kevinoli commited on
Commit
954df2f
1 Parent(s): a2585b3

Training in progress, step 8500, checkpoint

Browse files
checkpoint-8500/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5619fe2241cea92e8ac8b72d6459b577394edc96629f9180e50223fac24969ed
3
  size 1711848436
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:499ad615c0ef0c860efbb37c6f429c2bcaad55f3b45a5c94841e5cca6d51cd1a
3
  size 1711848436
checkpoint-8500/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:94601895527f60225e704f1d38c33eeb20c5f2a19bd78a23874a1d21412a8d2e
3
  size 3424043887
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b89dbe2b339eeff976777acc2acceb705512f5d47186322440c5095a38f72763
3
  size 3424043887
checkpoint-8500/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8cf41629f175b779a85078252e5aa3a6f6f055204ba65a695c647da5153e8288
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1bf1ae99ccab59ad210d611dc1a5be3f58fe2583add42e42c0569682d4d14ee
3
  size 623
checkpoint-8500/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 1.9636509418487549,
3
- "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l55-l/checkpoint-7000",
4
- "epoch": 1.565665868484067,
5
  "eval_steps": 500,
6
  "global_step": 8500,
7
  "is_hyper_param_search": false,
@@ -9,263 +9,263 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.09209799226376865,
13
- "grad_norm": 0.009597355499863625,
14
- "learning_rate": 4.846503346227052e-05,
15
- "loss": 0.7336,
16
  "step": 500
17
  },
18
  {
19
- "epoch": 0.09209799226376865,
20
- "eval_loss": 2.0780556201934814,
21
- "eval_runtime": 77.1404,
22
- "eval_samples_per_second": 15.634,
23
- "eval_steps_per_second": 1.957,
24
  "step": 500
25
  },
26
  {
27
- "epoch": 0.1841959845275373,
28
- "grad_norm": 0.10049372166395187,
29
- "learning_rate": 4.693006692454105e-05,
30
- "loss": 0.6953,
31
  "step": 1000
32
  },
33
  {
34
- "epoch": 0.1841959845275373,
35
- "eval_loss": 2.0782744884490967,
36
- "eval_runtime": 77.8965,
37
- "eval_samples_per_second": 15.482,
38
- "eval_steps_per_second": 1.938,
39
  "step": 1000
40
  },
41
  {
42
- "epoch": 0.27629397679130596,
43
- "grad_norm": 0.0041180383414030075,
44
- "learning_rate": 4.539510038681157e-05,
45
- "loss": 0.6953,
46
  "step": 1500
47
  },
48
  {
49
- "epoch": 0.27629397679130596,
50
- "eval_loss": 2.077699899673462,
51
- "eval_runtime": 78.9586,
52
- "eval_samples_per_second": 15.274,
53
- "eval_steps_per_second": 1.912,
54
  "step": 1500
55
  },
56
  {
57
- "epoch": 0.3683919690550746,
58
- "grad_norm": 0.29397013783454895,
59
- "learning_rate": 4.386013384908209e-05,
60
- "loss": 0.6932,
61
  "step": 2000
62
  },
63
  {
64
- "epoch": 0.3683919690550746,
65
- "eval_loss": 2.0721774101257324,
66
- "eval_runtime": 77.349,
67
- "eval_samples_per_second": 15.592,
68
- "eval_steps_per_second": 1.952,
69
  "step": 2000
70
  },
71
  {
72
- "epoch": 0.46048996131884323,
73
- "grad_norm": 0.9136261343955994,
74
- "learning_rate": 4.232516731135262e-05,
75
- "loss": 0.6907,
76
  "step": 2500
77
  },
78
  {
79
- "epoch": 0.46048996131884323,
80
- "eval_loss": 2.079211711883545,
81
- "eval_runtime": 79.3204,
82
- "eval_samples_per_second": 15.204,
83
- "eval_steps_per_second": 1.904,
84
  "step": 2500
85
  },
86
  {
87
- "epoch": 0.5525879535826119,
88
- "grad_norm": 4.484213829040527,
89
- "learning_rate": 4.079020077362314e-05,
90
- "loss": 0.7043,
91
  "step": 3000
92
  },
93
  {
94
- "epoch": 0.5525879535826119,
95
- "eval_loss": 2.0204577445983887,
96
- "eval_runtime": 78.612,
97
- "eval_samples_per_second": 15.341,
98
- "eval_steps_per_second": 1.921,
99
  "step": 3000
100
  },
101
  {
102
- "epoch": 0.6446859458463805,
103
- "grad_norm": 1.1021015644073486,
104
- "learning_rate": 3.9255234235893664e-05,
105
- "loss": 0.7078,
106
  "step": 3500
107
  },
108
  {
109
- "epoch": 0.6446859458463805,
110
- "eval_loss": 2.1256344318389893,
111
- "eval_runtime": 78.63,
112
- "eval_samples_per_second": 15.338,
113
- "eval_steps_per_second": 1.92,
114
  "step": 3500
115
  },
116
  {
117
- "epoch": 0.7367839381101492,
118
- "grad_norm": 16.864044189453125,
119
- "learning_rate": 3.772026769816418e-05,
120
- "loss": 0.7034,
121
  "step": 4000
122
  },
123
  {
124
- "epoch": 0.7367839381101492,
125
- "eval_loss": 2.0000314712524414,
126
- "eval_runtime": 78.6543,
127
- "eval_samples_per_second": 15.333,
128
- "eval_steps_per_second": 1.92,
129
  "step": 4000
130
  },
131
  {
132
- "epoch": 0.8288819303739179,
133
- "grad_norm": 20.071943283081055,
134
- "learning_rate": 3.6185301160434705e-05,
135
- "loss": 0.6671,
136
  "step": 4500
137
  },
138
  {
139
- "epoch": 0.8288819303739179,
140
- "eval_loss": 2.1607539653778076,
141
- "eval_runtime": 78.6404,
142
- "eval_samples_per_second": 15.336,
143
- "eval_steps_per_second": 1.92,
144
  "step": 4500
145
  },
146
  {
147
- "epoch": 0.9209799226376865,
148
- "grad_norm": 17.817607879638672,
149
- "learning_rate": 3.4650334622705225e-05,
150
- "loss": 0.6899,
151
  "step": 5000
152
  },
153
  {
154
- "epoch": 0.9209799226376865,
155
- "eval_loss": 2.093959331512451,
156
- "eval_runtime": 78.539,
157
- "eval_samples_per_second": 15.355,
158
- "eval_steps_per_second": 1.923,
159
  "step": 5000
160
  },
161
  {
162
- "epoch": 1.013077914901455,
163
- "grad_norm": 41.0746955871582,
164
- "learning_rate": 3.311536808497575e-05,
165
- "loss": 0.6649,
166
  "step": 5500
167
  },
168
  {
169
- "epoch": 1.013077914901455,
170
- "eval_loss": 1.9776026010513306,
171
- "eval_runtime": 78.5095,
172
- "eval_samples_per_second": 15.361,
173
- "eval_steps_per_second": 1.923,
174
  "step": 5500
175
  },
176
  {
177
- "epoch": 1.1051759071652238,
178
- "grad_norm": 35.426021575927734,
179
- "learning_rate": 3.158040154724627e-05,
180
- "loss": 0.6748,
181
  "step": 6000
182
  },
183
  {
184
- "epoch": 1.1051759071652238,
185
- "eval_loss": 1.9851746559143066,
186
- "eval_runtime": 78.5944,
187
- "eval_samples_per_second": 15.345,
188
- "eval_steps_per_second": 1.921,
189
  "step": 6000
190
  },
191
  {
192
- "epoch": 1.1972738994289924,
193
- "grad_norm": 0.8773899078369141,
194
- "learning_rate": 3.0045435009516792e-05,
195
- "loss": 0.6456,
196
  "step": 6500
197
  },
198
  {
199
- "epoch": 1.1972738994289924,
200
- "eval_loss": 1.9981426000595093,
201
- "eval_runtime": 78.4741,
202
- "eval_samples_per_second": 15.368,
203
- "eval_steps_per_second": 1.924,
204
  "step": 6500
205
  },
206
  {
207
- "epoch": 1.289371891692761,
208
- "grad_norm": 19.7215518951416,
209
- "learning_rate": 2.8510468471787316e-05,
210
- "loss": 0.6514,
211
  "step": 7000
212
  },
213
  {
214
- "epoch": 1.289371891692761,
215
- "eval_loss": 1.9636509418487549,
216
- "eval_runtime": 78.4704,
217
- "eval_samples_per_second": 15.369,
218
- "eval_steps_per_second": 1.924,
219
  "step": 7000
220
  },
221
  {
222
- "epoch": 1.3814698839565298,
223
- "grad_norm": 5.96924352645874,
224
- "learning_rate": 2.697550193405784e-05,
225
- "loss": 0.6179,
226
  "step": 7500
227
  },
228
  {
229
- "epoch": 1.3814698839565298,
230
- "eval_loss": 2.0034029483795166,
231
- "eval_runtime": 78.4771,
232
- "eval_samples_per_second": 15.368,
233
- "eval_steps_per_second": 1.924,
234
  "step": 7500
235
  },
236
  {
237
- "epoch": 1.4735678762202984,
238
- "grad_norm": 1.2145048379898071,
239
- "learning_rate": 2.5440535396328363e-05,
240
- "loss": 0.6561,
241
  "step": 8000
242
  },
243
  {
244
- "epoch": 1.4735678762202984,
245
- "eval_loss": 2.2864511013031006,
246
- "eval_runtime": 78.3606,
247
- "eval_samples_per_second": 15.39,
248
- "eval_steps_per_second": 1.927,
249
  "step": 8000
250
  },
251
  {
252
- "epoch": 1.565665868484067,
253
- "grad_norm": 15.816754341125488,
254
- "learning_rate": 2.3905568858598883e-05,
255
- "loss": 0.6328,
256
  "step": 8500
257
  },
258
  {
259
- "epoch": 1.565665868484067,
260
- "eval_loss": 2.4807703495025635,
261
- "eval_runtime": 77.0504,
262
- "eval_samples_per_second": 15.652,
263
- "eval_steps_per_second": 1.96,
264
  "step": 8500
265
  }
266
  ],
267
  "logging_steps": 500,
268
- "max_steps": 16287,
269
  "num_input_tokens_seen": 0,
270
  "num_train_epochs": 3,
271
  "save_steps": 500,
@@ -281,7 +281,7 @@
281
  "attributes": {}
282
  }
283
  },
284
- "total_flos": 3058485623236890.0,
285
  "train_batch_size": 2,
286
  "trial_name": null,
287
  "trial_params": null
 
1
  {
2
+ "best_metric": 1.9648858308792114,
3
+ "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l55-l/checkpoint-5500",
4
+ "epoch": 1.565954310980103,
5
  "eval_steps": 500,
6
  "global_step": 8500,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.09211495946941783,
13
+ "grad_norm": 0.030529862269759178,
14
+ "learning_rate": 4.8464750675509704e-05,
15
+ "loss": 0.7076,
16
  "step": 500
17
  },
18
  {
19
+ "epoch": 0.09211495946941783,
20
+ "eval_loss": 2.07871675491333,
21
+ "eval_runtime": 73.4415,
22
+ "eval_samples_per_second": 16.435,
23
+ "eval_steps_per_second": 2.056,
24
  "step": 500
25
  },
26
  {
27
+ "epoch": 0.18422991893883567,
28
+ "grad_norm": 0.02040918357670307,
29
+ "learning_rate": 4.6929501351019406e-05,
30
+ "loss": 0.7045,
31
  "step": 1000
32
  },
33
  {
34
+ "epoch": 0.18422991893883567,
35
+ "eval_loss": 2.0783257484436035,
36
+ "eval_runtime": 74.6654,
37
+ "eval_samples_per_second": 16.165,
38
+ "eval_steps_per_second": 2.022,
39
  "step": 1000
40
  },
41
  {
42
+ "epoch": 0.2763448784082535,
43
+ "grad_norm": 0.8010361194610596,
44
+ "learning_rate": 4.539425202652911e-05,
45
+ "loss": 0.7106,
46
  "step": 1500
47
  },
48
  {
49
+ "epoch": 0.2763448784082535,
50
+ "eval_loss": 2.0473792552948,
51
+ "eval_runtime": 75.4141,
52
+ "eval_samples_per_second": 16.005,
53
+ "eval_steps_per_second": 2.002,
54
  "step": 1500
55
  },
56
  {
57
+ "epoch": 0.36845983787767134,
58
+ "grad_norm": 1.1971429586410522,
59
+ "learning_rate": 4.3859002702038817e-05,
60
+ "loss": 0.704,
61
  "step": 2000
62
  },
63
  {
64
+ "epoch": 0.36845983787767134,
65
+ "eval_loss": 2.281594753265381,
66
+ "eval_runtime": 76.3969,
67
+ "eval_samples_per_second": 15.799,
68
+ "eval_steps_per_second": 1.977,
69
  "step": 2000
70
  },
71
  {
72
+ "epoch": 0.46057479734708917,
73
+ "grad_norm": 1.5381091833114624,
74
+ "learning_rate": 4.232375337754851e-05,
75
+ "loss": 0.6996,
76
  "step": 2500
77
  },
78
  {
79
+ "epoch": 0.46057479734708917,
80
+ "eval_loss": 2.045022964477539,
81
+ "eval_runtime": 76.3329,
82
+ "eval_samples_per_second": 15.812,
83
+ "eval_steps_per_second": 1.978,
84
  "step": 2500
85
  },
86
  {
87
+ "epoch": 0.552689756816507,
88
+ "grad_norm": 5.1257524490356445,
89
+ "learning_rate": 4.078850405305822e-05,
90
+ "loss": 0.6739,
91
  "step": 3000
92
  },
93
  {
94
+ "epoch": 0.552689756816507,
95
+ "eval_loss": 2.0969045162200928,
96
+ "eval_runtime": 76.2547,
97
+ "eval_samples_per_second": 15.829,
98
+ "eval_steps_per_second": 1.98,
99
  "step": 3000
100
  },
101
  {
102
+ "epoch": 0.6448047162859248,
103
+ "grad_norm": 6.156252861022949,
104
+ "learning_rate": 3.925325472856792e-05,
105
+ "loss": 0.6687,
106
  "step": 3500
107
  },
108
  {
109
+ "epoch": 0.6448047162859248,
110
+ "eval_loss": 2.0315287113189697,
111
+ "eval_runtime": 76.4108,
112
+ "eval_samples_per_second": 15.796,
113
+ "eval_steps_per_second": 1.976,
114
  "step": 3500
115
  },
116
  {
117
+ "epoch": 0.7369196757553427,
118
+ "grad_norm": 0.053296659141778946,
119
+ "learning_rate": 3.7718005404077624e-05,
120
+ "loss": 0.8387,
121
  "step": 4000
122
  },
123
  {
124
+ "epoch": 0.7369196757553427,
125
+ "eval_loss": 2.3810112476348877,
126
+ "eval_runtime": 74.9737,
127
+ "eval_samples_per_second": 16.099,
128
+ "eval_steps_per_second": 2.014,
129
  "step": 4000
130
  },
131
  {
132
+ "epoch": 0.8290346352247605,
133
+ "grad_norm": 0.1077703982591629,
134
+ "learning_rate": 3.6182756079587326e-05,
135
+ "loss": 0.6988,
136
  "step": 4500
137
  },
138
  {
139
+ "epoch": 0.8290346352247605,
140
+ "eval_loss": 2.0409226417541504,
141
+ "eval_runtime": 76.7206,
142
+ "eval_samples_per_second": 15.732,
143
+ "eval_steps_per_second": 1.968,
144
  "step": 4500
145
  },
146
  {
147
+ "epoch": 0.9211495946941783,
148
+ "grad_norm": 4.246299743652344,
149
+ "learning_rate": 3.4647506755097034e-05,
150
+ "loss": 0.6808,
151
  "step": 5000
152
  },
153
  {
154
+ "epoch": 0.9211495946941783,
155
+ "eval_loss": 1.9745088815689087,
156
+ "eval_runtime": 74.928,
157
+ "eval_samples_per_second": 16.109,
158
+ "eval_steps_per_second": 2.015,
159
  "step": 5000
160
  },
161
  {
162
+ "epoch": 1.013264554163596,
163
+ "grad_norm": 2.2185893058776855,
164
+ "learning_rate": 3.311225743060673e-05,
165
+ "loss": 0.6584,
166
  "step": 5500
167
  },
168
  {
169
+ "epoch": 1.013264554163596,
170
+ "eval_loss": 1.9648858308792114,
171
+ "eval_runtime": 76.0743,
172
+ "eval_samples_per_second": 15.866,
173
+ "eval_steps_per_second": 1.985,
174
  "step": 5500
175
  },
176
  {
177
+ "epoch": 1.105379513633014,
178
+ "grad_norm": 6.992915153503418,
179
+ "learning_rate": 3.157700810611644e-05,
180
+ "loss": 0.6567,
181
  "step": 6000
182
  },
183
  {
184
+ "epoch": 1.105379513633014,
185
+ "eval_loss": 2.0568628311157227,
186
+ "eval_runtime": 76.0502,
187
+ "eval_samples_per_second": 15.871,
188
+ "eval_steps_per_second": 1.986,
189
  "step": 6000
190
  },
191
  {
192
+ "epoch": 1.1974944731024317,
193
+ "grad_norm": 13.946557998657227,
194
+ "learning_rate": 3.004175878162614e-05,
195
+ "loss": 0.6542,
196
  "step": 6500
197
  },
198
  {
199
+ "epoch": 1.1974944731024317,
200
+ "eval_loss": 1.9788861274719238,
201
+ "eval_runtime": 75.9196,
202
+ "eval_samples_per_second": 15.898,
203
+ "eval_steps_per_second": 1.989,
204
  "step": 6500
205
  },
206
  {
207
+ "epoch": 1.2896094325718497,
208
+ "grad_norm": 2.426790237426758,
209
+ "learning_rate": 2.8506509457135838e-05,
210
+ "loss": 0.6536,
211
  "step": 7000
212
  },
213
  {
214
+ "epoch": 1.2896094325718497,
215
+ "eval_loss": 1.9676496982574463,
216
+ "eval_runtime": 75.155,
217
+ "eval_samples_per_second": 16.06,
218
+ "eval_steps_per_second": 2.009,
219
  "step": 7000
220
  },
221
  {
222
+ "epoch": 1.3817243920412676,
223
+ "grad_norm": 1.7985050678253174,
224
+ "learning_rate": 2.6971260132645544e-05,
225
+ "loss": 0.637,
226
  "step": 7500
227
  },
228
  {
229
+ "epoch": 1.3817243920412676,
230
+ "eval_loss": 2.003986358642578,
231
+ "eval_runtime": 74.9913,
232
+ "eval_samples_per_second": 16.095,
233
+ "eval_steps_per_second": 2.014,
234
  "step": 7500
235
  },
236
  {
237
+ "epoch": 1.4738393515106853,
238
+ "grad_norm": 22.398109436035156,
239
+ "learning_rate": 2.543601080815525e-05,
240
+ "loss": 0.6409,
241
  "step": 8000
242
  },
243
  {
244
+ "epoch": 1.4738393515106853,
245
+ "eval_loss": 1.9904581308364868,
246
+ "eval_runtime": 74.9317,
247
+ "eval_samples_per_second": 16.108,
248
+ "eval_steps_per_second": 2.015,
249
  "step": 8000
250
  },
251
  {
252
+ "epoch": 1.565954310980103,
253
+ "grad_norm": 4.255030155181885,
254
+ "learning_rate": 2.390076148366495e-05,
255
+ "loss": 0.6127,
256
  "step": 8500
257
  },
258
  {
259
+ "epoch": 1.565954310980103,
260
+ "eval_loss": 2.2638871669769287,
261
+ "eval_runtime": 74.9634,
262
+ "eval_samples_per_second": 16.101,
263
+ "eval_steps_per_second": 2.014,
264
  "step": 8500
265
  }
266
  ],
267
  "logging_steps": 500,
268
+ "max_steps": 16284,
269
  "num_input_tokens_seen": 0,
270
  "num_train_epochs": 3,
271
  "save_steps": 500,
 
281
  "attributes": {}
282
  }
283
  },
284
+ "total_flos": 3058665502412520.0,
285
  "train_batch_size": 2,
286
  "trial_name": null,
287
  "trial_params": null
checkpoint-8500/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d9885fd590ddfbffc4f8db1738c1993002dfa8c1603f7992a1cda9b93bbe7d34
3
  size 4847
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59e6c277b84a863e8b2dbf87019462e313e3d40a89afb31c644b26985cc0095f
3
  size 4847