Delete trainer_state.json
Browse files- trainer_state.json +0 -1293
trainer_state.json
DELETED
@@ -1,1293 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 2.4773519163763065,
|
5 |
-
"eval_steps": 500,
|
6 |
-
"global_step": 180,
|
7 |
-
"is_hyper_param_search": false,
|
8 |
-
"is_local_process_zero": true,
|
9 |
-
"is_world_process_zero": true,
|
10 |
-
"log_history": [
|
11 |
-
{
|
12 |
-
"epoch": 0.013937282229965157,
|
13 |
-
"grad_norm": 93.75749969482422,
|
14 |
-
"learning_rate": 2.2e-06,
|
15 |
-
"loss": 1.2675,
|
16 |
-
"step": 1
|
17 |
-
},
|
18 |
-
{
|
19 |
-
"epoch": 0.027874564459930314,
|
20 |
-
"grad_norm": 66.60814666748047,
|
21 |
-
"learning_rate": 4.4e-06,
|
22 |
-
"loss": 1.3582,
|
23 |
-
"step": 2
|
24 |
-
},
|
25 |
-
{
|
26 |
-
"epoch": 0.041811846689895474,
|
27 |
-
"grad_norm": 77.9828109741211,
|
28 |
-
"learning_rate": 6.5999999999999995e-06,
|
29 |
-
"loss": 1.2567,
|
30 |
-
"step": 3
|
31 |
-
},
|
32 |
-
{
|
33 |
-
"epoch": 0.05574912891986063,
|
34 |
-
"grad_norm": 17.5104923248291,
|
35 |
-
"learning_rate": 8.8e-06,
|
36 |
-
"loss": 1.1877,
|
37 |
-
"step": 4
|
38 |
-
},
|
39 |
-
{
|
40 |
-
"epoch": 0.06968641114982578,
|
41 |
-
"grad_norm": 10.045647621154785,
|
42 |
-
"learning_rate": 1.1e-05,
|
43 |
-
"loss": 1.1663,
|
44 |
-
"step": 5
|
45 |
-
},
|
46 |
-
{
|
47 |
-
"epoch": 0.08362369337979095,
|
48 |
-
"grad_norm": 12.031865119934082,
|
49 |
-
"learning_rate": 1.0999372667896238e-05,
|
50 |
-
"loss": 1.1769,
|
51 |
-
"step": 6
|
52 |
-
},
|
53 |
-
{
|
54 |
-
"epoch": 0.0975609756097561,
|
55 |
-
"grad_norm": 6.24857234954834,
|
56 |
-
"learning_rate": 1.0997490814692433e-05,
|
57 |
-
"loss": 1.1703,
|
58 |
-
"step": 7
|
59 |
-
},
|
60 |
-
{
|
61 |
-
"epoch": 0.11149825783972125,
|
62 |
-
"grad_norm": 4.496493816375732,
|
63 |
-
"learning_rate": 1.0994354869678378e-05,
|
64 |
-
"loss": 1.1608,
|
65 |
-
"step": 8
|
66 |
-
},
|
67 |
-
{
|
68 |
-
"epoch": 0.1254355400696864,
|
69 |
-
"grad_norm": 4.9181036949157715,
|
70 |
-
"learning_rate": 1.0989965548228246e-05,
|
71 |
-
"loss": 1.1165,
|
72 |
-
"step": 9
|
73 |
-
},
|
74 |
-
{
|
75 |
-
"epoch": 0.13937282229965156,
|
76 |
-
"grad_norm": 3.872187614440918,
|
77 |
-
"learning_rate": 1.0984323851637407e-05,
|
78 |
-
"loss": 1.1308,
|
79 |
-
"step": 10
|
80 |
-
},
|
81 |
-
{
|
82 |
-
"epoch": 0.15331010452961671,
|
83 |
-
"grad_norm": 4.341373920440674,
|
84 |
-
"learning_rate": 1.0977431066894e-05,
|
85 |
-
"loss": 1.1245,
|
86 |
-
"step": 11
|
87 |
-
},
|
88 |
-
{
|
89 |
-
"epoch": 0.1672473867595819,
|
90 |
-
"grad_norm": 3.867882013320923,
|
91 |
-
"learning_rate": 1.0969288766385357e-05,
|
92 |
-
"loss": 1.1486,
|
93 |
-
"step": 12
|
94 |
-
},
|
95 |
-
{
|
96 |
-
"epoch": 0.18118466898954705,
|
97 |
-
"grad_norm": 3.2795205116271973,
|
98 |
-
"learning_rate": 1.0959898807539295e-05,
|
99 |
-
"loss": 1.123,
|
100 |
-
"step": 13
|
101 |
-
},
|
102 |
-
{
|
103 |
-
"epoch": 0.1951219512195122,
|
104 |
-
"grad_norm": 3.429868698120117,
|
105 |
-
"learning_rate": 1.0949263332400415e-05,
|
106 |
-
"loss": 1.1169,
|
107 |
-
"step": 14
|
108 |
-
},
|
109 |
-
{
|
110 |
-
"epoch": 0.20905923344947736,
|
111 |
-
"grad_norm": 3.0912277698516846,
|
112 |
-
"learning_rate": 1.0937384767141438e-05,
|
113 |
-
"loss": 1.0832,
|
114 |
-
"step": 15
|
115 |
-
},
|
116 |
-
{
|
117 |
-
"epoch": 0.2229965156794425,
|
118 |
-
"grad_norm": 4.55407190322876,
|
119 |
-
"learning_rate": 1.0924265821509758e-05,
|
120 |
-
"loss": 1.1018,
|
121 |
-
"step": 16
|
122 |
-
},
|
123 |
-
{
|
124 |
-
"epoch": 0.23693379790940766,
|
125 |
-
"grad_norm": 3.4495749473571777,
|
126 |
-
"learning_rate": 1.090990948820929e-05,
|
127 |
-
"loss": 1.1186,
|
128 |
-
"step": 17
|
129 |
-
},
|
130 |
-
{
|
131 |
-
"epoch": 0.2508710801393728,
|
132 |
-
"grad_norm": 6.985757827758789,
|
133 |
-
"learning_rate": 1.0894319042217767e-05,
|
134 |
-
"loss": 1.1135,
|
135 |
-
"step": 18
|
136 |
-
},
|
137 |
-
{
|
138 |
-
"epoch": 0.26480836236933797,
|
139 |
-
"grad_norm": 3.7354323863983154,
|
140 |
-
"learning_rate": 1.0877498040039657e-05,
|
141 |
-
"loss": 1.1064,
|
142 |
-
"step": 19
|
143 |
-
},
|
144 |
-
{
|
145 |
-
"epoch": 0.2787456445993031,
|
146 |
-
"grad_norm": 4.375406265258789,
|
147 |
-
"learning_rate": 1.0859450318894847e-05,
|
148 |
-
"loss": 1.0767,
|
149 |
-
"step": 20
|
150 |
-
},
|
151 |
-
{
|
152 |
-
"epoch": 0.2926829268292683,
|
153 |
-
"grad_norm": 4.374971389770508,
|
154 |
-
"learning_rate": 1.0840179995843286e-05,
|
155 |
-
"loss": 1.1072,
|
156 |
-
"step": 21
|
157 |
-
},
|
158 |
-
{
|
159 |
-
"epoch": 0.30662020905923343,
|
160 |
-
"grad_norm": 3.129504442214966,
|
161 |
-
"learning_rate": 1.0819691466845815e-05,
|
162 |
-
"loss": 1.0414,
|
163 |
-
"step": 22
|
164 |
-
},
|
165 |
-
{
|
166 |
-
"epoch": 0.3205574912891986,
|
167 |
-
"grad_norm": 4.991695404052734,
|
168 |
-
"learning_rate": 1.079798940576134e-05,
|
169 |
-
"loss": 1.0773,
|
170 |
-
"step": 23
|
171 |
-
},
|
172 |
-
{
|
173 |
-
"epoch": 0.3344947735191638,
|
174 |
-
"grad_norm": 3.570375919342041,
|
175 |
-
"learning_rate": 1.0775078763280638e-05,
|
176 |
-
"loss": 1.0709,
|
177 |
-
"step": 24
|
178 |
-
},
|
179 |
-
{
|
180 |
-
"epoch": 0.34843205574912894,
|
181 |
-
"grad_norm": 7.718966007232666,
|
182 |
-
"learning_rate": 1.0750964765797001e-05,
|
183 |
-
"loss": 1.1019,
|
184 |
-
"step": 25
|
185 |
-
},
|
186 |
-
{
|
187 |
-
"epoch": 0.3623693379790941,
|
188 |
-
"grad_norm": 3.459160327911377,
|
189 |
-
"learning_rate": 1.072565291421398e-05,
|
190 |
-
"loss": 1.0505,
|
191 |
-
"step": 26
|
192 |
-
},
|
193 |
-
{
|
194 |
-
"epoch": 0.37630662020905925,
|
195 |
-
"grad_norm": 5.686498641967773,
|
196 |
-
"learning_rate": 1.069914898269052e-05,
|
197 |
-
"loss": 1.0897,
|
198 |
-
"step": 27
|
199 |
-
},
|
200 |
-
{
|
201 |
-
"epoch": 0.3902439024390244,
|
202 |
-
"grad_norm": 3.622288465499878,
|
203 |
-
"learning_rate": 1.067145901732376e-05,
|
204 |
-
"loss": 1.0817,
|
205 |
-
"step": 28
|
206 |
-
},
|
207 |
-
{
|
208 |
-
"epoch": 0.40418118466898956,
|
209 |
-
"grad_norm": 6.713829040527344,
|
210 |
-
"learning_rate": 1.0642589334769783e-05,
|
211 |
-
"loss": 1.0597,
|
212 |
-
"step": 29
|
213 |
-
},
|
214 |
-
{
|
215 |
-
"epoch": 0.4181184668989547,
|
216 |
-
"grad_norm": 3.2670044898986816,
|
217 |
-
"learning_rate": 1.061254652080265e-05,
|
218 |
-
"loss": 1.0631,
|
219 |
-
"step": 30
|
220 |
-
},
|
221 |
-
{
|
222 |
-
"epoch": 0.43205574912891986,
|
223 |
-
"grad_norm": 3.0448334217071533,
|
224 |
-
"learning_rate": 1.0581337428812077e-05,
|
225 |
-
"loss": 1.0697,
|
226 |
-
"step": 31
|
227 |
-
},
|
228 |
-
{
|
229 |
-
"epoch": 0.445993031358885,
|
230 |
-
"grad_norm": 4.583492755889893,
|
231 |
-
"learning_rate": 1.0548969178239997e-05,
|
232 |
-
"loss": 1.0564,
|
233 |
-
"step": 32
|
234 |
-
},
|
235 |
-
{
|
236 |
-
"epoch": 0.45993031358885017,
|
237 |
-
"grad_norm": 3.7002651691436768,
|
238 |
-
"learning_rate": 1.0515449152956496e-05,
|
239 |
-
"loss": 1.0852,
|
240 |
-
"step": 33
|
241 |
-
},
|
242 |
-
{
|
243 |
-
"epoch": 0.4738675958188153,
|
244 |
-
"grad_norm": 3.1550400257110596,
|
245 |
-
"learning_rate": 1.0480784999575381e-05,
|
246 |
-
"loss": 1.0576,
|
247 |
-
"step": 34
|
248 |
-
},
|
249 |
-
{
|
250 |
-
"epoch": 0.4878048780487805,
|
251 |
-
"grad_norm": 3.1954734325408936,
|
252 |
-
"learning_rate": 1.0444984625709842e-05,
|
253 |
-
"loss": 1.0965,
|
254 |
-
"step": 35
|
255 |
-
},
|
256 |
-
{
|
257 |
-
"epoch": 0.5017421602787456,
|
258 |
-
"grad_norm": 2.8396148681640625,
|
259 |
-
"learning_rate": 1.0408056198168555e-05,
|
260 |
-
"loss": 1.0401,
|
261 |
-
"step": 36
|
262 |
-
},
|
263 |
-
{
|
264 |
-
"epoch": 0.5156794425087108,
|
265 |
-
"grad_norm": 10.67453384399414,
|
266 |
-
"learning_rate": 1.0370008141092654e-05,
|
267 |
-
"loss": 1.0909,
|
268 |
-
"step": 37
|
269 |
-
},
|
270 |
-
{
|
271 |
-
"epoch": 0.5296167247386759,
|
272 |
-
"grad_norm": 3.554422378540039,
|
273 |
-
"learning_rate": 1.0330849134034033e-05,
|
274 |
-
"loss": 1.0534,
|
275 |
-
"step": 38
|
276 |
-
},
|
277 |
-
{
|
278 |
-
"epoch": 0.5435540069686411,
|
279 |
-
"grad_norm": 3.5049169063568115,
|
280 |
-
"learning_rate": 1.0290588109975334e-05,
|
281 |
-
"loss": 1.0366,
|
282 |
-
"step": 39
|
283 |
-
},
|
284 |
-
{
|
285 |
-
"epoch": 0.5574912891986062,
|
286 |
-
"grad_norm": 3.9955601692199707,
|
287 |
-
"learning_rate": 1.024923425329217e-05,
|
288 |
-
"loss": 1.0264,
|
289 |
-
"step": 40
|
290 |
-
},
|
291 |
-
{
|
292 |
-
"epoch": 0.5714285714285714,
|
293 |
-
"grad_norm": 3.2370924949645996,
|
294 |
-
"learning_rate": 1.0206796997657961e-05,
|
295 |
-
"loss": 1.0282,
|
296 |
-
"step": 41
|
297 |
-
},
|
298 |
-
{
|
299 |
-
"epoch": 0.5853658536585366,
|
300 |
-
"grad_norm": 3.0653462409973145,
|
301 |
-
"learning_rate": 1.0163286023891926e-05,
|
302 |
-
"loss": 1.066,
|
303 |
-
"step": 42
|
304 |
-
},
|
305 |
-
{
|
306 |
-
"epoch": 0.5993031358885017,
|
307 |
-
"grad_norm": 2.9653961658477783,
|
308 |
-
"learning_rate": 1.011871125775069e-05,
|
309 |
-
"loss": 1.068,
|
310 |
-
"step": 43
|
311 |
-
},
|
312 |
-
{
|
313 |
-
"epoch": 0.6132404181184669,
|
314 |
-
"grad_norm": 2.899329423904419,
|
315 |
-
"learning_rate": 1.0073082867664e-05,
|
316 |
-
"loss": 1.0075,
|
317 |
-
"step": 44
|
318 |
-
},
|
319 |
-
{
|
320 |
-
"epoch": 0.627177700348432,
|
321 |
-
"grad_norm": 3.231595277786255,
|
322 |
-
"learning_rate": 1.002641126241511e-05,
|
323 |
-
"loss": 1.0579,
|
324 |
-
"step": 45
|
325 |
-
},
|
326 |
-
{
|
327 |
-
"epoch": 0.6411149825783972,
|
328 |
-
"grad_norm": 3.7711715698242188,
|
329 |
-
"learning_rate": 9.978707088766316e-06,
|
330 |
-
"loss": 1.0511,
|
331 |
-
"step": 46
|
332 |
-
},
|
333 |
-
{
|
334 |
-
"epoch": 0.6550522648083623,
|
335 |
-
"grad_norm": 3.1697182655334473,
|
336 |
-
"learning_rate": 9.929981229030202e-06,
|
337 |
-
"loss": 1.0785,
|
338 |
-
"step": 47
|
339 |
-
},
|
340 |
-
{
|
341 |
-
"epoch": 0.6689895470383276,
|
342 |
-
"grad_norm": 2.8038201332092285,
|
343 |
-
"learning_rate": 9.88024479858717e-06,
|
344 |
-
"loss": 1.0389,
|
345 |
-
"step": 48
|
346 |
-
},
|
347 |
-
{
|
348 |
-
"epoch": 0.6829268292682927,
|
349 |
-
"grad_norm": 3.2184159755706787,
|
350 |
-
"learning_rate": 9.829509143349775e-06,
|
351 |
-
"loss": 1.0625,
|
352 |
-
"step": 49
|
353 |
-
},
|
354 |
-
{
|
355 |
-
"epoch": 0.6968641114982579,
|
356 |
-
"grad_norm": 2.9576430320739746,
|
357 |
-
"learning_rate": 9.77778583717451e-06,
|
358 |
-
"loss": 1.0217,
|
359 |
-
"step": 50
|
360 |
-
},
|
361 |
-
{
|
362 |
-
"epoch": 0.710801393728223,
|
363 |
-
"grad_norm": 2.918567657470703,
|
364 |
-
"learning_rate": 9.725086679221542e-06,
|
365 |
-
"loss": 1.0106,
|
366 |
-
"step": 51
|
367 |
-
},
|
368 |
-
{
|
369 |
-
"epoch": 0.7247386759581882,
|
370 |
-
"grad_norm": 2.8690834045410156,
|
371 |
-
"learning_rate": 9.671423691263104e-06,
|
372 |
-
"loss": 1.0427,
|
373 |
-
"step": 52
|
374 |
-
},
|
375 |
-
{
|
376 |
-
"epoch": 0.7386759581881533,
|
377 |
-
"grad_norm": 2.8049697875976562,
|
378 |
-
"learning_rate": 9.616809114941055e-06,
|
379 |
-
"loss": 1.0553,
|
380 |
-
"step": 53
|
381 |
-
},
|
382 |
-
{
|
383 |
-
"epoch": 0.7526132404181185,
|
384 |
-
"grad_norm": 3.5655879974365234,
|
385 |
-
"learning_rate": 9.561255408974332e-06,
|
386 |
-
"loss": 1.0308,
|
387 |
-
"step": 54
|
388 |
-
},
|
389 |
-
{
|
390 |
-
"epoch": 0.7665505226480837,
|
391 |
-
"grad_norm": 3.0086355209350586,
|
392 |
-
"learning_rate": 9.504775246316836e-06,
|
393 |
-
"loss": 1.0625,
|
394 |
-
"step": 55
|
395 |
-
},
|
396 |
-
{
|
397 |
-
"epoch": 0.7804878048780488,
|
398 |
-
"grad_norm": 3.2106897830963135,
|
399 |
-
"learning_rate": 9.447381511266482e-06,
|
400 |
-
"loss": 1.0582,
|
401 |
-
"step": 56
|
402 |
-
},
|
403 |
-
{
|
404 |
-
"epoch": 0.794425087108014,
|
405 |
-
"grad_norm": 3.013434410095215,
|
406 |
-
"learning_rate": 9.38908729652601e-06,
|
407 |
-
"loss": 1.0501,
|
408 |
-
"step": 57
|
409 |
-
},
|
410 |
-
{
|
411 |
-
"epoch": 0.8083623693379791,
|
412 |
-
"grad_norm": 3.194031000137329,
|
413 |
-
"learning_rate": 9.32990590021629e-06,
|
414 |
-
"loss": 1.0385,
|
415 |
-
"step": 58
|
416 |
-
},
|
417 |
-
{
|
418 |
-
"epoch": 0.8222996515679443,
|
419 |
-
"grad_norm": 2.9916939735412598,
|
420 |
-
"learning_rate": 9.269850822842717e-06,
|
421 |
-
"loss": 0.9978,
|
422 |
-
"step": 59
|
423 |
-
},
|
424 |
-
{
|
425 |
-
"epoch": 0.8362369337979094,
|
426 |
-
"grad_norm": 2.90720796585083,
|
427 |
-
"learning_rate": 9.208935764215487e-06,
|
428 |
-
"loss": 1.038,
|
429 |
-
"step": 60
|
430 |
-
},
|
431 |
-
{
|
432 |
-
"epoch": 0.8501742160278746,
|
433 |
-
"grad_norm": 2.983062505722046,
|
434 |
-
"learning_rate": 9.147174620324374e-06,
|
435 |
-
"loss": 1.0137,
|
436 |
-
"step": 61
|
437 |
-
},
|
438 |
-
{
|
439 |
-
"epoch": 0.8641114982578397,
|
440 |
-
"grad_norm": 3.021130084991455,
|
441 |
-
"learning_rate": 9.084581480168767e-06,
|
442 |
-
"loss": 1.0388,
|
443 |
-
"step": 62
|
444 |
-
},
|
445 |
-
{
|
446 |
-
"epoch": 0.8780487804878049,
|
447 |
-
"grad_norm": 3.831815242767334,
|
448 |
-
"learning_rate": 9.021170622543684e-06,
|
449 |
-
"loss": 1.0482,
|
450 |
-
"step": 63
|
451 |
-
},
|
452 |
-
{
|
453 |
-
"epoch": 0.89198606271777,
|
454 |
-
"grad_norm": 3.5826456546783447,
|
455 |
-
"learning_rate": 8.956956512782476e-06,
|
456 |
-
"loss": 1.0494,
|
457 |
-
"step": 64
|
458 |
-
},
|
459 |
-
{
|
460 |
-
"epoch": 0.9059233449477352,
|
461 |
-
"grad_norm": 3.445178747177124,
|
462 |
-
"learning_rate": 8.891953799456987e-06,
|
463 |
-
"loss": 0.9976,
|
464 |
-
"step": 65
|
465 |
-
},
|
466 |
-
{
|
467 |
-
"epoch": 0.9198606271777003,
|
468 |
-
"grad_norm": 2.8443377017974854,
|
469 |
-
"learning_rate": 8.826177311035906e-06,
|
470 |
-
"loss": 1.0646,
|
471 |
-
"step": 66
|
472 |
-
},
|
473 |
-
{
|
474 |
-
"epoch": 0.9337979094076655,
|
475 |
-
"grad_norm": 2.8210973739624023,
|
476 |
-
"learning_rate": 8.759642052502092e-06,
|
477 |
-
"loss": 1.0602,
|
478 |
-
"step": 67
|
479 |
-
},
|
480 |
-
{
|
481 |
-
"epoch": 0.9477351916376306,
|
482 |
-
"grad_norm": 2.9282021522521973,
|
483 |
-
"learning_rate": 8.692363201929623e-06,
|
484 |
-
"loss": 1.041,
|
485 |
-
"step": 68
|
486 |
-
},
|
487 |
-
{
|
488 |
-
"epoch": 0.9616724738675958,
|
489 |
-
"grad_norm": 2.916203260421753,
|
490 |
-
"learning_rate": 8.624356107021355e-06,
|
491 |
-
"loss": 1.035,
|
492 |
-
"step": 69
|
493 |
-
},
|
494 |
-
{
|
495 |
-
"epoch": 0.975609756097561,
|
496 |
-
"grad_norm": 2.7474899291992188,
|
497 |
-
"learning_rate": 8.555636281607811e-06,
|
498 |
-
"loss": 1.0443,
|
499 |
-
"step": 70
|
500 |
-
},
|
501 |
-
{
|
502 |
-
"epoch": 0.9895470383275261,
|
503 |
-
"grad_norm": 3.2409560680389404,
|
504 |
-
"learning_rate": 8.486219402108133e-06,
|
505 |
-
"loss": 1.0502,
|
506 |
-
"step": 71
|
507 |
-
},
|
508 |
-
{
|
509 |
-
"epoch": 1.0034843205574913,
|
510 |
-
"grad_norm": 3.2006499767303467,
|
511 |
-
"learning_rate": 8.416121303953973e-06,
|
512 |
-
"loss": 1.0337,
|
513 |
-
"step": 72
|
514 |
-
},
|
515 |
-
{
|
516 |
-
"epoch": 1.0017421602787457,
|
517 |
-
"grad_norm": 3.021101713180542,
|
518 |
-
"learning_rate": 8.345357977977113e-06,
|
519 |
-
"loss": 1.0164,
|
520 |
-
"step": 73
|
521 |
-
},
|
522 |
-
{
|
523 |
-
"epoch": 1.0156794425087108,
|
524 |
-
"grad_norm": 3.0675883293151855,
|
525 |
-
"learning_rate": 8.273945566761604e-06,
|
526 |
-
"loss": 0.9739,
|
527 |
-
"step": 74
|
528 |
-
},
|
529 |
-
{
|
530 |
-
"epoch": 1.029616724738676,
|
531 |
-
"grad_norm": 2.925316095352173,
|
532 |
-
"learning_rate": 8.201900360961325e-06,
|
533 |
-
"loss": 0.9607,
|
534 |
-
"step": 75
|
535 |
-
},
|
536 |
-
{
|
537 |
-
"epoch": 1.043554006968641,
|
538 |
-
"grad_norm": 3.2823657989501953,
|
539 |
-
"learning_rate": 8.12923879558374e-06,
|
540 |
-
"loss": 0.9313,
|
541 |
-
"step": 76
|
542 |
-
},
|
543 |
-
{
|
544 |
-
"epoch": 1.0574912891986064,
|
545 |
-
"grad_norm": 2.809762477874756,
|
546 |
-
"learning_rate": 8.055977446240727e-06,
|
547 |
-
"loss": 0.9682,
|
548 |
-
"step": 77
|
549 |
-
},
|
550 |
-
{
|
551 |
-
"epoch": 1.0714285714285714,
|
552 |
-
"grad_norm": 3.0586483478546143,
|
553 |
-
"learning_rate": 7.982133025367346e-06,
|
554 |
-
"loss": 0.9674,
|
555 |
-
"step": 78
|
556 |
-
},
|
557 |
-
{
|
558 |
-
"epoch": 1.0853658536585367,
|
559 |
-
"grad_norm": 2.9541120529174805,
|
560 |
-
"learning_rate": 7.907722378409371e-06,
|
561 |
-
"loss": 0.9691,
|
562 |
-
"step": 79
|
563 |
-
},
|
564 |
-
{
|
565 |
-
"epoch": 1.0993031358885017,
|
566 |
-
"grad_norm": 2.965197801589966,
|
567 |
-
"learning_rate": 7.83276247998052e-06,
|
568 |
-
"loss": 0.9537,
|
569 |
-
"step": 80
|
570 |
-
},
|
571 |
-
{
|
572 |
-
"epoch": 1.113240418118467,
|
573 |
-
"grad_norm": 3.096423625946045,
|
574 |
-
"learning_rate": 7.757270429990162e-06,
|
575 |
-
"loss": 0.9631,
|
576 |
-
"step": 81
|
577 |
-
},
|
578 |
-
{
|
579 |
-
"epoch": 1.127177700348432,
|
580 |
-
"grad_norm": 3.0303800106048584,
|
581 |
-
"learning_rate": 7.681263449742493e-06,
|
582 |
-
"loss": 0.9774,
|
583 |
-
"step": 82
|
584 |
-
},
|
585 |
-
{
|
586 |
-
"epoch": 1.1411149825783973,
|
587 |
-
"grad_norm": 3.2755138874053955,
|
588 |
-
"learning_rate": 7.604758878007994e-06,
|
589 |
-
"loss": 0.9589,
|
590 |
-
"step": 83
|
591 |
-
},
|
592 |
-
{
|
593 |
-
"epoch": 1.1550522648083623,
|
594 |
-
"grad_norm": 3.0583481788635254,
|
595 |
-
"learning_rate": 7.527774167068094e-06,
|
596 |
-
"loss": 0.9313,
|
597 |
-
"step": 84
|
598 |
-
},
|
599 |
-
{
|
600 |
-
"epoch": 1.1689895470383276,
|
601 |
-
"grad_norm": 2.8140718936920166,
|
602 |
-
"learning_rate": 7.4503268787339455e-06,
|
603 |
-
"loss": 0.9137,
|
604 |
-
"step": 85
|
605 |
-
},
|
606 |
-
{
|
607 |
-
"epoch": 1.1829268292682926,
|
608 |
-
"grad_norm": 3.0150392055511475,
|
609 |
-
"learning_rate": 7.372434680340213e-06,
|
610 |
-
"loss": 0.9324,
|
611 |
-
"step": 86
|
612 |
-
},
|
613 |
-
{
|
614 |
-
"epoch": 1.1968641114982579,
|
615 |
-
"grad_norm": 3.1816203594207764,
|
616 |
-
"learning_rate": 7.294115340714782e-06,
|
617 |
-
"loss": 0.9835,
|
618 |
-
"step": 87
|
619 |
-
},
|
620 |
-
{
|
621 |
-
"epoch": 1.210801393728223,
|
622 |
-
"grad_norm": 3.011570930480957,
|
623 |
-
"learning_rate": 7.215386726125319e-06,
|
624 |
-
"loss": 0.9395,
|
625 |
-
"step": 88
|
626 |
-
},
|
627 |
-
{
|
628 |
-
"epoch": 1.2247386759581882,
|
629 |
-
"grad_norm": 4.090625286102295,
|
630 |
-
"learning_rate": 7.1362667962036166e-06,
|
631 |
-
"loss": 0.9682,
|
632 |
-
"step": 89
|
633 |
-
},
|
634 |
-
{
|
635 |
-
"epoch": 1.2386759581881532,
|
636 |
-
"grad_norm": 3.290282964706421,
|
637 |
-
"learning_rate": 7.056773599848612e-06,
|
638 |
-
"loss": 0.9487,
|
639 |
-
"step": 90
|
640 |
-
},
|
641 |
-
{
|
642 |
-
"epoch": 1.2526132404181185,
|
643 |
-
"grad_norm": 3.567591667175293,
|
644 |
-
"learning_rate": 6.976925271109072e-06,
|
645 |
-
"loss": 0.9459,
|
646 |
-
"step": 91
|
647 |
-
},
|
648 |
-
{
|
649 |
-
"epoch": 1.2665505226480835,
|
650 |
-
"grad_norm": 2.817159414291382,
|
651 |
-
"learning_rate": 6.8967400250468335e-06,
|
652 |
-
"loss": 0.968,
|
653 |
-
"step": 92
|
654 |
-
},
|
655 |
-
{
|
656 |
-
"epoch": 1.2804878048780488,
|
657 |
-
"grad_norm": 3.8883535861968994,
|
658 |
-
"learning_rate": 6.816236153581568e-06,
|
659 |
-
"loss": 0.938,
|
660 |
-
"step": 93
|
661 |
-
},
|
662 |
-
{
|
663 |
-
"epoch": 1.294425087108014,
|
664 |
-
"grad_norm": 2.920037269592285,
|
665 |
-
"learning_rate": 6.735432021318023e-06,
|
666 |
-
"loss": 0.9401,
|
667 |
-
"step": 94
|
668 |
-
},
|
669 |
-
{
|
670 |
-
"epoch": 1.3083623693379791,
|
671 |
-
"grad_norm": 2.851327896118164,
|
672 |
-
"learning_rate": 6.654346061356661e-06,
|
673 |
-
"loss": 0.9636,
|
674 |
-
"step": 95
|
675 |
-
},
|
676 |
-
{
|
677 |
-
"epoch": 1.3222996515679442,
|
678 |
-
"grad_norm": 3.1404778957366943,
|
679 |
-
"learning_rate": 6.572996771088706e-06,
|
680 |
-
"loss": 0.9665,
|
681 |
-
"step": 96
|
682 |
-
},
|
683 |
-
{
|
684 |
-
"epoch": 1.3362369337979094,
|
685 |
-
"grad_norm": 4.205505847930908,
|
686 |
-
"learning_rate": 6.491402707976482e-06,
|
687 |
-
"loss": 0.945,
|
688 |
-
"step": 97
|
689 |
-
},
|
690 |
-
{
|
691 |
-
"epoch": 1.3501742160278747,
|
692 |
-
"grad_norm": 2.9891207218170166,
|
693 |
-
"learning_rate": 6.409582485320087e-06,
|
694 |
-
"loss": 0.9554,
|
695 |
-
"step": 98
|
696 |
-
},
|
697 |
-
{
|
698 |
-
"epoch": 1.3641114982578397,
|
699 |
-
"grad_norm": 2.9016053676605225,
|
700 |
-
"learning_rate": 6.327554768011307e-06,
|
701 |
-
"loss": 0.9613,
|
702 |
-
"step": 99
|
703 |
-
},
|
704 |
-
{
|
705 |
-
"epoch": 1.3780487804878048,
|
706 |
-
"grad_norm": 2.954751968383789,
|
707 |
-
"learning_rate": 6.245338268275765e-06,
|
708 |
-
"loss": 0.9358,
|
709 |
-
"step": 100
|
710 |
-
},
|
711 |
-
{
|
712 |
-
"epoch": 1.39198606271777,
|
713 |
-
"grad_norm": 3.2333548069000244,
|
714 |
-
"learning_rate": 6.162951741404276e-06,
|
715 |
-
"loss": 0.9573,
|
716 |
-
"step": 101
|
717 |
-
},
|
718 |
-
{
|
719 |
-
"epoch": 1.4059233449477353,
|
720 |
-
"grad_norm": 3.4226996898651123,
|
721 |
-
"learning_rate": 6.080413981474379e-06,
|
722 |
-
"loss": 0.9294,
|
723 |
-
"step": 102
|
724 |
-
},
|
725 |
-
{
|
726 |
-
"epoch": 1.4198606271777003,
|
727 |
-
"grad_norm": 3.169379711151123,
|
728 |
-
"learning_rate": 5.9977438170630085e-06,
|
729 |
-
"loss": 0.9195,
|
730 |
-
"step": 103
|
731 |
-
},
|
732 |
-
{
|
733 |
-
"epoch": 1.4337979094076654,
|
734 |
-
"grad_norm": 3.147132396697998,
|
735 |
-
"learning_rate": 5.914960106951313e-06,
|
736 |
-
"loss": 0.9524,
|
737 |
-
"step": 104
|
738 |
-
},
|
739 |
-
{
|
740 |
-
"epoch": 1.4477351916376306,
|
741 |
-
"grad_norm": 2.7678744792938232,
|
742 |
-
"learning_rate": 5.832081735822573e-06,
|
743 |
-
"loss": 0.9234,
|
744 |
-
"step": 105
|
745 |
-
},
|
746 |
-
{
|
747 |
-
"epoch": 1.461672473867596,
|
748 |
-
"grad_norm": 2.8116793632507324,
|
749 |
-
"learning_rate": 5.749127609954215e-06,
|
750 |
-
"loss": 0.9619,
|
751 |
-
"step": 106
|
752 |
-
},
|
753 |
-
{
|
754 |
-
"epoch": 1.475609756097561,
|
755 |
-
"grad_norm": 3.6426944732666016,
|
756 |
-
"learning_rate": 5.666116652904889e-06,
|
757 |
-
"loss": 0.9435,
|
758 |
-
"step": 107
|
759 |
-
},
|
760 |
-
{
|
761 |
-
"epoch": 1.489547038327526,
|
762 |
-
"grad_norm": 4.532324314117432,
|
763 |
-
"learning_rate": 5.5830678011976225e-06,
|
764 |
-
"loss": 0.9538,
|
765 |
-
"step": 108
|
766 |
-
},
|
767 |
-
{
|
768 |
-
"epoch": 1.5034843205574913,
|
769 |
-
"grad_norm": 3.7293591499328613,
|
770 |
-
"learning_rate": 5.5e-06,
|
771 |
-
"loss": 0.9595,
|
772 |
-
"step": 109
|
773 |
-
},
|
774 |
-
{
|
775 |
-
"epoch": 1.5174216027874565,
|
776 |
-
"grad_norm": 2.8792057037353516,
|
777 |
-
"learning_rate": 5.416932198802378e-06,
|
778 |
-
"loss": 0.9498,
|
779 |
-
"step": 110
|
780 |
-
},
|
781 |
-
{
|
782 |
-
"epoch": 1.5313588850174216,
|
783 |
-
"grad_norm": 2.9074580669403076,
|
784 |
-
"learning_rate": 5.333883347095112e-06,
|
785 |
-
"loss": 0.9572,
|
786 |
-
"step": 111
|
787 |
-
},
|
788 |
-
{
|
789 |
-
"epoch": 1.5452961672473866,
|
790 |
-
"grad_norm": 4.807355880737305,
|
791 |
-
"learning_rate": 5.250872390045787e-06,
|
792 |
-
"loss": 0.9527,
|
793 |
-
"step": 112
|
794 |
-
},
|
795 |
-
{
|
796 |
-
"epoch": 1.5592334494773519,
|
797 |
-
"grad_norm": 4.229637622833252,
|
798 |
-
"learning_rate": 5.167918264177426e-06,
|
799 |
-
"loss": 0.9275,
|
800 |
-
"step": 113
|
801 |
-
},
|
802 |
-
{
|
803 |
-
"epoch": 1.5731707317073171,
|
804 |
-
"grad_norm": 3.07944393157959,
|
805 |
-
"learning_rate": 5.085039893048687e-06,
|
806 |
-
"loss": 0.9565,
|
807 |
-
"step": 114
|
808 |
-
},
|
809 |
-
{
|
810 |
-
"epoch": 1.5871080139372822,
|
811 |
-
"grad_norm": 2.8650782108306885,
|
812 |
-
"learning_rate": 5.002256182936992e-06,
|
813 |
-
"loss": 0.9493,
|
814 |
-
"step": 115
|
815 |
-
},
|
816 |
-
{
|
817 |
-
"epoch": 1.6010452961672472,
|
818 |
-
"grad_norm": 2.928194284439087,
|
819 |
-
"learning_rate": 4.919586018525621e-06,
|
820 |
-
"loss": 0.9513,
|
821 |
-
"step": 116
|
822 |
-
},
|
823 |
-
{
|
824 |
-
"epoch": 1.6149825783972127,
|
825 |
-
"grad_norm": 3.433330535888672,
|
826 |
-
"learning_rate": 4.837048258595723e-06,
|
827 |
-
"loss": 0.978,
|
828 |
-
"step": 117
|
829 |
-
},
|
830 |
-
{
|
831 |
-
"epoch": 1.6289198606271778,
|
832 |
-
"grad_norm": 2.8581085205078125,
|
833 |
-
"learning_rate": 4.754661731724237e-06,
|
834 |
-
"loss": 0.9366,
|
835 |
-
"step": 118
|
836 |
-
},
|
837 |
-
{
|
838 |
-
"epoch": 1.6428571428571428,
|
839 |
-
"grad_norm": 2.897559404373169,
|
840 |
-
"learning_rate": 4.672445231988693e-06,
|
841 |
-
"loss": 0.9641,
|
842 |
-
"step": 119
|
843 |
-
},
|
844 |
-
{
|
845 |
-
"epoch": 1.656794425087108,
|
846 |
-
"grad_norm": 2.8979365825653076,
|
847 |
-
"learning_rate": 4.590417514679912e-06,
|
848 |
-
"loss": 0.9316,
|
849 |
-
"step": 120
|
850 |
-
},
|
851 |
-
{
|
852 |
-
"epoch": 1.6707317073170733,
|
853 |
-
"grad_norm": 3.090183734893799,
|
854 |
-
"learning_rate": 4.508597292023518e-06,
|
855 |
-
"loss": 0.9545,
|
856 |
-
"step": 121
|
857 |
-
},
|
858 |
-
{
|
859 |
-
"epoch": 1.6846689895470384,
|
860 |
-
"grad_norm": 2.9215645790100098,
|
861 |
-
"learning_rate": 4.427003228911295e-06,
|
862 |
-
"loss": 0.9169,
|
863 |
-
"step": 122
|
864 |
-
},
|
865 |
-
{
|
866 |
-
"epoch": 1.6986062717770034,
|
867 |
-
"grad_norm": 4.066997528076172,
|
868 |
-
"learning_rate": 4.345653938643339e-06,
|
869 |
-
"loss": 0.9357,
|
870 |
-
"step": 123
|
871 |
-
},
|
872 |
-
{
|
873 |
-
"epoch": 1.7125435540069687,
|
874 |
-
"grad_norm": 3.2360036373138428,
|
875 |
-
"learning_rate": 4.2645679786819796e-06,
|
876 |
-
"loss": 0.9205,
|
877 |
-
"step": 124
|
878 |
-
},
|
879 |
-
{
|
880 |
-
"epoch": 1.726480836236934,
|
881 |
-
"grad_norm": 2.890538215637207,
|
882 |
-
"learning_rate": 4.1837638464184334e-06,
|
883 |
-
"loss": 0.9355,
|
884 |
-
"step": 125
|
885 |
-
},
|
886 |
-
{
|
887 |
-
"epoch": 1.740418118466899,
|
888 |
-
"grad_norm": 3.4630892276763916,
|
889 |
-
"learning_rate": 4.103259974953166e-06,
|
890 |
-
"loss": 0.9263,
|
891 |
-
"step": 126
|
892 |
-
},
|
893 |
-
{
|
894 |
-
"epoch": 1.754355400696864,
|
895 |
-
"grad_norm": 3.326503276824951,
|
896 |
-
"learning_rate": 4.023074728890927e-06,
|
897 |
-
"loss": 0.9154,
|
898 |
-
"step": 127
|
899 |
-
},
|
900 |
-
{
|
901 |
-
"epoch": 1.7682926829268293,
|
902 |
-
"grad_norm": 3.6348044872283936,
|
903 |
-
"learning_rate": 3.943226400151388e-06,
|
904 |
-
"loss": 0.944,
|
905 |
-
"step": 128
|
906 |
-
},
|
907 |
-
{
|
908 |
-
"epoch": 1.7822299651567945,
|
909 |
-
"grad_norm": 3.0831995010375977,
|
910 |
-
"learning_rate": 3.863733203796385e-06,
|
911 |
-
"loss": 0.9457,
|
912 |
-
"step": 129
|
913 |
-
},
|
914 |
-
{
|
915 |
-
"epoch": 1.7961672473867596,
|
916 |
-
"grad_norm": 3.1765310764312744,
|
917 |
-
"learning_rate": 3.784613273874681e-06,
|
918 |
-
"loss": 0.9329,
|
919 |
-
"step": 130
|
920 |
-
},
|
921 |
-
{
|
922 |
-
"epoch": 1.8101045296167246,
|
923 |
-
"grad_norm": 2.767540693283081,
|
924 |
-
"learning_rate": 3.70588465928522e-06,
|
925 |
-
"loss": 0.9285,
|
926 |
-
"step": 131
|
927 |
-
},
|
928 |
-
{
|
929 |
-
"epoch": 1.82404181184669,
|
930 |
-
"grad_norm": 2.718041181564331,
|
931 |
-
"learning_rate": 3.6275653196597856e-06,
|
932 |
-
"loss": 0.9767,
|
933 |
-
"step": 132
|
934 |
-
},
|
935 |
-
{
|
936 |
-
"epoch": 1.8379790940766552,
|
937 |
-
"grad_norm": 4.337693691253662,
|
938 |
-
"learning_rate": 3.5496731212660538e-06,
|
939 |
-
"loss": 0.9351,
|
940 |
-
"step": 133
|
941 |
-
},
|
942 |
-
{
|
943 |
-
"epoch": 1.8519163763066202,
|
944 |
-
"grad_norm": 3.106288194656372,
|
945 |
-
"learning_rate": 3.472225832931907e-06,
|
946 |
-
"loss": 0.9554,
|
947 |
-
"step": 134
|
948 |
-
},
|
949 |
-
{
|
950 |
-
"epoch": 1.8658536585365852,
|
951 |
-
"grad_norm": 3.2049343585968018,
|
952 |
-
"learning_rate": 3.3952411219920066e-06,
|
953 |
-
"loss": 0.9601,
|
954 |
-
"step": 135
|
955 |
-
},
|
956 |
-
{
|
957 |
-
"epoch": 1.8797909407665505,
|
958 |
-
"grad_norm": 4.705850601196289,
|
959 |
-
"learning_rate": 3.318736550257507e-06,
|
960 |
-
"loss": 0.9595,
|
961 |
-
"step": 136
|
962 |
-
},
|
963 |
-
{
|
964 |
-
"epoch": 1.8937282229965158,
|
965 |
-
"grad_norm": 3.335909128189087,
|
966 |
-
"learning_rate": 3.2427295700098385e-06,
|
967 |
-
"loss": 0.9889,
|
968 |
-
"step": 137
|
969 |
-
},
|
970 |
-
{
|
971 |
-
"epoch": 1.9076655052264808,
|
972 |
-
"grad_norm": 2.938302993774414,
|
973 |
-
"learning_rate": 3.1672375200194797e-06,
|
974 |
-
"loss": 0.9306,
|
975 |
-
"step": 138
|
976 |
-
},
|
977 |
-
{
|
978 |
-
"epoch": 1.9216027874564459,
|
979 |
-
"grad_norm": 3.15813946723938,
|
980 |
-
"learning_rate": 3.092277621590627e-06,
|
981 |
-
"loss": 0.9332,
|
982 |
-
"step": 139
|
983 |
-
},
|
984 |
-
{
|
985 |
-
"epoch": 1.9355400696864111,
|
986 |
-
"grad_norm": 2.869309902191162,
|
987 |
-
"learning_rate": 3.0178669746326567e-06,
|
988 |
-
"loss": 0.9496,
|
989 |
-
"step": 140
|
990 |
-
},
|
991 |
-
{
|
992 |
-
"epoch": 1.9494773519163764,
|
993 |
-
"grad_norm": 2.8482539653778076,
|
994 |
-
"learning_rate": 2.9440225537592728e-06,
|
995 |
-
"loss": 0.9406,
|
996 |
-
"step": 141
|
997 |
-
},
|
998 |
-
{
|
999 |
-
"epoch": 1.9634146341463414,
|
1000 |
-
"grad_norm": 3.2523257732391357,
|
1001 |
-
"learning_rate": 2.8707612044162595e-06,
|
1002 |
-
"loss": 0.9377,
|
1003 |
-
"step": 142
|
1004 |
-
},
|
1005 |
-
{
|
1006 |
-
"epoch": 1.9773519163763065,
|
1007 |
-
"grad_norm": 3.42722487449646,
|
1008 |
-
"learning_rate": 2.7980996390386755e-06,
|
1009 |
-
"loss": 0.9622,
|
1010 |
-
"step": 143
|
1011 |
-
},
|
1012 |
-
{
|
1013 |
-
"epoch": 1.9912891986062717,
|
1014 |
-
"grad_norm": 3.1534297466278076,
|
1015 |
-
"learning_rate": 2.7260544332383964e-06,
|
1016 |
-
"loss": 0.9219,
|
1017 |
-
"step": 144
|
1018 |
-
},
|
1019 |
-
{
|
1020 |
-
"epoch": 2.005226480836237,
|
1021 |
-
"grad_norm": 2.9020955562591553,
|
1022 |
-
"learning_rate": 2.654642022022887e-06,
|
1023 |
-
"loss": 0.9233,
|
1024 |
-
"step": 145
|
1025 |
-
},
|
1026 |
-
{
|
1027 |
-
"epoch": 2.0034843205574915,
|
1028 |
-
"grad_norm": 6.1046247482299805,
|
1029 |
-
"learning_rate": 2.5838786960460267e-06,
|
1030 |
-
"loss": 0.9226,
|
1031 |
-
"step": 146
|
1032 |
-
},
|
1033 |
-
{
|
1034 |
-
"epoch": 2.0174216027874565,
|
1035 |
-
"grad_norm": 2.791771411895752,
|
1036 |
-
"learning_rate": 2.513780597891867e-06,
|
1037 |
-
"loss": 0.8854,
|
1038 |
-
"step": 147
|
1039 |
-
},
|
1040 |
-
{
|
1041 |
-
"epoch": 2.0313588850174216,
|
1042 |
-
"grad_norm": 4.039772033691406,
|
1043 |
-
"learning_rate": 2.444363718392189e-06,
|
1044 |
-
"loss": 0.9067,
|
1045 |
-
"step": 148
|
1046 |
-
},
|
1047 |
-
{
|
1048 |
-
"epoch": 2.0452961672473866,
|
1049 |
-
"grad_norm": 3.059945821762085,
|
1050 |
-
"learning_rate": 2.3756438929786434e-06,
|
1051 |
-
"loss": 0.872,
|
1052 |
-
"step": 149
|
1053 |
-
},
|
1054 |
-
{
|
1055 |
-
"epoch": 2.059233449477352,
|
1056 |
-
"grad_norm": 3.355379819869995,
|
1057 |
-
"learning_rate": 2.3076367980703774e-06,
|
1058 |
-
"loss": 0.8897,
|
1059 |
-
"step": 150
|
1060 |
-
},
|
1061 |
-
{
|
1062 |
-
"epoch": 2.073170731707317,
|
1063 |
-
"grad_norm": 3.4992244243621826,
|
1064 |
-
"learning_rate": 2.240357947497908e-06,
|
1065 |
-
"loss": 0.9425,
|
1066 |
-
"step": 151
|
1067 |
-
},
|
1068 |
-
{
|
1069 |
-
"epoch": 2.087108013937282,
|
1070 |
-
"grad_norm": 3.829111099243164,
|
1071 |
-
"learning_rate": 2.173822688964094e-06,
|
1072 |
-
"loss": 0.9141,
|
1073 |
-
"step": 152
|
1074 |
-
},
|
1075 |
-
{
|
1076 |
-
"epoch": 2.1010452961672472,
|
1077 |
-
"grad_norm": 2.9267561435699463,
|
1078 |
-
"learning_rate": 2.108046200543013e-06,
|
1079 |
-
"loss": 0.9238,
|
1080 |
-
"step": 153
|
1081 |
-
},
|
1082 |
-
{
|
1083 |
-
"epoch": 2.1149825783972127,
|
1084 |
-
"grad_norm": 3.307797908782959,
|
1085 |
-
"learning_rate": 2.0430434872175245e-06,
|
1086 |
-
"loss": 0.9014,
|
1087 |
-
"step": 154
|
1088 |
-
},
|
1089 |
-
{
|
1090 |
-
"epoch": 2.1289198606271778,
|
1091 |
-
"grad_norm": 3.0536694526672363,
|
1092 |
-
"learning_rate": 1.9788293774563163e-06,
|
1093 |
-
"loss": 0.8967,
|
1094 |
-
"step": 155
|
1095 |
-
},
|
1096 |
-
{
|
1097 |
-
"epoch": 2.142857142857143,
|
1098 |
-
"grad_norm": 3.2300519943237305,
|
1099 |
-
"learning_rate": 1.9154185198312327e-06,
|
1100 |
-
"loss": 0.8817,
|
1101 |
-
"step": 156
|
1102 |
-
},
|
1103 |
-
{
|
1104 |
-
"epoch": 2.156794425087108,
|
1105 |
-
"grad_norm": 3.410097599029541,
|
1106 |
-
"learning_rate": 1.8528253796756277e-06,
|
1107 |
-
"loss": 0.9151,
|
1108 |
-
"step": 157
|
1109 |
-
},
|
1110 |
-
{
|
1111 |
-
"epoch": 2.1707317073170733,
|
1112 |
-
"grad_norm": 3.1198341846466064,
|
1113 |
-
"learning_rate": 1.7910642357845122e-06,
|
1114 |
-
"loss": 0.9359,
|
1115 |
-
"step": 158
|
1116 |
-
},
|
1117 |
-
{
|
1118 |
-
"epoch": 2.1846689895470384,
|
1119 |
-
"grad_norm": 3.0908079147338867,
|
1120 |
-
"learning_rate": 1.7301491771572808e-06,
|
1121 |
-
"loss": 0.8697,
|
1122 |
-
"step": 159
|
1123 |
-
},
|
1124 |
-
{
|
1125 |
-
"epoch": 2.1986062717770034,
|
1126 |
-
"grad_norm": 45.06842803955078,
|
1127 |
-
"learning_rate": 1.67009409978371e-06,
|
1128 |
-
"loss": 0.8753,
|
1129 |
-
"step": 160
|
1130 |
-
},
|
1131 |
-
{
|
1132 |
-
"epoch": 2.2125435540069684,
|
1133 |
-
"grad_norm": 3.0929031372070312,
|
1134 |
-
"learning_rate": 1.610912703473989e-06,
|
1135 |
-
"loss": 0.9032,
|
1136 |
-
"step": 161
|
1137 |
-
},
|
1138 |
-
{
|
1139 |
-
"epoch": 2.226480836236934,
|
1140 |
-
"grad_norm": 3.174290180206299,
|
1141 |
-
"learning_rate": 1.5526184887335188e-06,
|
1142 |
-
"loss": 0.9066,
|
1143 |
-
"step": 162
|
1144 |
-
},
|
1145 |
-
{
|
1146 |
-
"epoch": 2.240418118466899,
|
1147 |
-
"grad_norm": 3.2078001499176025,
|
1148 |
-
"learning_rate": 1.4952247536831645e-06,
|
1149 |
-
"loss": 0.8793,
|
1150 |
-
"step": 163
|
1151 |
-
},
|
1152 |
-
{
|
1153 |
-
"epoch": 2.254355400696864,
|
1154 |
-
"grad_norm": 3.8318631649017334,
|
1155 |
-
"learning_rate": 1.438744591025668e-06,
|
1156 |
-
"loss": 0.8881,
|
1157 |
-
"step": 164
|
1158 |
-
},
|
1159 |
-
{
|
1160 |
-
"epoch": 2.2682926829268295,
|
1161 |
-
"grad_norm": 3.460146188735962,
|
1162 |
-
"learning_rate": 1.3831908850589433e-06,
|
1163 |
-
"loss": 0.9087,
|
1164 |
-
"step": 165
|
1165 |
-
},
|
1166 |
-
{
|
1167 |
-
"epoch": 2.2822299651567945,
|
1168 |
-
"grad_norm": 3.1915969848632812,
|
1169 |
-
"learning_rate": 1.3285763087368974e-06,
|
1170 |
-
"loss": 0.9007,
|
1171 |
-
"step": 166
|
1172 |
-
},
|
1173 |
-
{
|
1174 |
-
"epoch": 2.2961672473867596,
|
1175 |
-
"grad_norm": 3.0315561294555664,
|
1176 |
-
"learning_rate": 1.2749133207784575e-06,
|
1177 |
-
"loss": 0.879,
|
1178 |
-
"step": 167
|
1179 |
-
},
|
1180 |
-
{
|
1181 |
-
"epoch": 2.3101045296167246,
|
1182 |
-
"grad_norm": 4.681447505950928,
|
1183 |
-
"learning_rate": 1.2222141628254902e-06,
|
1184 |
-
"loss": 0.9268,
|
1185 |
-
"step": 168
|
1186 |
-
},
|
1187 |
-
{
|
1188 |
-
"epoch": 2.3240418118466897,
|
1189 |
-
"grad_norm": 2.946652889251709,
|
1190 |
-
"learning_rate": 1.1704908566502246e-06,
|
1191 |
-
"loss": 0.8952,
|
1192 |
-
"step": 169
|
1193 |
-
},
|
1194 |
-
{
|
1195 |
-
"epoch": 2.337979094076655,
|
1196 |
-
"grad_norm": 2.944952964782715,
|
1197 |
-
"learning_rate": 1.1197552014128314e-06,
|
1198 |
-
"loss": 0.8807,
|
1199 |
-
"step": 170
|
1200 |
-
},
|
1201 |
-
{
|
1202 |
-
"epoch": 2.35191637630662,
|
1203 |
-
"grad_norm": 3.2808735370635986,
|
1204 |
-
"learning_rate": 1.0700187709697969e-06,
|
1205 |
-
"loss": 0.8722,
|
1206 |
-
"step": 171
|
1207 |
-
},
|
1208 |
-
{
|
1209 |
-
"epoch": 2.3658536585365852,
|
1210 |
-
"grad_norm": 3.887355327606201,
|
1211 |
-
"learning_rate": 1.0212929112336848e-06,
|
1212 |
-
"loss": 0.8863,
|
1213 |
-
"step": 172
|
1214 |
-
},
|
1215 |
-
{
|
1216 |
-
"epoch": 2.3797909407665507,
|
1217 |
-
"grad_norm": 3.0687787532806396,
|
1218 |
-
"learning_rate": 9.7358873758489e-07,
|
1219 |
-
"loss": 0.9093,
|
1220 |
-
"step": 173
|
1221 |
-
},
|
1222 |
-
{
|
1223 |
-
"epoch": 2.3937282229965158,
|
1224 |
-
"grad_norm": 2.890662431716919,
|
1225 |
-
"learning_rate": 9.269171323360006e-07,
|
1226 |
-
"loss": 0.8987,
|
1227 |
-
"step": 174
|
1228 |
-
},
|
1229 |
-
{
|
1230 |
-
"epoch": 2.407665505226481,
|
1231 |
-
"grad_norm": 3.5224292278289795,
|
1232 |
-
"learning_rate": 8.812887422493117e-07,
|
1233 |
-
"loss": 0.9008,
|
1234 |
-
"step": 175
|
1235 |
-
},
|
1236 |
-
{
|
1237 |
-
"epoch": 2.421602787456446,
|
1238 |
-
"grad_norm": 7.938294410705566,
|
1239 |
-
"learning_rate": 8.367139761080734e-07,
|
1240 |
-
"loss": 0.8774,
|
1241 |
-
"step": 176
|
1242 |
-
},
|
1243 |
-
{
|
1244 |
-
"epoch": 2.435540069686411,
|
1245 |
-
"grad_norm": 3.240544557571411,
|
1246 |
-
"learning_rate": 7.932030023420393e-07,
|
1247 |
-
"loss": 0.9178,
|
1248 |
-
"step": 177
|
1249 |
-
},
|
1250 |
-
{
|
1251 |
-
"epoch": 2.4494773519163764,
|
1252 |
-
"grad_norm": 3.17809796333313,
|
1253 |
-
"learning_rate": 7.507657467078292e-07,
|
1254 |
-
"loss": 0.8881,
|
1255 |
-
"step": 178
|
1256 |
-
},
|
1257 |
-
{
|
1258 |
-
"epoch": 2.4634146341463414,
|
1259 |
-
"grad_norm": 3.075671911239624,
|
1260 |
-
"learning_rate": 7.094118900246642e-07,
|
1261 |
-
"loss": 0.94,
|
1262 |
-
"step": 179
|
1263 |
-
},
|
1264 |
-
{
|
1265 |
-
"epoch": 2.4773519163763065,
|
1266 |
-
"grad_norm": 3.6002347469329834,
|
1267 |
-
"learning_rate": 6.691508659659682e-07,
|
1268 |
-
"loss": 0.8895,
|
1269 |
-
"step": 180
|
1270 |
-
}
|
1271 |
-
],
|
1272 |
-
"logging_steps": 1,
|
1273 |
-
"max_steps": 213,
|
1274 |
-
"num_input_tokens_seen": 0,
|
1275 |
-
"num_train_epochs": 3,
|
1276 |
-
"save_steps": 36,
|
1277 |
-
"stateful_callbacks": {
|
1278 |
-
"TrainerControl": {
|
1279 |
-
"args": {
|
1280 |
-
"should_epoch_stop": false,
|
1281 |
-
"should_evaluate": false,
|
1282 |
-
"should_log": false,
|
1283 |
-
"should_save": true,
|
1284 |
-
"should_training_stop": false
|
1285 |
-
},
|
1286 |
-
"attributes": {}
|
1287 |
-
}
|
1288 |
-
},
|
1289 |
-
"total_flos": 6.164757741828047e+18,
|
1290 |
-
"train_batch_size": 1,
|
1291 |
-
"trial_name": null,
|
1292 |
-
"trial_params": null
|
1293 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|