KonradBRG commited on
Commit
96a0b97
·
verified ·
1 Parent(s): 552ede4

Training in progress, epoch 1

Browse files
Files changed (3) hide show
  1. config.json +1447 -0
  2. model.safetensors +3 -0
  3. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,1447 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "accumulate_gradients": 4,
3
+ "ae_steps": [],
4
+ "amp": 2,
5
+ "architectures": [
6
+ "XLMForSequenceClassification"
7
+ ],
8
+ "asm": false,
9
+ "attention_dropout": 0.1,
10
+ "batch_size": 16,
11
+ "beam_size": 1,
12
+ "bos_token_id": 0,
13
+ "bptt": 256,
14
+ "bt_src_langs": [],
15
+ "bt_steps": [],
16
+ "causal": false,
17
+ "clip_grad_norm": 1.0,
18
+ "clm_steps": [],
19
+ "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\"",
20
+ "context_size": 0,
21
+ "data_path": "/private/home/aconneau/projects/XLM/data/wiki/100/175k",
22
+ "debug": false,
23
+ "debug_slurm": false,
24
+ "debug_train": false,
25
+ "dropout": 0.1,
26
+ "dtype": "float32",
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_token_id": 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
+ "id2label": {
44
+ "0": "literal",
45
+ "1": "idiom",
46
+ "2": "metaphor",
47
+ "3": "simile"
48
+ },
49
+ "id2lang": {
50
+ "0": "af",
51
+ "1": "als",
52
+ "10": "be",
53
+ "11": "bg",
54
+ "12": "bn",
55
+ "13": "br",
56
+ "14": "bs",
57
+ "15": "ca",
58
+ "16": "ceb",
59
+ "17": "ckb",
60
+ "18": "cs",
61
+ "19": "cy",
62
+ "2": "am",
63
+ "20": "da",
64
+ "21": "de",
65
+ "22": "el",
66
+ "23": "en",
67
+ "24": "eo",
68
+ "25": "es",
69
+ "26": "et",
70
+ "27": "eu",
71
+ "28": "fa",
72
+ "29": "fi",
73
+ "3": "an",
74
+ "30": "fr",
75
+ "31": "fy",
76
+ "32": "ga",
77
+ "33": "gan",
78
+ "34": "gl",
79
+ "35": "gu",
80
+ "36": "he",
81
+ "37": "hi",
82
+ "38": "hr",
83
+ "39": "hu",
84
+ "4": "ang",
85
+ "40": "hy",
86
+ "41": "ia",
87
+ "42": "id",
88
+ "43": "is",
89
+ "44": "it",
90
+ "45": "ja",
91
+ "46": "jv",
92
+ "47": "ka",
93
+ "48": "kk",
94
+ "49": "kn",
95
+ "5": "ar",
96
+ "50": "ko",
97
+ "51": "ku",
98
+ "52": "la",
99
+ "53": "lb",
100
+ "54": "lt",
101
+ "55": "lv",
102
+ "56": "mk",
103
+ "57": "ml",
104
+ "58": "mn",
105
+ "59": "mr",
106
+ "6": "arz",
107
+ "60": "ms",
108
+ "61": "my",
109
+ "62": "nds",
110
+ "63": "ne",
111
+ "64": "nl",
112
+ "65": "nn",
113
+ "66": "no",
114
+ "67": "oc",
115
+ "68": "pl",
116
+ "69": "pt",
117
+ "7": "ast",
118
+ "70": "ro",
119
+ "71": "ru",
120
+ "72": "scn",
121
+ "73": "sco",
122
+ "74": "sh",
123
+ "75": "si",
124
+ "76": "simple",
125
+ "77": "sk",
126
+ "78": "sl",
127
+ "79": "sq",
128
+ "8": "az",
129
+ "80": "sr",
130
+ "81": "sv",
131
+ "82": "sw",
132
+ "83": "ta",
133
+ "84": "te",
134
+ "85": "th",
135
+ "86": "tl",
136
+ "87": "tr",
137
+ "88": "tt",
138
+ "89": "uk",
139
+ "9": "bar",
140
+ "90": "ur",
141
+ "91": "uz",
142
+ "92": "vi",
143
+ "93": "war",
144
+ "94": "wuu",
145
+ "95": "yi",
146
+ "96": "zh",
147
+ "97": "zh_classical",
148
+ "98": "zh_min_nan",
149
+ "99": "zh_yue"
150
+ },
151
+ "init_std": 0.02,
152
+ "is_encoder": true,
153
+ "is_master": true,
154
+ "is_slurm_job": true,
155
+ "label2id": {
156
+ "idiom": 1,
157
+ "literal": 0,
158
+ "metaphor": 2,
159
+ "simile": 3
160
+ },
161
+ "lambda_ae": 1.0,
162
+ "lambda_ae_config": null,
163
+ "lambda_bt": 1.0,
164
+ "lambda_bt_config": null,
165
+ "lambda_clm": 1.0,
166
+ "lambda_clm_config": null,
167
+ "lambda_mlm": 1.0,
168
+ "lambda_mlm_config": null,
169
+ "lambda_mt": 1.0,
170
+ "lambda_mt_config": null,
171
+ "lambda_pc": 1.0,
172
+ "lambda_pc_config": null,
173
+ "lang2id": {
174
+ "af": 0,
175
+ "als": 1,
176
+ "am": 2,
177
+ "an": 3,
178
+ "ang": 4,
179
+ "ar": 5,
180
+ "arz": 6,
181
+ "ast": 7,
182
+ "az": 8,
183
+ "bar": 9,
184
+ "be": 10,
185
+ "bg": 11,
186
+ "bn": 12,
187
+ "br": 13,
188
+ "bs": 14,
189
+ "ca": 15,
190
+ "ceb": 16,
191
+ "ckb": 17,
192
+ "cs": 18,
193
+ "cy": 19,
194
+ "da": 20,
195
+ "de": 21,
196
+ "el": 22,
197
+ "en": 23,
198
+ "eo": 24,
199
+ "es": 25,
200
+ "et": 26,
201
+ "eu": 27,
202
+ "fa": 28,
203
+ "fi": 29,
204
+ "fr": 30,
205
+ "fy": 31,
206
+ "ga": 32,
207
+ "gan": 33,
208
+ "gl": 34,
209
+ "gu": 35,
210
+ "he": 36,
211
+ "hi": 37,
212
+ "hr": 38,
213
+ "hu": 39,
214
+ "hy": 40,
215
+ "ia": 41,
216
+ "id": 42,
217
+ "is": 43,
218
+ "it": 44,
219
+ "ja": 45,
220
+ "jv": 46,
221
+ "ka": 47,
222
+ "kk": 48,
223
+ "kn": 49,
224
+ "ko": 50,
225
+ "ku": 51,
226
+ "la": 52,
227
+ "lb": 53,
228
+ "lt": 54,
229
+ "lv": 55,
230
+ "mk": 56,
231
+ "ml": 57,
232
+ "mn": 58,
233
+ "mr": 59,
234
+ "ms": 60,
235
+ "my": 61,
236
+ "nds": 62,
237
+ "ne": 63,
238
+ "nl": 64,
239
+ "nn": 65,
240
+ "no": 66,
241
+ "oc": 67,
242
+ "pl": 68,
243
+ "pt": 69,
244
+ "ro": 70,
245
+ "ru": 71,
246
+ "scn": 72,
247
+ "sco": 73,
248
+ "sh": 74,
249
+ "si": 75,
250
+ "simple": 76,
251
+ "sk": 77,
252
+ "sl": 78,
253
+ "sq": 79,
254
+ "sr": 80,
255
+ "sv": 81,
256
+ "sw": 82,
257
+ "ta": 83,
258
+ "te": 84,
259
+ "th": 85,
260
+ "tl": 86,
261
+ "tr": 87,
262
+ "tt": 88,
263
+ "uk": 89,
264
+ "ur": 90,
265
+ "uz": 91,
266
+ "vi": 92,
267
+ "war": 93,
268
+ "wuu": 94,
269
+ "yi": 95,
270
+ "zh": 96,
271
+ "zh_classical": 97,
272
+ "zh_min_nan": 98,
273
+ "zh_yue": 99
274
+ },
275
+ "lang_id": 0,
276
+ "langs": [
277
+ "en",
278
+ "es",
279
+ "fr",
280
+ "de",
281
+ "zh",
282
+ "ru",
283
+ "pt",
284
+ "it",
285
+ "ar",
286
+ "ja",
287
+ "id",
288
+ "tr",
289
+ "nl",
290
+ "pl",
291
+ "simple",
292
+ "fa",
293
+ "vi",
294
+ "sv",
295
+ "ko",
296
+ "he",
297
+ "ro",
298
+ "no",
299
+ "hi",
300
+ "uk",
301
+ "cs",
302
+ "fi",
303
+ "hu",
304
+ "th",
305
+ "da",
306
+ "ca",
307
+ "el",
308
+ "bg",
309
+ "sr",
310
+ "ms",
311
+ "bn",
312
+ "hr",
313
+ "sl",
314
+ "zh_yue",
315
+ "az",
316
+ "sk",
317
+ "eo",
318
+ "ta",
319
+ "sh",
320
+ "lt",
321
+ "et",
322
+ "ml",
323
+ "la",
324
+ "bs",
325
+ "sq",
326
+ "arz",
327
+ "af",
328
+ "ka",
329
+ "mr",
330
+ "eu",
331
+ "tl",
332
+ "ang",
333
+ "gl",
334
+ "nn",
335
+ "ur",
336
+ "kk",
337
+ "be",
338
+ "hy",
339
+ "te",
340
+ "lv",
341
+ "mk",
342
+ "zh_classical",
343
+ "als",
344
+ "is",
345
+ "wuu",
346
+ "my",
347
+ "sco",
348
+ "mn",
349
+ "ceb",
350
+ "ast",
351
+ "cy",
352
+ "kn",
353
+ "br",
354
+ "an",
355
+ "gu",
356
+ "bar",
357
+ "uz",
358
+ "lb",
359
+ "ne",
360
+ "si",
361
+ "war",
362
+ "jv",
363
+ "ga",
364
+ "zh_min_nan",
365
+ "oc",
366
+ "ku",
367
+ "sw",
368
+ "nds",
369
+ "ckb",
370
+ "ia",
371
+ "yi",
372
+ "fy",
373
+ "scn",
374
+ "gan",
375
+ "tt",
376
+ "am"
377
+ ],
378
+ "layer_norm_eps": 1e-12,
379
+ "lg_sampling_factor": 0.7,
380
+ "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",
381
+ "local_rank": 0,
382
+ "mask_index": 5,
383
+ "mask_token_id": 0,
384
+ "master_addr": "learnfair0332",
385
+ "master_port": 11363,
386
+ "max_batch_size": 0,
387
+ "max_epoch": 100000,
388
+ "max_len": 200,
389
+ "max_position_embeddings": 512,
390
+ "max_vocab": 200000,
391
+ "min_count": 0,
392
+ "mlm_steps": [
393
+ [
394
+ "en",
395
+ null
396
+ ],
397
+ [
398
+ "es",
399
+ null
400
+ ],
401
+ [
402
+ "fr",
403
+ null
404
+ ],
405
+ [
406
+ "de",
407
+ null
408
+ ],
409
+ [
410
+ "zh",
411
+ null
412
+ ],
413
+ [
414
+ "ru",
415
+ null
416
+ ],
417
+ [
418
+ "pt",
419
+ null
420
+ ],
421
+ [
422
+ "it",
423
+ null
424
+ ],
425
+ [
426
+ "ar",
427
+ null
428
+ ],
429
+ [
430
+ "ja",
431
+ null
432
+ ],
433
+ [
434
+ "id",
435
+ null
436
+ ],
437
+ [
438
+ "tr",
439
+ null
440
+ ],
441
+ [
442
+ "nl",
443
+ null
444
+ ],
445
+ [
446
+ "pl",
447
+ null
448
+ ],
449
+ [
450
+ "simple",
451
+ null
452
+ ],
453
+ [
454
+ "fa",
455
+ null
456
+ ],
457
+ [
458
+ "vi",
459
+ null
460
+ ],
461
+ [
462
+ "sv",
463
+ null
464
+ ],
465
+ [
466
+ "ko",
467
+ null
468
+ ],
469
+ [
470
+ "he",
471
+ null
472
+ ],
473
+ [
474
+ "ro",
475
+ null
476
+ ],
477
+ [
478
+ "no",
479
+ null
480
+ ],
481
+ [
482
+ "hi",
483
+ null
484
+ ],
485
+ [
486
+ "uk",
487
+ null
488
+ ],
489
+ [
490
+ "cs",
491
+ null
492
+ ],
493
+ [
494
+ "fi",
495
+ null
496
+ ],
497
+ [
498
+ "hu",
499
+ null
500
+ ],
501
+ [
502
+ "th",
503
+ null
504
+ ],
505
+ [
506
+ "da",
507
+ null
508
+ ],
509
+ [
510
+ "ca",
511
+ null
512
+ ],
513
+ [
514
+ "el",
515
+ null
516
+ ],
517
+ [
518
+ "bg",
519
+ null
520
+ ],
521
+ [
522
+ "sr",
523
+ null
524
+ ],
525
+ [
526
+ "ms",
527
+ null
528
+ ],
529
+ [
530
+ "bn",
531
+ null
532
+ ],
533
+ [
534
+ "hr",
535
+ null
536
+ ],
537
+ [
538
+ "sl",
539
+ null
540
+ ],
541
+ [
542
+ "zh_yue",
543
+ null
544
+ ],
545
+ [
546
+ "az",
547
+ null
548
+ ],
549
+ [
550
+ "sk",
551
+ null
552
+ ],
553
+ [
554
+ "eo",
555
+ null
556
+ ],
557
+ [
558
+ "ta",
559
+ null
560
+ ],
561
+ [
562
+ "sh",
563
+ null
564
+ ],
565
+ [
566
+ "lt",
567
+ null
568
+ ],
569
+ [
570
+ "et",
571
+ null
572
+ ],
573
+ [
574
+ "ml",
575
+ null
576
+ ],
577
+ [
578
+ "la",
579
+ null
580
+ ],
581
+ [
582
+ "bs",
583
+ null
584
+ ],
585
+ [
586
+ "sq",
587
+ null
588
+ ],
589
+ [
590
+ "arz",
591
+ null
592
+ ],
593
+ [
594
+ "af",
595
+ null
596
+ ],
597
+ [
598
+ "ka",
599
+ null
600
+ ],
601
+ [
602
+ "mr",
603
+ null
604
+ ],
605
+ [
606
+ "eu",
607
+ null
608
+ ],
609
+ [
610
+ "tl",
611
+ null
612
+ ],
613
+ [
614
+ "ang",
615
+ null
616
+ ],
617
+ [
618
+ "gl",
619
+ null
620
+ ],
621
+ [
622
+ "nn",
623
+ null
624
+ ],
625
+ [
626
+ "ur",
627
+ null
628
+ ],
629
+ [
630
+ "kk",
631
+ null
632
+ ],
633
+ [
634
+ "be",
635
+ null
636
+ ],
637
+ [
638
+ "hy",
639
+ null
640
+ ],
641
+ [
642
+ "te",
643
+ null
644
+ ],
645
+ [
646
+ "lv",
647
+ null
648
+ ],
649
+ [
650
+ "mk",
651
+ null
652
+ ],
653
+ [
654
+ "zh_classical",
655
+ null
656
+ ],
657
+ [
658
+ "als",
659
+ null
660
+ ],
661
+ [
662
+ "is",
663
+ null
664
+ ],
665
+ [
666
+ "wuu",
667
+ null
668
+ ],
669
+ [
670
+ "my",
671
+ null
672
+ ],
673
+ [
674
+ "sco",
675
+ null
676
+ ],
677
+ [
678
+ "mn",
679
+ null
680
+ ],
681
+ [
682
+ "ceb",
683
+ null
684
+ ],
685
+ [
686
+ "ast",
687
+ null
688
+ ],
689
+ [
690
+ "cy",
691
+ null
692
+ ],
693
+ [
694
+ "kn",
695
+ null
696
+ ],
697
+ [
698
+ "br",
699
+ null
700
+ ],
701
+ [
702
+ "an",
703
+ null
704
+ ],
705
+ [
706
+ "gu",
707
+ null
708
+ ],
709
+ [
710
+ "bar",
711
+ null
712
+ ],
713
+ [
714
+ "uz",
715
+ null
716
+ ],
717
+ [
718
+ "lb",
719
+ null
720
+ ],
721
+ [
722
+ "ne",
723
+ null
724
+ ],
725
+ [
726
+ "si",
727
+ null
728
+ ],
729
+ [
730
+ "war",
731
+ null
732
+ ],
733
+ [
734
+ "jv",
735
+ null
736
+ ],
737
+ [
738
+ "ga",
739
+ null
740
+ ],
741
+ [
742
+ "zh_min_nan",
743
+ null
744
+ ],
745
+ [
746
+ "oc",
747
+ null
748
+ ],
749
+ [
750
+ "ku",
751
+ null
752
+ ],
753
+ [
754
+ "sw",
755
+ null
756
+ ],
757
+ [
758
+ "nds",
759
+ null
760
+ ],
761
+ [
762
+ "ckb",
763
+ null
764
+ ],
765
+ [
766
+ "ia",
767
+ null
768
+ ],
769
+ [
770
+ "yi",
771
+ null
772
+ ],
773
+ [
774
+ "fy",
775
+ null
776
+ ],
777
+ [
778
+ "scn",
779
+ null
780
+ ],
781
+ [
782
+ "gan",
783
+ null
784
+ ],
785
+ [
786
+ "tt",
787
+ null
788
+ ],
789
+ [
790
+ "am",
791
+ null
792
+ ]
793
+ ],
794
+ "model_type": "xlm",
795
+ "mono_dataset": {
796
+ "af": {
797
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.af.pth",
798
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.af.pth",
799
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.af.pth"
800
+ },
801
+ "als": {
802
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.als.pth",
803
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.als.pth",
804
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.als.pth"
805
+ },
806
+ "am": {
807
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.am.pth",
808
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.am.pth",
809
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.am.pth"
810
+ },
811
+ "an": {
812
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.an.pth",
813
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.an.pth",
814
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.an.pth"
815
+ },
816
+ "ang": {
817
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ang.pth",
818
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ang.pth",
819
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ang.pth"
820
+ },
821
+ "ar": {
822
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ar.pth",
823
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ar.pth",
824
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ar.pth"
825
+ },
826
+ "arz": {
827
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.arz.pth",
828
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.arz.pth",
829
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.arz.pth"
830
+ },
831
+ "ast": {
832
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ast.pth",
833
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ast.pth",
834
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ast.pth"
835
+ },
836
+ "az": {
837
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.az.pth",
838
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.az.pth",
839
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.az.pth"
840
+ },
841
+ "bar": {
842
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bar.pth",
843
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bar.pth",
844
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bar.pth"
845
+ },
846
+ "be": {
847
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.be.pth",
848
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.be.pth",
849
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.be.pth"
850
+ },
851
+ "bg": {
852
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bg.pth",
853
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bg.pth",
854
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bg.pth"
855
+ },
856
+ "bn": {
857
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bn.pth",
858
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bn.pth",
859
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bn.pth"
860
+ },
861
+ "br": {
862
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.br.pth",
863
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.br.pth",
864
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.br.pth"
865
+ },
866
+ "bs": {
867
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.bs.pth",
868
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.bs.pth",
869
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.bs.pth"
870
+ },
871
+ "ca": {
872
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ca.pth",
873
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ca.pth",
874
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ca.pth"
875
+ },
876
+ "ceb": {
877
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ceb.pth",
878
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ceb.pth",
879
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ceb.pth"
880
+ },
881
+ "ckb": {
882
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ckb.pth",
883
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ckb.pth",
884
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ckb.pth"
885
+ },
886
+ "cs": {
887
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.cs.pth",
888
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.cs.pth",
889
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.cs.pth"
890
+ },
891
+ "cy": {
892
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.cy.pth",
893
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.cy.pth",
894
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.cy.pth"
895
+ },
896
+ "da": {
897
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.da.pth",
898
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.da.pth",
899
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.da.pth"
900
+ },
901
+ "de": {
902
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.de.pth",
903
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.de.pth",
904
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.de.pth"
905
+ },
906
+ "el": {
907
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.el.pth",
908
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.el.pth",
909
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.el.pth"
910
+ },
911
+ "en": {
912
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.en.pth",
913
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.en.pth",
914
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.en.pth"
915
+ },
916
+ "eo": {
917
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.eo.pth",
918
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.eo.pth",
919
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.eo.pth"
920
+ },
921
+ "es": {
922
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.es.pth",
923
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.es.pth",
924
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.es.pth"
925
+ },
926
+ "et": {
927
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.et.pth",
928
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.et.pth",
929
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.et.pth"
930
+ },
931
+ "eu": {
932
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.eu.pth",
933
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.eu.pth",
934
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.eu.pth"
935
+ },
936
+ "fa": {
937
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fa.pth",
938
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fa.pth",
939
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fa.pth"
940
+ },
941
+ "fi": {
942
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fi.pth",
943
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fi.pth",
944
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fi.pth"
945
+ },
946
+ "fr": {
947
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fr.pth",
948
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fr.pth",
949
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fr.pth"
950
+ },
951
+ "fy": {
952
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.fy.pth",
953
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.fy.pth",
954
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.fy.pth"
955
+ },
956
+ "ga": {
957
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ga.pth",
958
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ga.pth",
959
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ga.pth"
960
+ },
961
+ "gan": {
962
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gan.pth",
963
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gan.pth",
964
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gan.pth"
965
+ },
966
+ "gl": {
967
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gl.pth",
968
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gl.pth",
969
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gl.pth"
970
+ },
971
+ "gu": {
972
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.gu.pth",
973
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.gu.pth",
974
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.gu.pth"
975
+ },
976
+ "he": {
977
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.he.pth",
978
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.he.pth",
979
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.he.pth"
980
+ },
981
+ "hi": {
982
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hi.pth",
983
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hi.pth",
984
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hi.pth"
985
+ },
986
+ "hr": {
987
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hr.pth",
988
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hr.pth",
989
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hr.pth"
990
+ },
991
+ "hu": {
992
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hu.pth",
993
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hu.pth",
994
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hu.pth"
995
+ },
996
+ "hy": {
997
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.hy.pth",
998
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.hy.pth",
999
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.hy.pth"
1000
+ },
1001
+ "ia": {
1002
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ia.pth",
1003
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ia.pth",
1004
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ia.pth"
1005
+ },
1006
+ "id": {
1007
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.id.pth",
1008
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.id.pth",
1009
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.id.pth"
1010
+ },
1011
+ "is": {
1012
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.is.pth",
1013
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.is.pth",
1014
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.is.pth"
1015
+ },
1016
+ "it": {
1017
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.it.pth",
1018
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.it.pth",
1019
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.it.pth"
1020
+ },
1021
+ "ja": {
1022
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ja.pth",
1023
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ja.pth",
1024
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ja.pth"
1025
+ },
1026
+ "jv": {
1027
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.jv.pth",
1028
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.jv.pth",
1029
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.jv.pth"
1030
+ },
1031
+ "ka": {
1032
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ka.pth",
1033
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ka.pth",
1034
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ka.pth"
1035
+ },
1036
+ "kk": {
1037
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.kk.pth",
1038
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.kk.pth",
1039
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.kk.pth"
1040
+ },
1041
+ "kn": {
1042
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.kn.pth",
1043
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.kn.pth",
1044
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.kn.pth"
1045
+ },
1046
+ "ko": {
1047
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ko.pth",
1048
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ko.pth",
1049
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ko.pth"
1050
+ },
1051
+ "ku": {
1052
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ku.pth",
1053
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ku.pth",
1054
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ku.pth"
1055
+ },
1056
+ "la": {
1057
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.la.pth",
1058
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.la.pth",
1059
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.la.pth"
1060
+ },
1061
+ "lb": {
1062
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lb.pth",
1063
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lb.pth",
1064
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lb.pth"
1065
+ },
1066
+ "lt": {
1067
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lt.pth",
1068
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lt.pth",
1069
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lt.pth"
1070
+ },
1071
+ "lv": {
1072
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.lv.pth",
1073
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.lv.pth",
1074
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.lv.pth"
1075
+ },
1076
+ "mk": {
1077
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mk.pth",
1078
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mk.pth",
1079
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mk.pth"
1080
+ },
1081
+ "ml": {
1082
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ml.pth",
1083
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ml.pth",
1084
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ml.pth"
1085
+ },
1086
+ "mn": {
1087
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mn.pth",
1088
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mn.pth",
1089
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mn.pth"
1090
+ },
1091
+ "mr": {
1092
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.mr.pth",
1093
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.mr.pth",
1094
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.mr.pth"
1095
+ },
1096
+ "ms": {
1097
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ms.pth",
1098
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ms.pth",
1099
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ms.pth"
1100
+ },
1101
+ "my": {
1102
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.my.pth",
1103
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.my.pth",
1104
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.my.pth"
1105
+ },
1106
+ "nds": {
1107
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nds.pth",
1108
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nds.pth",
1109
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nds.pth"
1110
+ },
1111
+ "ne": {
1112
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ne.pth",
1113
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ne.pth",
1114
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ne.pth"
1115
+ },
1116
+ "nl": {
1117
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nl.pth",
1118
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nl.pth",
1119
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nl.pth"
1120
+ },
1121
+ "nn": {
1122
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.nn.pth",
1123
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.nn.pth",
1124
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.nn.pth"
1125
+ },
1126
+ "no": {
1127
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.no.pth",
1128
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.no.pth",
1129
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.no.pth"
1130
+ },
1131
+ "oc": {
1132
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.oc.pth",
1133
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.oc.pth",
1134
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.oc.pth"
1135
+ },
1136
+ "pl": {
1137
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.pl.pth",
1138
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.pl.pth",
1139
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.pl.pth"
1140
+ },
1141
+ "pt": {
1142
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.pt.pth",
1143
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.pt.pth",
1144
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.pt.pth"
1145
+ },
1146
+ "ro": {
1147
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ro.pth",
1148
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ro.pth",
1149
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ro.pth"
1150
+ },
1151
+ "ru": {
1152
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ru.pth",
1153
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ru.pth",
1154
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ru.pth"
1155
+ },
1156
+ "scn": {
1157
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.scn.pth",
1158
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.scn.pth",
1159
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.scn.pth"
1160
+ },
1161
+ "sco": {
1162
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sco.pth",
1163
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sco.pth",
1164
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sco.pth"
1165
+ },
1166
+ "sh": {
1167
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sh.pth",
1168
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sh.pth",
1169
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sh.pth"
1170
+ },
1171
+ "si": {
1172
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.si.pth",
1173
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.si.pth",
1174
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.si.pth"
1175
+ },
1176
+ "simple": {
1177
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.simple.pth",
1178
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.simple.pth",
1179
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.simple.pth"
1180
+ },
1181
+ "sk": {
1182
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sk.pth",
1183
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sk.pth",
1184
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sk.pth"
1185
+ },
1186
+ "sl": {
1187
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sl.pth",
1188
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sl.pth",
1189
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sl.pth"
1190
+ },
1191
+ "sq": {
1192
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sq.pth",
1193
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sq.pth",
1194
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sq.pth"
1195
+ },
1196
+ "sr": {
1197
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sr.pth",
1198
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sr.pth",
1199
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sr.pth"
1200
+ },
1201
+ "sv": {
1202
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sv.pth",
1203
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sv.pth",
1204
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sv.pth"
1205
+ },
1206
+ "sw": {
1207
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.sw.pth",
1208
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.sw.pth",
1209
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.sw.pth"
1210
+ },
1211
+ "ta": {
1212
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ta.pth",
1213
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ta.pth",
1214
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ta.pth"
1215
+ },
1216
+ "te": {
1217
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.te.pth",
1218
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.te.pth",
1219
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.te.pth"
1220
+ },
1221
+ "th": {
1222
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.th.pth",
1223
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.th.pth",
1224
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.th.pth"
1225
+ },
1226
+ "tl": {
1227
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tl.pth",
1228
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tl.pth",
1229
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tl.pth"
1230
+ },
1231
+ "tr": {
1232
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tr.pth",
1233
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tr.pth",
1234
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tr.pth"
1235
+ },
1236
+ "tt": {
1237
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.tt.pth",
1238
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.tt.pth",
1239
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.tt.pth"
1240
+ },
1241
+ "uk": {
1242
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.uk.pth",
1243
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.uk.pth",
1244
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.uk.pth"
1245
+ },
1246
+ "ur": {
1247
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.ur.pth",
1248
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.ur.pth",
1249
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.ur.pth"
1250
+ },
1251
+ "uz": {
1252
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.uz.pth",
1253
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.uz.pth",
1254
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.uz.pth"
1255
+ },
1256
+ "vi": {
1257
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.vi.pth",
1258
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.vi.pth",
1259
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.vi.pth"
1260
+ },
1261
+ "war": {
1262
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.war.pth",
1263
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.war.pth",
1264
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.war.pth"
1265
+ },
1266
+ "wuu": {
1267
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.wuu.pth",
1268
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.wuu.pth",
1269
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.wuu.pth"
1270
+ },
1271
+ "yi": {
1272
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.yi.pth",
1273
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.yi.pth",
1274
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.yi.pth"
1275
+ },
1276
+ "zh": {
1277
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh.pth",
1278
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh.pth",
1279
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh.pth"
1280
+ },
1281
+ "zh_classical": {
1282
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_classical.pth",
1283
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_classical.pth",
1284
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_classical.pth"
1285
+ },
1286
+ "zh_min_nan": {
1287
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_min_nan.pth",
1288
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_min_nan.pth",
1289
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_min_nan.pth"
1290
+ },
1291
+ "zh_yue": {
1292
+ "test": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/test.zh_yue.pth",
1293
+ "train": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/train.zh_yue.pth",
1294
+ "valid": "/private/home/aconneau/projects/XLM/data/wiki/100/175k/valid.zh_yue.pth"
1295
+ }
1296
+ },
1297
+ "mono_list": [
1298
+ "en",
1299
+ "es",
1300
+ "fr",
1301
+ "de",
1302
+ "zh",
1303
+ "ru",
1304
+ "pt",
1305
+ "it",
1306
+ "ar",
1307
+ "ja",
1308
+ "id",
1309
+ "tr",
1310
+ "nl",
1311
+ "pl",
1312
+ "simple",
1313
+ "fa",
1314
+ "vi",
1315
+ "sv",
1316
+ "ko",
1317
+ "he",
1318
+ "ro",
1319
+ "no",
1320
+ "hi",
1321
+ "uk",
1322
+ "cs",
1323
+ "fi",
1324
+ "hu",
1325
+ "th",
1326
+ "da",
1327
+ "ca",
1328
+ "el",
1329
+ "bg",
1330
+ "sr",
1331
+ "ms",
1332
+ "bn",
1333
+ "hr",
1334
+ "sl",
1335
+ "zh_yue",
1336
+ "az",
1337
+ "sk",
1338
+ "eo",
1339
+ "ta",
1340
+ "sh",
1341
+ "lt",
1342
+ "et",
1343
+ "ml",
1344
+ "la",
1345
+ "bs",
1346
+ "sq",
1347
+ "arz",
1348
+ "af",
1349
+ "ka",
1350
+ "mr",
1351
+ "eu",
1352
+ "tl",
1353
+ "ang",
1354
+ "gl",
1355
+ "nn",
1356
+ "ur",
1357
+ "kk",
1358
+ "be",
1359
+ "hy",
1360
+ "te",
1361
+ "lv",
1362
+ "mk",
1363
+ "zh_classical",
1364
+ "als",
1365
+ "is",
1366
+ "wuu",
1367
+ "my",
1368
+ "sco",
1369
+ "mn",
1370
+ "ceb",
1371
+ "ast",
1372
+ "cy",
1373
+ "kn",
1374
+ "br",
1375
+ "an",
1376
+ "gu",
1377
+ "bar",
1378
+ "uz",
1379
+ "lb",
1380
+ "ne",
1381
+ "si",
1382
+ "war",
1383
+ "jv",
1384
+ "ga",
1385
+ "zh_min_nan",
1386
+ "oc",
1387
+ "ku",
1388
+ "sw",
1389
+ "nds",
1390
+ "ckb",
1391
+ "ia",
1392
+ "yi",
1393
+ "fy",
1394
+ "scn",
1395
+ "gan",
1396
+ "tt",
1397
+ "am"
1398
+ ],
1399
+ "mt_steps": [],
1400
+ "multi_gpu": true,
1401
+ "multi_node": true,
1402
+ "n_gpu_per_node": 8,
1403
+ "n_heads": 16,
1404
+ "n_langs": 100,
1405
+ "n_layers": 16,
1406
+ "n_nodes": 4,
1407
+ "node_id": 0,
1408
+ "optimizer": "adam_inverse_sqrt,lr=0.00005,warmup_updates=30000,beta1=0.9,beta2=0.999,weight_decay=0.01,eps=0.000001",
1409
+ "pad_token_id": 2,
1410
+ "para_dataset": {},
1411
+ "para_list": [],
1412
+ "pc_steps": [],
1413
+ "ref_paths": {},
1414
+ "reload_checkpoint": "",
1415
+ "reload_emb": "",
1416
+ "reload_model": "/checkpoint/aconneau/dumped/xlm_17_100_240_big_model_upper.2/14884511/best-valid_zh_mlm_ppl.pth",
1417
+ "sample_alpha": 0.5,
1418
+ "save_periodic": 0,
1419
+ "share_inout_emb": true,
1420
+ "sinusoidal_embeddings": false,
1421
+ "split_data": true,
1422
+ "start_n_top": 5,
1423
+ "stopping_criterion": "_valid_zh_mlm_ppl,25",
1424
+ "summary_activation": null,
1425
+ "summary_first_dropout": 0.1,
1426
+ "summary_proj_to_labels": true,
1427
+ "summary_type": "first",
1428
+ "summary_use_proj": true,
1429
+ "tie_word_embeddings": true,
1430
+ "tokens_per_batch": -1,
1431
+ "transformers_version": "5.12.1",
1432
+ "unk_index": 3,
1433
+ "use_cache": false,
1434
+ "use_lang_emb": false,
1435
+ "use_memory": false,
1436
+ "validation_metrics": "_valid_en_mlm_ppl,_valid_mlm_ppl,_valid_zh_mlm_ppl",
1437
+ "vocab_size": 200000,
1438
+ "word_blank": 0.0,
1439
+ "word_dropout": 0.0,
1440
+ "word_keep": 0.1,
1441
+ "word_mask": 0.8,
1442
+ "word_mask_keep_rand": "0.8,0.1,0.1",
1443
+ "word_pred": 0.15,
1444
+ "word_rand": 0.1,
1445
+ "word_shuffle": 0.0,
1446
+ "world_size": 32
1447
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf02856cf1a51ae36a82da256551f68cb3f91cdc94424ec35008ff848c442061
3
+ size 2286036432
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cb6687ef0ba4df70d28717bc27379074a13579ddff21d475f69549167e87b7c
3
+ size 5265