kevinoli commited on
Commit
1a874c6
1 Parent(s): f64965c

Training in progress, step 10856, checkpoint

Browse files
checkpoint-10856/config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "openai/clip-vit-large-patch14-336",
3
+ "architectures": [
4
+ "CLIPModel"
5
+ ],
6
+ "initializer_factor": 1.0,
7
+ "logit_scale_init_value": 2.6592,
8
+ "model_type": "clip",
9
+ "projection_dim": 768,
10
+ "text_config": {
11
+ "dropout": 0.0,
12
+ "hidden_size": 768,
13
+ "intermediate_size": 3072,
14
+ "model_type": "clip_text_model",
15
+ "num_attention_heads": 12,
16
+ "projection_dim": 768
17
+ },
18
+ "torch_dtype": "float32",
19
+ "transformers_version": "4.45.0.dev0",
20
+ "vision_config": {
21
+ "dropout": 0.0,
22
+ "hidden_size": 1024,
23
+ "image_size": 336,
24
+ "intermediate_size": 4096,
25
+ "model_type": "clip_vision_model",
26
+ "num_attention_heads": 16,
27
+ "num_hidden_layers": 24,
28
+ "patch_size": 14,
29
+ "projection_dim": 768
30
+ }
31
+ }
checkpoint-10856/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4a99251b81ef252037521aafb4dffd51418da1b6a1ea61b02394753ca7b3b41
3
+ size 1711848436
checkpoint-10856/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adae5aee189ebec91faa264442dbda3765dbab97d56735059f70254fc3851ab6
3
+ size 3424043887
checkpoint-10856/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9be3731bd45b68aa3495049ac193f5be0636ae23217b87c22e9387c9699e277
3
+ size 14503
checkpoint-10856/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:84e184f25c803dd4323d591226a0ed6c476da5a70adf95880473dfc51d05386c
3
+ size 623
checkpoint-10856/trainer_state.json ADDED
@@ -0,0 +1,348 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 2.009561777114868,
3
+ "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e2l57-l/checkpoint-9500",
4
+ "epoch": 2.0,
5
+ "eval_steps": 500,
6
+ "global_step": 10856,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.09211495946941783,
13
+ "grad_norm": 0.00766215892508626,
14
+ "learning_rate": 4.769712601326456e-05,
15
+ "loss": 0.7207,
16
+ "step": 500
17
+ },
18
+ {
19
+ "epoch": 0.09211495946941783,
20
+ "eval_loss": 2.078455924987793,
21
+ "eval_runtime": 73.0449,
22
+ "eval_samples_per_second": 16.524,
23
+ "eval_steps_per_second": 2.067,
24
+ "step": 500
25
+ },
26
+ {
27
+ "epoch": 0.18422991893883567,
28
+ "grad_norm": 0.016227997839450836,
29
+ "learning_rate": 4.539425202652911e-05,
30
+ "loss": 0.702,
31
+ "step": 1000
32
+ },
33
+ {
34
+ "epoch": 0.18422991893883567,
35
+ "eval_loss": 2.0786073207855225,
36
+ "eval_runtime": 73.7466,
37
+ "eval_samples_per_second": 16.367,
38
+ "eval_steps_per_second": 2.048,
39
+ "step": 1000
40
+ },
41
+ {
42
+ "epoch": 0.2763448784082535,
43
+ "grad_norm": 3.326040029525757,
44
+ "learning_rate": 4.309137803979367e-05,
45
+ "loss": 0.7058,
46
+ "step": 1500
47
+ },
48
+ {
49
+ "epoch": 0.2763448784082535,
50
+ "eval_loss": 2.0656001567840576,
51
+ "eval_runtime": 75.4832,
52
+ "eval_samples_per_second": 15.99,
53
+ "eval_steps_per_second": 2.0,
54
+ "step": 1500
55
+ },
56
+ {
57
+ "epoch": 0.36845983787767134,
58
+ "grad_norm": 5.95470666885376,
59
+ "learning_rate": 4.078850405305822e-05,
60
+ "loss": 0.7117,
61
+ "step": 2000
62
+ },
63
+ {
64
+ "epoch": 0.36845983787767134,
65
+ "eval_loss": 2.113447427749634,
66
+ "eval_runtime": 75.7404,
67
+ "eval_samples_per_second": 15.936,
68
+ "eval_steps_per_second": 1.994,
69
+ "step": 2000
70
+ },
71
+ {
72
+ "epoch": 0.46057479734708917,
73
+ "grad_norm": 1.3789223432540894,
74
+ "learning_rate": 3.848563006632277e-05,
75
+ "loss": 0.7163,
76
+ "step": 2500
77
+ },
78
+ {
79
+ "epoch": 0.46057479734708917,
80
+ "eval_loss": 2.0306718349456787,
81
+ "eval_runtime": 75.7822,
82
+ "eval_samples_per_second": 15.927,
83
+ "eval_steps_per_second": 1.993,
84
+ "step": 2500
85
+ },
86
+ {
87
+ "epoch": 0.552689756816507,
88
+ "grad_norm": 17.15974998474121,
89
+ "learning_rate": 3.6182756079587326e-05,
90
+ "loss": 0.6794,
91
+ "step": 3000
92
+ },
93
+ {
94
+ "epoch": 0.552689756816507,
95
+ "eval_loss": 2.613356590270996,
96
+ "eval_runtime": 75.7313,
97
+ "eval_samples_per_second": 15.938,
98
+ "eval_steps_per_second": 1.994,
99
+ "step": 3000
100
+ },
101
+ {
102
+ "epoch": 0.6448047162859248,
103
+ "grad_norm": 6.376380443572998,
104
+ "learning_rate": 3.3879882092851885e-05,
105
+ "loss": 0.6818,
106
+ "step": 3500
107
+ },
108
+ {
109
+ "epoch": 0.6448047162859248,
110
+ "eval_loss": 2.0862319469451904,
111
+ "eval_runtime": 75.9967,
112
+ "eval_samples_per_second": 15.882,
113
+ "eval_steps_per_second": 1.987,
114
+ "step": 3500
115
+ },
116
+ {
117
+ "epoch": 0.7369196757553427,
118
+ "grad_norm": 4.5618391036987305,
119
+ "learning_rate": 3.157700810611644e-05,
120
+ "loss": 0.6693,
121
+ "step": 4000
122
+ },
123
+ {
124
+ "epoch": 0.7369196757553427,
125
+ "eval_loss": 2.0294671058654785,
126
+ "eval_runtime": 74.4695,
127
+ "eval_samples_per_second": 16.208,
128
+ "eval_steps_per_second": 2.028,
129
+ "step": 4000
130
+ },
131
+ {
132
+ "epoch": 0.8290346352247605,
133
+ "grad_norm": 4.356328964233398,
134
+ "learning_rate": 2.927413411938099e-05,
135
+ "loss": 0.6798,
136
+ "step": 4500
137
+ },
138
+ {
139
+ "epoch": 0.8290346352247605,
140
+ "eval_loss": 2.1398818492889404,
141
+ "eval_runtime": 76.5616,
142
+ "eval_samples_per_second": 15.765,
143
+ "eval_steps_per_second": 1.972,
144
+ "step": 4500
145
+ },
146
+ {
147
+ "epoch": 0.9211495946941783,
148
+ "grad_norm": 1.5961418151855469,
149
+ "learning_rate": 2.6971260132645544e-05,
150
+ "loss": 0.6915,
151
+ "step": 5000
152
+ },
153
+ {
154
+ "epoch": 0.9211495946941783,
155
+ "eval_loss": 2.0116002559661865,
156
+ "eval_runtime": 75.2588,
157
+ "eval_samples_per_second": 16.038,
158
+ "eval_steps_per_second": 2.006,
159
+ "step": 5000
160
+ },
161
+ {
162
+ "epoch": 1.013264554163596,
163
+ "grad_norm": 0.672991931438446,
164
+ "learning_rate": 2.4668386145910096e-05,
165
+ "loss": 0.6623,
166
+ "step": 5500
167
+ },
168
+ {
169
+ "epoch": 1.013264554163596,
170
+ "eval_loss": 2.1612467765808105,
171
+ "eval_runtime": 75.2829,
172
+ "eval_samples_per_second": 16.033,
173
+ "eval_steps_per_second": 2.006,
174
+ "step": 5500
175
+ },
176
+ {
177
+ "epoch": 1.105379513633014,
178
+ "grad_norm": 5.576838493347168,
179
+ "learning_rate": 2.2365512159174652e-05,
180
+ "loss": 0.6803,
181
+ "step": 6000
182
+ },
183
+ {
184
+ "epoch": 1.105379513633014,
185
+ "eval_loss": 2.0493156909942627,
186
+ "eval_runtime": 74.734,
187
+ "eval_samples_per_second": 16.151,
188
+ "eval_steps_per_second": 2.02,
189
+ "step": 6000
190
+ },
191
+ {
192
+ "epoch": 1.1974944731024317,
193
+ "grad_norm": 1.1342592239379883,
194
+ "learning_rate": 2.0062638172439205e-05,
195
+ "loss": 0.6729,
196
+ "step": 6500
197
+ },
198
+ {
199
+ "epoch": 1.1974944731024317,
200
+ "eval_loss": 2.051811933517456,
201
+ "eval_runtime": 74.1516,
202
+ "eval_samples_per_second": 16.277,
203
+ "eval_steps_per_second": 2.036,
204
+ "step": 6500
205
+ },
206
+ {
207
+ "epoch": 1.2896094325718497,
208
+ "grad_norm": 2.215428113937378,
209
+ "learning_rate": 1.7759764185703758e-05,
210
+ "loss": 0.6851,
211
+ "step": 7000
212
+ },
213
+ {
214
+ "epoch": 1.2896094325718497,
215
+ "eval_loss": 2.039102077484131,
216
+ "eval_runtime": 74.3318,
217
+ "eval_samples_per_second": 16.238,
218
+ "eval_steps_per_second": 2.031,
219
+ "step": 7000
220
+ },
221
+ {
222
+ "epoch": 1.3817243920412676,
223
+ "grad_norm": 4.479649066925049,
224
+ "learning_rate": 1.5456890198968314e-05,
225
+ "loss": 0.6686,
226
+ "step": 7500
227
+ },
228
+ {
229
+ "epoch": 1.3817243920412676,
230
+ "eval_loss": 2.06972074508667,
231
+ "eval_runtime": 74.3941,
232
+ "eval_samples_per_second": 16.224,
233
+ "eval_steps_per_second": 2.03,
234
+ "step": 7500
235
+ },
236
+ {
237
+ "epoch": 1.4738393515106853,
238
+ "grad_norm": 9.912164688110352,
239
+ "learning_rate": 1.3154016212232867e-05,
240
+ "loss": 0.6637,
241
+ "step": 8000
242
+ },
243
+ {
244
+ "epoch": 1.4738393515106853,
245
+ "eval_loss": 2.045731782913208,
246
+ "eval_runtime": 74.4167,
247
+ "eval_samples_per_second": 16.219,
248
+ "eval_steps_per_second": 2.029,
249
+ "step": 8000
250
+ },
251
+ {
252
+ "epoch": 1.565954310980103,
253
+ "grad_norm": 7.797455310821533,
254
+ "learning_rate": 1.0851142225497421e-05,
255
+ "loss": 0.6331,
256
+ "step": 8500
257
+ },
258
+ {
259
+ "epoch": 1.565954310980103,
260
+ "eval_loss": 2.1388187408447266,
261
+ "eval_runtime": 74.2931,
262
+ "eval_samples_per_second": 16.246,
263
+ "eval_steps_per_second": 2.032,
264
+ "step": 8500
265
+ },
266
+ {
267
+ "epoch": 1.658069270449521,
268
+ "grad_norm": 14.766168594360352,
269
+ "learning_rate": 8.548268238761976e-06,
270
+ "loss": 0.6529,
271
+ "step": 9000
272
+ },
273
+ {
274
+ "epoch": 1.658069270449521,
275
+ "eval_loss": 2.0181920528411865,
276
+ "eval_runtime": 74.6522,
277
+ "eval_samples_per_second": 16.168,
278
+ "eval_steps_per_second": 2.023,
279
+ "step": 9000
280
+ },
281
+ {
282
+ "epoch": 1.750184229918939,
283
+ "grad_norm": 10.570740699768066,
284
+ "learning_rate": 6.245394252026529e-06,
285
+ "loss": 0.6377,
286
+ "step": 9500
287
+ },
288
+ {
289
+ "epoch": 1.750184229918939,
290
+ "eval_loss": 2.009561777114868,
291
+ "eval_runtime": 74.3801,
292
+ "eval_samples_per_second": 16.227,
293
+ "eval_steps_per_second": 2.03,
294
+ "step": 9500
295
+ },
296
+ {
297
+ "epoch": 1.8422991893883567,
298
+ "grad_norm": 3.4046995639801025,
299
+ "learning_rate": 3.942520265291084e-06,
300
+ "loss": 0.6207,
301
+ "step": 10000
302
+ },
303
+ {
304
+ "epoch": 1.8422991893883567,
305
+ "eval_loss": 2.0490310192108154,
306
+ "eval_runtime": 74.7879,
307
+ "eval_samples_per_second": 16.139,
308
+ "eval_steps_per_second": 2.019,
309
+ "step": 10000
310
+ },
311
+ {
312
+ "epoch": 1.9344141488577744,
313
+ "grad_norm": 2.0113770961761475,
314
+ "learning_rate": 1.6396462785556374e-06,
315
+ "loss": 0.6626,
316
+ "step": 10500
317
+ },
318
+ {
319
+ "epoch": 1.9344141488577744,
320
+ "eval_loss": 2.0148627758026123,
321
+ "eval_runtime": 74.8678,
322
+ "eval_samples_per_second": 16.122,
323
+ "eval_steps_per_second": 2.017,
324
+ "step": 10500
325
+ }
326
+ ],
327
+ "logging_steps": 500,
328
+ "max_steps": 10856,
329
+ "num_input_tokens_seen": 0,
330
+ "num_train_epochs": 2,
331
+ "save_steps": 500,
332
+ "stateful_callbacks": {
333
+ "TrainerControl": {
334
+ "args": {
335
+ "should_epoch_stop": false,
336
+ "should_evaluate": false,
337
+ "should_log": false,
338
+ "should_save": true,
339
+ "should_training_stop": true
340
+ },
341
+ "attributes": {}
342
+ }
343
+ },
344
+ "total_flos": 3906256177981080.0,
345
+ "train_batch_size": 2,
346
+ "trial_name": null,
347
+ "trial_params": null
348
+ }
checkpoint-10856/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:113e5581618b2f31a20dabc0bbd4943905ebe2c2323e7b369e736bac42477916
3
+ size 4847