mnist-mobilevit
Browse files- README.md +83 -0
- all_results.json +13 -0
- config.json +71 -0
- model.safetensors +3 -0
- preprocessor_config.json +17 -0
- test_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +410 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
base_model: apple/mobilevit-xx-small
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
metrics:
|
7 |
+
- accuracy
|
8 |
+
model-index:
|
9 |
+
- name: mnist-mobilevit
|
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 |
+
# mnist-mobilevit
|
17 |
+
|
18 |
+
This model is a fine-tuned version of [apple/mobilevit-xx-small](https://huggingface.co/apple/mobilevit-xx-small) on an unknown dataset.
|
19 |
+
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 0.0185
|
21 |
+
- Accuracy: 0.9929
|
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: 0.0008
|
41 |
+
- train_batch_size: 128
|
42 |
+
- eval_batch_size: 128
|
43 |
+
- seed: 42
|
44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
45 |
+
- lr_scheduler_type: linear
|
46 |
+
- num_epochs: 5
|
47 |
+
- mixed_precision_training: Native AMP
|
48 |
+
|
49 |
+
### Training results
|
50 |
+
|
51 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
52 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|
|
53 |
+
| 0.026 | 0.2132 | 100 | 0.0458 | 0.9867 |
|
54 |
+
| 0.0232 | 0.4264 | 200 | 0.0453 | 0.9867 |
|
55 |
+
| 0.0277 | 0.6397 | 300 | 0.0484 | 0.9863 |
|
56 |
+
| 0.0293 | 0.8529 | 400 | 0.0469 | 0.9865 |
|
57 |
+
| 0.0235 | 1.0661 | 500 | 0.0288 | 0.9899 |
|
58 |
+
| 0.0203 | 1.2793 | 600 | 0.0253 | 0.9924 |
|
59 |
+
| 0.0182 | 1.4925 | 700 | 0.0286 | 0.9916 |
|
60 |
+
| 0.0205 | 1.7058 | 800 | 0.0203 | 0.9935 |
|
61 |
+
| 0.0162 | 1.9190 | 900 | 0.0238 | 0.9913 |
|
62 |
+
| 0.0118 | 2.1322 | 1000 | 0.0247 | 0.9916 |
|
63 |
+
| 0.0121 | 2.3454 | 1100 | 0.0194 | 0.9932 |
|
64 |
+
| 0.0154 | 2.5586 | 1200 | 0.0194 | 0.9933 |
|
65 |
+
| 0.015 | 2.7719 | 1300 | 0.0216 | 0.9933 |
|
66 |
+
| 0.0145 | 2.9851 | 1400 | 0.0238 | 0.9919 |
|
67 |
+
| 0.0098 | 3.1983 | 1500 | 0.0208 | 0.993 |
|
68 |
+
| 0.0093 | 3.4115 | 1600 | 0.0218 | 0.9929 |
|
69 |
+
| 0.0073 | 3.6247 | 1700 | 0.0189 | 0.9933 |
|
70 |
+
| 0.008 | 3.8380 | 1800 | 0.0194 | 0.9932 |
|
71 |
+
| 0.006 | 4.0512 | 1900 | 0.0183 | 0.9938 |
|
72 |
+
| 0.0063 | 4.2644 | 2000 | 0.0184 | 0.9934 |
|
73 |
+
| 0.0043 | 4.4776 | 2100 | 0.0184 | 0.9932 |
|
74 |
+
| 0.0035 | 4.6908 | 2200 | 0.0183 | 0.9931 |
|
75 |
+
| 0.0061 | 4.9041 | 2300 | 0.0184 | 0.9931 |
|
76 |
+
|
77 |
+
|
78 |
+
### Framework versions
|
79 |
+
|
80 |
+
- Transformers 4.42.4
|
81 |
+
- Pytorch 2.3.1+cu121
|
82 |
+
- Datasets 2.20.0
|
83 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 5.0,
|
3 |
+
"eval_accuracy": 0.9929,
|
4 |
+
"eval_loss": 0.018516644835472107,
|
5 |
+
"eval_runtime": 5.3033,
|
6 |
+
"eval_samples_per_second": 1885.618,
|
7 |
+
"eval_steps_per_second": 14.896,
|
8 |
+
"total_flos": 1346208595200000.0,
|
9 |
+
"train_loss": 0.014373551363121472,
|
10 |
+
"train_runtime": 335.5598,
|
11 |
+
"train_samples_per_second": 894.028,
|
12 |
+
"train_steps_per_second": 6.988
|
13 |
+
}
|
config.json
ADDED
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "apple/mobilevit-xx-small",
|
3 |
+
"architectures": [
|
4 |
+
"MobileViTForImageClassification"
|
5 |
+
],
|
6 |
+
"aspp_dropout_prob": 0.1,
|
7 |
+
"aspp_out_channels": 256,
|
8 |
+
"atrous_rates": [
|
9 |
+
6,
|
10 |
+
12,
|
11 |
+
18
|
12 |
+
],
|
13 |
+
"attention_probs_dropout_prob": 0.0,
|
14 |
+
"classifier_dropout_prob": 0.1,
|
15 |
+
"conv_kernel_size": 3,
|
16 |
+
"expand_ratio": 2.0,
|
17 |
+
"hidden_act": "silu",
|
18 |
+
"hidden_dropout_prob": 0.05,
|
19 |
+
"hidden_sizes": [
|
20 |
+
64,
|
21 |
+
80,
|
22 |
+
96
|
23 |
+
],
|
24 |
+
"id2label": {
|
25 |
+
"0": "0",
|
26 |
+
"1": "1",
|
27 |
+
"2": "2",
|
28 |
+
"3": "3",
|
29 |
+
"4": "4",
|
30 |
+
"5": "5",
|
31 |
+
"6": "6",
|
32 |
+
"7": "7",
|
33 |
+
"8": "8",
|
34 |
+
"9": "9"
|
35 |
+
},
|
36 |
+
"image_size": 28,
|
37 |
+
"initializer_range": 0.02,
|
38 |
+
"label2id": {
|
39 |
+
"0": "0",
|
40 |
+
"1": "1",
|
41 |
+
"2": "2",
|
42 |
+
"3": "3",
|
43 |
+
"4": "4",
|
44 |
+
"5": "5",
|
45 |
+
"6": "6",
|
46 |
+
"7": "7",
|
47 |
+
"8": "8",
|
48 |
+
"9": "9"
|
49 |
+
},
|
50 |
+
"layer_norm_eps": 1e-05,
|
51 |
+
"mlp_ratio": 2.0,
|
52 |
+
"model_type": "mobilevit",
|
53 |
+
"neck_hidden_sizes": [
|
54 |
+
16,
|
55 |
+
16,
|
56 |
+
24,
|
57 |
+
48,
|
58 |
+
64,
|
59 |
+
80,
|
60 |
+
320
|
61 |
+
],
|
62 |
+
"num_attention_heads": 4,
|
63 |
+
"num_channels": 1,
|
64 |
+
"output_stride": 32,
|
65 |
+
"patch_size": 2,
|
66 |
+
"problem_type": "single_label_classification",
|
67 |
+
"qkv_bias": true,
|
68 |
+
"semantic_loss_ignore_index": 255,
|
69 |
+
"torch_dtype": "float32",
|
70 |
+
"transformers_version": "4.42.4"
|
71 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0119f9f34f9a51d5c21e72608ca4c85e344b686635be5105a81b7590e73a2b03
|
3 |
+
size 3878480
|
preprocessor_config.json
ADDED
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"crop_size": {
|
3 |
+
"height": 28,
|
4 |
+
"width": 28
|
5 |
+
},
|
6 |
+
"do_center_crop": true,
|
7 |
+
"do_convert_rgb": false,
|
8 |
+
"do_flip_channel_order": false,
|
9 |
+
"do_rescale": true,
|
10 |
+
"do_resize": true,
|
11 |
+
"image_processor_type": "MobileViTImageProcessor",
|
12 |
+
"resample": 2,
|
13 |
+
"rescale_factor": 0.00392156862745098,
|
14 |
+
"size": {
|
15 |
+
"shortest_edge": 28
|
16 |
+
}
|
17 |
+
}
|
test_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 5.0,
|
3 |
+
"eval_accuracy": 0.9929,
|
4 |
+
"eval_loss": 0.018516644835472107,
|
5 |
+
"eval_runtime": 5.3033,
|
6 |
+
"eval_samples_per_second": 1885.618,
|
7 |
+
"eval_steps_per_second": 14.896
|
8 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 5.0,
|
3 |
+
"total_flos": 1346208595200000.0,
|
4 |
+
"train_loss": 0.014373551363121472,
|
5 |
+
"train_runtime": 335.5598,
|
6 |
+
"train_samples_per_second": 894.028,
|
7 |
+
"train_steps_per_second": 6.988
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 5.0,
|
5 |
+
"eval_steps": 100,
|
6 |
+
"global_step": 2345,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.21321961620469082,
|
13 |
+
"grad_norm": 1.6159120798110962,
|
14 |
+
"learning_rate": 0.0007658848614072496,
|
15 |
+
"loss": 0.026,
|
16 |
+
"step": 100
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21321961620469082,
|
20 |
+
"eval_accuracy": 0.9867,
|
21 |
+
"eval_loss": 0.04582102224230766,
|
22 |
+
"eval_runtime": 5.154,
|
23 |
+
"eval_samples_per_second": 1940.235,
|
24 |
+
"eval_steps_per_second": 15.328,
|
25 |
+
"step": 100
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 0.42643923240938164,
|
29 |
+
"grad_norm": 0.22664281725883484,
|
30 |
+
"learning_rate": 0.000731769722814499,
|
31 |
+
"loss": 0.0232,
|
32 |
+
"step": 200
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.42643923240938164,
|
36 |
+
"eval_accuracy": 0.9867,
|
37 |
+
"eval_loss": 0.04527696222066879,
|
38 |
+
"eval_runtime": 5.4041,
|
39 |
+
"eval_samples_per_second": 1850.446,
|
40 |
+
"eval_steps_per_second": 14.619,
|
41 |
+
"step": 200
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 0.6396588486140725,
|
45 |
+
"grad_norm": 1.3475489616394043,
|
46 |
+
"learning_rate": 0.0006976545842217485,
|
47 |
+
"loss": 0.0277,
|
48 |
+
"step": 300
|
49 |
+
},
|
50 |
+
{
|
51 |
+
"epoch": 0.6396588486140725,
|
52 |
+
"eval_accuracy": 0.9863,
|
53 |
+
"eval_loss": 0.04839787632226944,
|
54 |
+
"eval_runtime": 5.3353,
|
55 |
+
"eval_samples_per_second": 1874.323,
|
56 |
+
"eval_steps_per_second": 14.807,
|
57 |
+
"step": 300
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.8528784648187633,
|
61 |
+
"grad_norm": 0.9408140182495117,
|
62 |
+
"learning_rate": 0.0006635394456289979,
|
63 |
+
"loss": 0.0293,
|
64 |
+
"step": 400
|
65 |
+
},
|
66 |
+
{
|
67 |
+
"epoch": 0.8528784648187633,
|
68 |
+
"eval_accuracy": 0.9865,
|
69 |
+
"eval_loss": 0.046898942440748215,
|
70 |
+
"eval_runtime": 5.1935,
|
71 |
+
"eval_samples_per_second": 1925.498,
|
72 |
+
"eval_steps_per_second": 15.211,
|
73 |
+
"step": 400
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 1.0660980810234542,
|
77 |
+
"grad_norm": 0.6018700003623962,
|
78 |
+
"learning_rate": 0.0006294243070362473,
|
79 |
+
"loss": 0.0235,
|
80 |
+
"step": 500
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 1.0660980810234542,
|
84 |
+
"eval_accuracy": 0.9899,
|
85 |
+
"eval_loss": 0.028819510713219643,
|
86 |
+
"eval_runtime": 5.1407,
|
87 |
+
"eval_samples_per_second": 1945.277,
|
88 |
+
"eval_steps_per_second": 15.368,
|
89 |
+
"step": 500
|
90 |
+
},
|
91 |
+
{
|
92 |
+
"epoch": 1.279317697228145,
|
93 |
+
"grad_norm": 0.7029837369918823,
|
94 |
+
"learning_rate": 0.0005953091684434968,
|
95 |
+
"loss": 0.0203,
|
96 |
+
"step": 600
|
97 |
+
},
|
98 |
+
{
|
99 |
+
"epoch": 1.279317697228145,
|
100 |
+
"eval_accuracy": 0.9924,
|
101 |
+
"eval_loss": 0.02526690438389778,
|
102 |
+
"eval_runtime": 4.869,
|
103 |
+
"eval_samples_per_second": 2053.81,
|
104 |
+
"eval_steps_per_second": 16.225,
|
105 |
+
"step": 600
|
106 |
+
},
|
107 |
+
{
|
108 |
+
"epoch": 1.4925373134328357,
|
109 |
+
"grad_norm": 0.17393378913402557,
|
110 |
+
"learning_rate": 0.0005611940298507463,
|
111 |
+
"loss": 0.0182,
|
112 |
+
"step": 700
|
113 |
+
},
|
114 |
+
{
|
115 |
+
"epoch": 1.4925373134328357,
|
116 |
+
"eval_accuracy": 0.9916,
|
117 |
+
"eval_loss": 0.028603948652744293,
|
118 |
+
"eval_runtime": 4.6847,
|
119 |
+
"eval_samples_per_second": 2134.586,
|
120 |
+
"eval_steps_per_second": 16.863,
|
121 |
+
"step": 700
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 1.7057569296375266,
|
125 |
+
"grad_norm": 0.9470173120498657,
|
126 |
+
"learning_rate": 0.0005270788912579957,
|
127 |
+
"loss": 0.0205,
|
128 |
+
"step": 800
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 1.7057569296375266,
|
132 |
+
"eval_accuracy": 0.9935,
|
133 |
+
"eval_loss": 0.02031560428440571,
|
134 |
+
"eval_runtime": 5.0835,
|
135 |
+
"eval_samples_per_second": 1967.155,
|
136 |
+
"eval_steps_per_second": 15.541,
|
137 |
+
"step": 800
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.9189765458422174,
|
141 |
+
"grad_norm": 1.4821853637695312,
|
142 |
+
"learning_rate": 0.0004929637526652453,
|
143 |
+
"loss": 0.0162,
|
144 |
+
"step": 900
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.9189765458422174,
|
148 |
+
"eval_accuracy": 0.9913,
|
149 |
+
"eval_loss": 0.023791342973709106,
|
150 |
+
"eval_runtime": 5.2196,
|
151 |
+
"eval_samples_per_second": 1915.863,
|
152 |
+
"eval_steps_per_second": 15.135,
|
153 |
+
"step": 900
|
154 |
+
},
|
155 |
+
{
|
156 |
+
"epoch": 2.1321961620469083,
|
157 |
+
"grad_norm": 1.0076653957366943,
|
158 |
+
"learning_rate": 0.0004588486140724947,
|
159 |
+
"loss": 0.0118,
|
160 |
+
"step": 1000
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.1321961620469083,
|
164 |
+
"eval_accuracy": 0.9916,
|
165 |
+
"eval_loss": 0.024731909856200218,
|
166 |
+
"eval_runtime": 5.8209,
|
167 |
+
"eval_samples_per_second": 1717.944,
|
168 |
+
"eval_steps_per_second": 13.572,
|
169 |
+
"step": 1000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 2.345415778251599,
|
173 |
+
"grad_norm": 0.39255017042160034,
|
174 |
+
"learning_rate": 0.0004247334754797441,
|
175 |
+
"loss": 0.0121,
|
176 |
+
"step": 1100
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 2.345415778251599,
|
180 |
+
"eval_accuracy": 0.9932,
|
181 |
+
"eval_loss": 0.019426949322223663,
|
182 |
+
"eval_runtime": 5.5038,
|
183 |
+
"eval_samples_per_second": 1816.931,
|
184 |
+
"eval_steps_per_second": 14.354,
|
185 |
+
"step": 1100
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 2.55863539445629,
|
189 |
+
"grad_norm": 0.4671665132045746,
|
190 |
+
"learning_rate": 0.0003906183368869936,
|
191 |
+
"loss": 0.0154,
|
192 |
+
"step": 1200
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"epoch": 2.55863539445629,
|
196 |
+
"eval_accuracy": 0.9933,
|
197 |
+
"eval_loss": 0.01936325989663601,
|
198 |
+
"eval_runtime": 5.4304,
|
199 |
+
"eval_samples_per_second": 1841.5,
|
200 |
+
"eval_steps_per_second": 14.548,
|
201 |
+
"step": 1200
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 2.771855010660981,
|
205 |
+
"grad_norm": 0.17253048717975616,
|
206 |
+
"learning_rate": 0.0003565031982942431,
|
207 |
+
"loss": 0.015,
|
208 |
+
"step": 1300
|
209 |
+
},
|
210 |
+
{
|
211 |
+
"epoch": 2.771855010660981,
|
212 |
+
"eval_accuracy": 0.9933,
|
213 |
+
"eval_loss": 0.02162059210240841,
|
214 |
+
"eval_runtime": 5.3729,
|
215 |
+
"eval_samples_per_second": 1861.177,
|
216 |
+
"eval_steps_per_second": 14.703,
|
217 |
+
"step": 1300
|
218 |
+
},
|
219 |
+
{
|
220 |
+
"epoch": 2.9850746268656714,
|
221 |
+
"grad_norm": 0.8343185782432556,
|
222 |
+
"learning_rate": 0.00032238805970149256,
|
223 |
+
"loss": 0.0145,
|
224 |
+
"step": 1400
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 2.9850746268656714,
|
228 |
+
"eval_accuracy": 0.9919,
|
229 |
+
"eval_loss": 0.02381654642522335,
|
230 |
+
"eval_runtime": 5.1099,
|
231 |
+
"eval_samples_per_second": 1957.0,
|
232 |
+
"eval_steps_per_second": 15.46,
|
233 |
+
"step": 1400
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 3.1982942430703627,
|
237 |
+
"grad_norm": 0.3106481432914734,
|
238 |
+
"learning_rate": 0.000288272921108742,
|
239 |
+
"loss": 0.0098,
|
240 |
+
"step": 1500
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 3.1982942430703627,
|
244 |
+
"eval_accuracy": 0.993,
|
245 |
+
"eval_loss": 0.020756520330905914,
|
246 |
+
"eval_runtime": 4.9767,
|
247 |
+
"eval_samples_per_second": 2009.353,
|
248 |
+
"eval_steps_per_second": 15.874,
|
249 |
+
"step": 1500
|
250 |
+
},
|
251 |
+
{
|
252 |
+
"epoch": 3.411513859275053,
|
253 |
+
"grad_norm": 0.03616774454712868,
|
254 |
+
"learning_rate": 0.0002541577825159915,
|
255 |
+
"loss": 0.0093,
|
256 |
+
"step": 1600
|
257 |
+
},
|
258 |
+
{
|
259 |
+
"epoch": 3.411513859275053,
|
260 |
+
"eval_accuracy": 0.9929,
|
261 |
+
"eval_loss": 0.021822581067681313,
|
262 |
+
"eval_runtime": 5.4965,
|
263 |
+
"eval_samples_per_second": 1819.356,
|
264 |
+
"eval_steps_per_second": 14.373,
|
265 |
+
"step": 1600
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 3.624733475479744,
|
269 |
+
"grad_norm": 0.6045345067977905,
|
270 |
+
"learning_rate": 0.00022004264392324095,
|
271 |
+
"loss": 0.0073,
|
272 |
+
"step": 1700
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 3.624733475479744,
|
276 |
+
"eval_accuracy": 0.9933,
|
277 |
+
"eval_loss": 0.018862707540392876,
|
278 |
+
"eval_runtime": 5.7766,
|
279 |
+
"eval_samples_per_second": 1731.12,
|
280 |
+
"eval_steps_per_second": 13.676,
|
281 |
+
"step": 1700
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 3.837953091684435,
|
285 |
+
"grad_norm": 0.631215512752533,
|
286 |
+
"learning_rate": 0.0001859275053304904,
|
287 |
+
"loss": 0.008,
|
288 |
+
"step": 1800
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 3.837953091684435,
|
292 |
+
"eval_accuracy": 0.9932,
|
293 |
+
"eval_loss": 0.01944512128829956,
|
294 |
+
"eval_runtime": 5.1522,
|
295 |
+
"eval_samples_per_second": 1940.922,
|
296 |
+
"eval_steps_per_second": 15.333,
|
297 |
+
"step": 1800
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 4.051172707889126,
|
301 |
+
"grad_norm": 0.5996153950691223,
|
302 |
+
"learning_rate": 0.0001518123667377399,
|
303 |
+
"loss": 0.006,
|
304 |
+
"step": 1900
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 4.051172707889126,
|
308 |
+
"eval_accuracy": 0.9938,
|
309 |
+
"eval_loss": 0.018317226320505142,
|
310 |
+
"eval_runtime": 4.9544,
|
311 |
+
"eval_samples_per_second": 2018.428,
|
312 |
+
"eval_steps_per_second": 15.946,
|
313 |
+
"step": 1900
|
314 |
+
},
|
315 |
+
{
|
316 |
+
"epoch": 4.264392324093817,
|
317 |
+
"grad_norm": 0.0069321137852966785,
|
318 |
+
"learning_rate": 0.00011769722814498933,
|
319 |
+
"loss": 0.0063,
|
320 |
+
"step": 2000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 4.264392324093817,
|
324 |
+
"eval_accuracy": 0.9934,
|
325 |
+
"eval_loss": 0.018384862691164017,
|
326 |
+
"eval_runtime": 4.7636,
|
327 |
+
"eval_samples_per_second": 2099.232,
|
328 |
+
"eval_steps_per_second": 16.584,
|
329 |
+
"step": 2000
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 4.477611940298507,
|
333 |
+
"grad_norm": 0.06509185582399368,
|
334 |
+
"learning_rate": 8.392324093816631e-05,
|
335 |
+
"loss": 0.0043,
|
336 |
+
"step": 2100
|
337 |
+
},
|
338 |
+
{
|
339 |
+
"epoch": 4.477611940298507,
|
340 |
+
"eval_accuracy": 0.9932,
|
341 |
+
"eval_loss": 0.018380142748355865,
|
342 |
+
"eval_runtime": 4.6951,
|
343 |
+
"eval_samples_per_second": 2129.883,
|
344 |
+
"eval_steps_per_second": 16.826,
|
345 |
+
"step": 2100
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 4.690831556503198,
|
349 |
+
"grad_norm": 0.17103737592697144,
|
350 |
+
"learning_rate": 4.980810234541578e-05,
|
351 |
+
"loss": 0.0035,
|
352 |
+
"step": 2200
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 4.690831556503198,
|
356 |
+
"eval_accuracy": 0.9931,
|
357 |
+
"eval_loss": 0.018344268202781677,
|
358 |
+
"eval_runtime": 4.6229,
|
359 |
+
"eval_samples_per_second": 2163.133,
|
360 |
+
"eval_steps_per_second": 17.089,
|
361 |
+
"step": 2200
|
362 |
+
},
|
363 |
+
{
|
364 |
+
"epoch": 4.904051172707889,
|
365 |
+
"grad_norm": 0.6465263962745667,
|
366 |
+
"learning_rate": 1.5692963752665246e-05,
|
367 |
+
"loss": 0.0061,
|
368 |
+
"step": 2300
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 4.904051172707889,
|
372 |
+
"eval_accuracy": 0.9931,
|
373 |
+
"eval_loss": 0.018412619829177856,
|
374 |
+
"eval_runtime": 4.7593,
|
375 |
+
"eval_samples_per_second": 2101.13,
|
376 |
+
"eval_steps_per_second": 16.599,
|
377 |
+
"step": 2300
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 5.0,
|
381 |
+
"step": 2345,
|
382 |
+
"total_flos": 1346208595200000.0,
|
383 |
+
"train_loss": 0.014373551363121472,
|
384 |
+
"train_runtime": 335.5598,
|
385 |
+
"train_samples_per_second": 894.028,
|
386 |
+
"train_steps_per_second": 6.988
|
387 |
+
}
|
388 |
+
],
|
389 |
+
"logging_steps": 100,
|
390 |
+
"max_steps": 2345,
|
391 |
+
"num_input_tokens_seen": 0,
|
392 |
+
"num_train_epochs": 5,
|
393 |
+
"save_steps": 3000,
|
394 |
+
"stateful_callbacks": {
|
395 |
+
"TrainerControl": {
|
396 |
+
"args": {
|
397 |
+
"should_epoch_stop": false,
|
398 |
+
"should_evaluate": false,
|
399 |
+
"should_log": false,
|
400 |
+
"should_save": true,
|
401 |
+
"should_training_stop": true
|
402 |
+
},
|
403 |
+
"attributes": {}
|
404 |
+
}
|
405 |
+
},
|
406 |
+
"total_flos": 1346208595200000.0,
|
407 |
+
"train_batch_size": 128,
|
408 |
+
"trial_name": null,
|
409 |
+
"trial_params": null
|
410 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f025d41c388918472644cf05d640f1bc355f2bc4b3d2ce3674bdc0e5ceeb4001
|
3 |
+
size 5112
|