Hugging Face's logo

darthraider
/
vit-base-fruit-punch

Image Classification
Transformers
TensorBoard
Safetensors
vit
Generated from Trainer
Model card Files
xet
 Metrics Community
darthraider commited on
Commit
38bd784
·
verified ·
1 Parent(s): ed80e67

Training in progress, step 100

Browse files
Files changed (14) hide show
  1. README.md +72 -0
  2. all_results.json +13 -0
  3. config.json +36 -0
  4. eval_results.json +8 -0
  5. model.safetensors +3 -0
  6. preprocessor_config.json +23 -0
  7. runs/Jan24_16-46-06_d9578ace0713/events.out.tfevents.1737737210.d9578ace0713.31.0 +3 -0
  8. runs/Jan24_16-46-06_d9578ace0713/events.out.tfevents.1737737787.d9578ace0713.31.1 +3 -0
  9. runs/Jan24_17-05-18_7865a1b6aba1/events.out.tfevents.1737738320.7865a1b6aba1.31.0 +3 -0
  10. runs/Jan24_17-05-18_7865a1b6aba1/events.out.tfevents.1737739352.7865a1b6aba1.31.1 +3 -0
  11. runs/Mar04_18-25-10_2cb32fa4c9b0/events.out.tfevents.1741112714.2cb32fa4c9b0.31.0 +3 -0
  12. train_results.json +8 -0
  13. trainer_state.json +832 -0
  14. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: apache-2.0
4
+ base_model: google/vit-base-patch16-224-in21k
5
+ tags:
6
+ - image-classification
7
+ - generated_from_trainer
8
+ metrics:
9
+ - accuracy
10
+ model-index:
11
+ - name: vit-base-fruit-punch
12
+ results: []
13
+ ---
14
+
15
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
+ should probably proofread and complete it, then remove this comment. -->
17
+
18
+ # vit-base-fruit-punch
19
+
20
+ This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the darthraider/fruit-ripeness-detection-dataset dataset.
21
+ It achieves the following results on the evaluation set:
22
+ - Loss: 0.0004
23
+ - Accuracy: 1.0
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 5e-05
43
+ - train_batch_size: 32
44
+ - eval_batch_size: 16
45
+ - seed: 42
46
+ - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
47
+ - lr_scheduler_type: linear
48
+ - num_epochs: 8
49
+ - mixed_precision_training: Native AMP
50
+
51
+ ### Training results
52
+
53
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
54
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
55
+ | 0.046 | 0.8 | 100 | 0.0443 | 1.0 |
56
+ | 0.0164 | 1.6 | 200 | 0.0188 | 0.999 |
57
+ | 0.007 | 2.4 | 300 | 0.0071 | 1.0 |
58
+ | 0.0033 | 3.2 | 400 | 0.0035 | 1.0 |
59
+ | 0.0017 | 4.0 | 500 | 0.0021 | 1.0 |
60
+ | 0.001 | 4.8 | 600 | 0.0011 | 1.0 |
61
+ | 0.0006 | 5.6 | 700 | 0.0007 | 1.0 |
62
+ | 0.0004 | 6.4 | 800 | 0.0005 | 1.0 |
63
+ | 0.0004 | 7.2 | 900 | 0.0004 | 1.0 |
64
+ | 0.0003 | 8.0 | 1000 | 0.0004 | 1.0 |
65
+
66
+
67
+ ### Framework versions
68
+
69
+ - Transformers 4.47.0
70
+ - Pytorch 2.5.1+cu121
71
+ - Datasets 3.2.0
72
+ - Tokenizers 0.21.0
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "eval_accuracy": 1.0,
4
+ "eval_loss": 0.00038815097650513053,
5
+ "eval_runtime": 13.0148,
6
+ "eval_samples_per_second": 76.836,
7
+ "eval_steps_per_second": 4.841,
8
+ "total_flos": 2.479168170953736e+18,
9
+ "train_loss": 0.03568338290625252,
10
+ "train_runtime": 1001.3794,
11
+ "train_samples_per_second": 31.948,
12
+ "train_steps_per_second": 0.999
13
+ }
config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Raw_Banana",
13
+ "1": "Raw_Mango",
14
+ "2": "Ripe_Banana",
15
+ "3": "Ripe_Mango"
16
+ },
17
+ "image_size": 224,
18
+ "initializer_range": 0.02,
19
+ "intermediate_size": 3072,
20
+ "label2id": {
21
+ "Raw_Banana": "0",
22
+ "Raw_Mango": "1",
23
+ "Ripe_Banana": "2",
24
+ "Ripe_Mango": "3"
25
+ },
26
+ "layer_norm_eps": 1e-12,
27
+ "model_type": "vit",
28
+ "num_attention_heads": 12,
29
+ "num_channels": 3,
30
+ "num_hidden_layers": 12,
31
+ "patch_size": 16,
32
+ "problem_type": "single_label_classification",
33
+ "qkv_bias": true,
34
+ "torch_dtype": "float32",
35
+ "transformers_version": "4.47.0"
36
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "eval_accuracy": 1.0,
4
+ "eval_loss": 0.00038815097650513053,
5
+ "eval_runtime": 13.0148,
6
+ "eval_samples_per_second": 76.836,
7
+ "eval_steps_per_second": 4.841
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:00d34b9dc6c81929431f59aca85621c9fb4799d49707f1d941215a8a56ee7739
3
+ size 343230128
preprocessor_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "ViTFeatureExtractor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "resample": 2,
18
+ "rescale_factor": 0.00392156862745098,
19
+ "size": {
20
+ "height": 224,
21
+ "width": 224
22
+ }
23
+ }
runs/Jan24_16-46-06_d9578ace0713/events.out.tfevents.1737737210.d9578ace0713.31.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fdbd7cdcc8cc5f489e74d170012c72d7406c763ce85816f9468c697b6a81024a
3
+ size 17624
runs/Jan24_16-46-06_d9578ace0713/events.out.tfevents.1737737787.d9578ace0713.31.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12e8f72e8cc8cace2f86e7ed830f4ca3e12b941d0362fed96e8b0d6943e34be0
3
+ size 411
runs/Jan24_17-05-18_7865a1b6aba1/events.out.tfevents.1737738320.7865a1b6aba1.31.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:548b70b3b94947d57d436cc1f8aa3d23e1e9f47aa4803786123eac8d3896369e
3
+ size 29788
runs/Jan24_17-05-18_7865a1b6aba1/events.out.tfevents.1737739352.7865a1b6aba1.31.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47fe68f9a15da60a6e582528aacd5d9baa0697c2a2c05cf43ce5a1bb01e2860d
3
+ size 411
runs/Mar04_18-25-10_2cb32fa4c9b0/events.out.tfevents.1741112714.2cb32fa4c9b0.31.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:475aec770a04db7eaf756b3dbf6d0b6467b7bbce5731a1b8c774229e3d1215ba
3
+ size 7545
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "total_flos": 2.479168170953736e+18,
4
+ "train_loss": 0.03568338290625252,
5
+ "train_runtime": 1001.3794,
6
+ "train_samples_per_second": 31.948,
7
+ "train_steps_per_second": 0.999
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,832 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.00038815097650513053,
3
+ "best_model_checkpoint": "./vit-base-fruit-punch/checkpoint-1000",
4
+ "epoch": 8.0,
5
+ "eval_steps": 100,
6
+ "global_step": 1000,
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": 106393.265625,
14
+ "learning_rate": 4.9500000000000004e-05,
15
+ "loss": 1.1459,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.16,
20
+ "grad_norm": 88276.09375,
21
+ "learning_rate": 4.9e-05,
22
+ "loss": 0.7296,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.24,
27
+ "grad_norm": 63037.91796875,
28
+ "learning_rate": 4.85e-05,
29
+ "loss": 0.4458,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.32,
34
+ "grad_norm": 42146.3203125,
35
+ "learning_rate": 4.8e-05,
36
+ "loss": 0.2647,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.4,
41
+ "grad_norm": 27606.990234375,
42
+ "learning_rate": 4.75e-05,
43
+ "loss": 0.1553,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.48,
48
+ "grad_norm": 25297.625,
49
+ "learning_rate": 4.7e-05,
50
+ "loss": 0.1053,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.56,
55
+ "grad_norm": 19000.822265625,
56
+ "learning_rate": 4.6500000000000005e-05,
57
+ "loss": 0.0844,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.64,
62
+ "grad_norm": 26219.373046875,
63
+ "learning_rate": 4.600000000000001e-05,
64
+ "loss": 0.0644,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.72,
69
+ "grad_norm": 13452.0966796875,
70
+ "learning_rate": 4.55e-05,
71
+ "loss": 0.0583,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.8,
76
+ "grad_norm": 11122.3671875,
77
+ "learning_rate": 4.5e-05,
78
+ "loss": 0.046,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.8,
83
+ "eval_accuracy": 1.0,
84
+ "eval_loss": 0.04426969215273857,
85
+ "eval_runtime": 12.7701,
86
+ "eval_samples_per_second": 78.308,
87
+ "eval_steps_per_second": 4.933,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 0.88,
92
+ "grad_norm": 10542.482421875,
93
+ "learning_rate": 4.4500000000000004e-05,
94
+ "loss": 0.0397,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 0.96,
99
+ "grad_norm": 10209.0732421875,
100
+ "learning_rate": 4.4000000000000006e-05,
101
+ "loss": 0.0357,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 1.04,
106
+ "grad_norm": 8613.568359375,
107
+ "learning_rate": 4.35e-05,
108
+ "loss": 0.0325,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 1.12,
113
+ "grad_norm": 7610.2255859375,
114
+ "learning_rate": 4.3e-05,
115
+ "loss": 0.0289,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 1.2,
120
+ "grad_norm": 6793.42236328125,
121
+ "learning_rate": 4.25e-05,
122
+ "loss": 0.0259,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 1.28,
127
+ "grad_norm": 6208.12890625,
128
+ "learning_rate": 4.2e-05,
129
+ "loss": 0.0234,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 1.3599999999999999,
134
+ "grad_norm": 8605.1796875,
135
+ "learning_rate": 4.15e-05,
136
+ "loss": 0.0214,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 1.44,
141
+ "grad_norm": 5137.94921875,
142
+ "learning_rate": 4.1e-05,
143
+ "loss": 0.0197,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 1.52,
148
+ "grad_norm": 4862.05859375,
149
+ "learning_rate": 4.05e-05,
150
+ "loss": 0.0179,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 1.6,
155
+ "grad_norm": 4239.48583984375,
156
+ "learning_rate": 4e-05,
157
+ "loss": 0.0164,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 1.6,
162
+ "eval_accuracy": 0.999,
163
+ "eval_loss": 0.018771709874272346,
164
+ "eval_runtime": 12.8207,
165
+ "eval_samples_per_second": 77.999,
166
+ "eval_steps_per_second": 4.914,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 1.6800000000000002,
171
+ "grad_norm": 4013.29931640625,
172
+ "learning_rate": 3.9500000000000005e-05,
173
+ "loss": 0.0149,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 1.76,
178
+ "grad_norm": 3654.03173828125,
179
+ "learning_rate": 3.9000000000000006e-05,
180
+ "loss": 0.0137,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 1.8399999999999999,
185
+ "grad_norm": 3286.04052734375,
186
+ "learning_rate": 3.85e-05,
187
+ "loss": 0.0125,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 1.92,
192
+ "grad_norm": 3118.806396484375,
193
+ "learning_rate": 3.8e-05,
194
+ "loss": 0.0115,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 2.0,
199
+ "grad_norm": 2893.08203125,
200
+ "learning_rate": 3.7500000000000003e-05,
201
+ "loss": 0.0105,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 2.08,
206
+ "grad_norm": 2609.7880859375,
207
+ "learning_rate": 3.7e-05,
208
+ "loss": 0.0098,
209
+ "step": 260
210
+ },
211
+ {
212
+ "epoch": 2.16,
213
+ "grad_norm": 2484.2294921875,
214
+ "learning_rate": 3.65e-05,
215
+ "loss": 0.0089,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 2.24,
220
+ "grad_norm": 2335.24072265625,
221
+ "learning_rate": 3.6e-05,
222
+ "loss": 0.0082,
223
+ "step": 280
224
+ },
225
+ {
226
+ "epoch": 2.32,
227
+ "grad_norm": 2343.471435546875,
228
+ "learning_rate": 3.55e-05,
229
+ "loss": 0.0076,
230
+ "step": 290
231
+ },
232
+ {
233
+ "epoch": 2.4,
234
+ "grad_norm": 1955.1356201171875,
235
+ "learning_rate": 3.5e-05,
236
+ "loss": 0.007,
237
+ "step": 300
238
+ },
239
+ {
240
+ "epoch": 2.4,
241
+ "eval_accuracy": 1.0,
242
+ "eval_loss": 0.007058804389089346,
243
+ "eval_runtime": 12.8867,
244
+ "eval_samples_per_second": 77.599,
245
+ "eval_steps_per_second": 4.889,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 2.48,
250
+ "grad_norm": 1792.7567138671875,
251
+ "learning_rate": 3.45e-05,
252
+ "loss": 0.0065,
253
+ "step": 310
254
+ },
255
+ {
256
+ "epoch": 2.56,
257
+ "grad_norm": 1647.1790771484375,
258
+ "learning_rate": 3.4000000000000007e-05,
259
+ "loss": 0.006,
260
+ "step": 320
261
+ },
262
+ {
263
+ "epoch": 2.64,
264
+ "grad_norm": 1525.8948974609375,
265
+ "learning_rate": 3.35e-05,
266
+ "loss": 0.0055,
267
+ "step": 330
268
+ },
269
+ {
270
+ "epoch": 2.7199999999999998,
271
+ "grad_norm": 1406.7808837890625,
272
+ "learning_rate": 3.3e-05,
273
+ "loss": 0.0051,
274
+ "step": 340
275
+ },
276
+ {
277
+ "epoch": 2.8,
278
+ "grad_norm": 1297.923583984375,
279
+ "learning_rate": 3.2500000000000004e-05,
280
+ "loss": 0.0048,
281
+ "step": 350
282
+ },
283
+ {
284
+ "epoch": 2.88,
285
+ "grad_norm": 1231.1033935546875,
286
+ "learning_rate": 3.2000000000000005e-05,
287
+ "loss": 0.0044,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 2.96,
292
+ "grad_norm": 1162.668212890625,
293
+ "learning_rate": 3.15e-05,
294
+ "loss": 0.0041,
295
+ "step": 370
296
+ },
297
+ {
298
+ "epoch": 3.04,
299
+ "grad_norm": 1072.34814453125,
300
+ "learning_rate": 3.1e-05,
301
+ "loss": 0.0038,
302
+ "step": 380
303
+ },
304
+ {
305
+ "epoch": 3.12,
306
+ "grad_norm": 981.4453125,
307
+ "learning_rate": 3.05e-05,
308
+ "loss": 0.0035,
309
+ "step": 390
310
+ },
311
+ {
312
+ "epoch": 3.2,
313
+ "grad_norm": 917.4797973632812,
314
+ "learning_rate": 3e-05,
315
+ "loss": 0.0033,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 3.2,
320
+ "eval_accuracy": 1.0,
321
+ "eval_loss": 0.0035334331914782524,
322
+ "eval_runtime": 12.9877,
323
+ "eval_samples_per_second": 76.996,
324
+ "eval_steps_per_second": 4.851,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 3.2800000000000002,
329
+ "grad_norm": 911.9354858398438,
330
+ "learning_rate": 2.95e-05,
331
+ "loss": 0.0031,
332
+ "step": 410
333
+ },
334
+ {
335
+ "epoch": 3.36,
336
+ "grad_norm": 814.162841796875,
337
+ "learning_rate": 2.9e-05,
338
+ "loss": 0.0029,
339
+ "step": 420
340
+ },
341
+ {
342
+ "epoch": 3.44,
343
+ "grad_norm": 755.6058959960938,
344
+ "learning_rate": 2.8499999999999998e-05,
345
+ "loss": 0.0027,
346
+ "step": 430
347
+ },
348
+ {
349
+ "epoch": 3.52,
350
+ "grad_norm": 703.8270263671875,
351
+ "learning_rate": 2.8000000000000003e-05,
352
+ "loss": 0.0025,
353
+ "step": 440
354
+ },
355
+ {
356
+ "epoch": 3.6,
357
+ "grad_norm": 666.2708740234375,
358
+ "learning_rate": 2.7500000000000004e-05,
359
+ "loss": 0.0023,
360
+ "step": 450
361
+ },
362
+ {
363
+ "epoch": 3.68,
364
+ "grad_norm": 613.1011962890625,
365
+ "learning_rate": 2.7000000000000002e-05,
366
+ "loss": 0.0022,
367
+ "step": 460
368
+ },
369
+ {
370
+ "epoch": 3.76,
371
+ "grad_norm": 577.6843872070312,
372
+ "learning_rate": 2.6500000000000004e-05,
373
+ "loss": 0.002,
374
+ "step": 470
375
+ },
376
+ {
377
+ "epoch": 3.84,
378
+ "grad_norm": 554.882568359375,
379
+ "learning_rate": 2.6000000000000002e-05,
380
+ "loss": 0.0019,
381
+ "step": 480
382
+ },
383
+ {
384
+ "epoch": 3.92,
385
+ "grad_norm": 528.9329223632812,
386
+ "learning_rate": 2.5500000000000003e-05,
387
+ "loss": 0.0018,
388
+ "step": 490
389
+ },
390
+ {
391
+ "epoch": 4.0,
392
+ "grad_norm": 483.04278564453125,
393
+ "learning_rate": 2.5e-05,
394
+ "loss": 0.0017,
395
+ "step": 500
396
+ },
397
+ {
398
+ "epoch": 4.0,
399
+ "eval_accuracy": 1.0,
400
+ "eval_loss": 0.002120121382176876,
401
+ "eval_runtime": 13.1662,
402
+ "eval_samples_per_second": 75.952,
403
+ "eval_steps_per_second": 4.785,
404
+ "step": 500
405
+ },
406
+ {
407
+ "epoch": 4.08,
408
+ "grad_norm": 484.7864074707031,
409
+ "learning_rate": 2.45e-05,
410
+ "loss": 0.0016,
411
+ "step": 510
412
+ },
413
+ {
414
+ "epoch": 4.16,
415
+ "grad_norm": 438.2588806152344,
416
+ "learning_rate": 2.4e-05,
417
+ "loss": 0.0015,
418
+ "step": 520
419
+ },
420
+ {
421
+ "epoch": 4.24,
422
+ "grad_norm": 422.3269348144531,
423
+ "learning_rate": 2.35e-05,
424
+ "loss": 0.0014,
425
+ "step": 530
426
+ },
427
+ {
428
+ "epoch": 4.32,
429
+ "grad_norm": 391.5030822753906,
430
+ "learning_rate": 2.3000000000000003e-05,
431
+ "loss": 0.0013,
432
+ "step": 540
433
+ },
434
+ {
435
+ "epoch": 4.4,
436
+ "grad_norm": 370.3298645019531,
437
+ "learning_rate": 2.25e-05,
438
+ "loss": 0.0013,
439
+ "step": 550
440
+ },
441
+ {
442
+ "epoch": 4.48,
443
+ "grad_norm": 364.08148193359375,
444
+ "learning_rate": 2.2000000000000003e-05,
445
+ "loss": 0.0012,
446
+ "step": 560
447
+ },
448
+ {
449
+ "epoch": 4.5600000000000005,
450
+ "grad_norm": 336.76055908203125,
451
+ "learning_rate": 2.15e-05,
452
+ "loss": 0.0011,
453
+ "step": 570
454
+ },
455
+ {
456
+ "epoch": 4.64,
457
+ "grad_norm": 307.3483581542969,
458
+ "learning_rate": 2.1e-05,
459
+ "loss": 0.0011,
460
+ "step": 580
461
+ },
462
+ {
463
+ "epoch": 4.72,
464
+ "grad_norm": 294.3090515136719,
465
+ "learning_rate": 2.05e-05,
466
+ "loss": 0.001,
467
+ "step": 590
468
+ },
469
+ {
470
+ "epoch": 4.8,
471
+ "grad_norm": 281.6009521484375,
472
+ "learning_rate": 2e-05,
473
+ "loss": 0.001,
474
+ "step": 600
475
+ },
476
+ {
477
+ "epoch": 4.8,
478
+ "eval_accuracy": 1.0,
479
+ "eval_loss": 0.0011029670713469386,
480
+ "eval_runtime": 12.8305,
481
+ "eval_samples_per_second": 77.939,
482
+ "eval_steps_per_second": 4.91,
483
+ "step": 600
484
+ },
485
+ {
486
+ "epoch": 4.88,
487
+ "grad_norm": 271.7886657714844,
488
+ "learning_rate": 1.9500000000000003e-05,
489
+ "loss": 0.0009,
490
+ "step": 610
491
+ },
492
+ {
493
+ "epoch": 4.96,
494
+ "grad_norm": 253.9346923828125,
495
+ "learning_rate": 1.9e-05,
496
+ "loss": 0.0009,
497
+ "step": 620
498
+ },
499
+ {
500
+ "epoch": 5.04,
501
+ "grad_norm": 248.29092407226562,
502
+ "learning_rate": 1.85e-05,
503
+ "loss": 0.0008,
504
+ "step": 630
505
+ },
506
+ {
507
+ "epoch": 5.12,
508
+ "grad_norm": 230.48448181152344,
509
+ "learning_rate": 1.8e-05,
510
+ "loss": 0.0008,
511
+ "step": 640
512
+ },
513
+ {
514
+ "epoch": 5.2,
515
+ "grad_norm": 229.324462890625,
516
+ "learning_rate": 1.75e-05,
517
+ "loss": 0.0008,
518
+ "step": 650
519
+ },
520
+ {
521
+ "epoch": 5.28,
522
+ "grad_norm": 212.48696899414062,
523
+ "learning_rate": 1.7000000000000003e-05,
524
+ "loss": 0.0007,
525
+ "step": 660
526
+ },
527
+ {
528
+ "epoch": 5.36,
529
+ "grad_norm": 210.8821258544922,
530
+ "learning_rate": 1.65e-05,
531
+ "loss": 0.0007,
532
+ "step": 670
533
+ },
534
+ {
535
+ "epoch": 5.44,
536
+ "grad_norm": 198.33792114257812,
537
+ "learning_rate": 1.6000000000000003e-05,
538
+ "loss": 0.0007,
539
+ "step": 680
540
+ },
541
+ {
542
+ "epoch": 5.52,
543
+ "grad_norm": 193.6271514892578,
544
+ "learning_rate": 1.55e-05,
545
+ "loss": 0.0006,
546
+ "step": 690
547
+ },
548
+ {
549
+ "epoch": 5.6,
550
+ "grad_norm": 182.02919006347656,
551
+ "learning_rate": 1.5e-05,
552
+ "loss": 0.0006,
553
+ "step": 700
554
+ },
555
+ {
556
+ "epoch": 5.6,
557
+ "eval_accuracy": 1.0,
558
+ "eval_loss": 0.0007458859472535551,
559
+ "eval_runtime": 12.8435,
560
+ "eval_samples_per_second": 77.86,
561
+ "eval_steps_per_second": 4.905,
562
+ "step": 700
563
+ },
564
+ {
565
+ "epoch": 5.68,
566
+ "grad_norm": 178.0136260986328,
567
+ "learning_rate": 1.45e-05,
568
+ "loss": 0.0006,
569
+ "step": 710
570
+ },
571
+ {
572
+ "epoch": 5.76,
573
+ "grad_norm": 169.41432189941406,
574
+ "learning_rate": 1.4000000000000001e-05,
575
+ "loss": 0.0006,
576
+ "step": 720
577
+ },
578
+ {
579
+ "epoch": 5.84,
580
+ "grad_norm": 176.4602813720703,
581
+ "learning_rate": 1.3500000000000001e-05,
582
+ "loss": 0.0005,
583
+ "step": 730
584
+ },
585
+ {
586
+ "epoch": 5.92,
587
+ "grad_norm": 160.14578247070312,
588
+ "learning_rate": 1.3000000000000001e-05,
589
+ "loss": 0.0005,
590
+ "step": 740
591
+ },
592
+ {
593
+ "epoch": 6.0,
594
+ "grad_norm": 166.63352966308594,
595
+ "learning_rate": 1.25e-05,
596
+ "loss": 0.0005,
597
+ "step": 750
598
+ },
599
+ {
600
+ "epoch": 6.08,
601
+ "grad_norm": 152.0838623046875,
602
+ "learning_rate": 1.2e-05,
603
+ "loss": 0.0005,
604
+ "step": 760
605
+ },
606
+ {
607
+ "epoch": 6.16,
608
+ "grad_norm": 147.35086059570312,
609
+ "learning_rate": 1.1500000000000002e-05,
610
+ "loss": 0.0005,
611
+ "step": 770
612
+ },
613
+ {
614
+ "epoch": 6.24,
615
+ "grad_norm": 140.54298400878906,
616
+ "learning_rate": 1.1000000000000001e-05,
617
+ "loss": 0.0005,
618
+ "step": 780
619
+ },
620
+ {
621
+ "epoch": 6.32,
622
+ "grad_norm": 141.50022888183594,
623
+ "learning_rate": 1.05e-05,
624
+ "loss": 0.0005,
625
+ "step": 790
626
+ },
627
+ {
628
+ "epoch": 6.4,
629
+ "grad_norm": 133.0969696044922,
630
+ "learning_rate": 1e-05,
631
+ "loss": 0.0004,
632
+ "step": 800
633
+ },
634
+ {
635
+ "epoch": 6.4,
636
+ "eval_accuracy": 1.0,
637
+ "eval_loss": 0.000529205659404397,
638
+ "eval_runtime": 13.0142,
639
+ "eval_samples_per_second": 76.839,
640
+ "eval_steps_per_second": 4.841,
641
+ "step": 800
642
+ },
643
+ {
644
+ "epoch": 6.48,
645
+ "grad_norm": 132.95997619628906,
646
+ "learning_rate": 9.5e-06,
647
+ "loss": 0.0004,
648
+ "step": 810
649
+ },
650
+ {
651
+ "epoch": 6.5600000000000005,
652
+ "grad_norm": 134.52874755859375,
653
+ "learning_rate": 9e-06,
654
+ "loss": 0.0004,
655
+ "step": 820
656
+ },
657
+ {
658
+ "epoch": 6.64,
659
+ "grad_norm": 126.07892608642578,
660
+ "learning_rate": 8.500000000000002e-06,
661
+ "loss": 0.0004,
662
+ "step": 830
663
+ },
664
+ {
665
+ "epoch": 6.72,
666
+ "grad_norm": 122.06675720214844,
667
+ "learning_rate": 8.000000000000001e-06,
668
+ "loss": 0.0004,
669
+ "step": 840
670
+ },
671
+ {
672
+ "epoch": 6.8,
673
+ "grad_norm": 119.95358276367188,
674
+ "learning_rate": 7.5e-06,
675
+ "loss": 0.0004,
676
+ "step": 850
677
+ },
678
+ {
679
+ "epoch": 6.88,
680
+ "grad_norm": 120.792236328125,
681
+ "learning_rate": 7.000000000000001e-06,
682
+ "loss": 0.0004,
683
+ "step": 860
684
+ },
685
+ {
686
+ "epoch": 6.96,
687
+ "grad_norm": 121.48966217041016,
688
+ "learning_rate": 6.5000000000000004e-06,
689
+ "loss": 0.0004,
690
+ "step": 870
691
+ },
692
+ {
693
+ "epoch": 7.04,
694
+ "grad_norm": 116.1141128540039,
695
+ "learning_rate": 6e-06,
696
+ "loss": 0.0004,
697
+ "step": 880
698
+ },
699
+ {
700
+ "epoch": 7.12,
701
+ "grad_norm": 112.2109375,
702
+ "learning_rate": 5.500000000000001e-06,
703
+ "loss": 0.0004,
704
+ "step": 890
705
+ },
706
+ {
707
+ "epoch": 7.2,
708
+ "grad_norm": 111.11483001708984,
709
+ "learning_rate": 5e-06,
710
+ "loss": 0.0004,
711
+ "step": 900
712
+ },
713
+ {
714
+ "epoch": 7.2,
715
+ "eval_accuracy": 1.0,
716
+ "eval_loss": 0.00041427274118177593,
717
+ "eval_runtime": 12.8429,
718
+ "eval_samples_per_second": 77.864,
719
+ "eval_steps_per_second": 4.905,
720
+ "step": 900
721
+ },
722
+ {
723
+ "epoch": 7.28,
724
+ "grad_norm": 111.32091522216797,
725
+ "learning_rate": 4.5e-06,
726
+ "loss": 0.0004,
727
+ "step": 910
728
+ },
729
+ {
730
+ "epoch": 7.36,
731
+ "grad_norm": 107.73372650146484,
732
+ "learning_rate": 4.000000000000001e-06,
733
+ "loss": 0.0004,
734
+ "step": 920
735
+ },
736
+ {
737
+ "epoch": 7.44,
738
+ "grad_norm": 110.24082946777344,
739
+ "learning_rate": 3.5000000000000004e-06,
740
+ "loss": 0.0004,
741
+ "step": 930
742
+ },
743
+ {
744
+ "epoch": 7.52,
745
+ "grad_norm": 111.30231475830078,
746
+ "learning_rate": 3e-06,
747
+ "loss": 0.0003,
748
+ "step": 940
749
+ },
750
+ {
751
+ "epoch": 7.6,
752
+ "grad_norm": 104.49162292480469,
753
+ "learning_rate": 2.5e-06,
754
+ "loss": 0.0003,
755
+ "step": 950
756
+ },
757
+ {
758
+ "epoch": 7.68,
759
+ "grad_norm": 107.62512969970703,
760
+ "learning_rate": 2.0000000000000003e-06,
761
+ "loss": 0.0003,
762
+ "step": 960
763
+ },
764
+ {
765
+ "epoch": 7.76,
766
+ "grad_norm": 106.17529296875,
767
+ "learning_rate": 1.5e-06,
768
+ "loss": 0.0003,
769
+ "step": 970
770
+ },
771
+ {
772
+ "epoch": 7.84,
773
+ "grad_norm": 104.55513000488281,
774
+ "learning_rate": 1.0000000000000002e-06,
775
+ "loss": 0.0003,
776
+ "step": 980
777
+ },
778
+ {
779
+ "epoch": 7.92,
780
+ "grad_norm": 104.8958969116211,
781
+ "learning_rate": 5.000000000000001e-07,
782
+ "loss": 0.0003,
783
+ "step": 990
784
+ },
785
+ {
786
+ "epoch": 8.0,
787
+ "grad_norm": 110.42820739746094,
788
+ "learning_rate": 0.0,
789
+ "loss": 0.0003,
790
+ "step": 1000
791
+ },
792
+ {
793
+ "epoch": 8.0,
794
+ "eval_accuracy": 1.0,
795
+ "eval_loss": 0.00038815097650513053,
796
+ "eval_runtime": 12.8005,
797
+ "eval_samples_per_second": 78.122,
798
+ "eval_steps_per_second": 4.922,
799
+ "step": 1000
800
+ },
801
+ {
802
+ "epoch": 8.0,
803
+ "step": 1000,
804
+ "total_flos": 2.479168170953736e+18,
805
+ "train_loss": 0.03568338290625252,
806
+ "train_runtime": 1001.3794,
807
+ "train_samples_per_second": 31.948,
808
+ "train_steps_per_second": 0.999
809
+ }
810
+ ],
811
+ "logging_steps": 10,
812
+ "max_steps": 1000,
813
+ "num_input_tokens_seen": 0,
814
+ "num_train_epochs": 8,
815
+ "save_steps": 100,
816
+ "stateful_callbacks": {
817
+ "TrainerControl": {
818
+ "args": {
819
+ "should_epoch_stop": false,
820
+ "should_evaluate": false,
821
+ "should_log": false,
822
+ "should_save": true,
823
+ "should_training_stop": true
824
+ },
825
+ "attributes": {}
826
+ }
827
+ },
828
+ "total_flos": 2.479168170953736e+18,
829
+ "train_batch_size": 32,
830
+ "trial_name": null,
831
+ "trial_params": null
832
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c90d2e87a33e49393c2aa6e2655fd60211e97d72800909f42601002c4b0d812
3
+ size 5304