Upload 8 files
Browse files- config.json +60 -0
- optimizer.pt +3 -0
- preprocessor_config.json +29 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +2047 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/beit-base-patch16-224-pt22k-ft22k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BeitForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"auxiliary_channels": 256,
|
| 8 |
+
"auxiliary_concat_input": false,
|
| 9 |
+
"auxiliary_loss_weight": 0.4,
|
| 10 |
+
"auxiliary_num_convs": 1,
|
| 11 |
+
"drop_path_rate": 0.1,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.0,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": "adenocarcinoma",
|
| 17 |
+
"1": "large.cell",
|
| 18 |
+
"2": "normal",
|
| 19 |
+
"3": "squamous.cell"
|
| 20 |
+
},
|
| 21 |
+
"image_size": 224,
|
| 22 |
+
"initializer_range": 0.02,
|
| 23 |
+
"intermediate_size": 3072,
|
| 24 |
+
"label2id": {
|
| 25 |
+
"adenocarcinoma": 0,
|
| 26 |
+
"large.cell": 1,
|
| 27 |
+
"normal": 2,
|
| 28 |
+
"squamous.cell": 3
|
| 29 |
+
},
|
| 30 |
+
"layer_norm_eps": 1e-12,
|
| 31 |
+
"layer_scale_init_value": 0.1,
|
| 32 |
+
"model_type": "beit",
|
| 33 |
+
"num_attention_heads": 12,
|
| 34 |
+
"num_channels": 3,
|
| 35 |
+
"num_hidden_layers": 12,
|
| 36 |
+
"out_indices": [
|
| 37 |
+
3,
|
| 38 |
+
5,
|
| 39 |
+
7,
|
| 40 |
+
11
|
| 41 |
+
],
|
| 42 |
+
"patch_size": 16,
|
| 43 |
+
"pool_scales": [
|
| 44 |
+
1,
|
| 45 |
+
2,
|
| 46 |
+
3,
|
| 47 |
+
6
|
| 48 |
+
],
|
| 49 |
+
"problem_type": "single_label_classification",
|
| 50 |
+
"semantic_loss_ignore_index": 255,
|
| 51 |
+
"torch_dtype": "float32",
|
| 52 |
+
"transformers_version": "4.33.0",
|
| 53 |
+
"use_absolute_position_embeddings": false,
|
| 54 |
+
"use_auxiliary_head": true,
|
| 55 |
+
"use_mask_token": false,
|
| 56 |
+
"use_mean_pooling": true,
|
| 57 |
+
"use_relative_position_bias": true,
|
| 58 |
+
"use_shared_relative_position_bias": false,
|
| 59 |
+
"vocab_size": 8192
|
| 60 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca4c6019eee4c994eddd28a567fc9cd5db53a57cfeb6ba8a0c0ffb16572ce355
|
| 3 |
+
size 686306437
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 224,
|
| 4 |
+
"width": 224
|
| 5 |
+
},
|
| 6 |
+
"do_center_crop": false,
|
| 7 |
+
"do_normalize": true,
|
| 8 |
+
"do_reduce_labels": false,
|
| 9 |
+
"do_rescale": true,
|
| 10 |
+
"do_resize": true,
|
| 11 |
+
"feature_extractor_type": "BeitFeatureExtractor",
|
| 12 |
+
"image_mean": [
|
| 13 |
+
0.5,
|
| 14 |
+
0.5,
|
| 15 |
+
0.5
|
| 16 |
+
],
|
| 17 |
+
"image_processor_type": "BeitImageProcessor",
|
| 18 |
+
"image_std": [
|
| 19 |
+
0.5,
|
| 20 |
+
0.5,
|
| 21 |
+
0.5
|
| 22 |
+
],
|
| 23 |
+
"resample": 2,
|
| 24 |
+
"rescale_factor": 0.00392156862745098,
|
| 25 |
+
"size": {
|
| 26 |
+
"height": 224,
|
| 27 |
+
"width": 224
|
| 28 |
+
}
|
| 29 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abc2032c47243bd99fb36a857773ff5b53396b0934cf0f1e74a12f2e0dc5f7f7
|
| 3 |
+
size 343136393
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c5d90cfcd2db4dcff5325014f91ea781e58f84df60353e1b6771643d75ece2a
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8344ae90a4dd6860b9c65ccaa955daa156231ed363e25ba26f9074c4b6836d8
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2047 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9925,
|
| 3 |
+
"best_model_checkpoint": "./BeiT/checkpoint-3058",
|
| 4 |
+
"epoch": 22.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3058,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.07,
|
| 13 |
+
"learning_rate": 4.796163069544365e-07,
|
| 14 |
+
"loss": 1.3973,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.14,
|
| 19 |
+
"learning_rate": 9.59232613908873e-07,
|
| 20 |
+
"loss": 1.3593,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.22,
|
| 25 |
+
"learning_rate": 1.4388489208633094e-06,
|
| 26 |
+
"loss": 1.2572,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.29,
|
| 31 |
+
"learning_rate": 1.918465227817746e-06,
|
| 32 |
+
"loss": 1.1821,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.36,
|
| 37 |
+
"learning_rate": 2.3980815347721824e-06,
|
| 38 |
+
"loss": 1.0659,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.43,
|
| 43 |
+
"learning_rate": 2.877697841726619e-06,
|
| 44 |
+
"loss": 0.926,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.5,
|
| 49 |
+
"learning_rate": 3.3573141486810557e-06,
|
| 50 |
+
"loss": 0.8208,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.58,
|
| 55 |
+
"learning_rate": 3.836930455635492e-06,
|
| 56 |
+
"loss": 0.7399,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.65,
|
| 61 |
+
"learning_rate": 4.316546762589928e-06,
|
| 62 |
+
"loss": 0.7633,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.72,
|
| 67 |
+
"learning_rate": 4.796163069544365e-06,
|
| 68 |
+
"loss": 0.7109,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.79,
|
| 73 |
+
"learning_rate": 5.275779376498802e-06,
|
| 74 |
+
"loss": 0.6965,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.86,
|
| 79 |
+
"learning_rate": 5.755395683453238e-06,
|
| 80 |
+
"loss": 0.6409,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.94,
|
| 85 |
+
"learning_rate": 6.2350119904076745e-06,
|
| 86 |
+
"loss": 0.6413,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.0,
|
| 91 |
+
"eval_accuracy": 0.74,
|
| 92 |
+
"eval_loss": 0.6694741249084473,
|
| 93 |
+
"eval_runtime": 7.2563,
|
| 94 |
+
"eval_samples_per_second": 55.124,
|
| 95 |
+
"eval_steps_per_second": 3.445,
|
| 96 |
+
"step": 139
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.01,
|
| 100 |
+
"learning_rate": 6.714628297362111e-06,
|
| 101 |
+
"loss": 0.6174,
|
| 102 |
+
"step": 140
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.08,
|
| 106 |
+
"learning_rate": 7.194244604316547e-06,
|
| 107 |
+
"loss": 0.5881,
|
| 108 |
+
"step": 150
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.15,
|
| 112 |
+
"learning_rate": 7.673860911270984e-06,
|
| 113 |
+
"loss": 0.5689,
|
| 114 |
+
"step": 160
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.22,
|
| 118 |
+
"learning_rate": 8.153477218225421e-06,
|
| 119 |
+
"loss": 0.5596,
|
| 120 |
+
"step": 170
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 1.29,
|
| 124 |
+
"learning_rate": 8.633093525179856e-06,
|
| 125 |
+
"loss": 0.5571,
|
| 126 |
+
"step": 180
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.37,
|
| 130 |
+
"learning_rate": 9.112709832134293e-06,
|
| 131 |
+
"loss": 0.4756,
|
| 132 |
+
"step": 190
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.44,
|
| 136 |
+
"learning_rate": 9.59232613908873e-06,
|
| 137 |
+
"loss": 0.4701,
|
| 138 |
+
"step": 200
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.51,
|
| 142 |
+
"learning_rate": 1.0071942446043167e-05,
|
| 143 |
+
"loss": 0.4945,
|
| 144 |
+
"step": 210
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.58,
|
| 148 |
+
"learning_rate": 1.0551558752997603e-05,
|
| 149 |
+
"loss": 0.4683,
|
| 150 |
+
"step": 220
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.65,
|
| 154 |
+
"learning_rate": 1.103117505995204e-05,
|
| 155 |
+
"loss": 0.4497,
|
| 156 |
+
"step": 230
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.73,
|
| 160 |
+
"learning_rate": 1.1510791366906475e-05,
|
| 161 |
+
"loss": 0.4282,
|
| 162 |
+
"step": 240
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.8,
|
| 166 |
+
"learning_rate": 1.1990407673860914e-05,
|
| 167 |
+
"loss": 0.3876,
|
| 168 |
+
"step": 250
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.87,
|
| 172 |
+
"learning_rate": 1.2470023980815349e-05,
|
| 173 |
+
"loss": 0.3782,
|
| 174 |
+
"step": 260
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 1.94,
|
| 178 |
+
"learning_rate": 1.2949640287769784e-05,
|
| 179 |
+
"loss": 0.391,
|
| 180 |
+
"step": 270
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 2.0,
|
| 184 |
+
"eval_accuracy": 0.8575,
|
| 185 |
+
"eval_loss": 0.37282153964042664,
|
| 186 |
+
"eval_runtime": 6.8753,
|
| 187 |
+
"eval_samples_per_second": 58.18,
|
| 188 |
+
"eval_steps_per_second": 3.636,
|
| 189 |
+
"step": 278
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 2.01,
|
| 193 |
+
"learning_rate": 1.3429256594724223e-05,
|
| 194 |
+
"loss": 0.3699,
|
| 195 |
+
"step": 280
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 2.09,
|
| 199 |
+
"learning_rate": 1.3908872901678658e-05,
|
| 200 |
+
"loss": 0.3421,
|
| 201 |
+
"step": 290
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.16,
|
| 205 |
+
"learning_rate": 1.4388489208633095e-05,
|
| 206 |
+
"loss": 0.3257,
|
| 207 |
+
"step": 300
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.23,
|
| 211 |
+
"learning_rate": 1.4868105515587531e-05,
|
| 212 |
+
"loss": 0.2969,
|
| 213 |
+
"step": 310
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.3,
|
| 217 |
+
"learning_rate": 1.534772182254197e-05,
|
| 218 |
+
"loss": 0.2498,
|
| 219 |
+
"step": 320
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 2.37,
|
| 223 |
+
"learning_rate": 1.5827338129496403e-05,
|
| 224 |
+
"loss": 0.2428,
|
| 225 |
+
"step": 330
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 2.45,
|
| 229 |
+
"learning_rate": 1.6306954436450842e-05,
|
| 230 |
+
"loss": 0.2272,
|
| 231 |
+
"step": 340
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 2.52,
|
| 235 |
+
"learning_rate": 1.6786570743405277e-05,
|
| 236 |
+
"loss": 0.267,
|
| 237 |
+
"step": 350
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.59,
|
| 241 |
+
"learning_rate": 1.7266187050359712e-05,
|
| 242 |
+
"loss": 0.2261,
|
| 243 |
+
"step": 360
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 2.66,
|
| 247 |
+
"learning_rate": 1.774580335731415e-05,
|
| 248 |
+
"loss": 0.2052,
|
| 249 |
+
"step": 370
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.73,
|
| 253 |
+
"learning_rate": 1.8225419664268586e-05,
|
| 254 |
+
"loss": 0.236,
|
| 255 |
+
"step": 380
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.81,
|
| 259 |
+
"learning_rate": 1.870503597122302e-05,
|
| 260 |
+
"loss": 0.2256,
|
| 261 |
+
"step": 390
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 2.88,
|
| 265 |
+
"learning_rate": 1.918465227817746e-05,
|
| 266 |
+
"loss": 0.1794,
|
| 267 |
+
"step": 400
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 2.95,
|
| 271 |
+
"learning_rate": 1.9664268585131895e-05,
|
| 272 |
+
"loss": 0.2041,
|
| 273 |
+
"step": 410
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 3.0,
|
| 277 |
+
"eval_accuracy": 0.9075,
|
| 278 |
+
"eval_loss": 0.2558615207672119,
|
| 279 |
+
"eval_runtime": 6.9006,
|
| 280 |
+
"eval_samples_per_second": 57.966,
|
| 281 |
+
"eval_steps_per_second": 3.623,
|
| 282 |
+
"step": 417
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 3.02,
|
| 286 |
+
"learning_rate": 1.9984012789768187e-05,
|
| 287 |
+
"loss": 0.2144,
|
| 288 |
+
"step": 420
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 3.09,
|
| 292 |
+
"learning_rate": 1.993072208899547e-05,
|
| 293 |
+
"loss": 0.2121,
|
| 294 |
+
"step": 430
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.17,
|
| 298 |
+
"learning_rate": 1.9877431388222757e-05,
|
| 299 |
+
"loss": 0.1434,
|
| 300 |
+
"step": 440
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 3.24,
|
| 304 |
+
"learning_rate": 1.982414068745004e-05,
|
| 305 |
+
"loss": 0.1592,
|
| 306 |
+
"step": 450
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 3.31,
|
| 310 |
+
"learning_rate": 1.9770849986677328e-05,
|
| 311 |
+
"loss": 0.1343,
|
| 312 |
+
"step": 460
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 3.38,
|
| 316 |
+
"learning_rate": 1.971755928590461e-05,
|
| 317 |
+
"loss": 0.1119,
|
| 318 |
+
"step": 470
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.45,
|
| 322 |
+
"learning_rate": 1.9664268585131895e-05,
|
| 323 |
+
"loss": 0.0998,
|
| 324 |
+
"step": 480
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 3.53,
|
| 328 |
+
"learning_rate": 1.961097788435918e-05,
|
| 329 |
+
"loss": 0.1291,
|
| 330 |
+
"step": 490
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 3.6,
|
| 334 |
+
"learning_rate": 1.955768718358647e-05,
|
| 335 |
+
"loss": 0.105,
|
| 336 |
+
"step": 500
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.67,
|
| 340 |
+
"learning_rate": 1.950439648281375e-05,
|
| 341 |
+
"loss": 0.0977,
|
| 342 |
+
"step": 510
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 3.74,
|
| 346 |
+
"learning_rate": 1.9451105782041035e-05,
|
| 347 |
+
"loss": 0.0733,
|
| 348 |
+
"step": 520
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 3.81,
|
| 352 |
+
"learning_rate": 1.9397815081268322e-05,
|
| 353 |
+
"loss": 0.0885,
|
| 354 |
+
"step": 530
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 3.88,
|
| 358 |
+
"learning_rate": 1.9344524380495602e-05,
|
| 359 |
+
"loss": 0.1262,
|
| 360 |
+
"step": 540
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 3.96,
|
| 364 |
+
"learning_rate": 1.929123367972289e-05,
|
| 365 |
+
"loss": 0.1423,
|
| 366 |
+
"step": 550
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 4.0,
|
| 370 |
+
"eval_accuracy": 0.94,
|
| 371 |
+
"eval_loss": 0.18216529488563538,
|
| 372 |
+
"eval_runtime": 7.2699,
|
| 373 |
+
"eval_samples_per_second": 55.022,
|
| 374 |
+
"eval_steps_per_second": 3.439,
|
| 375 |
+
"step": 556
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 4.03,
|
| 379 |
+
"learning_rate": 1.9237942978950176e-05,
|
| 380 |
+
"loss": 0.1161,
|
| 381 |
+
"step": 560
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 4.1,
|
| 385 |
+
"learning_rate": 1.918465227817746e-05,
|
| 386 |
+
"loss": 0.0915,
|
| 387 |
+
"step": 570
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 4.17,
|
| 391 |
+
"learning_rate": 1.9131361577404743e-05,
|
| 392 |
+
"loss": 0.0604,
|
| 393 |
+
"step": 580
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 4.24,
|
| 397 |
+
"learning_rate": 1.907807087663203e-05,
|
| 398 |
+
"loss": 0.0902,
|
| 399 |
+
"step": 590
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 4.32,
|
| 403 |
+
"learning_rate": 1.9024780175859313e-05,
|
| 404 |
+
"loss": 0.0442,
|
| 405 |
+
"step": 600
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 4.39,
|
| 409 |
+
"learning_rate": 1.89714894750866e-05,
|
| 410 |
+
"loss": 0.046,
|
| 411 |
+
"step": 610
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 4.46,
|
| 415 |
+
"learning_rate": 1.8918198774313884e-05,
|
| 416 |
+
"loss": 0.0934,
|
| 417 |
+
"step": 620
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 4.53,
|
| 421 |
+
"learning_rate": 1.8864908073541167e-05,
|
| 422 |
+
"loss": 0.0748,
|
| 423 |
+
"step": 630
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 4.6,
|
| 427 |
+
"learning_rate": 1.8811617372768454e-05,
|
| 428 |
+
"loss": 0.0561,
|
| 429 |
+
"step": 640
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 4.68,
|
| 433 |
+
"learning_rate": 1.8758326671995738e-05,
|
| 434 |
+
"loss": 0.0618,
|
| 435 |
+
"step": 650
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 4.75,
|
| 439 |
+
"learning_rate": 1.870503597122302e-05,
|
| 440 |
+
"loss": 0.0415,
|
| 441 |
+
"step": 660
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.82,
|
| 445 |
+
"learning_rate": 1.8651745270450308e-05,
|
| 446 |
+
"loss": 0.073,
|
| 447 |
+
"step": 670
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 4.89,
|
| 451 |
+
"learning_rate": 1.8598454569677595e-05,
|
| 452 |
+
"loss": 0.0632,
|
| 453 |
+
"step": 680
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 4.96,
|
| 457 |
+
"learning_rate": 1.8545163868904878e-05,
|
| 458 |
+
"loss": 0.0568,
|
| 459 |
+
"step": 690
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 5.0,
|
| 463 |
+
"eval_accuracy": 0.975,
|
| 464 |
+
"eval_loss": 0.06315237283706665,
|
| 465 |
+
"eval_runtime": 7.0385,
|
| 466 |
+
"eval_samples_per_second": 56.83,
|
| 467 |
+
"eval_steps_per_second": 3.552,
|
| 468 |
+
"step": 695
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 5.04,
|
| 472 |
+
"learning_rate": 1.8491873168132162e-05,
|
| 473 |
+
"loss": 0.0407,
|
| 474 |
+
"step": 700
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 5.11,
|
| 478 |
+
"learning_rate": 1.843858246735945e-05,
|
| 479 |
+
"loss": 0.0445,
|
| 480 |
+
"step": 710
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 5.18,
|
| 484 |
+
"learning_rate": 1.8385291766586732e-05,
|
| 485 |
+
"loss": 0.029,
|
| 486 |
+
"step": 720
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 5.25,
|
| 490 |
+
"learning_rate": 1.833200106581402e-05,
|
| 491 |
+
"loss": 0.031,
|
| 492 |
+
"step": 730
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 5.32,
|
| 496 |
+
"learning_rate": 1.8278710365041302e-05,
|
| 497 |
+
"loss": 0.0548,
|
| 498 |
+
"step": 740
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.4,
|
| 502 |
+
"learning_rate": 1.8225419664268586e-05,
|
| 503 |
+
"loss": 0.0242,
|
| 504 |
+
"step": 750
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 5.47,
|
| 508 |
+
"learning_rate": 1.8172128963495873e-05,
|
| 509 |
+
"loss": 0.0266,
|
| 510 |
+
"step": 760
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 5.54,
|
| 514 |
+
"learning_rate": 1.8118838262723156e-05,
|
| 515 |
+
"loss": 0.0263,
|
| 516 |
+
"step": 770
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 5.61,
|
| 520 |
+
"learning_rate": 1.806554756195044e-05,
|
| 521 |
+
"loss": 0.0468,
|
| 522 |
+
"step": 780
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 5.68,
|
| 526 |
+
"learning_rate": 1.8012256861177727e-05,
|
| 527 |
+
"loss": 0.0375,
|
| 528 |
+
"step": 790
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 5.76,
|
| 532 |
+
"learning_rate": 1.795896616040501e-05,
|
| 533 |
+
"loss": 0.0203,
|
| 534 |
+
"step": 800
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 5.83,
|
| 538 |
+
"learning_rate": 1.7905675459632294e-05,
|
| 539 |
+
"loss": 0.0131,
|
| 540 |
+
"step": 810
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 5.9,
|
| 544 |
+
"learning_rate": 1.785238475885958e-05,
|
| 545 |
+
"loss": 0.0722,
|
| 546 |
+
"step": 820
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 5.97,
|
| 550 |
+
"learning_rate": 1.7799094058086864e-05,
|
| 551 |
+
"loss": 0.059,
|
| 552 |
+
"step": 830
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 6.0,
|
| 556 |
+
"eval_accuracy": 0.9625,
|
| 557 |
+
"eval_loss": 0.1519964635372162,
|
| 558 |
+
"eval_runtime": 7.1264,
|
| 559 |
+
"eval_samples_per_second": 56.129,
|
| 560 |
+
"eval_steps_per_second": 3.508,
|
| 561 |
+
"step": 834
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 6.04,
|
| 565 |
+
"learning_rate": 1.774580335731415e-05,
|
| 566 |
+
"loss": 0.0351,
|
| 567 |
+
"step": 840
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 6.12,
|
| 571 |
+
"learning_rate": 1.7692512656541434e-05,
|
| 572 |
+
"loss": 0.0197,
|
| 573 |
+
"step": 850
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 6.19,
|
| 577 |
+
"learning_rate": 1.763922195576872e-05,
|
| 578 |
+
"loss": 0.0217,
|
| 579 |
+
"step": 860
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 6.26,
|
| 583 |
+
"learning_rate": 1.7585931254996005e-05,
|
| 584 |
+
"loss": 0.0217,
|
| 585 |
+
"step": 870
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 6.33,
|
| 589 |
+
"learning_rate": 1.753264055422329e-05,
|
| 590 |
+
"loss": 0.02,
|
| 591 |
+
"step": 880
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 6.4,
|
| 595 |
+
"learning_rate": 1.7479349853450575e-05,
|
| 596 |
+
"loss": 0.0236,
|
| 597 |
+
"step": 890
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 6.47,
|
| 601 |
+
"learning_rate": 1.742605915267786e-05,
|
| 602 |
+
"loss": 0.0118,
|
| 603 |
+
"step": 900
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 6.55,
|
| 607 |
+
"learning_rate": 1.7372768451905145e-05,
|
| 608 |
+
"loss": 0.0266,
|
| 609 |
+
"step": 910
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 6.62,
|
| 613 |
+
"learning_rate": 1.731947775113243e-05,
|
| 614 |
+
"loss": 0.0183,
|
| 615 |
+
"step": 920
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 6.69,
|
| 619 |
+
"learning_rate": 1.7266187050359712e-05,
|
| 620 |
+
"loss": 0.0454,
|
| 621 |
+
"step": 930
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 6.76,
|
| 625 |
+
"learning_rate": 1.7212896349587e-05,
|
| 626 |
+
"loss": 0.0166,
|
| 627 |
+
"step": 940
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 6.83,
|
| 631 |
+
"learning_rate": 1.7159605648814283e-05,
|
| 632 |
+
"loss": 0.0099,
|
| 633 |
+
"step": 950
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 6.91,
|
| 637 |
+
"learning_rate": 1.710631494804157e-05,
|
| 638 |
+
"loss": 0.0106,
|
| 639 |
+
"step": 960
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 6.98,
|
| 643 |
+
"learning_rate": 1.7053024247268853e-05,
|
| 644 |
+
"loss": 0.0396,
|
| 645 |
+
"step": 970
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 7.0,
|
| 649 |
+
"eval_accuracy": 0.98,
|
| 650 |
+
"eval_loss": 0.09595155715942383,
|
| 651 |
+
"eval_runtime": 6.9691,
|
| 652 |
+
"eval_samples_per_second": 57.396,
|
| 653 |
+
"eval_steps_per_second": 3.587,
|
| 654 |
+
"step": 973
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 7.05,
|
| 658 |
+
"learning_rate": 1.6999733546496136e-05,
|
| 659 |
+
"loss": 0.0152,
|
| 660 |
+
"step": 980
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 7.12,
|
| 664 |
+
"learning_rate": 1.6946442845723423e-05,
|
| 665 |
+
"loss": 0.0171,
|
| 666 |
+
"step": 990
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 7.19,
|
| 670 |
+
"learning_rate": 1.689315214495071e-05,
|
| 671 |
+
"loss": 0.0058,
|
| 672 |
+
"step": 1000
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 7.27,
|
| 676 |
+
"learning_rate": 1.683986144417799e-05,
|
| 677 |
+
"loss": 0.0287,
|
| 678 |
+
"step": 1010
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 7.34,
|
| 682 |
+
"learning_rate": 1.6786570743405277e-05,
|
| 683 |
+
"loss": 0.0342,
|
| 684 |
+
"step": 1020
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 7.41,
|
| 688 |
+
"learning_rate": 1.6733280042632564e-05,
|
| 689 |
+
"loss": 0.0331,
|
| 690 |
+
"step": 1030
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 7.48,
|
| 694 |
+
"learning_rate": 1.6679989341859847e-05,
|
| 695 |
+
"loss": 0.0153,
|
| 696 |
+
"step": 1040
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 7.55,
|
| 700 |
+
"learning_rate": 1.662669864108713e-05,
|
| 701 |
+
"loss": 0.0444,
|
| 702 |
+
"step": 1050
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 7.63,
|
| 706 |
+
"learning_rate": 1.6573407940314418e-05,
|
| 707 |
+
"loss": 0.009,
|
| 708 |
+
"step": 1060
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 7.7,
|
| 712 |
+
"learning_rate": 1.65201172395417e-05,
|
| 713 |
+
"loss": 0.012,
|
| 714 |
+
"step": 1070
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 7.77,
|
| 718 |
+
"learning_rate": 1.6466826538768985e-05,
|
| 719 |
+
"loss": 0.0158,
|
| 720 |
+
"step": 1080
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 7.84,
|
| 724 |
+
"learning_rate": 1.641353583799627e-05,
|
| 725 |
+
"loss": 0.0328,
|
| 726 |
+
"step": 1090
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 7.91,
|
| 730 |
+
"learning_rate": 1.6360245137223555e-05,
|
| 731 |
+
"loss": 0.0159,
|
| 732 |
+
"step": 1100
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 7.99,
|
| 736 |
+
"learning_rate": 1.6306954436450842e-05,
|
| 737 |
+
"loss": 0.0076,
|
| 738 |
+
"step": 1110
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 8.0,
|
| 742 |
+
"eval_accuracy": 0.98,
|
| 743 |
+
"eval_loss": 0.07062595337629318,
|
| 744 |
+
"eval_runtime": 6.9534,
|
| 745 |
+
"eval_samples_per_second": 57.526,
|
| 746 |
+
"eval_steps_per_second": 3.595,
|
| 747 |
+
"step": 1112
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 8.06,
|
| 751 |
+
"learning_rate": 1.6253663735678125e-05,
|
| 752 |
+
"loss": 0.0208,
|
| 753 |
+
"step": 1120
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 8.13,
|
| 757 |
+
"learning_rate": 1.620037303490541e-05,
|
| 758 |
+
"loss": 0.0138,
|
| 759 |
+
"step": 1130
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 8.2,
|
| 763 |
+
"learning_rate": 1.6147082334132696e-05,
|
| 764 |
+
"loss": 0.0404,
|
| 765 |
+
"step": 1140
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 8.27,
|
| 769 |
+
"learning_rate": 1.609379163335998e-05,
|
| 770 |
+
"loss": 0.0395,
|
| 771 |
+
"step": 1150
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 8.35,
|
| 775 |
+
"learning_rate": 1.6040500932587263e-05,
|
| 776 |
+
"loss": 0.0044,
|
| 777 |
+
"step": 1160
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 8.42,
|
| 781 |
+
"learning_rate": 1.598721023181455e-05,
|
| 782 |
+
"loss": 0.0055,
|
| 783 |
+
"step": 1170
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 8.49,
|
| 787 |
+
"learning_rate": 1.5933919531041836e-05,
|
| 788 |
+
"loss": 0.017,
|
| 789 |
+
"step": 1180
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 8.56,
|
| 793 |
+
"learning_rate": 1.588062883026912e-05,
|
| 794 |
+
"loss": 0.0405,
|
| 795 |
+
"step": 1190
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 8.63,
|
| 799 |
+
"learning_rate": 1.5827338129496403e-05,
|
| 800 |
+
"loss": 0.0405,
|
| 801 |
+
"step": 1200
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 8.71,
|
| 805 |
+
"learning_rate": 1.577404742872369e-05,
|
| 806 |
+
"loss": 0.0289,
|
| 807 |
+
"step": 1210
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 8.78,
|
| 811 |
+
"learning_rate": 1.5720756727950974e-05,
|
| 812 |
+
"loss": 0.0217,
|
| 813 |
+
"step": 1220
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 8.85,
|
| 817 |
+
"learning_rate": 1.566746602717826e-05,
|
| 818 |
+
"loss": 0.0319,
|
| 819 |
+
"step": 1230
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 8.92,
|
| 823 |
+
"learning_rate": 1.5614175326405544e-05,
|
| 824 |
+
"loss": 0.0176,
|
| 825 |
+
"step": 1240
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 8.99,
|
| 829 |
+
"learning_rate": 1.5560884625632828e-05,
|
| 830 |
+
"loss": 0.0316,
|
| 831 |
+
"step": 1250
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 9.0,
|
| 835 |
+
"eval_accuracy": 0.9775,
|
| 836 |
+
"eval_loss": 0.1436285525560379,
|
| 837 |
+
"eval_runtime": 6.9746,
|
| 838 |
+
"eval_samples_per_second": 57.351,
|
| 839 |
+
"eval_steps_per_second": 3.584,
|
| 840 |
+
"step": 1251
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 9.06,
|
| 844 |
+
"learning_rate": 1.5507593924860114e-05,
|
| 845 |
+
"loss": 0.0269,
|
| 846 |
+
"step": 1260
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 9.14,
|
| 850 |
+
"learning_rate": 1.5454303224087398e-05,
|
| 851 |
+
"loss": 0.016,
|
| 852 |
+
"step": 1270
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 9.21,
|
| 856 |
+
"learning_rate": 1.540101252331468e-05,
|
| 857 |
+
"loss": 0.0046,
|
| 858 |
+
"step": 1280
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 9.28,
|
| 862 |
+
"learning_rate": 1.534772182254197e-05,
|
| 863 |
+
"loss": 0.0437,
|
| 864 |
+
"step": 1290
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 9.35,
|
| 868 |
+
"learning_rate": 1.5294431121769252e-05,
|
| 869 |
+
"loss": 0.0109,
|
| 870 |
+
"step": 1300
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 9.42,
|
| 874 |
+
"learning_rate": 1.5241140420996537e-05,
|
| 875 |
+
"loss": 0.011,
|
| 876 |
+
"step": 1310
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 9.5,
|
| 880 |
+
"learning_rate": 1.5187849720223822e-05,
|
| 881 |
+
"loss": 0.0045,
|
| 882 |
+
"step": 1320
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 9.57,
|
| 886 |
+
"learning_rate": 1.5134559019451107e-05,
|
| 887 |
+
"loss": 0.0157,
|
| 888 |
+
"step": 1330
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 9.64,
|
| 892 |
+
"learning_rate": 1.5081268318678392e-05,
|
| 893 |
+
"loss": 0.0077,
|
| 894 |
+
"step": 1340
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 9.71,
|
| 898 |
+
"learning_rate": 1.5027977617905676e-05,
|
| 899 |
+
"loss": 0.0638,
|
| 900 |
+
"step": 1350
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 9.78,
|
| 904 |
+
"learning_rate": 1.4974686917132961e-05,
|
| 905 |
+
"loss": 0.0166,
|
| 906 |
+
"step": 1360
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 9.86,
|
| 910 |
+
"learning_rate": 1.4921396216360246e-05,
|
| 911 |
+
"loss": 0.0189,
|
| 912 |
+
"step": 1370
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 9.93,
|
| 916 |
+
"learning_rate": 1.4868105515587531e-05,
|
| 917 |
+
"loss": 0.0171,
|
| 918 |
+
"step": 1380
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 10.0,
|
| 922 |
+
"learning_rate": 1.4814814814814815e-05,
|
| 923 |
+
"loss": 0.0306,
|
| 924 |
+
"step": 1390
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 10.0,
|
| 928 |
+
"eval_accuracy": 0.9075,
|
| 929 |
+
"eval_loss": 0.604502260684967,
|
| 930 |
+
"eval_runtime": 6.8184,
|
| 931 |
+
"eval_samples_per_second": 58.664,
|
| 932 |
+
"eval_steps_per_second": 3.667,
|
| 933 |
+
"step": 1390
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 10.07,
|
| 937 |
+
"learning_rate": 1.47615241140421e-05,
|
| 938 |
+
"loss": 0.0172,
|
| 939 |
+
"step": 1400
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 10.14,
|
| 943 |
+
"learning_rate": 1.4708233413269385e-05,
|
| 944 |
+
"loss": 0.0334,
|
| 945 |
+
"step": 1410
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 10.22,
|
| 949 |
+
"learning_rate": 1.4654942712496672e-05,
|
| 950 |
+
"loss": 0.0366,
|
| 951 |
+
"step": 1420
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 10.29,
|
| 955 |
+
"learning_rate": 1.4601652011723954e-05,
|
| 956 |
+
"loss": 0.0109,
|
| 957 |
+
"step": 1430
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 10.36,
|
| 961 |
+
"learning_rate": 1.454836131095124e-05,
|
| 962 |
+
"loss": 0.032,
|
| 963 |
+
"step": 1440
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 10.43,
|
| 967 |
+
"learning_rate": 1.4495070610178526e-05,
|
| 968 |
+
"loss": 0.0078,
|
| 969 |
+
"step": 1450
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 10.5,
|
| 973 |
+
"learning_rate": 1.4441779909405811e-05,
|
| 974 |
+
"loss": 0.024,
|
| 975 |
+
"step": 1460
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 10.58,
|
| 979 |
+
"learning_rate": 1.4388489208633095e-05,
|
| 980 |
+
"loss": 0.0126,
|
| 981 |
+
"step": 1470
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 10.65,
|
| 985 |
+
"learning_rate": 1.433519850786038e-05,
|
| 986 |
+
"loss": 0.0014,
|
| 987 |
+
"step": 1480
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 10.72,
|
| 991 |
+
"learning_rate": 1.4281907807087665e-05,
|
| 992 |
+
"loss": 0.0308,
|
| 993 |
+
"step": 1490
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 10.79,
|
| 997 |
+
"learning_rate": 1.422861710631495e-05,
|
| 998 |
+
"loss": 0.023,
|
| 999 |
+
"step": 1500
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 10.86,
|
| 1003 |
+
"learning_rate": 1.4175326405542234e-05,
|
| 1004 |
+
"loss": 0.0044,
|
| 1005 |
+
"step": 1510
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 10.94,
|
| 1009 |
+
"learning_rate": 1.4122035704769519e-05,
|
| 1010 |
+
"loss": 0.0093,
|
| 1011 |
+
"step": 1520
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 11.0,
|
| 1015 |
+
"eval_accuracy": 0.99,
|
| 1016 |
+
"eval_loss": 0.09465105831623077,
|
| 1017 |
+
"eval_runtime": 7.036,
|
| 1018 |
+
"eval_samples_per_second": 56.85,
|
| 1019 |
+
"eval_steps_per_second": 3.553,
|
| 1020 |
+
"step": 1529
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 11.01,
|
| 1024 |
+
"learning_rate": 1.4068745003996804e-05,
|
| 1025 |
+
"loss": 0.0059,
|
| 1026 |
+
"step": 1530
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 11.08,
|
| 1030 |
+
"learning_rate": 1.4015454303224087e-05,
|
| 1031 |
+
"loss": 0.0108,
|
| 1032 |
+
"step": 1540
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 11.15,
|
| 1036 |
+
"learning_rate": 1.3962163602451373e-05,
|
| 1037 |
+
"loss": 0.0021,
|
| 1038 |
+
"step": 1550
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 11.22,
|
| 1042 |
+
"learning_rate": 1.3908872901678658e-05,
|
| 1043 |
+
"loss": 0.0233,
|
| 1044 |
+
"step": 1560
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 11.29,
|
| 1048 |
+
"learning_rate": 1.3855582200905943e-05,
|
| 1049 |
+
"loss": 0.0138,
|
| 1050 |
+
"step": 1570
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 11.37,
|
| 1054 |
+
"learning_rate": 1.3802291500133226e-05,
|
| 1055 |
+
"loss": 0.0021,
|
| 1056 |
+
"step": 1580
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 11.44,
|
| 1060 |
+
"learning_rate": 1.3749000799360512e-05,
|
| 1061 |
+
"loss": 0.0431,
|
| 1062 |
+
"step": 1590
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 11.51,
|
| 1066 |
+
"learning_rate": 1.3695710098587799e-05,
|
| 1067 |
+
"loss": 0.0433,
|
| 1068 |
+
"step": 1600
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 11.58,
|
| 1072 |
+
"learning_rate": 1.3642419397815084e-05,
|
| 1073 |
+
"loss": 0.0074,
|
| 1074 |
+
"step": 1610
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 11.65,
|
| 1078 |
+
"learning_rate": 1.3589128697042367e-05,
|
| 1079 |
+
"loss": 0.0163,
|
| 1080 |
+
"step": 1620
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 11.73,
|
| 1084 |
+
"learning_rate": 1.3535837996269652e-05,
|
| 1085 |
+
"loss": 0.0118,
|
| 1086 |
+
"step": 1630
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 11.8,
|
| 1090 |
+
"learning_rate": 1.3482547295496938e-05,
|
| 1091 |
+
"loss": 0.0139,
|
| 1092 |
+
"step": 1640
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 11.87,
|
| 1096 |
+
"learning_rate": 1.3429256594724223e-05,
|
| 1097 |
+
"loss": 0.0021,
|
| 1098 |
+
"step": 1650
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 11.94,
|
| 1102 |
+
"learning_rate": 1.3375965893951506e-05,
|
| 1103 |
+
"loss": 0.0268,
|
| 1104 |
+
"step": 1660
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 12.0,
|
| 1108 |
+
"eval_accuracy": 0.98,
|
| 1109 |
+
"eval_loss": 0.18784460425376892,
|
| 1110 |
+
"eval_runtime": 6.9369,
|
| 1111 |
+
"eval_samples_per_second": 57.662,
|
| 1112 |
+
"eval_steps_per_second": 3.604,
|
| 1113 |
+
"step": 1668
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 12.01,
|
| 1117 |
+
"learning_rate": 1.3322675193178791e-05,
|
| 1118 |
+
"loss": 0.0057,
|
| 1119 |
+
"step": 1670
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 12.09,
|
| 1123 |
+
"learning_rate": 1.3269384492406077e-05,
|
| 1124 |
+
"loss": 0.0175,
|
| 1125 |
+
"step": 1680
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 12.16,
|
| 1129 |
+
"learning_rate": 1.3216093791633362e-05,
|
| 1130 |
+
"loss": 0.0159,
|
| 1131 |
+
"step": 1690
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 12.23,
|
| 1135 |
+
"learning_rate": 1.3162803090860645e-05,
|
| 1136 |
+
"loss": 0.0209,
|
| 1137 |
+
"step": 1700
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 12.3,
|
| 1141 |
+
"learning_rate": 1.310951239008793e-05,
|
| 1142 |
+
"loss": 0.0383,
|
| 1143 |
+
"step": 1710
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 12.37,
|
| 1147 |
+
"learning_rate": 1.3056221689315216e-05,
|
| 1148 |
+
"loss": 0.0291,
|
| 1149 |
+
"step": 1720
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 12.45,
|
| 1153 |
+
"learning_rate": 1.30029309885425e-05,
|
| 1154 |
+
"loss": 0.0002,
|
| 1155 |
+
"step": 1730
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 12.52,
|
| 1159 |
+
"learning_rate": 1.2949640287769784e-05,
|
| 1160 |
+
"loss": 0.0009,
|
| 1161 |
+
"step": 1740
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 12.59,
|
| 1165 |
+
"learning_rate": 1.289634958699707e-05,
|
| 1166 |
+
"loss": 0.004,
|
| 1167 |
+
"step": 1750
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 12.66,
|
| 1171 |
+
"learning_rate": 1.2843058886224356e-05,
|
| 1172 |
+
"loss": 0.003,
|
| 1173 |
+
"step": 1760
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 12.73,
|
| 1177 |
+
"learning_rate": 1.2789768185451638e-05,
|
| 1178 |
+
"loss": 0.0116,
|
| 1179 |
+
"step": 1770
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 12.81,
|
| 1183 |
+
"learning_rate": 1.2736477484678925e-05,
|
| 1184 |
+
"loss": 0.04,
|
| 1185 |
+
"step": 1780
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 12.88,
|
| 1189 |
+
"learning_rate": 1.268318678390621e-05,
|
| 1190 |
+
"loss": 0.0225,
|
| 1191 |
+
"step": 1790
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 12.95,
|
| 1195 |
+
"learning_rate": 1.2629896083133495e-05,
|
| 1196 |
+
"loss": 0.0037,
|
| 1197 |
+
"step": 1800
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 13.0,
|
| 1201 |
+
"eval_accuracy": 0.9825,
|
| 1202 |
+
"eval_loss": 0.11539778858423233,
|
| 1203 |
+
"eval_runtime": 6.9266,
|
| 1204 |
+
"eval_samples_per_second": 57.749,
|
| 1205 |
+
"eval_steps_per_second": 3.609,
|
| 1206 |
+
"step": 1807
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 13.02,
|
| 1210 |
+
"learning_rate": 1.2576605382360779e-05,
|
| 1211 |
+
"loss": 0.0465,
|
| 1212 |
+
"step": 1810
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 13.09,
|
| 1216 |
+
"learning_rate": 1.2523314681588064e-05,
|
| 1217 |
+
"loss": 0.0189,
|
| 1218 |
+
"step": 1820
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 13.17,
|
| 1222 |
+
"learning_rate": 1.2470023980815349e-05,
|
| 1223 |
+
"loss": 0.0393,
|
| 1224 |
+
"step": 1830
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 13.24,
|
| 1228 |
+
"learning_rate": 1.2416733280042634e-05,
|
| 1229 |
+
"loss": 0.0043,
|
| 1230 |
+
"step": 1840
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 13.31,
|
| 1234 |
+
"learning_rate": 1.2363442579269918e-05,
|
| 1235 |
+
"loss": 0.0283,
|
| 1236 |
+
"step": 1850
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 13.38,
|
| 1240 |
+
"learning_rate": 1.2310151878497203e-05,
|
| 1241 |
+
"loss": 0.0546,
|
| 1242 |
+
"step": 1860
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 13.45,
|
| 1246 |
+
"learning_rate": 1.2256861177724488e-05,
|
| 1247 |
+
"loss": 0.0078,
|
| 1248 |
+
"step": 1870
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 13.53,
|
| 1252 |
+
"learning_rate": 1.2203570476951773e-05,
|
| 1253 |
+
"loss": 0.0057,
|
| 1254 |
+
"step": 1880
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 13.6,
|
| 1258 |
+
"learning_rate": 1.2150279776179057e-05,
|
| 1259 |
+
"loss": 0.002,
|
| 1260 |
+
"step": 1890
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 13.67,
|
| 1264 |
+
"learning_rate": 1.2096989075406342e-05,
|
| 1265 |
+
"loss": 0.0221,
|
| 1266 |
+
"step": 1900
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 13.74,
|
| 1270 |
+
"learning_rate": 1.2043698374633627e-05,
|
| 1271 |
+
"loss": 0.0107,
|
| 1272 |
+
"step": 1910
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 13.81,
|
| 1276 |
+
"learning_rate": 1.1990407673860914e-05,
|
| 1277 |
+
"loss": 0.0088,
|
| 1278 |
+
"step": 1920
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 13.88,
|
| 1282 |
+
"learning_rate": 1.1937116973088196e-05,
|
| 1283 |
+
"loss": 0.0363,
|
| 1284 |
+
"step": 1930
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 13.96,
|
| 1288 |
+
"learning_rate": 1.1883826272315483e-05,
|
| 1289 |
+
"loss": 0.0005,
|
| 1290 |
+
"step": 1940
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 14.0,
|
| 1294 |
+
"eval_accuracy": 0.9875,
|
| 1295 |
+
"eval_loss": 0.13052797317504883,
|
| 1296 |
+
"eval_runtime": 7.0568,
|
| 1297 |
+
"eval_samples_per_second": 56.683,
|
| 1298 |
+
"eval_steps_per_second": 3.543,
|
| 1299 |
+
"step": 1946
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 14.03,
|
| 1303 |
+
"learning_rate": 1.1830535571542768e-05,
|
| 1304 |
+
"loss": 0.0045,
|
| 1305 |
+
"step": 1950
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 14.1,
|
| 1309 |
+
"learning_rate": 1.1777244870770053e-05,
|
| 1310 |
+
"loss": 0.0345,
|
| 1311 |
+
"step": 1960
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 14.17,
|
| 1315 |
+
"learning_rate": 1.1723954169997336e-05,
|
| 1316 |
+
"loss": 0.0001,
|
| 1317 |
+
"step": 1970
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 14.24,
|
| 1321 |
+
"learning_rate": 1.1670663469224622e-05,
|
| 1322 |
+
"loss": 0.0215,
|
| 1323 |
+
"step": 1980
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 14.32,
|
| 1327 |
+
"learning_rate": 1.1617372768451907e-05,
|
| 1328 |
+
"loss": 0.0264,
|
| 1329 |
+
"step": 1990
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 14.39,
|
| 1333 |
+
"learning_rate": 1.156408206767919e-05,
|
| 1334 |
+
"loss": 0.0002,
|
| 1335 |
+
"step": 2000
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 14.46,
|
| 1339 |
+
"learning_rate": 1.1510791366906475e-05,
|
| 1340 |
+
"loss": 0.0017,
|
| 1341 |
+
"step": 2010
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 14.53,
|
| 1345 |
+
"learning_rate": 1.145750066613376e-05,
|
| 1346 |
+
"loss": 0.01,
|
| 1347 |
+
"step": 2020
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 14.6,
|
| 1351 |
+
"learning_rate": 1.1404209965361046e-05,
|
| 1352 |
+
"loss": 0.0,
|
| 1353 |
+
"step": 2030
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 14.68,
|
| 1357 |
+
"learning_rate": 1.135091926458833e-05,
|
| 1358 |
+
"loss": 0.0333,
|
| 1359 |
+
"step": 2040
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 14.75,
|
| 1363 |
+
"learning_rate": 1.1297628563815614e-05,
|
| 1364 |
+
"loss": 0.0002,
|
| 1365 |
+
"step": 2050
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 14.82,
|
| 1369 |
+
"learning_rate": 1.12443378630429e-05,
|
| 1370 |
+
"loss": 0.0147,
|
| 1371 |
+
"step": 2060
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 14.89,
|
| 1375 |
+
"learning_rate": 1.1191047162270185e-05,
|
| 1376 |
+
"loss": 0.0652,
|
| 1377 |
+
"step": 2070
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 14.96,
|
| 1381 |
+
"learning_rate": 1.1137756461497468e-05,
|
| 1382 |
+
"loss": 0.0232,
|
| 1383 |
+
"step": 2080
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 15.0,
|
| 1387 |
+
"eval_accuracy": 0.9825,
|
| 1388 |
+
"eval_loss": 0.1760598123073578,
|
| 1389 |
+
"eval_runtime": 6.7923,
|
| 1390 |
+
"eval_samples_per_second": 58.89,
|
| 1391 |
+
"eval_steps_per_second": 3.681,
|
| 1392 |
+
"step": 2085
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 15.04,
|
| 1396 |
+
"learning_rate": 1.1084465760724753e-05,
|
| 1397 |
+
"loss": 0.0012,
|
| 1398 |
+
"step": 2090
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 15.11,
|
| 1402 |
+
"learning_rate": 1.103117505995204e-05,
|
| 1403 |
+
"loss": 0.0,
|
| 1404 |
+
"step": 2100
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 15.18,
|
| 1408 |
+
"learning_rate": 1.0977884359179325e-05,
|
| 1409 |
+
"loss": 0.0399,
|
| 1410 |
+
"step": 2110
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 15.25,
|
| 1414 |
+
"learning_rate": 1.0924593658406609e-05,
|
| 1415 |
+
"loss": 0.0176,
|
| 1416 |
+
"step": 2120
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 15.32,
|
| 1420 |
+
"learning_rate": 1.0871302957633894e-05,
|
| 1421 |
+
"loss": 0.0102,
|
| 1422 |
+
"step": 2130
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 15.4,
|
| 1426 |
+
"learning_rate": 1.081801225686118e-05,
|
| 1427 |
+
"loss": 0.0043,
|
| 1428 |
+
"step": 2140
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 15.47,
|
| 1432 |
+
"learning_rate": 1.0764721556088464e-05,
|
| 1433 |
+
"loss": 0.0036,
|
| 1434 |
+
"step": 2150
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 15.54,
|
| 1438 |
+
"learning_rate": 1.0711430855315748e-05,
|
| 1439 |
+
"loss": 0.0034,
|
| 1440 |
+
"step": 2160
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 15.61,
|
| 1444 |
+
"learning_rate": 1.0658140154543033e-05,
|
| 1445 |
+
"loss": 0.0045,
|
| 1446 |
+
"step": 2170
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 15.68,
|
| 1450 |
+
"learning_rate": 1.0604849453770318e-05,
|
| 1451 |
+
"loss": 0.0002,
|
| 1452 |
+
"step": 2180
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 15.76,
|
| 1456 |
+
"learning_rate": 1.0551558752997603e-05,
|
| 1457 |
+
"loss": 0.0264,
|
| 1458 |
+
"step": 2190
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 15.83,
|
| 1462 |
+
"learning_rate": 1.0498268052224887e-05,
|
| 1463 |
+
"loss": 0.0149,
|
| 1464 |
+
"step": 2200
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 15.9,
|
| 1468 |
+
"learning_rate": 1.0444977351452172e-05,
|
| 1469 |
+
"loss": 0.0015,
|
| 1470 |
+
"step": 2210
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 15.97,
|
| 1474 |
+
"learning_rate": 1.0391686650679457e-05,
|
| 1475 |
+
"loss": 0.0048,
|
| 1476 |
+
"step": 2220
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 16.0,
|
| 1480 |
+
"eval_accuracy": 0.98,
|
| 1481 |
+
"eval_loss": 0.28786617517471313,
|
| 1482 |
+
"eval_runtime": 7.1065,
|
| 1483 |
+
"eval_samples_per_second": 56.287,
|
| 1484 |
+
"eval_steps_per_second": 3.518,
|
| 1485 |
+
"step": 2224
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 16.04,
|
| 1489 |
+
"learning_rate": 1.0338395949906742e-05,
|
| 1490 |
+
"loss": 0.0,
|
| 1491 |
+
"step": 2230
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 16.12,
|
| 1495 |
+
"learning_rate": 1.0285105249134026e-05,
|
| 1496 |
+
"loss": 0.0135,
|
| 1497 |
+
"step": 2240
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 16.19,
|
| 1501 |
+
"learning_rate": 1.0231814548361311e-05,
|
| 1502 |
+
"loss": 0.0054,
|
| 1503 |
+
"step": 2250
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 16.26,
|
| 1507 |
+
"learning_rate": 1.0178523847588598e-05,
|
| 1508 |
+
"loss": 0.0005,
|
| 1509 |
+
"step": 2260
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 16.33,
|
| 1513 |
+
"learning_rate": 1.012523314681588e-05,
|
| 1514 |
+
"loss": 0.0045,
|
| 1515 |
+
"step": 2270
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 16.4,
|
| 1519 |
+
"learning_rate": 1.0071942446043167e-05,
|
| 1520 |
+
"loss": 0.0101,
|
| 1521 |
+
"step": 2280
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 16.47,
|
| 1525 |
+
"learning_rate": 1.0018651745270452e-05,
|
| 1526 |
+
"loss": 0.0241,
|
| 1527 |
+
"step": 2290
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 16.55,
|
| 1531 |
+
"learning_rate": 9.965361044497735e-06,
|
| 1532 |
+
"loss": 0.0035,
|
| 1533 |
+
"step": 2300
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 16.62,
|
| 1537 |
+
"learning_rate": 9.91207034372502e-06,
|
| 1538 |
+
"loss": 0.0315,
|
| 1539 |
+
"step": 2310
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 16.69,
|
| 1543 |
+
"learning_rate": 9.858779642952306e-06,
|
| 1544 |
+
"loss": 0.0049,
|
| 1545 |
+
"step": 2320
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 16.76,
|
| 1549 |
+
"learning_rate": 9.80548894217959e-06,
|
| 1550 |
+
"loss": 0.0,
|
| 1551 |
+
"step": 2330
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 16.83,
|
| 1555 |
+
"learning_rate": 9.752198241406874e-06,
|
| 1556 |
+
"loss": 0.0378,
|
| 1557 |
+
"step": 2340
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 16.91,
|
| 1561 |
+
"learning_rate": 9.698907540634161e-06,
|
| 1562 |
+
"loss": 0.0,
|
| 1563 |
+
"step": 2350
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 16.98,
|
| 1567 |
+
"learning_rate": 9.645616839861445e-06,
|
| 1568 |
+
"loss": 0.0,
|
| 1569 |
+
"step": 2360
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 17.0,
|
| 1573 |
+
"eval_accuracy": 0.985,
|
| 1574 |
+
"eval_loss": 0.20865210890769958,
|
| 1575 |
+
"eval_runtime": 6.9281,
|
| 1576 |
+
"eval_samples_per_second": 57.736,
|
| 1577 |
+
"eval_steps_per_second": 3.608,
|
| 1578 |
+
"step": 2363
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 17.05,
|
| 1582 |
+
"learning_rate": 9.59232613908873e-06,
|
| 1583 |
+
"loss": 0.0168,
|
| 1584 |
+
"step": 2370
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 17.12,
|
| 1588 |
+
"learning_rate": 9.539035438316015e-06,
|
| 1589 |
+
"loss": 0.0048,
|
| 1590 |
+
"step": 2380
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 17.19,
|
| 1594 |
+
"learning_rate": 9.4857447375433e-06,
|
| 1595 |
+
"loss": 0.0071,
|
| 1596 |
+
"step": 2390
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 17.27,
|
| 1600 |
+
"learning_rate": 9.432454036770584e-06,
|
| 1601 |
+
"loss": 0.0,
|
| 1602 |
+
"step": 2400
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 17.34,
|
| 1606 |
+
"learning_rate": 9.379163335997869e-06,
|
| 1607 |
+
"loss": 0.0,
|
| 1608 |
+
"step": 2410
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 17.41,
|
| 1612 |
+
"learning_rate": 9.325872635225154e-06,
|
| 1613 |
+
"loss": 0.0,
|
| 1614 |
+
"step": 2420
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 17.48,
|
| 1618 |
+
"learning_rate": 9.272581934452439e-06,
|
| 1619 |
+
"loss": 0.0,
|
| 1620 |
+
"step": 2430
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 17.55,
|
| 1624 |
+
"learning_rate": 9.219291233679724e-06,
|
| 1625 |
+
"loss": 0.0061,
|
| 1626 |
+
"step": 2440
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 17.63,
|
| 1630 |
+
"learning_rate": 9.16600053290701e-06,
|
| 1631 |
+
"loss": 0.0108,
|
| 1632 |
+
"step": 2450
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 17.7,
|
| 1636 |
+
"learning_rate": 9.112709832134293e-06,
|
| 1637 |
+
"loss": 0.014,
|
| 1638 |
+
"step": 2460
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 17.77,
|
| 1642 |
+
"learning_rate": 9.059419131361578e-06,
|
| 1643 |
+
"loss": 0.003,
|
| 1644 |
+
"step": 2470
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 17.84,
|
| 1648 |
+
"learning_rate": 9.006128430588863e-06,
|
| 1649 |
+
"loss": 0.0273,
|
| 1650 |
+
"step": 2480
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 17.91,
|
| 1654 |
+
"learning_rate": 8.952837729816147e-06,
|
| 1655 |
+
"loss": 0.0187,
|
| 1656 |
+
"step": 2490
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 17.99,
|
| 1660 |
+
"learning_rate": 8.899547029043432e-06,
|
| 1661 |
+
"loss": 0.0055,
|
| 1662 |
+
"step": 2500
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 18.0,
|
| 1666 |
+
"eval_accuracy": 0.985,
|
| 1667 |
+
"eval_loss": 0.2575609087944031,
|
| 1668 |
+
"eval_runtime": 7.0331,
|
| 1669 |
+
"eval_samples_per_second": 56.874,
|
| 1670 |
+
"eval_steps_per_second": 3.555,
|
| 1671 |
+
"step": 2502
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 18.06,
|
| 1675 |
+
"learning_rate": 8.846256328270717e-06,
|
| 1676 |
+
"loss": 0.0108,
|
| 1677 |
+
"step": 2510
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 18.13,
|
| 1681 |
+
"learning_rate": 8.792965627498002e-06,
|
| 1682 |
+
"loss": 0.034,
|
| 1683 |
+
"step": 2520
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 18.2,
|
| 1687 |
+
"learning_rate": 8.739674926725287e-06,
|
| 1688 |
+
"loss": 0.0001,
|
| 1689 |
+
"step": 2530
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 18.27,
|
| 1693 |
+
"learning_rate": 8.686384225952573e-06,
|
| 1694 |
+
"loss": 0.0035,
|
| 1695 |
+
"step": 2540
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 18.35,
|
| 1699 |
+
"learning_rate": 8.633093525179856e-06,
|
| 1700 |
+
"loss": 0.0278,
|
| 1701 |
+
"step": 2550
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 18.42,
|
| 1705 |
+
"learning_rate": 8.579802824407141e-06,
|
| 1706 |
+
"loss": 0.0025,
|
| 1707 |
+
"step": 2560
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 18.49,
|
| 1711 |
+
"learning_rate": 8.526512123634426e-06,
|
| 1712 |
+
"loss": 0.0,
|
| 1713 |
+
"step": 2570
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 18.56,
|
| 1717 |
+
"learning_rate": 8.473221422861712e-06,
|
| 1718 |
+
"loss": 0.0199,
|
| 1719 |
+
"step": 2580
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 18.63,
|
| 1723 |
+
"learning_rate": 8.419930722088995e-06,
|
| 1724 |
+
"loss": 0.0,
|
| 1725 |
+
"step": 2590
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 18.71,
|
| 1729 |
+
"learning_rate": 8.366640021316282e-06,
|
| 1730 |
+
"loss": 0.0171,
|
| 1731 |
+
"step": 2600
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 18.78,
|
| 1735 |
+
"learning_rate": 8.313349320543565e-06,
|
| 1736 |
+
"loss": 0.0,
|
| 1737 |
+
"step": 2610
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 18.85,
|
| 1741 |
+
"learning_rate": 8.26005861977085e-06,
|
| 1742 |
+
"loss": 0.0189,
|
| 1743 |
+
"step": 2620
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 18.92,
|
| 1747 |
+
"learning_rate": 8.206767918998136e-06,
|
| 1748 |
+
"loss": 0.0,
|
| 1749 |
+
"step": 2630
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 18.99,
|
| 1753 |
+
"learning_rate": 8.153477218225421e-06,
|
| 1754 |
+
"loss": 0.0,
|
| 1755 |
+
"step": 2640
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 19.0,
|
| 1759 |
+
"eval_accuracy": 0.985,
|
| 1760 |
+
"eval_loss": 0.23741036653518677,
|
| 1761 |
+
"eval_runtime": 6.8944,
|
| 1762 |
+
"eval_samples_per_second": 58.019,
|
| 1763 |
+
"eval_steps_per_second": 3.626,
|
| 1764 |
+
"step": 2641
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 19.06,
|
| 1768 |
+
"learning_rate": 8.100186517452704e-06,
|
| 1769 |
+
"loss": 0.0,
|
| 1770 |
+
"step": 2650
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 19.14,
|
| 1774 |
+
"learning_rate": 8.04689581667999e-06,
|
| 1775 |
+
"loss": 0.0,
|
| 1776 |
+
"step": 2660
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 19.21,
|
| 1780 |
+
"learning_rate": 7.993605115907275e-06,
|
| 1781 |
+
"loss": 0.0,
|
| 1782 |
+
"step": 2670
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 19.28,
|
| 1786 |
+
"learning_rate": 7.94031441513456e-06,
|
| 1787 |
+
"loss": 0.0126,
|
| 1788 |
+
"step": 2680
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 19.35,
|
| 1792 |
+
"learning_rate": 7.887023714361845e-06,
|
| 1793 |
+
"loss": 0.0208,
|
| 1794 |
+
"step": 2690
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 19.42,
|
| 1798 |
+
"learning_rate": 7.83373301358913e-06,
|
| 1799 |
+
"loss": 0.0007,
|
| 1800 |
+
"step": 2700
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 19.5,
|
| 1804 |
+
"learning_rate": 7.780442312816414e-06,
|
| 1805 |
+
"loss": 0.035,
|
| 1806 |
+
"step": 2710
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 19.57,
|
| 1810 |
+
"learning_rate": 7.727151612043699e-06,
|
| 1811 |
+
"loss": 0.0,
|
| 1812 |
+
"step": 2720
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 19.64,
|
| 1816 |
+
"learning_rate": 7.673860911270984e-06,
|
| 1817 |
+
"loss": 0.0,
|
| 1818 |
+
"step": 2730
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 19.71,
|
| 1822 |
+
"learning_rate": 7.6205702104982685e-06,
|
| 1823 |
+
"loss": 0.0,
|
| 1824 |
+
"step": 2740
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 19.78,
|
| 1828 |
+
"learning_rate": 7.567279509725554e-06,
|
| 1829 |
+
"loss": 0.0225,
|
| 1830 |
+
"step": 2750
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 19.86,
|
| 1834 |
+
"learning_rate": 7.513988808952838e-06,
|
| 1835 |
+
"loss": 0.0,
|
| 1836 |
+
"step": 2760
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 19.93,
|
| 1840 |
+
"learning_rate": 7.460698108180123e-06,
|
| 1841 |
+
"loss": 0.0157,
|
| 1842 |
+
"step": 2770
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 20.0,
|
| 1846 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 1847 |
+
"loss": 0.0034,
|
| 1848 |
+
"step": 2780
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 20.0,
|
| 1852 |
+
"eval_accuracy": 0.99,
|
| 1853 |
+
"eval_loss": 0.16036586463451385,
|
| 1854 |
+
"eval_runtime": 6.9057,
|
| 1855 |
+
"eval_samples_per_second": 57.923,
|
| 1856 |
+
"eval_steps_per_second": 3.62,
|
| 1857 |
+
"step": 2780
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 20.07,
|
| 1861 |
+
"learning_rate": 7.354116706634693e-06,
|
| 1862 |
+
"loss": 0.0,
|
| 1863 |
+
"step": 2790
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 20.14,
|
| 1867 |
+
"learning_rate": 7.300826005861977e-06,
|
| 1868 |
+
"loss": 0.0061,
|
| 1869 |
+
"step": 2800
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 20.22,
|
| 1873 |
+
"learning_rate": 7.247535305089263e-06,
|
| 1874 |
+
"loss": 0.0,
|
| 1875 |
+
"step": 2810
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 20.29,
|
| 1879 |
+
"learning_rate": 7.194244604316547e-06,
|
| 1880 |
+
"loss": 0.0,
|
| 1881 |
+
"step": 2820
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 20.36,
|
| 1885 |
+
"learning_rate": 7.1409539035438325e-06,
|
| 1886 |
+
"loss": 0.0303,
|
| 1887 |
+
"step": 2830
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 20.43,
|
| 1891 |
+
"learning_rate": 7.087663202771117e-06,
|
| 1892 |
+
"loss": 0.0,
|
| 1893 |
+
"step": 2840
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 20.5,
|
| 1897 |
+
"learning_rate": 7.034372501998402e-06,
|
| 1898 |
+
"loss": 0.0,
|
| 1899 |
+
"step": 2850
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 20.58,
|
| 1903 |
+
"learning_rate": 6.981081801225686e-06,
|
| 1904 |
+
"loss": 0.0126,
|
| 1905 |
+
"step": 2860
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 20.65,
|
| 1909 |
+
"learning_rate": 6.9277911004529715e-06,
|
| 1910 |
+
"loss": 0.0299,
|
| 1911 |
+
"step": 2870
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 20.72,
|
| 1915 |
+
"learning_rate": 6.874500399680256e-06,
|
| 1916 |
+
"loss": 0.01,
|
| 1917 |
+
"step": 2880
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 20.79,
|
| 1921 |
+
"learning_rate": 6.821209698907542e-06,
|
| 1922 |
+
"loss": 0.0,
|
| 1923 |
+
"step": 2890
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 20.86,
|
| 1927 |
+
"learning_rate": 6.767918998134826e-06,
|
| 1928 |
+
"loss": 0.0025,
|
| 1929 |
+
"step": 2900
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 20.94,
|
| 1933 |
+
"learning_rate": 6.714628297362111e-06,
|
| 1934 |
+
"loss": 0.0303,
|
| 1935 |
+
"step": 2910
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 21.0,
|
| 1939 |
+
"eval_accuracy": 0.9875,
|
| 1940 |
+
"eval_loss": 0.14314711093902588,
|
| 1941 |
+
"eval_runtime": 7.0921,
|
| 1942 |
+
"eval_samples_per_second": 56.401,
|
| 1943 |
+
"eval_steps_per_second": 3.525,
|
| 1944 |
+
"step": 2919
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 21.01,
|
| 1948 |
+
"learning_rate": 6.661337596589396e-06,
|
| 1949 |
+
"loss": 0.0203,
|
| 1950 |
+
"step": 2920
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 21.08,
|
| 1954 |
+
"learning_rate": 6.608046895816681e-06,
|
| 1955 |
+
"loss": 0.0,
|
| 1956 |
+
"step": 2930
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 21.15,
|
| 1960 |
+
"learning_rate": 6.554756195043965e-06,
|
| 1961 |
+
"loss": 0.0,
|
| 1962 |
+
"step": 2940
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 21.22,
|
| 1966 |
+
"learning_rate": 6.50146549427125e-06,
|
| 1967 |
+
"loss": 0.0186,
|
| 1968 |
+
"step": 2950
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 21.29,
|
| 1972 |
+
"learning_rate": 6.448174793498535e-06,
|
| 1973 |
+
"loss": 0.0071,
|
| 1974 |
+
"step": 2960
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 21.37,
|
| 1978 |
+
"learning_rate": 6.394884092725819e-06,
|
| 1979 |
+
"loss": 0.0005,
|
| 1980 |
+
"step": 2970
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 21.44,
|
| 1984 |
+
"learning_rate": 6.341593391953105e-06,
|
| 1985 |
+
"loss": 0.0055,
|
| 1986 |
+
"step": 2980
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 21.51,
|
| 1990 |
+
"learning_rate": 6.288302691180389e-06,
|
| 1991 |
+
"loss": 0.0,
|
| 1992 |
+
"step": 2990
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 21.58,
|
| 1996 |
+
"learning_rate": 6.2350119904076745e-06,
|
| 1997 |
+
"loss": 0.0331,
|
| 1998 |
+
"step": 3000
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 21.65,
|
| 2002 |
+
"learning_rate": 6.181721289634959e-06,
|
| 2003 |
+
"loss": 0.0,
|
| 2004 |
+
"step": 3010
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 21.73,
|
| 2008 |
+
"learning_rate": 6.128430588862244e-06,
|
| 2009 |
+
"loss": 0.002,
|
| 2010 |
+
"step": 3020
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 21.8,
|
| 2014 |
+
"learning_rate": 6.075139888089528e-06,
|
| 2015 |
+
"loss": 0.0132,
|
| 2016 |
+
"step": 3030
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 21.87,
|
| 2020 |
+
"learning_rate": 6.0218491873168135e-06,
|
| 2021 |
+
"loss": 0.0,
|
| 2022 |
+
"step": 3040
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 21.94,
|
| 2026 |
+
"learning_rate": 5.968558486544098e-06,
|
| 2027 |
+
"loss": 0.0,
|
| 2028 |
+
"step": 3050
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 22.0,
|
| 2032 |
+
"eval_accuracy": 0.9925,
|
| 2033 |
+
"eval_loss": 0.14593777060508728,
|
| 2034 |
+
"eval_runtime": 6.9363,
|
| 2035 |
+
"eval_samples_per_second": 57.668,
|
| 2036 |
+
"eval_steps_per_second": 3.604,
|
| 2037 |
+
"step": 3058
|
| 2038 |
+
}
|
| 2039 |
+
],
|
| 2040 |
+
"logging_steps": 10,
|
| 2041 |
+
"max_steps": 4170,
|
| 2042 |
+
"num_train_epochs": 30,
|
| 2043 |
+
"save_steps": 500,
|
| 2044 |
+
"total_flos": 1.7055354463120835e+19,
|
| 2045 |
+
"trial_name": null,
|
| 2046 |
+
"trial_params": null
|
| 2047 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:91122b47221a17b58d9411bdd2b40f3a8cb92f9202d2f778c36c7c26005d1d7f
|
| 3 |
+
size 4027
|