |
-
},
|
1055 |
-
{
|
1056 |
-
"epoch": 0.2002856422750567,
|
1057 |
-
"grad_norm": 0.94921875,
|
1058 |
-
"learning_rate": 9.138652325437326e-06,
|
1059 |
-
"loss": 0.689,
|
1060 |
-
"step": 149
|
1061 |
-
},
|
1062 |
-
{
|
1063 |
-
"epoch": 0.20162984121650004,
|
1064 |
-
"grad_norm": 1.015625,
|
1065 |
-
"learning_rate": 9.12658959580959e-06,
|
1066 |
-
"loss": 0.6874,
|
1067 |
-
"step": 150
|
1068 |
-
},
|
1069 |
-
{
|
1070 |
-
"epoch": 0.20297404015794338,
|
1071 |
-
"grad_norm": 0.90234375,
|
1072 |
-
"learning_rate": 9.114451063877152e-06,
|
1073 |
-
"loss": 0.6617,
|
1074 |
-
"step": 151
|
1075 |
-
},
|
1076 |
-
{
|
1077 |
-
"epoch": 0.2043182390993867,
|
1078 |
-
"grad_norm": 1.0,
|
1079 |
-
"learning_rate": 9.102236952615588e-06,
|
1080 |
-
"loss": 0.6882,
|
1081 |
-
"step": 152
|
1082 |
-
},
|
1083 |
-
{
|
1084 |
-
"epoch": 0.20566243804083004,
|
1085 |
-
"grad_norm": 1.03125,
|
1086 |
-
"learning_rate": 9.08994748638881e-06,
|
1087 |
-
"loss": 0.6919,
|
1088 |
-
"step": 153
|
1089 |
-
},
|
1090 |
-
{
|
1091 |
-
"epoch": 0.20700663698227337,
|
1092 |
-
"grad_norm": 0.97265625,
|
1093 |
-
"learning_rate": 9.077582890944945e-06,
|
1094 |
-
"loss": 0.6713,
|
1095 |
-
"step": 154
|
1096 |
-
},
|
1097 |
-
{
|
1098 |
-
"epoch": 0.2083508359237167,
|
1099 |
-
"grad_norm": 0.921875,
|
1100 |
-
"learning_rate": 9.065143393412179e-06,
|
1101 |
-
"loss": 0.6504,
|
1102 |
-
"step": 155
|
1103 |
-
},
|
1104 |
-
{
|
1105 |
-
"epoch": 0.20969503486516006,
|
1106 |
-
"grad_norm": 0.98046875,
|
1107 |
-
"learning_rate": 9.052629222294605e-06,
|
1108 |
-
"loss": 0.6674,
|
1109 |
-
"step": 156
|
1110 |
-
},
|
1111 |
-
{
|
1112 |
-
"epoch": 0.2110392338066034,
|
1113 |
-
"grad_norm": 0.9375,
|
1114 |
-
"learning_rate": 9.040040607467999e-06,
|
1115 |
-
"loss": 0.7111,
|
1116 |
-
"step": 157
|
1117 |
-
},
|
1118 |
-
{
|
1119 |
-
"epoch": 0.21238343274804672,
|
1120 |
-
"grad_norm": 1.03125,
|
1121 |
-
"learning_rate": 9.02737778017562e-06,
|
1122 |
-
"loss": 0.6596,
|
1123 |
-
"step": 158
|
1124 |
-
},
|
1125 |
-
{
|
1126 |
-
"epoch": 0.21372763168949005,
|
1127 |
-
"grad_norm": 0.921875,
|
1128 |
-
"learning_rate": 9.014640973023951e-06,
|
1129 |
-
"loss": 0.6846,
|
1130 |
-
"step": 159
|
1131 |
-
},
|
1132 |
-
{
|
1133 |
-
"epoch": 0.21507183063093338,
|
1134 |
-
"grad_norm": 0.94921875,
|
1135 |
-
"learning_rate": 9.00183041997843e-06,
|
1136 |
-
"loss": 0.6788,
|
1137 |
-
"step": 160
|
1138 |
-
},
|
1139 |
-
{
|
1140 |
-
"epoch": 0.2164160295723767,
|
1141 |
-
"grad_norm": 0.890625,
|
1142 |
-
"learning_rate": 8.988946356359147e-06,
|
1143 |
-
"loss": 0.6888,
|
1144 |
-
"step": 161
|
1145 |
-
},
|
1146 |
-
{
|
1147 |
-
"epoch": 0.21776022851382004,
|
1148 |
-
"grad_norm": 0.92578125,
|
1149 |
-
"learning_rate": 8.97598901883653e-06,
|
1150 |
-
"loss": 0.6854,
|
1151 |
-
"step": 162
|
1152 |
-
},
|
1153 |
-
{
|
1154 |
-
"epoch": 0.21910442745526337,
|
1155 |
-
"grad_norm": 0.94140625,
|
1156 |
-
"learning_rate": 8.962958645426989e-06,
|
1157 |
-
"loss": 0.7137,
|
1158 |
-
"step": 163
|
1159 |
-
},
|
1160 |
-
{
|
1161 |
-
"epoch": 0.2204486263967067,
|
1162 |
-
"grad_norm": 0.8984375,
|
1163 |
-
"learning_rate": 8.949855475488549e-06,
|
1164 |
-
"loss": 0.7041,
|
1165 |
-
"step": 164
|
1166 |
-
},
|
1167 |
-
{
|
1168 |
-
"epoch": 0.22179282533815003,
|
1169 |
-
"grad_norm": 0.98046875,
|
1170 |
-
"learning_rate": 8.936679749716452e-06,
|
1171 |
-
"loss": 0.6573,
|
1172 |
-
"step": 165
|
1173 |
-
},
|
1174 |
-
{
|
1175 |
-
"epoch": 0.2231370242795934,
|
1176 |
-
"grad_norm": 0.8984375,
|
1177 |
-
"learning_rate": 8.923431710138735e-06,
|
1178 |
-
"loss": 0.6694,
|
1179 |
-
"step": 166
|
1180 |
-
},
|
1181 |
-
{
|
1182 |
-
"epoch": 0.22448122322103672,
|
1183 |
-
"grad_norm": 0.94140625,
|
1184 |
-
"learning_rate": 8.910111600111786e-06,
|
1185 |
-
"loss": 0.6904,
|
1186 |
-
"step": 167
|
1187 |
-
},
|
1188 |
-
{
|
1189 |
-
"epoch": 0.22582542216248006,
|
1190 |
-
"grad_norm": 0.953125,
|
1191 |
-
"learning_rate": 8.896719664315866e-06,
|
1192 |
-
"loss": 0.679,
|
1193 |
-
"step": 168
|
1194 |
-
},
|
1195 |
-
{
|
1196 |
-
"epoch": 0.22716962110392339,
|
1197 |
-
"grad_norm": 0.9140625,
|
1198 |
-
"learning_rate": 8.883256148750634e-06,
|
1199 |
-
"loss": 0.6631,
|
1200 |
-
"step": 169
|
1201 |
-
},
|
1202 |
-
{
|
1203 |
-
"epoch": 0.22851382004536672,
|
1204 |
-
"grad_norm": 0.9765625,
|
1205 |
-
"learning_rate": 8.869721300730596e-06,
|
1206 |
-
"loss": 0.6576,
|
1207 |
-
"step": 170
|
1208 |
-
},
|
1209 |
-
{
|
1210 |
-
"epoch": 0.22985801898681005,
|
1211 |
-
"grad_norm": 0.9296875,
|
1212 |
-
"learning_rate": 8.856115368880598e-06,
|
1213 |
-
"loss": 0.6714,
|
1214 |
-
"step": 171
|
1215 |
-
},
|
1216 |
-
{
|
1217 |
-
"epoch": 0.23120221792825338,
|
1218 |
-
"grad_norm": 0.97265625,
|
1219 |
-
"learning_rate": 8.842438603131232e-06,
|
1220 |
-
"loss": 0.6572,
|
1221 |
-
"step": 172
|
1222 |
-
},
|
1223 |
-
{
|
1224 |
-
"epoch": 0.2325464168696967,
|
1225 |
-
"grad_norm": 0.91796875,
|
1226 |
-
"learning_rate": 8.828691254714259e-06,
|
1227 |
-
"loss": 0.6798,
|
1228 |
-
"step": 173
|
1229 |
-
},
|
1230 |
-
{
|
1231 |
-
"epoch": 0.23389061581114004,
|
1232 |
-
"grad_norm": 0.90234375,
|
1233 |
-
"learning_rate": 8.814873576157988e-06,
|
1234 |
-
"loss": 0.6934,
|
1235 |
-
"step": 174
|
1236 |
-
},
|
1237 |
-
{
|
1238 |
-
"epoch": 0.23523481475258337,
|
1239 |
-
"grad_norm": 1.0390625,
|
1240 |
-
"learning_rate": 8.800985821282637e-06,
|
1241 |
-
"loss": 0.6648,
|
1242 |
-
"step": 175
|
1243 |
-
},
|
1244 |
-
{
|
1245 |
-
"epoch": 0.23657901369402673,
|
1246 |
-
"grad_norm": 0.8203125,
|
1247 |
-
"learning_rate": 8.787028245195676e-06,
|
1248 |
-
"loss": 0.6705,
|
1249 |
-
"step": 176
|
1250 |
-
},
|
1251 |
-
{
|
1252 |
-
"epoch": 0.23792321263547006,
|
1253 |
-
"grad_norm": 0.8984375,
|
1254 |
-
"learning_rate": 8.773001104287137e-06,
|
1255 |
-
"loss": 0.6661,
|
1256 |
-
"step": 177
|
1257 |
-
},
|
1258 |
-
{
|
1259 |
-
"epoch": 0.2392674115769134,
|
1260 |
-
"grad_norm": 0.94140625,
|
1261 |
-
"learning_rate": 8.758904656224904e-06,
|
1262 |
-
"loss": 0.6968,
|
1263 |
-
"step": 178
|
1264 |
-
},
|
1265 |
-
{
|
1266 |
-
"epoch": 0.24061161051835672,
|
1267 |
-
"grad_norm": 0.94140625,
|
1268 |
-
"learning_rate": 8.744739159949981e-06,
|
1269 |
-
"loss": 0.6817,
|
1270 |
-
"step": 179
|
1271 |
-
},
|
1272 |
-
{
|
1273 |
-
"epoch": 0.24195580945980005,
|
1274 |
-
"grad_norm": 0.9296875,
|
1275 |
-
"learning_rate": 8.730504875671732e-06,
|
1276 |
-
"loss": 0.6704,
|
1277 |
-
"step": 180
|
1278 |
-
},
|
1279 |
-
{
|
1280 |
-
"epoch": 0.24330000840124338,
|
1281 |
-
"grad_norm": 0.90625,
|
1282 |
-
"learning_rate": 8.71620206486311e-06,
|
1283 |
-
"loss": 0.6484,
|
1284 |
-
"step": 181
|
1285 |
-
},
|
1286 |
-
{
|
1287 |
-
"epoch": 0.24464420734268671,
|
1288 |
-
"grad_norm": 0.8515625,
|
1289 |
-
"learning_rate": 8.701830990255843e-06,
|
1290 |
-
"loss": 0.6711,
|
1291 |
-
"step": 182
|
1292 |
-
},
|
1293 |
-
{
|
1294 |
-
"epoch": 0.24598840628413005,
|
1295 |
-
"grad_norm": 1.0234375,
|
1296 |
-
"learning_rate": 8.687391915835617e-06,
|
1297 |
-
"loss": 0.6526,
|
1298 |
-
"step": 183
|
1299 |
-
},
|
1300 |
-
{
|
1301 |
-
"epoch": 0.24733260522557338,
|
1302 |
-
"grad_norm": 0.87890625,
|
1303 |
-
"learning_rate": 8.672885106837216e-06,
|
1304 |
-
"loss": 0.6907,
|
1305 |
-
"step": 184
|
1306 |
-
},
|
1307 |
-
{
|
1308 |
-
"epoch": 0.2486768041670167,
|
1309 |
-
"grad_norm": 0.9609375,
|
1310 |
-
"learning_rate": 8.658310829739666e-06,
|
1311 |
-
"loss": 0.7003,
|
1312 |
-
"step": 185
|
1313 |
-
},
|
1314 |
-
{
|
1315 |
-
"epoch": 0.25002100310846004,
|
1316 |
-
"grad_norm": 0.85546875,
|
1317 |
-
"learning_rate": 8.643669352261321e-06,
|
1318 |
-
"loss": 0.6353,
|
1319 |
-
"step": 186
|
1320 |
-
},
|
1321 |
-
{
|
1322 |
-
"epoch": 0.25136520204990337,
|
1323 |
-
"grad_norm": 0.921875,
|
1324 |
-
"learning_rate": 8.628960943354965e-06,
|
1325 |
-
"loss": 0.6566,
|
1326 |
-
"step": 187
|
1327 |
-
},
|
1328 |
-
{
|
1329 |
-
"epoch": 0.2527094009913467,
|
1330 |
-
"grad_norm": 0.92578125,
|
1331 |
-
"learning_rate": 8.614185873202852e-06,
|
1332 |
-
"loss": 0.6676,
|
1333 |
-
"step": 188
|
1334 |
-
},
|
1335 |
-
{
|
1336 |
-
"epoch": 0.25405359993279003,
|
1337 |
-
"grad_norm": 0.953125,
|
1338 |
-
"learning_rate": 8.599344413211755e-06,
|
1339 |
-
"loss": 0.6647,
|
1340 |
-
"step": 189
|
1341 |
-
},
|
1342 |
-
{
|
1343 |
-
"epoch": 0.25539779887423336,
|
1344 |
-
"grad_norm": 0.9140625,
|
1345 |
-
"learning_rate": 8.58443683600798e-06,
|
1346 |
-
"loss": 0.6771,
|
1347 |
-
"step": 190
|
1348 |
-
},
|
1349 |
-
{
|
1350 |
-
"epoch": 0.2567419978156767,
|
1351 |
-
"grad_norm": 0.90234375,
|
1352 |
-
"learning_rate": 8.569463415432356e-06,
|
1353 |
-
"loss": 0.6629,
|
1354 |
-
"step": 191
|
1355 |
-
},
|
1356 |
-
{
|
1357 |
-
"epoch": 0.2580861967571201,
|
1358 |
-
"grad_norm": 0.91015625,
|
1359 |
-
"learning_rate": 8.554424426535202e-06,
|
1360 |
-
"loss": 0.7327,
|
1361 |
-
"step": 192
|
1362 |
-
},
|
1363 |
-
{
|
1364 |
-
"epoch": 0.2594303956985634,
|
1365 |
-
"grad_norm": 0.94921875,
|
1366 |
-
"learning_rate": 8.539320145571277e-06,
|
1367 |
-
"loss": 0.6809,
|
1368 |
-
"step": 193
|
1369 |
-
},
|
1370 |
-
{
|
1371 |
-
"epoch": 0.26077459464000674,
|
1372 |
-
"grad_norm": 0.875,
|
1373 |
-
"learning_rate": 8.524150849994708e-06,
|
1374 |
-
"loss": 0.6895,
|
1375 |
-
"step": 194
|
1376 |
-
},
|
1377 |
-
{
|
1378 |
-
"epoch": 0.26211879358145007,
|
1379 |
-
"grad_norm": 0.9140625,
|
1380 |
-
"learning_rate": 8.50891681845389e-06,
|
1381 |
-
"loss": 0.6739,
|
1382 |
-
"step": 195
|
1383 |
-
},
|
1384 |
-
{
|
1385 |
-
"epoch": 0.2634629925228934,
|
1386 |
-
"grad_norm": 0.96875,
|
1387 |
-
"learning_rate": 8.493618330786365e-06,
|
1388 |
-
"loss": 0.7063,
|
1389 |
-
"step": 196
|
1390 |
-
},
|
1391 |
-
{
|
1392 |
-
"epoch": 0.26480719146433673,
|
1393 |
-
"grad_norm": 0.85546875,
|
1394 |
-
"learning_rate": 8.47825566801369e-06,
|
1395 |
-
"loss": 0.6676,
|
1396 |
-
"step": 197
|
1397 |
-
},
|
1398 |
-
{
|
1399 |
-
"epoch": 0.26615139040578006,
|
1400 |
-
"grad_norm": 0.81640625,
|
1401 |
-
"learning_rate": 8.462829112336266e-06,
|
1402 |
-
"loss": 0.6842,
|
1403 |
-
"step": 198
|
1404 |
-
},
|
1405 |
-
{
|
1406 |
-
"epoch": 0.2674955893472234,
|
1407 |
-
"grad_norm": 0.90625,
|
1408 |
-
"learning_rate": 8.44733894712816e-06,
|
1409 |
-
"loss": 0.7045,
|
1410 |
-
"step": 199
|
1411 |
-
},
|
1412 |
-
{
|
1413 |
-
"epoch": 0.2688397882886667,
|
1414 |
-
"grad_norm": 1.015625,
|
1415 |
-
"learning_rate": 8.431785456931898e-06,
|
1416 |
-
"loss": 0.6569,
|
1417 |
-
"step": 200
|
1418 |
-
},
|
1419 |
-
{
|
1420 |
-
"epoch": 0.27018398723011006,
|
1421 |
-
"grad_norm": 0.91015625,
|
1422 |
-
"learning_rate": 8.416168927453237e-06,
|
1423 |
-
"loss": 0.679,
|
1424 |
-
"step": 201
|
1425 |
-
},
|
1426 |
-
{
|
1427 |
-
"epoch": 0.2715281861715534,
|
1428 |
-
"grad_norm": 0.8671875,
|
1429 |
-
"learning_rate": 8.400489645555914e-06,
|
1430 |
-
"loss": 0.6452,
|
1431 |
-
"step": 202
|
1432 |
-
},
|
1433 |
-
{
|
1434 |
-
"epoch": 0.2728723851129967,
|
1435 |
-
"grad_norm": 0.9296875,
|
1436 |
-
"learning_rate": 8.384747899256386e-06,
|
1437 |
-
"loss": 0.6744,
|
1438 |
-
"step": 203
|
1439 |
-
},
|
1440 |
-
{
|
1441 |
-
"epoch": 0.27421658405444005,
|
1442 |
-
"grad_norm": 0.91015625,
|
1443 |
-
"learning_rate": 8.368943977718528e-06,
|
1444 |
-
"loss": 0.6501,
|
1445 |
-
"step": 204
|
1446 |
-
},
|
1447 |
-
{
|
1448 |
-
"epoch": 0.2755607829958834,
|
1449 |
-
"grad_norm": 0.890625,
|
1450 |
-
"learning_rate": 8.353078171248335e-06,
|
1451 |
-
"loss": 0.666,
|
1452 |
-
"step": 205
|
1453 |
-
},
|
1454 |
-
{
|
1455 |
-
"epoch": 0.2769049819373267,
|
1456 |
-
"grad_norm": 0.87109375,
|
1457 |
-
"learning_rate": 8.337150771288571e-06,
|
1458 |
-
"loss": 0.6663,
|
1459 |
-
"step": 206
|
1460 |
-
},
|
1461 |
-
{
|
1462 |
-
"epoch": 0.27824918087877004,
|
1463 |
-
"grad_norm": 0.9296875,
|
1464 |
-
"learning_rate": 8.32116207041343e-06,
|
1465 |
-
"loss": 0.6481,
|
1466 |
-
"step": 207
|
1467 |
-
},
|
1468 |
-
{
|
1469 |
-
"epoch": 0.2795933798202134,
|
1470 |
-
"grad_norm": 0.921875,
|
1471 |
-
"learning_rate": 8.30511236232316e-06,
|
1472 |
-
"loss": 0.6884,
|
1473 |
-
"step": 208
|
1474 |
-
},
|
1475 |
-
{
|
1476 |
-
"epoch": 0.2809375787616567,
|
1477 |
-
"grad_norm": 0.890625,
|
1478 |
-
"learning_rate": 8.289001941838659e-06,
|
1479 |
-
"loss": 0.6562,
|
1480 |
-
"step": 209
|
1481 |
-
},
|
1482 |
-
{
|
1483 |
-
"epoch": 0.28228177770310003,
|
1484 |
-
"grad_norm": 0.8984375,
|
1485 |
-
"learning_rate": 8.27283110489607e-06,
|
1486 |
-
"loss": 0.7106,
|
1487 |
-
"step": 210
|
1488 |
-
},
|
1489 |
-
{
|
1490 |
-
"epoch": 0.28362597664454336,
|
1491 |
-
"grad_norm": 0.98046875,
|
1492 |
-
"learning_rate": 8.256600148541339e-06,
|
1493 |
-
"loss": 0.689,
|
1494 |
-
"step": 211
|
1495 |
-
},
|
1496 |
-
{
|
1497 |
-
"epoch": 0.28497017558598675,
|
1498 |
-
"grad_norm": 0.96484375,
|
1499 |
-
"learning_rate": 8.240309370924758e-06,
|
1500 |
-
"loss": 0.6683,
|
1501 |
-
"step": 212
|
1502 |
-
},
|
1503 |
-
{
|
1504 |
-
"epoch": 0.2863143745274301,
|
1505 |
-
"grad_norm": 0.921875,
|
1506 |
-
"learning_rate": 8.223959071295492e-06,
|
1507 |
-
"loss": 0.6866,
|
1508 |
-
"step": 213
|
1509 |
-
},
|
1510 |
-
{
|
1511 |
-
"epoch": 0.2876585734688734,
|
1512 |
-
"grad_norm": 0.94140625,
|
1513 |
-
"learning_rate": 8.207549549996083e-06,
|
1514 |
-
"loss": 0.6688,
|
1515 |
-
"step": 214
|
1516 |
-
},
|
1517 |
-
{
|
1518 |
-
"epoch": 0.28900277241031674,
|
1519 |
-
"grad_norm": 0.91796875,
|
1520 |
-
"learning_rate": 8.191081108456922e-06,
|
1521 |
-
"loss": 0.6721,
|
1522 |
-
"step": 215
|
1523 |
-
},
|
1524 |
-
{
|
1525 |
-
"epoch": 0.2903469713517601,
|
1526 |
-
"grad_norm": 0.89453125,
|
1527 |
-
"learning_rate": 8.174554049190726e-06,
|
1528 |
-
"loss": 0.6579,
|
1529 |
-
"step": 216
|
1530 |
-
},
|
1531 |
-
{
|
1532 |
-
"epoch": 0.2916911702932034,
|
1533 |
-
"grad_norm": 0.9609375,
|
1534 |
-
"learning_rate": 8.157968675786971e-06,
|
1535 |
-
"loss": 0.6841,
|
1536 |
-
"step": 217
|
1537 |
-
},
|
1538 |
-
{
|
1539 |
-
"epoch": 0.29303536923464674,
|
1540 |
-
"grad_norm": 0.93359375,
|
1541 |
-
"learning_rate": 8.141325292906325e-06,
|
1542 |
-
"loss": 0.6739,
|
1543 |
-
"step": 218
|
1544 |
-
},
|
1545 |
-
{
|
1546 |
-
"epoch": 0.29437956817609007,
|
1547 |
-
"grad_norm": 0.875,
|
1548 |
-
"learning_rate": 8.124624206275041e-06,
|
1549 |
-
"loss": 0.6544,
|
1550 |
-
"step": 219
|
1551 |
-
},
|
1552 |
-
{
|
1553 |
-
"epoch": 0.2957237671175334,
|
1554 |
-
"grad_norm": 0.97265625,
|
1555 |
-
"learning_rate": 8.107865722679347e-06,
|
1556 |
-
"loss": 0.6525,
|
1557 |
-
"step": 220
|
1558 |
-
},
|
1559 |
-
{
|
1560 |
-
"epoch": 0.29706796605897673,
|
1561 |
-
"grad_norm": 0.890625,
|
1562 |
-
"learning_rate": 8.091050149959808e-06,
|
1563 |
-
"loss": 0.6394,
|
1564 |
-
"step": 221
|
1565 |
-
},
|
1566 |
-
{
|
1567 |
-
"epoch": 0.29841216500042006,
|
1568 |
-
"grad_norm": 0.87890625,
|
1569 |
-
"learning_rate": 8.074177797005677e-06,
|
1570 |
-
"loss": 0.6674,
|
1571 |
-
"step": 222
|
1572 |
-
},
|
1573 |
-
{
|
1574 |
-
"epoch": 0.2997563639418634,
|
1575 |
-
"grad_norm": 0.984375,
|
1576 |
-
"learning_rate": 8.057248973749216e-06,
|
1577 |
-
"loss": 0.6749,
|
1578 |
-
"step": 223
|
1579 |
-
},
|
1580 |
-
{
|
1581 |
-
"epoch": 0.3011005628833067,
|
1582 |
-
"grad_norm": 1.0234375,
|
1583 |
-
"learning_rate": 8.040263991159996e-06,
|
1584 |
-
"loss": 0.7223,
|
1585 |
-
"step": 224
|
1586 |
-
},
|
1587 |
-
{
|
1588 |
-
"epoch": 0.30244476182475005,
|
1589 |
-
"grad_norm": 0.859375,
|
1590 |
-
"learning_rate": 8.0232231612392e-06,
|
1591 |
-
"loss": 0.6861,
|
1592 |
-
"step": 225
|
1593 |
-
},
|
1594 |
-
{
|
1595 |
-
"epoch": 0.3037889607661934,
|
1596 |
-
"grad_norm": 0.9140625,
|
1597 |
-
"learning_rate": 8.006126797013884e-06,
|
1598 |
-
"loss": 0.6883,
|
1599 |
-
"step": 226
|
1600 |
-
},
|
1601 |
-
{
|
1602 |
-
"epoch": 0.3051331597076367,
|
1603 |
-
"grad_norm": 0.94140625,
|
1604 |
-
"learning_rate": 7.98897521253122e-06,
|
1605 |
-
"loss": 0.655,
|
1606 |
-
"step": 227
|
1607 |
-
},
|
1608 |
-
{
|
1609 |
-
"epoch": 0.30647735864908004,
|
1610 |
-
"grad_norm": 0.8984375,
|
1611 |
-
"learning_rate": 7.971768722852741e-06,
|
1612 |
-
"loss": 0.6696,
|
1613 |
-
"step": 228
|
1614 |
-
},
|
1615 |
-
{
|
1616 |
-
"epoch": 0.3078215575905234,
|
1617 |
-
"grad_norm": 0.875,
|
1618 |
-
"learning_rate": 7.954507644048544e-06,
|
1619 |
-
"loss": 0.6676,
|
1620 |
-
"step": 229
|
1621 |
-
},
|
1622 |
-
{
|
1623 |
-
"epoch": 0.3091657565319667,
|
1624 |
-
"grad_norm": 0.8828125,
|
1625 |
-
"learning_rate": 7.937192293191485e-06,
|
1626 |
-
"loss": 0.6508,
|
1627 |
-
"step": 230
|
1628 |
-
},
|
1629 |
-
{
|
1630 |
-
"epoch": 0.31050995547341004,
|
1631 |
-
"grad_norm": 0.9375,
|
1632 |
-
"learning_rate": 7.919822988351359e-06,
|
1633 |
-
"loss": 0.6676,
|
1634 |
-
"step": 231
|
1635 |
-
},
|
1636 |
-
{
|
1637 |
-
"epoch": 0.3118541544148534,
|
1638 |
-
"grad_norm": 0.859375,
|
1639 |
-
"learning_rate": 7.902400048589051e-06,
|
1640 |
-
"loss": 0.6512,
|
1641 |
-
"step": 232
|
1642 |
-
},
|
1643 |
-
{
|
1644 |
-
"epoch": 0.31319835335629675,
|
1645 |
-
"grad_norm": 0.94140625,
|
1646 |
-
"learning_rate": 7.884923793950684e-06,
|
1647 |
-
"loss": 0.6846,
|
1648 |
-
"step": 233
|
1649 |
-
},
|
1650 |
-
{
|
1651 |
-
"epoch": 0.3145425522977401,
|
1652 |
-
"grad_norm": 0.8515625,
|
1653 |
-
"learning_rate": 7.86739454546173e-06,
|
1654 |
-
"loss": 0.6644,
|
1655 |
-
"step": 234
|
1656 |
-
},
|
1657 |
-
{
|
1658 |
-
"epoch": 0.3158867512391834,
|
1659 |
-
"grad_norm": 0.8828125,
|
1660 |
-
"learning_rate": 7.849812625121122e-06,
|
1661 |
-
"loss": 0.6701,
|
1662 |
-
"step": 235
|
1663 |
-
},
|
1664 |
-
{
|
1665 |
-
"epoch": 0.31723095018062675,
|
1666 |
-
"grad_norm": 0.94921875,
|
1667 |
-
"learning_rate": 7.832178355895327e-06,
|
1668 |
-
"loss": 0.6749,
|
1669 |
-
"step": 236
|
1670 |
-
},
|
1671 |
-
{
|
1672 |
-
"epoch": 0.3185751491220701,
|
1673 |
-
"grad_norm": 0.87890625,
|
1674 |
-
"learning_rate": 7.81449206171243e-06,
|
1675 |
-
"loss": 0.6584,
|
1676 |
-
"step": 237
|
1677 |
-
},
|
1678 |
-
{
|
1679 |
-
"epoch": 0.3199193480635134,
|
1680 |
-
"grad_norm": 0.890625,
|
1681 |
-
"learning_rate": 7.796754067456168e-06,
|
1682 |
-
"loss": 0.6685,
|
1683 |
-
"step": 238
|
1684 |
-
},
|
1685 |
-
{
|
1686 |
-
"epoch": 0.32126354700495674,
|
1687 |
-
"grad_norm": 0.8671875,
|
1688 |
-
"learning_rate": 7.778964698959973e-06,
|
1689 |
-
"loss": 0.6659,
|
1690 |
-
"step": 239
|
1691 |
-
},
|
1692 |
-
{
|
1693 |
-
"epoch": 0.32260774594640007,
|
1694 |
-
"grad_norm": 0.8671875,
|
1695 |
-
"learning_rate": 7.761124283000982e-06,
|
1696 |
-
"loss": 0.6647,
|
1697 |
-
"step": 240
|
1698 |
-
},
|
1699 |
-
{
|
1700 |
-
"epoch": 0.3239519448878434,
|
1701 |
-
"grad_norm": 0.9140625,
|
1702 |
-
"learning_rate": 7.743233147294036e-06,
|
1703 |
-
"loss": 0.6793,
|
1704 |
-
"step": 241
|
1705 |
-
},
|
1706 |
-
{
|
1707 |
-
"epoch": 0.32529614382928673,
|
1708 |
-
"grad_norm": 0.87109375,
|
1709 |
-
"learning_rate": 7.725291620485653e-06,
|
1710 |
-
"loss": 0.6633,
|
1711 |
-
"step": 242
|
1712 |
-
},
|
1713 |
-
{
|
1714 |
-
"epoch": 0.32664034277073006,
|
1715 |
-
"grad_norm": 0.90625,
|
1716 |
-
"learning_rate": 7.707300032148004e-06,
|
1717 |
-
"loss": 0.6812,
|
1718 |
-
"step": 243
|
1719 |
-
},
|
1720 |
-
{
|
1721 |
-
"epoch": 0.3279845417121734,
|
1722 |
-
"grad_norm": 0.88671875,
|
1723 |
-
"learning_rate": 7.689258712772851e-06,
|
1724 |
-
"loss": 0.6734,
|
1725 |
-
"step": 244
|
1726 |
-
},
|
1727 |
-
{
|
1728 |
-
"epoch": 0.3293287406536167,
|
1729 |
-
"grad_norm": 0.859375,
|
1730 |
-
"learning_rate": 7.671167993765474e-06,
|
1731 |
-
"loss": 0.6775,
|
1732 |
-
"step": 245
|
1733 |
-
},
|
1734 |
-
{
|
1735 |
-
"epoch": 0.33067293959506006,
|
1736 |
-
"grad_norm": 0.84375,
|
1737 |
-
"learning_rate": 7.653028207438588e-06,
|
1738 |
-
"loss": 0.6723,
|
1739 |
-
"step": 246
|
1740 |
-
},
|
1741 |
-
{
|
1742 |
-
"epoch": 0.3320171385365034,
|
1743 |
-
"grad_norm": 0.90625,
|
1744 |
-
"learning_rate": 7.634839687006242e-06,
|
1745 |
-
"loss": 0.6814,
|
1746 |
-
"step": 247
|
1747 |
-
},
|
1748 |
-
{
|
1749 |
-
"epoch": 0.3333613374779467,
|
1750 |
-
"grad_norm": 0.984375,
|
1751 |
-
"learning_rate": 7.616602766577683e-06,
|
1752 |
-
"loss": 0.6816,
|
1753 |
-
"step": 248
|
1754 |
-
},
|
1755 |
-
{
|
1756 |
-
"epoch": 0.3333613374779467,
|
1757 |
-
"eval_loss": 2.7341208457946777,
|
1758 |
-
"eval_runtime": 5734.9058,
|
1759 |
-
"eval_samples_per_second": 3.861,
|
1760 |
-
"eval_steps_per_second": 0.483,
|
1761 |
-
"step": 248
|
1762 |
-
}
|
1763 |
-
],
|
1764 |
-
"logging_steps": 1,
|
1765 |
-
"max_steps": 743,
|
1766 |
-
"num_input_tokens_seen": 0,
|
1767 |
-
"num_train_epochs": 1,
|
1768 |
-
"save_steps": 248,
|
1769 |
-
"stateful_callbacks": {
|
1770 |
-
"TrainerControl": {
|
1771 |
-
"args": {
|
1772 |
-
"should_epoch_stop": false,
|
1773 |
-
"should_evaluate": false,
|
1774 |
-
"should_log": false,
|
1775 |
-
"should_save": true,
|
1776 |
-
"should_training_stop": false
|
1777 |
-
},
|
1778 |
-
"attributes": {}
|
1779 |
-
}
|
1780 |
-
},
|
1781 |
-
"total_flos": 5.854896016293102e+18,
|
1782 |
-
"train_batch_size": 8,
|
1783 |
-
"trial_name": null,
|
1784 |
-
"trial_params": null
|
1785 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|