Training in progress, step 200
Browse files- config.json +591 -0
- model.safetensors +3 -0
- preprocessor_config.json +25 -0
- runs/Aug23_06-39-32_d2292a02e520/events.out.tfevents.1724395184.d2292a02e520.203.0 +3 -0
- runs/Aug23_06-41-55_d2292a02e520/events.out.tfevents.1724395323.d2292a02e520.203.1 +3 -0
- runs/Aug23_06-47-08_d2292a02e520/events.out.tfevents.1724395633.d2292a02e520.203.2 +3 -0
- runs/Aug23_06-51-31_d2292a02e520/events.out.tfevents.1724395896.d2292a02e520.203.3 +3 -0
- runs/Aug23_07-12-24_d2292a02e520/events.out.tfevents.1724397146.d2292a02e520.203.4 +3 -0
- runs/Aug23_07-15-44_d2292a02e520/events.out.tfevents.1724397349.d2292a02e520.203.5 +3 -0
- runs/Aug23_07-18-04_d2292a02e520/events.out.tfevents.1724397488.d2292a02e520.203.6 +3 -0
- runs/Aug23_07-24-56_d2292a02e520/events.out.tfevents.1724397902.d2292a02e520.203.7 +3 -0
- special_tokens_map.json +37 -0
- tokenizer.json +0 -0
- tokenizer_config.json +62 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
@@ -0,0 +1,591 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Salesforce/blip-vqa-base",
|
3 |
+
"architectures": [
|
4 |
+
"ViltForQuestionAnswering"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.0,
|
7 |
+
"hidden_act": "gelu",
|
8 |
+
"hidden_dropout_prob": 0.0,
|
9 |
+
"hidden_size": 768,
|
10 |
+
"id2label": {
|
11 |
+
"0": 0,
|
12 |
+
"1": 1,
|
13 |
+
"2": 2,
|
14 |
+
"3": 3,
|
15 |
+
"4": 4,
|
16 |
+
"5": 5,
|
17 |
+
"6": 6,
|
18 |
+
"7": 7,
|
19 |
+
"8": 8,
|
20 |
+
"9": 9,
|
21 |
+
"10": 10,
|
22 |
+
"11": 11,
|
23 |
+
"12": 12,
|
24 |
+
"13": 13,
|
25 |
+
"14": 14,
|
26 |
+
"15": 15,
|
27 |
+
"16": 16,
|
28 |
+
"17": 17,
|
29 |
+
"18": 18,
|
30 |
+
"19": 19,
|
31 |
+
"20": 20,
|
32 |
+
"21": 21,
|
33 |
+
"22": 22,
|
34 |
+
"23": 23,
|
35 |
+
"24": 24,
|
36 |
+
"25": 25,
|
37 |
+
"26": 26,
|
38 |
+
"27": 27,
|
39 |
+
"28": 28,
|
40 |
+
"29": 29,
|
41 |
+
"30": 30,
|
42 |
+
"31": 31,
|
43 |
+
"32": 32,
|
44 |
+
"33": 33,
|
45 |
+
"34": 34,
|
46 |
+
"35": 35,
|
47 |
+
"36": 36,
|
48 |
+
"37": 37,
|
49 |
+
"38": 38,
|
50 |
+
"39": 39,
|
51 |
+
"40": 40,
|
52 |
+
"41": 41,
|
53 |
+
"42": 42,
|
54 |
+
"43": 43,
|
55 |
+
"44": 44,
|
56 |
+
"45": 45,
|
57 |
+
"46": 46,
|
58 |
+
"47": 47,
|
59 |
+
"48": 48,
|
60 |
+
"49": 49,
|
61 |
+
"50": 50,
|
62 |
+
"51": 51,
|
63 |
+
"52": 52,
|
64 |
+
"53": 53,
|
65 |
+
"54": 54,
|
66 |
+
"55": 55,
|
67 |
+
"56": 56,
|
68 |
+
"57": 57,
|
69 |
+
"58": 58,
|
70 |
+
"59": 59,
|
71 |
+
"60": 60,
|
72 |
+
"61": 61,
|
73 |
+
"62": 62,
|
74 |
+
"63": 63,
|
75 |
+
"64": 64,
|
76 |
+
"65": 65,
|
77 |
+
"66": 66,
|
78 |
+
"67": 67,
|
79 |
+
"68": 68,
|
80 |
+
"69": 69,
|
81 |
+
"70": 70,
|
82 |
+
"71": 71,
|
83 |
+
"72": 72,
|
84 |
+
"73": 73,
|
85 |
+
"74": 74,
|
86 |
+
"75": 75,
|
87 |
+
"76": 76,
|
88 |
+
"77": 77,
|
89 |
+
"78": 78,
|
90 |
+
"79": 79,
|
91 |
+
"80": 80,
|
92 |
+
"81": 81,
|
93 |
+
"82": 82,
|
94 |
+
"83": 83,
|
95 |
+
"84": 84,
|
96 |
+
"85": 85,
|
97 |
+
"86": 86,
|
98 |
+
"87": 87,
|
99 |
+
"88": 88,
|
100 |
+
"89": 89,
|
101 |
+
"90": 90,
|
102 |
+
"91": 91,
|
103 |
+
"92": 92,
|
104 |
+
"93": 93,
|
105 |
+
"94": 94,
|
106 |
+
"95": 95,
|
107 |
+
"96": 96,
|
108 |
+
"97": 97,
|
109 |
+
"98": 98,
|
110 |
+
"99": 99,
|
111 |
+
"100": 100,
|
112 |
+
"101": 101,
|
113 |
+
"102": 102,
|
114 |
+
"103": 103,
|
115 |
+
"104": 104,
|
116 |
+
"105": 105,
|
117 |
+
"106": 106,
|
118 |
+
"107": 107,
|
119 |
+
"108": 108,
|
120 |
+
"109": 109,
|
121 |
+
"110": 110,
|
122 |
+
"111": 111,
|
123 |
+
"112": 112,
|
124 |
+
"113": 113,
|
125 |
+
"114": 114,
|
126 |
+
"115": 115,
|
127 |
+
"116": 116,
|
128 |
+
"117": 117,
|
129 |
+
"118": 118,
|
130 |
+
"119": 119,
|
131 |
+
"120": 120,
|
132 |
+
"121": 121,
|
133 |
+
"122": 122,
|
134 |
+
"123": 123,
|
135 |
+
"124": 124,
|
136 |
+
"125": 125,
|
137 |
+
"126": 126,
|
138 |
+
"127": 127,
|
139 |
+
"128": 128,
|
140 |
+
"129": 129,
|
141 |
+
"130": 130,
|
142 |
+
"131": 131,
|
143 |
+
"132": 132,
|
144 |
+
"133": 133,
|
145 |
+
"134": 134,
|
146 |
+
"135": 135,
|
147 |
+
"136": 136,
|
148 |
+
"137": 137,
|
149 |
+
"138": 138,
|
150 |
+
"139": 139,
|
151 |
+
"140": 140,
|
152 |
+
"141": 141,
|
153 |
+
"142": 142,
|
154 |
+
"143": 143,
|
155 |
+
"144": 144,
|
156 |
+
"145": 145,
|
157 |
+
"146": 146,
|
158 |
+
"147": 147,
|
159 |
+
"148": 148,
|
160 |
+
"149": 149,
|
161 |
+
"150": 150,
|
162 |
+
"151": 151,
|
163 |
+
"152": 152,
|
164 |
+
"153": 153,
|
165 |
+
"154": 154,
|
166 |
+
"155": 155,
|
167 |
+
"156": 156,
|
168 |
+
"157": 157,
|
169 |
+
"158": 158,
|
170 |
+
"159": 159,
|
171 |
+
"160": 160,
|
172 |
+
"161": 161,
|
173 |
+
"162": 162,
|
174 |
+
"163": 163,
|
175 |
+
"164": 164,
|
176 |
+
"165": 165,
|
177 |
+
"166": 166,
|
178 |
+
"167": 167,
|
179 |
+
"168": 168,
|
180 |
+
"169": 169,
|
181 |
+
"170": 170,
|
182 |
+
"171": 171,
|
183 |
+
"172": 172,
|
184 |
+
"173": 173,
|
185 |
+
"174": 174,
|
186 |
+
"175": 175,
|
187 |
+
"176": 176,
|
188 |
+
"177": 177,
|
189 |
+
"178": 178,
|
190 |
+
"179": 179,
|
191 |
+
"180": 180,
|
192 |
+
"181": 181,
|
193 |
+
"182": 182,
|
194 |
+
"183": 183,
|
195 |
+
"184": 184,
|
196 |
+
"185": 185,
|
197 |
+
"186": 186,
|
198 |
+
"187": 187,
|
199 |
+
"188": 188,
|
200 |
+
"189": 189,
|
201 |
+
"190": 190,
|
202 |
+
"191": 191,
|
203 |
+
"192": 192,
|
204 |
+
"193": 193,
|
205 |
+
"194": 194,
|
206 |
+
"195": 195,
|
207 |
+
"196": 196,
|
208 |
+
"197": 197,
|
209 |
+
"198": 198
|
210 |
+
},
|
211 |
+
"image_size": 384,
|
212 |
+
"image_text_hidden_size": 256,
|
213 |
+
"initializer_factor": 1.0,
|
214 |
+
"initializer_range": 0.02,
|
215 |
+
"intermediate_size": 3072,
|
216 |
+
"label2id": {
|
217 |
+
"0": 0,
|
218 |
+
"1": 1,
|
219 |
+
"2": 2,
|
220 |
+
"3": 3,
|
221 |
+
"4": 4,
|
222 |
+
"5": 5,
|
223 |
+
"6": 6,
|
224 |
+
"7": 7,
|
225 |
+
"8": 8,
|
226 |
+
"9": 9,
|
227 |
+
"10": 10,
|
228 |
+
"11": 11,
|
229 |
+
"12": 12,
|
230 |
+
"13": 13,
|
231 |
+
"14": 14,
|
232 |
+
"15": 15,
|
233 |
+
"16": 16,
|
234 |
+
"17": 17,
|
235 |
+
"18": 18,
|
236 |
+
"19": 19,
|
237 |
+
"20": 20,
|
238 |
+
"21": 21,
|
239 |
+
"22": 22,
|
240 |
+
"23": 23,
|
241 |
+
"24": 24,
|
242 |
+
"25": 25,
|
243 |
+
"26": 26,
|
244 |
+
"27": 27,
|
245 |
+
"28": 28,
|
246 |
+
"29": 29,
|
247 |
+
"30": 30,
|
248 |
+
"31": 31,
|
249 |
+
"32": 32,
|
250 |
+
"33": 33,
|
251 |
+
"34": 34,
|
252 |
+
"35": 35,
|
253 |
+
"36": 36,
|
254 |
+
"37": 37,
|
255 |
+
"38": 38,
|
256 |
+
"39": 39,
|
257 |
+
"40": 40,
|
258 |
+
"41": 41,
|
259 |
+
"42": 42,
|
260 |
+
"43": 43,
|
261 |
+
"44": 44,
|
262 |
+
"45": 45,
|
263 |
+
"46": 46,
|
264 |
+
"47": 47,
|
265 |
+
"48": 48,
|
266 |
+
"49": 49,
|
267 |
+
"50": 50,
|
268 |
+
"51": 51,
|
269 |
+
"52": 52,
|
270 |
+
"53": 53,
|
271 |
+
"54": 54,
|
272 |
+
"55": 55,
|
273 |
+
"56": 56,
|
274 |
+
"57": 57,
|
275 |
+
"58": 58,
|
276 |
+
"59": 59,
|
277 |
+
"60": 60,
|
278 |
+
"61": 61,
|
279 |
+
"62": 62,
|
280 |
+
"63": 63,
|
281 |
+
"64": 64,
|
282 |
+
"65": 65,
|
283 |
+
"66": 66,
|
284 |
+
"67": 67,
|
285 |
+
"68": 68,
|
286 |
+
"69": 69,
|
287 |
+
"70": 70,
|
288 |
+
"71": 71,
|
289 |
+
"72": 72,
|
290 |
+
"73": 73,
|
291 |
+
"74": 74,
|
292 |
+
"75": 75,
|
293 |
+
"76": 76,
|
294 |
+
"77": 77,
|
295 |
+
"78": 78,
|
296 |
+
"79": 79,
|
297 |
+
"80": 80,
|
298 |
+
"81": 81,
|
299 |
+
"82": 82,
|
300 |
+
"83": 83,
|
301 |
+
"84": 84,
|
302 |
+
"85": 85,
|
303 |
+
"86": 86,
|
304 |
+
"87": 87,
|
305 |
+
"88": 88,
|
306 |
+
"89": 89,
|
307 |
+
"90": 90,
|
308 |
+
"91": 91,
|
309 |
+
"92": 92,
|
310 |
+
"93": 93,
|
311 |
+
"94": 94,
|
312 |
+
"95": 95,
|
313 |
+
"96": 96,
|
314 |
+
"97": 97,
|
315 |
+
"98": 98,
|
316 |
+
"99": 99,
|
317 |
+
"100": 100,
|
318 |
+
"101": 101,
|
319 |
+
"102": 102,
|
320 |
+
"103": 103,
|
321 |
+
"104": 104,
|
322 |
+
"105": 105,
|
323 |
+
"106": 106,
|
324 |
+
"107": 107,
|
325 |
+
"108": 108,
|
326 |
+
"109": 109,
|
327 |
+
"110": 110,
|
328 |
+
"111": 111,
|
329 |
+
"112": 112,
|
330 |
+
"113": 113,
|
331 |
+
"114": 114,
|
332 |
+
"115": 115,
|
333 |
+
"116": 116,
|
334 |
+
"117": 117,
|
335 |
+
"118": 118,
|
336 |
+
"119": 119,
|
337 |
+
"120": 120,
|
338 |
+
"121": 121,
|
339 |
+
"122": 122,
|
340 |
+
"123": 123,
|
341 |
+
"124": 124,
|
342 |
+
"125": 125,
|
343 |
+
"126": 126,
|
344 |
+
"127": 127,
|
345 |
+
"128": 128,
|
346 |
+
"129": 129,
|
347 |
+
"130": 130,
|
348 |
+
"131": 131,
|
349 |
+
"132": 132,
|
350 |
+
"133": 133,
|
351 |
+
"134": 134,
|
352 |
+
"135": 135,
|
353 |
+
"136": 136,
|
354 |
+
"137": 137,
|
355 |
+
"138": 138,
|
356 |
+
"139": 139,
|
357 |
+
"140": 140,
|
358 |
+
"141": 141,
|
359 |
+
"142": 142,
|
360 |
+
"143": 143,
|
361 |
+
"144": 144,
|
362 |
+
"145": 145,
|
363 |
+
"146": 146,
|
364 |
+
"147": 147,
|
365 |
+
"148": 148,
|
366 |
+
"149": 149,
|
367 |
+
"150": 150,
|
368 |
+
"151": 151,
|
369 |
+
"152": 152,
|
370 |
+
"153": 153,
|
371 |
+
"154": 154,
|
372 |
+
"155": 155,
|
373 |
+
"156": 156,
|
374 |
+
"157": 157,
|
375 |
+
"158": 158,
|
376 |
+
"159": 159,
|
377 |
+
"160": 160,
|
378 |
+
"161": 161,
|
379 |
+
"162": 162,
|
380 |
+
"163": 163,
|
381 |
+
"164": 164,
|
382 |
+
"165": 165,
|
383 |
+
"166": 166,
|
384 |
+
"167": 167,
|
385 |
+
"168": 168,
|
386 |
+
"169": 169,
|
387 |
+
"170": 170,
|
388 |
+
"171": 171,
|
389 |
+
"172": 172,
|
390 |
+
"173": 173,
|
391 |
+
"174": 174,
|
392 |
+
"175": 175,
|
393 |
+
"176": 176,
|
394 |
+
"177": 177,
|
395 |
+
"178": 178,
|
396 |
+
"179": 179,
|
397 |
+
"180": 180,
|
398 |
+
"181": 181,
|
399 |
+
"182": 182,
|
400 |
+
"183": 183,
|
401 |
+
"184": 184,
|
402 |
+
"185": 185,
|
403 |
+
"186": 186,
|
404 |
+
"187": 187,
|
405 |
+
"188": 188,
|
406 |
+
"189": 189,
|
407 |
+
"190": 190,
|
408 |
+
"191": 191,
|
409 |
+
"192": 192,
|
410 |
+
"193": 193,
|
411 |
+
"194": 194,
|
412 |
+
"195": 195,
|
413 |
+
"196": 196,
|
414 |
+
"197": 197,
|
415 |
+
"198": 198
|
416 |
+
},
|
417 |
+
"layer_norm_eps": 1e-12,
|
418 |
+
"logit_scale_init_value": 2.6592,
|
419 |
+
"max_image_length": -1,
|
420 |
+
"max_position_embeddings": 40,
|
421 |
+
"modality_type_vocab_size": 2,
|
422 |
+
"model_type": "vilt",
|
423 |
+
"num_attention_heads": 12,
|
424 |
+
"num_channels": 3,
|
425 |
+
"num_hidden_layers": 12,
|
426 |
+
"num_images": -1,
|
427 |
+
"patch_size": 32,
|
428 |
+
"projection_dim": 512,
|
429 |
+
"qkv_bias": true,
|
430 |
+
"text_config": {
|
431 |
+
"_name_or_path": "",
|
432 |
+
"add_cross_attention": false,
|
433 |
+
"architectures": null,
|
434 |
+
"attention_probs_dropout_prob": 0.0,
|
435 |
+
"bad_words_ids": null,
|
436 |
+
"begin_suppress_tokens": null,
|
437 |
+
"bos_token_id": 30522,
|
438 |
+
"chunk_size_feed_forward": 0,
|
439 |
+
"cross_attention_hidden_size": null,
|
440 |
+
"decoder_start_token_id": null,
|
441 |
+
"diversity_penalty": 0.0,
|
442 |
+
"do_sample": false,
|
443 |
+
"early_stopping": false,
|
444 |
+
"encoder_no_repeat_ngram_size": 0,
|
445 |
+
"eos_token_id": 2,
|
446 |
+
"exponential_decay_length_penalty": null,
|
447 |
+
"finetuning_task": null,
|
448 |
+
"forced_bos_token_id": null,
|
449 |
+
"forced_eos_token_id": null,
|
450 |
+
"hidden_act": "gelu",
|
451 |
+
"hidden_dropout_prob": 0.0,
|
452 |
+
"hidden_size": 768,
|
453 |
+
"id2label": {
|
454 |
+
"0": "LABEL_0",
|
455 |
+
"1": "LABEL_1"
|
456 |
+
},
|
457 |
+
"initializer_factor": 1.0,
|
458 |
+
"initializer_range": 0.02,
|
459 |
+
"intermediate_size": 3072,
|
460 |
+
"is_decoder": true,
|
461 |
+
"is_encoder_decoder": false,
|
462 |
+
"label2id": {
|
463 |
+
"LABEL_0": 0,
|
464 |
+
"LABEL_1": 1
|
465 |
+
},
|
466 |
+
"layer_norm_eps": 1e-12,
|
467 |
+
"length_penalty": 1.0,
|
468 |
+
"max_length": 20,
|
469 |
+
"max_position_embeddings": 512,
|
470 |
+
"min_length": 0,
|
471 |
+
"model_type": "blip_text_model",
|
472 |
+
"no_repeat_ngram_size": 0,
|
473 |
+
"num_attention_heads": 12,
|
474 |
+
"num_beam_groups": 1,
|
475 |
+
"num_beams": 1,
|
476 |
+
"num_hidden_layers": 12,
|
477 |
+
"num_return_sequences": 1,
|
478 |
+
"output_attentions": false,
|
479 |
+
"output_hidden_states": false,
|
480 |
+
"output_scores": false,
|
481 |
+
"pad_token_id": 0,
|
482 |
+
"prefix": null,
|
483 |
+
"problem_type": null,
|
484 |
+
"projection_dim": 768,
|
485 |
+
"pruned_heads": {},
|
486 |
+
"remove_invalid_values": false,
|
487 |
+
"repetition_penalty": 1.0,
|
488 |
+
"return_dict": true,
|
489 |
+
"return_dict_in_generate": false,
|
490 |
+
"sep_token_id": 102,
|
491 |
+
"suppress_tokens": null,
|
492 |
+
"task_specific_params": null,
|
493 |
+
"temperature": 1.0,
|
494 |
+
"tf_legacy_loss": false,
|
495 |
+
"tie_encoder_decoder": false,
|
496 |
+
"tie_word_embeddings": true,
|
497 |
+
"tokenizer_class": null,
|
498 |
+
"top_k": 50,
|
499 |
+
"top_p": 1.0,
|
500 |
+
"torch_dtype": null,
|
501 |
+
"torchscript": false,
|
502 |
+
"transformers_version": "4.26.0.dev0",
|
503 |
+
"typical_p": 1.0,
|
504 |
+
"use_bfloat16": false,
|
505 |
+
"use_cache": true,
|
506 |
+
"vocab_size": 30524
|
507 |
+
},
|
508 |
+
"tie_word_embeddings": false,
|
509 |
+
"torch_dtype": "float32",
|
510 |
+
"transformers_version": "4.42.4",
|
511 |
+
"type_vocab_size": 2,
|
512 |
+
"vision_config": {
|
513 |
+
"_name_or_path": "",
|
514 |
+
"add_cross_attention": false,
|
515 |
+
"architectures": null,
|
516 |
+
"attention_dropout": 0.0,
|
517 |
+
"bad_words_ids": null,
|
518 |
+
"begin_suppress_tokens": null,
|
519 |
+
"bos_token_id": null,
|
520 |
+
"chunk_size_feed_forward": 0,
|
521 |
+
"cross_attention_hidden_size": null,
|
522 |
+
"decoder_start_token_id": null,
|
523 |
+
"diversity_penalty": 0.0,
|
524 |
+
"do_sample": false,
|
525 |
+
"dropout": 0.0,
|
526 |
+
"early_stopping": false,
|
527 |
+
"encoder_no_repeat_ngram_size": 0,
|
528 |
+
"eos_token_id": null,
|
529 |
+
"exponential_decay_length_penalty": null,
|
530 |
+
"finetuning_task": null,
|
531 |
+
"forced_bos_token_id": null,
|
532 |
+
"forced_eos_token_id": null,
|
533 |
+
"hidden_act": "gelu",
|
534 |
+
"hidden_size": 768,
|
535 |
+
"id2label": {
|
536 |
+
"0": "LABEL_0",
|
537 |
+
"1": "LABEL_1"
|
538 |
+
},
|
539 |
+
"image_size": 384,
|
540 |
+
"initializer_factor": 1.0,
|
541 |
+
"initializer_range": 0.02,
|
542 |
+
"intermediate_size": 3072,
|
543 |
+
"is_decoder": false,
|
544 |
+
"is_encoder_decoder": false,
|
545 |
+
"label2id": {
|
546 |
+
"LABEL_0": 0,
|
547 |
+
"LABEL_1": 1
|
548 |
+
},
|
549 |
+
"layer_norm_eps": 1e-05,
|
550 |
+
"length_penalty": 1.0,
|
551 |
+
"max_length": 20,
|
552 |
+
"min_length": 0,
|
553 |
+
"model_type": "blip_vision_model",
|
554 |
+
"no_repeat_ngram_size": 0,
|
555 |
+
"num_attention_heads": 12,
|
556 |
+
"num_beam_groups": 1,
|
557 |
+
"num_beams": 1,
|
558 |
+
"num_channels": 3,
|
559 |
+
"num_hidden_layers": 12,
|
560 |
+
"num_return_sequences": 1,
|
561 |
+
"output_attentions": false,
|
562 |
+
"output_hidden_states": false,
|
563 |
+
"output_scores": false,
|
564 |
+
"pad_token_id": null,
|
565 |
+
"patch_size": 16,
|
566 |
+
"prefix": null,
|
567 |
+
"problem_type": null,
|
568 |
+
"projection_dim": 512,
|
569 |
+
"pruned_heads": {},
|
570 |
+
"remove_invalid_values": false,
|
571 |
+
"repetition_penalty": 1.0,
|
572 |
+
"return_dict": true,
|
573 |
+
"return_dict_in_generate": false,
|
574 |
+
"sep_token_id": null,
|
575 |
+
"suppress_tokens": null,
|
576 |
+
"task_specific_params": null,
|
577 |
+
"temperature": 1.0,
|
578 |
+
"tf_legacy_loss": false,
|
579 |
+
"tie_encoder_decoder": false,
|
580 |
+
"tie_word_embeddings": true,
|
581 |
+
"tokenizer_class": null,
|
582 |
+
"top_k": 50,
|
583 |
+
"top_p": 1.0,
|
584 |
+
"torch_dtype": null,
|
585 |
+
"torchscript": false,
|
586 |
+
"transformers_version": "4.26.0.dev0",
|
587 |
+
"typical_p": 1.0,
|
588 |
+
"use_bfloat16": false
|
589 |
+
},
|
590 |
+
"vocab_size": 30522
|
591 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a2af93ea4c11397458233ff4a43c98bc02cc98d953543d71c98b0cfd39e0a158
|
3 |
+
size 452365324
|
preprocessor_config.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"do_pad": true,
|
4 |
+
"do_rescale": true,
|
5 |
+
"do_resize": true,
|
6 |
+
"image_mean": [
|
7 |
+
0.48145466,
|
8 |
+
0.4578275,
|
9 |
+
0.40821073
|
10 |
+
],
|
11 |
+
"image_processor_type": "ViltImageProcessor",
|
12 |
+
"image_std": [
|
13 |
+
0.26862954,
|
14 |
+
0.26130258,
|
15 |
+
0.27577711
|
16 |
+
],
|
17 |
+
"processor_class": "ViltProcessor",
|
18 |
+
"resample": 3,
|
19 |
+
"rescale_factor": 0.00392156862745098,
|
20 |
+
"size": {
|
21 |
+
"height": 384,
|
22 |
+
"width": 384
|
23 |
+
},
|
24 |
+
"size_divisor": 32
|
25 |
+
}
|
runs/Aug23_06-39-32_d2292a02e520/events.out.tfevents.1724395184.d2292a02e520.203.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:26bd17d904c4bb1250a49acd077206cd2d975333b9514efd421663c075bd0a53
|
3 |
+
size 17119
|
runs/Aug23_06-41-55_d2292a02e520/events.out.tfevents.1724395323.d2292a02e520.203.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:52c54fa72422cc7e60ba91f482df7c2339c2611186a9a5edf231a59c9517b569
|
3 |
+
size 17119
|
runs/Aug23_06-47-08_d2292a02e520/events.out.tfevents.1724395633.d2292a02e520.203.2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4ac574c7ffe592e6bd6ebd4a1fc6456329d592c7de776f5deb62ed4c6ee062e
|
3 |
+
size 17119
|
runs/Aug23_06-51-31_d2292a02e520/events.out.tfevents.1724395896.d2292a02e520.203.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f98cca7395bd63ca0e3ca4e881c8b1395b4d3b396fc1e71c50048db81a868bc4
|
3 |
+
size 17119
|
runs/Aug23_07-12-24_d2292a02e520/events.out.tfevents.1724397146.d2292a02e520.203.4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a15fa6d1f71498aa71c7060b65453f94175ba15aa76cd4e9f1a0a32ce294b4b6
|
3 |
+
size 17119
|
runs/Aug23_07-15-44_d2292a02e520/events.out.tfevents.1724397349.d2292a02e520.203.5
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0289669657d4cb27af5ec51dcfea637d47e7ee135fd5be481c500786022fe2ec
|
3 |
+
size 17119
|
runs/Aug23_07-18-04_d2292a02e520/events.out.tfevents.1724397488.d2292a02e520.203.6
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8223a117a49ff4e806c278989fe878c618a72e8b76f0261e0b15983ba3ac4ad5
|
3 |
+
size 15297
|
runs/Aug23_07-24-56_d2292a02e520/events.out.tfevents.1724397902.d2292a02e520.203.7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f7061ff4f2d22f31db52a5d3cc155a32a2aec5d060eb167cabb9c2c70f94b86d
|
3 |
+
size 16133
|
special_tokens_map.json
ADDED
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": {
|
3 |
+
"content": "[CLS]",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"mask_token": {
|
10 |
+
"content": "[MASK]",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": {
|
17 |
+
"content": "[PAD]",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"sep_token": {
|
24 |
+
"content": "[SEP]",
|
25 |
+
"lstrip": false,
|
26 |
+
"normalized": false,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": false
|
29 |
+
},
|
30 |
+
"unk_token": {
|
31 |
+
"content": "[UNK]",
|
32 |
+
"lstrip": false,
|
33 |
+
"normalized": false,
|
34 |
+
"rstrip": false,
|
35 |
+
"single_word": false
|
36 |
+
}
|
37 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"added_tokens_decoder": {
|
3 |
+
"0": {
|
4 |
+
"content": "[PAD]",
|
5 |
+
"lstrip": false,
|
6 |
+
"normalized": false,
|
7 |
+
"rstrip": false,
|
8 |
+
"single_word": false,
|
9 |
+
"special": true
|
10 |
+
},
|
11 |
+
"100": {
|
12 |
+
"content": "[UNK]",
|
13 |
+
"lstrip": false,
|
14 |
+
"normalized": false,
|
15 |
+
"rstrip": false,
|
16 |
+
"single_word": false,
|
17 |
+
"special": true
|
18 |
+
},
|
19 |
+
"101": {
|
20 |
+
"content": "[CLS]",
|
21 |
+
"lstrip": false,
|
22 |
+
"normalized": false,
|
23 |
+
"rstrip": false,
|
24 |
+
"single_word": false,
|
25 |
+
"special": true
|
26 |
+
},
|
27 |
+
"102": {
|
28 |
+
"content": "[SEP]",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false,
|
33 |
+
"special": true
|
34 |
+
},
|
35 |
+
"103": {
|
36 |
+
"content": "[MASK]",
|
37 |
+
"lstrip": false,
|
38 |
+
"normalized": false,
|
39 |
+
"rstrip": false,
|
40 |
+
"single_word": false,
|
41 |
+
"special": true
|
42 |
+
}
|
43 |
+
},
|
44 |
+
"clean_up_tokenization_spaces": true,
|
45 |
+
"cls_token": "[CLS]",
|
46 |
+
"do_basic_tokenize": true,
|
47 |
+
"do_lower_case": true,
|
48 |
+
"mask_token": "[MASK]",
|
49 |
+
"model_input_names": [
|
50 |
+
"input_ids",
|
51 |
+
"attention_mask"
|
52 |
+
],
|
53 |
+
"model_max_length": 512,
|
54 |
+
"never_split": null,
|
55 |
+
"pad_token": "[PAD]",
|
56 |
+
"processor_class": "ViltProcessor",
|
57 |
+
"sep_token": "[SEP]",
|
58 |
+
"strip_accents": null,
|
59 |
+
"tokenize_chinese_chars": true,
|
60 |
+
"tokenizer_class": "BertTokenizer",
|
61 |
+
"unk_token": "[UNK]"
|
62 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0f1e41af2def3c4612b0d14a4baf74aa6496108ae04b0b4c6c4613e700e5770e
|
3 |
+
size 5112
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|