kevinoli commited on
Commit
91f6f2d
1 Parent(s): 9021536

Training in progress, step 7000, checkpoint

Browse files
checkpoint-7000/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c474a25d66b60395b7998d904083307c4231d3c6e72506d5116afdb7962c58bb
3
  size 1711848436
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0142364f978126efa7e67a89547cb98f2a0d6f9bd160513701f0d7975d87ff86
3
  size 1711848436
checkpoint-7000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:92743dc1209adeea9879018e333d282f1e1bd25310518ab64c314bbcad932e03
3
  size 3424043887
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a08ab3cc21222d06d6be5ae1d578f753ed0e8e977db01279b32cbb2deed24e71
3
  size 3424043887
checkpoint-7000/trainer_state.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
- "best_metric": 0.6204637289047241,
3
- "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-l/checkpoint-7000",
4
  "epoch": 0.7465870307167235,
5
  "eval_steps": 500,
6
  "global_step": 7000,
@@ -10,212 +10,212 @@
10
  "log_history": [
11
  {
12
  "epoch": 0.05332764505119454,
13
- "grad_norm": 414.2552795410156,
14
  "learning_rate": 4.911120591581342e-07,
15
- "loss": 0.3809,
16
  "step": 500
17
  },
18
  {
19
  "epoch": 0.05332764505119454,
20
- "eval_loss": 1.1109352111816406,
21
- "eval_runtime": 63.6734,
22
- "eval_samples_per_second": 15.501,
23
- "eval_steps_per_second": 1.947,
24
  "step": 500
25
  },
26
  {
27
  "epoch": 0.10665529010238908,
28
- "grad_norm": 45.880027770996094,
29
  "learning_rate": 4.822241183162685e-07,
30
- "loss": 0.2649,
31
  "step": 1000
32
  },
33
  {
34
  "epoch": 0.10665529010238908,
35
- "eval_loss": 0.9588962197303772,
36
- "eval_runtime": 62.6825,
37
- "eval_samples_per_second": 15.746,
38
- "eval_steps_per_second": 1.978,
39
  "step": 1000
40
  },
41
  {
42
  "epoch": 0.1599829351535836,
43
- "grad_norm": 432.9915466308594,
44
  "learning_rate": 4.733361774744027e-07,
45
- "loss": 0.2104,
46
  "step": 1500
47
  },
48
  {
49
  "epoch": 0.1599829351535836,
50
- "eval_loss": 0.90110844373703,
51
- "eval_runtime": 62.6757,
52
- "eval_samples_per_second": 15.748,
53
- "eval_steps_per_second": 1.978,
54
  "step": 1500
55
  },
56
  {
57
  "epoch": 0.21331058020477817,
58
- "grad_norm": 0.014208819717168808,
59
  "learning_rate": 4.6444823663253695e-07,
60
- "loss": 0.1849,
61
  "step": 2000
62
  },
63
  {
64
  "epoch": 0.21331058020477817,
65
- "eval_loss": 0.8570474982261658,
66
- "eval_runtime": 63.7296,
67
- "eval_samples_per_second": 15.487,
68
- "eval_steps_per_second": 1.946,
69
  "step": 2000
70
  },
71
  {
72
  "epoch": 0.2666382252559727,
73
- "grad_norm": 174.73324584960938,
74
  "learning_rate": 4.5556029579067116e-07,
75
- "loss": 0.2056,
76
  "step": 2500
77
  },
78
  {
79
  "epoch": 0.2666382252559727,
80
- "eval_loss": 0.7895939350128174,
81
- "eval_runtime": 63.8387,
82
- "eval_samples_per_second": 15.461,
83
- "eval_steps_per_second": 1.942,
84
  "step": 2500
85
  },
86
  {
87
  "epoch": 0.3199658703071672,
88
- "grad_norm": 9.852073823424234e-09,
89
  "learning_rate": 4.4667235494880547e-07,
90
- "loss": 0.159,
91
  "step": 3000
92
  },
93
  {
94
  "epoch": 0.3199658703071672,
95
- "eval_loss": 0.7647623419761658,
96
- "eval_runtime": 63.9155,
97
- "eval_samples_per_second": 15.442,
98
- "eval_steps_per_second": 1.94,
99
  "step": 3000
100
  },
101
  {
102
  "epoch": 0.37329351535836175,
103
- "grad_norm": 0.0025215104687958956,
104
  "learning_rate": 4.377844141069397e-07,
105
- "loss": 0.1696,
106
  "step": 3500
107
  },
108
  {
109
  "epoch": 0.37329351535836175,
110
- "eval_loss": 0.7638036012649536,
111
- "eval_runtime": 63.8003,
112
- "eval_samples_per_second": 15.47,
113
- "eval_steps_per_second": 1.944,
114
  "step": 3500
115
  },
116
  {
117
  "epoch": 0.42662116040955633,
118
- "grad_norm": 2.4611830711364746,
119
  "learning_rate": 4.2889647326507393e-07,
120
- "loss": 0.1125,
121
  "step": 4000
122
  },
123
  {
124
  "epoch": 0.42662116040955633,
125
- "eval_loss": 0.7486374378204346,
126
- "eval_runtime": 63.9847,
127
- "eval_samples_per_second": 15.426,
128
- "eval_steps_per_second": 1.938,
129
  "step": 4000
130
  },
131
  {
132
  "epoch": 0.47994880546075086,
133
- "grad_norm": 5.230295658111572,
134
  "learning_rate": 4.2000853242320814e-07,
135
- "loss": 0.1322,
136
  "step": 4500
137
  },
138
  {
139
  "epoch": 0.47994880546075086,
140
- "eval_loss": 0.7251659035682678,
141
- "eval_runtime": 64.4782,
142
- "eval_samples_per_second": 15.308,
143
- "eval_steps_per_second": 1.923,
144
  "step": 4500
145
  },
146
  {
147
  "epoch": 0.5332764505119454,
148
- "grad_norm": 3.821002974291332e-05,
149
  "learning_rate": 4.1112059158134245e-07,
150
- "loss": 0.1633,
151
  "step": 5000
152
  },
153
  {
154
  "epoch": 0.5332764505119454,
155
- "eval_loss": 0.7035414576530457,
156
- "eval_runtime": 62.6186,
157
- "eval_samples_per_second": 15.762,
158
- "eval_steps_per_second": 1.98,
159
  "step": 5000
160
  },
161
  {
162
  "epoch": 0.58660409556314,
163
- "grad_norm": 5.916939244343666e-06,
164
  "learning_rate": 4.0223265073947665e-07,
165
- "loss": 0.125,
166
  "step": 5500
167
  },
168
  {
169
  "epoch": 0.58660409556314,
170
- "eval_loss": 0.6792380809783936,
171
- "eval_runtime": 65.0551,
172
- "eval_samples_per_second": 15.172,
173
- "eval_steps_per_second": 1.906,
174
  "step": 5500
175
  },
176
  {
177
  "epoch": 0.6399317406143344,
178
- "grad_norm": 92.8386459350586,
179
  "learning_rate": 3.933447098976109e-07,
180
- "loss": 0.1297,
181
  "step": 6000
182
  },
183
  {
184
  "epoch": 0.6399317406143344,
185
- "eval_loss": 0.6549482345581055,
186
- "eval_runtime": 64.0827,
187
- "eval_samples_per_second": 15.402,
188
- "eval_steps_per_second": 1.935,
189
  "step": 6000
190
  },
191
  {
192
  "epoch": 0.693259385665529,
193
- "grad_norm": 0.00014717792510055006,
194
  "learning_rate": 3.8445676905574517e-07,
195
- "loss": 0.1114,
196
  "step": 6500
197
  },
198
  {
199
  "epoch": 0.693259385665529,
200
- "eval_loss": 0.6223254799842834,
201
- "eval_runtime": 64.89,
202
- "eval_samples_per_second": 15.21,
203
- "eval_steps_per_second": 1.911,
204
  "step": 6500
205
  },
206
  {
207
  "epoch": 0.7465870307167235,
208
- "grad_norm": 4.789559397977428e-07,
209
  "learning_rate": 3.755688282138794e-07,
210
- "loss": 0.144,
211
  "step": 7000
212
  },
213
  {
214
  "epoch": 0.7465870307167235,
215
- "eval_loss": 0.6204637289047241,
216
- "eval_runtime": 62.7674,
217
- "eval_samples_per_second": 15.725,
218
- "eval_steps_per_second": 1.976,
219
  "step": 7000
220
  }
221
  ],
 
1
  {
2
+ "best_metric": 0.6227805614471436,
3
+ "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l57-l/checkpoint-6500",
4
  "epoch": 0.7465870307167235,
5
  "eval_steps": 500,
6
  "global_step": 7000,
 
10
  "log_history": [
11
  {
12
  "epoch": 0.05332764505119454,
13
+ "grad_norm": 401.61187744140625,
14
  "learning_rate": 4.911120591581342e-07,
15
+ "loss": 0.3808,
16
  "step": 500
17
  },
18
  {
19
  "epoch": 0.05332764505119454,
20
+ "eval_loss": 1.1107642650604248,
21
+ "eval_runtime": 62.4169,
22
+ "eval_samples_per_second": 15.813,
23
+ "eval_steps_per_second": 1.987,
24
  "step": 500
25
  },
26
  {
27
  "epoch": 0.10665529010238908,
28
+ "grad_norm": 42.57551193237305,
29
  "learning_rate": 4.822241183162685e-07,
30
+ "loss": 0.2666,
31
  "step": 1000
32
  },
33
  {
34
  "epoch": 0.10665529010238908,
35
+ "eval_loss": 0.9644754528999329,
36
+ "eval_runtime": 63.4304,
37
+ "eval_samples_per_second": 15.56,
38
+ "eval_steps_per_second": 1.955,
39
  "step": 1000
40
  },
41
  {
42
  "epoch": 0.1599829351535836,
43
+ "grad_norm": 518.2286376953125,
44
  "learning_rate": 4.733361774744027e-07,
45
+ "loss": 0.2085,
46
  "step": 1500
47
  },
48
  {
49
  "epoch": 0.1599829351535836,
50
+ "eval_loss": 0.9051361680030823,
51
+ "eval_runtime": 63.9009,
52
+ "eval_samples_per_second": 15.446,
53
+ "eval_steps_per_second": 1.941,
54
  "step": 1500
55
  },
56
  {
57
  "epoch": 0.21331058020477817,
58
+ "grad_norm": 0.0223398394882679,
59
  "learning_rate": 4.6444823663253695e-07,
60
+ "loss": 0.1867,
61
  "step": 2000
62
  },
63
  {
64
  "epoch": 0.21331058020477817,
65
+ "eval_loss": 0.8611643314361572,
66
+ "eval_runtime": 63.9649,
67
+ "eval_samples_per_second": 15.43,
68
+ "eval_steps_per_second": 1.939,
69
  "step": 2000
70
  },
71
  {
72
  "epoch": 0.2666382252559727,
73
+ "grad_norm": 175.0380096435547,
74
  "learning_rate": 4.5556029579067116e-07,
75
+ "loss": 0.2068,
76
  "step": 2500
77
  },
78
  {
79
  "epoch": 0.2666382252559727,
80
+ "eval_loss": 0.7947371006011963,
81
+ "eval_runtime": 64.0455,
82
+ "eval_samples_per_second": 15.411,
83
+ "eval_steps_per_second": 1.936,
84
  "step": 2500
85
  },
86
  {
87
  "epoch": 0.3199658703071672,
88
+ "grad_norm": 9.192651084788395e-09,
89
  "learning_rate": 4.4667235494880547e-07,
90
+ "loss": 0.163,
91
  "step": 3000
92
  },
93
  {
94
  "epoch": 0.3199658703071672,
95
+ "eval_loss": 0.7677457332611084,
96
+ "eval_runtime": 64.1593,
97
+ "eval_samples_per_second": 15.384,
98
+ "eval_steps_per_second": 1.933,
99
  "step": 3000
100
  },
101
  {
102
  "epoch": 0.37329351535836175,
103
+ "grad_norm": 0.002516463166102767,
104
  "learning_rate": 4.377844141069397e-07,
105
+ "loss": 0.1729,
106
  "step": 3500
107
  },
108
  {
109
  "epoch": 0.37329351535836175,
110
+ "eval_loss": 0.7635349035263062,
111
+ "eval_runtime": 62.8257,
112
+ "eval_samples_per_second": 15.71,
113
+ "eval_steps_per_second": 1.974,
114
  "step": 3500
115
  },
116
  {
117
  "epoch": 0.42662116040955633,
118
+ "grad_norm": 7.581352710723877,
119
  "learning_rate": 4.2889647326507393e-07,
120
+ "loss": 0.1149,
121
  "step": 4000
122
  },
123
  {
124
  "epoch": 0.42662116040955633,
125
+ "eval_loss": 0.7458549737930298,
126
+ "eval_runtime": 64.2184,
127
+ "eval_samples_per_second": 15.369,
128
+ "eval_steps_per_second": 1.931,
129
  "step": 4000
130
  },
131
  {
132
  "epoch": 0.47994880546075086,
133
+ "grad_norm": 1.3885610103607178,
134
  "learning_rate": 4.2000853242320814e-07,
135
+ "loss": 0.1264,
136
  "step": 4500
137
  },
138
  {
139
  "epoch": 0.47994880546075086,
140
+ "eval_loss": 0.7185410261154175,
141
+ "eval_runtime": 62.6998,
142
+ "eval_samples_per_second": 15.742,
143
+ "eval_steps_per_second": 1.978,
144
  "step": 4500
145
  },
146
  {
147
  "epoch": 0.5332764505119454,
148
+ "grad_norm": 4.22987541242037e-05,
149
  "learning_rate": 4.1112059158134245e-07,
150
+ "loss": 0.1638,
151
  "step": 5000
152
  },
153
  {
154
  "epoch": 0.5332764505119454,
155
+ "eval_loss": 0.7118513584136963,
156
+ "eval_runtime": 63.9674,
157
+ "eval_samples_per_second": 15.43,
158
+ "eval_steps_per_second": 1.938,
159
  "step": 5000
160
  },
161
  {
162
  "epoch": 0.58660409556314,
163
+ "grad_norm": 3.7000099837314337e-06,
164
  "learning_rate": 4.0223265073947665e-07,
165
+ "loss": 0.1287,
166
  "step": 5500
167
  },
168
  {
169
  "epoch": 0.58660409556314,
170
+ "eval_loss": 0.6782647967338562,
171
+ "eval_runtime": 62.7899,
172
+ "eval_samples_per_second": 15.719,
173
+ "eval_steps_per_second": 1.975,
174
  "step": 5500
175
  },
176
  {
177
  "epoch": 0.6399317406143344,
178
+ "grad_norm": 43.72269058227539,
179
  "learning_rate": 3.933447098976109e-07,
180
+ "loss": 0.1324,
181
  "step": 6000
182
  },
183
  {
184
  "epoch": 0.6399317406143344,
185
+ "eval_loss": 0.6554597020149231,
186
+ "eval_runtime": 64.1964,
187
+ "eval_samples_per_second": 15.375,
188
+ "eval_steps_per_second": 1.932,
189
  "step": 6000
190
  },
191
  {
192
  "epoch": 0.693259385665529,
193
+ "grad_norm": 9.483234316576272e-05,
194
  "learning_rate": 3.8445676905574517e-07,
195
+ "loss": 0.1051,
196
  "step": 6500
197
  },
198
  {
199
  "epoch": 0.693259385665529,
200
+ "eval_loss": 0.6227805614471436,
201
+ "eval_runtime": 63.0075,
202
+ "eval_samples_per_second": 15.665,
203
+ "eval_steps_per_second": 1.968,
204
  "step": 6500
205
  },
206
  {
207
  "epoch": 0.7465870307167235,
208
+ "grad_norm": 9.806918797039543e-07,
209
  "learning_rate": 3.755688282138794e-07,
210
+ "loss": 0.1486,
211
  "step": 7000
212
  },
213
  {
214
  "epoch": 0.7465870307167235,
215
+ "eval_loss": 0.6244157552719116,
216
+ "eval_runtime": 62.7149,
217
+ "eval_samples_per_second": 15.738,
218
+ "eval_steps_per_second": 1.977,
219
  "step": 7000
220
  }
221
  ],
checkpoint-7000/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:057f6bbf1ff3e34abda0863071bbf5fdcc7edb1102f04dba05c7a15087bc3396
3
  size 4847
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54c5f1c69a0ead76c168a8c74b80758f8f8eb77228d72de2bfc13a6b8cdc784b
3
  size 4847