Training in progress, epoch 1
Browse files- config.json +1 -0
- logs/events.out.tfevents.1740281118.c59d799b8436.377.0 +2 -2
- model.safetensors +1 -1
- trainer_state.json +7 -152
config.json
CHANGED
@@ -57,6 +57,7 @@
|
|
57 |
],
|
58 |
"patch_size": 4,
|
59 |
"path_norm": true,
|
|
|
60 |
"qkv_bias": true,
|
61 |
"stage_names": [
|
62 |
"stem",
|
|
|
57 |
],
|
58 |
"patch_size": 4,
|
59 |
"path_norm": true,
|
60 |
+
"problem_type": "single_label_classification",
|
61 |
"qkv_bias": true,
|
62 |
"stage_names": [
|
63 |
"stem",
|
logs/events.out.tfevents.1740281118.c59d799b8436.377.0
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:896215c191341dd1a8205cbc6455a3ed9524f79b681f3931847aeb26f08c2587
|
3 |
+
size 6589
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 347523416
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c66f92bdf1001f104cec70620455422d1479bde4f8f6728a50cf0765622332fd
|
3 |
size 347523416
|
trainer_state.json
CHANGED
@@ -1,158 +1,13 @@
|
|
1 |
{
|
2 |
-
"best_metric":
|
3 |
-
"best_model_checkpoint":
|
4 |
-
"epoch":
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
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 |
-
"grad_norm": 7.30280876159668,
|
14 |
-
"learning_rate": 6.577540106951872e-06,
|
15 |
-
"loss": 1.6649,
|
16 |
-
"step": 125
|
17 |
-
},
|
18 |
-
{
|
19 |
-
"epoch": 1.0,
|
20 |
-
"eval_accuracy": 0.7351190476190477,
|
21 |
-
"eval_loss": 0.8050427436828613,
|
22 |
-
"eval_runtime": 14.2308,
|
23 |
-
"eval_samples_per_second": 23.611,
|
24 |
-
"eval_steps_per_second": 0.422,
|
25 |
-
"step": 125
|
26 |
-
},
|
27 |
-
{
|
28 |
-
"epoch": 2.0,
|
29 |
-
"grad_norm": 6.75390100479126,
|
30 |
-
"learning_rate": 9.871664184437119e-06,
|
31 |
-
"loss": 0.431,
|
32 |
-
"step": 250
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 2.0,
|
36 |
-
"eval_accuracy": 0.8511904761904762,
|
37 |
-
"eval_loss": 0.45273008942604065,
|
38 |
-
"eval_runtime": 13.8146,
|
39 |
-
"eval_samples_per_second": 24.322,
|
40 |
-
"eval_steps_per_second": 0.434,
|
41 |
-
"step": 250
|
42 |
-
},
|
43 |
-
{
|
44 |
-
"epoch": 3.0,
|
45 |
-
"grad_norm": 8.129321098327637,
|
46 |
-
"learning_rate": 8.848767222994402e-06,
|
47 |
-
"loss": 0.1233,
|
48 |
-
"step": 375
|
49 |
-
},
|
50 |
-
{
|
51 |
-
"epoch": 3.0,
|
52 |
-
"eval_accuracy": 0.8928571428571429,
|
53 |
-
"eval_loss": 0.3934735655784607,
|
54 |
-
"eval_runtime": 14.0064,
|
55 |
-
"eval_samples_per_second": 23.989,
|
56 |
-
"eval_steps_per_second": 0.428,
|
57 |
-
"step": 375
|
58 |
-
},
|
59 |
-
{
|
60 |
-
"epoch": 4.0,
|
61 |
-
"grad_norm": 1.9338823556900024,
|
62 |
-
"learning_rate": 7.007580044706666e-06,
|
63 |
-
"loss": 0.0326,
|
64 |
-
"step": 500
|
65 |
-
},
|
66 |
-
{
|
67 |
-
"epoch": 4.0,
|
68 |
-
"eval_accuracy": 0.9077380952380952,
|
69 |
-
"eval_loss": 0.4275733530521393,
|
70 |
-
"eval_runtime": 13.7209,
|
71 |
-
"eval_samples_per_second": 24.488,
|
72 |
-
"eval_steps_per_second": 0.437,
|
73 |
-
"step": 500
|
74 |
-
},
|
75 |
-
{
|
76 |
-
"epoch": 5.0,
|
77 |
-
"grad_norm": 0.1951848417520523,
|
78 |
-
"learning_rate": 4.739559281809819e-06,
|
79 |
-
"loss": 0.01,
|
80 |
-
"step": 625
|
81 |
-
},
|
82 |
-
{
|
83 |
-
"epoch": 5.0,
|
84 |
-
"eval_accuracy": 0.9166666666666666,
|
85 |
-
"eval_loss": 0.45519202947616577,
|
86 |
-
"eval_runtime": 13.6317,
|
87 |
-
"eval_samples_per_second": 24.648,
|
88 |
-
"eval_steps_per_second": 0.44,
|
89 |
-
"step": 625
|
90 |
-
},
|
91 |
-
{
|
92 |
-
"epoch": 6.0,
|
93 |
-
"grad_norm": 0.07782671600580215,
|
94 |
-
"learning_rate": 2.5674558357271096e-06,
|
95 |
-
"loss": 0.0049,
|
96 |
-
"step": 750
|
97 |
-
},
|
98 |
-
{
|
99 |
-
"epoch": 6.0,
|
100 |
-
"eval_accuracy": 0.9077380952380952,
|
101 |
-
"eval_loss": 0.4804078936576843,
|
102 |
-
"eval_runtime": 12.6134,
|
103 |
-
"eval_samples_per_second": 26.638,
|
104 |
-
"eval_steps_per_second": 0.476,
|
105 |
-
"step": 750
|
106 |
-
},
|
107 |
-
{
|
108 |
-
"epoch": 7.0,
|
109 |
-
"grad_norm": 0.7393015623092651,
|
110 |
-
"learning_rate": 9.974770025621516e-06,
|
111 |
-
"loss": 0.0024,
|
112 |
-
"step": 875
|
113 |
-
},
|
114 |
-
{
|
115 |
-
"epoch": 7.0,
|
116 |
-
"eval_accuracy": 0.9166666666666666,
|
117 |
-
"eval_loss": 0.4863254427909851,
|
118 |
-
"eval_runtime": 12.7829,
|
119 |
-
"eval_samples_per_second": 26.285,
|
120 |
-
"eval_steps_per_second": 0.469,
|
121 |
-
"step": 875
|
122 |
-
},
|
123 |
-
{
|
124 |
-
"epoch": 8.0,
|
125 |
-
"grad_norm": 0.14619985222816467,
|
126 |
-
"learning_rate": 6.511415799458836e-06,
|
127 |
-
"loss": 0.005,
|
128 |
-
"step": 1000
|
129 |
-
},
|
130 |
-
{
|
131 |
-
"epoch": 8.0,
|
132 |
-
"eval_accuracy": 0.9166666666666666,
|
133 |
-
"eval_loss": 0.5142239332199097,
|
134 |
-
"eval_runtime": 12.7007,
|
135 |
-
"eval_samples_per_second": 26.455,
|
136 |
-
"eval_steps_per_second": 0.472,
|
137 |
-
"step": 1000
|
138 |
-
},
|
139 |
-
{
|
140 |
-
"epoch": 9.0,
|
141 |
-
"grad_norm": 0.16888085007667542,
|
142 |
-
"learning_rate": 1.2234689624958324e-06,
|
143 |
-
"loss": 0.0026,
|
144 |
-
"step": 1125
|
145 |
-
},
|
146 |
-
{
|
147 |
-
"epoch": 9.0,
|
148 |
-
"eval_accuracy": 0.9166666666666666,
|
149 |
-
"eval_loss": 0.5233145952224731,
|
150 |
-
"eval_runtime": 12.5815,
|
151 |
-
"eval_samples_per_second": 26.706,
|
152 |
-
"eval_steps_per_second": 0.477,
|
153 |
-
"step": 1125
|
154 |
-
}
|
155 |
-
],
|
156 |
"logging_steps": 500,
|
157 |
"max_steps": 1875,
|
158 |
"num_input_tokens_seen": 0,
|
@@ -164,13 +19,13 @@
|
|
164 |
"should_epoch_stop": false,
|
165 |
"should_evaluate": false,
|
166 |
"should_log": false,
|
167 |
-
"should_save":
|
168 |
"should_training_stop": false
|
169 |
},
|
170 |
"attributes": {}
|
171 |
}
|
172 |
},
|
173 |
-
"total_flos":
|
174 |
"train_batch_size": 64,
|
175 |
"trial_name": null,
|
176 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 0,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
10 |
+
"log_history": [],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
"logging_steps": 500,
|
12 |
"max_steps": 1875,
|
13 |
"num_input_tokens_seen": 0,
|
|
|
19 |
"should_epoch_stop": false,
|
20 |
"should_evaluate": false,
|
21 |
"should_log": false,
|
22 |
+
"should_save": false,
|
23 |
"should_training_stop": false
|
24 |
},
|
25 |
"attributes": {}
|
26 |
}
|
27 |
},
|
28 |
+
"total_flos": 0,
|
29 |
"train_batch_size": 64,
|
30 |
"trial_name": null,
|
31 |
"trial_params": null
|