kevinoli commited on
Commit
b000a40
1 Parent(s): 005d734

Training in progress, step 9000, checkpoint

Browse files
checkpoint-9000/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bacc106f805d1664efd928c9de16ef7bd05d70e877fc274be1702920584e0b7a
3
  size 1711848436
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d638f870076f59bcf62bc721789ad802073f21550aac7d765264bc60fd66a2e
3
  size 1711848436
checkpoint-9000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:038bccbd5f3cd685d2a5b1c0ab8fbe1b1b4be3becb119348efdf8adeedd1fcba
3
  size 3424043887
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d67b2e9ae7e533e1fd76e1e7c17674d11caac51f4a4fc5710cea07ea31bd5831
3
  size 3424043887
checkpoint-9000/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75e7a66179ff03d27874bf0d0676cd79fe7137c908da603458cd4b183abc4a42
3
  size 14503
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26c99df967694e66808fb8d4e4814367449c627fb6d4991b78dafe1ced9ade3e
3
  size 14503
checkpoint-9000/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:030cdf3785ad8c7848d8573bd5a175cb6354ae7c6ff01cab22da4837592126e3
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5a26111b00cb995c2867625ba8cd506ff6bcfe61cacafd273950252577ca9da
3
  size 623
checkpoint-9000/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 0.5879408121109009,
3
- "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l57-l/checkpoint-8500",
4
- "epoch": 1.658069270449521,
5
  "eval_steps": 500,
6
  "global_step": 9000,
7
  "is_hyper_param_search": false,
@@ -9,278 +9,278 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.09211495946941783,
13
- "grad_norm": 2.6780805587768555,
14
- "learning_rate": 4.84647506755097e-07,
15
- "loss": 0.2705,
16
  "step": 500
17
  },
18
  {
19
- "epoch": 0.09211495946941783,
20
- "eval_loss": 1.0680681467056274,
21
- "eval_runtime": 74.5723,
22
- "eval_samples_per_second": 16.186,
23
- "eval_steps_per_second": 2.025,
24
  "step": 500
25
  },
26
  {
27
- "epoch": 0.18422991893883567,
28
- "grad_norm": 8.301130294799805,
29
- "learning_rate": 4.69295013510194e-07,
30
- "loss": 0.2545,
31
  "step": 1000
32
  },
33
  {
34
- "epoch": 0.18422991893883567,
35
- "eval_loss": 0.9443553686141968,
36
- "eval_runtime": 74.9643,
37
- "eval_samples_per_second": 16.101,
38
- "eval_steps_per_second": 2.014,
39
  "step": 1000
40
  },
41
  {
42
- "epoch": 0.2763448784082535,
43
- "grad_norm": 0.018898434937000275,
44
- "learning_rate": 4.5394252026529107e-07,
45
- "loss": 0.234,
46
  "step": 1500
47
  },
48
  {
49
- "epoch": 0.2763448784082535,
50
- "eval_loss": 0.876915693283081,
51
- "eval_runtime": 75.6401,
52
- "eval_samples_per_second": 15.957,
53
- "eval_steps_per_second": 1.996,
54
  "step": 1500
55
  },
56
  {
57
- "epoch": 0.36845983787767134,
58
- "grad_norm": 0.00010984255641233176,
59
- "learning_rate": 4.385900270203881e-07,
60
- "loss": 0.1539,
61
  "step": 2000
62
  },
63
  {
64
- "epoch": 0.36845983787767134,
65
- "eval_loss": 0.8415330648422241,
66
- "eval_runtime": 76.9959,
67
- "eval_samples_per_second": 15.676,
68
- "eval_steps_per_second": 1.961,
69
  "step": 2000
70
  },
71
  {
72
- "epoch": 0.46057479734708917,
73
- "grad_norm": 395.890625,
74
- "learning_rate": 4.232375337754851e-07,
75
- "loss": 0.1766,
76
  "step": 2500
77
  },
78
  {
79
- "epoch": 0.46057479734708917,
80
- "eval_loss": 0.7660450339317322,
81
- "eval_runtime": 76.733,
82
- "eval_samples_per_second": 15.73,
83
- "eval_steps_per_second": 1.968,
84
  "step": 2500
85
  },
86
  {
87
- "epoch": 0.552689756816507,
88
- "grad_norm": 363.6924743652344,
89
- "learning_rate": 4.0788504053058217e-07,
90
- "loss": 0.1679,
91
  "step": 3000
92
  },
93
  {
94
- "epoch": 0.552689756816507,
95
- "eval_loss": 0.7269155383110046,
96
- "eval_runtime": 76.7928,
97
- "eval_samples_per_second": 15.718,
98
- "eval_steps_per_second": 1.966,
99
  "step": 3000
100
  },
101
  {
102
- "epoch": 0.6448047162859248,
103
- "grad_norm": 28.914026260375977,
104
- "learning_rate": 3.925325472856792e-07,
105
- "loss": 0.1104,
106
  "step": 3500
107
  },
108
  {
109
- "epoch": 0.6448047162859248,
110
- "eval_loss": 0.7097663283348083,
111
- "eval_runtime": 76.6897,
112
- "eval_samples_per_second": 15.739,
113
- "eval_steps_per_second": 1.969,
114
  "step": 3500
115
  },
116
  {
117
- "epoch": 0.7369196757553427,
118
- "grad_norm": 116.85651397705078,
119
- "learning_rate": 3.771800540407762e-07,
120
- "loss": 0.1367,
121
  "step": 4000
122
  },
123
  {
124
- "epoch": 0.7369196757553427,
125
- "eval_loss": 0.6968969106674194,
126
- "eval_runtime": 77.1872,
127
- "eval_samples_per_second": 15.637,
128
- "eval_steps_per_second": 1.956,
129
  "step": 4000
130
  },
131
  {
132
- "epoch": 0.8290346352247605,
133
- "grad_norm": 391.0343322753906,
134
- "learning_rate": 3.618275607958732e-07,
135
- "loss": 0.1129,
136
  "step": 4500
137
  },
138
  {
139
- "epoch": 0.8290346352247605,
140
- "eval_loss": 0.6776800751686096,
141
- "eval_runtime": 75.0351,
142
- "eval_samples_per_second": 16.086,
143
- "eval_steps_per_second": 2.012,
144
  "step": 4500
145
  },
146
  {
147
- "epoch": 0.9211495946941783,
148
- "grad_norm": 2.0514346033451147e-05,
149
- "learning_rate": 3.464750675509703e-07,
150
- "loss": 0.1125,
151
  "step": 5000
152
  },
153
  {
154
- "epoch": 0.9211495946941783,
155
- "eval_loss": 0.6658498048782349,
156
- "eval_runtime": 77.111,
157
- "eval_samples_per_second": 15.653,
158
- "eval_steps_per_second": 1.958,
159
  "step": 5000
160
  },
161
  {
162
- "epoch": 1.013264554163596,
163
- "grad_norm": 0.000568103976547718,
164
- "learning_rate": 3.311225743060673e-07,
165
- "loss": 0.1071,
166
  "step": 5500
167
  },
168
  {
169
- "epoch": 1.013264554163596,
170
- "eval_loss": 0.6465296149253845,
171
- "eval_runtime": 75.2951,
172
- "eval_samples_per_second": 16.03,
173
- "eval_steps_per_second": 2.005,
174
  "step": 5500
175
  },
176
  {
177
- "epoch": 1.105379513633014,
178
- "grad_norm": 0.02698369137942791,
179
- "learning_rate": 3.157700810611643e-07,
180
- "loss": 0.0553,
181
  "step": 6000
182
  },
183
  {
184
- "epoch": 1.105379513633014,
185
- "eval_loss": 0.6357071995735168,
186
- "eval_runtime": 76.6726,
187
- "eval_samples_per_second": 15.742,
188
- "eval_steps_per_second": 1.969,
189
  "step": 6000
190
  },
191
  {
192
- "epoch": 1.1974944731024317,
193
- "grad_norm": 0.0005627564387395978,
194
- "learning_rate": 3.0041758781626137e-07,
195
- "loss": 0.0729,
196
  "step": 6500
197
  },
198
  {
199
- "epoch": 1.1974944731024317,
200
- "eval_loss": 0.628384530544281,
201
- "eval_runtime": 75.6565,
202
- "eval_samples_per_second": 15.954,
203
- "eval_steps_per_second": 1.996,
204
  "step": 6500
205
  },
206
  {
207
- "epoch": 1.2896094325718497,
208
- "grad_norm": 4.7338733111246256e-07,
209
- "learning_rate": 2.8506509457135833e-07,
210
- "loss": 0.0476,
211
  "step": 7000
212
  },
213
  {
214
- "epoch": 1.2896094325718497,
215
- "eval_loss": 0.6259528994560242,
216
- "eval_runtime": 75.2674,
217
- "eval_samples_per_second": 16.036,
218
- "eval_steps_per_second": 2.006,
219
  "step": 7000
220
  },
221
  {
222
- "epoch": 1.3817243920412676,
223
- "grad_norm": 28.43704605102539,
224
- "learning_rate": 2.697126013264554e-07,
225
- "loss": 0.0756,
226
  "step": 7500
227
  },
228
  {
229
- "epoch": 1.3817243920412676,
230
- "eval_loss": 0.6102285981178284,
231
- "eval_runtime": 76.1062,
232
- "eval_samples_per_second": 15.859,
233
- "eval_steps_per_second": 1.984,
234
  "step": 7500
235
  },
236
  {
237
- "epoch": 1.4738393515106853,
238
- "grad_norm": 0.0002958567056339234,
239
- "learning_rate": 2.5436010808155247e-07,
240
- "loss": 0.0797,
241
  "step": 8000
242
  },
243
  {
244
- "epoch": 1.4738393515106853,
245
- "eval_loss": 0.6023213267326355,
246
- "eval_runtime": 75.1644,
247
- "eval_samples_per_second": 16.058,
248
- "eval_steps_per_second": 2.009,
249
  "step": 8000
250
  },
251
  {
252
- "epoch": 1.565954310980103,
253
- "grad_norm": 0.0021068258211016655,
254
- "learning_rate": 2.390076148366495e-07,
255
- "loss": 0.0536,
256
  "step": 8500
257
  },
258
  {
259
- "epoch": 1.565954310980103,
260
- "eval_loss": 0.5879408121109009,
261
- "eval_runtime": 76.0196,
262
- "eval_samples_per_second": 15.877,
263
- "eval_steps_per_second": 1.986,
264
  "step": 8500
265
  },
266
  {
267
- "epoch": 1.658069270449521,
268
- "grad_norm": 4.1152270569000393e-07,
269
- "learning_rate": 2.236551215917465e-07,
270
- "loss": 0.0784,
271
  "step": 9000
272
  },
273
  {
274
- "epoch": 1.658069270449521,
275
- "eval_loss": 0.5879626274108887,
276
- "eval_runtime": 75.6453,
277
- "eval_samples_per_second": 15.956,
278
- "eval_steps_per_second": 1.996,
279
  "step": 9000
280
  }
281
  ],
282
  "logging_steps": 500,
283
- "max_steps": 16284,
284
  "num_input_tokens_seen": 0,
285
  "num_train_epochs": 3,
286
  "save_steps": 500,
 
1
  {
2
+ "best_metric": 0.5558739304542542,
3
+ "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e3l57-l/checkpoint-9000",
4
+ "epoch": 0.9598976109215017,
5
  "eval_steps": 500,
6
  "global_step": 9000,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.05332764505119454,
13
+ "grad_norm": 414.62841796875,
14
+ "learning_rate": 4.911120591581342e-07,
15
+ "loss": 0.3812,
16
  "step": 500
17
  },
18
  {
19
+ "epoch": 0.05332764505119454,
20
+ "eval_loss": 1.1163015365600586,
21
+ "eval_runtime": 61.8962,
22
+ "eval_samples_per_second": 15.946,
23
+ "eval_steps_per_second": 2.003,
24
  "step": 500
25
  },
26
  {
27
+ "epoch": 0.10665529010238908,
28
+ "grad_norm": 34.08418655395508,
29
+ "learning_rate": 4.822241183162685e-07,
30
+ "loss": 0.2683,
31
  "step": 1000
32
  },
33
  {
34
+ "epoch": 0.10665529010238908,
35
+ "eval_loss": 0.9684447646141052,
36
+ "eval_runtime": 61.821,
37
+ "eval_samples_per_second": 15.965,
38
+ "eval_steps_per_second": 2.006,
39
  "step": 1000
40
  },
41
  {
42
+ "epoch": 0.1599829351535836,
43
+ "grad_norm": 350.4574890136719,
44
+ "learning_rate": 4.733361774744027e-07,
45
+ "loss": 0.2119,
46
  "step": 1500
47
  },
48
  {
49
+ "epoch": 0.1599829351535836,
50
+ "eval_loss": 0.9099885821342468,
51
+ "eval_runtime": 61.9606,
52
+ "eval_samples_per_second": 15.929,
53
+ "eval_steps_per_second": 2.001,
54
  "step": 1500
55
  },
56
  {
57
+ "epoch": 0.21331058020477817,
58
+ "grad_norm": 0.01787523180246353,
59
+ "learning_rate": 4.6444823663253695e-07,
60
+ "loss": 0.1889,
61
  "step": 2000
62
  },
63
  {
64
+ "epoch": 0.21331058020477817,
65
+ "eval_loss": 0.8620074987411499,
66
+ "eval_runtime": 63.2824,
67
+ "eval_samples_per_second": 15.597,
68
+ "eval_steps_per_second": 1.959,
69
  "step": 2000
70
  },
71
  {
72
+ "epoch": 0.2666382252559727,
73
+ "grad_norm": 118.66907501220703,
74
+ "learning_rate": 4.5556029579067116e-07,
75
+ "loss": 0.2071,
76
  "step": 2500
77
  },
78
  {
79
+ "epoch": 0.2666382252559727,
80
+ "eval_loss": 0.791786253452301,
81
+ "eval_runtime": 63.7932,
82
+ "eval_samples_per_second": 15.472,
83
+ "eval_steps_per_second": 1.944,
84
  "step": 2500
85
  },
86
  {
87
+ "epoch": 0.3199658703071672,
88
+ "grad_norm": 7.21087323185543e-09,
89
+ "learning_rate": 4.4667235494880547e-07,
90
+ "loss": 0.1588,
91
  "step": 3000
92
  },
93
  {
94
+ "epoch": 0.3199658703071672,
95
+ "eval_loss": 0.7657254338264465,
96
+ "eval_runtime": 62.4738,
97
+ "eval_samples_per_second": 15.799,
98
+ "eval_steps_per_second": 1.985,
99
  "step": 3000
100
  },
101
  {
102
+ "epoch": 0.37329351535836175,
103
+ "grad_norm": 0.0014836661284789443,
104
+ "learning_rate": 4.377844141069397e-07,
105
+ "loss": 0.1718,
106
  "step": 3500
107
  },
108
  {
109
+ "epoch": 0.37329351535836175,
110
+ "eval_loss": 0.7610095739364624,
111
+ "eval_runtime": 63.9118,
112
+ "eval_samples_per_second": 15.443,
113
+ "eval_steps_per_second": 1.94,
114
  "step": 3500
115
  },
116
  {
117
+ "epoch": 0.42662116040955633,
118
+ "grad_norm": 3.2739710807800293,
119
+ "learning_rate": 4.2889647326507393e-07,
120
+ "loss": 0.1113,
121
  "step": 4000
122
  },
123
  {
124
+ "epoch": 0.42662116040955633,
125
+ "eval_loss": 0.7458393573760986,
126
+ "eval_runtime": 62.1803,
127
+ "eval_samples_per_second": 15.873,
128
+ "eval_steps_per_second": 1.994,
129
  "step": 4000
130
  },
131
  {
132
+ "epoch": 0.47994880546075086,
133
+ "grad_norm": 1.788787841796875,
134
+ "learning_rate": 4.2000853242320814e-07,
135
+ "loss": 0.1313,
136
  "step": 4500
137
  },
138
  {
139
+ "epoch": 0.47994880546075086,
140
+ "eval_loss": 0.7168156504631042,
141
+ "eval_runtime": 62.1477,
142
+ "eval_samples_per_second": 15.882,
143
+ "eval_steps_per_second": 1.995,
144
  "step": 4500
145
  },
146
  {
147
+ "epoch": 0.5332764505119454,
148
+ "grad_norm": 4.031916250823997e-05,
149
+ "learning_rate": 4.1112059158134245e-07,
150
+ "loss": 0.1649,
151
  "step": 5000
152
  },
153
  {
154
+ "epoch": 0.5332764505119454,
155
+ "eval_loss": 0.7019046545028687,
156
+ "eval_runtime": 62.2092,
157
+ "eval_samples_per_second": 15.866,
158
+ "eval_steps_per_second": 1.993,
159
  "step": 5000
160
  },
161
  {
162
+ "epoch": 0.58660409556314,
163
+ "grad_norm": 3.883213139488362e-06,
164
+ "learning_rate": 4.0223265073947665e-07,
165
+ "loss": 0.1245,
166
  "step": 5500
167
  },
168
  {
169
+ "epoch": 0.58660409556314,
170
+ "eval_loss": 0.6812178492546082,
171
+ "eval_runtime": 62.492,
172
+ "eval_samples_per_second": 15.794,
173
+ "eval_steps_per_second": 1.984,
174
  "step": 5500
175
  },
176
  {
177
+ "epoch": 0.6399317406143344,
178
+ "grad_norm": 80.17021942138672,
179
+ "learning_rate": 3.933447098976109e-07,
180
+ "loss": 0.1286,
181
  "step": 6000
182
  },
183
  {
184
+ "epoch": 0.6399317406143344,
185
+ "eval_loss": 0.6501584649085999,
186
+ "eval_runtime": 62.5236,
187
+ "eval_samples_per_second": 15.786,
188
+ "eval_steps_per_second": 1.983,
189
  "step": 6000
190
  },
191
  {
192
+ "epoch": 0.693259385665529,
193
+ "grad_norm": 0.0001418297761119902,
194
+ "learning_rate": 3.8445676905574517e-07,
195
+ "loss": 0.1076,
196
  "step": 6500
197
  },
198
  {
199
+ "epoch": 0.693259385665529,
200
+ "eval_loss": 0.6153982281684875,
201
+ "eval_runtime": 62.2927,
202
+ "eval_samples_per_second": 15.845,
203
+ "eval_steps_per_second": 1.991,
204
  "step": 6500
205
  },
206
  {
207
+ "epoch": 0.7465870307167235,
208
+ "grad_norm": 5.804526495012396e-07,
209
+ "learning_rate": 3.755688282138794e-07,
210
+ "loss": 0.1477,
211
  "step": 7000
212
  },
213
  {
214
+ "epoch": 0.7465870307167235,
215
+ "eval_loss": 0.6118016839027405,
216
+ "eval_runtime": 63.1588,
217
+ "eval_samples_per_second": 15.627,
218
+ "eval_steps_per_second": 1.963,
219
  "step": 7000
220
  },
221
  {
222
+ "epoch": 0.7999146757679181,
223
+ "grad_norm": 2.66489315032959,
224
+ "learning_rate": 3.6668088737201363e-07,
225
+ "loss": 0.1315,
226
  "step": 7500
227
  },
228
  {
229
+ "epoch": 0.7999146757679181,
230
+ "eval_loss": 0.601563036441803,
231
+ "eval_runtime": 62.9163,
232
+ "eval_samples_per_second": 15.688,
233
+ "eval_steps_per_second": 1.971,
234
  "step": 7500
235
  },
236
  {
237
+ "epoch": 0.8532423208191127,
238
+ "grad_norm": 851.2547607421875,
239
+ "learning_rate": 3.5779294653014783e-07,
240
+ "loss": 0.1413,
241
  "step": 8000
242
  },
243
  {
244
+ "epoch": 0.8532423208191127,
245
+ "eval_loss": 0.584888756275177,
246
+ "eval_runtime": 62.6735,
247
+ "eval_samples_per_second": 15.748,
248
+ "eval_steps_per_second": 1.979,
249
  "step": 8000
250
  },
251
  {
252
+ "epoch": 0.9065699658703071,
253
+ "grad_norm": 6.041275628376752e-05,
254
+ "learning_rate": 3.4890500568828214e-07,
255
+ "loss": 0.124,
256
  "step": 8500
257
  },
258
  {
259
+ "epoch": 0.9065699658703071,
260
+ "eval_loss": 0.5766209363937378,
261
+ "eval_runtime": 63.3493,
262
+ "eval_samples_per_second": 15.58,
263
+ "eval_steps_per_second": 1.957,
264
  "step": 8500
265
  },
266
  {
267
+ "epoch": 0.9598976109215017,
268
+ "grad_norm": 3.0901598930358887,
269
+ "learning_rate": 3.4001706484641635e-07,
270
+ "loss": 0.1215,
271
  "step": 9000
272
  },
273
  {
274
+ "epoch": 0.9598976109215017,
275
+ "eval_loss": 0.5558739304542542,
276
+ "eval_runtime": 63.0862,
277
+ "eval_samples_per_second": 15.645,
278
+ "eval_steps_per_second": 1.966,
279
  "step": 9000
280
  }
281
  ],
282
  "logging_steps": 500,
283
+ "max_steps": 28128,
284
  "num_input_tokens_seen": 0,
285
  "num_train_epochs": 3,
286
  "save_steps": 500,
checkpoint-9000/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ea34fd4333686a44a60d290faa4ae8e00df783131842d9ae3ef1b9e874a39349
3
  size 4847
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd16a1abd112ae60a2700f590eb3bd73e96acd9a0a37b30bbfa3f8ee2dc23bc1
3
  size 4847