End of training
Browse files- README.md +2 -2
- all_results.json +8 -0
- runs/Jul02_19-45-22_c6d9a0630dcb/events.out.tfevents.1719950183.c6d9a0630dcb.4133.5 +2 -2
- test_results.json +8 -0
- trainer_state.json +189 -0
README.md
CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
17 |
|
18 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 0.
|
21 |
-
- Accuracy: 0.
|
22 |
|
23 |
## Model description
|
24 |
|
|
|
17 |
|
18 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 0.6847
|
21 |
+
- Accuracy: 0.75
|
22 |
|
23 |
## Model description
|
24 |
|
all_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 4.2,
|
3 |
+
"eval_accuracy": 0.75,
|
4 |
+
"eval_loss": 0.6846519112586975,
|
5 |
+
"eval_runtime": 14.0555,
|
6 |
+
"eval_samples_per_second": 0.854,
|
7 |
+
"eval_steps_per_second": 0.427
|
8 |
+
}
|
runs/Jul02_19-45-22_c6d9a0630dcb/events.out.tfevents.1719950183.c6d9a0630dcb.4133.5
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:ab438901554fb6f68f8f2b184c3b5e8a747d69a21c4031db162e88e4ddd688d3
|
3 |
+
size 722
|
test_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 4.2,
|
3 |
+
"eval_accuracy": 0.75,
|
4 |
+
"eval_loss": 0.6846519112586975,
|
5 |
+
"eval_runtime": 14.0555,
|
6 |
+
"eval_samples_per_second": 0.854,
|
7 |
+
"eval_steps_per_second": 0.427
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,189 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 1.0,
|
3 |
+
"best_model_checkpoint": "videomae-base-finetuned-Risky-situations/checkpoint-25",
|
4 |
+
"epoch": 4.2,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 125,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.08,
|
13 |
+
"grad_norm": 8.516080856323242,
|
14 |
+
"learning_rate": 3.846153846153846e-05,
|
15 |
+
"loss": 0.5136,
|
16 |
+
"step": 10
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.16,
|
20 |
+
"grad_norm": 40.52229690551758,
|
21 |
+
"learning_rate": 4.6875e-05,
|
22 |
+
"loss": 0.4618,
|
23 |
+
"step": 20
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.2,
|
27 |
+
"eval_accuracy": 1.0,
|
28 |
+
"eval_loss": 0.010277315974235535,
|
29 |
+
"eval_runtime": 11.9488,
|
30 |
+
"eval_samples_per_second": 1.004,
|
31 |
+
"eval_steps_per_second": 0.502,
|
32 |
+
"step": 25
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.04,
|
36 |
+
"grad_norm": 110.08763122558594,
|
37 |
+
"learning_rate": 4.2410714285714285e-05,
|
38 |
+
"loss": 0.0748,
|
39 |
+
"step": 30
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 1.12,
|
43 |
+
"grad_norm": 14.86829662322998,
|
44 |
+
"learning_rate": 3.794642857142857e-05,
|
45 |
+
"loss": 0.6214,
|
46 |
+
"step": 40
|
47 |
+
},
|
48 |
+
{
|
49 |
+
"epoch": 1.2,
|
50 |
+
"grad_norm": 0.27806639671325684,
|
51 |
+
"learning_rate": 3.348214285714286e-05,
|
52 |
+
"loss": 0.3548,
|
53 |
+
"step": 50
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 1.2,
|
57 |
+
"eval_accuracy": 1.0,
|
58 |
+
"eval_loss": 0.002936301752924919,
|
59 |
+
"eval_runtime": 11.6674,
|
60 |
+
"eval_samples_per_second": 1.029,
|
61 |
+
"eval_steps_per_second": 0.514,
|
62 |
+
"step": 50
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 2.08,
|
66 |
+
"grad_norm": 0.032773926854133606,
|
67 |
+
"learning_rate": 2.9017857142857146e-05,
|
68 |
+
"loss": 0.2253,
|
69 |
+
"step": 60
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 2.16,
|
73 |
+
"grad_norm": 0.036098264157772064,
|
74 |
+
"learning_rate": 2.455357142857143e-05,
|
75 |
+
"loss": 0.2182,
|
76 |
+
"step": 70
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 2.2,
|
80 |
+
"eval_accuracy": 1.0,
|
81 |
+
"eval_loss": 0.0008991466020233929,
|
82 |
+
"eval_runtime": 11.8302,
|
83 |
+
"eval_samples_per_second": 1.014,
|
84 |
+
"eval_steps_per_second": 0.507,
|
85 |
+
"step": 75
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 3.04,
|
89 |
+
"grad_norm": 0.034782785922288895,
|
90 |
+
"learning_rate": 2.0089285714285717e-05,
|
91 |
+
"loss": 0.0011,
|
92 |
+
"step": 80
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 3.12,
|
96 |
+
"grad_norm": 0.4517887532711029,
|
97 |
+
"learning_rate": 1.5625e-05,
|
98 |
+
"loss": 0.3581,
|
99 |
+
"step": 90
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 3.2,
|
103 |
+
"grad_norm": 3.7396240234375,
|
104 |
+
"learning_rate": 1.1160714285714287e-05,
|
105 |
+
"loss": 0.2513,
|
106 |
+
"step": 100
|
107 |
+
},
|
108 |
+
{
|
109 |
+
"epoch": 3.2,
|
110 |
+
"eval_accuracy": 1.0,
|
111 |
+
"eval_loss": 0.003270552260801196,
|
112 |
+
"eval_runtime": 11.5723,
|
113 |
+
"eval_samples_per_second": 1.037,
|
114 |
+
"eval_steps_per_second": 0.518,
|
115 |
+
"step": 100
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 4.08,
|
119 |
+
"grad_norm": 357.18890380859375,
|
120 |
+
"learning_rate": 6.696428571428572e-06,
|
121 |
+
"loss": 0.2286,
|
122 |
+
"step": 110
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 4.16,
|
126 |
+
"grad_norm": 0.019339008256793022,
|
127 |
+
"learning_rate": 2.2321428571428573e-06,
|
128 |
+
"loss": 0.0021,
|
129 |
+
"step": 120
|
130 |
+
},
|
131 |
+
{
|
132 |
+
"epoch": 4.2,
|
133 |
+
"eval_accuracy": 1.0,
|
134 |
+
"eval_loss": 0.0007905661477707326,
|
135 |
+
"eval_runtime": 14.2671,
|
136 |
+
"eval_samples_per_second": 0.841,
|
137 |
+
"eval_steps_per_second": 0.421,
|
138 |
+
"step": 125
|
139 |
+
},
|
140 |
+
{
|
141 |
+
"epoch": 4.2,
|
142 |
+
"step": 125,
|
143 |
+
"total_flos": 3.11516147367936e+17,
|
144 |
+
"train_loss": 0.26490905868075787,
|
145 |
+
"train_runtime": 483.7689,
|
146 |
+
"train_samples_per_second": 0.517,
|
147 |
+
"train_steps_per_second": 0.258
|
148 |
+
},
|
149 |
+
{
|
150 |
+
"epoch": 4.2,
|
151 |
+
"eval_accuracy": 0.7692307692307693,
|
152 |
+
"eval_loss": 0.6328043341636658,
|
153 |
+
"eval_runtime": 16.8871,
|
154 |
+
"eval_samples_per_second": 0.77,
|
155 |
+
"eval_steps_per_second": 0.415,
|
156 |
+
"step": 125
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 4.2,
|
160 |
+
"eval_accuracy": 0.75,
|
161 |
+
"eval_loss": 0.6846519112586975,
|
162 |
+
"eval_runtime": 14.0555,
|
163 |
+
"eval_samples_per_second": 0.854,
|
164 |
+
"eval_steps_per_second": 0.427,
|
165 |
+
"step": 125
|
166 |
+
}
|
167 |
+
],
|
168 |
+
"logging_steps": 10,
|
169 |
+
"max_steps": 125,
|
170 |
+
"num_input_tokens_seen": 0,
|
171 |
+
"num_train_epochs": 9223372036854775807,
|
172 |
+
"save_steps": 500,
|
173 |
+
"stateful_callbacks": {
|
174 |
+
"TrainerControl": {
|
175 |
+
"args": {
|
176 |
+
"should_epoch_stop": false,
|
177 |
+
"should_evaluate": false,
|
178 |
+
"should_log": false,
|
179 |
+
"should_save": true,
|
180 |
+
"should_training_stop": true
|
181 |
+
},
|
182 |
+
"attributes": {}
|
183 |
+
}
|
184 |
+
},
|
185 |
+
"total_flos": 3.11516147367936e+17,
|
186 |
+
"train_batch_size": 2,
|
187 |
+
"trial_name": null,
|
188 |
+
"trial_params": null
|
189 |
+
}
|