upload tokenizer
Browse files- cvlm_llama2_tokenizer_100img_and_224loc_addpatch/added_tokens.json +332 -0
- cvlm_llama2_tokenizer_100img_and_224loc_addpatch/special_tokens_map.json +24 -0
- cvlm_llama2_tokenizer_100img_and_224loc_addpatch/tokenizer.model +3 -0
- cvlm_llama2_tokenizer_100img_and_224loc_addpatch/tokenizer_config.json +34 -0
cvlm_llama2_tokenizer_100img_and_224loc_addpatch/added_tokens.json
ADDED
@@ -0,0 +1,332 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"</img>": 32101,
|
3 |
+
"</patch>": 32103,
|
4 |
+
"<box_end>": 32329,
|
5 |
+
"<box_start>": 32328,
|
6 |
+
"<img>": 32100,
|
7 |
+
"<img_00000>": 32000,
|
8 |
+
"<img_00001>": 32001,
|
9 |
+
"<img_00002>": 32002,
|
10 |
+
"<img_00003>": 32003,
|
11 |
+
"<img_00004>": 32004,
|
12 |
+
"<img_00005>": 32005,
|
13 |
+
"<img_00006>": 32006,
|
14 |
+
"<img_00007>": 32007,
|
15 |
+
"<img_00008>": 32008,
|
16 |
+
"<img_00009>": 32009,
|
17 |
+
"<img_00010>": 32010,
|
18 |
+
"<img_00011>": 32011,
|
19 |
+
"<img_00012>": 32012,
|
20 |
+
"<img_00013>": 32013,
|
21 |
+
"<img_00014>": 32014,
|
22 |
+
"<img_00015>": 32015,
|
23 |
+
"<img_00016>": 32016,
|
24 |
+
"<img_00017>": 32017,
|
25 |
+
"<img_00018>": 32018,
|
26 |
+
"<img_00019>": 32019,
|
27 |
+
"<img_00020>": 32020,
|
28 |
+
"<img_00021>": 32021,
|
29 |
+
"<img_00022>": 32022,
|
30 |
+
"<img_00023>": 32023,
|
31 |
+
"<img_00024>": 32024,
|
32 |
+
"<img_00025>": 32025,
|
33 |
+
"<img_00026>": 32026,
|
34 |
+
"<img_00027>": 32027,
|
35 |
+
"<img_00028>": 32028,
|
36 |
+
"<img_00029>": 32029,
|
37 |
+
"<img_00030>": 32030,
|
38 |
+
"<img_00031>": 32031,
|
39 |
+
"<img_00032>": 32032,
|
40 |
+
"<img_00033>": 32033,
|
41 |
+
"<img_00034>": 32034,
|
42 |
+
"<img_00035>": 32035,
|
43 |
+
"<img_00036>": 32036,
|
44 |
+
"<img_00037>": 32037,
|
45 |
+
"<img_00038>": 32038,
|
46 |
+
"<img_00039>": 32039,
|
47 |
+
"<img_00040>": 32040,
|
48 |
+
"<img_00041>": 32041,
|
49 |
+
"<img_00042>": 32042,
|
50 |
+
"<img_00043>": 32043,
|
51 |
+
"<img_00044>": 32044,
|
52 |
+
"<img_00045>": 32045,
|
53 |
+
"<img_00046>": 32046,
|
54 |
+
"<img_00047>": 32047,
|
55 |
+
"<img_00048>": 32048,
|
56 |
+
"<img_00049>": 32049,
|
57 |
+
"<img_00050>": 32050,
|
58 |
+
"<img_00051>": 32051,
|
59 |
+
"<img_00052>": 32052,
|
60 |
+
"<img_00053>": 32053,
|
61 |
+
"<img_00054>": 32054,
|
62 |
+
"<img_00055>": 32055,
|
63 |
+
"<img_00056>": 32056,
|
64 |
+
"<img_00057>": 32057,
|
65 |
+
"<img_00058>": 32058,
|
66 |
+
"<img_00059>": 32059,
|
67 |
+
"<img_00060>": 32060,
|
68 |
+
"<img_00061>": 32061,
|
69 |
+
"<img_00062>": 32062,
|
70 |
+
"<img_00063>": 32063,
|
71 |
+
"<img_00064>": 32064,
|
72 |
+
"<img_00065>": 32065,
|
73 |
+
"<img_00066>": 32066,
|
74 |
+
"<img_00067>": 32067,
|
75 |
+
"<img_00068>": 32068,
|
76 |
+
"<img_00069>": 32069,
|
77 |
+
"<img_00070>": 32070,
|
78 |
+
"<img_00071>": 32071,
|
79 |
+
"<img_00072>": 32072,
|
80 |
+
"<img_00073>": 32073,
|
81 |
+
"<img_00074>": 32074,
|
82 |
+
"<img_00075>": 32075,
|
83 |
+
"<img_00076>": 32076,
|
84 |
+
"<img_00077>": 32077,
|
85 |
+
"<img_00078>": 32078,
|
86 |
+
"<img_00079>": 32079,
|
87 |
+
"<img_00080>": 32080,
|
88 |
+
"<img_00081>": 32081,
|
89 |
+
"<img_00082>": 32082,
|
90 |
+
"<img_00083>": 32083,
|
91 |
+
"<img_00084>": 32084,
|
92 |
+
"<img_00085>": 32085,
|
93 |
+
"<img_00086>": 32086,
|
94 |
+
"<img_00087>": 32087,
|
95 |
+
"<img_00088>": 32088,
|
96 |
+
"<img_00089>": 32089,
|
97 |
+
"<img_00090>": 32090,
|
98 |
+
"<img_00091>": 32091,
|
99 |
+
"<img_00092>": 32092,
|
100 |
+
"<img_00093>": 32093,
|
101 |
+
"<img_00094>": 32094,
|
102 |
+
"<img_00095>": 32095,
|
103 |
+
"<img_00096>": 32096,
|
104 |
+
"<img_00097>": 32097,
|
105 |
+
"<img_00098>": 32098,
|
106 |
+
"<img_00099>": 32099,
|
107 |
+
"<loc-0>": 32104,
|
108 |
+
"<loc-100>": 32204,
|
109 |
+
"<loc-101>": 32205,
|
110 |
+
"<loc-102>": 32206,
|
111 |
+
"<loc-103>": 32207,
|
112 |
+
"<loc-104>": 32208,
|
113 |
+
"<loc-105>": 32209,
|
114 |
+
"<loc-106>": 32210,
|
115 |
+
"<loc-107>": 32211,
|
116 |
+
"<loc-108>": 32212,
|
117 |
+
"<loc-109>": 32213,
|
118 |
+
"<loc-10>": 32114,
|
119 |
+
"<loc-110>": 32214,
|
120 |
+
"<loc-111>": 32215,
|
121 |
+
"<loc-112>": 32216,
|
122 |
+
"<loc-113>": 32217,
|
123 |
+
"<loc-114>": 32218,
|
124 |
+
"<loc-115>": 32219,
|
125 |
+
"<loc-116>": 32220,
|
126 |
+
"<loc-117>": 32221,
|
127 |
+
"<loc-118>": 32222,
|
128 |
+
"<loc-119>": 32223,
|
129 |
+
"<loc-11>": 32115,
|
130 |
+
"<loc-120>": 32224,
|
131 |
+
"<loc-121>": 32225,
|
132 |
+
"<loc-122>": 32226,
|
133 |
+
"<loc-123>": 32227,
|
134 |
+
"<loc-124>": 32228,
|
135 |
+
"<loc-125>": 32229,
|
136 |
+
"<loc-126>": 32230,
|
137 |
+
"<loc-127>": 32231,
|
138 |
+
"<loc-128>": 32232,
|
139 |
+
"<loc-129>": 32233,
|
140 |
+
"<loc-12>": 32116,
|
141 |
+
"<loc-130>": 32234,
|
142 |
+
"<loc-131>": 32235,
|
143 |
+
"<loc-132>": 32236,
|
144 |
+
"<loc-133>": 32237,
|
145 |
+
"<loc-134>": 32238,
|
146 |
+
"<loc-135>": 32239,
|
147 |
+
"<loc-136>": 32240,
|
148 |
+
"<loc-137>": 32241,
|
149 |
+
"<loc-138>": 32242,
|
150 |
+
"<loc-139>": 32243,
|
151 |
+
"<loc-13>": 32117,
|
152 |
+
"<loc-140>": 32244,
|
153 |
+
"<loc-141>": 32245,
|
154 |
+
"<loc-142>": 32246,
|
155 |
+
"<loc-143>": 32247,
|
156 |
+
"<loc-144>": 32248,
|
157 |
+
"<loc-145>": 32249,
|
158 |
+
"<loc-146>": 32250,
|
159 |
+
"<loc-147>": 32251,
|
160 |
+
"<loc-148>": 32252,
|
161 |
+
"<loc-149>": 32253,
|
162 |
+
"<loc-14>": 32118,
|
163 |
+
"<loc-150>": 32254,
|
164 |
+
"<loc-151>": 32255,
|
165 |
+
"<loc-152>": 32256,
|
166 |
+
"<loc-153>": 32257,
|
167 |
+
"<loc-154>": 32258,
|
168 |
+
"<loc-155>": 32259,
|
169 |
+
"<loc-156>": 32260,
|
170 |
+
"<loc-157>": 32261,
|
171 |
+
"<loc-158>": 32262,
|
172 |
+
"<loc-159>": 32263,
|
173 |
+
"<loc-15>": 32119,
|
174 |
+
"<loc-160>": 32264,
|
175 |
+
"<loc-161>": 32265,
|
176 |
+
"<loc-162>": 32266,
|
177 |
+
"<loc-163>": 32267,
|
178 |
+
"<loc-164>": 32268,
|
179 |
+
"<loc-165>": 32269,
|
180 |
+
"<loc-166>": 32270,
|
181 |
+
"<loc-167>": 32271,
|
182 |
+
"<loc-168>": 32272,
|
183 |
+
"<loc-169>": 32273,
|
184 |
+
"<loc-16>": 32120,
|
185 |
+
"<loc-170>": 32274,
|
186 |
+
"<loc-171>": 32275,
|
187 |
+
"<loc-172>": 32276,
|
188 |
+
"<loc-173>": 32277,
|
189 |
+
"<loc-174>": 32278,
|
190 |
+
"<loc-175>": 32279,
|
191 |
+
"<loc-176>": 32280,
|
192 |
+
"<loc-177>": 32281,
|
193 |
+
"<loc-178>": 32282,
|
194 |
+
"<loc-179>": 32283,
|
195 |
+
"<loc-17>": 32121,
|
196 |
+
"<loc-180>": 32284,
|
197 |
+
"<loc-181>": 32285,
|
198 |
+
"<loc-182>": 32286,
|
199 |
+
"<loc-183>": 32287,
|
200 |
+
"<loc-184>": 32288,
|
201 |
+
"<loc-185>": 32289,
|
202 |
+
"<loc-186>": 32290,
|
203 |
+
"<loc-187>": 32291,
|
204 |
+
"<loc-188>": 32292,
|
205 |
+
"<loc-189>": 32293,
|
206 |
+
"<loc-18>": 32122,
|
207 |
+
"<loc-190>": 32294,
|
208 |
+
"<loc-191>": 32295,
|
209 |
+
"<loc-192>": 32296,
|
210 |
+
"<loc-193>": 32297,
|
211 |
+
"<loc-194>": 32298,
|
212 |
+
"<loc-195>": 32299,
|
213 |
+
"<loc-196>": 32300,
|
214 |
+
"<loc-197>": 32301,
|
215 |
+
"<loc-198>": 32302,
|
216 |
+
"<loc-199>": 32303,
|
217 |
+
"<loc-19>": 32123,
|
218 |
+
"<loc-1>": 32105,
|
219 |
+
"<loc-200>": 32304,
|
220 |
+
"<loc-201>": 32305,
|
221 |
+
"<loc-202>": 32306,
|
222 |
+
"<loc-203>": 32307,
|
223 |
+
"<loc-204>": 32308,
|
224 |
+
"<loc-205>": 32309,
|
225 |
+
"<loc-206>": 32310,
|
226 |
+
"<loc-207>": 32311,
|
227 |
+
"<loc-208>": 32312,
|
228 |
+
"<loc-209>": 32313,
|
229 |
+
"<loc-20>": 32124,
|
230 |
+
"<loc-210>": 32314,
|
231 |
+
"<loc-211>": 32315,
|
232 |
+
"<loc-212>": 32316,
|
233 |
+
"<loc-213>": 32317,
|
234 |
+
"<loc-214>": 32318,
|
235 |
+
"<loc-215>": 32319,
|
236 |
+
"<loc-216>": 32320,
|
237 |
+
"<loc-217>": 32321,
|
238 |
+
"<loc-218>": 32322,
|
239 |
+
"<loc-219>": 32323,
|
240 |
+
"<loc-21>": 32125,
|
241 |
+
"<loc-220>": 32324,
|
242 |
+
"<loc-221>": 32325,
|
243 |
+
"<loc-222>": 32326,
|
244 |
+
"<loc-223>": 32327,
|
245 |
+
"<loc-22>": 32126,
|
246 |
+
"<loc-23>": 32127,
|
247 |
+
"<loc-24>": 32128,
|
248 |
+
"<loc-25>": 32129,
|
249 |
+
"<loc-26>": 32130,
|
250 |
+
"<loc-27>": 32131,
|
251 |
+
"<loc-28>": 32132,
|
252 |
+
"<loc-29>": 32133,
|
253 |
+
"<loc-2>": 32106,
|
254 |
+
"<loc-30>": 32134,
|
255 |
+
"<loc-31>": 32135,
|
256 |
+
"<loc-32>": 32136,
|
257 |
+
"<loc-33>": 32137,
|
258 |
+
"<loc-34>": 32138,
|
259 |
+
"<loc-35>": 32139,
|
260 |
+
"<loc-36>": 32140,
|
261 |
+
"<loc-37>": 32141,
|
262 |
+
"<loc-38>": 32142,
|
263 |
+
"<loc-39>": 32143,
|
264 |
+
"<loc-3>": 32107,
|
265 |
+
"<loc-40>": 32144,
|
266 |
+
"<loc-41>": 32145,
|
267 |
+
"<loc-42>": 32146,
|
268 |
+
"<loc-43>": 32147,
|
269 |
+
"<loc-44>": 32148,
|
270 |
+
"<loc-45>": 32149,
|
271 |
+
"<loc-46>": 32150,
|
272 |
+
"<loc-47>": 32151,
|
273 |
+
"<loc-48>": 32152,
|
274 |
+
"<loc-49>": 32153,
|
275 |
+
"<loc-4>": 32108,
|
276 |
+
"<loc-50>": 32154,
|
277 |
+
"<loc-51>": 32155,
|
278 |
+
"<loc-52>": 32156,
|
279 |
+
"<loc-53>": 32157,
|
280 |
+
"<loc-54>": 32158,
|
281 |
+
"<loc-55>": 32159,
|
282 |
+
"<loc-56>": 32160,
|
283 |
+
"<loc-57>": 32161,
|
284 |
+
"<loc-58>": 32162,
|
285 |
+
"<loc-59>": 32163,
|
286 |
+
"<loc-5>": 32109,
|
287 |
+
"<loc-60>": 32164,
|
288 |
+
"<loc-61>": 32165,
|
289 |
+
"<loc-62>": 32166,
|
290 |
+
"<loc-63>": 32167,
|
291 |
+
"<loc-64>": 32168,
|
292 |
+
"<loc-65>": 32169,
|
293 |
+
"<loc-66>": 32170,
|
294 |
+
"<loc-67>": 32171,
|
295 |
+
"<loc-68>": 32172,
|
296 |
+
"<loc-69>": 32173,
|
297 |
+
"<loc-6>": 32110,
|
298 |
+
"<loc-70>": 32174,
|
299 |
+
"<loc-71>": 32175,
|
300 |
+
"<loc-72>": 32176,
|
301 |
+
"<loc-73>": 32177,
|
302 |
+
"<loc-74>": 32178,
|
303 |
+
"<loc-75>": 32179,
|
304 |
+
"<loc-76>": 32180,
|
305 |
+
"<loc-77>": 32181,
|
306 |
+
"<loc-78>": 32182,
|
307 |
+
"<loc-79>": 32183,
|
308 |
+
"<loc-7>": 32111,
|
309 |
+
"<loc-80>": 32184,
|
310 |
+
"<loc-81>": 32185,
|
311 |
+
"<loc-82>": 32186,
|
312 |
+
"<loc-83>": 32187,
|
313 |
+
"<loc-84>": 32188,
|
314 |
+
"<loc-85>": 32189,
|
315 |
+
"<loc-86>": 32190,
|
316 |
+
"<loc-87>": 32191,
|
317 |
+
"<loc-88>": 32192,
|
318 |
+
"<loc-89>": 32193,
|
319 |
+
"<loc-8>": 32112,
|
320 |
+
"<loc-90>": 32194,
|
321 |
+
"<loc-91>": 32195,
|
322 |
+
"<loc-92>": 32196,
|
323 |
+
"<loc-93>": 32197,
|
324 |
+
"<loc-94>": 32198,
|
325 |
+
"<loc-95>": 32199,
|
326 |
+
"<loc-96>": 32200,
|
327 |
+
"<loc-97>": 32201,
|
328 |
+
"<loc-98>": 32202,
|
329 |
+
"<loc-99>": 32203,
|
330 |
+
"<loc-9>": 32113,
|
331 |
+
"<patch>": 32102
|
332 |
+
}
|
cvlm_llama2_tokenizer_100img_and_224loc_addpatch/special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": "<unk>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
cvlm_llama2_tokenizer_100img_and_224loc_addpatch/tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
3 |
+
size 499723
|
cvlm_llama2_tokenizer_100img_and_224loc_addpatch/tokenizer_config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": true,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"bos_token": {
|
5 |
+
"__type": "AddedToken",
|
6 |
+
"content": "<s>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": false,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"clean_up_tokenization_spaces": false,
|
13 |
+
"eos_token": {
|
14 |
+
"__type": "AddedToken",
|
15 |
+
"content": "</s>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": false,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false
|
20 |
+
},
|
21 |
+
"legacy": false,
|
22 |
+
"model_max_length": 1000000000000000019884624838656,
|
23 |
+
"pad_token": null,
|
24 |
+
"sp_model_kwargs": {},
|
25 |
+
"tokenizer_class": "LlamaTokenizer",
|
26 |
+
"unk_token": {
|
27 |
+
"__type": "AddedToken",
|
28 |
+
"content": "<unk>",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false
|
33 |
+
}
|
34 |
+
}
|