Commit
•
66879e1
1
Parent(s):
695e79e
🍻 cheers
Browse files- README.md +10 -6
- all_results.json +16 -0
- config.json +1 -1
- eval_results.json +12 -0
- runs/Jan16_07-24-46_datavisu2/events.out.tfevents.1705407350.datavisu2.52767.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +351 -0
README.md
CHANGED
@@ -1,7 +1,11 @@
|
|
1 |
---
|
|
|
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/dinov2-large
|
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 |
# dinov2-large-2024_01_14-without_data_aug_batch-size32_epochs20_freeze
|
17 |
|
18 |
-
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) 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: 0.0001
|
26 |
|
27 |
## Model description
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- eng
|
4 |
license: apache-2.0
|
5 |
base_model: facebook/dinov2-large
|
6 |
tags:
|
7 |
+
- multilabel-image-classification
|
8 |
+
- multilabel
|
9 |
- generated_from_trainer
|
10 |
metrics:
|
11 |
- accuracy
|
|
|
19 |
|
20 |
# dinov2-large-2024_01_14-without_data_aug_batch-size32_epochs20_freeze
|
21 |
|
22 |
+
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the multilabel_complete_dataset dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
+
- Loss: 0.0904
|
25 |
+
- F1 Micro: 0.8447
|
26 |
+
- F1 Macro: 0.7156
|
27 |
+
- Roc Auc: 0.9011
|
28 |
+
- Accuracy: 0.5459
|
29 |
- Learning Rate: 0.0001
|
30 |
|
31 |
## Model description
|
all_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"eval_accuracy": 0.5458683473389355,
|
4 |
+
"eval_f1_macro": 0.7156058376343973,
|
5 |
+
"eval_f1_micro": 0.8447395301327885,
|
6 |
+
"eval_loss": 0.09043727070093155,
|
7 |
+
"eval_roc_auc": 0.9011367813249752,
|
8 |
+
"eval_runtime": 688.8904,
|
9 |
+
"eval_samples_per_second": 4.146,
|
10 |
+
"eval_steps_per_second": 0.131,
|
11 |
+
"learning_rate": 0.0001,
|
12 |
+
"train_loss": 0.03709733433967089,
|
13 |
+
"train_runtime": 20246.0375,
|
14 |
+
"train_samples_per_second": 8.66,
|
15 |
+
"train_steps_per_second": 0.271
|
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-large2024_01_16",
|
3 |
"apply_layernorm": true,
|
4 |
"architectures": [
|
5 |
"NewheadDinov2ForImageClassification"
|
eval_results.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"eval_accuracy": 0.5458683473389355,
|
4 |
+
"eval_f1_macro": 0.7156058376343973,
|
5 |
+
"eval_f1_micro": 0.8447395301327885,
|
6 |
+
"eval_loss": 0.09043727070093155,
|
7 |
+
"eval_roc_auc": 0.9011367813249752,
|
8 |
+
"eval_runtime": 688.8904,
|
9 |
+
"eval_samples_per_second": 4.146,
|
10 |
+
"eval_steps_per_second": 0.131,
|
11 |
+
"learning_rate": 0.0001
|
12 |
+
}
|
runs/Jan16_07-24-46_datavisu2/events.out.tfevents.1705407350.datavisu2.52767.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c2f507aa669562943a7b7adfaa8134cae603525e7721c16afef134cc74ee0842
|
3 |
+
size 624
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 20.0,
|
3 |
+
"learning_rate": 0.0001,
|
4 |
+
"train_loss": 0.03709733433967089,
|
5 |
+
"train_runtime": 20246.0375,
|
6 |
+
"train_samples_per_second": 8.66,
|
7 |
+
"train_steps_per_second": 0.271
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,351 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.08830931037664413,
|
3 |
+
"best_model_checkpoint": "/home1/datawork/mcontini/models/multilabel/huggingface/dinov2-large-2024_01_14-without_data_aug_batch-size32_epochs20_freeze/checkpoint-5480",
|
4 |
+
"epoch": 20.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 5480,
|
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": 0.4696441032798325,
|
14 |
+
"eval_f1_macro": 0.5445473909109761,
|
15 |
+
"eval_f1_micro": 0.7663056558363417,
|
16 |
+
"eval_loss": 0.1264539361000061,
|
17 |
+
"eval_roc_auc": 0.8400878306196953,
|
18 |
+
"eval_runtime": 1036.6447,
|
19 |
+
"eval_samples_per_second": 2.765,
|
20 |
+
"eval_steps_per_second": 0.087,
|
21 |
+
"learning_rate": 0.001,
|
22 |
+
"step": 274
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"epoch": 1.82,
|
26 |
+
"learning_rate": 0.001,
|
27 |
+
"loss": 0.2337,
|
28 |
+
"step": 500
|
29 |
+
},
|
30 |
+
{
|
31 |
+
"epoch": 2.0,
|
32 |
+
"eval_accuracy": 0.5115143056524774,
|
33 |
+
"eval_f1_macro": 0.6503987559811281,
|
34 |
+
"eval_f1_micro": 0.8025544267053701,
|
35 |
+
"eval_loss": 0.11075052618980408,
|
36 |
+
"eval_roc_auc": 0.8693871473003357,
|
37 |
+
"eval_runtime": 1011.3901,
|
38 |
+
"eval_samples_per_second": 2.834,
|
39 |
+
"eval_steps_per_second": 0.089,
|
40 |
+
"learning_rate": 0.001,
|
41 |
+
"step": 548
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 3.0,
|
45 |
+
"eval_accuracy": 0.5177948360083741,
|
46 |
+
"eval_f1_macro": 0.6963127404974011,
|
47 |
+
"eval_f1_micro": 0.8184487821730869,
|
48 |
+
"eval_loss": 0.1064244732260704,
|
49 |
+
"eval_roc_auc": 0.88043822362814,
|
50 |
+
"eval_runtime": 1010.6496,
|
51 |
+
"eval_samples_per_second": 2.836,
|
52 |
+
"eval_steps_per_second": 0.089,
|
53 |
+
"learning_rate": 0.001,
|
54 |
+
"step": 822
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 3.65,
|
58 |
+
"learning_rate": 0.001,
|
59 |
+
"loss": 0.1259,
|
60 |
+
"step": 1000
|
61 |
+
},
|
62 |
+
{
|
63 |
+
"epoch": 4.0,
|
64 |
+
"eval_accuracy": 0.5188415910676901,
|
65 |
+
"eval_f1_macro": 0.6838420416441195,
|
66 |
+
"eval_f1_micro": 0.8164364720041263,
|
67 |
+
"eval_loss": 0.10615076124668121,
|
68 |
+
"eval_roc_auc": 0.8808295127544562,
|
69 |
+
"eval_runtime": 1012.7035,
|
70 |
+
"eval_samples_per_second": 2.83,
|
71 |
+
"eval_steps_per_second": 0.089,
|
72 |
+
"learning_rate": 0.001,
|
73 |
+
"step": 1096
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 5.0,
|
77 |
+
"eval_accuracy": 0.4965108164689463,
|
78 |
+
"eval_f1_macro": 0.6749223162423283,
|
79 |
+
"eval_f1_micro": 0.8156853881794552,
|
80 |
+
"eval_loss": 0.10690969228744507,
|
81 |
+
"eval_roc_auc": 0.8848957723927348,
|
82 |
+
"eval_runtime": 1011.7793,
|
83 |
+
"eval_samples_per_second": 2.833,
|
84 |
+
"eval_steps_per_second": 0.089,
|
85 |
+
"learning_rate": 0.001,
|
86 |
+
"step": 1370
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 5.47,
|
90 |
+
"learning_rate": 0.001,
|
91 |
+
"loss": 0.1181,
|
92 |
+
"step": 1500
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 6.0,
|
96 |
+
"eval_accuracy": 0.5212840195394278,
|
97 |
+
"eval_f1_macro": 0.6806419028627124,
|
98 |
+
"eval_f1_micro": 0.82229280970257,
|
99 |
+
"eval_loss": 0.10277453809976578,
|
100 |
+
"eval_roc_auc": 0.8815633385282128,
|
101 |
+
"eval_runtime": 1010.7838,
|
102 |
+
"eval_samples_per_second": 2.835,
|
103 |
+
"eval_steps_per_second": 0.089,
|
104 |
+
"learning_rate": 0.001,
|
105 |
+
"step": 1644
|
106 |
+
},
|
107 |
+
{
|
108 |
+
"epoch": 7.0,
|
109 |
+
"eval_accuracy": 0.5268667131891137,
|
110 |
+
"eval_f1_macro": 0.6718237657927368,
|
111 |
+
"eval_f1_micro": 0.8252749770852429,
|
112 |
+
"eval_loss": 0.09917270392179489,
|
113 |
+
"eval_roc_auc": 0.8857388585296528,
|
114 |
+
"eval_runtime": 1012.914,
|
115 |
+
"eval_samples_per_second": 2.829,
|
116 |
+
"eval_steps_per_second": 0.089,
|
117 |
+
"learning_rate": 0.001,
|
118 |
+
"step": 1918
|
119 |
+
},
|
120 |
+
{
|
121 |
+
"epoch": 7.3,
|
122 |
+
"learning_rate": 0.001,
|
123 |
+
"loss": 0.1146,
|
124 |
+
"step": 2000
|
125 |
+
},
|
126 |
+
{
|
127 |
+
"epoch": 8.0,
|
128 |
+
"eval_accuracy": 0.5216329378925332,
|
129 |
+
"eval_f1_macro": 0.6810819602181296,
|
130 |
+
"eval_f1_micro": 0.8223672809196465,
|
131 |
+
"eval_loss": 0.10044297575950623,
|
132 |
+
"eval_roc_auc": 0.8815306397502212,
|
133 |
+
"eval_runtime": 1019.7188,
|
134 |
+
"eval_samples_per_second": 2.811,
|
135 |
+
"eval_steps_per_second": 0.088,
|
136 |
+
"learning_rate": 0.001,
|
137 |
+
"step": 2192
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 9.0,
|
141 |
+
"eval_accuracy": 0.5230286113049546,
|
142 |
+
"eval_f1_macro": 0.6845263067090662,
|
143 |
+
"eval_f1_micro": 0.8302484858776249,
|
144 |
+
"eval_loss": 0.10188507288694382,
|
145 |
+
"eval_roc_auc": 0.8923158419824271,
|
146 |
+
"eval_runtime": 1018.1313,
|
147 |
+
"eval_samples_per_second": 2.815,
|
148 |
+
"eval_steps_per_second": 0.088,
|
149 |
+
"learning_rate": 0.001,
|
150 |
+
"step": 2466
|
151 |
+
},
|
152 |
+
{
|
153 |
+
"epoch": 9.12,
|
154 |
+
"learning_rate": 0.001,
|
155 |
+
"loss": 0.1123,
|
156 |
+
"step": 2500
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 10.0,
|
160 |
+
"eval_accuracy": 0.5279134682484299,
|
161 |
+
"eval_f1_macro": 0.6775112836458498,
|
162 |
+
"eval_f1_micro": 0.8181073809802316,
|
163 |
+
"eval_loss": 0.10208763182163239,
|
164 |
+
"eval_roc_auc": 0.8754023666126408,
|
165 |
+
"eval_runtime": 1019.3535,
|
166 |
+
"eval_samples_per_second": 2.812,
|
167 |
+
"eval_steps_per_second": 0.088,
|
168 |
+
"learning_rate": 0.001,
|
169 |
+
"step": 2740
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 10.95,
|
173 |
+
"learning_rate": 0.001,
|
174 |
+
"loss": 0.11,
|
175 |
+
"step": 3000
|
176 |
+
},
|
177 |
+
{
|
178 |
+
"epoch": 11.0,
|
179 |
+
"eval_accuracy": 0.5429169574319609,
|
180 |
+
"eval_f1_macro": 0.6897433634697611,
|
181 |
+
"eval_f1_micro": 0.8290004089501665,
|
182 |
+
"eval_loss": 0.09596865624189377,
|
183 |
+
"eval_roc_auc": 0.8815259520118676,
|
184 |
+
"eval_runtime": 1024.4911,
|
185 |
+
"eval_samples_per_second": 2.797,
|
186 |
+
"eval_steps_per_second": 0.088,
|
187 |
+
"learning_rate": 0.001,
|
188 |
+
"step": 3014
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 12.0,
|
192 |
+
"eval_accuracy": 0.537334263782275,
|
193 |
+
"eval_f1_macro": 0.682540187560418,
|
194 |
+
"eval_f1_micro": 0.8315506492020821,
|
195 |
+
"eval_loss": 0.0966910719871521,
|
196 |
+
"eval_roc_auc": 0.8896335469441946,
|
197 |
+
"eval_runtime": 1022.2342,
|
198 |
+
"eval_samples_per_second": 2.804,
|
199 |
+
"eval_steps_per_second": 0.088,
|
200 |
+
"learning_rate": 0.001,
|
201 |
+
"step": 3288
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 12.77,
|
205 |
+
"learning_rate": 0.001,
|
206 |
+
"loss": 0.1098,
|
207 |
+
"step": 3500
|
208 |
+
},
|
209 |
+
{
|
210 |
+
"epoch": 13.0,
|
211 |
+
"eval_accuracy": 0.5327983251919051,
|
212 |
+
"eval_f1_macro": 0.6961160759945566,
|
213 |
+
"eval_f1_micro": 0.8254151291512914,
|
214 |
+
"eval_loss": 0.10087021440267563,
|
215 |
+
"eval_roc_auc": 0.8837691574163886,
|
216 |
+
"eval_runtime": 1018.0187,
|
217 |
+
"eval_samples_per_second": 2.815,
|
218 |
+
"eval_steps_per_second": 0.088,
|
219 |
+
"learning_rate": 0.001,
|
220 |
+
"step": 3562
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 14.0,
|
224 |
+
"eval_accuracy": 0.5331472435450104,
|
225 |
+
"eval_f1_macro": 0.7092151646695873,
|
226 |
+
"eval_f1_micro": 0.8278460468351495,
|
227 |
+
"eval_loss": 0.09922181814908981,
|
228 |
+
"eval_roc_auc": 0.8833891332859308,
|
229 |
+
"eval_runtime": 691.5955,
|
230 |
+
"eval_samples_per_second": 4.144,
|
231 |
+
"eval_steps_per_second": 0.13,
|
232 |
+
"learning_rate": 0.001,
|
233 |
+
"step": 3836
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 14.6,
|
237 |
+
"learning_rate": 0.001,
|
238 |
+
"loss": 0.1166,
|
239 |
+
"step": 4000
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 15.0,
|
243 |
+
"eval_accuracy": 0.5334961618981159,
|
244 |
+
"eval_f1_macro": 0.6923192725695103,
|
245 |
+
"eval_f1_micro": 0.8229681773227064,
|
246 |
+
"eval_loss": 0.09913520514965057,
|
247 |
+
"eval_roc_auc": 0.8796342115933741,
|
248 |
+
"eval_runtime": 682.6982,
|
249 |
+
"eval_samples_per_second": 4.198,
|
250 |
+
"eval_steps_per_second": 0.132,
|
251 |
+
"learning_rate": 0.001,
|
252 |
+
"step": 4110
|
253 |
+
},
|
254 |
+
{
|
255 |
+
"epoch": 16.0,
|
256 |
+
"eval_accuracy": 0.5390788555478018,
|
257 |
+
"eval_f1_macro": 0.7123997578521556,
|
258 |
+
"eval_f1_micro": 0.8394566401100476,
|
259 |
+
"eval_loss": 0.09176727384328842,
|
260 |
+
"eval_roc_auc": 0.8932457594574617,
|
261 |
+
"eval_runtime": 684.6393,
|
262 |
+
"eval_samples_per_second": 4.186,
|
263 |
+
"eval_steps_per_second": 0.131,
|
264 |
+
"learning_rate": 0.0001,
|
265 |
+
"step": 4384
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 16.42,
|
269 |
+
"learning_rate": 0.0001,
|
270 |
+
"loss": 0.1072,
|
271 |
+
"step": 4500
|
272 |
+
},
|
273 |
+
{
|
274 |
+
"epoch": 17.0,
|
275 |
+
"eval_accuracy": 0.5471039776692254,
|
276 |
+
"eval_f1_macro": 0.7179397631117517,
|
277 |
+
"eval_f1_micro": 0.8438971798218237,
|
278 |
+
"eval_loss": 0.09066394716501236,
|
279 |
+
"eval_roc_auc": 0.898963054967641,
|
280 |
+
"eval_runtime": 690.5455,
|
281 |
+
"eval_samples_per_second": 4.15,
|
282 |
+
"eval_steps_per_second": 0.13,
|
283 |
+
"learning_rate": 0.0001,
|
284 |
+
"step": 4658
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 18.0,
|
288 |
+
"eval_accuracy": 0.5446615491974878,
|
289 |
+
"eval_f1_macro": 0.7236976914894626,
|
290 |
+
"eval_f1_micro": 0.8441121068554142,
|
291 |
+
"eval_loss": 0.08911634236574173,
|
292 |
+
"eval_roc_auc": 0.8971207593709188,
|
293 |
+
"eval_runtime": 685.8059,
|
294 |
+
"eval_samples_per_second": 4.179,
|
295 |
+
"eval_steps_per_second": 0.131,
|
296 |
+
"learning_rate": 0.0001,
|
297 |
+
"step": 4932
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 18.25,
|
301 |
+
"learning_rate": 0.0001,
|
302 |
+
"loss": 0.1017,
|
303 |
+
"step": 5000
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 19.0,
|
307 |
+
"eval_accuracy": 0.5478018143754362,
|
308 |
+
"eval_f1_macro": 0.727715453888227,
|
309 |
+
"eval_f1_micro": 0.8479439674649797,
|
310 |
+
"eval_loss": 0.08924739807844162,
|
311 |
+
"eval_roc_auc": 0.9027101012597639,
|
312 |
+
"eval_runtime": 697.8818,
|
313 |
+
"eval_samples_per_second": 4.107,
|
314 |
+
"eval_steps_per_second": 0.129,
|
315 |
+
"learning_rate": 0.0001,
|
316 |
+
"step": 5206
|
317 |
+
},
|
318 |
+
{
|
319 |
+
"epoch": 20.0,
|
320 |
+
"eval_accuracy": 0.5509420795533845,
|
321 |
+
"eval_f1_macro": 0.7309632047081221,
|
322 |
+
"eval_f1_micro": 0.848399617483265,
|
323 |
+
"eval_loss": 0.08830931037664413,
|
324 |
+
"eval_roc_auc": 0.9043454872974694,
|
325 |
+
"eval_runtime": 680.9035,
|
326 |
+
"eval_samples_per_second": 4.209,
|
327 |
+
"eval_steps_per_second": 0.132,
|
328 |
+
"learning_rate": 0.0001,
|
329 |
+
"step": 5480
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 20.0,
|
333 |
+
"learning_rate": 0.0001,
|
334 |
+
"step": 5480,
|
335 |
+
"total_flos": 2.5999409342512156e+20,
|
336 |
+
"train_loss": 0.03709733433967089,
|
337 |
+
"train_runtime": 20246.0375,
|
338 |
+
"train_samples_per_second": 8.66,
|
339 |
+
"train_steps_per_second": 0.271
|
340 |
+
}
|
341 |
+
],
|
342 |
+
"logging_steps": 500,
|
343 |
+
"max_steps": 5480,
|
344 |
+
"num_input_tokens_seen": 0,
|
345 |
+
"num_train_epochs": 20,
|
346 |
+
"save_steps": 500,
|
347 |
+
"total_flos": 2.5999409342512156e+20,
|
348 |
+
"train_batch_size": 32,
|
349 |
+
"trial_name": null,
|
350 |
+
"trial_params": null
|
351 |
+
}
|