Hugging Face's logo

DylanJHJ
/
blip-vqa-base-product2query-v1

Model card Files Community
DylanJHJ commited on
Commit
ee07c0f
•
1 Parent(s): 2efc2ff

update checkpoint

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +9 -0
  2. checkpoint-10000/config.json +26 -0
  3. checkpoint-10000/generation_config.json +7 -0
  4. checkpoint-10000/optimizer.pt +3 -0
  5. checkpoint-10000/pytorch_model.bin +3 -0
  6. checkpoint-10000/rng_state.pth +3 -0
  7. checkpoint-10000/scheduler.pt +3 -0
  8. checkpoint-10000/trainer_state.json +399 -0
  9. checkpoint-10000/training_args.bin +3 -0
  10. checkpoint-12500/config.json +26 -0
  11. checkpoint-12500/generation_config.json +7 -0
  12. checkpoint-12500/optimizer.pt +3 -0
  13. checkpoint-12500/pytorch_model.bin +3 -0
  14. checkpoint-12500/rng_state.pth +3 -0
  15. checkpoint-12500/scheduler.pt +3 -0
  16. checkpoint-12500/trainer_state.json +494 -0
  17. checkpoint-12500/training_args.bin +3 -0
  18. checkpoint-15000/config.json +26 -0
  19. checkpoint-15000/generation_config.json +7 -0
  20. checkpoint-15000/optimizer.pt +3 -0
  21. checkpoint-15000/pytorch_model.bin +3 -0
  22. checkpoint-15000/rng_state.pth +3 -0
  23. checkpoint-15000/scheduler.pt +3 -0
  24. checkpoint-15000/trainer_state.json +589 -0
  25. checkpoint-15000/training_args.bin +3 -0
  26. checkpoint-17500/config.json +26 -0
  27. checkpoint-17500/generation_config.json +7 -0
  28. checkpoint-17500/optimizer.pt +3 -0
  29. checkpoint-17500/pytorch_model.bin +3 -0
  30. checkpoint-17500/rng_state.pth +3 -0
  31. checkpoint-17500/scheduler.pt +3 -0
  32. checkpoint-17500/trainer_state.json +684 -0
  33. checkpoint-17500/training_args.bin +3 -0
  34. checkpoint-20000/config.json +26 -0
  35. checkpoint-20000/generation_config.json +7 -0
  36. checkpoint-20000/optimizer.pt +3 -0
  37. checkpoint-20000/pytorch_model.bin +3 -0
  38. checkpoint-20000/rng_state.pth +3 -0
  39. checkpoint-20000/scheduler.pt +3 -0
  40. checkpoint-20000/trainer_state.json +779 -0
  41. checkpoint-20000/training_args.bin +3 -0
  42. checkpoint-22500/config.json +26 -0
  43. checkpoint-22500/generation_config.json +7 -0
  44. checkpoint-22500/optimizer.pt +3 -0
  45. checkpoint-22500/pytorch_model.bin +3 -0
  46. checkpoint-22500/rng_state.pth +3 -0
  47. checkpoint-22500/scheduler.pt +3 -0
  48. checkpoint-22500/trainer_state.json +874 -0
  49. checkpoint-22500/training_args.bin +3 -0
  50. checkpoint-25000/config.json +26 -0
.gitattributes CHANGED
@@ -33,3 +33,12 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ checkpoint-20000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
37
+ checkpoint-22500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
38
+ checkpoint-25000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
39
+ checkpoint-27500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
40
+ checkpoint-12500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
41
+ checkpoint-17500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
42
+ checkpoint-7500/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
43
+ checkpoint-10000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
44
+ checkpoint-15000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
checkpoint-10000/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-10000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-10000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27465b58163aae26fbf301b92a6f9d7cc830108c6c7250ac807633e5278e4a4d
3
+ size 3078042062
checkpoint-10000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:92d249507f796a36c2c26a19deb8a34a2915cdfc9d572702141845bcea0dd59a
3
+ size 1538969242
checkpoint-10000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5f38f6446ba43bb9ae0be4911c150b41c5adfbe2712844ea3cb854b62ee2432
3
+ size 14244
checkpoint-10000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73fd0d7947c346b952f2fd5927f6ac1598903b1942bc135c6555d3aee039930c
3
+ size 1064
checkpoint-10000/trainer_state.json ADDED
@@ -0,0 +1,399 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.4443561782335523,
5
+ "eval_steps": 500,
6
+ "global_step": 10000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ }
391
+ ],
392
+ "logging_steps": 500,
393
+ "max_steps": 30000,
394
+ "num_train_epochs": 5,
395
+ "save_steps": 2500,
396
+ "total_flos": 4.113331352053805e+16,
397
+ "trial_name": null,
398
+ "trial_params": null
399
+ }
checkpoint-10000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-12500/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-12500/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-12500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee97fcdb8c080ac2caa5fcf2373e95541d2445a9454adf858e057536e89dc639
3
+ size 3078042062
checkpoint-12500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9972c892ef1ff70e3e00dd712cae70e31cbaabdebc7090055cf647fc19348fb
3
+ size 1538969242
checkpoint-12500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f32c25f8ac5f19710a6ed752bf6968914f1573258beefeff5964234566a45c0a
3
+ size 14244
checkpoint-12500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5242ff2f4901c608786cd6ac1f4792276106232603b8ffc6173b6cb5b054b802
3
+ size 1064
checkpoint-12500/trainer_state.json ADDED
@@ -0,0 +1,494 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.8054452227919406,
5
+ "eval_steps": 500,
6
+ "global_step": 12500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ },
391
+ {
392
+ "epoch": 1.52,
393
+ "learning_rate": 1.4539904997395468e-05,
394
+ "loss": 1.6827,
395
+ "step": 10500
396
+ },
397
+ {
398
+ "epoch": 1.52,
399
+ "eval_gen_len": 6.267666666666667,
400
+ "eval_loss": 1.8478548526763916,
401
+ "eval_rouge1": 75.7687,
402
+ "eval_rouge2": 62.3363,
403
+ "eval_rougeL": 75.7808,
404
+ "eval_rougeLsum": 75.7533,
405
+ "eval_runtime": 1520.7572,
406
+ "eval_samples_per_second": 1.973,
407
+ "eval_steps_per_second": 0.493,
408
+ "step": 10500
409
+ },
410
+ {
411
+ "epoch": 1.59,
412
+ "learning_rate": 1.4067366430758004e-05,
413
+ "loss": 1.6808,
414
+ "step": 11000
415
+ },
416
+ {
417
+ "epoch": 1.59,
418
+ "eval_gen_len": 6.267666666666667,
419
+ "eval_loss": 1.8268691301345825,
420
+ "eval_rouge1": 76.5635,
421
+ "eval_rouge2": 63.5572,
422
+ "eval_rougeL": 76.5759,
423
+ "eval_rougeLsum": 76.5281,
424
+ "eval_runtime": 1529.0949,
425
+ "eval_samples_per_second": 1.962,
426
+ "eval_steps_per_second": 0.49,
427
+ "step": 11000
428
+ },
429
+ {
430
+ "epoch": 1.66,
431
+ "learning_rate": 1.3583679495453e-05,
432
+ "loss": 1.6747,
433
+ "step": 11500
434
+ },
435
+ {
436
+ "epoch": 1.66,
437
+ "eval_gen_len": 6.267666666666667,
438
+ "eval_loss": 1.8204782009124756,
439
+ "eval_rouge1": 76.0665,
440
+ "eval_rouge2": 62.7626,
441
+ "eval_rougeL": 76.0441,
442
+ "eval_rougeLsum": 76.0155,
443
+ "eval_runtime": 1521.0486,
444
+ "eval_samples_per_second": 1.972,
445
+ "eval_steps_per_second": 0.493,
446
+ "step": 11500
447
+ },
448
+ {
449
+ "epoch": 1.73,
450
+ "learning_rate": 1.3090169943749475e-05,
451
+ "loss": 1.6628,
452
+ "step": 12000
453
+ },
454
+ {
455
+ "epoch": 1.73,
456
+ "eval_gen_len": 6.267666666666667,
457
+ "eval_loss": 1.8059669733047485,
458
+ "eval_rouge1": 76.6687,
459
+ "eval_rouge2": 63.8953,
460
+ "eval_rougeL": 76.678,
461
+ "eval_rougeLsum": 76.6457,
462
+ "eval_runtime": 1519.6043,
463
+ "eval_samples_per_second": 1.974,
464
+ "eval_steps_per_second": 0.494,
465
+ "step": 12000
466
+ },
467
+ {
468
+ "epoch": 1.81,
469
+ "learning_rate": 1.2588190451025209e-05,
470
+ "loss": 1.6517,
471
+ "step": 12500
472
+ },
473
+ {
474
+ "epoch": 1.81,
475
+ "eval_gen_len": 6.267666666666667,
476
+ "eval_loss": 1.7973886728286743,
477
+ "eval_rouge1": 77.5432,
478
+ "eval_rouge2": 64.831,
479
+ "eval_rougeL": 77.5477,
480
+ "eval_rougeLsum": 77.5112,
481
+ "eval_runtime": 1514.3733,
482
+ "eval_samples_per_second": 1.981,
483
+ "eval_steps_per_second": 0.495,
484
+ "step": 12500
485
+ }
486
+ ],
487
+ "logging_steps": 500,
488
+ "max_steps": 30000,
489
+ "num_train_epochs": 5,
490
+ "save_steps": 2500,
491
+ "total_flos": 5.140320363779078e+16,
492
+ "trial_name": null,
493
+ "trial_params": null
494
+ }
checkpoint-12500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-15000/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-15000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-15000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fdf0d895932edee4066f27c6e79b7327a9eb1290adb260bf36ce4e9ea78b27fa
3
+ size 3078042062
checkpoint-15000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2cf21ac142ab8371603f7aacb99f6cbc7b993e078fd4569c3cc4ccaa1df32dec
3
+ size 1538969242
checkpoint-15000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f86f749dc73bc16a2502fef2f98f5c00b4400cb2c67fbe62653b7ed104d13779
3
+ size 14244
checkpoint-15000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0791ab28ed85fc2b7bf19fb7882ce2711b9dea36e12032576612122561cfd5f1
3
+ size 1064
checkpoint-15000/trainer_state.json ADDED
@@ -0,0 +1,589 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.1665342673503285,
5
+ "eval_steps": 500,
6
+ "global_step": 15000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ },
391
+ {
392
+ "epoch": 1.52,
393
+ "learning_rate": 1.4539904997395468e-05,
394
+ "loss": 1.6827,
395
+ "step": 10500
396
+ },
397
+ {
398
+ "epoch": 1.52,
399
+ "eval_gen_len": 6.267666666666667,
400
+ "eval_loss": 1.8478548526763916,
401
+ "eval_rouge1": 75.7687,
402
+ "eval_rouge2": 62.3363,
403
+ "eval_rougeL": 75.7808,
404
+ "eval_rougeLsum": 75.7533,
405
+ "eval_runtime": 1520.7572,
406
+ "eval_samples_per_second": 1.973,
407
+ "eval_steps_per_second": 0.493,
408
+ "step": 10500
409
+ },
410
+ {
411
+ "epoch": 1.59,
412
+ "learning_rate": 1.4067366430758004e-05,
413
+ "loss": 1.6808,
414
+ "step": 11000
415
+ },
416
+ {
417
+ "epoch": 1.59,
418
+ "eval_gen_len": 6.267666666666667,
419
+ "eval_loss": 1.8268691301345825,
420
+ "eval_rouge1": 76.5635,
421
+ "eval_rouge2": 63.5572,
422
+ "eval_rougeL": 76.5759,
423
+ "eval_rougeLsum": 76.5281,
424
+ "eval_runtime": 1529.0949,
425
+ "eval_samples_per_second": 1.962,
426
+ "eval_steps_per_second": 0.49,
427
+ "step": 11000
428
+ },
429
+ {
430
+ "epoch": 1.66,
431
+ "learning_rate": 1.3583679495453e-05,
432
+ "loss": 1.6747,
433
+ "step": 11500
434
+ },
435
+ {
436
+ "epoch": 1.66,
437
+ "eval_gen_len": 6.267666666666667,
438
+ "eval_loss": 1.8204782009124756,
439
+ "eval_rouge1": 76.0665,
440
+ "eval_rouge2": 62.7626,
441
+ "eval_rougeL": 76.0441,
442
+ "eval_rougeLsum": 76.0155,
443
+ "eval_runtime": 1521.0486,
444
+ "eval_samples_per_second": 1.972,
445
+ "eval_steps_per_second": 0.493,
446
+ "step": 11500
447
+ },
448
+ {
449
+ "epoch": 1.73,
450
+ "learning_rate": 1.3090169943749475e-05,
451
+ "loss": 1.6628,
452
+ "step": 12000
453
+ },
454
+ {
455
+ "epoch": 1.73,
456
+ "eval_gen_len": 6.267666666666667,
457
+ "eval_loss": 1.8059669733047485,
458
+ "eval_rouge1": 76.6687,
459
+ "eval_rouge2": 63.8953,
460
+ "eval_rougeL": 76.678,
461
+ "eval_rougeLsum": 76.6457,
462
+ "eval_runtime": 1519.6043,
463
+ "eval_samples_per_second": 1.974,
464
+ "eval_steps_per_second": 0.494,
465
+ "step": 12000
466
+ },
467
+ {
468
+ "epoch": 1.81,
469
+ "learning_rate": 1.2588190451025209e-05,
470
+ "loss": 1.6517,
471
+ "step": 12500
472
+ },
473
+ {
474
+ "epoch": 1.81,
475
+ "eval_gen_len": 6.267666666666667,
476
+ "eval_loss": 1.7973886728286743,
477
+ "eval_rouge1": 77.5432,
478
+ "eval_rouge2": 64.831,
479
+ "eval_rougeL": 77.5477,
480
+ "eval_rougeLsum": 77.5112,
481
+ "eval_runtime": 1514.3733,
482
+ "eval_samples_per_second": 1.981,
483
+ "eval_steps_per_second": 0.495,
484
+ "step": 12500
485
+ },
486
+ {
487
+ "epoch": 1.88,
488
+ "learning_rate": 1.2079116908177592e-05,
489
+ "loss": 1.6493,
490
+ "step": 13000
491
+ },
492
+ {
493
+ "epoch": 1.88,
494
+ "eval_gen_len": 6.267666666666667,
495
+ "eval_loss": 1.7867239713668823,
496
+ "eval_rouge1": 77.6195,
497
+ "eval_rouge2": 64.9393,
498
+ "eval_rougeL": 77.5908,
499
+ "eval_rougeLsum": 77.5498,
500
+ "eval_runtime": 1514.8609,
501
+ "eval_samples_per_second": 1.98,
502
+ "eval_steps_per_second": 0.495,
503
+ "step": 13000
504
+ },
505
+ {
506
+ "epoch": 1.95,
507
+ "learning_rate": 1.156434465040231e-05,
508
+ "loss": 1.6435,
509
+ "step": 13500
510
+ },
511
+ {
512
+ "epoch": 1.95,
513
+ "eval_gen_len": 6.267666666666667,
514
+ "eval_loss": 1.780102252960205,
515
+ "eval_rouge1": 77.8512,
516
+ "eval_rouge2": 65.2471,
517
+ "eval_rougeL": 77.8573,
518
+ "eval_rougeLsum": 77.8118,
519
+ "eval_runtime": 1516.9222,
520
+ "eval_samples_per_second": 1.978,
521
+ "eval_steps_per_second": 0.494,
522
+ "step": 13500
523
+ },
524
+ {
525
+ "epoch": 2.02,
526
+ "learning_rate": 1.1045284632676535e-05,
527
+ "loss": 1.6189,
528
+ "step": 14000
529
+ },
530
+ {
531
+ "epoch": 2.02,
532
+ "eval_gen_len": 6.267666666666667,
533
+ "eval_loss": 1.7743651866912842,
534
+ "eval_rouge1": 78.5967,
535
+ "eval_rouge2": 66.4212,
536
+ "eval_rougeL": 78.6125,
537
+ "eval_rougeLsum": 78.5741,
538
+ "eval_runtime": 1518.2132,
539
+ "eval_samples_per_second": 1.976,
540
+ "eval_steps_per_second": 0.494,
541
+ "step": 14000
542
+ },
543
+ {
544
+ "epoch": 2.09,
545
+ "learning_rate": 1.0523359562429441e-05,
546
+ "loss": 1.5746,
547
+ "step": 14500
548
+ },
549
+ {
550
+ "epoch": 2.09,
551
+ "eval_gen_len": 6.267666666666667,
552
+ "eval_loss": 1.7696142196655273,
553
+ "eval_rouge1": 78.7375,
554
+ "eval_rouge2": 67.0042,
555
+ "eval_rougeL": 78.7704,
556
+ "eval_rougeLsum": 78.6912,
557
+ "eval_runtime": 1504.5573,
558
+ "eval_samples_per_second": 1.994,
559
+ "eval_steps_per_second": 0.498,
560
+ "step": 14500
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 1e-05,
565
+ "loss": 1.5767,
566
+ "step": 15000
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "eval_gen_len": 6.267666666666667,
571
+ "eval_loss": 1.7685412168502808,
572
+ "eval_rouge1": 78.1583,
573
+ "eval_rouge2": 66.0184,
574
+ "eval_rougeL": 78.1256,
575
+ "eval_rougeLsum": 78.1273,
576
+ "eval_runtime": 1504.2381,
577
+ "eval_samples_per_second": 1.994,
578
+ "eval_steps_per_second": 0.499,
579
+ "step": 15000
580
+ }
581
+ ],
582
+ "logging_steps": 500,
583
+ "max_steps": 30000,
584
+ "num_train_epochs": 5,
585
+ "save_steps": 2500,
586
+ "total_flos": 6.170901674306002e+16,
587
+ "trial_name": null,
588
+ "trial_params": null
589
+ }
checkpoint-15000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-17500/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-17500/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-17500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b22d5595a6b6b2ce5850bea856c52ea5d51f7dadc41829440a11555634756ecf
3
+ size 3078042062
checkpoint-17500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ba7c2ad848fb38ee3bbdad2a75266c10d4fff522cb2fa8bdacf29c65a05ca36
3
+ size 1538969242
checkpoint-17500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3aff5959553392abd501b3f79410659c7c7572b63703f17ccfc9d3476abd5dff
3
+ size 14244
checkpoint-17500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46f2e688c027625ed6ebaf2a9dbd51b21a84d53bcc8869641b8b55a3efbd5578
3
+ size 1064
checkpoint-17500/trainer_state.json ADDED
@@ -0,0 +1,684 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.527623311908717,
5
+ "eval_steps": 500,
6
+ "global_step": 17500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ },
391
+ {
392
+ "epoch": 1.52,
393
+ "learning_rate": 1.4539904997395468e-05,
394
+ "loss": 1.6827,
395
+ "step": 10500
396
+ },
397
+ {
398
+ "epoch": 1.52,
399
+ "eval_gen_len": 6.267666666666667,
400
+ "eval_loss": 1.8478548526763916,
401
+ "eval_rouge1": 75.7687,
402
+ "eval_rouge2": 62.3363,
403
+ "eval_rougeL": 75.7808,
404
+ "eval_rougeLsum": 75.7533,
405
+ "eval_runtime": 1520.7572,
406
+ "eval_samples_per_second": 1.973,
407
+ "eval_steps_per_second": 0.493,
408
+ "step": 10500
409
+ },
410
+ {
411
+ "epoch": 1.59,
412
+ "learning_rate": 1.4067366430758004e-05,
413
+ "loss": 1.6808,
414
+ "step": 11000
415
+ },
416
+ {
417
+ "epoch": 1.59,
418
+ "eval_gen_len": 6.267666666666667,
419
+ "eval_loss": 1.8268691301345825,
420
+ "eval_rouge1": 76.5635,
421
+ "eval_rouge2": 63.5572,
422
+ "eval_rougeL": 76.5759,
423
+ "eval_rougeLsum": 76.5281,
424
+ "eval_runtime": 1529.0949,
425
+ "eval_samples_per_second": 1.962,
426
+ "eval_steps_per_second": 0.49,
427
+ "step": 11000
428
+ },
429
+ {
430
+ "epoch": 1.66,
431
+ "learning_rate": 1.3583679495453e-05,
432
+ "loss": 1.6747,
433
+ "step": 11500
434
+ },
435
+ {
436
+ "epoch": 1.66,
437
+ "eval_gen_len": 6.267666666666667,
438
+ "eval_loss": 1.8204782009124756,
439
+ "eval_rouge1": 76.0665,
440
+ "eval_rouge2": 62.7626,
441
+ "eval_rougeL": 76.0441,
442
+ "eval_rougeLsum": 76.0155,
443
+ "eval_runtime": 1521.0486,
444
+ "eval_samples_per_second": 1.972,
445
+ "eval_steps_per_second": 0.493,
446
+ "step": 11500
447
+ },
448
+ {
449
+ "epoch": 1.73,
450
+ "learning_rate": 1.3090169943749475e-05,
451
+ "loss": 1.6628,
452
+ "step": 12000
453
+ },
454
+ {
455
+ "epoch": 1.73,
456
+ "eval_gen_len": 6.267666666666667,
457
+ "eval_loss": 1.8059669733047485,
458
+ "eval_rouge1": 76.6687,
459
+ "eval_rouge2": 63.8953,
460
+ "eval_rougeL": 76.678,
461
+ "eval_rougeLsum": 76.6457,
462
+ "eval_runtime": 1519.6043,
463
+ "eval_samples_per_second": 1.974,
464
+ "eval_steps_per_second": 0.494,
465
+ "step": 12000
466
+ },
467
+ {
468
+ "epoch": 1.81,
469
+ "learning_rate": 1.2588190451025209e-05,
470
+ "loss": 1.6517,
471
+ "step": 12500
472
+ },
473
+ {
474
+ "epoch": 1.81,
475
+ "eval_gen_len": 6.267666666666667,
476
+ "eval_loss": 1.7973886728286743,
477
+ "eval_rouge1": 77.5432,
478
+ "eval_rouge2": 64.831,
479
+ "eval_rougeL": 77.5477,
480
+ "eval_rougeLsum": 77.5112,
481
+ "eval_runtime": 1514.3733,
482
+ "eval_samples_per_second": 1.981,
483
+ "eval_steps_per_second": 0.495,
484
+ "step": 12500
485
+ },
486
+ {
487
+ "epoch": 1.88,
488
+ "learning_rate": 1.2079116908177592e-05,
489
+ "loss": 1.6493,
490
+ "step": 13000
491
+ },
492
+ {
493
+ "epoch": 1.88,
494
+ "eval_gen_len": 6.267666666666667,
495
+ "eval_loss": 1.7867239713668823,
496
+ "eval_rouge1": 77.6195,
497
+ "eval_rouge2": 64.9393,
498
+ "eval_rougeL": 77.5908,
499
+ "eval_rougeLsum": 77.5498,
500
+ "eval_runtime": 1514.8609,
501
+ "eval_samples_per_second": 1.98,
502
+ "eval_steps_per_second": 0.495,
503
+ "step": 13000
504
+ },
505
+ {
506
+ "epoch": 1.95,
507
+ "learning_rate": 1.156434465040231e-05,
508
+ "loss": 1.6435,
509
+ "step": 13500
510
+ },
511
+ {
512
+ "epoch": 1.95,
513
+ "eval_gen_len": 6.267666666666667,
514
+ "eval_loss": 1.780102252960205,
515
+ "eval_rouge1": 77.8512,
516
+ "eval_rouge2": 65.2471,
517
+ "eval_rougeL": 77.8573,
518
+ "eval_rougeLsum": 77.8118,
519
+ "eval_runtime": 1516.9222,
520
+ "eval_samples_per_second": 1.978,
521
+ "eval_steps_per_second": 0.494,
522
+ "step": 13500
523
+ },
524
+ {
525
+ "epoch": 2.02,
526
+ "learning_rate": 1.1045284632676535e-05,
527
+ "loss": 1.6189,
528
+ "step": 14000
529
+ },
530
+ {
531
+ "epoch": 2.02,
532
+ "eval_gen_len": 6.267666666666667,
533
+ "eval_loss": 1.7743651866912842,
534
+ "eval_rouge1": 78.5967,
535
+ "eval_rouge2": 66.4212,
536
+ "eval_rougeL": 78.6125,
537
+ "eval_rougeLsum": 78.5741,
538
+ "eval_runtime": 1518.2132,
539
+ "eval_samples_per_second": 1.976,
540
+ "eval_steps_per_second": 0.494,
541
+ "step": 14000
542
+ },
543
+ {
544
+ "epoch": 2.09,
545
+ "learning_rate": 1.0523359562429441e-05,
546
+ "loss": 1.5746,
547
+ "step": 14500
548
+ },
549
+ {
550
+ "epoch": 2.09,
551
+ "eval_gen_len": 6.267666666666667,
552
+ "eval_loss": 1.7696142196655273,
553
+ "eval_rouge1": 78.7375,
554
+ "eval_rouge2": 67.0042,
555
+ "eval_rougeL": 78.7704,
556
+ "eval_rougeLsum": 78.6912,
557
+ "eval_runtime": 1504.5573,
558
+ "eval_samples_per_second": 1.994,
559
+ "eval_steps_per_second": 0.498,
560
+ "step": 14500
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 1e-05,
565
+ "loss": 1.5767,
566
+ "step": 15000
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "eval_gen_len": 6.267666666666667,
571
+ "eval_loss": 1.7685412168502808,
572
+ "eval_rouge1": 78.1583,
573
+ "eval_rouge2": 66.0184,
574
+ "eval_rougeL": 78.1256,
575
+ "eval_rougeLsum": 78.1273,
576
+ "eval_runtime": 1504.2381,
577
+ "eval_samples_per_second": 1.994,
578
+ "eval_steps_per_second": 0.499,
579
+ "step": 15000
580
+ },
581
+ {
582
+ "epoch": 2.24,
583
+ "learning_rate": 9.476640437570562e-06,
584
+ "loss": 1.5713,
585
+ "step": 15500
586
+ },
587
+ {
588
+ "epoch": 2.24,
589
+ "eval_gen_len": 6.267666666666667,
590
+ "eval_loss": 1.7626044750213623,
591
+ "eval_rouge1": 78.6909,
592
+ "eval_rouge2": 66.7862,
593
+ "eval_rougeL": 78.7168,
594
+ "eval_rougeLsum": 78.671,
595
+ "eval_runtime": 1509.684,
596
+ "eval_samples_per_second": 1.987,
597
+ "eval_steps_per_second": 0.497,
598
+ "step": 15500
599
+ },
600
+ {
601
+ "epoch": 2.31,
602
+ "learning_rate": 8.954715367323468e-06,
603
+ "loss": 1.5731,
604
+ "step": 16000
605
+ },
606
+ {
607
+ "epoch": 2.31,
608
+ "eval_gen_len": 6.267666666666667,
609
+ "eval_loss": 1.752744436264038,
610
+ "eval_rouge1": 78.9605,
611
+ "eval_rouge2": 67.084,
612
+ "eval_rougeL": 78.9504,
613
+ "eval_rougeLsum": 78.9289,
614
+ "eval_runtime": 1512.7838,
615
+ "eval_samples_per_second": 1.983,
616
+ "eval_steps_per_second": 0.496,
617
+ "step": 16000
618
+ },
619
+ {
620
+ "epoch": 2.38,
621
+ "learning_rate": 8.43565534959769e-06,
622
+ "loss": 1.5683,
623
+ "step": 16500
624
+ },
625
+ {
626
+ "epoch": 2.38,
627
+ "eval_gen_len": 6.267666666666667,
628
+ "eval_loss": 1.748329758644104,
629
+ "eval_rouge1": 79.0099,
630
+ "eval_rouge2": 67.1613,
631
+ "eval_rougeL": 79.0228,
632
+ "eval_rougeLsum": 79.0093,
633
+ "eval_runtime": 1516.6027,
634
+ "eval_samples_per_second": 1.978,
635
+ "eval_steps_per_second": 0.495,
636
+ "step": 16500
637
+ },
638
+ {
639
+ "epoch": 2.46,
640
+ "learning_rate": 7.92088309182241e-06,
641
+ "loss": 1.5684,
642
+ "step": 17000
643
+ },
644
+ {
645
+ "epoch": 2.46,
646
+ "eval_gen_len": 6.267666666666667,
647
+ "eval_loss": 1.7434362173080444,
648
+ "eval_rouge1": 78.9853,
649
+ "eval_rouge2": 67.0167,
650
+ "eval_rougeL": 78.9846,
651
+ "eval_rougeLsum": 78.9721,
652
+ "eval_runtime": 1523.8295,
653
+ "eval_samples_per_second": 1.969,
654
+ "eval_steps_per_second": 0.492,
655
+ "step": 17000
656
+ },
657
+ {
658
+ "epoch": 2.53,
659
+ "learning_rate": 7.411809548974792e-06,
660
+ "loss": 1.5653,
661
+ "step": 17500
662
+ },
663
+ {
664
+ "epoch": 2.53,
665
+ "eval_gen_len": 6.267666666666667,
666
+ "eval_loss": 1.7393466234207153,
667
+ "eval_rouge1": 79.413,
668
+ "eval_rouge2": 67.6295,
669
+ "eval_rougeL": 79.4249,
670
+ "eval_rougeLsum": 79.4027,
671
+ "eval_runtime": 1511.8366,
672
+ "eval_samples_per_second": 1.984,
673
+ "eval_steps_per_second": 0.496,
674
+ "step": 17500
675
+ }
676
+ ],
677
+ "logging_steps": 500,
678
+ "max_steps": 30000,
679
+ "num_train_epochs": 5,
680
+ "save_steps": 2500,
681
+ "total_flos": 7.200705320688427e+16,
682
+ "trial_name": null,
683
+ "trial_params": null
684
+ }
checkpoint-17500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-20000/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-20000/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-20000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60cda14d36a1663797cad9056ed621a9258c78b47f011f50831f00299a996e9e
3
+ size 3078042062
checkpoint-20000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a20d7a13c7a22af4be20468338bf7d925ab7ecd21a0498eb444bfe365cb3aa79
3
+ size 1538969242
checkpoint-20000/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23c03a4de3b86963ec287522a2181427c3374e2dc539cb28bf29e0a8cece77ca
3
+ size 14244
checkpoint-20000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c757df9d55d2304b27d457415a826c98701e7a2c5fbdf3ea465123fd1c57387
3
+ size 1064
checkpoint-20000/trainer_state.json ADDED
@@ -0,0 +1,779 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.8887123564671047,
5
+ "eval_steps": 500,
6
+ "global_step": 20000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ },
391
+ {
392
+ "epoch": 1.52,
393
+ "learning_rate": 1.4539904997395468e-05,
394
+ "loss": 1.6827,
395
+ "step": 10500
396
+ },
397
+ {
398
+ "epoch": 1.52,
399
+ "eval_gen_len": 6.267666666666667,
400
+ "eval_loss": 1.8478548526763916,
401
+ "eval_rouge1": 75.7687,
402
+ "eval_rouge2": 62.3363,
403
+ "eval_rougeL": 75.7808,
404
+ "eval_rougeLsum": 75.7533,
405
+ "eval_runtime": 1520.7572,
406
+ "eval_samples_per_second": 1.973,
407
+ "eval_steps_per_second": 0.493,
408
+ "step": 10500
409
+ },
410
+ {
411
+ "epoch": 1.59,
412
+ "learning_rate": 1.4067366430758004e-05,
413
+ "loss": 1.6808,
414
+ "step": 11000
415
+ },
416
+ {
417
+ "epoch": 1.59,
418
+ "eval_gen_len": 6.267666666666667,
419
+ "eval_loss": 1.8268691301345825,
420
+ "eval_rouge1": 76.5635,
421
+ "eval_rouge2": 63.5572,
422
+ "eval_rougeL": 76.5759,
423
+ "eval_rougeLsum": 76.5281,
424
+ "eval_runtime": 1529.0949,
425
+ "eval_samples_per_second": 1.962,
426
+ "eval_steps_per_second": 0.49,
427
+ "step": 11000
428
+ },
429
+ {
430
+ "epoch": 1.66,
431
+ "learning_rate": 1.3583679495453e-05,
432
+ "loss": 1.6747,
433
+ "step": 11500
434
+ },
435
+ {
436
+ "epoch": 1.66,
437
+ "eval_gen_len": 6.267666666666667,
438
+ "eval_loss": 1.8204782009124756,
439
+ "eval_rouge1": 76.0665,
440
+ "eval_rouge2": 62.7626,
441
+ "eval_rougeL": 76.0441,
442
+ "eval_rougeLsum": 76.0155,
443
+ "eval_runtime": 1521.0486,
444
+ "eval_samples_per_second": 1.972,
445
+ "eval_steps_per_second": 0.493,
446
+ "step": 11500
447
+ },
448
+ {
449
+ "epoch": 1.73,
450
+ "learning_rate": 1.3090169943749475e-05,
451
+ "loss": 1.6628,
452
+ "step": 12000
453
+ },
454
+ {
455
+ "epoch": 1.73,
456
+ "eval_gen_len": 6.267666666666667,
457
+ "eval_loss": 1.8059669733047485,
458
+ "eval_rouge1": 76.6687,
459
+ "eval_rouge2": 63.8953,
460
+ "eval_rougeL": 76.678,
461
+ "eval_rougeLsum": 76.6457,
462
+ "eval_runtime": 1519.6043,
463
+ "eval_samples_per_second": 1.974,
464
+ "eval_steps_per_second": 0.494,
465
+ "step": 12000
466
+ },
467
+ {
468
+ "epoch": 1.81,
469
+ "learning_rate": 1.2588190451025209e-05,
470
+ "loss": 1.6517,
471
+ "step": 12500
472
+ },
473
+ {
474
+ "epoch": 1.81,
475
+ "eval_gen_len": 6.267666666666667,
476
+ "eval_loss": 1.7973886728286743,
477
+ "eval_rouge1": 77.5432,
478
+ "eval_rouge2": 64.831,
479
+ "eval_rougeL": 77.5477,
480
+ "eval_rougeLsum": 77.5112,
481
+ "eval_runtime": 1514.3733,
482
+ "eval_samples_per_second": 1.981,
483
+ "eval_steps_per_second": 0.495,
484
+ "step": 12500
485
+ },
486
+ {
487
+ "epoch": 1.88,
488
+ "learning_rate": 1.2079116908177592e-05,
489
+ "loss": 1.6493,
490
+ "step": 13000
491
+ },
492
+ {
493
+ "epoch": 1.88,
494
+ "eval_gen_len": 6.267666666666667,
495
+ "eval_loss": 1.7867239713668823,
496
+ "eval_rouge1": 77.6195,
497
+ "eval_rouge2": 64.9393,
498
+ "eval_rougeL": 77.5908,
499
+ "eval_rougeLsum": 77.5498,
500
+ "eval_runtime": 1514.8609,
501
+ "eval_samples_per_second": 1.98,
502
+ "eval_steps_per_second": 0.495,
503
+ "step": 13000
504
+ },
505
+ {
506
+ "epoch": 1.95,
507
+ "learning_rate": 1.156434465040231e-05,
508
+ "loss": 1.6435,
509
+ "step": 13500
510
+ },
511
+ {
512
+ "epoch": 1.95,
513
+ "eval_gen_len": 6.267666666666667,
514
+ "eval_loss": 1.780102252960205,
515
+ "eval_rouge1": 77.8512,
516
+ "eval_rouge2": 65.2471,
517
+ "eval_rougeL": 77.8573,
518
+ "eval_rougeLsum": 77.8118,
519
+ "eval_runtime": 1516.9222,
520
+ "eval_samples_per_second": 1.978,
521
+ "eval_steps_per_second": 0.494,
522
+ "step": 13500
523
+ },
524
+ {
525
+ "epoch": 2.02,
526
+ "learning_rate": 1.1045284632676535e-05,
527
+ "loss": 1.6189,
528
+ "step": 14000
529
+ },
530
+ {
531
+ "epoch": 2.02,
532
+ "eval_gen_len": 6.267666666666667,
533
+ "eval_loss": 1.7743651866912842,
534
+ "eval_rouge1": 78.5967,
535
+ "eval_rouge2": 66.4212,
536
+ "eval_rougeL": 78.6125,
537
+ "eval_rougeLsum": 78.5741,
538
+ "eval_runtime": 1518.2132,
539
+ "eval_samples_per_second": 1.976,
540
+ "eval_steps_per_second": 0.494,
541
+ "step": 14000
542
+ },
543
+ {
544
+ "epoch": 2.09,
545
+ "learning_rate": 1.0523359562429441e-05,
546
+ "loss": 1.5746,
547
+ "step": 14500
548
+ },
549
+ {
550
+ "epoch": 2.09,
551
+ "eval_gen_len": 6.267666666666667,
552
+ "eval_loss": 1.7696142196655273,
553
+ "eval_rouge1": 78.7375,
554
+ "eval_rouge2": 67.0042,
555
+ "eval_rougeL": 78.7704,
556
+ "eval_rougeLsum": 78.6912,
557
+ "eval_runtime": 1504.5573,
558
+ "eval_samples_per_second": 1.994,
559
+ "eval_steps_per_second": 0.498,
560
+ "step": 14500
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 1e-05,
565
+ "loss": 1.5767,
566
+ "step": 15000
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "eval_gen_len": 6.267666666666667,
571
+ "eval_loss": 1.7685412168502808,
572
+ "eval_rouge1": 78.1583,
573
+ "eval_rouge2": 66.0184,
574
+ "eval_rougeL": 78.1256,
575
+ "eval_rougeLsum": 78.1273,
576
+ "eval_runtime": 1504.2381,
577
+ "eval_samples_per_second": 1.994,
578
+ "eval_steps_per_second": 0.499,
579
+ "step": 15000
580
+ },
581
+ {
582
+ "epoch": 2.24,
583
+ "learning_rate": 9.476640437570562e-06,
584
+ "loss": 1.5713,
585
+ "step": 15500
586
+ },
587
+ {
588
+ "epoch": 2.24,
589
+ "eval_gen_len": 6.267666666666667,
590
+ "eval_loss": 1.7626044750213623,
591
+ "eval_rouge1": 78.6909,
592
+ "eval_rouge2": 66.7862,
593
+ "eval_rougeL": 78.7168,
594
+ "eval_rougeLsum": 78.671,
595
+ "eval_runtime": 1509.684,
596
+ "eval_samples_per_second": 1.987,
597
+ "eval_steps_per_second": 0.497,
598
+ "step": 15500
599
+ },
600
+ {
601
+ "epoch": 2.31,
602
+ "learning_rate": 8.954715367323468e-06,
603
+ "loss": 1.5731,
604
+ "step": 16000
605
+ },
606
+ {
607
+ "epoch": 2.31,
608
+ "eval_gen_len": 6.267666666666667,
609
+ "eval_loss": 1.752744436264038,
610
+ "eval_rouge1": 78.9605,
611
+ "eval_rouge2": 67.084,
612
+ "eval_rougeL": 78.9504,
613
+ "eval_rougeLsum": 78.9289,
614
+ "eval_runtime": 1512.7838,
615
+ "eval_samples_per_second": 1.983,
616
+ "eval_steps_per_second": 0.496,
617
+ "step": 16000
618
+ },
619
+ {
620
+ "epoch": 2.38,
621
+ "learning_rate": 8.43565534959769e-06,
622
+ "loss": 1.5683,
623
+ "step": 16500
624
+ },
625
+ {
626
+ "epoch": 2.38,
627
+ "eval_gen_len": 6.267666666666667,
628
+ "eval_loss": 1.748329758644104,
629
+ "eval_rouge1": 79.0099,
630
+ "eval_rouge2": 67.1613,
631
+ "eval_rougeL": 79.0228,
632
+ "eval_rougeLsum": 79.0093,
633
+ "eval_runtime": 1516.6027,
634
+ "eval_samples_per_second": 1.978,
635
+ "eval_steps_per_second": 0.495,
636
+ "step": 16500
637
+ },
638
+ {
639
+ "epoch": 2.46,
640
+ "learning_rate": 7.92088309182241e-06,
641
+ "loss": 1.5684,
642
+ "step": 17000
643
+ },
644
+ {
645
+ "epoch": 2.46,
646
+ "eval_gen_len": 6.267666666666667,
647
+ "eval_loss": 1.7434362173080444,
648
+ "eval_rouge1": 78.9853,
649
+ "eval_rouge2": 67.0167,
650
+ "eval_rougeL": 78.9846,
651
+ "eval_rougeLsum": 78.9721,
652
+ "eval_runtime": 1523.8295,
653
+ "eval_samples_per_second": 1.969,
654
+ "eval_steps_per_second": 0.492,
655
+ "step": 17000
656
+ },
657
+ {
658
+ "epoch": 2.53,
659
+ "learning_rate": 7.411809548974792e-06,
660
+ "loss": 1.5653,
661
+ "step": 17500
662
+ },
663
+ {
664
+ "epoch": 2.53,
665
+ "eval_gen_len": 6.267666666666667,
666
+ "eval_loss": 1.7393466234207153,
667
+ "eval_rouge1": 79.413,
668
+ "eval_rouge2": 67.6295,
669
+ "eval_rougeL": 79.4249,
670
+ "eval_rougeLsum": 79.4027,
671
+ "eval_runtime": 1511.8366,
672
+ "eval_samples_per_second": 1.984,
673
+ "eval_steps_per_second": 0.496,
674
+ "step": 17500
675
+ },
676
+ {
677
+ "epoch": 2.6,
678
+ "learning_rate": 6.909830056250527e-06,
679
+ "loss": 1.5584,
680
+ "step": 18000
681
+ },
682
+ {
683
+ "epoch": 2.6,
684
+ "eval_gen_len": 6.267666666666667,
685
+ "eval_loss": 1.735812783241272,
686
+ "eval_rouge1": 79.3295,
687
+ "eval_rouge2": 67.4641,
688
+ "eval_rougeL": 79.3042,
689
+ "eval_rougeLsum": 79.3028,
690
+ "eval_runtime": 1515.6868,
691
+ "eval_samples_per_second": 1.979,
692
+ "eval_steps_per_second": 0.495,
693
+ "step": 18000
694
+ },
695
+ {
696
+ "epoch": 2.67,
697
+ "learning_rate": 6.4163205045469975e-06,
698
+ "loss": 1.5567,
699
+ "step": 18500
700
+ },
701
+ {
702
+ "epoch": 2.67,
703
+ "eval_gen_len": 6.267666666666667,
704
+ "eval_loss": 1.7302848100662231,
705
+ "eval_rouge1": 80.0165,
706
+ "eval_rouge2": 68.3289,
707
+ "eval_rougeL": 80.0436,
708
+ "eval_rougeLsum": 80.019,
709
+ "eval_runtime": 1522.3815,
710
+ "eval_samples_per_second": 1.971,
711
+ "eval_steps_per_second": 0.493,
712
+ "step": 18500
713
+ },
714
+ {
715
+ "epoch": 2.74,
716
+ "learning_rate": 5.932633569242e-06,
717
+ "loss": 1.5564,
718
+ "step": 19000
719
+ },
720
+ {
721
+ "epoch": 2.74,
722
+ "eval_gen_len": 6.267666666666667,
723
+ "eval_loss": 1.7278697490692139,
724
+ "eval_rouge1": 79.8733,
725
+ "eval_rouge2": 68.1584,
726
+ "eval_rougeL": 79.868,
727
+ "eval_rougeLsum": 79.8511,
728
+ "eval_runtime": 1516.0739,
729
+ "eval_samples_per_second": 1.979,
730
+ "eval_steps_per_second": 0.495,
731
+ "step": 19000
732
+ },
733
+ {
734
+ "epoch": 2.82,
735
+ "learning_rate": 5.460095002604533e-06,
736
+ "loss": 1.553,
737
+ "step": 19500
738
+ },
739
+ {
740
+ "epoch": 2.82,
741
+ "eval_gen_len": 6.267666666666667,
742
+ "eval_loss": 1.7226474285125732,
743
+ "eval_rouge1": 79.9727,
744
+ "eval_rouge2": 68.2612,
745
+ "eval_rougeL": 79.9901,
746
+ "eval_rougeLsum": 79.9554,
747
+ "eval_runtime": 1525.8223,
748
+ "eval_samples_per_second": 1.966,
749
+ "eval_steps_per_second": 0.492,
750
+ "step": 19500
751
+ },
752
+ {
753
+ "epoch": 2.89,
754
+ "learning_rate": 5.000000000000003e-06,
755
+ "loss": 1.5498,
756
+ "step": 20000
757
+ },
758
+ {
759
+ "epoch": 2.89,
760
+ "eval_gen_len": 6.267666666666667,
761
+ "eval_loss": 1.720814824104309,
762
+ "eval_rouge1": 79.9937,
763
+ "eval_rouge2": 68.221,
764
+ "eval_rougeL": 79.9703,
765
+ "eval_rougeLsum": 79.9696,
766
+ "eval_runtime": 1519.3409,
767
+ "eval_samples_per_second": 1.975,
768
+ "eval_steps_per_second": 0.494,
769
+ "step": 20000
770
+ }
771
+ ],
772
+ "logging_steps": 500,
773
+ "max_steps": 30000,
774
+ "num_train_epochs": 5,
775
+ "save_steps": 2500,
776
+ "total_flos": 8.228710189060075e+16,
777
+ "trial_name": null,
778
+ "trial_params": null
779
+ }
checkpoint-20000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-22500/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }
checkpoint-22500/generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 30522,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.33.1"
7
+ }
checkpoint-22500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e5e92c16b59440226cfa5763559e07af79bdacd9bfd9748b1378353caa1f33f
3
+ size 3078042062
checkpoint-22500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffa5247febe0a8a38bf74db793ce3230f6892f44be9a2de37106de2033b27b89
3
+ size 1538969242
checkpoint-22500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76c0b77b75b104d52df7208449e713d15585ed674305a174ce34842d80400b09
3
+ size 14244
checkpoint-22500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32ee6664ee727851cfca6e4938df7a940cb94828611adc0d09a73350ad5ca022
3
+ size 1064
checkpoint-22500/trainer_state.json ADDED
@@ -0,0 +1,874 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.249801401025493,
5
+ "eval_steps": 500,
6
+ "global_step": 22500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.07,
13
+ "learning_rate": 1.9986295347545738e-05,
14
+ "loss": 4.4247,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.07,
19
+ "eval_gen_len": 6.267,
20
+ "eval_loss": 3.5801377296447754,
21
+ "eval_rouge1": 27.9882,
22
+ "eval_rouge2": 8.9331,
23
+ "eval_rougeL": 27.6011,
24
+ "eval_rougeLsum": 27.5353,
25
+ "eval_runtime": 1554.6657,
26
+ "eval_samples_per_second": 1.93,
27
+ "eval_steps_per_second": 0.482,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 0.14,
32
+ "learning_rate": 1.9945218953682736e-05,
33
+ "loss": 2.959,
34
+ "step": 1000
35
+ },
36
+ {
37
+ "epoch": 0.14,
38
+ "eval_gen_len": 6.267333333333333,
39
+ "eval_loss": 3.238922119140625,
40
+ "eval_rouge1": 34.4913,
41
+ "eval_rouge2": 14.1739,
42
+ "eval_rougeL": 34.2505,
43
+ "eval_rougeLsum": 34.1717,
44
+ "eval_runtime": 1549.8046,
45
+ "eval_samples_per_second": 1.936,
46
+ "eval_steps_per_second": 0.484,
47
+ "step": 1000
48
+ },
49
+ {
50
+ "epoch": 0.22,
51
+ "learning_rate": 1.9876883405951378e-05,
52
+ "loss": 2.7367,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 0.22,
57
+ "eval_gen_len": 6.267666666666667,
58
+ "eval_loss": 2.983262300491333,
59
+ "eval_rouge1": 40.5052,
60
+ "eval_rouge2": 18.725,
61
+ "eval_rougeL": 40.1333,
62
+ "eval_rougeLsum": 40.0554,
63
+ "eval_runtime": 1542.1597,
64
+ "eval_samples_per_second": 1.945,
65
+ "eval_steps_per_second": 0.486,
66
+ "step": 1500
67
+ },
68
+ {
69
+ "epoch": 0.29,
70
+ "learning_rate": 1.9781476007338058e-05,
71
+ "loss": 2.553,
72
+ "step": 2000
73
+ },
74
+ {
75
+ "epoch": 0.29,
76
+ "eval_gen_len": 6.267333333333333,
77
+ "eval_loss": 2.781419277191162,
78
+ "eval_rouge1": 45.972,
79
+ "eval_rouge2": 24.75,
80
+ "eval_rougeL": 45.6989,
81
+ "eval_rougeLsum": 45.6257,
82
+ "eval_runtime": 1547.6212,
83
+ "eval_samples_per_second": 1.938,
84
+ "eval_steps_per_second": 0.485,
85
+ "step": 2000
86
+ },
87
+ {
88
+ "epoch": 0.36,
89
+ "learning_rate": 1.9659258262890683e-05,
90
+ "loss": 2.3988,
91
+ "step": 2500
92
+ },
93
+ {
94
+ "epoch": 0.36,
95
+ "eval_gen_len": 6.267666666666667,
96
+ "eval_loss": 2.6185145378112793,
97
+ "eval_rouge1": 51.3237,
98
+ "eval_rouge2": 30.8584,
99
+ "eval_rougeL": 51.0697,
100
+ "eval_rougeLsum": 50.9947,
101
+ "eval_runtime": 1538.5542,
102
+ "eval_samples_per_second": 1.95,
103
+ "eval_steps_per_second": 0.487,
104
+ "step": 2500
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 1.9510565162951538e-05,
109
+ "loss": 2.2788,
110
+ "step": 3000
111
+ },
112
+ {
113
+ "epoch": 0.43,
114
+ "eval_gen_len": 6.267666666666667,
115
+ "eval_loss": 2.4821877479553223,
116
+ "eval_rouge1": 55.054,
117
+ "eval_rouge2": 35.9558,
118
+ "eval_rougeL": 54.885,
119
+ "eval_rougeLsum": 54.8263,
120
+ "eval_runtime": 1535.9732,
121
+ "eval_samples_per_second": 1.953,
122
+ "eval_steps_per_second": 0.488,
123
+ "step": 3000
124
+ },
125
+ {
126
+ "epoch": 0.51,
127
+ "learning_rate": 1.9335804264972018e-05,
128
+ "loss": 2.185,
129
+ "step": 3500
130
+ },
131
+ {
132
+ "epoch": 0.51,
133
+ "eval_gen_len": 6.267666666666667,
134
+ "eval_loss": 2.372607469558716,
135
+ "eval_rouge1": 58.4125,
136
+ "eval_rouge2": 39.7017,
137
+ "eval_rougeL": 58.2864,
138
+ "eval_rougeLsum": 58.2323,
139
+ "eval_runtime": 1555.498,
140
+ "eval_samples_per_second": 1.929,
141
+ "eval_steps_per_second": 0.482,
142
+ "step": 3500
143
+ },
144
+ {
145
+ "epoch": 0.58,
146
+ "learning_rate": 1.913545457642601e-05,
147
+ "loss": 2.1024,
148
+ "step": 4000
149
+ },
150
+ {
151
+ "epoch": 0.58,
152
+ "eval_gen_len": 6.267666666666667,
153
+ "eval_loss": 2.2966315746307373,
154
+ "eval_rouge1": 60.2269,
155
+ "eval_rouge2": 42.343,
156
+ "eval_rougeL": 60.1064,
157
+ "eval_rougeLsum": 60.015,
158
+ "eval_runtime": 1539.5016,
159
+ "eval_samples_per_second": 1.949,
160
+ "eval_steps_per_second": 0.487,
161
+ "step": 4000
162
+ },
163
+ {
164
+ "epoch": 0.65,
165
+ "learning_rate": 1.891006524188368e-05,
166
+ "loss": 2.0395,
167
+ "step": 4500
168
+ },
169
+ {
170
+ "epoch": 0.65,
171
+ "eval_gen_len": 6.267666666666667,
172
+ "eval_loss": 2.2079367637634277,
173
+ "eval_rouge1": 63.9442,
174
+ "eval_rouge2": 47.1262,
175
+ "eval_rougeL": 63.8226,
176
+ "eval_rougeLsum": 63.783,
177
+ "eval_runtime": 1534.6718,
178
+ "eval_samples_per_second": 1.955,
179
+ "eval_steps_per_second": 0.489,
180
+ "step": 4500
181
+ },
182
+ {
183
+ "epoch": 0.72,
184
+ "learning_rate": 1.866025403784439e-05,
185
+ "loss": 1.9929,
186
+ "step": 5000
187
+ },
188
+ {
189
+ "epoch": 0.72,
190
+ "eval_gen_len": 6.267666666666667,
191
+ "eval_loss": 2.142320156097412,
192
+ "eval_rouge1": 66.0535,
193
+ "eval_rouge2": 49.403,
194
+ "eval_rougeL": 65.9837,
195
+ "eval_rougeLsum": 65.9295,
196
+ "eval_runtime": 1531.6986,
197
+ "eval_samples_per_second": 1.959,
198
+ "eval_steps_per_second": 0.49,
199
+ "step": 5000
200
+ },
201
+ {
202
+ "epoch": 0.79,
203
+ "learning_rate": 1.8386705679454243e-05,
204
+ "loss": 1.944,
205
+ "step": 5500
206
+ },
207
+ {
208
+ "epoch": 0.79,
209
+ "eval_gen_len": 6.267666666666667,
210
+ "eval_loss": 2.094369888305664,
211
+ "eval_rouge1": 67.4662,
212
+ "eval_rouge2": 50.9242,
213
+ "eval_rougeL": 67.3906,
214
+ "eval_rougeLsum": 67.3416,
215
+ "eval_runtime": 1527.1607,
216
+ "eval_samples_per_second": 1.964,
217
+ "eval_steps_per_second": 0.491,
218
+ "step": 5500
219
+ },
220
+ {
221
+ "epoch": 0.87,
222
+ "learning_rate": 1.8090169943749477e-05,
223
+ "loss": 1.902,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 0.87,
228
+ "eval_gen_len": 6.267666666666667,
229
+ "eval_loss": 2.0382838249206543,
230
+ "eval_rouge1": 69.3267,
231
+ "eval_rouge2": 53.8395,
232
+ "eval_rougeL": 69.2759,
233
+ "eval_rougeLsum": 69.2151,
234
+ "eval_runtime": 1532.3017,
235
+ "eval_samples_per_second": 1.958,
236
+ "eval_steps_per_second": 0.489,
237
+ "step": 6000
238
+ },
239
+ {
240
+ "epoch": 0.94,
241
+ "learning_rate": 1.777145961456971e-05,
242
+ "loss": 1.8712,
243
+ "step": 6500
244
+ },
245
+ {
246
+ "epoch": 0.94,
247
+ "eval_gen_len": 6.267666666666667,
248
+ "eval_loss": 2.010657548904419,
249
+ "eval_rouge1": 70.2771,
250
+ "eval_rouge2": 54.5208,
251
+ "eval_rougeL": 70.2492,
252
+ "eval_rougeLsum": 70.2095,
253
+ "eval_runtime": 1563.3117,
254
+ "eval_samples_per_second": 1.919,
255
+ "eval_steps_per_second": 0.48,
256
+ "step": 6500
257
+ },
258
+ {
259
+ "epoch": 1.01,
260
+ "learning_rate": 1.7431448254773943e-05,
261
+ "loss": 1.8387,
262
+ "step": 7000
263
+ },
264
+ {
265
+ "epoch": 1.01,
266
+ "eval_gen_len": 6.267666666666667,
267
+ "eval_loss": 1.9761910438537598,
268
+ "eval_rouge1": 71.145,
269
+ "eval_rouge2": 56.4319,
270
+ "eval_rougeL": 71.1008,
271
+ "eval_rougeLsum": 71.0876,
272
+ "eval_runtime": 1552.4923,
273
+ "eval_samples_per_second": 1.932,
274
+ "eval_steps_per_second": 0.483,
275
+ "step": 7000
276
+ },
277
+ {
278
+ "epoch": 1.08,
279
+ "learning_rate": 1.7071067811865477e-05,
280
+ "loss": 1.7558,
281
+ "step": 7500
282
+ },
283
+ {
284
+ "epoch": 1.08,
285
+ "eval_gen_len": 6.267666666666667,
286
+ "eval_loss": 1.9575979709625244,
287
+ "eval_rouge1": 72.527,
288
+ "eval_rouge2": 58.0461,
289
+ "eval_rougeL": 72.4801,
290
+ "eval_rougeLsum": 72.461,
291
+ "eval_runtime": 1522.5751,
292
+ "eval_samples_per_second": 1.97,
293
+ "eval_steps_per_second": 0.493,
294
+ "step": 7500
295
+ },
296
+ {
297
+ "epoch": 1.16,
298
+ "learning_rate": 1.6691306063588583e-05,
299
+ "loss": 1.7363,
300
+ "step": 8000
301
+ },
302
+ {
303
+ "epoch": 1.16,
304
+ "eval_gen_len": 6.267666666666667,
305
+ "eval_loss": 1.9305455684661865,
306
+ "eval_rouge1": 73.3884,
307
+ "eval_rouge2": 59.6248,
308
+ "eval_rougeL": 73.3396,
309
+ "eval_rougeLsum": 73.3362,
310
+ "eval_runtime": 1541.3792,
311
+ "eval_samples_per_second": 1.946,
312
+ "eval_steps_per_second": 0.487,
313
+ "step": 8000
314
+ },
315
+ {
316
+ "epoch": 1.23,
317
+ "learning_rate": 1.6293203910498375e-05,
318
+ "loss": 1.7245,
319
+ "step": 8500
320
+ },
321
+ {
322
+ "epoch": 1.23,
323
+ "eval_gen_len": 6.267666666666667,
324
+ "eval_loss": 1.9158198833465576,
325
+ "eval_rouge1": 73.3565,
326
+ "eval_rouge2": 58.7585,
327
+ "eval_rougeL": 73.3541,
328
+ "eval_rougeLsum": 73.3668,
329
+ "eval_runtime": 1546.9117,
330
+ "eval_samples_per_second": 1.939,
331
+ "eval_steps_per_second": 0.485,
332
+ "step": 8500
333
+ },
334
+ {
335
+ "epoch": 1.3,
336
+ "learning_rate": 1.5877852522924733e-05,
337
+ "loss": 1.7147,
338
+ "step": 9000
339
+ },
340
+ {
341
+ "epoch": 1.3,
342
+ "eval_gen_len": 6.267666666666667,
343
+ "eval_loss": 1.898065447807312,
344
+ "eval_rouge1": 74.1688,
345
+ "eval_rouge2": 59.9465,
346
+ "eval_rougeL": 74.1934,
347
+ "eval_rougeLsum": 74.1447,
348
+ "eval_runtime": 1537.688,
349
+ "eval_samples_per_second": 1.951,
350
+ "eval_steps_per_second": 0.488,
351
+ "step": 9000
352
+ },
353
+ {
354
+ "epoch": 1.37,
355
+ "learning_rate": 1.5446390350150272e-05,
356
+ "loss": 1.7013,
357
+ "step": 9500
358
+ },
359
+ {
360
+ "epoch": 1.37,
361
+ "eval_gen_len": 6.267666666666667,
362
+ "eval_loss": 1.8747327327728271,
363
+ "eval_rouge1": 75.1266,
364
+ "eval_rouge2": 61.394,
365
+ "eval_rougeL": 75.128,
366
+ "eval_rougeLsum": 75.0856,
367
+ "eval_runtime": 1526.768,
368
+ "eval_samples_per_second": 1.965,
369
+ "eval_steps_per_second": 0.491,
370
+ "step": 9500
371
+ },
372
+ {
373
+ "epoch": 1.44,
374
+ "learning_rate": 1.5000000000000002e-05,
375
+ "loss": 1.6906,
376
+ "step": 10000
377
+ },
378
+ {
379
+ "epoch": 1.44,
380
+ "eval_gen_len": 6.267666666666667,
381
+ "eval_loss": 1.8567513227462769,
382
+ "eval_rouge1": 75.8503,
383
+ "eval_rouge2": 62.2772,
384
+ "eval_rougeL": 75.8449,
385
+ "eval_rougeLsum": 75.8138,
386
+ "eval_runtime": 1518.2689,
387
+ "eval_samples_per_second": 1.976,
388
+ "eval_steps_per_second": 0.494,
389
+ "step": 10000
390
+ },
391
+ {
392
+ "epoch": 1.52,
393
+ "learning_rate": 1.4539904997395468e-05,
394
+ "loss": 1.6827,
395
+ "step": 10500
396
+ },
397
+ {
398
+ "epoch": 1.52,
399
+ "eval_gen_len": 6.267666666666667,
400
+ "eval_loss": 1.8478548526763916,
401
+ "eval_rouge1": 75.7687,
402
+ "eval_rouge2": 62.3363,
403
+ "eval_rougeL": 75.7808,
404
+ "eval_rougeLsum": 75.7533,
405
+ "eval_runtime": 1520.7572,
406
+ "eval_samples_per_second": 1.973,
407
+ "eval_steps_per_second": 0.493,
408
+ "step": 10500
409
+ },
410
+ {
411
+ "epoch": 1.59,
412
+ "learning_rate": 1.4067366430758004e-05,
413
+ "loss": 1.6808,
414
+ "step": 11000
415
+ },
416
+ {
417
+ "epoch": 1.59,
418
+ "eval_gen_len": 6.267666666666667,
419
+ "eval_loss": 1.8268691301345825,
420
+ "eval_rouge1": 76.5635,
421
+ "eval_rouge2": 63.5572,
422
+ "eval_rougeL": 76.5759,
423
+ "eval_rougeLsum": 76.5281,
424
+ "eval_runtime": 1529.0949,
425
+ "eval_samples_per_second": 1.962,
426
+ "eval_steps_per_second": 0.49,
427
+ "step": 11000
428
+ },
429
+ {
430
+ "epoch": 1.66,
431
+ "learning_rate": 1.3583679495453e-05,
432
+ "loss": 1.6747,
433
+ "step": 11500
434
+ },
435
+ {
436
+ "epoch": 1.66,
437
+ "eval_gen_len": 6.267666666666667,
438
+ "eval_loss": 1.8204782009124756,
439
+ "eval_rouge1": 76.0665,
440
+ "eval_rouge2": 62.7626,
441
+ "eval_rougeL": 76.0441,
442
+ "eval_rougeLsum": 76.0155,
443
+ "eval_runtime": 1521.0486,
444
+ "eval_samples_per_second": 1.972,
445
+ "eval_steps_per_second": 0.493,
446
+ "step": 11500
447
+ },
448
+ {
449
+ "epoch": 1.73,
450
+ "learning_rate": 1.3090169943749475e-05,
451
+ "loss": 1.6628,
452
+ "step": 12000
453
+ },
454
+ {
455
+ "epoch": 1.73,
456
+ "eval_gen_len": 6.267666666666667,
457
+ "eval_loss": 1.8059669733047485,
458
+ "eval_rouge1": 76.6687,
459
+ "eval_rouge2": 63.8953,
460
+ "eval_rougeL": 76.678,
461
+ "eval_rougeLsum": 76.6457,
462
+ "eval_runtime": 1519.6043,
463
+ "eval_samples_per_second": 1.974,
464
+ "eval_steps_per_second": 0.494,
465
+ "step": 12000
466
+ },
467
+ {
468
+ "epoch": 1.81,
469
+ "learning_rate": 1.2588190451025209e-05,
470
+ "loss": 1.6517,
471
+ "step": 12500
472
+ },
473
+ {
474
+ "epoch": 1.81,
475
+ "eval_gen_len": 6.267666666666667,
476
+ "eval_loss": 1.7973886728286743,
477
+ "eval_rouge1": 77.5432,
478
+ "eval_rouge2": 64.831,
479
+ "eval_rougeL": 77.5477,
480
+ "eval_rougeLsum": 77.5112,
481
+ "eval_runtime": 1514.3733,
482
+ "eval_samples_per_second": 1.981,
483
+ "eval_steps_per_second": 0.495,
484
+ "step": 12500
485
+ },
486
+ {
487
+ "epoch": 1.88,
488
+ "learning_rate": 1.2079116908177592e-05,
489
+ "loss": 1.6493,
490
+ "step": 13000
491
+ },
492
+ {
493
+ "epoch": 1.88,
494
+ "eval_gen_len": 6.267666666666667,
495
+ "eval_loss": 1.7867239713668823,
496
+ "eval_rouge1": 77.6195,
497
+ "eval_rouge2": 64.9393,
498
+ "eval_rougeL": 77.5908,
499
+ "eval_rougeLsum": 77.5498,
500
+ "eval_runtime": 1514.8609,
501
+ "eval_samples_per_second": 1.98,
502
+ "eval_steps_per_second": 0.495,
503
+ "step": 13000
504
+ },
505
+ {
506
+ "epoch": 1.95,
507
+ "learning_rate": 1.156434465040231e-05,
508
+ "loss": 1.6435,
509
+ "step": 13500
510
+ },
511
+ {
512
+ "epoch": 1.95,
513
+ "eval_gen_len": 6.267666666666667,
514
+ "eval_loss": 1.780102252960205,
515
+ "eval_rouge1": 77.8512,
516
+ "eval_rouge2": 65.2471,
517
+ "eval_rougeL": 77.8573,
518
+ "eval_rougeLsum": 77.8118,
519
+ "eval_runtime": 1516.9222,
520
+ "eval_samples_per_second": 1.978,
521
+ "eval_steps_per_second": 0.494,
522
+ "step": 13500
523
+ },
524
+ {
525
+ "epoch": 2.02,
526
+ "learning_rate": 1.1045284632676535e-05,
527
+ "loss": 1.6189,
528
+ "step": 14000
529
+ },
530
+ {
531
+ "epoch": 2.02,
532
+ "eval_gen_len": 6.267666666666667,
533
+ "eval_loss": 1.7743651866912842,
534
+ "eval_rouge1": 78.5967,
535
+ "eval_rouge2": 66.4212,
536
+ "eval_rougeL": 78.6125,
537
+ "eval_rougeLsum": 78.5741,
538
+ "eval_runtime": 1518.2132,
539
+ "eval_samples_per_second": 1.976,
540
+ "eval_steps_per_second": 0.494,
541
+ "step": 14000
542
+ },
543
+ {
544
+ "epoch": 2.09,
545
+ "learning_rate": 1.0523359562429441e-05,
546
+ "loss": 1.5746,
547
+ "step": 14500
548
+ },
549
+ {
550
+ "epoch": 2.09,
551
+ "eval_gen_len": 6.267666666666667,
552
+ "eval_loss": 1.7696142196655273,
553
+ "eval_rouge1": 78.7375,
554
+ "eval_rouge2": 67.0042,
555
+ "eval_rougeL": 78.7704,
556
+ "eval_rougeLsum": 78.6912,
557
+ "eval_runtime": 1504.5573,
558
+ "eval_samples_per_second": 1.994,
559
+ "eval_steps_per_second": 0.498,
560
+ "step": 14500
561
+ },
562
+ {
563
+ "epoch": 2.17,
564
+ "learning_rate": 1e-05,
565
+ "loss": 1.5767,
566
+ "step": 15000
567
+ },
568
+ {
569
+ "epoch": 2.17,
570
+ "eval_gen_len": 6.267666666666667,
571
+ "eval_loss": 1.7685412168502808,
572
+ "eval_rouge1": 78.1583,
573
+ "eval_rouge2": 66.0184,
574
+ "eval_rougeL": 78.1256,
575
+ "eval_rougeLsum": 78.1273,
576
+ "eval_runtime": 1504.2381,
577
+ "eval_samples_per_second": 1.994,
578
+ "eval_steps_per_second": 0.499,
579
+ "step": 15000
580
+ },
581
+ {
582
+ "epoch": 2.24,
583
+ "learning_rate": 9.476640437570562e-06,
584
+ "loss": 1.5713,
585
+ "step": 15500
586
+ },
587
+ {
588
+ "epoch": 2.24,
589
+ "eval_gen_len": 6.267666666666667,
590
+ "eval_loss": 1.7626044750213623,
591
+ "eval_rouge1": 78.6909,
592
+ "eval_rouge2": 66.7862,
593
+ "eval_rougeL": 78.7168,
594
+ "eval_rougeLsum": 78.671,
595
+ "eval_runtime": 1509.684,
596
+ "eval_samples_per_second": 1.987,
597
+ "eval_steps_per_second": 0.497,
598
+ "step": 15500
599
+ },
600
+ {
601
+ "epoch": 2.31,
602
+ "learning_rate": 8.954715367323468e-06,
603
+ "loss": 1.5731,
604
+ "step": 16000
605
+ },
606
+ {
607
+ "epoch": 2.31,
608
+ "eval_gen_len": 6.267666666666667,
609
+ "eval_loss": 1.752744436264038,
610
+ "eval_rouge1": 78.9605,
611
+ "eval_rouge2": 67.084,
612
+ "eval_rougeL": 78.9504,
613
+ "eval_rougeLsum": 78.9289,
614
+ "eval_runtime": 1512.7838,
615
+ "eval_samples_per_second": 1.983,
616
+ "eval_steps_per_second": 0.496,
617
+ "step": 16000
618
+ },
619
+ {
620
+ "epoch": 2.38,
621
+ "learning_rate": 8.43565534959769e-06,
622
+ "loss": 1.5683,
623
+ "step": 16500
624
+ },
625
+ {
626
+ "epoch": 2.38,
627
+ "eval_gen_len": 6.267666666666667,
628
+ "eval_loss": 1.748329758644104,
629
+ "eval_rouge1": 79.0099,
630
+ "eval_rouge2": 67.1613,
631
+ "eval_rougeL": 79.0228,
632
+ "eval_rougeLsum": 79.0093,
633
+ "eval_runtime": 1516.6027,
634
+ "eval_samples_per_second": 1.978,
635
+ "eval_steps_per_second": 0.495,
636
+ "step": 16500
637
+ },
638
+ {
639
+ "epoch": 2.46,
640
+ "learning_rate": 7.92088309182241e-06,
641
+ "loss": 1.5684,
642
+ "step": 17000
643
+ },
644
+ {
645
+ "epoch": 2.46,
646
+ "eval_gen_len": 6.267666666666667,
647
+ "eval_loss": 1.7434362173080444,
648
+ "eval_rouge1": 78.9853,
649
+ "eval_rouge2": 67.0167,
650
+ "eval_rougeL": 78.9846,
651
+ "eval_rougeLsum": 78.9721,
652
+ "eval_runtime": 1523.8295,
653
+ "eval_samples_per_second": 1.969,
654
+ "eval_steps_per_second": 0.492,
655
+ "step": 17000
656
+ },
657
+ {
658
+ "epoch": 2.53,
659
+ "learning_rate": 7.411809548974792e-06,
660
+ "loss": 1.5653,
661
+ "step": 17500
662
+ },
663
+ {
664
+ "epoch": 2.53,
665
+ "eval_gen_len": 6.267666666666667,
666
+ "eval_loss": 1.7393466234207153,
667
+ "eval_rouge1": 79.413,
668
+ "eval_rouge2": 67.6295,
669
+ "eval_rougeL": 79.4249,
670
+ "eval_rougeLsum": 79.4027,
671
+ "eval_runtime": 1511.8366,
672
+ "eval_samples_per_second": 1.984,
673
+ "eval_steps_per_second": 0.496,
674
+ "step": 17500
675
+ },
676
+ {
677
+ "epoch": 2.6,
678
+ "learning_rate": 6.909830056250527e-06,
679
+ "loss": 1.5584,
680
+ "step": 18000
681
+ },
682
+ {
683
+ "epoch": 2.6,
684
+ "eval_gen_len": 6.267666666666667,
685
+ "eval_loss": 1.735812783241272,
686
+ "eval_rouge1": 79.3295,
687
+ "eval_rouge2": 67.4641,
688
+ "eval_rougeL": 79.3042,
689
+ "eval_rougeLsum": 79.3028,
690
+ "eval_runtime": 1515.6868,
691
+ "eval_samples_per_second": 1.979,
692
+ "eval_steps_per_second": 0.495,
693
+ "step": 18000
694
+ },
695
+ {
696
+ "epoch": 2.67,
697
+ "learning_rate": 6.4163205045469975e-06,
698
+ "loss": 1.5567,
699
+ "step": 18500
700
+ },
701
+ {
702
+ "epoch": 2.67,
703
+ "eval_gen_len": 6.267666666666667,
704
+ "eval_loss": 1.7302848100662231,
705
+ "eval_rouge1": 80.0165,
706
+ "eval_rouge2": 68.3289,
707
+ "eval_rougeL": 80.0436,
708
+ "eval_rougeLsum": 80.019,
709
+ "eval_runtime": 1522.3815,
710
+ "eval_samples_per_second": 1.971,
711
+ "eval_steps_per_second": 0.493,
712
+ "step": 18500
713
+ },
714
+ {
715
+ "epoch": 2.74,
716
+ "learning_rate": 5.932633569242e-06,
717
+ "loss": 1.5564,
718
+ "step": 19000
719
+ },
720
+ {
721
+ "epoch": 2.74,
722
+ "eval_gen_len": 6.267666666666667,
723
+ "eval_loss": 1.7278697490692139,
724
+ "eval_rouge1": 79.8733,
725
+ "eval_rouge2": 68.1584,
726
+ "eval_rougeL": 79.868,
727
+ "eval_rougeLsum": 79.8511,
728
+ "eval_runtime": 1516.0739,
729
+ "eval_samples_per_second": 1.979,
730
+ "eval_steps_per_second": 0.495,
731
+ "step": 19000
732
+ },
733
+ {
734
+ "epoch": 2.82,
735
+ "learning_rate": 5.460095002604533e-06,
736
+ "loss": 1.553,
737
+ "step": 19500
738
+ },
739
+ {
740
+ "epoch": 2.82,
741
+ "eval_gen_len": 6.267666666666667,
742
+ "eval_loss": 1.7226474285125732,
743
+ "eval_rouge1": 79.9727,
744
+ "eval_rouge2": 68.2612,
745
+ "eval_rougeL": 79.9901,
746
+ "eval_rougeLsum": 79.9554,
747
+ "eval_runtime": 1525.8223,
748
+ "eval_samples_per_second": 1.966,
749
+ "eval_steps_per_second": 0.492,
750
+ "step": 19500
751
+ },
752
+ {
753
+ "epoch": 2.89,
754
+ "learning_rate": 5.000000000000003e-06,
755
+ "loss": 1.5498,
756
+ "step": 20000
757
+ },
758
+ {
759
+ "epoch": 2.89,
760
+ "eval_gen_len": 6.267666666666667,
761
+ "eval_loss": 1.720814824104309,
762
+ "eval_rouge1": 79.9937,
763
+ "eval_rouge2": 68.221,
764
+ "eval_rougeL": 79.9703,
765
+ "eval_rougeLsum": 79.9696,
766
+ "eval_runtime": 1519.3409,
767
+ "eval_samples_per_second": 1.975,
768
+ "eval_steps_per_second": 0.494,
769
+ "step": 20000
770
+ },
771
+ {
772
+ "epoch": 2.96,
773
+ "learning_rate": 4.5536096498497295e-06,
774
+ "loss": 1.5482,
775
+ "step": 20500
776
+ },
777
+ {
778
+ "epoch": 2.96,
779
+ "eval_gen_len": 6.267666666666667,
780
+ "eval_loss": 1.7179197072982788,
781
+ "eval_rouge1": 79.8959,
782
+ "eval_rouge2": 68.4771,
783
+ "eval_rougeL": 79.913,
784
+ "eval_rougeLsum": 79.8608,
785
+ "eval_runtime": 1520.8168,
786
+ "eval_samples_per_second": 1.973,
787
+ "eval_steps_per_second": 0.493,
788
+ "step": 20500
789
+ },
790
+ {
791
+ "epoch": 3.03,
792
+ "learning_rate": 4.12214747707527e-06,
793
+ "loss": 1.5305,
794
+ "step": 21000
795
+ },
796
+ {
797
+ "epoch": 3.03,
798
+ "eval_gen_len": 6.267666666666667,
799
+ "eval_loss": 1.7194596529006958,
800
+ "eval_rouge1": 80.1371,
801
+ "eval_rouge2": 68.4834,
802
+ "eval_rougeL": 80.1291,
803
+ "eval_rougeLsum": 80.1164,
804
+ "eval_runtime": 1522.408,
805
+ "eval_samples_per_second": 1.971,
806
+ "eval_steps_per_second": 0.493,
807
+ "step": 21000
808
+ },
809
+ {
810
+ "epoch": 3.11,
811
+ "learning_rate": 3.7067960895016277e-06,
812
+ "loss": 1.5057,
813
+ "step": 21500
814
+ },
815
+ {
816
+ "epoch": 3.11,
817
+ "eval_gen_len": 6.267666666666667,
818
+ "eval_loss": 1.7195872068405151,
819
+ "eval_rouge1": 80.143,
820
+ "eval_rouge2": 68.5801,
821
+ "eval_rougeL": 80.1614,
822
+ "eval_rougeLsum": 80.1476,
823
+ "eval_runtime": 1525.9564,
824
+ "eval_samples_per_second": 1.966,
825
+ "eval_steps_per_second": 0.491,
826
+ "step": 21500
827
+ },
828
+ {
829
+ "epoch": 3.18,
830
+ "learning_rate": 3.308693936411421e-06,
831
+ "loss": 1.5057,
832
+ "step": 22000
833
+ },
834
+ {
835
+ "epoch": 3.18,
836
+ "eval_gen_len": 6.267666666666667,
837
+ "eval_loss": 1.7173995971679688,
838
+ "eval_rouge1": 80.3721,
839
+ "eval_rouge2": 69.0264,
840
+ "eval_rougeL": 80.3868,
841
+ "eval_rougeLsum": 80.3636,
842
+ "eval_runtime": 1526.4809,
843
+ "eval_samples_per_second": 1.965,
844
+ "eval_steps_per_second": 0.491,
845
+ "step": 22000
846
+ },
847
+ {
848
+ "epoch": 3.25,
849
+ "learning_rate": 2.9289321881345257e-06,
850
+ "loss": 1.5052,
851
+ "step": 22500
852
+ },
853
+ {
854
+ "epoch": 3.25,
855
+ "eval_gen_len": 6.267666666666667,
856
+ "eval_loss": 1.7151570320129395,
857
+ "eval_rouge1": 80.3483,
858
+ "eval_rouge2": 69.2677,
859
+ "eval_rougeL": 80.333,
860
+ "eval_rougeLsum": 80.319,
861
+ "eval_runtime": 1527.1337,
862
+ "eval_samples_per_second": 1.964,
863
+ "eval_steps_per_second": 0.491,
864
+ "step": 22500
865
+ }
866
+ ],
867
+ "logging_steps": 500,
868
+ "max_steps": 30000,
869
+ "num_train_epochs": 5,
870
+ "save_steps": 2500,
871
+ "total_flos": 9.256721527856222e+16,
872
+ "trial_name": null,
873
+ "trial_params": null
874
+ }
checkpoint-22500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9810d57e4da355ac8aaf202fa72b6c52feff13a2bae2de7c074719dbb3c5e06b
3
+ size 4728
checkpoint-25000/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/blip-vqa-base",
3
+ "architectures": [
4
+ "BlipForQuestionAnswering"
5
+ ],
6
+ "image_text_hidden_size": 256,
7
+ "initializer_factor": 1.0,
8
+ "initializer_range": 0.02,
9
+ "logit_scale_init_value": 2.6592,
10
+ "model_type": "blip",
11
+ "projection_dim": 512,
12
+ "text_config": {
13
+ "initializer_factor": 1.0,
14
+ "model_type": "blip_text_model",
15
+ "num_attention_heads": 12
16
+ },
17
+ "torch_dtype": "float32",
18
+ "transformers_version": "4.33.1",
19
+ "vision_config": {
20
+ "dropout": 0.0,
21
+ "initializer_factor": 1.0,
22
+ "initializer_range": 0.02,
23
+ "model_type": "blip_vision_model",
24
+ "num_channels": 3
25
+ }
26
+ }