Commit
•
c4f012d
1
Parent(s):
2d6adc1
🍻 cheers
Browse files- README.md +10 -6
- all_results.json +16 -0
- config.json +1 -1
- eval_results.json +12 -0
- train_results.json +8 -0
- trainer_state.json +349 -0
README.md
CHANGED
@@ -1,7 +1,11 @@
|
|
1 |
---
|
|
|
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/dinov2-base
|
4 |
tags:
|
|
|
|
|
5 |
- generated_from_trainer
|
6 |
metrics:
|
7 |
- accuracy
|
@@ -15,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# dino-base-2023_12_18-kornia_img-size518_batch-size32_epochs20
|
17 |
|
18 |
-
This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on the
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 0.
|
21 |
-
- F1 Micro: 0.
|
22 |
-
- F1 Macro: 0.
|
23 |
-
- Roc Auc: 0.
|
24 |
-
- Accuracy: 0.
|
25 |
- Learning Rate: 0.01
|
26 |
|
27 |
## Model description
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- eng
|
4 |
license: apache-2.0
|
5 |
base_model: facebook/dinov2-base
|
6 |
tags:
|
7 |
+
- multilabel-image-classification
|
8 |
+
- multilabel
|
9 |
- generated_from_trainer
|
10 |
metrics:
|
11 |
- accuracy
|
|
|
19 |
|
20 |
# dino-base-2023_12_18-kornia_img-size518_batch-size32_epochs20
|
21 |
|
22 |
+
This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on the multilabel_complete_dataset dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
+
- Loss: 0.1422
|
25 |
+
- F1 Micro: 0.7786
|
26 |
+
- F1 Macro: 0.7231
|
27 |
+
- Roc Auc: 0.8542
|
28 |
+
- Accuracy: 0.4586
|
29 |
- Learning Rate: 0.01
|
30 |
|
31 |
## Model description
|
all_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"eval_accuracy": 0.4586413234776487,
|
4 |
+
"eval_f1_macro": 0.7230670728714794,
|
5 |
+
"eval_f1_micro": 0.7785516178736519,
|
6 |
+
"eval_loss": 0.14224284887313843,
|
7 |
+
"eval_roc_auc": 0.854225641152639,
|
8 |
+
"eval_runtime": 453.9537,
|
9 |
+
"eval_samples_per_second": 6.258,
|
10 |
+
"eval_steps_per_second": 0.196,
|
11 |
+
"learning_rate": 0.01,
|
12 |
+
"train_loss": 0.19166865989343443,
|
13 |
+
"train_runtime": 38664.1493,
|
14 |
+
"train_samples_per_second": 4.432,
|
15 |
+
"train_steps_per_second": 0.139
|
16 |
+
}
|
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "facebook/dinov2-
|
3 |
"apply_layernorm": true,
|
4 |
"architectures": [
|
5 |
"NewheadDinov2ForImageClassification"
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "facebook/dinov2-base2023_12_18",
|
3 |
"apply_layernorm": true,
|
4 |
"architectures": [
|
5 |
"NewheadDinov2ForImageClassification"
|
eval_results.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"eval_accuracy": 0.4586413234776487,
|
4 |
+
"eval_f1_macro": 0.7230670728714794,
|
5 |
+
"eval_f1_micro": 0.7785516178736519,
|
6 |
+
"eval_loss": 0.14224284887313843,
|
7 |
+
"eval_roc_auc": 0.854225641152639,
|
8 |
+
"eval_runtime": 453.9537,
|
9 |
+
"eval_samples_per_second": 6.258,
|
10 |
+
"eval_steps_per_second": 0.196,
|
11 |
+
"learning_rate": 0.01
|
12 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"learning_rate": 0.01,
|
4 |
+
"train_loss": 0.19166865989343443,
|
5 |
+
"train_runtime": 38664.1493,
|
6 |
+
"train_samples_per_second": 4.432,
|
7 |
+
"train_steps_per_second": 0.139
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,349 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.1429145485162735,
|
3 |
+
"best_model_checkpoint": "/home1/datawork/mcontini/models/multilabel/huggingface/dino-base-2023_12_18-kornia_img-size518_batch-size32_epochs20/checkpoint-4288",
|
4 |
+
"epoch": 20.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 5360,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 1.0,
|
13 |
+
"eval_accuracy": 0.40836012861736337,
|
14 |
+
"eval_f1_macro": 0.5724832983501338,
|
15 |
+
"eval_f1_micro": 0.66917502787068,
|
16 |
+
"eval_loss": 0.18527662754058838,
|
17 |
+
"eval_roc_auc": 0.7700155485850371,
|
18 |
+
"eval_runtime": 452.1812,
|
19 |
+
"eval_samples_per_second": 6.19,
|
20 |
+
"eval_steps_per_second": 0.195,
|
21 |
+
"learning_rate": 0.01,
|
22 |
+
"step": 268
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"epoch": 1.87,
|
26 |
+
"learning_rate": 0.01,
|
27 |
+
"loss": 0.2132,
|
28 |
+
"step": 500
|
29 |
+
},
|
30 |
+
{
|
31 |
+
"epoch": 2.0,
|
32 |
+
"eval_accuracy": 0.41336191496963204,
|
33 |
+
"eval_f1_macro": 0.7066155100478595,
|
34 |
+
"eval_f1_micro": 0.7630167597765364,
|
35 |
+
"eval_loss": 0.1677507609128952,
|
36 |
+
"eval_roc_auc": 0.8728909820573381,
|
37 |
+
"eval_runtime": 452.3557,
|
38 |
+
"eval_samples_per_second": 6.188,
|
39 |
+
"eval_steps_per_second": 0.195,
|
40 |
+
"learning_rate": 0.01,
|
41 |
+
"step": 536
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 3.0,
|
45 |
+
"eval_accuracy": 0.4540907466952483,
|
46 |
+
"eval_f1_macro": 0.7007992492814997,
|
47 |
+
"eval_f1_micro": 0.7664760009947775,
|
48 |
+
"eval_loss": 0.15684403479099274,
|
49 |
+
"eval_roc_auc": 0.8456356297999449,
|
50 |
+
"eval_runtime": 449.5285,
|
51 |
+
"eval_samples_per_second": 6.227,
|
52 |
+
"eval_steps_per_second": 0.196,
|
53 |
+
"learning_rate": 0.01,
|
54 |
+
"step": 804
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 3.73,
|
58 |
+
"learning_rate": 0.01,
|
59 |
+
"loss": 0.1913,
|
60 |
+
"step": 1000
|
61 |
+
},
|
62 |
+
{
|
63 |
+
"epoch": 4.0,
|
64 |
+
"eval_accuracy": 0.45087531261164704,
|
65 |
+
"eval_f1_macro": 0.7401496915878392,
|
66 |
+
"eval_f1_micro": 0.7900162149640955,
|
67 |
+
"eval_loss": 0.14960218966007233,
|
68 |
+
"eval_roc_auc": 0.8786412916916798,
|
69 |
+
"eval_runtime": 447.0415,
|
70 |
+
"eval_samples_per_second": 6.261,
|
71 |
+
"eval_steps_per_second": 0.197,
|
72 |
+
"learning_rate": 0.01,
|
73 |
+
"step": 1072
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 5.0,
|
77 |
+
"eval_accuracy": 0.4090746695248303,
|
78 |
+
"eval_f1_macro": 0.7354739664497006,
|
79 |
+
"eval_f1_micro": 0.7692135710692527,
|
80 |
+
"eval_loss": 0.16684336960315704,
|
81 |
+
"eval_roc_auc": 0.8765264813882132,
|
82 |
+
"eval_runtime": 449.4387,
|
83 |
+
"eval_samples_per_second": 6.228,
|
84 |
+
"eval_steps_per_second": 0.196,
|
85 |
+
"learning_rate": 0.01,
|
86 |
+
"step": 1340
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 5.6,
|
90 |
+
"learning_rate": 0.01,
|
91 |
+
"loss": 0.1899,
|
92 |
+
"step": 1500
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 6.0,
|
96 |
+
"eval_accuracy": 0.46052161486245086,
|
97 |
+
"eval_f1_macro": 0.6472064428456584,
|
98 |
+
"eval_f1_micro": 0.7618986529826812,
|
99 |
+
"eval_loss": 0.1519184708595276,
|
100 |
+
"eval_roc_auc": 0.8352141402262778,
|
101 |
+
"eval_runtime": 452.3131,
|
102 |
+
"eval_samples_per_second": 6.188,
|
103 |
+
"eval_steps_per_second": 0.195,
|
104 |
+
"learning_rate": 0.01,
|
105 |
+
"step": 1608
|
106 |
+
},
|
107 |
+
{
|
108 |
+
"epoch": 7.0,
|
109 |
+
"eval_accuracy": 0.4390853876384423,
|
110 |
+
"eval_f1_macro": 0.6880839155457914,
|
111 |
+
"eval_f1_micro": 0.7724533380141596,
|
112 |
+
"eval_loss": 0.15900948643684387,
|
113 |
+
"eval_roc_auc": 0.8654028748541202,
|
114 |
+
"eval_runtime": 449.1274,
|
115 |
+
"eval_samples_per_second": 6.232,
|
116 |
+
"eval_steps_per_second": 0.196,
|
117 |
+
"learning_rate": 0.01,
|
118 |
+
"step": 1876
|
119 |
+
},
|
120 |
+
{
|
121 |
+
"epoch": 7.46,
|
122 |
+
"learning_rate": 0.01,
|
123 |
+
"loss": 0.188,
|
124 |
+
"step": 2000
|
125 |
+
},
|
126 |
+
{
|
127 |
+
"epoch": 8.0,
|
128 |
+
"eval_accuracy": 0.44587352625937837,
|
129 |
+
"eval_f1_macro": 0.6946270568821779,
|
130 |
+
"eval_f1_micro": 0.7812088569718731,
|
131 |
+
"eval_loss": 0.14899832010269165,
|
132 |
+
"eval_roc_auc": 0.8641522922073566,
|
133 |
+
"eval_runtime": 449.6916,
|
134 |
+
"eval_samples_per_second": 6.224,
|
135 |
+
"eval_steps_per_second": 0.196,
|
136 |
+
"learning_rate": 0.01,
|
137 |
+
"step": 2144
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 9.0,
|
141 |
+
"eval_accuracy": 0.4669524830296534,
|
142 |
+
"eval_f1_macro": 0.7114660813932495,
|
143 |
+
"eval_f1_micro": 0.7886651925986268,
|
144 |
+
"eval_loss": 0.14934837818145752,
|
145 |
+
"eval_roc_auc": 0.8764578257293321,
|
146 |
+
"eval_runtime": 456.6917,
|
147 |
+
"eval_samples_per_second": 6.129,
|
148 |
+
"eval_steps_per_second": 0.193,
|
149 |
+
"learning_rate": 0.01,
|
150 |
+
"step": 2412
|
151 |
+
},
|
152 |
+
{
|
153 |
+
"epoch": 9.33,
|
154 |
+
"learning_rate": 0.01,
|
155 |
+
"loss": 0.1888,
|
156 |
+
"step": 2500
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 10.0,
|
160 |
+
"eval_accuracy": 0.47195426938192214,
|
161 |
+
"eval_f1_macro": 0.7013528531179296,
|
162 |
+
"eval_f1_micro": 0.7743778567800914,
|
163 |
+
"eval_loss": 0.14436914026737213,
|
164 |
+
"eval_roc_auc": 0.8445106252570078,
|
165 |
+
"eval_runtime": 457.228,
|
166 |
+
"eval_samples_per_second": 6.122,
|
167 |
+
"eval_steps_per_second": 0.192,
|
168 |
+
"learning_rate": 0.01,
|
169 |
+
"step": 2680
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 11.0,
|
173 |
+
"eval_accuracy": 0.4347981421936406,
|
174 |
+
"eval_f1_macro": 0.6894667259492441,
|
175 |
+
"eval_f1_micro": 0.7651653850848285,
|
176 |
+
"eval_loss": 0.158202663064003,
|
177 |
+
"eval_roc_auc": 0.8497870587564459,
|
178 |
+
"eval_runtime": 451.0784,
|
179 |
+
"eval_samples_per_second": 6.205,
|
180 |
+
"eval_steps_per_second": 0.195,
|
181 |
+
"learning_rate": 0.01,
|
182 |
+
"step": 2948
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 11.19,
|
186 |
+
"learning_rate": 0.01,
|
187 |
+
"loss": 0.1888,
|
188 |
+
"step": 3000
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 12.0,
|
192 |
+
"eval_accuracy": 0.4615934262236513,
|
193 |
+
"eval_f1_macro": 0.6945655735071663,
|
194 |
+
"eval_f1_micro": 0.7490938380990738,
|
195 |
+
"eval_loss": 0.15359686315059662,
|
196 |
+
"eval_roc_auc": 0.8176277103144143,
|
197 |
+
"eval_runtime": 456.0533,
|
198 |
+
"eval_samples_per_second": 6.137,
|
199 |
+
"eval_steps_per_second": 0.193,
|
200 |
+
"learning_rate": 0.01,
|
201 |
+
"step": 3216
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 13.0,
|
205 |
+
"eval_accuracy": 0.4555198285101822,
|
206 |
+
"eval_f1_macro": 0.691978470964451,
|
207 |
+
"eval_f1_micro": 0.7727919787089188,
|
208 |
+
"eval_loss": 0.15135115385055542,
|
209 |
+
"eval_roc_auc": 0.8502953769902751,
|
210 |
+
"eval_runtime": 449.1842,
|
211 |
+
"eval_samples_per_second": 6.231,
|
212 |
+
"eval_steps_per_second": 0.196,
|
213 |
+
"learning_rate": 0.01,
|
214 |
+
"step": 3484
|
215 |
+
},
|
216 |
+
{
|
217 |
+
"epoch": 13.06,
|
218 |
+
"learning_rate": 0.01,
|
219 |
+
"loss": 0.1886,
|
220 |
+
"step": 3500
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 14.0,
|
224 |
+
"eval_accuracy": 0.4355126831011075,
|
225 |
+
"eval_f1_macro": 0.5592534839805278,
|
226 |
+
"eval_f1_micro": 0.686313973548016,
|
227 |
+
"eval_loss": 0.16679102182388306,
|
228 |
+
"eval_roc_auc": 0.7724894168127243,
|
229 |
+
"eval_runtime": 450.5535,
|
230 |
+
"eval_samples_per_second": 6.212,
|
231 |
+
"eval_steps_per_second": 0.195,
|
232 |
+
"learning_rate": 0.01,
|
233 |
+
"step": 3752
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 14.93,
|
237 |
+
"learning_rate": 0.01,
|
238 |
+
"loss": 0.1906,
|
239 |
+
"step": 4000
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 15.0,
|
243 |
+
"eval_accuracy": 0.4533762057877814,
|
244 |
+
"eval_f1_macro": 0.6660464840620275,
|
245 |
+
"eval_f1_micro": 0.7589447960398522,
|
246 |
+
"eval_loss": 0.15235908329486847,
|
247 |
+
"eval_roc_auc": 0.8395164037430609,
|
248 |
+
"eval_runtime": 453.9816,
|
249 |
+
"eval_samples_per_second": 6.165,
|
250 |
+
"eval_steps_per_second": 0.194,
|
251 |
+
"learning_rate": 0.01,
|
252 |
+
"step": 4020
|
253 |
+
},
|
254 |
+
{
|
255 |
+
"epoch": 16.0,
|
256 |
+
"eval_accuracy": 0.47624151482672383,
|
257 |
+
"eval_f1_macro": 0.7239567261315109,
|
258 |
+
"eval_f1_micro": 0.784867024597328,
|
259 |
+
"eval_loss": 0.1429145485162735,
|
260 |
+
"eval_roc_auc": 0.8546121997155703,
|
261 |
+
"eval_runtime": 451.7346,
|
262 |
+
"eval_samples_per_second": 6.196,
|
263 |
+
"eval_steps_per_second": 0.195,
|
264 |
+
"learning_rate": 0.01,
|
265 |
+
"step": 4288
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 16.79,
|
269 |
+
"learning_rate": 0.01,
|
270 |
+
"loss": 0.1879,
|
271 |
+
"step": 4500
|
272 |
+
},
|
273 |
+
{
|
274 |
+
"epoch": 17.0,
|
275 |
+
"eval_accuracy": 0.4122901036084316,
|
276 |
+
"eval_f1_macro": 0.6093139734586531,
|
277 |
+
"eval_f1_micro": 0.7452841241944933,
|
278 |
+
"eval_loss": 0.17112578451633453,
|
279 |
+
"eval_roc_auc": 0.8492466394980158,
|
280 |
+
"eval_runtime": 451.5598,
|
281 |
+
"eval_samples_per_second": 6.199,
|
282 |
+
"eval_steps_per_second": 0.195,
|
283 |
+
"learning_rate": 0.01,
|
284 |
+
"step": 4556
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 18.0,
|
288 |
+
"eval_accuracy": 0.43729903536977494,
|
289 |
+
"eval_f1_macro": 0.5856617023433228,
|
290 |
+
"eval_f1_micro": 0.7304124060404957,
|
291 |
+
"eval_loss": 0.15878728032112122,
|
292 |
+
"eval_roc_auc": 0.8061813000254404,
|
293 |
+
"eval_runtime": 453.038,
|
294 |
+
"eval_samples_per_second": 6.178,
|
295 |
+
"eval_steps_per_second": 0.194,
|
296 |
+
"learning_rate": 0.01,
|
297 |
+
"step": 4824
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 18.66,
|
301 |
+
"learning_rate": 0.01,
|
302 |
+
"loss": 0.1888,
|
303 |
+
"step": 5000
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 19.0,
|
307 |
+
"eval_accuracy": 0.4465880671668453,
|
308 |
+
"eval_f1_macro": 0.69497597215645,
|
309 |
+
"eval_f1_micro": 0.7428348605966075,
|
310 |
+
"eval_loss": 0.16344697773456573,
|
311 |
+
"eval_roc_auc": 0.821961971607228,
|
312 |
+
"eval_runtime": 451.5696,
|
313 |
+
"eval_samples_per_second": 6.198,
|
314 |
+
"eval_steps_per_second": 0.195,
|
315 |
+
"learning_rate": 0.01,
|
316 |
+
"step": 5092
|
317 |
+
},
|
318 |
+
{
|
319 |
+
"epoch": 20.0,
|
320 |
+
"eval_accuracy": 0.4644515898535191,
|
321 |
+
"eval_f1_macro": 0.7272077268448706,
|
322 |
+
"eval_f1_micro": 0.7788184879093969,
|
323 |
+
"eval_loss": 0.1484629064798355,
|
324 |
+
"eval_roc_auc": 0.856568523463588,
|
325 |
+
"eval_runtime": 450.0153,
|
326 |
+
"eval_samples_per_second": 6.22,
|
327 |
+
"eval_steps_per_second": 0.196,
|
328 |
+
"learning_rate": 0.01,
|
329 |
+
"step": 5360
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 20.0,
|
333 |
+
"learning_rate": 0.01,
|
334 |
+
"step": 5360,
|
335 |
+
"total_flos": 7.288742611336668e+19,
|
336 |
+
"train_loss": 0.19166865989343443,
|
337 |
+
"train_runtime": 38664.1493,
|
338 |
+
"train_samples_per_second": 4.432,
|
339 |
+
"train_steps_per_second": 0.139
|
340 |
+
}
|
341 |
+
],
|
342 |
+
"logging_steps": 500,
|
343 |
+
"max_steps": 5360,
|
344 |
+
"num_train_epochs": 20,
|
345 |
+
"save_steps": 500,
|
346 |
+
"total_flos": 7.288742611336668e+19,
|
347 |
+
"trial_name": null,
|
348 |
+
"trial_params": null
|
349 |
+
}
|