Upload 8 files
Browse files- config.json +96 -0
- optimizer.pt +3 -0
- preprocessor_config.json +9 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1750 -0
- training_args.bin +3 -0
config.json
ADDED
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "ntu-spml/distilhubert",
|
3 |
+
"activation_dropout": 0.1,
|
4 |
+
"apply_spec_augment": false,
|
5 |
+
"architectures": [
|
6 |
+
"HubertForSequenceClassification"
|
7 |
+
],
|
8 |
+
"attention_dropout": 0.1,
|
9 |
+
"bos_token_id": 1,
|
10 |
+
"classifier_proj_size": 256,
|
11 |
+
"conv_bias": false,
|
12 |
+
"conv_dim": [
|
13 |
+
512,
|
14 |
+
512,
|
15 |
+
512,
|
16 |
+
512,
|
17 |
+
512,
|
18 |
+
512,
|
19 |
+
512
|
20 |
+
],
|
21 |
+
"conv_kernel": [
|
22 |
+
10,
|
23 |
+
3,
|
24 |
+
3,
|
25 |
+
3,
|
26 |
+
3,
|
27 |
+
2,
|
28 |
+
2
|
29 |
+
],
|
30 |
+
"conv_stride": [
|
31 |
+
5,
|
32 |
+
2,
|
33 |
+
2,
|
34 |
+
2,
|
35 |
+
2,
|
36 |
+
2,
|
37 |
+
2
|
38 |
+
],
|
39 |
+
"ctc_loss_reduction": "sum",
|
40 |
+
"ctc_zero_infinity": false,
|
41 |
+
"do_stable_layer_norm": false,
|
42 |
+
"eos_token_id": 2,
|
43 |
+
"feat_extract_activation": "gelu",
|
44 |
+
"feat_extract_norm": "group",
|
45 |
+
"feat_proj_dropout": 0.0,
|
46 |
+
"feat_proj_layer_norm": false,
|
47 |
+
"final_dropout": 0.0,
|
48 |
+
"hidden_act": "gelu",
|
49 |
+
"hidden_dropout": 0.1,
|
50 |
+
"hidden_size": 768,
|
51 |
+
"id2label": {
|
52 |
+
"0": "blues",
|
53 |
+
"1": "classical",
|
54 |
+
"2": "country",
|
55 |
+
"3": "disco",
|
56 |
+
"4": "hiphop",
|
57 |
+
"5": "jazz",
|
58 |
+
"6": "metal",
|
59 |
+
"7": "pop",
|
60 |
+
"8": "reggae",
|
61 |
+
"9": "rock"
|
62 |
+
},
|
63 |
+
"initializer_range": 0.02,
|
64 |
+
"intermediate_size": 3072,
|
65 |
+
"label2id": {
|
66 |
+
"blues": "0",
|
67 |
+
"classical": "1",
|
68 |
+
"country": "2",
|
69 |
+
"disco": "3",
|
70 |
+
"hiphop": "4",
|
71 |
+
"jazz": "5",
|
72 |
+
"metal": "6",
|
73 |
+
"pop": "7",
|
74 |
+
"reggae": "8",
|
75 |
+
"rock": "9"
|
76 |
+
},
|
77 |
+
"layer_norm_eps": 1e-05,
|
78 |
+
"layerdrop": 0.0,
|
79 |
+
"mask_feature_length": 10,
|
80 |
+
"mask_feature_min_masks": 0,
|
81 |
+
"mask_feature_prob": 0.0,
|
82 |
+
"mask_time_length": 10,
|
83 |
+
"mask_time_min_masks": 2,
|
84 |
+
"mask_time_prob": 0.05,
|
85 |
+
"model_type": "hubert",
|
86 |
+
"num_attention_heads": 12,
|
87 |
+
"num_conv_pos_embedding_groups": 16,
|
88 |
+
"num_conv_pos_embeddings": 128,
|
89 |
+
"num_feat_extract_layers": 7,
|
90 |
+
"num_hidden_layers": 2,
|
91 |
+
"pad_token_id": 0,
|
92 |
+
"torch_dtype": "float32",
|
93 |
+
"transformers_version": "4.31.0.dev0",
|
94 |
+
"use_weighted_layer_sum": false,
|
95 |
+
"vocab_size": 32
|
96 |
+
}
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd5def78fcf35ce8197a0a89b2cddd0d0ab2a7e652f880a0f267e8082c15dcc1
|
3 |
+
size 189555589
|
preprocessor_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000
|
9 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:213bef9ea1cb8c00dc924977331618ea4ae9823f2d2b164e92ee138a1a71b505
|
3 |
+
size 94783376
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:98fe4e41e502c92480c84ccee907b3ce52b631e614a8ed347276406cc29ef887
|
3 |
+
size 14575
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:27fb51ce21283dcc0c99fed0d74cb955b323f5a3c1836121af7be1af641d3763
|
3 |
+
size 627
|
trainer_state.json
ADDED
@@ -0,0 +1,1750 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.87,
|
3 |
+
"best_model_checkpoint": "distilhubert-finetuned-gtzan/checkpoint-1356",
|
4 |
+
"epoch": 12.0,
|
5 |
+
"global_step": 1356,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.04,
|
12 |
+
"learning_rate": 1.1061946902654867e-06,
|
13 |
+
"loss": 2.3091,
|
14 |
+
"step": 5
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.09,
|
18 |
+
"learning_rate": 2.2123893805309734e-06,
|
19 |
+
"loss": 2.3013,
|
20 |
+
"step": 10
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.13,
|
24 |
+
"learning_rate": 3.3185840707964607e-06,
|
25 |
+
"loss": 2.304,
|
26 |
+
"step": 15
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.18,
|
30 |
+
"learning_rate": 4.424778761061947e-06,
|
31 |
+
"loss": 2.307,
|
32 |
+
"step": 20
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.22,
|
36 |
+
"learning_rate": 5.5309734513274346e-06,
|
37 |
+
"loss": 2.3035,
|
38 |
+
"step": 25
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.27,
|
42 |
+
"learning_rate": 6.6371681415929215e-06,
|
43 |
+
"loss": 2.2975,
|
44 |
+
"step": 30
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.31,
|
48 |
+
"learning_rate": 7.743362831858407e-06,
|
49 |
+
"loss": 2.2995,
|
50 |
+
"step": 35
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.35,
|
54 |
+
"learning_rate": 8.849557522123894e-06,
|
55 |
+
"loss": 2.2945,
|
56 |
+
"step": 40
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.4,
|
60 |
+
"learning_rate": 9.95575221238938e-06,
|
61 |
+
"loss": 2.2868,
|
62 |
+
"step": 45
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.44,
|
66 |
+
"learning_rate": 1.1061946902654869e-05,
|
67 |
+
"loss": 2.2764,
|
68 |
+
"step": 50
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.49,
|
72 |
+
"learning_rate": 1.2168141592920354e-05,
|
73 |
+
"loss": 2.2802,
|
74 |
+
"step": 55
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.53,
|
78 |
+
"learning_rate": 1.3274336283185843e-05,
|
79 |
+
"loss": 2.2647,
|
80 |
+
"step": 60
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.58,
|
84 |
+
"learning_rate": 1.4380530973451328e-05,
|
85 |
+
"loss": 2.2792,
|
86 |
+
"step": 65
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.62,
|
90 |
+
"learning_rate": 1.5486725663716813e-05,
|
91 |
+
"loss": 2.2499,
|
92 |
+
"step": 70
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.66,
|
96 |
+
"learning_rate": 1.6592920353982302e-05,
|
97 |
+
"loss": 2.2809,
|
98 |
+
"step": 75
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.71,
|
102 |
+
"learning_rate": 1.7699115044247787e-05,
|
103 |
+
"loss": 2.2683,
|
104 |
+
"step": 80
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.75,
|
108 |
+
"learning_rate": 1.8805309734513272e-05,
|
109 |
+
"loss": 2.2492,
|
110 |
+
"step": 85
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.8,
|
114 |
+
"learning_rate": 1.991150442477876e-05,
|
115 |
+
"loss": 2.239,
|
116 |
+
"step": 90
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.84,
|
120 |
+
"learning_rate": 2.101769911504425e-05,
|
121 |
+
"loss": 2.2239,
|
122 |
+
"step": 95
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.88,
|
126 |
+
"learning_rate": 2.2123893805309738e-05,
|
127 |
+
"loss": 2.1822,
|
128 |
+
"step": 100
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.93,
|
132 |
+
"learning_rate": 2.3230088495575223e-05,
|
133 |
+
"loss": 2.1901,
|
134 |
+
"step": 105
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 0.97,
|
138 |
+
"learning_rate": 2.433628318584071e-05,
|
139 |
+
"loss": 2.2018,
|
140 |
+
"step": 110
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 1.0,
|
144 |
+
"eval_accuracy": 0.37,
|
145 |
+
"eval_loss": 2.124462842941284,
|
146 |
+
"eval_runtime": 37.2852,
|
147 |
+
"eval_samples_per_second": 2.682,
|
148 |
+
"eval_steps_per_second": 0.349,
|
149 |
+
"step": 113
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 1.02,
|
153 |
+
"learning_rate": 2.5442477876106197e-05,
|
154 |
+
"loss": 2.1462,
|
155 |
+
"step": 115
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 1.06,
|
159 |
+
"learning_rate": 2.6548672566371686e-05,
|
160 |
+
"loss": 2.1083,
|
161 |
+
"step": 120
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 1.11,
|
165 |
+
"learning_rate": 2.7654867256637168e-05,
|
166 |
+
"loss": 2.0828,
|
167 |
+
"step": 125
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 1.15,
|
171 |
+
"learning_rate": 2.8761061946902656e-05,
|
172 |
+
"loss": 2.0594,
|
173 |
+
"step": 130
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 1.19,
|
177 |
+
"learning_rate": 2.9867256637168145e-05,
|
178 |
+
"loss": 2.0813,
|
179 |
+
"step": 135
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 1.24,
|
183 |
+
"learning_rate": 3.097345132743363e-05,
|
184 |
+
"loss": 2.0167,
|
185 |
+
"step": 140
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 1.28,
|
189 |
+
"learning_rate": 3.185840707964602e-05,
|
190 |
+
"loss": 1.9729,
|
191 |
+
"step": 145
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 1.33,
|
195 |
+
"learning_rate": 3.296460176991151e-05,
|
196 |
+
"loss": 1.9886,
|
197 |
+
"step": 150
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 1.37,
|
201 |
+
"learning_rate": 3.407079646017699e-05,
|
202 |
+
"loss": 1.9951,
|
203 |
+
"step": 155
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 1.42,
|
207 |
+
"learning_rate": 3.517699115044248e-05,
|
208 |
+
"loss": 1.9228,
|
209 |
+
"step": 160
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 1.46,
|
213 |
+
"learning_rate": 3.628318584070797e-05,
|
214 |
+
"loss": 1.9038,
|
215 |
+
"step": 165
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 1.5,
|
219 |
+
"learning_rate": 3.7389380530973455e-05,
|
220 |
+
"loss": 1.857,
|
221 |
+
"step": 170
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 1.55,
|
225 |
+
"learning_rate": 3.849557522123894e-05,
|
226 |
+
"loss": 1.922,
|
227 |
+
"step": 175
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 1.59,
|
231 |
+
"learning_rate": 3.9601769911504426e-05,
|
232 |
+
"loss": 1.8138,
|
233 |
+
"step": 180
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.64,
|
237 |
+
"learning_rate": 4.0707964601769914e-05,
|
238 |
+
"loss": 1.7358,
|
239 |
+
"step": 185
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.68,
|
243 |
+
"learning_rate": 4.1814159292035396e-05,
|
244 |
+
"loss": 1.6802,
|
245 |
+
"step": 190
|
246 |
+
},
|
247 |
+
{
|
248 |
+
"epoch": 1.73,
|
249 |
+
"learning_rate": 4.2920353982300885e-05,
|
250 |
+
"loss": 1.6438,
|
251 |
+
"step": 195
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 1.77,
|
255 |
+
"learning_rate": 4.4026548672566373e-05,
|
256 |
+
"loss": 1.67,
|
257 |
+
"step": 200
|
258 |
+
},
|
259 |
+
{
|
260 |
+
"epoch": 1.81,
|
261 |
+
"learning_rate": 4.491150442477876e-05,
|
262 |
+
"loss": 1.8038,
|
263 |
+
"step": 205
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 1.86,
|
267 |
+
"learning_rate": 4.579646017699115e-05,
|
268 |
+
"loss": 1.7205,
|
269 |
+
"step": 210
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 1.9,
|
273 |
+
"learning_rate": 4.690265486725664e-05,
|
274 |
+
"loss": 1.7188,
|
275 |
+
"step": 215
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.95,
|
279 |
+
"learning_rate": 4.800884955752213e-05,
|
280 |
+
"loss": 1.5285,
|
281 |
+
"step": 220
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 1.99,
|
285 |
+
"learning_rate": 4.911504424778761e-05,
|
286 |
+
"loss": 1.5528,
|
287 |
+
"step": 225
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 2.0,
|
291 |
+
"eval_accuracy": 0.56,
|
292 |
+
"eval_loss": 1.5200048685073853,
|
293 |
+
"eval_runtime": 36.9619,
|
294 |
+
"eval_samples_per_second": 2.705,
|
295 |
+
"eval_steps_per_second": 0.352,
|
296 |
+
"step": 226
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 2.04,
|
300 |
+
"learning_rate": 4.997541789577188e-05,
|
301 |
+
"loss": 1.5776,
|
302 |
+
"step": 230
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.08,
|
306 |
+
"learning_rate": 4.985250737463127e-05,
|
307 |
+
"loss": 1.5497,
|
308 |
+
"step": 235
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 2.12,
|
312 |
+
"learning_rate": 4.972959685349066e-05,
|
313 |
+
"loss": 1.4853,
|
314 |
+
"step": 240
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.17,
|
318 |
+
"learning_rate": 4.960668633235005e-05,
|
319 |
+
"loss": 1.5263,
|
320 |
+
"step": 245
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.21,
|
324 |
+
"learning_rate": 4.9483775811209444e-05,
|
325 |
+
"loss": 1.4766,
|
326 |
+
"step": 250
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 2.26,
|
330 |
+
"learning_rate": 4.936086529006883e-05,
|
331 |
+
"loss": 1.4554,
|
332 |
+
"step": 255
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 2.3,
|
336 |
+
"learning_rate": 4.923795476892822e-05,
|
337 |
+
"loss": 1.3716,
|
338 |
+
"step": 260
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.35,
|
342 |
+
"learning_rate": 4.911504424778761e-05,
|
343 |
+
"loss": 1.3669,
|
344 |
+
"step": 265
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 2.39,
|
348 |
+
"learning_rate": 4.8992133726647006e-05,
|
349 |
+
"loss": 1.4348,
|
350 |
+
"step": 270
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 2.43,
|
354 |
+
"learning_rate": 4.8869223205506394e-05,
|
355 |
+
"loss": 1.275,
|
356 |
+
"step": 275
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 2.48,
|
360 |
+
"learning_rate": 4.874631268436578e-05,
|
361 |
+
"loss": 1.4206,
|
362 |
+
"step": 280
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 2.52,
|
366 |
+
"learning_rate": 4.862340216322517e-05,
|
367 |
+
"loss": 1.4962,
|
368 |
+
"step": 285
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 2.57,
|
372 |
+
"learning_rate": 4.850049164208457e-05,
|
373 |
+
"loss": 1.2934,
|
374 |
+
"step": 290
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 2.61,
|
378 |
+
"learning_rate": 4.8377581120943956e-05,
|
379 |
+
"loss": 1.263,
|
380 |
+
"step": 295
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 2.65,
|
384 |
+
"learning_rate": 4.825467059980335e-05,
|
385 |
+
"loss": 1.2592,
|
386 |
+
"step": 300
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 2.7,
|
390 |
+
"learning_rate": 4.813176007866273e-05,
|
391 |
+
"loss": 1.1302,
|
392 |
+
"step": 305
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 2.74,
|
396 |
+
"learning_rate": 4.800884955752213e-05,
|
397 |
+
"loss": 1.242,
|
398 |
+
"step": 310
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 2.79,
|
402 |
+
"learning_rate": 4.788593903638152e-05,
|
403 |
+
"loss": 1.2154,
|
404 |
+
"step": 315
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 2.83,
|
408 |
+
"learning_rate": 4.776302851524091e-05,
|
409 |
+
"loss": 1.1873,
|
410 |
+
"step": 320
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 2.88,
|
414 |
+
"learning_rate": 4.7640117994100294e-05,
|
415 |
+
"loss": 1.2688,
|
416 |
+
"step": 325
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 2.92,
|
420 |
+
"learning_rate": 4.751720747295968e-05,
|
421 |
+
"loss": 1.1488,
|
422 |
+
"step": 330
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 2.96,
|
426 |
+
"learning_rate": 4.739429695181908e-05,
|
427 |
+
"loss": 1.2506,
|
428 |
+
"step": 335
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 3.0,
|
432 |
+
"eval_accuracy": 0.71,
|
433 |
+
"eval_loss": 1.1110644340515137,
|
434 |
+
"eval_runtime": 37.2461,
|
435 |
+
"eval_samples_per_second": 2.685,
|
436 |
+
"eval_steps_per_second": 0.349,
|
437 |
+
"step": 339
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 3.01,
|
441 |
+
"learning_rate": 4.727138643067847e-05,
|
442 |
+
"loss": 1.2124,
|
443 |
+
"step": 340
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 3.05,
|
447 |
+
"learning_rate": 4.714847590953786e-05,
|
448 |
+
"loss": 1.1278,
|
449 |
+
"step": 345
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 3.1,
|
453 |
+
"learning_rate": 4.7025565388397244e-05,
|
454 |
+
"loss": 1.1197,
|
455 |
+
"step": 350
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 3.14,
|
459 |
+
"learning_rate": 4.690265486725664e-05,
|
460 |
+
"loss": 0.9486,
|
461 |
+
"step": 355
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 3.19,
|
465 |
+
"learning_rate": 4.677974434611603e-05,
|
466 |
+
"loss": 1.0949,
|
467 |
+
"step": 360
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 3.23,
|
471 |
+
"learning_rate": 4.6656833824975424e-05,
|
472 |
+
"loss": 1.1837,
|
473 |
+
"step": 365
|
474 |
+
},
|
475 |
+
{
|
476 |
+
"epoch": 3.27,
|
477 |
+
"learning_rate": 4.655850540806293e-05,
|
478 |
+
"loss": 1.0059,
|
479 |
+
"step": 370
|
480 |
+
},
|
481 |
+
{
|
482 |
+
"epoch": 3.32,
|
483 |
+
"learning_rate": 4.643559488692232e-05,
|
484 |
+
"loss": 1.0395,
|
485 |
+
"step": 375
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 3.36,
|
489 |
+
"learning_rate": 4.631268436578171e-05,
|
490 |
+
"loss": 1.1674,
|
491 |
+
"step": 380
|
492 |
+
},
|
493 |
+
{
|
494 |
+
"epoch": 3.41,
|
495 |
+
"learning_rate": 4.6189773844641105e-05,
|
496 |
+
"loss": 0.9608,
|
497 |
+
"step": 385
|
498 |
+
},
|
499 |
+
{
|
500 |
+
"epoch": 3.45,
|
501 |
+
"learning_rate": 4.606686332350049e-05,
|
502 |
+
"loss": 1.2654,
|
503 |
+
"step": 390
|
504 |
+
},
|
505 |
+
{
|
506 |
+
"epoch": 3.5,
|
507 |
+
"learning_rate": 4.594395280235988e-05,
|
508 |
+
"loss": 0.8788,
|
509 |
+
"step": 395
|
510 |
+
},
|
511 |
+
{
|
512 |
+
"epoch": 3.54,
|
513 |
+
"learning_rate": 4.582104228121927e-05,
|
514 |
+
"loss": 0.9846,
|
515 |
+
"step": 400
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 3.58,
|
519 |
+
"learning_rate": 4.5698131760078666e-05,
|
520 |
+
"loss": 0.8597,
|
521 |
+
"step": 405
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 3.63,
|
525 |
+
"learning_rate": 4.5575221238938055e-05,
|
526 |
+
"loss": 1.1485,
|
527 |
+
"step": 410
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 3.67,
|
531 |
+
"learning_rate": 4.545231071779745e-05,
|
532 |
+
"loss": 1.0469,
|
533 |
+
"step": 415
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 3.72,
|
537 |
+
"learning_rate": 4.532940019665683e-05,
|
538 |
+
"loss": 0.9581,
|
539 |
+
"step": 420
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 3.76,
|
543 |
+
"learning_rate": 4.520648967551623e-05,
|
544 |
+
"loss": 0.9821,
|
545 |
+
"step": 425
|
546 |
+
},
|
547 |
+
{
|
548 |
+
"epoch": 3.81,
|
549 |
+
"learning_rate": 4.5083579154375616e-05,
|
550 |
+
"loss": 1.0122,
|
551 |
+
"step": 430
|
552 |
+
},
|
553 |
+
{
|
554 |
+
"epoch": 3.85,
|
555 |
+
"learning_rate": 4.496066863323501e-05,
|
556 |
+
"loss": 0.8543,
|
557 |
+
"step": 435
|
558 |
+
},
|
559 |
+
{
|
560 |
+
"epoch": 3.89,
|
561 |
+
"learning_rate": 4.48377581120944e-05,
|
562 |
+
"loss": 0.8246,
|
563 |
+
"step": 440
|
564 |
+
},
|
565 |
+
{
|
566 |
+
"epoch": 3.94,
|
567 |
+
"learning_rate": 4.471484759095379e-05,
|
568 |
+
"loss": 0.891,
|
569 |
+
"step": 445
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 3.98,
|
573 |
+
"learning_rate": 4.459193706981318e-05,
|
574 |
+
"loss": 1.0517,
|
575 |
+
"step": 450
|
576 |
+
},
|
577 |
+
{
|
578 |
+
"epoch": 4.0,
|
579 |
+
"eval_accuracy": 0.74,
|
580 |
+
"eval_loss": 0.950592041015625,
|
581 |
+
"eval_runtime": 37.0565,
|
582 |
+
"eval_samples_per_second": 2.699,
|
583 |
+
"eval_steps_per_second": 0.351,
|
584 |
+
"step": 452
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 4.03,
|
588 |
+
"learning_rate": 4.446902654867257e-05,
|
589 |
+
"loss": 0.7031,
|
590 |
+
"step": 455
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 4.07,
|
594 |
+
"learning_rate": 4.434611602753196e-05,
|
595 |
+
"loss": 0.6235,
|
596 |
+
"step": 460
|
597 |
+
},
|
598 |
+
{
|
599 |
+
"epoch": 4.12,
|
600 |
+
"learning_rate": 4.422320550639135e-05,
|
601 |
+
"loss": 0.8002,
|
602 |
+
"step": 465
|
603 |
+
},
|
604 |
+
{
|
605 |
+
"epoch": 4.16,
|
606 |
+
"learning_rate": 4.410029498525074e-05,
|
607 |
+
"loss": 0.9738,
|
608 |
+
"step": 470
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 4.2,
|
612 |
+
"learning_rate": 4.397738446411013e-05,
|
613 |
+
"loss": 0.8511,
|
614 |
+
"step": 475
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 4.25,
|
618 |
+
"learning_rate": 4.385447394296952e-05,
|
619 |
+
"loss": 0.6335,
|
620 |
+
"step": 480
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 4.29,
|
624 |
+
"learning_rate": 4.373156342182891e-05,
|
625 |
+
"loss": 0.644,
|
626 |
+
"step": 485
|
627 |
+
},
|
628 |
+
{
|
629 |
+
"epoch": 4.34,
|
630 |
+
"learning_rate": 4.36086529006883e-05,
|
631 |
+
"loss": 0.8481,
|
632 |
+
"step": 490
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 4.38,
|
636 |
+
"learning_rate": 4.348574237954769e-05,
|
637 |
+
"loss": 0.6972,
|
638 |
+
"step": 495
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 4.42,
|
642 |
+
"learning_rate": 4.3362831858407084e-05,
|
643 |
+
"loss": 0.8044,
|
644 |
+
"step": 500
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 4.47,
|
648 |
+
"learning_rate": 4.323992133726647e-05,
|
649 |
+
"loss": 0.7073,
|
650 |
+
"step": 505
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 4.51,
|
654 |
+
"learning_rate": 4.311701081612586e-05,
|
655 |
+
"loss": 0.8473,
|
656 |
+
"step": 510
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 4.56,
|
660 |
+
"learning_rate": 4.301868239921337e-05,
|
661 |
+
"loss": 0.7983,
|
662 |
+
"step": 515
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 4.6,
|
666 |
+
"learning_rate": 4.2895771878072765e-05,
|
667 |
+
"loss": 0.5702,
|
668 |
+
"step": 520
|
669 |
+
},
|
670 |
+
{
|
671 |
+
"epoch": 4.65,
|
672 |
+
"learning_rate": 4.2772861356932154e-05,
|
673 |
+
"loss": 0.649,
|
674 |
+
"step": 525
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 4.69,
|
678 |
+
"learning_rate": 4.264995083579155e-05,
|
679 |
+
"loss": 0.6669,
|
680 |
+
"step": 530
|
681 |
+
},
|
682 |
+
{
|
683 |
+
"epoch": 4.73,
|
684 |
+
"learning_rate": 4.252704031465093e-05,
|
685 |
+
"loss": 0.5829,
|
686 |
+
"step": 535
|
687 |
+
},
|
688 |
+
{
|
689 |
+
"epoch": 4.78,
|
690 |
+
"learning_rate": 4.2404129793510327e-05,
|
691 |
+
"loss": 0.65,
|
692 |
+
"step": 540
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 4.82,
|
696 |
+
"learning_rate": 4.2281219272369715e-05,
|
697 |
+
"loss": 0.81,
|
698 |
+
"step": 545
|
699 |
+
},
|
700 |
+
{
|
701 |
+
"epoch": 4.87,
|
702 |
+
"learning_rate": 4.215830875122911e-05,
|
703 |
+
"loss": 0.7507,
|
704 |
+
"step": 550
|
705 |
+
},
|
706 |
+
{
|
707 |
+
"epoch": 4.91,
|
708 |
+
"learning_rate": 4.20353982300885e-05,
|
709 |
+
"loss": 0.6655,
|
710 |
+
"step": 555
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 4.96,
|
714 |
+
"learning_rate": 4.191248770894789e-05,
|
715 |
+
"loss": 0.6181,
|
716 |
+
"step": 560
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 5.0,
|
720 |
+
"learning_rate": 4.178957718780728e-05,
|
721 |
+
"loss": 0.8272,
|
722 |
+
"step": 565
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 5.0,
|
726 |
+
"eval_accuracy": 0.83,
|
727 |
+
"eval_loss": 0.7394164800643921,
|
728 |
+
"eval_runtime": 37.0651,
|
729 |
+
"eval_samples_per_second": 2.698,
|
730 |
+
"eval_steps_per_second": 0.351,
|
731 |
+
"step": 565
|
732 |
+
},
|
733 |
+
{
|
734 |
+
"epoch": 5.04,
|
735 |
+
"learning_rate": 4.166666666666667e-05,
|
736 |
+
"loss": 0.6379,
|
737 |
+
"step": 570
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 5.09,
|
741 |
+
"learning_rate": 4.154375614552606e-05,
|
742 |
+
"loss": 0.5593,
|
743 |
+
"step": 575
|
744 |
+
},
|
745 |
+
{
|
746 |
+
"epoch": 5.13,
|
747 |
+
"learning_rate": 4.142084562438545e-05,
|
748 |
+
"loss": 0.524,
|
749 |
+
"step": 580
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 5.18,
|
753 |
+
"learning_rate": 4.129793510324484e-05,
|
754 |
+
"loss": 0.4926,
|
755 |
+
"step": 585
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 5.22,
|
759 |
+
"learning_rate": 4.1175024582104233e-05,
|
760 |
+
"loss": 0.6712,
|
761 |
+
"step": 590
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 5.27,
|
765 |
+
"learning_rate": 4.105211406096362e-05,
|
766 |
+
"loss": 0.59,
|
767 |
+
"step": 595
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 5.31,
|
771 |
+
"learning_rate": 4.092920353982301e-05,
|
772 |
+
"loss": 0.3684,
|
773 |
+
"step": 600
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 5.35,
|
777 |
+
"learning_rate": 4.08062930186824e-05,
|
778 |
+
"loss": 0.5484,
|
779 |
+
"step": 605
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 5.4,
|
783 |
+
"learning_rate": 4.068338249754179e-05,
|
784 |
+
"loss": 0.4718,
|
785 |
+
"step": 610
|
786 |
+
},
|
787 |
+
{
|
788 |
+
"epoch": 5.44,
|
789 |
+
"learning_rate": 4.0560471976401183e-05,
|
790 |
+
"loss": 0.5205,
|
791 |
+
"step": 615
|
792 |
+
},
|
793 |
+
{
|
794 |
+
"epoch": 5.49,
|
795 |
+
"learning_rate": 4.043756145526057e-05,
|
796 |
+
"loss": 0.4103,
|
797 |
+
"step": 620
|
798 |
+
},
|
799 |
+
{
|
800 |
+
"epoch": 5.53,
|
801 |
+
"learning_rate": 4.031465093411996e-05,
|
802 |
+
"loss": 0.6238,
|
803 |
+
"step": 625
|
804 |
+
},
|
805 |
+
{
|
806 |
+
"epoch": 5.58,
|
807 |
+
"learning_rate": 4.019174041297935e-05,
|
808 |
+
"loss": 0.391,
|
809 |
+
"step": 630
|
810 |
+
},
|
811 |
+
{
|
812 |
+
"epoch": 5.62,
|
813 |
+
"learning_rate": 4.0068829891838745e-05,
|
814 |
+
"loss": 0.626,
|
815 |
+
"step": 635
|
816 |
+
},
|
817 |
+
{
|
818 |
+
"epoch": 5.66,
|
819 |
+
"learning_rate": 3.9945919370698133e-05,
|
820 |
+
"loss": 0.5142,
|
821 |
+
"step": 640
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 5.71,
|
825 |
+
"learning_rate": 3.982300884955752e-05,
|
826 |
+
"loss": 0.5664,
|
827 |
+
"step": 645
|
828 |
+
},
|
829 |
+
{
|
830 |
+
"epoch": 5.75,
|
831 |
+
"learning_rate": 3.970009832841691e-05,
|
832 |
+
"loss": 0.3242,
|
833 |
+
"step": 650
|
834 |
+
},
|
835 |
+
{
|
836 |
+
"epoch": 5.8,
|
837 |
+
"learning_rate": 3.9577187807276306e-05,
|
838 |
+
"loss": 0.4262,
|
839 |
+
"step": 655
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 5.84,
|
843 |
+
"learning_rate": 3.9454277286135695e-05,
|
844 |
+
"loss": 0.6169,
|
845 |
+
"step": 660
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 5.88,
|
849 |
+
"learning_rate": 3.9331366764995083e-05,
|
850 |
+
"loss": 0.436,
|
851 |
+
"step": 665
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 5.93,
|
855 |
+
"learning_rate": 3.920845624385447e-05,
|
856 |
+
"loss": 0.2636,
|
857 |
+
"step": 670
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 5.97,
|
861 |
+
"learning_rate": 3.908554572271387e-05,
|
862 |
+
"loss": 0.3728,
|
863 |
+
"step": 675
|
864 |
+
},
|
865 |
+
{
|
866 |
+
"epoch": 6.0,
|
867 |
+
"eval_accuracy": 0.81,
|
868 |
+
"eval_loss": 0.6849124431610107,
|
869 |
+
"eval_runtime": 37.0876,
|
870 |
+
"eval_samples_per_second": 2.696,
|
871 |
+
"eval_steps_per_second": 0.351,
|
872 |
+
"step": 678
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 6.02,
|
876 |
+
"learning_rate": 3.8962635201573256e-05,
|
877 |
+
"loss": 0.4045,
|
878 |
+
"step": 680
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 6.06,
|
882 |
+
"learning_rate": 3.883972468043265e-05,
|
883 |
+
"loss": 0.2084,
|
884 |
+
"step": 685
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 6.11,
|
888 |
+
"learning_rate": 3.8716814159292034e-05,
|
889 |
+
"loss": 0.3604,
|
890 |
+
"step": 690
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 6.15,
|
894 |
+
"learning_rate": 3.859390363815143e-05,
|
895 |
+
"loss": 0.3663,
|
896 |
+
"step": 695
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 6.19,
|
900 |
+
"learning_rate": 3.847099311701082e-05,
|
901 |
+
"loss": 0.3161,
|
902 |
+
"step": 700
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 6.24,
|
906 |
+
"learning_rate": 3.834808259587021e-05,
|
907 |
+
"loss": 0.4456,
|
908 |
+
"step": 705
|
909 |
+
},
|
910 |
+
{
|
911 |
+
"epoch": 6.28,
|
912 |
+
"learning_rate": 3.82251720747296e-05,
|
913 |
+
"loss": 0.2935,
|
914 |
+
"step": 710
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 6.33,
|
918 |
+
"learning_rate": 3.810226155358899e-05,
|
919 |
+
"loss": 0.2577,
|
920 |
+
"step": 715
|
921 |
+
},
|
922 |
+
{
|
923 |
+
"epoch": 6.37,
|
924 |
+
"learning_rate": 3.797935103244838e-05,
|
925 |
+
"loss": 0.2841,
|
926 |
+
"step": 720
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 6.42,
|
930 |
+
"learning_rate": 3.7856440511307774e-05,
|
931 |
+
"loss": 0.2892,
|
932 |
+
"step": 725
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 6.46,
|
936 |
+
"learning_rate": 3.773352999016716e-05,
|
937 |
+
"loss": 0.2578,
|
938 |
+
"step": 730
|
939 |
+
},
|
940 |
+
{
|
941 |
+
"epoch": 6.5,
|
942 |
+
"learning_rate": 3.7610619469026545e-05,
|
943 |
+
"loss": 0.2938,
|
944 |
+
"step": 735
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 6.55,
|
948 |
+
"learning_rate": 3.748770894788594e-05,
|
949 |
+
"loss": 0.2936,
|
950 |
+
"step": 740
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 6.59,
|
954 |
+
"learning_rate": 3.736479842674533e-05,
|
955 |
+
"loss": 0.426,
|
956 |
+
"step": 745
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 6.64,
|
960 |
+
"learning_rate": 3.7241887905604724e-05,
|
961 |
+
"loss": 0.2328,
|
962 |
+
"step": 750
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 6.68,
|
966 |
+
"learning_rate": 3.711897738446411e-05,
|
967 |
+
"loss": 0.2715,
|
968 |
+
"step": 755
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 6.73,
|
972 |
+
"learning_rate": 3.69960668633235e-05,
|
973 |
+
"loss": 0.3068,
|
974 |
+
"step": 760
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 6.77,
|
978 |
+
"learning_rate": 3.687315634218289e-05,
|
979 |
+
"loss": 0.4009,
|
980 |
+
"step": 765
|
981 |
+
},
|
982 |
+
{
|
983 |
+
"epoch": 6.81,
|
984 |
+
"learning_rate": 3.6750245821042286e-05,
|
985 |
+
"loss": 0.3685,
|
986 |
+
"step": 770
|
987 |
+
},
|
988 |
+
{
|
989 |
+
"epoch": 6.86,
|
990 |
+
"learning_rate": 3.6627335299901674e-05,
|
991 |
+
"loss": 0.3065,
|
992 |
+
"step": 775
|
993 |
+
},
|
994 |
+
{
|
995 |
+
"epoch": 6.9,
|
996 |
+
"learning_rate": 3.650442477876106e-05,
|
997 |
+
"loss": 0.2677,
|
998 |
+
"step": 780
|
999 |
+
},
|
1000 |
+
{
|
1001 |
+
"epoch": 6.95,
|
1002 |
+
"learning_rate": 3.638151425762045e-05,
|
1003 |
+
"loss": 0.3096,
|
1004 |
+
"step": 785
|
1005 |
+
},
|
1006 |
+
{
|
1007 |
+
"epoch": 6.99,
|
1008 |
+
"learning_rate": 3.625860373647985e-05,
|
1009 |
+
"loss": 0.3878,
|
1010 |
+
"step": 790
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 7.0,
|
1014 |
+
"eval_accuracy": 0.8,
|
1015 |
+
"eval_loss": 0.647432267665863,
|
1016 |
+
"eval_runtime": 36.7252,
|
1017 |
+
"eval_samples_per_second": 2.723,
|
1018 |
+
"eval_steps_per_second": 0.354,
|
1019 |
+
"step": 791
|
1020 |
+
},
|
1021 |
+
{
|
1022 |
+
"epoch": 7.04,
|
1023 |
+
"learning_rate": 3.6135693215339236e-05,
|
1024 |
+
"loss": 0.1663,
|
1025 |
+
"step": 795
|
1026 |
+
},
|
1027 |
+
{
|
1028 |
+
"epoch": 7.08,
|
1029 |
+
"learning_rate": 3.6012782694198624e-05,
|
1030 |
+
"loss": 0.2787,
|
1031 |
+
"step": 800
|
1032 |
+
},
|
1033 |
+
{
|
1034 |
+
"epoch": 7.12,
|
1035 |
+
"learning_rate": 3.588987217305801e-05,
|
1036 |
+
"loss": 0.1559,
|
1037 |
+
"step": 805
|
1038 |
+
},
|
1039 |
+
{
|
1040 |
+
"epoch": 7.17,
|
1041 |
+
"learning_rate": 3.576696165191741e-05,
|
1042 |
+
"loss": 0.1331,
|
1043 |
+
"step": 810
|
1044 |
+
},
|
1045 |
+
{
|
1046 |
+
"epoch": 7.21,
|
1047 |
+
"learning_rate": 3.56440511307768e-05,
|
1048 |
+
"loss": 0.1643,
|
1049 |
+
"step": 815
|
1050 |
+
},
|
1051 |
+
{
|
1052 |
+
"epoch": 7.26,
|
1053 |
+
"learning_rate": 3.5521140609636186e-05,
|
1054 |
+
"loss": 0.1594,
|
1055 |
+
"step": 820
|
1056 |
+
},
|
1057 |
+
{
|
1058 |
+
"epoch": 7.3,
|
1059 |
+
"learning_rate": 3.54228121927237e-05,
|
1060 |
+
"loss": 0.1871,
|
1061 |
+
"step": 825
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 7.35,
|
1065 |
+
"learning_rate": 3.529990167158309e-05,
|
1066 |
+
"loss": 0.1745,
|
1067 |
+
"step": 830
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 7.39,
|
1071 |
+
"learning_rate": 3.517699115044248e-05,
|
1072 |
+
"loss": 0.125,
|
1073 |
+
"step": 835
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 7.43,
|
1077 |
+
"learning_rate": 3.5054080629301874e-05,
|
1078 |
+
"loss": 0.1259,
|
1079 |
+
"step": 840
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 7.48,
|
1083 |
+
"learning_rate": 3.493117010816126e-05,
|
1084 |
+
"loss": 0.2476,
|
1085 |
+
"step": 845
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 7.52,
|
1089 |
+
"learning_rate": 3.480825958702065e-05,
|
1090 |
+
"loss": 0.0858,
|
1091 |
+
"step": 850
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 7.57,
|
1095 |
+
"learning_rate": 3.468534906588004e-05,
|
1096 |
+
"loss": 0.1988,
|
1097 |
+
"step": 855
|
1098 |
+
},
|
1099 |
+
{
|
1100 |
+
"epoch": 7.61,
|
1101 |
+
"learning_rate": 3.4562438544739435e-05,
|
1102 |
+
"loss": 0.1398,
|
1103 |
+
"step": 860
|
1104 |
+
},
|
1105 |
+
{
|
1106 |
+
"epoch": 7.65,
|
1107 |
+
"learning_rate": 3.4439528023598824e-05,
|
1108 |
+
"loss": 0.1934,
|
1109 |
+
"step": 865
|
1110 |
+
},
|
1111 |
+
{
|
1112 |
+
"epoch": 7.7,
|
1113 |
+
"learning_rate": 3.431661750245821e-05,
|
1114 |
+
"loss": 0.0942,
|
1115 |
+
"step": 870
|
1116 |
+
},
|
1117 |
+
{
|
1118 |
+
"epoch": 7.74,
|
1119 |
+
"learning_rate": 3.41937069813176e-05,
|
1120 |
+
"loss": 0.1991,
|
1121 |
+
"step": 875
|
1122 |
+
},
|
1123 |
+
{
|
1124 |
+
"epoch": 7.79,
|
1125 |
+
"learning_rate": 3.407079646017699e-05,
|
1126 |
+
"loss": 0.1436,
|
1127 |
+
"step": 880
|
1128 |
+
},
|
1129 |
+
{
|
1130 |
+
"epoch": 7.83,
|
1131 |
+
"learning_rate": 3.3947885939036385e-05,
|
1132 |
+
"loss": 0.0865,
|
1133 |
+
"step": 885
|
1134 |
+
},
|
1135 |
+
{
|
1136 |
+
"epoch": 7.88,
|
1137 |
+
"learning_rate": 3.3824975417895774e-05,
|
1138 |
+
"loss": 0.1405,
|
1139 |
+
"step": 890
|
1140 |
+
},
|
1141 |
+
{
|
1142 |
+
"epoch": 7.92,
|
1143 |
+
"learning_rate": 3.370206489675516e-05,
|
1144 |
+
"loss": 0.285,
|
1145 |
+
"step": 895
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 7.96,
|
1149 |
+
"learning_rate": 3.357915437561455e-05,
|
1150 |
+
"loss": 0.1119,
|
1151 |
+
"step": 900
|
1152 |
+
},
|
1153 |
+
{
|
1154 |
+
"epoch": 8.0,
|
1155 |
+
"eval_accuracy": 0.83,
|
1156 |
+
"eval_loss": 0.5846229791641235,
|
1157 |
+
"eval_runtime": 37.7892,
|
1158 |
+
"eval_samples_per_second": 2.646,
|
1159 |
+
"eval_steps_per_second": 0.344,
|
1160 |
+
"step": 904
|
1161 |
+
},
|
1162 |
+
{
|
1163 |
+
"epoch": 8.01,
|
1164 |
+
"learning_rate": 3.3456243854473946e-05,
|
1165 |
+
"loss": 0.1696,
|
1166 |
+
"step": 905
|
1167 |
+
},
|
1168 |
+
{
|
1169 |
+
"epoch": 8.05,
|
1170 |
+
"learning_rate": 3.3333333333333335e-05,
|
1171 |
+
"loss": 0.1067,
|
1172 |
+
"step": 910
|
1173 |
+
},
|
1174 |
+
{
|
1175 |
+
"epoch": 8.1,
|
1176 |
+
"learning_rate": 3.3210422812192724e-05,
|
1177 |
+
"loss": 0.0852,
|
1178 |
+
"step": 915
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 8.14,
|
1182 |
+
"learning_rate": 3.308751229105211e-05,
|
1183 |
+
"loss": 0.1015,
|
1184 |
+
"step": 920
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 8.19,
|
1188 |
+
"learning_rate": 3.296460176991151e-05,
|
1189 |
+
"loss": 0.1699,
|
1190 |
+
"step": 925
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 8.23,
|
1194 |
+
"learning_rate": 3.2841691248770896e-05,
|
1195 |
+
"loss": 0.0691,
|
1196 |
+
"step": 930
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 8.27,
|
1200 |
+
"learning_rate": 3.2718780727630285e-05,
|
1201 |
+
"loss": 0.0661,
|
1202 |
+
"step": 935
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 8.32,
|
1206 |
+
"learning_rate": 3.2595870206489674e-05,
|
1207 |
+
"loss": 0.0512,
|
1208 |
+
"step": 940
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 8.36,
|
1212 |
+
"learning_rate": 3.247295968534907e-05,
|
1213 |
+
"loss": 0.1003,
|
1214 |
+
"step": 945
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 8.41,
|
1218 |
+
"learning_rate": 3.235004916420846e-05,
|
1219 |
+
"loss": 0.0523,
|
1220 |
+
"step": 950
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 8.45,
|
1224 |
+
"learning_rate": 3.222713864306785e-05,
|
1225 |
+
"loss": 0.0893,
|
1226 |
+
"step": 955
|
1227 |
+
},
|
1228 |
+
{
|
1229 |
+
"epoch": 8.5,
|
1230 |
+
"learning_rate": 3.2104228121927235e-05,
|
1231 |
+
"loss": 0.053,
|
1232 |
+
"step": 960
|
1233 |
+
},
|
1234 |
+
{
|
1235 |
+
"epoch": 8.54,
|
1236 |
+
"learning_rate": 3.198131760078663e-05,
|
1237 |
+
"loss": 0.0592,
|
1238 |
+
"step": 965
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 8.58,
|
1242 |
+
"learning_rate": 3.185840707964602e-05,
|
1243 |
+
"loss": 0.0779,
|
1244 |
+
"step": 970
|
1245 |
+
},
|
1246 |
+
{
|
1247 |
+
"epoch": 8.63,
|
1248 |
+
"learning_rate": 3.1735496558505414e-05,
|
1249 |
+
"loss": 0.0611,
|
1250 |
+
"step": 975
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 8.67,
|
1254 |
+
"learning_rate": 3.16125860373648e-05,
|
1255 |
+
"loss": 0.0776,
|
1256 |
+
"step": 980
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 8.72,
|
1260 |
+
"learning_rate": 3.148967551622419e-05,
|
1261 |
+
"loss": 0.1722,
|
1262 |
+
"step": 985
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 8.76,
|
1266 |
+
"learning_rate": 3.136676499508358e-05,
|
1267 |
+
"loss": 0.0809,
|
1268 |
+
"step": 990
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 8.81,
|
1272 |
+
"learning_rate": 3.124385447394297e-05,
|
1273 |
+
"loss": 0.185,
|
1274 |
+
"step": 995
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 8.85,
|
1278 |
+
"learning_rate": 3.1120943952802364e-05,
|
1279 |
+
"loss": 0.0357,
|
1280 |
+
"step": 1000
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 8.89,
|
1284 |
+
"learning_rate": 3.0998033431661746e-05,
|
1285 |
+
"loss": 0.082,
|
1286 |
+
"step": 1005
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 8.94,
|
1290 |
+
"learning_rate": 3.087512291052114e-05,
|
1291 |
+
"loss": 0.1928,
|
1292 |
+
"step": 1010
|
1293 |
+
},
|
1294 |
+
{
|
1295 |
+
"epoch": 8.98,
|
1296 |
+
"learning_rate": 3.075221238938053e-05,
|
1297 |
+
"loss": 0.1312,
|
1298 |
+
"step": 1015
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 9.0,
|
1302 |
+
"eval_accuracy": 0.84,
|
1303 |
+
"eval_loss": 0.5652107000350952,
|
1304 |
+
"eval_runtime": 36.9689,
|
1305 |
+
"eval_samples_per_second": 2.705,
|
1306 |
+
"eval_steps_per_second": 0.352,
|
1307 |
+
"step": 1017
|
1308 |
+
},
|
1309 |
+
{
|
1310 |
+
"epoch": 9.03,
|
1311 |
+
"learning_rate": 3.0629301868239926e-05,
|
1312 |
+
"loss": 0.0416,
|
1313 |
+
"step": 1020
|
1314 |
+
},
|
1315 |
+
{
|
1316 |
+
"epoch": 9.07,
|
1317 |
+
"learning_rate": 3.050639134709931e-05,
|
1318 |
+
"loss": 0.1309,
|
1319 |
+
"step": 1025
|
1320 |
+
},
|
1321 |
+
{
|
1322 |
+
"epoch": 9.12,
|
1323 |
+
"learning_rate": 3.0383480825958703e-05,
|
1324 |
+
"loss": 0.1201,
|
1325 |
+
"step": 1030
|
1326 |
+
},
|
1327 |
+
{
|
1328 |
+
"epoch": 9.16,
|
1329 |
+
"learning_rate": 3.0260570304818092e-05,
|
1330 |
+
"loss": 0.0265,
|
1331 |
+
"step": 1035
|
1332 |
+
},
|
1333 |
+
{
|
1334 |
+
"epoch": 9.2,
|
1335 |
+
"learning_rate": 3.0137659783677484e-05,
|
1336 |
+
"loss": 0.0291,
|
1337 |
+
"step": 1040
|
1338 |
+
},
|
1339 |
+
{
|
1340 |
+
"epoch": 9.25,
|
1341 |
+
"learning_rate": 3.0014749262536872e-05,
|
1342 |
+
"loss": 0.0306,
|
1343 |
+
"step": 1045
|
1344 |
+
},
|
1345 |
+
{
|
1346 |
+
"epoch": 9.29,
|
1347 |
+
"learning_rate": 2.9891838741396268e-05,
|
1348 |
+
"loss": 0.0262,
|
1349 |
+
"step": 1050
|
1350 |
+
},
|
1351 |
+
{
|
1352 |
+
"epoch": 9.34,
|
1353 |
+
"learning_rate": 2.9768928220255653e-05,
|
1354 |
+
"loss": 0.0262,
|
1355 |
+
"step": 1055
|
1356 |
+
},
|
1357 |
+
{
|
1358 |
+
"epoch": 9.38,
|
1359 |
+
"learning_rate": 2.964601769911505e-05,
|
1360 |
+
"loss": 0.0243,
|
1361 |
+
"step": 1060
|
1362 |
+
},
|
1363 |
+
{
|
1364 |
+
"epoch": 9.42,
|
1365 |
+
"learning_rate": 2.9523107177974434e-05,
|
1366 |
+
"loss": 0.0314,
|
1367 |
+
"step": 1065
|
1368 |
+
},
|
1369 |
+
{
|
1370 |
+
"epoch": 9.47,
|
1371 |
+
"learning_rate": 2.940019665683383e-05,
|
1372 |
+
"loss": 0.0218,
|
1373 |
+
"step": 1070
|
1374 |
+
},
|
1375 |
+
{
|
1376 |
+
"epoch": 9.51,
|
1377 |
+
"learning_rate": 2.9277286135693215e-05,
|
1378 |
+
"loss": 0.0294,
|
1379 |
+
"step": 1075
|
1380 |
+
},
|
1381 |
+
{
|
1382 |
+
"epoch": 9.56,
|
1383 |
+
"learning_rate": 2.915437561455261e-05,
|
1384 |
+
"loss": 0.0187,
|
1385 |
+
"step": 1080
|
1386 |
+
},
|
1387 |
+
{
|
1388 |
+
"epoch": 9.6,
|
1389 |
+
"learning_rate": 2.9031465093412e-05,
|
1390 |
+
"loss": 0.0614,
|
1391 |
+
"step": 1085
|
1392 |
+
},
|
1393 |
+
{
|
1394 |
+
"epoch": 9.65,
|
1395 |
+
"learning_rate": 2.890855457227139e-05,
|
1396 |
+
"loss": 0.0269,
|
1397 |
+
"step": 1090
|
1398 |
+
},
|
1399 |
+
{
|
1400 |
+
"epoch": 9.69,
|
1401 |
+
"learning_rate": 2.878564405113078e-05,
|
1402 |
+
"loss": 0.0591,
|
1403 |
+
"step": 1095
|
1404 |
+
},
|
1405 |
+
{
|
1406 |
+
"epoch": 9.73,
|
1407 |
+
"learning_rate": 2.866273352999017e-05,
|
1408 |
+
"loss": 0.0282,
|
1409 |
+
"step": 1100
|
1410 |
+
},
|
1411 |
+
{
|
1412 |
+
"epoch": 9.78,
|
1413 |
+
"learning_rate": 2.853982300884956e-05,
|
1414 |
+
"loss": 0.0242,
|
1415 |
+
"step": 1105
|
1416 |
+
},
|
1417 |
+
{
|
1418 |
+
"epoch": 9.82,
|
1419 |
+
"learning_rate": 2.8416912487708945e-05,
|
1420 |
+
"loss": 0.0709,
|
1421 |
+
"step": 1110
|
1422 |
+
},
|
1423 |
+
{
|
1424 |
+
"epoch": 9.87,
|
1425 |
+
"learning_rate": 2.829400196656834e-05,
|
1426 |
+
"loss": 0.0158,
|
1427 |
+
"step": 1115
|
1428 |
+
},
|
1429 |
+
{
|
1430 |
+
"epoch": 9.91,
|
1431 |
+
"learning_rate": 2.8171091445427726e-05,
|
1432 |
+
"loss": 0.1044,
|
1433 |
+
"step": 1120
|
1434 |
+
},
|
1435 |
+
{
|
1436 |
+
"epoch": 9.96,
|
1437 |
+
"learning_rate": 2.804818092428712e-05,
|
1438 |
+
"loss": 0.0404,
|
1439 |
+
"step": 1125
|
1440 |
+
},
|
1441 |
+
{
|
1442 |
+
"epoch": 10.0,
|
1443 |
+
"learning_rate": 2.792527040314651e-05,
|
1444 |
+
"loss": 0.0149,
|
1445 |
+
"step": 1130
|
1446 |
+
},
|
1447 |
+
{
|
1448 |
+
"epoch": 10.0,
|
1449 |
+
"eval_accuracy": 0.84,
|
1450 |
+
"eval_loss": 0.707106351852417,
|
1451 |
+
"eval_runtime": 37.5782,
|
1452 |
+
"eval_samples_per_second": 2.661,
|
1453 |
+
"eval_steps_per_second": 0.346,
|
1454 |
+
"step": 1130
|
1455 |
+
},
|
1456 |
+
{
|
1457 |
+
"epoch": 10.04,
|
1458 |
+
"learning_rate": 2.7802359882005902e-05,
|
1459 |
+
"loss": 0.1165,
|
1460 |
+
"step": 1135
|
1461 |
+
},
|
1462 |
+
{
|
1463 |
+
"epoch": 10.09,
|
1464 |
+
"learning_rate": 2.767944936086529e-05,
|
1465 |
+
"loss": 0.0129,
|
1466 |
+
"step": 1140
|
1467 |
+
},
|
1468 |
+
{
|
1469 |
+
"epoch": 10.13,
|
1470 |
+
"learning_rate": 2.7556538839724683e-05,
|
1471 |
+
"loss": 0.0624,
|
1472 |
+
"step": 1145
|
1473 |
+
},
|
1474 |
+
{
|
1475 |
+
"epoch": 10.18,
|
1476 |
+
"learning_rate": 2.743362831858407e-05,
|
1477 |
+
"loss": 0.0272,
|
1478 |
+
"step": 1150
|
1479 |
+
},
|
1480 |
+
{
|
1481 |
+
"epoch": 10.22,
|
1482 |
+
"learning_rate": 2.7310717797443463e-05,
|
1483 |
+
"loss": 0.0176,
|
1484 |
+
"step": 1155
|
1485 |
+
},
|
1486 |
+
{
|
1487 |
+
"epoch": 10.27,
|
1488 |
+
"learning_rate": 2.7187807276302852e-05,
|
1489 |
+
"loss": 0.0138,
|
1490 |
+
"step": 1160
|
1491 |
+
},
|
1492 |
+
{
|
1493 |
+
"epoch": 10.31,
|
1494 |
+
"learning_rate": 2.7064896755162244e-05,
|
1495 |
+
"loss": 0.0138,
|
1496 |
+
"step": 1165
|
1497 |
+
},
|
1498 |
+
{
|
1499 |
+
"epoch": 10.35,
|
1500 |
+
"learning_rate": 2.6941986234021633e-05,
|
1501 |
+
"loss": 0.0177,
|
1502 |
+
"step": 1170
|
1503 |
+
},
|
1504 |
+
{
|
1505 |
+
"epoch": 10.4,
|
1506 |
+
"learning_rate": 2.6819075712881025e-05,
|
1507 |
+
"loss": 0.024,
|
1508 |
+
"step": 1175
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 10.44,
|
1512 |
+
"learning_rate": 2.6696165191740413e-05,
|
1513 |
+
"loss": 0.0676,
|
1514 |
+
"step": 1180
|
1515 |
+
},
|
1516 |
+
{
|
1517 |
+
"epoch": 10.49,
|
1518 |
+
"learning_rate": 2.6573254670599805e-05,
|
1519 |
+
"loss": 0.0112,
|
1520 |
+
"step": 1185
|
1521 |
+
},
|
1522 |
+
{
|
1523 |
+
"epoch": 10.53,
|
1524 |
+
"learning_rate": 2.6450344149459194e-05,
|
1525 |
+
"loss": 0.011,
|
1526 |
+
"step": 1190
|
1527 |
+
},
|
1528 |
+
{
|
1529 |
+
"epoch": 10.58,
|
1530 |
+
"learning_rate": 2.6327433628318586e-05,
|
1531 |
+
"loss": 0.0118,
|
1532 |
+
"step": 1195
|
1533 |
+
},
|
1534 |
+
{
|
1535 |
+
"epoch": 10.62,
|
1536 |
+
"learning_rate": 2.6204523107177975e-05,
|
1537 |
+
"loss": 0.018,
|
1538 |
+
"step": 1200
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 10.66,
|
1542 |
+
"learning_rate": 2.6081612586037367e-05,
|
1543 |
+
"loss": 0.017,
|
1544 |
+
"step": 1205
|
1545 |
+
},
|
1546 |
+
{
|
1547 |
+
"epoch": 10.71,
|
1548 |
+
"learning_rate": 2.5958702064896756e-05,
|
1549 |
+
"loss": 0.0117,
|
1550 |
+
"step": 1210
|
1551 |
+
},
|
1552 |
+
{
|
1553 |
+
"epoch": 10.75,
|
1554 |
+
"learning_rate": 2.583579154375615e-05,
|
1555 |
+
"loss": 0.01,
|
1556 |
+
"step": 1215
|
1557 |
+
},
|
1558 |
+
{
|
1559 |
+
"epoch": 10.8,
|
1560 |
+
"learning_rate": 2.5712881022615536e-05,
|
1561 |
+
"loss": 0.0101,
|
1562 |
+
"step": 1220
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 10.84,
|
1566 |
+
"learning_rate": 2.558997050147493e-05,
|
1567 |
+
"loss": 0.017,
|
1568 |
+
"step": 1225
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 10.88,
|
1572 |
+
"learning_rate": 2.5467059980334317e-05,
|
1573 |
+
"loss": 0.0094,
|
1574 |
+
"step": 1230
|
1575 |
+
},
|
1576 |
+
{
|
1577 |
+
"epoch": 10.93,
|
1578 |
+
"learning_rate": 2.5344149459193706e-05,
|
1579 |
+
"loss": 0.0085,
|
1580 |
+
"step": 1235
|
1581 |
+
},
|
1582 |
+
{
|
1583 |
+
"epoch": 10.97,
|
1584 |
+
"learning_rate": 2.5221238938053098e-05,
|
1585 |
+
"loss": 0.0154,
|
1586 |
+
"step": 1240
|
1587 |
+
},
|
1588 |
+
{
|
1589 |
+
"epoch": 11.0,
|
1590 |
+
"eval_accuracy": 0.84,
|
1591 |
+
"eval_loss": 0.7669206261634827,
|
1592 |
+
"eval_runtime": 37.5505,
|
1593 |
+
"eval_samples_per_second": 2.663,
|
1594 |
+
"eval_steps_per_second": 0.346,
|
1595 |
+
"step": 1243
|
1596 |
+
},
|
1597 |
+
{
|
1598 |
+
"epoch": 11.02,
|
1599 |
+
"learning_rate": 2.5098328416912486e-05,
|
1600 |
+
"loss": 0.0085,
|
1601 |
+
"step": 1245
|
1602 |
+
},
|
1603 |
+
{
|
1604 |
+
"epoch": 11.06,
|
1605 |
+
"learning_rate": 2.4975417895771878e-05,
|
1606 |
+
"loss": 0.0091,
|
1607 |
+
"step": 1250
|
1608 |
+
},
|
1609 |
+
{
|
1610 |
+
"epoch": 11.11,
|
1611 |
+
"learning_rate": 2.485250737463127e-05,
|
1612 |
+
"loss": 0.0086,
|
1613 |
+
"step": 1255
|
1614 |
+
},
|
1615 |
+
{
|
1616 |
+
"epoch": 11.15,
|
1617 |
+
"learning_rate": 2.4729596853490662e-05,
|
1618 |
+
"loss": 0.0087,
|
1619 |
+
"step": 1260
|
1620 |
+
},
|
1621 |
+
{
|
1622 |
+
"epoch": 11.19,
|
1623 |
+
"learning_rate": 2.460668633235005e-05,
|
1624 |
+
"loss": 0.008,
|
1625 |
+
"step": 1265
|
1626 |
+
},
|
1627 |
+
{
|
1628 |
+
"epoch": 11.24,
|
1629 |
+
"learning_rate": 2.4483775811209443e-05,
|
1630 |
+
"loss": 0.0078,
|
1631 |
+
"step": 1270
|
1632 |
+
},
|
1633 |
+
{
|
1634 |
+
"epoch": 11.28,
|
1635 |
+
"learning_rate": 2.436086529006883e-05,
|
1636 |
+
"loss": 0.0076,
|
1637 |
+
"step": 1275
|
1638 |
+
},
|
1639 |
+
{
|
1640 |
+
"epoch": 11.33,
|
1641 |
+
"learning_rate": 2.4237954768928224e-05,
|
1642 |
+
"loss": 0.0074,
|
1643 |
+
"step": 1280
|
1644 |
+
},
|
1645 |
+
{
|
1646 |
+
"epoch": 11.37,
|
1647 |
+
"learning_rate": 2.411504424778761e-05,
|
1648 |
+
"loss": 0.0076,
|
1649 |
+
"step": 1285
|
1650 |
+
},
|
1651 |
+
{
|
1652 |
+
"epoch": 11.42,
|
1653 |
+
"learning_rate": 2.3992133726647e-05,
|
1654 |
+
"loss": 0.0071,
|
1655 |
+
"step": 1290
|
1656 |
+
},
|
1657 |
+
{
|
1658 |
+
"epoch": 11.46,
|
1659 |
+
"learning_rate": 2.3869223205506393e-05,
|
1660 |
+
"loss": 0.0078,
|
1661 |
+
"step": 1295
|
1662 |
+
},
|
1663 |
+
{
|
1664 |
+
"epoch": 11.5,
|
1665 |
+
"learning_rate": 2.374631268436578e-05,
|
1666 |
+
"loss": 0.0069,
|
1667 |
+
"step": 1300
|
1668 |
+
},
|
1669 |
+
{
|
1670 |
+
"epoch": 11.55,
|
1671 |
+
"learning_rate": 2.3623402163225174e-05,
|
1672 |
+
"loss": 0.0071,
|
1673 |
+
"step": 1305
|
1674 |
+
},
|
1675 |
+
{
|
1676 |
+
"epoch": 11.59,
|
1677 |
+
"learning_rate": 2.3500491642084562e-05,
|
1678 |
+
"loss": 0.0086,
|
1679 |
+
"step": 1310
|
1680 |
+
},
|
1681 |
+
{
|
1682 |
+
"epoch": 11.64,
|
1683 |
+
"learning_rate": 2.3377581120943954e-05,
|
1684 |
+
"loss": 0.0075,
|
1685 |
+
"step": 1315
|
1686 |
+
},
|
1687 |
+
{
|
1688 |
+
"epoch": 11.68,
|
1689 |
+
"learning_rate": 2.3254670599803343e-05,
|
1690 |
+
"loss": 0.0072,
|
1691 |
+
"step": 1320
|
1692 |
+
},
|
1693 |
+
{
|
1694 |
+
"epoch": 11.73,
|
1695 |
+
"learning_rate": 2.3131760078662735e-05,
|
1696 |
+
"loss": 0.0078,
|
1697 |
+
"step": 1325
|
1698 |
+
},
|
1699 |
+
{
|
1700 |
+
"epoch": 11.77,
|
1701 |
+
"learning_rate": 2.3008849557522124e-05,
|
1702 |
+
"loss": 0.0398,
|
1703 |
+
"step": 1330
|
1704 |
+
},
|
1705 |
+
{
|
1706 |
+
"epoch": 11.81,
|
1707 |
+
"learning_rate": 2.2885939036381516e-05,
|
1708 |
+
"loss": 0.0069,
|
1709 |
+
"step": 1335
|
1710 |
+
},
|
1711 |
+
{
|
1712 |
+
"epoch": 11.86,
|
1713 |
+
"learning_rate": 2.2763028515240904e-05,
|
1714 |
+
"loss": 0.0069,
|
1715 |
+
"step": 1340
|
1716 |
+
},
|
1717 |
+
{
|
1718 |
+
"epoch": 11.9,
|
1719 |
+
"learning_rate": 2.2640117994100296e-05,
|
1720 |
+
"loss": 0.0057,
|
1721 |
+
"step": 1345
|
1722 |
+
},
|
1723 |
+
{
|
1724 |
+
"epoch": 11.95,
|
1725 |
+
"learning_rate": 2.2517207472959685e-05,
|
1726 |
+
"loss": 0.0108,
|
1727 |
+
"step": 1350
|
1728 |
+
},
|
1729 |
+
{
|
1730 |
+
"epoch": 11.99,
|
1731 |
+
"learning_rate": 2.2394296951819077e-05,
|
1732 |
+
"loss": 0.1238,
|
1733 |
+
"step": 1355
|
1734 |
+
},
|
1735 |
+
{
|
1736 |
+
"epoch": 12.0,
|
1737 |
+
"eval_accuracy": 0.87,
|
1738 |
+
"eval_loss": 0.7153857350349426,
|
1739 |
+
"eval_runtime": 37.3772,
|
1740 |
+
"eval_samples_per_second": 2.675,
|
1741 |
+
"eval_steps_per_second": 0.348,
|
1742 |
+
"step": 1356
|
1743 |
+
}
|
1744 |
+
],
|
1745 |
+
"max_steps": 2260,
|
1746 |
+
"num_train_epochs": 20,
|
1747 |
+
"total_flos": 3.6803929647744e+17,
|
1748 |
+
"trial_name": null,
|
1749 |
+
"trial_params": null
|
1750 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5a8354dc326f843935d0365262dd9a2d2c1c9ecddc6b717d50fa49ba59768733
|
3 |
+
size 3963
|