HF中国镜像站's logo

Tianjiao-Yu
/
videomae-huge

Video Classification
Transformers
TensorBoard
Safetensors
videomae
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Tianjiao-Yu commited on
Commit
3bf783b
·
verified ·
1 Parent(s): 3d2980b

End of training

Browse files
Files changed (10) hide show
  1. README.md +100 -0
  2. all_results.json +8 -0
  3. config.json +63 -0
  4. model.safetensors +3 -0
  5. preprocessor_config.json +26 -0
  6. runs/Feb20_20-21-03_plan.cs.vt.edu/events.out.tfevents.1708489271.plan.cs.vt.edu.4158866.0 +3 -0
  7. runs/Feb20_20-21-03_plan.cs.vt.edu/events.out.tfevents.1708491367.plan.cs.vt.edu.4158866.1 +3 -0
  8. test_results.json +8 -0
  9. trainer_state.json +738 -0
  10. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ base_model: MCG-NJU/videomae-huge-finetuned-kinetics
4
+ tags:
5
+ - generated_from_trainer
6
+ metrics:
7
+ - accuracy
8
+ model-index:
9
+ - name: videomae-huge
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # videomae-huge
17
+
18
+ This model is a fine-tuned version of [MCG-NJU/videomae-huge-finetuned-kinetics](https://huggingface.co/MCG-NJU/videomae-huge-finetuned-kinetics) on an unknown dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 1.5347
21
+ - Accuracy: 0.4286
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 5e-05
41
+ - train_batch_size: 16
42
+ - eval_batch_size: 16
43
+ - seed: 42
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_ratio: 0.1
47
+ - training_steps: 275
48
+
49
+ ### Training results
50
+
51
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
52
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
53
+ | 2.6465 | 0.03 | 7 | 2.6473 | 0.0714 |
54
+ | 2.6262 | 1.03 | 14 | 2.6245 | 0.1429 |
55
+ | 2.5179 | 2.03 | 21 | 2.5627 | 0.1786 |
56
+ | 2.4135 | 3.03 | 28 | 2.4946 | 0.1429 |
57
+ | 2.1202 | 4.03 | 35 | 2.4093 | 0.3214 |
58
+ | 1.848 | 5.03 | 42 | 2.3353 | 0.2857 |
59
+ | 1.726 | 6.03 | 49 | 2.2536 | 0.3214 |
60
+ | 1.3251 | 7.03 | 56 | 2.1674 | 0.3571 |
61
+ | 1.174 | 8.03 | 63 | 2.1128 | 0.3571 |
62
+ | 0.9108 | 9.03 | 70 | 2.0246 | 0.4286 |
63
+ | 0.7467 | 10.03 | 77 | 1.9720 | 0.3571 |
64
+ | 0.6224 | 11.03 | 84 | 1.9130 | 0.3929 |
65
+ | 0.4737 | 12.03 | 91 | 1.9136 | 0.3929 |
66
+ | 0.3682 | 13.03 | 98 | 1.8184 | 0.4643 |
67
+ | 0.2657 | 14.03 | 105 | 1.8734 | 0.3571 |
68
+ | 0.2085 | 15.03 | 112 | 1.8544 | 0.4286 |
69
+ | 0.175 | 16.03 | 119 | 1.8411 | 0.3929 |
70
+ | 0.1327 | 17.03 | 126 | 1.7585 | 0.3929 |
71
+ | 0.1001 | 18.03 | 133 | 1.8193 | 0.3929 |
72
+ | 0.0832 | 19.03 | 140 | 1.7595 | 0.3929 |
73
+ | 0.085 | 20.03 | 147 | 1.7836 | 0.4286 |
74
+ | 0.0614 | 21.03 | 154 | 1.7182 | 0.4286 |
75
+ | 0.0496 | 22.03 | 161 | 1.7845 | 0.4286 |
76
+ | 0.0474 | 23.03 | 168 | 1.7712 | 0.4643 |
77
+ | 0.0403 | 24.03 | 175 | 1.7409 | 0.4643 |
78
+ | 0.0395 | 25.03 | 182 | 1.7425 | 0.4643 |
79
+ | 0.0369 | 26.03 | 189 | 1.7532 | 0.4643 |
80
+ | 0.0346 | 27.03 | 196 | 1.7388 | 0.5 |
81
+ | 0.0321 | 28.03 | 203 | 1.7391 | 0.5 |
82
+ | 0.0314 | 29.03 | 210 | 1.7470 | 0.5 |
83
+ | 0.0313 | 30.03 | 217 | 1.7349 | 0.4643 |
84
+ | 0.0307 | 31.03 | 224 | 1.7574 | 0.4643 |
85
+ | 0.0283 | 32.03 | 231 | 1.7857 | 0.4286 |
86
+ | 0.0276 | 33.03 | 238 | 1.7865 | 0.4643 |
87
+ | 0.0257 | 34.03 | 245 | 1.7707 | 0.4643 |
88
+ | 0.0264 | 35.03 | 252 | 1.7683 | 0.4643 |
89
+ | 0.0254 | 36.03 | 259 | 1.7712 | 0.4643 |
90
+ | 0.0257 | 37.03 | 266 | 1.7736 | 0.4643 |
91
+ | 0.0246 | 38.03 | 273 | 1.7748 | 0.4643 |
92
+ | 0.0242 | 39.01 | 275 | 1.7750 | 0.4643 |
93
+
94
+
95
+ ### Framework versions
96
+
97
+ - Transformers 4.37.2
98
+ - Pytorch 2.1.0+cu121
99
+ - Datasets 2.17.1
100
+ - Tokenizers 0.15.2
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.01,
3
+ "eval_accuracy": 0.42857142857142855,
4
+ "eval_loss": 1.5347131490707397,
5
+ "eval_runtime": 4.3115,
6
+ "eval_samples_per_second": 3.247,
7
+ "eval_steps_per_second": 0.232
8
+ }
config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "MCG-NJU/videomae-huge-finetuned-kinetics",
3
+ "architectures": [
4
+ "VideoMAEForVideoClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "decoder_hidden_size": 640,
8
+ "decoder_intermediate_size": 2560,
9
+ "decoder_num_attention_heads": 8,
10
+ "decoder_num_hidden_layers": 12,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.0,
13
+ "hidden_size": 1280,
14
+ "id2label": {
15
+ "0": "climb",
16
+ "1": "crawl",
17
+ "2": "grasp",
18
+ "3": "hiding",
19
+ "4": "jump",
20
+ "5": "pick up",
21
+ "6": "pull",
22
+ "7": "push",
23
+ "8": "put down",
24
+ "9": "roll",
25
+ "10": "running",
26
+ "11": "slide",
27
+ "12": "walking",
28
+ "13": "zibaroon"
29
+ },
30
+ "image_size": 224,
31
+ "initializer_range": 0.02,
32
+ "intermediate_size": 5120,
33
+ "label2id": {
34
+ "climb": 0,
35
+ "crawl": 1,
36
+ "grasp": 2,
37
+ "hiding": 3,
38
+ "jump": 4,
39
+ "pick up": 5,
40
+ "pull": 6,
41
+ "push": 7,
42
+ "put down": 8,
43
+ "roll": 9,
44
+ "running": 10,
45
+ "slide": 11,
46
+ "walking": 12,
47
+ "zibaroon": 13
48
+ },
49
+ "layer_norm_eps": 1e-12,
50
+ "model_type": "videomae",
51
+ "norm_pix_loss": true,
52
+ "num_attention_heads": 16,
53
+ "num_channels": 3,
54
+ "num_frames": 16,
55
+ "num_hidden_layers": 32,
56
+ "patch_size": 16,
57
+ "problem_type": "single_label_classification",
58
+ "qkv_bias": true,
59
+ "torch_dtype": "float32",
60
+ "transformers_version": "4.37.2",
61
+ "tubelet_size": 2,
62
+ "use_mean_pooling": true
63
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae3205b26f2006d84fdf5f7a0f4d4a1aa6d9b0de42bb0c4f46cb1f6e782af747
3
+ size 2526560304
preprocessor_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": true,
7
+ "do_normalize": true,
8
+ "do_rescale": true,
9
+ "do_resize": true,
10
+ "image_mean": [
11
+ 0.485,
12
+ 0.456,
13
+ 0.406
14
+ ],
15
+ "image_processor_type": "VideoMAEImageProcessor",
16
+ "image_std": [
17
+ 0.229,
18
+ 0.224,
19
+ 0.225
20
+ ],
21
+ "resample": 2,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "shortest_edge": 224
25
+ }
26
+ }
runs/Feb20_20-21-03_plan.cs.vt.edu/events.out.tfevents.1708489271.plan.cs.vt.edu.4158866.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03b76a965492c64e263a5e24d51287d8d2b340420ff20a13f85aab0019a3d129
3
+ size 26802
runs/Feb20_20-21-03_plan.cs.vt.edu/events.out.tfevents.1708491367.plan.cs.vt.edu.4158866.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:501216b5b76b5259fcbe97bb0fbcd017c0d8c17ba11277140f094f0749469bf9
3
+ size 734
test_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.01,
3
+ "eval_accuracy": 0.42857142857142855,
4
+ "eval_loss": 1.5347131490707397,
5
+ "eval_runtime": 4.3115,
6
+ "eval_samples_per_second": 3.247,
7
+ "eval_steps_per_second": 0.232
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,738 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.5,
3
+ "best_model_checkpoint": "Tianjiao-Yu/videomae-huge/checkpoint-196",
4
+ "epoch": 39.00727272727273,
5
+ "eval_steps": 500,
6
+ "global_step": 275,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02,
13
+ "learning_rate": 8.92857142857143e-06,
14
+ "loss": 2.6465,
15
+ "step": 5
16
+ },
17
+ {
18
+ "epoch": 0.03,
19
+ "eval_accuracy": 0.07142857142857142,
20
+ "eval_loss": 2.647310972213745,
21
+ "eval_runtime": 8.9709,
22
+ "eval_samples_per_second": 3.121,
23
+ "eval_steps_per_second": 0.223,
24
+ "step": 7
25
+ },
26
+ {
27
+ "epoch": 1.01,
28
+ "learning_rate": 1.785714285714286e-05,
29
+ "loss": 2.6262,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 1.03,
34
+ "eval_accuracy": 0.14285714285714285,
35
+ "eval_loss": 2.624509811401367,
36
+ "eval_runtime": 7.3449,
37
+ "eval_samples_per_second": 3.812,
38
+ "eval_steps_per_second": 0.272,
39
+ "step": 14
40
+ },
41
+ {
42
+ "epoch": 2.0,
43
+ "learning_rate": 2.6785714285714288e-05,
44
+ "loss": 2.5945,
45
+ "step": 15
46
+ },
47
+ {
48
+ "epoch": 2.02,
49
+ "learning_rate": 3.571428571428572e-05,
50
+ "loss": 2.5179,
51
+ "step": 20
52
+ },
53
+ {
54
+ "epoch": 2.03,
55
+ "eval_accuracy": 0.17857142857142858,
56
+ "eval_loss": 2.562652111053467,
57
+ "eval_runtime": 8.1508,
58
+ "eval_samples_per_second": 3.435,
59
+ "eval_steps_per_second": 0.245,
60
+ "step": 21
61
+ },
62
+ {
63
+ "epoch": 3.01,
64
+ "learning_rate": 4.464285714285715e-05,
65
+ "loss": 2.4135,
66
+ "step": 25
67
+ },
68
+ {
69
+ "epoch": 3.03,
70
+ "eval_accuracy": 0.14285714285714285,
71
+ "eval_loss": 2.494605302810669,
72
+ "eval_runtime": 7.2598,
73
+ "eval_samples_per_second": 3.857,
74
+ "eval_steps_per_second": 0.275,
75
+ "step": 28
76
+ },
77
+ {
78
+ "epoch": 4.01,
79
+ "learning_rate": 4.9595141700404864e-05,
80
+ "loss": 2.182,
81
+ "step": 30
82
+ },
83
+ {
84
+ "epoch": 4.03,
85
+ "learning_rate": 4.8582995951417004e-05,
86
+ "loss": 2.1202,
87
+ "step": 35
88
+ },
89
+ {
90
+ "epoch": 4.03,
91
+ "eval_accuracy": 0.32142857142857145,
92
+ "eval_loss": 2.4093017578125,
93
+ "eval_runtime": 7.6554,
94
+ "eval_samples_per_second": 3.658,
95
+ "eval_steps_per_second": 0.261,
96
+ "step": 35
97
+ },
98
+ {
99
+ "epoch": 5.02,
100
+ "learning_rate": 4.757085020242915e-05,
101
+ "loss": 1.848,
102
+ "step": 40
103
+ },
104
+ {
105
+ "epoch": 5.03,
106
+ "eval_accuracy": 0.2857142857142857,
107
+ "eval_loss": 2.335253953933716,
108
+ "eval_runtime": 7.517,
109
+ "eval_samples_per_second": 3.725,
110
+ "eval_steps_per_second": 0.266,
111
+ "step": 42
112
+ },
113
+ {
114
+ "epoch": 6.01,
115
+ "learning_rate": 4.65587044534413e-05,
116
+ "loss": 1.726,
117
+ "step": 45
118
+ },
119
+ {
120
+ "epoch": 6.03,
121
+ "eval_accuracy": 0.32142857142857145,
122
+ "eval_loss": 2.2536072731018066,
123
+ "eval_runtime": 7.0278,
124
+ "eval_samples_per_second": 3.984,
125
+ "eval_steps_per_second": 0.285,
126
+ "step": 49
127
+ },
128
+ {
129
+ "epoch": 7.0,
130
+ "learning_rate": 4.5546558704453443e-05,
131
+ "loss": 1.5657,
132
+ "step": 50
133
+ },
134
+ {
135
+ "epoch": 7.02,
136
+ "learning_rate": 4.453441295546559e-05,
137
+ "loss": 1.3251,
138
+ "step": 55
139
+ },
140
+ {
141
+ "epoch": 7.03,
142
+ "eval_accuracy": 0.35714285714285715,
143
+ "eval_loss": 2.1673505306243896,
144
+ "eval_runtime": 8.0208,
145
+ "eval_samples_per_second": 3.491,
146
+ "eval_steps_per_second": 0.249,
147
+ "step": 56
148
+ },
149
+ {
150
+ "epoch": 8.01,
151
+ "learning_rate": 4.3522267206477737e-05,
152
+ "loss": 1.174,
153
+ "step": 60
154
+ },
155
+ {
156
+ "epoch": 8.03,
157
+ "eval_accuracy": 0.35714285714285715,
158
+ "eval_loss": 2.1128125190734863,
159
+ "eval_runtime": 8.0792,
160
+ "eval_samples_per_second": 3.466,
161
+ "eval_steps_per_second": 0.248,
162
+ "step": 63
163
+ },
164
+ {
165
+ "epoch": 9.01,
166
+ "learning_rate": 4.251012145748988e-05,
167
+ "loss": 0.9546,
168
+ "step": 65
169
+ },
170
+ {
171
+ "epoch": 9.03,
172
+ "learning_rate": 4.149797570850202e-05,
173
+ "loss": 0.9108,
174
+ "step": 70
175
+ },
176
+ {
177
+ "epoch": 9.03,
178
+ "eval_accuracy": 0.42857142857142855,
179
+ "eval_loss": 2.0245614051818848,
180
+ "eval_runtime": 7.8233,
181
+ "eval_samples_per_second": 3.579,
182
+ "eval_steps_per_second": 0.256,
183
+ "step": 70
184
+ },
185
+ {
186
+ "epoch": 10.02,
187
+ "learning_rate": 4.048582995951417e-05,
188
+ "loss": 0.7467,
189
+ "step": 75
190
+ },
191
+ {
192
+ "epoch": 10.03,
193
+ "eval_accuracy": 0.35714285714285715,
194
+ "eval_loss": 1.9719778299331665,
195
+ "eval_runtime": 7.6011,
196
+ "eval_samples_per_second": 3.684,
197
+ "eval_steps_per_second": 0.263,
198
+ "step": 77
199
+ },
200
+ {
201
+ "epoch": 11.01,
202
+ "learning_rate": 3.9473684210526316e-05,
203
+ "loss": 0.6224,
204
+ "step": 80
205
+ },
206
+ {
207
+ "epoch": 11.03,
208
+ "eval_accuracy": 0.39285714285714285,
209
+ "eval_loss": 1.9129974842071533,
210
+ "eval_runtime": 7.7956,
211
+ "eval_samples_per_second": 3.592,
212
+ "eval_steps_per_second": 0.257,
213
+ "step": 84
214
+ },
215
+ {
216
+ "epoch": 12.0,
217
+ "learning_rate": 3.846153846153846e-05,
218
+ "loss": 0.5523,
219
+ "step": 85
220
+ },
221
+ {
222
+ "epoch": 12.02,
223
+ "learning_rate": 3.744939271255061e-05,
224
+ "loss": 0.4737,
225
+ "step": 90
226
+ },
227
+ {
228
+ "epoch": 12.03,
229
+ "eval_accuracy": 0.39285714285714285,
230
+ "eval_loss": 1.9135581254959106,
231
+ "eval_runtime": 8.3212,
232
+ "eval_samples_per_second": 3.365,
233
+ "eval_steps_per_second": 0.24,
234
+ "step": 91
235
+ },
236
+ {
237
+ "epoch": 13.01,
238
+ "learning_rate": 3.6437246963562756e-05,
239
+ "loss": 0.3682,
240
+ "step": 95
241
+ },
242
+ {
243
+ "epoch": 13.03,
244
+ "eval_accuracy": 0.4642857142857143,
245
+ "eval_loss": 1.8183634281158447,
246
+ "eval_runtime": 8.3757,
247
+ "eval_samples_per_second": 3.343,
248
+ "eval_steps_per_second": 0.239,
249
+ "step": 98
250
+ },
251
+ {
252
+ "epoch": 14.01,
253
+ "learning_rate": 3.54251012145749e-05,
254
+ "loss": 0.3151,
255
+ "step": 100
256
+ },
257
+ {
258
+ "epoch": 14.03,
259
+ "learning_rate": 3.441295546558704e-05,
260
+ "loss": 0.2657,
261
+ "step": 105
262
+ },
263
+ {
264
+ "epoch": 14.03,
265
+ "eval_accuracy": 0.35714285714285715,
266
+ "eval_loss": 1.8734323978424072,
267
+ "eval_runtime": 8.0062,
268
+ "eval_samples_per_second": 3.497,
269
+ "eval_steps_per_second": 0.25,
270
+ "step": 105
271
+ },
272
+ {
273
+ "epoch": 15.02,
274
+ "learning_rate": 3.340080971659919e-05,
275
+ "loss": 0.2085,
276
+ "step": 110
277
+ },
278
+ {
279
+ "epoch": 15.03,
280
+ "eval_accuracy": 0.42857142857142855,
281
+ "eval_loss": 1.8544093370437622,
282
+ "eval_runtime": 7.1501,
283
+ "eval_samples_per_second": 3.916,
284
+ "eval_steps_per_second": 0.28,
285
+ "step": 112
286
+ },
287
+ {
288
+ "epoch": 16.01,
289
+ "learning_rate": 3.2388663967611336e-05,
290
+ "loss": 0.175,
291
+ "step": 115
292
+ },
293
+ {
294
+ "epoch": 16.03,
295
+ "eval_accuracy": 0.39285714285714285,
296
+ "eval_loss": 1.8410834074020386,
297
+ "eval_runtime": 7.4699,
298
+ "eval_samples_per_second": 3.748,
299
+ "eval_steps_per_second": 0.268,
300
+ "step": 119
301
+ },
302
+ {
303
+ "epoch": 17.0,
304
+ "learning_rate": 3.137651821862348e-05,
305
+ "loss": 0.1416,
306
+ "step": 120
307
+ },
308
+ {
309
+ "epoch": 17.02,
310
+ "learning_rate": 3.0364372469635626e-05,
311
+ "loss": 0.1327,
312
+ "step": 125
313
+ },
314
+ {
315
+ "epoch": 17.03,
316
+ "eval_accuracy": 0.39285714285714285,
317
+ "eval_loss": 1.7585405111312866,
318
+ "eval_runtime": 7.2596,
319
+ "eval_samples_per_second": 3.857,
320
+ "eval_steps_per_second": 0.275,
321
+ "step": 126
322
+ },
323
+ {
324
+ "epoch": 18.01,
325
+ "learning_rate": 2.9352226720647776e-05,
326
+ "loss": 0.1001,
327
+ "step": 130
328
+ },
329
+ {
330
+ "epoch": 18.03,
331
+ "eval_accuracy": 0.39285714285714285,
332
+ "eval_loss": 1.8193204402923584,
333
+ "eval_runtime": 7.6478,
334
+ "eval_samples_per_second": 3.661,
335
+ "eval_steps_per_second": 0.262,
336
+ "step": 133
337
+ },
338
+ {
339
+ "epoch": 19.01,
340
+ "learning_rate": 2.8340080971659922e-05,
341
+ "loss": 0.094,
342
+ "step": 135
343
+ },
344
+ {
345
+ "epoch": 19.03,
346
+ "learning_rate": 2.732793522267207e-05,
347
+ "loss": 0.0832,
348
+ "step": 140
349
+ },
350
+ {
351
+ "epoch": 19.03,
352
+ "eval_accuracy": 0.39285714285714285,
353
+ "eval_loss": 1.759466528892517,
354
+ "eval_runtime": 7.1794,
355
+ "eval_samples_per_second": 3.9,
356
+ "eval_steps_per_second": 0.279,
357
+ "step": 140
358
+ },
359
+ {
360
+ "epoch": 20.02,
361
+ "learning_rate": 2.6315789473684212e-05,
362
+ "loss": 0.085,
363
+ "step": 145
364
+ },
365
+ {
366
+ "epoch": 20.03,
367
+ "eval_accuracy": 0.42857142857142855,
368
+ "eval_loss": 1.7836247682571411,
369
+ "eval_runtime": 7.7103,
370
+ "eval_samples_per_second": 3.631,
371
+ "eval_steps_per_second": 0.259,
372
+ "step": 147
373
+ },
374
+ {
375
+ "epoch": 21.01,
376
+ "learning_rate": 2.530364372469636e-05,
377
+ "loss": 0.0614,
378
+ "step": 150
379
+ },
380
+ {
381
+ "epoch": 21.03,
382
+ "eval_accuracy": 0.42857142857142855,
383
+ "eval_loss": 1.7181528806686401,
384
+ "eval_runtime": 7.4571,
385
+ "eval_samples_per_second": 3.755,
386
+ "eval_steps_per_second": 0.268,
387
+ "step": 154
388
+ },
389
+ {
390
+ "epoch": 22.0,
391
+ "learning_rate": 2.4291497975708502e-05,
392
+ "loss": 0.0604,
393
+ "step": 155
394
+ },
395
+ {
396
+ "epoch": 22.02,
397
+ "learning_rate": 2.327935222672065e-05,
398
+ "loss": 0.0496,
399
+ "step": 160
400
+ },
401
+ {
402
+ "epoch": 22.03,
403
+ "eval_accuracy": 0.42857142857142855,
404
+ "eval_loss": 1.7844511270523071,
405
+ "eval_runtime": 7.281,
406
+ "eval_samples_per_second": 3.846,
407
+ "eval_steps_per_second": 0.275,
408
+ "step": 161
409
+ },
410
+ {
411
+ "epoch": 23.01,
412
+ "learning_rate": 2.2267206477732795e-05,
413
+ "loss": 0.0474,
414
+ "step": 165
415
+ },
416
+ {
417
+ "epoch": 23.03,
418
+ "eval_accuracy": 0.4642857142857143,
419
+ "eval_loss": 1.771189570426941,
420
+ "eval_runtime": 7.0825,
421
+ "eval_samples_per_second": 3.953,
422
+ "eval_steps_per_second": 0.282,
423
+ "step": 168
424
+ },
425
+ {
426
+ "epoch": 24.01,
427
+ "learning_rate": 2.125506072874494e-05,
428
+ "loss": 0.0423,
429
+ "step": 170
430
+ },
431
+ {
432
+ "epoch": 24.03,
433
+ "learning_rate": 2.0242914979757085e-05,
434
+ "loss": 0.0403,
435
+ "step": 175
436
+ },
437
+ {
438
+ "epoch": 24.03,
439
+ "eval_accuracy": 0.4642857142857143,
440
+ "eval_loss": 1.740900993347168,
441
+ "eval_runtime": 6.8813,
442
+ "eval_samples_per_second": 4.069,
443
+ "eval_steps_per_second": 0.291,
444
+ "step": 175
445
+ },
446
+ {
447
+ "epoch": 25.02,
448
+ "learning_rate": 1.923076923076923e-05,
449
+ "loss": 0.0395,
450
+ "step": 180
451
+ },
452
+ {
453
+ "epoch": 25.03,
454
+ "eval_accuracy": 0.4642857142857143,
455
+ "eval_loss": 1.7425283193588257,
456
+ "eval_runtime": 7.2597,
457
+ "eval_samples_per_second": 3.857,
458
+ "eval_steps_per_second": 0.275,
459
+ "step": 182
460
+ },
461
+ {
462
+ "epoch": 26.01,
463
+ "learning_rate": 1.8218623481781378e-05,
464
+ "loss": 0.0369,
465
+ "step": 185
466
+ },
467
+ {
468
+ "epoch": 26.03,
469
+ "eval_accuracy": 0.4642857142857143,
470
+ "eval_loss": 1.7531570196151733,
471
+ "eval_runtime": 6.8458,
472
+ "eval_samples_per_second": 4.09,
473
+ "eval_steps_per_second": 0.292,
474
+ "step": 189
475
+ },
476
+ {
477
+ "epoch": 27.0,
478
+ "learning_rate": 1.720647773279352e-05,
479
+ "loss": 0.0358,
480
+ "step": 190
481
+ },
482
+ {
483
+ "epoch": 27.02,
484
+ "learning_rate": 1.6194331983805668e-05,
485
+ "loss": 0.0346,
486
+ "step": 195
487
+ },
488
+ {
489
+ "epoch": 27.03,
490
+ "eval_accuracy": 0.5,
491
+ "eval_loss": 1.7388352155685425,
492
+ "eval_runtime": 7.2714,
493
+ "eval_samples_per_second": 3.851,
494
+ "eval_steps_per_second": 0.275,
495
+ "step": 196
496
+ },
497
+ {
498
+ "epoch": 28.01,
499
+ "learning_rate": 1.5182186234817813e-05,
500
+ "loss": 0.0321,
501
+ "step": 200
502
+ },
503
+ {
504
+ "epoch": 28.03,
505
+ "eval_accuracy": 0.5,
506
+ "eval_loss": 1.7390631437301636,
507
+ "eval_runtime": 7.0464,
508
+ "eval_samples_per_second": 3.974,
509
+ "eval_steps_per_second": 0.284,
510
+ "step": 203
511
+ },
512
+ {
513
+ "epoch": 29.01,
514
+ "learning_rate": 1.4170040485829961e-05,
515
+ "loss": 0.0351,
516
+ "step": 205
517
+ },
518
+ {
519
+ "epoch": 29.03,
520
+ "learning_rate": 1.3157894736842106e-05,
521
+ "loss": 0.0314,
522
+ "step": 210
523
+ },
524
+ {
525
+ "epoch": 29.03,
526
+ "eval_accuracy": 0.5,
527
+ "eval_loss": 1.7469881772994995,
528
+ "eval_runtime": 6.6833,
529
+ "eval_samples_per_second": 4.19,
530
+ "eval_steps_per_second": 0.299,
531
+ "step": 210
532
+ },
533
+ {
534
+ "epoch": 30.02,
535
+ "learning_rate": 1.2145748987854251e-05,
536
+ "loss": 0.0313,
537
+ "step": 215
538
+ },
539
+ {
540
+ "epoch": 30.03,
541
+ "eval_accuracy": 0.4642857142857143,
542
+ "eval_loss": 1.7349082231521606,
543
+ "eval_runtime": 6.7759,
544
+ "eval_samples_per_second": 4.132,
545
+ "eval_steps_per_second": 0.295,
546
+ "step": 217
547
+ },
548
+ {
549
+ "epoch": 31.01,
550
+ "learning_rate": 1.1133603238866398e-05,
551
+ "loss": 0.0307,
552
+ "step": 220
553
+ },
554
+ {
555
+ "epoch": 31.03,
556
+ "eval_accuracy": 0.4642857142857143,
557
+ "eval_loss": 1.7574304342269897,
558
+ "eval_runtime": 6.501,
559
+ "eval_samples_per_second": 4.307,
560
+ "eval_steps_per_second": 0.308,
561
+ "step": 224
562
+ },
563
+ {
564
+ "epoch": 32.0,
565
+ "learning_rate": 1.0121457489878542e-05,
566
+ "loss": 0.0296,
567
+ "step": 225
568
+ },
569
+ {
570
+ "epoch": 32.02,
571
+ "learning_rate": 9.109311740890689e-06,
572
+ "loss": 0.0283,
573
+ "step": 230
574
+ },
575
+ {
576
+ "epoch": 32.03,
577
+ "eval_accuracy": 0.42857142857142855,
578
+ "eval_loss": 1.7856690883636475,
579
+ "eval_runtime": 6.5016,
580
+ "eval_samples_per_second": 4.307,
581
+ "eval_steps_per_second": 0.308,
582
+ "step": 231
583
+ },
584
+ {
585
+ "epoch": 33.01,
586
+ "learning_rate": 8.097165991902834e-06,
587
+ "loss": 0.0276,
588
+ "step": 235
589
+ },
590
+ {
591
+ "epoch": 33.03,
592
+ "eval_accuracy": 0.4642857142857143,
593
+ "eval_loss": 1.786450982093811,
594
+ "eval_runtime": 7.4349,
595
+ "eval_samples_per_second": 3.766,
596
+ "eval_steps_per_second": 0.269,
597
+ "step": 238
598
+ },
599
+ {
600
+ "epoch": 34.01,
601
+ "learning_rate": 7.0850202429149805e-06,
602
+ "loss": 0.0282,
603
+ "step": 240
604
+ },
605
+ {
606
+ "epoch": 34.03,
607
+ "learning_rate": 6.0728744939271254e-06,
608
+ "loss": 0.0257,
609
+ "step": 245
610
+ },
611
+ {
612
+ "epoch": 34.03,
613
+ "eval_accuracy": 0.4642857142857143,
614
+ "eval_loss": 1.7707021236419678,
615
+ "eval_runtime": 7.9939,
616
+ "eval_samples_per_second": 3.503,
617
+ "eval_steps_per_second": 0.25,
618
+ "step": 245
619
+ },
620
+ {
621
+ "epoch": 35.02,
622
+ "learning_rate": 5.060728744939271e-06,
623
+ "loss": 0.0264,
624
+ "step": 250
625
+ },
626
+ {
627
+ "epoch": 35.03,
628
+ "eval_accuracy": 0.4642857142857143,
629
+ "eval_loss": 1.7682857513427734,
630
+ "eval_runtime": 7.2901,
631
+ "eval_samples_per_second": 3.841,
632
+ "eval_steps_per_second": 0.274,
633
+ "step": 252
634
+ },
635
+ {
636
+ "epoch": 36.01,
637
+ "learning_rate": 4.048582995951417e-06,
638
+ "loss": 0.0254,
639
+ "step": 255
640
+ },
641
+ {
642
+ "epoch": 36.03,
643
+ "eval_accuracy": 0.4642857142857143,
644
+ "eval_loss": 1.7711747884750366,
645
+ "eval_runtime": 7.529,
646
+ "eval_samples_per_second": 3.719,
647
+ "eval_steps_per_second": 0.266,
648
+ "step": 259
649
+ },
650
+ {
651
+ "epoch": 37.0,
652
+ "learning_rate": 3.0364372469635627e-06,
653
+ "loss": 0.0247,
654
+ "step": 260
655
+ },
656
+ {
657
+ "epoch": 37.02,
658
+ "learning_rate": 2.0242914979757085e-06,
659
+ "loss": 0.0257,
660
+ "step": 265
661
+ },
662
+ {
663
+ "epoch": 37.03,
664
+ "eval_accuracy": 0.4642857142857143,
665
+ "eval_loss": 1.7735536098480225,
666
+ "eval_runtime": 7.5429,
667
+ "eval_samples_per_second": 3.712,
668
+ "eval_steps_per_second": 0.265,
669
+ "step": 266
670
+ },
671
+ {
672
+ "epoch": 38.01,
673
+ "learning_rate": 1.0121457489878542e-06,
674
+ "loss": 0.0246,
675
+ "step": 270
676
+ },
677
+ {
678
+ "epoch": 38.03,
679
+ "eval_accuracy": 0.4642857142857143,
680
+ "eval_loss": 1.7748419046401978,
681
+ "eval_runtime": 7.4873,
682
+ "eval_samples_per_second": 3.74,
683
+ "eval_steps_per_second": 0.267,
684
+ "step": 273
685
+ },
686
+ {
687
+ "epoch": 39.01,
688
+ "learning_rate": 0.0,
689
+ "loss": 0.0242,
690
+ "step": 275
691
+ },
692
+ {
693
+ "epoch": 39.01,
694
+ "eval_accuracy": 0.4642857142857143,
695
+ "eval_loss": 1.774975061416626,
696
+ "eval_runtime": 7.7693,
697
+ "eval_samples_per_second": 3.604,
698
+ "eval_steps_per_second": 0.257,
699
+ "step": 275
700
+ },
701
+ {
702
+ "epoch": 39.01,
703
+ "step": 275,
704
+ "total_flos": 3.980465870018229e+19,
705
+ "train_loss": 0.5794233712283048,
706
+ "train_runtime": 2037.2017,
707
+ "train_samples_per_second": 2.16,
708
+ "train_steps_per_second": 0.135
709
+ },
710
+ {
711
+ "epoch": 39.01,
712
+ "eval_accuracy": 0.42857142857142855,
713
+ "eval_loss": 1.5347131490707397,
714
+ "eval_runtime": 5.4942,
715
+ "eval_samples_per_second": 2.548,
716
+ "eval_steps_per_second": 0.182,
717
+ "step": 275
718
+ },
719
+ {
720
+ "epoch": 39.01,
721
+ "eval_accuracy": 0.42857142857142855,
722
+ "eval_loss": 1.5347131490707397,
723
+ "eval_runtime": 4.3115,
724
+ "eval_samples_per_second": 3.247,
725
+ "eval_steps_per_second": 0.232,
726
+ "step": 275
727
+ }
728
+ ],
729
+ "logging_steps": 5,
730
+ "max_steps": 275,
731
+ "num_input_tokens_seen": 0,
732
+ "num_train_epochs": 9223372036854775807,
733
+ "save_steps": 500,
734
+ "total_flos": 3.980465870018229e+19,
735
+ "train_batch_size": 16,
736
+ "trial_name": null,
737
+ "trial_params": null
738
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5da55cdf3ae599f60b6c1ca815936709b3bb7791b63b743fed24cd1f23ce505
3
+ size 4728