Update config.json
Browse files- config.json +966 -950
config.json
CHANGED
@@ -1,1316 +1,1287 @@
|
|
1 |
{
|
|
|
|
|
|
|
2 |
"architectures": [
|
3 |
"XLMWithLMHeadModel"
|
4 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
"dump_path": "/checkpoint/aconneau/dumped/xlm_17_100_big.3/16656234",
|
6 |
-
"
|
7 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
"exp_id": "16656234",
|
|
|
9 |
"fp16": true,
|
10 |
-
"amp": 2,
|
11 |
-
"encoder_only": true,
|
12 |
-
"emb_dim": 1280,
|
13 |
-
"n_layers": 16,
|
14 |
-
"n_heads": 16,
|
15 |
-
"dropout": 0.1,
|
16 |
-
"attention_dropout": 0.1,
|
17 |
"gelu_activation": true,
|
18 |
-
"
|
19 |
-
"sinusoidal_embeddings": false,
|
20 |
-
"use_lang_emb": false,
|
21 |
-
"use_memory": false,
|
22 |
-
"asm": false,
|
23 |
-
"context_size": 0,
|
24 |
-
"word_pred": 0.15,
|
25 |
-
"sample_alpha": 0.5,
|
26 |
-
"word_mask_keep_rand": "0.8,0.1,0.1",
|
27 |
-
"word_shuffle": 0.0,
|
28 |
-
"word_dropout": 0.0,
|
29 |
-
"word_blank": 0.0,
|
30 |
-
"data_path": "/private/home/aconneau/projects/XLM/data/wiki/100/175k",
|
31 |
-
"lgs": "en-es-fr-de-zh-ru-pt-it-ar-ja-id-tr-nl-pl-simple-fa-vi-sv-ko-he-ro-no-hi-uk-cs-fi-hu-th-da-ca-el-bg-sr-ms-bn-hr-sl-zh_yue-az-sk-eo-ta-sh-lt-et-ml-la-bs-sq-arz-af-ka-mr-eu-tl-ang-gl-nn-ur-kk-be-hy-te-lv-mk-zh_classical-als-is-wuu-my-sco-mn-ceb-ast-cy-kn-br-an-gu-bar-uz-lb-ne-si-war-jv-ga-zh_min_nan-oc-ku-sw-nds-ckb-ia-yi-fy-scn-gan-tt-am",
|
32 |
-
"max_vocab": 200000,
|
33 |
-
"min_count": 0,
|
34 |
-
"lg_sampling_factor": 0.7,
|
35 |
-
"bptt": 256,
|
36 |
-
"max_len": 200,
|
37 |
"group_by_size": true,
|
38 |
-
"
|
39 |
-
"
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
"lambda_ae": 1.0,
|
53 |
-
"
|
54 |
"lambda_bt": 1.0,
|
55 |
-
"
|
56 |
-
"
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
|
107 |
-
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
|
116 |
-
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
|
133 |
-
|
134 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
null
|
136 |
],
|
137 |
[
|
138 |
-
"
|
139 |
null
|
140 |
],
|
141 |
[
|
142 |
-
"
|
143 |
null
|
144 |
],
|
145 |
[
|
146 |
-
"
|
147 |
null
|
148 |
],
|
149 |
[
|
150 |
-
"
|
151 |
null
|
152 |
],
|
153 |
[
|
154 |
-
"
|
155 |
null
|
156 |
],
|
157 |
[
|
158 |
-
"
|
159 |
null
|
160 |
],
|
161 |
[
|
162 |
-
"
|
163 |
null
|
164 |
],
|
165 |
[
|
166 |
-
"
|
167 |
null
|
168 |
],
|
169 |
[
|
170 |
-
"
|
171 |
null
|
172 |
],
|
173 |
[
|
174 |
-
"
|
175 |
null
|
176 |
],
|
177 |
[
|
178 |
-
"
|
179 |
null
|
180 |
],
|
181 |
[
|
182 |
-
"
|
183 |
null
|
184 |
],
|
185 |
[
|
186 |
-
"
|
187 |
null
|
188 |
],
|
189 |
[
|
190 |
-
"
|
191 |
null
|
192 |
],
|
193 |
[
|
194 |
-
"
|
195 |
null
|
196 |
],
|
197 |
[
|
198 |
-
"
|
199 |
null
|
200 |
],
|
201 |
[
|
202 |
-
"
|
203 |
null
|
204 |
],
|
205 |
[
|
206 |
-
"
|
207 |
null
|
208 |
],
|
209 |
[
|
210 |
-
"
|
211 |
null
|
212 |
],
|
213 |
[
|
214 |
-
"
|
215 |
null
|
216 |
],
|
217 |
[
|
218 |
-
"
|
219 |
null
|
220 |
],
|
221 |
[
|
222 |
-
"
|
223 |
null
|
224 |
],
|
225 |
[
|
226 |
-
"
|
227 |
null
|
228 |
],
|
229 |
[
|
230 |
-
"
|
231 |
null
|
232 |
],
|
233 |
[
|
234 |
-
"
|
235 |
null
|
236 |
],
|
237 |
[
|
238 |
-
"
|
239 |
null
|
240 |
],
|
241 |
[
|
242 |
-
"
|
243 |
null
|
244 |
],
|
245 |
[
|
246 |
-
"
|
247 |
null
|
248 |
],
|
249 |
[
|
250 |
-
"
|
251 |
null
|
252 |
],
|
253 |
[
|
254 |
-
"
|
255 |
null
|
256 |
],
|
257 |
[
|
258 |
-
"
|
259 |
null
|
260 |
],
|
261 |
[
|
262 |
-
"
|
263 |
null
|
264 |
],
|
265 |
[
|
266 |
-
"
|
267 |
null
|
268 |
],
|
269 |
[
|
270 |
-
"
|
271 |
null
|
272 |
],
|
273 |
[
|
274 |
-
"
|
275 |
null
|
276 |
],
|
277 |
[
|
278 |
-
"
|
279 |
null
|
280 |
],
|
281 |
[
|
282 |
-
"
|
283 |
null
|
284 |
],
|
285 |
[
|
286 |
-
"
|
287 |
null
|
288 |
],
|
289 |
[
|
290 |
-
"
|
291 |
null
|
292 |
],
|
293 |
[
|
294 |
-
"
|
295 |
null
|
296 |
],
|
297 |
[
|
298 |
-
"
|
299 |
null
|
300 |
],
|
301 |
[
|
302 |
-
"
|
303 |
null
|
304 |
],
|
305 |
[
|
306 |
-
"
|
307 |
null
|
308 |
],
|
309 |
[
|
310 |
-
"
|
311 |
null
|
312 |
],
|
313 |
[
|
314 |
-
"
|
315 |
null
|
316 |
],
|
317 |
[
|
318 |
-
"
|
319 |
null
|
320 |
],
|
321 |
[
|
322 |
-
"
|
323 |
null
|
324 |
],
|
325 |
[
|
326 |
-
"
|
327 |
null
|
328 |
],
|
329 |
[
|
330 |
-
"
|
331 |
null
|
332 |
],
|
333 |
[
|
334 |
-
"
|
335 |
null
|
336 |
],
|
337 |
[
|
338 |
-
"
|
339 |
null
|
340 |
],
|
341 |
[
|
342 |
-
"
|
343 |
null
|
344 |
],
|
345 |
[
|
346 |
-
"
|
347 |
null
|
348 |
],
|
349 |
[
|
350 |
-
"
|
351 |
null
|
352 |
],
|
353 |
[
|
354 |
-
"
|
355 |
null
|
356 |
],
|
357 |
[
|
358 |
-
"
|
359 |
null
|
360 |
],
|
361 |
[
|
362 |
-
"
|
363 |
null
|
364 |
],
|
365 |
[
|
366 |
-
"
|
367 |
null
|
368 |
],
|
369 |
[
|
370 |
-
"
|
371 |
null
|
372 |
],
|
373 |
[
|
374 |
-
"
|
375 |
null
|
376 |
],
|
377 |
[
|
378 |
-
"
|
379 |
null
|
380 |
],
|
381 |
[
|
382 |
-
"
|
383 |
null
|
384 |
],
|
385 |
[
|
386 |
-
"
|
387 |
null
|
388 |
],
|
389 |
[
|
390 |
-
"
|
391 |
null
|
392 |
],
|
393 |
[
|
394 |
-
"
|
395 |
null
|
396 |
],
|
397 |
[
|
398 |
-
"
|
399 |
null
|
400 |
],
|
401 |
[
|
402 |
-
"
|
403 |
null
|
404 |
],
|
405 |
[
|
406 |
-
"
|
407 |
null
|
408 |
],
|
409 |
[
|
410 |
-
"
|
411 |
null
|
412 |
],
|
413 |
[
|
414 |
-
"
|
415 |
null
|
416 |
],
|
417 |
[
|
418 |
-
"
|
419 |
null
|
420 |
],
|
421 |
[
|
422 |
-
"
|
423 |
null
|
424 |
],
|
425 |
[
|
426 |
-
"
|
427 |
null
|
428 |
],
|
429 |
[
|
430 |
-
"
|
431 |
null
|
432 |
],
|
433 |
[
|
434 |
-
"
|
435 |
null
|
436 |
],
|
437 |
[
|
438 |
-
"
|
439 |
null
|
440 |
],
|
441 |
[
|
442 |
-
"
|
443 |
null
|
444 |
],
|
445 |
[
|
446 |
-
"
|
447 |
null
|
448 |
],
|
449 |
[
|
450 |
-
"
|
451 |
null
|
452 |
],
|
453 |
[
|
454 |
-
"
|
455 |
null
|
456 |
-
]
|
457 |
-
|
458 |
-
|
459 |
-
|
460 |
-
|
461 |
-
|
462 |
-
|
463 |
-
|
464 |
-
|
465 |
-
|
466 |
-
|
467 |
-
|
468 |
-
|
469 |
-
|
470 |
-
|
471 |
-
|
472 |
-
|
473 |
-
|
474 |
-
|
475 |
-
|
476 |
-
|
477 |
-
|
478 |
-
|
479 |
-
|
480 |
-
|
481 |
-
|
482 |
-
|
483 |
-
|
484 |
-
|
485 |
-
|
486 |
-
|
487 |
-
|
488 |
-
|
489 |
-
|
490 |
-
|
491 |
-
|
492 |
-
|
493 |
-
|
494 |
-
|
495 |
-
|
496 |
-
|
497 |
-
|
498 |
-
|
499 |
-
|
500 |
-
|
501 |
-
|
502 |
-
|
503 |
-
|
504 |
-
|
505 |
-
|
506 |
-
|
507 |
-
|
508 |
-
|
509 |
-
|
510 |
-
|
511 |
-
|
512 |
-
|
513 |
-
|
514 |
-
|
515 |
-
|
516 |
-
|
517 |
-
|
518 |
-
|
519 |
-
|
520 |
-
|
521 |
-
|
522 |
-
|
523 |
-
|
524 |
-
|
525 |
-
|
526 |
-
|
527 |
-
|
528 |
-
|
529 |
-
|
530 |
-
|
531 |
-
|
532 |
-
|
533 |
-
|
534 |
-
|
535 |
-
"kk",
|
536 |
-
"be",
|
537 |
-
"hy",
|
538 |
-
"te",
|
539 |
-
"lv",
|
540 |
-
"mk",
|
541 |
-
"zh_classical",
|
542 |
-
"als",
|
543 |
-
"is",
|
544 |
-
"wuu",
|
545 |
-
"my",
|
546 |
-
"sco",
|
547 |
-
"mn",
|
548 |
-
"ceb",
|
549 |
-
"ast",
|
550 |
-
"cy",
|
551 |
-
"kn",
|
552 |
-
"br",
|
553 |
-
"an",
|
554 |
-
"gu",
|
555 |
-
"bar",
|
556 |
-
"uz",
|
557 |
-
"lb",
|
558 |
-
"ne",
|
559 |
-
"si",
|
560 |
-
"war",
|
561 |
-
"jv",
|
562 |
-
"ga",
|
563 |
-
"zh_min_nan",
|
564 |
-
"oc",
|
565 |
-
"ku",
|
566 |
-
"sw",
|
567 |
-
"nds",
|
568 |
-
"ckb",
|
569 |
-
"ia",
|
570 |
-
"yi",
|
571 |
-
"fy",
|
572 |
-
"scn",
|
573 |
-
"gan",
|
574 |
-
"tt",
|
575 |
-
"am"
|
576 |
-
],
|
577 |
-
"id2lang": {
|
578 |
-
"0": "af",
|
579 |
-
"1": "als",
|
580 |
-
"2": "am",
|
581 |
-
"3": "an",
|
582 |
-
"4": "ang",
|
583 |
-
"5": "ar",
|
584 |
-
"6": "arz",
|
585 |
-
"7": "ast",
|
586 |
-
"8": "az",
|
587 |
-
"9": "bar",
|
588 |
-
"10": "be",
|
589 |
-
"11": "bg",
|
590 |
-
"12": "bn",
|
591 |
-
"13": "br",
|
592 |
-
"14": "bs",
|
593 |
-
"15": "ca",
|
594 |
-
"16": "ceb",
|
595 |
-
"17": "ckb",
|
596 |
-
"18": "cs",
|
597 |
-
"19": "cy",
|
598 |
-
"20": "da",
|
599 |
-
"21": "de",
|
600 |
-
"22": "el",
|
601 |
-
"23": "en",
|
602 |
-
"24": "eo",
|
603 |
-
"25": "es",
|
604 |
-
"26": "et",
|
605 |
-
"27": "eu",
|
606 |
-
"28": "fa",
|
607 |
-
"29": "fi",
|
608 |
-
"30": "fr",
|
609 |
-
"31": "fy",
|
610 |
-
"32": "ga",
|
611 |
-
"33": "gan",
|
612 |
-
"34": "gl",
|
613 |
-
"35": "gu",
|
614 |
-
"36": "he",
|
615 |
-
"37": "hi",
|
616 |
-
"38": "hr",
|
617 |
-
"39": "hu",
|
618 |
-
"40": "hy",
|
619 |
-
"41": "ia",
|
620 |
-
"42": "id",
|
621 |
-
"43": "is",
|
622 |
-
"44": "it",
|
623 |
-
"45": "ja",
|
624 |
-
"46": "jv",
|
625 |
-
"47": "ka",
|
626 |
-
"48": "kk",
|
627 |
-
"49": "kn",
|
628 |
-
"50": "ko",
|
629 |
-
"51": "ku",
|
630 |
-
"52": "la",
|
631 |
-
"53": "lb",
|
632 |
-
"54": "lt",
|
633 |
-
"55": "lv",
|
634 |
-
"56": "mk",
|
635 |
-
"57": "ml",
|
636 |
-
"58": "mn",
|
637 |
-
"59": "mr",
|
638 |
-
"60": "ms",
|
639 |
-
"61": "my",
|
640 |
-
"62": "nds",
|
641 |
-
"63": "ne",
|
642 |
-
"64": "nl",
|
643 |
-
"65": "nn",
|
644 |
-
"66": "no",
|
645 |
-
"67": "oc",
|
646 |
-
"68": "pl",
|
647 |
-
"69": "pt",
|
648 |
-
"70": "ro",
|
649 |
-
"71": "ru",
|
650 |
-
"72": "scn",
|
651 |
-
"73": "sco",
|
652 |
-
"74": "sh",
|
653 |
-
"75": "si",
|
654 |
-
"76": "simple",
|
655 |
-
"77": "sk",
|
656 |
-
"78": "sl",
|
657 |
-
"79": "sq",
|
658 |
-
"80": "sr",
|
659 |
-
"81": "sv",
|
660 |
-
"82": "sw",
|
661 |
-
"83": "ta",
|
662 |
-
"84": "te",
|
663 |
-
"85": "th",
|
664 |
-
"86": "tl",
|
665 |
-
"87": "tr",
|
666 |
-
"88": "tt",
|
667 |
-
"89": "uk",
|
668 |
-
"90": "ur",
|
669 |
-
"91": "uz",
|
670 |
-
"92": "vi",
|
671 |
-
"93": "war",
|
672 |
-
"94": "wuu",
|
673 |
-
"95": "yi",
|
674 |
-
"96": "zh",
|
675 |
-
"97": "zh_classical",
|
676 |
-
"98": "zh_min_nan",
|
677 |
-
"99": "zh_yue"
|
678 |
-
},
|
679 |
-
"lang2id": {
|
680 |
-
"af": 0,
|
681 |
-
"als": 1,
|
682 |
-
"am": 2,
|
683 |
-
"an": 3,
|
684 |
-
"ang": 4,
|
685 |
-
"ar": 5,
|
686 |
-
"arz": 6,
|
687 |
-
"ast": 7,
|
688 |
-
"az": 8,
|
689 |
-
"bar": 9,
|
690 |
-
"be": 10,
|
691 |
-
"bg": 11,
|
692 |
-
"bn": 12,
|
693 |
-
"br": 13,
|
694 |
-
"bs": 14,
|
695 |
-
"ca": 15,
|
696 |
-
"ceb": 16,
|
697 |
-
"ckb": 17,
|
698 |
-
"cs": 18,
|
699 |
-
"cy": 19,
|
700 |
-
"da": 20,
|
701 |
-
"de": 21,
|
702 |
-
"el": 22,
|
703 |
-
"en": 23,
|
704 |
-
"eo": 24,
|
705 |
-
"es": 25,
|
706 |
-
"et": 26,
|
707 |
-
"eu": 27,
|
708 |
-
"fa": 28,
|
709 |
-
"fi": 29,
|
710 |
-
"fr": 30,
|
711 |
-
"fy": 31,
|
712 |
-
"ga": 32,
|
713 |
-
"gan": 33,
|
714 |
-
"gl": 34,
|
715 |
-
"gu": 35,
|
716 |
-
"he": 36,
|
717 |
-
"hi": 37,
|
718 |
-
"hr": 38,
|
719 |
-
"hu": 39,
|
720 |
-
"hy": 40,
|
721 |
-
"ia": 41,
|
722 |
-
"id": 42,
|
723 |
-
"is": 43,
|
724 |
-
"it": 44,
|
725 |
-
"ja": 45,
|
726 |
-
"jv": 46,
|
727 |
-
"ka": 47,
|
728 |
-
"kk": 48,
|
729 |
-
"kn": 49,
|
730 |
-
"ko": 50,
|
731 |
-
"ku": 51,
|
732 |
-
"la": 52,
|
733 |
-
"lb": 53,
|
734 |
-
"lt": 54,
|
735 |
-
"lv": 55,
|
736 |
-
"mk": 56,
|
737 |
-
"ml": 57,
|
738 |
-
"mn": 58,
|
739 |
-
"mr": 59,
|
740 |
-
"ms": 60,
|
741 |
-
"my": 61,
|
742 |
-
"nds": 62,
|
743 |
-
"ne": 63,
|
744 |
-
"nl": 64,
|
745 |
-
"nn": 65,
|
746 |
-
"no": 66,
|
747 |
-
"oc": 67,
|
748 |
-
"pl": 68,
|
749 |
-
"pt": 69,
|
750 |
-
"ro": 70,
|
751 |
-
"ru": 71,
|
752 |
-
"scn": 72,
|
753 |
-
"sco": 73,
|
754 |
-
"sh": 74,
|
755 |
-
"si": 75,
|
756 |
-
"simple": 76,
|
757 |
-
"sk": 77,
|
758 |
-
"sl": 78,
|
759 |
-
"sq": 79,
|
760 |
-
"sr": 80,
|
761 |
-
"sv": 81,
|
762 |
-
"sw": 82,
|
763 |
-
"ta": 83,
|
764 |
-
"te": 84,
|
765 |
-
"th": 85,
|
766 |
-
"tl": 86,
|
767 |
-
"tr": 87,
|
768 |
-
"tt": 88,
|
769 |
-
"uk": 89,
|
770 |
-
"ur": 90,
|
771 |
-
"uz": 91,
|
772 |
-
"vi": 92,
|
773 |
-
"war": 93,
|
774 |
-
"wuu": 94,
|
775 |
-
"yi": 95,
|
776 |
-
"zh": 96,
|
777 |
-
"zh_classical": 97,
|
778 |
-
"zh_min_nan": 98,
|
779 |
-
"zh_yue": 99
|
780 |
-
},
|
781 |
-
"n_langs": 100,
|
782 |
-
"bt_src_langs": [],
|
783 |
"mono_dataset": {
|
784 |
-
"
|
785 |
-
"
|
786 |
-
"
|
787 |
-
"
|
788 |
-
},
|
789 |
-
"es": {
|
790 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.es.pth",
|
791 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.es.pth",
|
792 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.es.pth"
|
793 |
-
},
|
794 |
-
"fr": {
|
795 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fr.pth",
|
796 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fr.pth",
|
797 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fr.pth"
|
798 |
-
},
|
799 |
-
"de": {
|
800 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.de.pth",
|
801 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.de.pth",
|
802 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.de.pth"
|
803 |
},
|
804 |
-
"
|
805 |
-
"
|
806 |
-
"
|
807 |
-
"
|
808 |
},
|
809 |
-
"
|
810 |
-
"
|
811 |
-
"
|
812 |
-
"
|
813 |
},
|
814 |
-
"
|
815 |
-
"
|
816 |
-
"
|
817 |
-
"
|
818 |
},
|
819 |
-
"
|
820 |
-
"
|
821 |
-
"
|
822 |
-
"
|
823 |
},
|
824 |
"ar": {
|
|
|
825 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ar.pth",
|
826 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ar.pth"
|
827 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ar.pth"
|
828 |
-
},
|
829 |
-
"ja": {
|
830 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ja.pth",
|
831 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ja.pth",
|
832 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ja.pth"
|
833 |
-
},
|
834 |
-
"id": {
|
835 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.id.pth",
|
836 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.id.pth",
|
837 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.id.pth"
|
838 |
},
|
839 |
-
"
|
840 |
-
"
|
841 |
-
"
|
842 |
-
"
|
843 |
-
},
|
844 |
-
"nl": {
|
845 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nl.pth",
|
846 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nl.pth",
|
847 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nl.pth"
|
848 |
},
|
849 |
-
"
|
850 |
-
"
|
851 |
-
"
|
852 |
-
"
|
853 |
},
|
854 |
-
"
|
855 |
-
"
|
856 |
-
"
|
857 |
-
"
|
858 |
},
|
859 |
-
"
|
860 |
-
"
|
861 |
-
"
|
862 |
-
"
|
863 |
},
|
864 |
-
"
|
865 |
-
"
|
866 |
-
"
|
867 |
-
"
|
868 |
},
|
869 |
-
"
|
870 |
-
"
|
871 |
-
"
|
872 |
-
"
|
873 |
},
|
874 |
-
"
|
875 |
-
"
|
876 |
-
"
|
877 |
-
"
|
878 |
},
|
879 |
-
"
|
880 |
-
"
|
881 |
-
"
|
882 |
-
"
|
883 |
},
|
884 |
-
"
|
885 |
-
"
|
886 |
-
"
|
887 |
-
"
|
888 |
},
|
889 |
-
"
|
890 |
-
"
|
891 |
-
"
|
892 |
-
"
|
893 |
},
|
894 |
-
"
|
895 |
-
"
|
896 |
-
"
|
897 |
-
"
|
898 |
},
|
899 |
-
"
|
900 |
-
"
|
901 |
-
"
|
902 |
-
"
|
903 |
},
|
904 |
"cs": {
|
|
|
905 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.cs.pth",
|
906 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.cs.pth"
|
907 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.cs.pth"
|
908 |
-
},
|
909 |
-
"fi": {
|
910 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fi.pth",
|
911 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fi.pth",
|
912 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fi.pth"
|
913 |
-
},
|
914 |
-
"hu": {
|
915 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hu.pth",
|
916 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hu.pth",
|
917 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hu.pth"
|
918 |
},
|
919 |
-
"
|
920 |
-
"
|
921 |
-
"
|
922 |
-
"
|
923 |
},
|
924 |
"da": {
|
|
|
925 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.da.pth",
|
926 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.da.pth"
|
927 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.da.pth"
|
928 |
},
|
929 |
-
"
|
930 |
-
"
|
931 |
-
"
|
932 |
-
"
|
933 |
},
|
934 |
"el": {
|
|
|
935 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.el.pth",
|
936 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.el.pth"
|
937 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.el.pth"
|
938 |
-
},
|
939 |
-
"bg": {
|
940 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bg.pth",
|
941 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bg.pth",
|
942 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bg.pth"
|
943 |
-
},
|
944 |
-
"sr": {
|
945 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sr.pth",
|
946 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sr.pth",
|
947 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sr.pth"
|
948 |
-
},
|
949 |
-
"ms": {
|
950 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ms.pth",
|
951 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ms.pth",
|
952 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ms.pth"
|
953 |
},
|
954 |
-
"
|
955 |
-
"
|
956 |
-
"
|
957 |
-
"
|
958 |
},
|
959 |
-
"
|
960 |
-
"
|
961 |
-
"
|
962 |
-
"
|
963 |
},
|
964 |
-
"
|
965 |
-
"
|
966 |
-
"
|
967 |
-
"
|
968 |
},
|
969 |
-
"
|
970 |
-
"
|
971 |
-
"
|
972 |
-
"
|
973 |
},
|
974 |
-
"
|
975 |
-
"
|
976 |
-
"
|
977 |
-
"
|
978 |
},
|
979 |
-
"
|
980 |
-
"
|
981 |
-
"
|
982 |
-
"
|
983 |
},
|
984 |
-
"
|
985 |
-
"
|
986 |
-
"
|
987 |
-
"
|
988 |
},
|
989 |
-
"
|
990 |
-
"
|
991 |
-
"
|
992 |
-
"
|
993 |
},
|
994 |
-
"
|
995 |
-
"
|
996 |
-
"
|
997 |
-
"
|
998 |
},
|
999 |
-
"
|
1000 |
-
"
|
1001 |
-
"
|
1002 |
-
"
|
1003 |
},
|
1004 |
-
"
|
1005 |
-
"
|
1006 |
-
"
|
1007 |
-
"
|
1008 |
},
|
1009 |
-
"
|
1010 |
-
"
|
1011 |
-
"
|
1012 |
-
"
|
1013 |
},
|
1014 |
-
"
|
1015 |
-
"
|
1016 |
-
"
|
1017 |
-
"
|
1018 |
},
|
1019 |
-
"
|
1020 |
-
"
|
1021 |
-
"
|
1022 |
-
"
|
1023 |
},
|
1024 |
-
"
|
1025 |
-
"
|
1026 |
-
"
|
1027 |
-
"
|
1028 |
},
|
1029 |
-
"
|
1030 |
-
"
|
1031 |
-
"
|
1032 |
-
"
|
1033 |
},
|
1034 |
-
"
|
1035 |
-
"
|
1036 |
-
"
|
1037 |
-
"
|
1038 |
},
|
1039 |
-
"
|
1040 |
-
"
|
1041 |
-
"
|
1042 |
-
"
|
1043 |
},
|
1044 |
-
"
|
1045 |
-
"
|
1046 |
-
"
|
1047 |
-
"
|
1048 |
},
|
1049 |
-
"
|
1050 |
-
"
|
1051 |
-
"
|
1052 |
-
"
|
1053 |
},
|
1054 |
-
"
|
1055 |
-
"
|
1056 |
-
"
|
1057 |
-
"
|
1058 |
},
|
1059 |
-
"
|
1060 |
-
"
|
1061 |
-
"
|
1062 |
-
"
|
1063 |
},
|
1064 |
-
"
|
1065 |
-
"
|
1066 |
-
"
|
1067 |
-
"
|
1068 |
},
|
1069 |
-
"
|
1070 |
-
"
|
1071 |
-
"
|
1072 |
-
"
|
1073 |
},
|
1074 |
-
"
|
1075 |
-
"
|
1076 |
-
"
|
1077 |
-
"
|
1078 |
},
|
1079 |
"kk": {
|
|
|
1080 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.kk.pth",
|
1081 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.kk.pth"
|
1082 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.kk.pth"
|
1083 |
},
|
1084 |
-
"
|
1085 |
-
"
|
1086 |
-
"
|
1087 |
-
"
|
1088 |
},
|
1089 |
-
"
|
1090 |
-
"
|
1091 |
-
"
|
1092 |
-
"
|
1093 |
},
|
1094 |
-
"
|
1095 |
-
"
|
1096 |
-
"
|
1097 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1098 |
},
|
1099 |
"lv": {
|
|
|
1100 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lv.pth",
|
1101 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lv.pth"
|
1102 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lv.pth"
|
1103 |
},
|
1104 |
"mk": {
|
|
|
1105 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mk.pth",
|
1106 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mk.pth"
|
1107 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mk.pth"
|
1108 |
},
|
1109 |
-
"
|
1110 |
-
"
|
1111 |
-
"
|
1112 |
-
"
|
1113 |
},
|
1114 |
-
"
|
1115 |
-
"
|
1116 |
-
"
|
1117 |
-
"
|
1118 |
},
|
1119 |
-
"
|
1120 |
-
"
|
1121 |
-
"
|
1122 |
-
"
|
1123 |
},
|
1124 |
-
"
|
1125 |
-
"
|
1126 |
-
"
|
1127 |
-
"
|
1128 |
},
|
1129 |
"my": {
|
|
|
1130 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.my.pth",
|
1131 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.my.pth"
|
1132 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.my.pth"
|
1133 |
},
|
1134 |
-
"
|
1135 |
-
"
|
1136 |
-
"
|
1137 |
-
"
|
1138 |
},
|
1139 |
-
"
|
1140 |
-
"
|
1141 |
-
"
|
1142 |
-
"
|
1143 |
},
|
1144 |
-
"
|
1145 |
-
"
|
1146 |
-
"
|
1147 |
-
"
|
1148 |
},
|
1149 |
-
"
|
1150 |
-
"
|
1151 |
-
"
|
1152 |
-
"
|
1153 |
},
|
1154 |
-
"
|
1155 |
-
"
|
1156 |
-
"
|
1157 |
-
"
|
1158 |
},
|
1159 |
-
"
|
1160 |
-
"
|
1161 |
-
"
|
1162 |
-
"
|
1163 |
},
|
1164 |
-
"
|
1165 |
-
"
|
1166 |
-
"
|
1167 |
-
"
|
1168 |
},
|
1169 |
-
"
|
1170 |
-
"
|
1171 |
-
"
|
1172 |
-
"
|
1173 |
},
|
1174 |
-
"
|
1175 |
-
"
|
1176 |
-
"
|
1177 |
-
"
|
1178 |
},
|
1179 |
-
"
|
1180 |
-
"
|
1181 |
-
"
|
1182 |
-
"
|
1183 |
},
|
1184 |
-
"
|
1185 |
-
"
|
1186 |
-
"
|
1187 |
-
"
|
1188 |
},
|
1189 |
-
"
|
1190 |
-
"
|
1191 |
-
"
|
1192 |
-
"
|
1193 |
},
|
1194 |
-
"
|
1195 |
-
"
|
1196 |
-
"
|
1197 |
-
"
|
1198 |
},
|
1199 |
"si": {
|
|
|
1200 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.si.pth",
|
1201 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.si.pth"
|
1202 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.si.pth"
|
1203 |
},
|
1204 |
-
"
|
1205 |
-
"
|
1206 |
-
"
|
1207 |
-
"
|
1208 |
},
|
1209 |
-
"
|
1210 |
-
"
|
1211 |
-
"
|
1212 |
-
"
|
1213 |
},
|
1214 |
-
"
|
1215 |
-
"
|
1216 |
-
"
|
1217 |
-
"
|
1218 |
},
|
1219 |
-
"
|
1220 |
-
"
|
1221 |
-
"
|
1222 |
-
"
|
1223 |
},
|
1224 |
-
"
|
1225 |
-
"
|
1226 |
-
"
|
1227 |
-
"
|
1228 |
},
|
1229 |
-
"
|
1230 |
-
"
|
1231 |
-
"
|
1232 |
-
"
|
1233 |
},
|
1234 |
"sw": {
|
|
|
1235 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sw.pth",
|
1236 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sw.pth"
|
1237 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sw.pth"
|
1238 |
},
|
1239 |
-
"
|
1240 |
-
"
|
1241 |
-
"
|
1242 |
-
"
|
1243 |
},
|
1244 |
-
"
|
1245 |
-
"
|
1246 |
-
"
|
1247 |
-
"
|
1248 |
},
|
1249 |
-
"
|
1250 |
-
"
|
1251 |
-
"
|
1252 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1253 |
},
|
1254 |
"yi": {
|
|
|
1255 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.yi.pth",
|
1256 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.yi.pth"
|
1257 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.yi.pth"
|
1258 |
-
},
|
1259 |
-
"fy": {
|
1260 |
-
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fy.pth",
|
1261 |
-
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fy.pth",
|
1262 |
-
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fy.pth"
|
1263 |
},
|
1264 |
-
"
|
1265 |
-
"
|
1266 |
-
"
|
1267 |
-
"
|
1268 |
},
|
1269 |
-
"
|
1270 |
-
"
|
1271 |
-
"
|
1272 |
-
"
|
1273 |
},
|
1274 |
-
"
|
1275 |
-
"
|
1276 |
-
"
|
1277 |
-
"
|
1278 |
},
|
1279 |
-
"
|
1280 |
-
"
|
1281 |
-
"
|
1282 |
-
"
|
1283 |
}
|
1284 |
},
|
1285 |
-
"para_dataset": {},
|
1286 |
-
"word_mask": 0.8,
|
1287 |
-
"word_keep": 0.1,
|
1288 |
-
"word_rand": 0.1,
|
1289 |
-
"is_slurm_job": true,
|
1290 |
-
"n_nodes": 4,
|
1291 |
-
"node_id": 0,
|
1292 |
-
"global_rank": 0,
|
1293 |
-
"world_size": 32,
|
1294 |
-
"n_gpu_per_node": 8,
|
1295 |
-
"master_addr": "learnfair0332",
|
1296 |
-
"is_master": true,
|
1297 |
-
"multi_node": true,
|
1298 |
-
"multi_gpu": true,
|
1299 |
-
"command": "python /private/home/aconneau/workdir/xlm_17_100_big.3/2019_08_10_19_23_42/train.py --n_heads 16 --bt_steps '' --max_vocab 200000 --word_mask_keep_rand '0.8,0.1,0.1' --use_lang_emb false --data_path '/private/home/aconneau/projects/XLM/data/wiki/100/175k' --save_periodic 0 --max_len 200 --bptt 256 --ae_steps '' --fp16 true --share_inout_emb true --sinusoidal_embeddings false --word_shuffle 0 --tokens_per_batch '-1' --accumulate_gradients 4 --validation_metrics '_valid_en_mlm_ppl,_valid_mlm_ppl,_valid_zh_mlm_ppl' --attention_dropout '0.1' --split_data true --max_epoch 100000 --stopping_criterion '_valid_zh_mlm_ppl,25' --dump_path '/checkpoint/aconneau/dumped' --epoch_size 200000 --word_blank 0 --gelu_activation true --n_layers 16 --optimizer 'adam_inverse_sqrt,lr=0.00005,warmup_updates=30000,beta1=0.9,beta2=0.999,weight_decay=0.01,eps=0.000001' --mlm_steps 'en,es,fr,de,zh,ru,pt,it,ar,ja,id,tr,nl,pl,simple,fa,vi,sv,ko,he,ro,no,hi,uk,cs,fi,hu,th,da,ca,el,bg,sr,ms,bn,hr,sl,zh_yue,az,sk,eo,ta,sh,lt,et,ml,la,bs,sq,arz,af,ka,mr,eu,tl,ang,gl,nn,ur,kk,be,hy,te,lv,mk,zh_classical,als,is,wuu,my,sco,mn,ceb,ast,cy,kn,br,an,gu,bar,uz,lb,ne,si,war,jv,ga,zh_min_nan,oc,ku,sw,nds,ckb,ia,yi,fy,scn,gan,tt,am' --eval_bleu false --dropout '0.1' --mt_steps '' --batch_size 16 --word_dropout 0 --reload_model '/checkpoint/aconneau/dumped/xlm_17_100_240_big_model_upper.2/14884511/best-valid_zh_mlm_ppl.pth' --min_count 0 --amp 2 --group_by_size true --asm false --sample_alpha '0.5' --word_pred '0.15' --clip_grad_norm 1 --emb_dim 1280 --encoder_only true --lgs 'en-es-fr-de-zh-ru-pt-it-ar-ja-id-tr-nl-pl-simple-fa-vi-sv-ko-he-ro-no-hi-uk-cs-fi-hu-th-da-ca-el-bg-sr-ms-bn-hr-sl-zh_yue-az-sk-eo-ta-sh-lt-et-ml-la-bs-sq-arz-af-ka-mr-eu-tl-ang-gl-nn-ur-kk-be-hy-te-lv-mk-zh_classical-als-is-wuu-my-sco-mn-ceb-ast-cy-kn-br-an-gu-bar-uz-lb-ne-si-war-jv-ga-zh_min_nan-oc-ku-sw-nds-ckb-ia-yi-fy-scn-gan-tt-am' --clm_steps '' --exp_name 'xlm_17_100_big.3' --lg_sampling_factor '0.7' --eval_only false --exp_id 16656234 --master_port 11363 --exp_id \"16656234\"",
|
1300 |
-
"n_words": 200000,
|
1301 |
-
"bos_index": 0,
|
1302 |
-
"eos_index": 1,
|
1303 |
-
"pad_index": 2,
|
1304 |
-
"unk_index": 3,
|
1305 |
-
"mask_index": 5,
|
1306 |
-
"lambda_clm_config": null,
|
1307 |
-
"lambda_mlm_config": null,
|
1308 |
-
"lambda_pc_config": null,
|
1309 |
-
"lambda_ae_config": null,
|
1310 |
-
"lambda_mt_config": null,
|
1311 |
-
"lambda_bt_config": null,
|
1312 |
-
"hyp_path": "/checkpoint/aconneau/dumped/xlm_17_100_big.3/16656234/hypotheses",
|
1313 |
-
"ref_paths": {},
|
1314 |
"mono_list": [
|
1315 |
"en",
|
1316 |
"es",
|
@@ -1413,5 +1384,50 @@
|
|
1413 |
"tt",
|
1414 |
"am"
|
1415 |
],
|
1416 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1417 |
}
|
|
|
1 |
{
|
2 |
+
"accumulate_gradients": 4,
|
3 |
+
"ae_steps": [],
|
4 |
+
"amp": 2,
|
5 |
"architectures": [
|
6 |
"XLMWithLMHeadModel"
|
7 |
],
|
8 |
+
"asm": false,
|
9 |
+
"attention_dropout": 0.1,
|
10 |
+
"batch_size": 16,
|
11 |
+
"beam_size": 1,
|
12 |
+
"bos_index": 0,
|
13 |
+
"bos_token_id": 0,
|
14 |
+
"bptt": 256,
|
15 |
+
"bt_src_langs": [],
|
16 |
+
"bt_steps": [],
|
17 |
+
"causal": false,
|
18 |
+
"clip_grad_norm": 1.0,
|
19 |
+
"clm_steps": [],
|
20 |
+
"command": "python /private/home/aconneau/workdir/xlm_17_100_big.3/2019_08_10_19_23_42/train.py --n_heads 16 --bt_steps '' --max_vocab 200000 --word_mask_keep_rand '0.8,0.1,0.1' --use_lang_emb false --data_path '/private/home/aconneau/projects/XLM/data/wiki/100/175k' --save_periodic 0 --max_len 200 --bptt 256 --ae_steps '' --fp16 true --share_inout_emb true --sinusoidal_embeddings false --word_shuffle 0 --tokens_per_batch '-1' --accumulate_gradients 4 --validation_metrics '_valid_en_mlm_ppl,_valid_mlm_ppl,_valid_zh_mlm_ppl' --attention_dropout '0.1' --split_data true --max_epoch 100000 --stopping_criterion '_valid_zh_mlm_ppl,25' --dump_path '/checkpoint/aconneau/dumped' --epoch_size 200000 --word_blank 0 --gelu_activation true --n_layers 16 --optimizer 'adam_inverse_sqrt,lr=0.00005,warmup_updates=30000,beta1=0.9,beta2=0.999,weight_decay=0.01,eps=0.000001' --mlm_steps 'en,es,fr,de,zh,ru,pt,it,ar,ja,id,tr,nl,pl,simple,fa,vi,sv,ko,he,ro,no,hi,uk,cs,fi,hu,th,da,ca,el,bg,sr,ms,bn,hr,sl,zh_yue,az,sk,eo,ta,sh,lt,et,ml,la,bs,sq,arz,af,ka,mr,eu,tl,ang,gl,nn,ur,kk,be,hy,te,lv,mk,zh_classical,als,is,wuu,my,sco,mn,ceb,ast,cy,kn,br,an,gu,bar,uz,lb,ne,si,war,jv,ga,zh_min_nan,oc,ku,sw,nds,ckb,ia,yi,fy,scn,gan,tt,am' --eval_bleu false --dropout '0.1' --mt_steps '' --batch_size 16 --word_dropout 0 --reload_model '/checkpoint/aconneau/dumped/xlm_17_100_240_big_model_upper.2/14884511/best-valid_zh_mlm_ppl.pth' --min_count 0 --amp 2 --group_by_size true --asm false --sample_alpha '0.5' --word_pred '0.15' --clip_grad_norm 1 --emb_dim 1280 --encoder_only true --lgs 'en-es-fr-de-zh-ru-pt-it-ar-ja-id-tr-nl-pl-simple-fa-vi-sv-ko-he-ro-no-hi-uk-cs-fi-hu-th-da-ca-el-bg-sr-ms-bn-hr-sl-zh_yue-az-sk-eo-ta-sh-lt-et-ml-la-bs-sq-arz-af-ka-mr-eu-tl-ang-gl-nn-ur-kk-be-hy-te-lv-mk-zh_classical-als-is-wuu-my-sco-mn-ceb-ast-cy-kn-br-an-gu-bar-uz-lb-ne-si-war-jv-ga-zh_min_nan-oc-ku-sw-nds-ckb-ia-yi-fy-scn-gan-tt-am' --clm_steps '' --exp_name 'xlm_17_100_big.3' --lg_sampling_factor '0.7' --eval_only false --exp_id 16656234 --master_port 11363 --exp_id \"16656234\"",
|
21 |
+
"context_size": 0,
|
22 |
+
"data_path": "/private/home/aconneau/projects/XLM/data/wiki/100/175k",
|
23 |
+
"debug": false,
|
24 |
+
"debug_slurm": false,
|
25 |
+
"debug_train": false,
|
26 |
+
"dropout": 0.1,
|
27 |
"dump_path": "/checkpoint/aconneau/dumped/xlm_17_100_big.3/16656234",
|
28 |
+
"emb_dim": 1280,
|
29 |
+
"embed_init_std": 0.02209708691207961,
|
30 |
+
"encoder_only": true,
|
31 |
+
"end_n_top": 5,
|
32 |
+
"eos_index": 1,
|
33 |
+
"epoch_size": 200000,
|
34 |
+
"eval_bleu": false,
|
35 |
+
"eval_only": false,
|
36 |
"exp_id": "16656234",
|
37 |
+
"exp_name": "xlm_17_100_big.3",
|
38 |
"fp16": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
"gelu_activation": true,
|
40 |
+
"global_rank": 0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
41 |
"group_by_size": true,
|
42 |
+
"hyp_path": "/checkpoint/aconneau/dumped/xlm_17_100_big.3/16656234/hypotheses",
|
43 |
+
"id2lang": {
|
44 |
+
"0": "af",
|
45 |
+
"1": "als",
|
46 |
+
"10": "be",
|
47 |
+
"11": "bg",
|
48 |
+
"12": "bn",
|
49 |
+
"13": "br",
|
50 |
+
"14": "bs",
|
51 |
+
"15": "ca",
|
52 |
+
"16": "ceb",
|
53 |
+
"17": "ckb",
|
54 |
+
"18": "cs",
|
55 |
+
"19": "cy",
|
56 |
+
"2": "am",
|
57 |
+
"20": "da",
|
58 |
+
"21": "de",
|
59 |
+
"22": "el",
|
60 |
+
"23": "en",
|
61 |
+
"24": "eo",
|
62 |
+
"25": "es",
|
63 |
+
"26": "et",
|
64 |
+
"27": "eu",
|
65 |
+
"28": "fa",
|
66 |
+
"29": "fi",
|
67 |
+
"3": "an",
|
68 |
+
"30": "fr",
|
69 |
+
"31": "fy",
|
70 |
+
"32": "ga",
|
71 |
+
"33": "gan",
|
72 |
+
"34": "gl",
|
73 |
+
"35": "gu",
|
74 |
+
"36": "he",
|
75 |
+
"37": "hi",
|
76 |
+
"38": "hr",
|
77 |
+
"39": "hu",
|
78 |
+
"4": "ang",
|
79 |
+
"40": "hy",
|
80 |
+
"41": "ia",
|
81 |
+
"42": "id",
|
82 |
+
"43": "is",
|
83 |
+
"44": "it",
|
84 |
+
"45": "ja",
|
85 |
+
"46": "jv",
|
86 |
+
"47": "ka",
|
87 |
+
"48": "kk",
|
88 |
+
"49": "kn",
|
89 |
+
"5": "ar",
|
90 |
+
"50": "ko",
|
91 |
+
"51": "ku",
|
92 |
+
"52": "la",
|
93 |
+
"53": "lb",
|
94 |
+
"54": "lt",
|
95 |
+
"55": "lv",
|
96 |
+
"56": "mk",
|
97 |
+
"57": "ml",
|
98 |
+
"58": "mn",
|
99 |
+
"59": "mr",
|
100 |
+
"6": "arz",
|
101 |
+
"60": "ms",
|
102 |
+
"61": "my",
|
103 |
+
"62": "nds",
|
104 |
+
"63": "ne",
|
105 |
+
"64": "nl",
|
106 |
+
"65": "nn",
|
107 |
+
"66": "no",
|
108 |
+
"67": "oc",
|
109 |
+
"68": "pl",
|
110 |
+
"69": "pt",
|
111 |
+
"7": "ast",
|
112 |
+
"70": "ro",
|
113 |
+
"71": "ru",
|
114 |
+
"72": "scn",
|
115 |
+
"73": "sco",
|
116 |
+
"74": "sh",
|
117 |
+
"75": "si",
|
118 |
+
"76": "simple",
|
119 |
+
"77": "sk",
|
120 |
+
"78": "sl",
|
121 |
+
"79": "sq",
|
122 |
+
"8": "az",
|
123 |
+
"80": "sr",
|
124 |
+
"81": "sv",
|
125 |
+
"82": "sw",
|
126 |
+
"83": "ta",
|
127 |
+
"84": "te",
|
128 |
+
"85": "th",
|
129 |
+
"86": "tl",
|
130 |
+
"87": "tr",
|
131 |
+
"88": "tt",
|
132 |
+
"89": "uk",
|
133 |
+
"9": "bar",
|
134 |
+
"90": "ur",
|
135 |
+
"91": "uz",
|
136 |
+
"92": "vi",
|
137 |
+
"93": "war",
|
138 |
+
"94": "wuu",
|
139 |
+
"95": "yi",
|
140 |
+
"96": "zh",
|
141 |
+
"97": "zh_classical",
|
142 |
+
"98": "zh_min_nan",
|
143 |
+
"99": "zh_yue"
|
144 |
+
},
|
145 |
+
"init_std": 0.02,
|
146 |
+
"is_encoder": true,
|
147 |
+
"is_master": true,
|
148 |
+
"is_slurm_job": true,
|
149 |
"lambda_ae": 1.0,
|
150 |
+
"lambda_ae_config": null,
|
151 |
"lambda_bt": 1.0,
|
152 |
+
"lambda_bt_config": null,
|
153 |
+
"lambda_clm": 1.0,
|
154 |
+
"lambda_clm_config": null,
|
155 |
+
"lambda_mlm": 1.0,
|
156 |
+
"lambda_mlm_config": null,
|
157 |
+
"lambda_mt": 1.0,
|
158 |
+
"lambda_mt_config": null,
|
159 |
+
"lambda_pc": 1.0,
|
160 |
+
"lambda_pc_config": null,
|
161 |
+
"lang2id": {
|
162 |
+
"af": 0,
|
163 |
+
"als": 1,
|
164 |
+
"am": 2,
|
165 |
+
"an": 3,
|
166 |
+
"ang": 4,
|
167 |
+
"ar": 5,
|
168 |
+
"arz": 6,
|
169 |
+
"ast": 7,
|
170 |
+
"az": 8,
|
171 |
+
"bar": 9,
|
172 |
+
"be": 10,
|
173 |
+
"bg": 11,
|
174 |
+
"bn": 12,
|
175 |
+
"br": 13,
|
176 |
+
"bs": 14,
|
177 |
+
"ca": 15,
|
178 |
+
"ceb": 16,
|
179 |
+
"ckb": 17,
|
180 |
+
"cs": 18,
|
181 |
+
"cy": 19,
|
182 |
+
"da": 20,
|
183 |
+
"de": 21,
|
184 |
+
"el": 22,
|
185 |
+
"en": 23,
|
186 |
+
"eo": 24,
|
187 |
+
"es": 25,
|
188 |
+
"et": 26,
|
189 |
+
"eu": 27,
|
190 |
+
"fa": 28,
|
191 |
+
"fi": 29,
|
192 |
+
"fr": 30,
|
193 |
+
"fy": 31,
|
194 |
+
"ga": 32,
|
195 |
+
"gan": 33,
|
196 |
+
"gl": 34,
|
197 |
+
"gu": 35,
|
198 |
+
"he": 36,
|
199 |
+
"hi": 37,
|
200 |
+
"hr": 38,
|
201 |
+
"hu": 39,
|
202 |
+
"hy": 40,
|
203 |
+
"ia": 41,
|
204 |
+
"id": 42,
|
205 |
+
"is": 43,
|
206 |
+
"it": 44,
|
207 |
+
"ja": 45,
|
208 |
+
"jv": 46,
|
209 |
+
"ka": 47,
|
210 |
+
"kk": 48,
|
211 |
+
"kn": 49,
|
212 |
+
"ko": 50,
|
213 |
+
"ku": 51,
|
214 |
+
"la": 52,
|
215 |
+
"lb": 53,
|
216 |
+
"lt": 54,
|
217 |
+
"lv": 55,
|
218 |
+
"mk": 56,
|
219 |
+
"ml": 57,
|
220 |
+
"mn": 58,
|
221 |
+
"mr": 59,
|
222 |
+
"ms": 60,
|
223 |
+
"my": 61,
|
224 |
+
"nds": 62,
|
225 |
+
"ne": 63,
|
226 |
+
"nl": 64,
|
227 |
+
"nn": 65,
|
228 |
+
"no": 66,
|
229 |
+
"oc": 67,
|
230 |
+
"pl": 68,
|
231 |
+
"pt": 69,
|
232 |
+
"ro": 70,
|
233 |
+
"ru": 71,
|
234 |
+
"scn": 72,
|
235 |
+
"sco": 73,
|
236 |
+
"sh": 74,
|
237 |
+
"si": 75,
|
238 |
+
"simple": 76,
|
239 |
+
"sk": 77,
|
240 |
+
"sl": 78,
|
241 |
+
"sq": 79,
|
242 |
+
"sr": 80,
|
243 |
+
"sv": 81,
|
244 |
+
"sw": 82,
|
245 |
+
"ta": 83,
|
246 |
+
"te": 84,
|
247 |
+
"th": 85,
|
248 |
+
"tl": 86,
|
249 |
+
"tr": 87,
|
250 |
+
"tt": 88,
|
251 |
+
"uk": 89,
|
252 |
+
"ur": 90,
|
253 |
+
"uz": 91,
|
254 |
+
"vi": 92,
|
255 |
+
"war": 93,
|
256 |
+
"wuu": 94,
|
257 |
+
"yi": 95,
|
258 |
+
"zh": 96,
|
259 |
+
"zh_classical": 97,
|
260 |
+
"zh_min_nan": 98,
|
261 |
+
"zh_yue": 99
|
262 |
+
},
|
263 |
+
"lang_id": 0,
|
264 |
+
"langs": [
|
265 |
+
"en",
|
266 |
+
"es",
|
267 |
+
"fr",
|
268 |
+
"de",
|
269 |
+
"zh",
|
270 |
+
"ru",
|
271 |
+
"pt",
|
272 |
+
"it",
|
273 |
+
"ar",
|
274 |
+
"ja",
|
275 |
+
"id",
|
276 |
+
"tr",
|
277 |
+
"nl",
|
278 |
+
"pl",
|
279 |
+
"simple",
|
280 |
+
"fa",
|
281 |
+
"vi",
|
282 |
+
"sv",
|
283 |
+
"ko",
|
284 |
+
"he",
|
285 |
+
"ro",
|
286 |
+
"no",
|
287 |
+
"hi",
|
288 |
+
"uk",
|
289 |
+
"cs",
|
290 |
+
"fi",
|
291 |
+
"hu",
|
292 |
+
"th",
|
293 |
+
"da",
|
294 |
+
"ca",
|
295 |
+
"el",
|
296 |
+
"bg",
|
297 |
+
"sr",
|
298 |
+
"ms",
|
299 |
+
"bn",
|
300 |
+
"hr",
|
301 |
+
"sl",
|
302 |
+
"zh_yue",
|
303 |
+
"az",
|
304 |
+
"sk",
|
305 |
+
"eo",
|
306 |
+
"ta",
|
307 |
+
"sh",
|
308 |
+
"lt",
|
309 |
+
"et",
|
310 |
+
"ml",
|
311 |
+
"la",
|
312 |
+
"bs",
|
313 |
+
"sq",
|
314 |
+
"arz",
|
315 |
+
"af",
|
316 |
+
"ka",
|
317 |
+
"mr",
|
318 |
+
"eu",
|
319 |
+
"tl",
|
320 |
+
"ang",
|
321 |
+
"gl",
|
322 |
+
"nn",
|
323 |
+
"ur",
|
324 |
+
"kk",
|
325 |
+
"be",
|
326 |
+
"hy",
|
327 |
+
"te",
|
328 |
+
"lv",
|
329 |
+
"mk",
|
330 |
+
"zh_classical",
|
331 |
+
"als",
|
332 |
+
"is",
|
333 |
+
"wuu",
|
334 |
+
"my",
|
335 |
+
"sco",
|
336 |
+
"mn",
|
337 |
+
"ceb",
|
338 |
+
"ast",
|
339 |
+
"cy",
|
340 |
+
"kn",
|
341 |
+
"br",
|
342 |
+
"an",
|
343 |
+
"gu",
|
344 |
+
"bar",
|
345 |
+
"uz",
|
346 |
+
"lb",
|
347 |
+
"ne",
|
348 |
+
"si",
|
349 |
+
"war",
|
350 |
+
"jv",
|
351 |
+
"ga",
|
352 |
+
"zh_min_nan",
|
353 |
+
"oc",
|
354 |
+
"ku",
|
355 |
+
"sw",
|
356 |
+
"nds",
|
357 |
+
"ckb",
|
358 |
+
"ia",
|
359 |
+
"yi",
|
360 |
+
"fy",
|
361 |
+
"scn",
|
362 |
+
"gan",
|
363 |
+
"tt",
|
364 |
+
"am"
|
365 |
+
],
|
366 |
+
"layer_norm_eps": 1e-12,
|
367 |
+
"lg_sampling_factor": 0.7,
|
368 |
+
"lgs": "en-es-fr-de-zh-ru-pt-it-ar-ja-id-tr-nl-pl-simple-fa-vi-sv-ko-he-ro-no-hi-uk-cs-fi-hu-th-da-ca-el-bg-sr-ms-bn-hr-sl-zh_yue-az-sk-eo-ta-sh-lt-et-ml-la-bs-sq-arz-af-ka-mr-eu-tl-ang-gl-nn-ur-kk-be-hy-te-lv-mk-zh_classical-als-is-wuu-my-sco-mn-ceb-ast-cy-kn-br-an-gu-bar-uz-lb-ne-si-war-jv-ga-zh_min_nan-oc-ku-sw-nds-ckb-ia-yi-fy-scn-gan-tt-am",
|
369 |
+
"local_rank": 0,
|
370 |
+
"mask_index": 5,
|
371 |
+
"mask_token_id": 0,
|
372 |
+
"master_addr": "learnfair0332",
|
373 |
+
"master_port": 11363,
|
374 |
+
"max_batch_size": 0,
|
375 |
+
"max_epoch": 100000,
|
376 |
+
"max_len": 200,
|
377 |
+
"max_position_embeddings": 512,
|
378 |
+
"max_vocab": 200000,
|
379 |
+
"min_count": 0,
|
380 |
+
"mlm_steps": [
|
381 |
+
[
|
382 |
+
"en",
|
383 |
null
|
384 |
],
|
385 |
[
|
386 |
+
"es",
|
387 |
null
|
388 |
],
|
389 |
[
|
390 |
+
"fr",
|
391 |
null
|
392 |
],
|
393 |
[
|
394 |
+
"de",
|
395 |
null
|
396 |
],
|
397 |
[
|
398 |
+
"zh",
|
399 |
null
|
400 |
],
|
401 |
[
|
402 |
+
"ru",
|
403 |
null
|
404 |
],
|
405 |
[
|
406 |
+
"pt",
|
407 |
null
|
408 |
],
|
409 |
[
|
410 |
+
"it",
|
411 |
null
|
412 |
],
|
413 |
[
|
414 |
+
"ar",
|
415 |
null
|
416 |
],
|
417 |
[
|
418 |
+
"ja",
|
419 |
null
|
420 |
],
|
421 |
[
|
422 |
+
"id",
|
423 |
null
|
424 |
],
|
425 |
[
|
426 |
+
"tr",
|
427 |
null
|
428 |
],
|
429 |
[
|
430 |
+
"nl",
|
431 |
null
|
432 |
],
|
433 |
[
|
434 |
+
"pl",
|
435 |
null
|
436 |
],
|
437 |
[
|
438 |
+
"simple",
|
439 |
null
|
440 |
],
|
441 |
[
|
442 |
+
"fa",
|
443 |
null
|
444 |
],
|
445 |
[
|
446 |
+
"vi",
|
447 |
null
|
448 |
],
|
449 |
[
|
450 |
+
"sv",
|
451 |
null
|
452 |
],
|
453 |
[
|
454 |
+
"ko",
|
455 |
null
|
456 |
],
|
457 |
[
|
458 |
+
"he",
|
459 |
null
|
460 |
],
|
461 |
[
|
462 |
+
"ro",
|
463 |
null
|
464 |
],
|
465 |
[
|
466 |
+
"no",
|
467 |
null
|
468 |
],
|
469 |
[
|
470 |
+
"hi",
|
471 |
null
|
472 |
],
|
473 |
[
|
474 |
+
"uk",
|
475 |
null
|
476 |
],
|
477 |
[
|
478 |
+
"cs",
|
479 |
null
|
480 |
],
|
481 |
[
|
482 |
+
"fi",
|
483 |
null
|
484 |
],
|
485 |
[
|
486 |
+
"hu",
|
487 |
null
|
488 |
],
|
489 |
[
|
490 |
+
"th",
|
491 |
null
|
492 |
],
|
493 |
[
|
494 |
+
"da",
|
495 |
null
|
496 |
],
|
497 |
[
|
498 |
+
"ca",
|
499 |
null
|
500 |
],
|
501 |
[
|
502 |
+
"el",
|
503 |
null
|
504 |
],
|
505 |
[
|
506 |
+
"bg",
|
507 |
null
|
508 |
],
|
509 |
[
|
510 |
+
"sr",
|
511 |
null
|
512 |
],
|
513 |
[
|
514 |
+
"ms",
|
515 |
null
|
516 |
],
|
517 |
[
|
518 |
+
"bn",
|
519 |
null
|
520 |
],
|
521 |
[
|
522 |
+
"hr",
|
523 |
null
|
524 |
],
|
525 |
[
|
526 |
+
"sl",
|
527 |
null
|
528 |
],
|
529 |
[
|
530 |
+
"zh_yue",
|
531 |
null
|
532 |
],
|
533 |
[
|
534 |
+
"az",
|
535 |
null
|
536 |
],
|
537 |
[
|
538 |
+
"sk",
|
539 |
null
|
540 |
],
|
541 |
[
|
542 |
+
"eo",
|
543 |
null
|
544 |
],
|
545 |
[
|
546 |
+
"ta",
|
547 |
null
|
548 |
],
|
549 |
[
|
550 |
+
"sh",
|
551 |
null
|
552 |
],
|
553 |
[
|
554 |
+
"lt",
|
555 |
null
|
556 |
],
|
557 |
[
|
558 |
+
"et",
|
559 |
null
|
560 |
],
|
561 |
[
|
562 |
+
"ml",
|
563 |
null
|
564 |
],
|
565 |
[
|
566 |
+
"la",
|
567 |
null
|
568 |
],
|
569 |
[
|
570 |
+
"bs",
|
571 |
null
|
572 |
],
|
573 |
[
|
574 |
+
"sq",
|
575 |
null
|
576 |
],
|
577 |
[
|
578 |
+
"arz",
|
579 |
null
|
580 |
],
|
581 |
[
|
582 |
+
"af",
|
583 |
null
|
584 |
],
|
585 |
[
|
586 |
+
"ka",
|
587 |
null
|
588 |
],
|
589 |
[
|
590 |
+
"mr",
|
591 |
null
|
592 |
],
|
593 |
[
|
594 |
+
"eu",
|
595 |
null
|
596 |
],
|
597 |
[
|
598 |
+
"tl",
|
599 |
null
|
600 |
],
|
601 |
[
|
602 |
+
"ang",
|
603 |
null
|
604 |
],
|
605 |
[
|
606 |
+
"gl",
|
607 |
null
|
608 |
],
|
609 |
[
|
610 |
+
"nn",
|
611 |
null
|
612 |
],
|
613 |
[
|
614 |
+
"ur",
|
615 |
null
|
616 |
],
|
617 |
[
|
618 |
+
"kk",
|
619 |
null
|
620 |
],
|
621 |
[
|
622 |
+
"be",
|
623 |
null
|
624 |
],
|
625 |
[
|
626 |
+
"hy",
|
627 |
null
|
628 |
],
|
629 |
[
|
630 |
+
"te",
|
631 |
null
|
632 |
],
|
633 |
[
|
634 |
+
"lv",
|
635 |
null
|
636 |
],
|
637 |
[
|
638 |
+
"mk",
|
639 |
null
|
640 |
],
|
641 |
[
|
642 |
+
"zh_classical",
|
643 |
null
|
644 |
],
|
645 |
[
|
646 |
+
"als",
|
647 |
null
|
648 |
],
|
649 |
[
|
650 |
+
"is",
|
651 |
null
|
652 |
],
|
653 |
[
|
654 |
+
"wuu",
|
655 |
null
|
656 |
],
|
657 |
[
|
658 |
+
"my",
|
659 |
null
|
660 |
],
|
661 |
[
|
662 |
+
"sco",
|
663 |
null
|
664 |
],
|
665 |
[
|
666 |
+
"mn",
|
667 |
null
|
668 |
],
|
669 |
[
|
670 |
+
"ceb",
|
671 |
null
|
672 |
],
|
673 |
[
|
674 |
+
"ast",
|
675 |
null
|
676 |
],
|
677 |
[
|
678 |
+
"cy",
|
679 |
null
|
680 |
],
|
681 |
[
|
682 |
+
"kn",
|
683 |
null
|
684 |
],
|
685 |
[
|
686 |
+
"br",
|
687 |
null
|
688 |
],
|
689 |
[
|
690 |
+
"an",
|
691 |
null
|
692 |
],
|
693 |
[
|
694 |
+
"gu",
|
695 |
null
|
696 |
],
|
697 |
[
|
698 |
+
"bar",
|
699 |
null
|
700 |
],
|
701 |
[
|
702 |
+
"uz",
|
703 |
null
|
704 |
+
],
|
705 |
+
[
|
706 |
+
"lb",
|
707 |
+
null
|
708 |
+
],
|
709 |
+
[
|
710 |
+
"ne",
|
711 |
+
null
|
712 |
+
],
|
713 |
+
[
|
714 |
+
"si",
|
715 |
+
null
|
716 |
+
],
|
717 |
+
[
|
718 |
+
"war",
|
719 |
+
null
|
720 |
+
],
|
721 |
+
[
|
722 |
+
"jv",
|
723 |
+
null
|
724 |
+
],
|
725 |
+
[
|
726 |
+
"ga",
|
727 |
+
null
|
728 |
+
],
|
729 |
+
[
|
730 |
+
"zh_min_nan",
|
731 |
+
null
|
732 |
+
],
|
733 |
+
[
|
734 |
+
"oc",
|
735 |
+
null
|
736 |
+
],
|
737 |
+
[
|
738 |
+
"ku",
|
739 |
+
null
|
740 |
+
],
|
741 |
+
[
|
742 |
+
"sw",
|
743 |
+
null
|
744 |
+
],
|
745 |
+
[
|
746 |
+
"nds",
|
747 |
+
null
|
748 |
+
],
|
749 |
+
[
|
750 |
+
"ckb",
|
751 |
+
null
|
752 |
+
],
|
753 |
+
[
|
754 |
+
"ia",
|
755 |
+
null
|
756 |
+
],
|
757 |
+
[
|
758 |
+
"yi",
|
759 |
+
null
|
760 |
+
],
|
761 |
+
[
|
762 |
+
"fy",
|
763 |
+
null
|
764 |
+
],
|
765 |
+
[
|
766 |
+
"scn",
|
767 |
+
null
|
768 |
+
],
|
769 |
+
[
|
770 |
+
"gan",
|
771 |
+
null
|
772 |
+
],
|
773 |
+
[
|
774 |
+
"tt",
|
775 |
+
null
|
776 |
+
],
|
777 |
+
[
|
778 |
+
"am",
|
779 |
+
null
|
780 |
+
]
|
781 |
+
],
|
782 |
+
"model_type": "xlm",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
783 |
"mono_dataset": {
|
784 |
+
"af": {
|
785 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.af.pth",
|
786 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.af.pth",
|
787 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.af.pth"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
788 |
},
|
789 |
+
"als": {
|
790 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.als.pth",
|
791 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.als.pth",
|
792 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.als.pth"
|
793 |
},
|
794 |
+
"am": {
|
795 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.am.pth",
|
796 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.am.pth",
|
797 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.am.pth"
|
798 |
},
|
799 |
+
"an": {
|
800 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.an.pth",
|
801 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.an.pth",
|
802 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.an.pth"
|
803 |
},
|
804 |
+
"ang": {
|
805 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ang.pth",
|
806 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ang.pth",
|
807 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ang.pth"
|
808 |
},
|
809 |
"ar": {
|
810 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ar.pth",
|
811 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ar.pth",
|
812 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ar.pth"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
813 |
},
|
814 |
+
"arz": {
|
815 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.arz.pth",
|
816 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.arz.pth",
|
817 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.arz.pth"
|
|
|
|
|
|
|
|
|
|
|
818 |
},
|
819 |
+
"ast": {
|
820 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ast.pth",
|
821 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ast.pth",
|
822 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ast.pth"
|
823 |
},
|
824 |
+
"az": {
|
825 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.az.pth",
|
826 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.az.pth",
|
827 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.az.pth"
|
828 |
},
|
829 |
+
"bar": {
|
830 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bar.pth",
|
831 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bar.pth",
|
832 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bar.pth"
|
833 |
},
|
834 |
+
"be": {
|
835 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.be.pth",
|
836 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.be.pth",
|
837 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.be.pth"
|
838 |
},
|
839 |
+
"bg": {
|
840 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bg.pth",
|
841 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bg.pth",
|
842 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bg.pth"
|
843 |
},
|
844 |
+
"bn": {
|
845 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bn.pth",
|
846 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bn.pth",
|
847 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bn.pth"
|
848 |
},
|
849 |
+
"br": {
|
850 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.br.pth",
|
851 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.br.pth",
|
852 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.br.pth"
|
853 |
},
|
854 |
+
"bs": {
|
855 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bs.pth",
|
856 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bs.pth",
|
857 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bs.pth"
|
858 |
},
|
859 |
+
"ca": {
|
860 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ca.pth",
|
861 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ca.pth",
|
862 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ca.pth"
|
863 |
},
|
864 |
+
"ceb": {
|
865 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ceb.pth",
|
866 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ceb.pth",
|
867 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ceb.pth"
|
868 |
},
|
869 |
+
"ckb": {
|
870 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ckb.pth",
|
871 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ckb.pth",
|
872 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ckb.pth"
|
873 |
},
|
874 |
"cs": {
|
875 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.cs.pth",
|
876 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.cs.pth",
|
877 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.cs.pth"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
878 |
},
|
879 |
+
"cy": {
|
880 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.cy.pth",
|
881 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.cy.pth",
|
882 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.cy.pth"
|
883 |
},
|
884 |
"da": {
|
885 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.da.pth",
|
886 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.da.pth",
|
887 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.da.pth"
|
|
|
888 |
},
|
889 |
+
"de": {
|
890 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.de.pth",
|
891 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.de.pth",
|
892 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.de.pth"
|
893 |
},
|
894 |
"el": {
|
895 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.el.pth",
|
896 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.el.pth",
|
897 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.el.pth"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
898 |
},
|
899 |
+
"en": {
|
900 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.en.pth",
|
901 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.en.pth",
|
902 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.en.pth"
|
903 |
},
|
904 |
+
"eo": {
|
905 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.eo.pth",
|
906 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.eo.pth",
|
907 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.eo.pth"
|
908 |
},
|
909 |
+
"es": {
|
910 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.es.pth",
|
911 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.es.pth",
|
912 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.es.pth"
|
913 |
},
|
914 |
+
"et": {
|
915 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.et.pth",
|
916 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.et.pth",
|
917 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.et.pth"
|
918 |
},
|
919 |
+
"eu": {
|
920 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.eu.pth",
|
921 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.eu.pth",
|
922 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.eu.pth"
|
923 |
},
|
924 |
+
"fa": {
|
925 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fa.pth",
|
926 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fa.pth",
|
927 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fa.pth"
|
928 |
},
|
929 |
+
"fi": {
|
930 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fi.pth",
|
931 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fi.pth",
|
932 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fi.pth"
|
933 |
},
|
934 |
+
"fr": {
|
935 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fr.pth",
|
936 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fr.pth",
|
937 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fr.pth"
|
938 |
},
|
939 |
+
"fy": {
|
940 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fy.pth",
|
941 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fy.pth",
|
942 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fy.pth"
|
943 |
},
|
944 |
+
"ga": {
|
945 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ga.pth",
|
946 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ga.pth",
|
947 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ga.pth"
|
948 |
},
|
949 |
+
"gan": {
|
950 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gan.pth",
|
951 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gan.pth",
|
952 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gan.pth"
|
953 |
},
|
954 |
+
"gl": {
|
955 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gl.pth",
|
956 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gl.pth",
|
957 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gl.pth"
|
958 |
},
|
959 |
+
"gu": {
|
960 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gu.pth",
|
961 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gu.pth",
|
962 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gu.pth"
|
963 |
},
|
964 |
+
"he": {
|
965 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.he.pth",
|
966 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.he.pth",
|
967 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.he.pth"
|
968 |
},
|
969 |
+
"hi": {
|
970 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hi.pth",
|
971 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hi.pth",
|
972 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hi.pth"
|
973 |
},
|
974 |
+
"hr": {
|
975 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hr.pth",
|
976 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hr.pth",
|
977 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hr.pth"
|
978 |
},
|
979 |
+
"hu": {
|
980 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hu.pth",
|
981 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hu.pth",
|
982 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hu.pth"
|
983 |
},
|
984 |
+
"hy": {
|
985 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hy.pth",
|
986 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hy.pth",
|
987 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hy.pth"
|
988 |
},
|
989 |
+
"ia": {
|
990 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ia.pth",
|
991 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ia.pth",
|
992 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ia.pth"
|
993 |
},
|
994 |
+
"id": {
|
995 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.id.pth",
|
996 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.id.pth",
|
997 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.id.pth"
|
998 |
},
|
999 |
+
"is": {
|
1000 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.is.pth",
|
1001 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.is.pth",
|
1002 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.is.pth"
|
1003 |
},
|
1004 |
+
"it": {
|
1005 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.it.pth",
|
1006 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.it.pth",
|
1007 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.it.pth"
|
1008 |
},
|
1009 |
+
"ja": {
|
1010 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ja.pth",
|
1011 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ja.pth",
|
1012 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ja.pth"
|
1013 |
},
|
1014 |
+
"jv": {
|
1015 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.jv.pth",
|
1016 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.jv.pth",
|
1017 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.jv.pth"
|
1018 |
},
|
1019 |
+
"ka": {
|
1020 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ka.pth",
|
1021 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ka.pth",
|
1022 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ka.pth"
|
1023 |
},
|
1024 |
"kk": {
|
1025 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.kk.pth",
|
1026 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.kk.pth",
|
1027 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.kk.pth"
|
|
|
1028 |
},
|
1029 |
+
"kn": {
|
1030 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.kn.pth",
|
1031 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.kn.pth",
|
1032 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.kn.pth"
|
1033 |
},
|
1034 |
+
"ko": {
|
1035 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ko.pth",
|
1036 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ko.pth",
|
1037 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ko.pth"
|
1038 |
},
|
1039 |
+
"ku": {
|
1040 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ku.pth",
|
1041 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ku.pth",
|
1042 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ku.pth"
|
1043 |
+
},
|
1044 |
+
"la": {
|
1045 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.la.pth",
|
1046 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.la.pth",
|
1047 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.la.pth"
|
1048 |
+
},
|
1049 |
+
"lb": {
|
1050 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lb.pth",
|
1051 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lb.pth",
|
1052 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lb.pth"
|
1053 |
+
},
|
1054 |
+
"lt": {
|
1055 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lt.pth",
|
1056 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lt.pth",
|
1057 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lt.pth"
|
1058 |
},
|
1059 |
"lv": {
|
1060 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lv.pth",
|
1061 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lv.pth",
|
1062 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lv.pth"
|
|
|
1063 |
},
|
1064 |
"mk": {
|
1065 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mk.pth",
|
1066 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mk.pth",
|
1067 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mk.pth"
|
|
|
1068 |
},
|
1069 |
+
"ml": {
|
1070 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ml.pth",
|
1071 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ml.pth",
|
1072 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ml.pth"
|
1073 |
},
|
1074 |
+
"mn": {
|
1075 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mn.pth",
|
1076 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mn.pth",
|
1077 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mn.pth"
|
1078 |
},
|
1079 |
+
"mr": {
|
1080 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mr.pth",
|
1081 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mr.pth",
|
1082 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mr.pth"
|
1083 |
},
|
1084 |
+
"ms": {
|
1085 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ms.pth",
|
1086 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ms.pth",
|
1087 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ms.pth"
|
1088 |
},
|
1089 |
"my": {
|
1090 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.my.pth",
|
1091 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.my.pth",
|
1092 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.my.pth"
|
|
|
1093 |
},
|
1094 |
+
"nds": {
|
1095 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nds.pth",
|
1096 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nds.pth",
|
1097 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nds.pth"
|
1098 |
},
|
1099 |
+
"ne": {
|
1100 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ne.pth",
|
1101 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ne.pth",
|
1102 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ne.pth"
|
1103 |
},
|
1104 |
+
"nl": {
|
1105 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nl.pth",
|
1106 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nl.pth",
|
1107 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nl.pth"
|
1108 |
},
|
1109 |
+
"nn": {
|
1110 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nn.pth",
|
1111 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nn.pth",
|
1112 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nn.pth"
|
1113 |
},
|
1114 |
+
"no": {
|
1115 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.no.pth",
|
1116 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.no.pth",
|
1117 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.no.pth"
|
1118 |
},
|
1119 |
+
"oc": {
|
1120 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.oc.pth",
|
1121 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.oc.pth",
|
1122 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.oc.pth"
|
1123 |
},
|
1124 |
+
"pl": {
|
1125 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.pl.pth",
|
1126 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.pl.pth",
|
1127 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.pl.pth"
|
1128 |
},
|
1129 |
+
"pt": {
|
1130 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.pt.pth",
|
1131 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.pt.pth",
|
1132 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.pt.pth"
|
1133 |
},
|
1134 |
+
"ro": {
|
1135 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ro.pth",
|
1136 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ro.pth",
|
1137 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ro.pth"
|
1138 |
},
|
1139 |
+
"ru": {
|
1140 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ru.pth",
|
1141 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ru.pth",
|
1142 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ru.pth"
|
1143 |
},
|
1144 |
+
"scn": {
|
1145 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.scn.pth",
|
1146 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.scn.pth",
|
1147 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.scn.pth"
|
1148 |
},
|
1149 |
+
"sco": {
|
1150 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sco.pth",
|
1151 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sco.pth",
|
1152 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sco.pth"
|
1153 |
},
|
1154 |
+
"sh": {
|
1155 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sh.pth",
|
1156 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sh.pth",
|
1157 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sh.pth"
|
1158 |
},
|
1159 |
"si": {
|
1160 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.si.pth",
|
1161 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.si.pth",
|
1162 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.si.pth"
|
|
|
1163 |
},
|
1164 |
+
"simple": {
|
1165 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.simple.pth",
|
1166 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.simple.pth",
|
1167 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.simple.pth"
|
1168 |
},
|
1169 |
+
"sk": {
|
1170 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sk.pth",
|
1171 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sk.pth",
|
1172 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sk.pth"
|
1173 |
},
|
1174 |
+
"sl": {
|
1175 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sl.pth",
|
1176 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sl.pth",
|
1177 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sl.pth"
|
1178 |
},
|
1179 |
+
"sq": {
|
1180 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sq.pth",
|
1181 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sq.pth",
|
1182 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sq.pth"
|
1183 |
},
|
1184 |
+
"sr": {
|
1185 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sr.pth",
|
1186 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sr.pth",
|
1187 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sr.pth"
|
1188 |
},
|
1189 |
+
"sv": {
|
1190 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sv.pth",
|
1191 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sv.pth",
|
1192 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sv.pth"
|
1193 |
},
|
1194 |
"sw": {
|
1195 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sw.pth",
|
1196 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sw.pth",
|
1197 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sw.pth"
|
|
|
1198 |
},
|
1199 |
+
"ta": {
|
1200 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ta.pth",
|
1201 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ta.pth",
|
1202 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ta.pth"
|
1203 |
},
|
1204 |
+
"te": {
|
1205 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.te.pth",
|
1206 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.te.pth",
|
1207 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.te.pth"
|
1208 |
},
|
1209 |
+
"th": {
|
1210 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.th.pth",
|
1211 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.th.pth",
|
1212 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.th.pth"
|
1213 |
+
},
|
1214 |
+
"tl": {
|
1215 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tl.pth",
|
1216 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tl.pth",
|
1217 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tl.pth"
|
1218 |
+
},
|
1219 |
+
"tr": {
|
1220 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tr.pth",
|
1221 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tr.pth",
|
1222 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tr.pth"
|
1223 |
+
},
|
1224 |
+
"tt": {
|
1225 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tt.pth",
|
1226 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tt.pth",
|
1227 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tt.pth"
|
1228 |
+
},
|
1229 |
+
"uk": {
|
1230 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.uk.pth",
|
1231 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.uk.pth",
|
1232 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.uk.pth"
|
1233 |
+
},
|
1234 |
+
"ur": {
|
1235 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ur.pth",
|
1236 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ur.pth",
|
1237 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ur.pth"
|
1238 |
+
},
|
1239 |
+
"uz": {
|
1240 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.uz.pth",
|
1241 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.uz.pth",
|
1242 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.uz.pth"
|
1243 |
+
},
|
1244 |
+
"vi": {
|
1245 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.vi.pth",
|
1246 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.vi.pth",
|
1247 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.vi.pth"
|
1248 |
+
},
|
1249 |
+
"war": {
|
1250 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.war.pth",
|
1251 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.war.pth",
|
1252 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.war.pth"
|
1253 |
+
},
|
1254 |
+
"wuu": {
|
1255 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.wuu.pth",
|
1256 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.wuu.pth",
|
1257 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.wuu.pth"
|
1258 |
},
|
1259 |
"yi": {
|
1260 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.yi.pth",
|
1261 |
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.yi.pth",
|
1262 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.yi.pth"
|
|
|
|
|
|
|
|
|
|
|
|
|
1263 |
},
|
1264 |
+
"zh": {
|
1265 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh.pth",
|
1266 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh.pth",
|
1267 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh.pth"
|
1268 |
},
|
1269 |
+
"zh_classical": {
|
1270 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_classical.pth",
|
1271 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_classical.pth",
|
1272 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_classical.pth"
|
1273 |
},
|
1274 |
+
"zh_min_nan": {
|
1275 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_min_nan.pth",
|
1276 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_min_nan.pth",
|
1277 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_min_nan.pth"
|
1278 |
},
|
1279 |
+
"zh_yue": {
|
1280 |
+
"test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_yue.pth",
|
1281 |
+
"train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_yue.pth",
|
1282 |
+
"valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_yue.pth"
|
1283 |
}
|
1284 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1285 |
"mono_list": [
|
1286 |
"en",
|
1287 |
"es",
|
|
|
1384 |
"tt",
|
1385 |
"am"
|
1386 |
],
|
1387 |
+
"mt_steps": [],
|
1388 |
+
"multi_gpu": true,
|
1389 |
+
"multi_node": true,
|
1390 |
+
"n_gpu_per_node": 8,
|
1391 |
+
"n_heads": 16,
|
1392 |
+
"n_langs": 100,
|
1393 |
+
"n_layers": 16,
|
1394 |
+
"n_nodes": 4,
|
1395 |
+
"node_id": 0,
|
1396 |
+
"optimizer": "adam_inverse_sqrt,lr=0.00005,warmup_updates=30000,beta1=0.9,beta2=0.999,weight_decay=0.01,eps=0.000001",
|
1397 |
+
"pad_index": 2,
|
1398 |
+
"pad_token_id": 2,
|
1399 |
+
"para_dataset": {},
|
1400 |
+
"para_list": [],
|
1401 |
+
"pc_steps": [],
|
1402 |
+
"ref_paths": {},
|
1403 |
+
"reload_checkpoint": "",
|
1404 |
+
"reload_emb": "",
|
1405 |
+
"reload_model": "/checkpoint/aconneau/dumped/xlm_17_100_240_big_model_upper.2/14884511/best-valid_zh_mlm_ppl.pth",
|
1406 |
+
"sample_alpha": 0.5,
|
1407 |
+
"save_periodic": 0,
|
1408 |
+
"share_inout_emb": true,
|
1409 |
+
"sinusoidal_embeddings": false,
|
1410 |
+
"split_data": true,
|
1411 |
+
"start_n_top": 5,
|
1412 |
+
"stopping_criterion": "_valid_zh_mlm_ppl,25",
|
1413 |
+
"summary_activation": null,
|
1414 |
+
"summary_first_dropout": 0.1,
|
1415 |
+
"summary_proj_to_labels": true,
|
1416 |
+
"summary_type": "first",
|
1417 |
+
"summary_use_proj": true,
|
1418 |
+
"tokens_per_batch": -1,
|
1419 |
+
"unk_index": 3,
|
1420 |
+
"use_lang_emb": false,
|
1421 |
+
"use_memory": false,
|
1422 |
+
"validation_metrics": "_valid_en_mlm_ppl,_valid_mlm_ppl,_valid_zh_mlm_ppl",
|
1423 |
+
"vocab_size": 200000,
|
1424 |
+
"word_blank": 0.0,
|
1425 |
+
"word_dropout": 0.0,
|
1426 |
+
"word_keep": 0.1,
|
1427 |
+
"word_mask": 0.8,
|
1428 |
+
"word_mask_keep_rand": "0.8,0.1,0.1",
|
1429 |
+
"word_pred": 0.15,
|
1430 |
+
"word_rand": 0.1,
|
1431 |
+
"word_shuffle": 0.0,
|
1432 |
+
"world_size": 32
|
1433 |
}
|