Commit
•
65ff560
1
Parent(s):
d3e8af1
added all set up
Browse files- config.json +63 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- preprocessor_config.json +22 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1419 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "umm-maybe/AI-image-detector",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"SwinForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"depths": [
|
| 8 |
+
2,
|
| 9 |
+
2,
|
| 10 |
+
18,
|
| 11 |
+
2
|
| 12 |
+
],
|
| 13 |
+
"drop_path_rate": 0.1,
|
| 14 |
+
"embed_dim": 128,
|
| 15 |
+
"encoder_stride": 32,
|
| 16 |
+
"hidden_act": "gelu",
|
| 17 |
+
"hidden_dropout_prob": 0.0,
|
| 18 |
+
"hidden_size": 1024,
|
| 19 |
+
"id2label": {
|
| 20 |
+
"0": "ai_gen",
|
| 21 |
+
"1": "human"
|
| 22 |
+
},
|
| 23 |
+
"image_size": 224,
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"label2id": {
|
| 26 |
+
"ai_gen": "0",
|
| 27 |
+
"human": "1"
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-05,
|
| 30 |
+
"max_length": 128,
|
| 31 |
+
"mlp_ratio": 4.0,
|
| 32 |
+
"model_type": "swin",
|
| 33 |
+
"num_channels": 3,
|
| 34 |
+
"num_heads": [
|
| 35 |
+
4,
|
| 36 |
+
8,
|
| 37 |
+
16,
|
| 38 |
+
32
|
| 39 |
+
],
|
| 40 |
+
"num_layers": 4,
|
| 41 |
+
"out_features": [
|
| 42 |
+
"stage4"
|
| 43 |
+
],
|
| 44 |
+
"out_indices": [
|
| 45 |
+
4
|
| 46 |
+
],
|
| 47 |
+
"padding": "max_length",
|
| 48 |
+
"patch_size": 4,
|
| 49 |
+
"path_norm": true,
|
| 50 |
+
"problem_type": "single_label_classification",
|
| 51 |
+
"qkv_bias": true,
|
| 52 |
+
"stage_names": [
|
| 53 |
+
"stem",
|
| 54 |
+
"stage1",
|
| 55 |
+
"stage2",
|
| 56 |
+
"stage3",
|
| 57 |
+
"stage4"
|
| 58 |
+
],
|
| 59 |
+
"torch_dtype": "float32",
|
| 60 |
+
"transformers_version": "4.37.2",
|
| 61 |
+
"use_absolute_embeddings": false,
|
| 62 |
+
"window_size": 7
|
| 63 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58457fe35502bf9d591620e0971895d68bea5fae61544c2b2cae888dc603e822
|
| 3 |
+
size 347498816
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:76ff02edeebd91b23634b24589add183c0e2344206543246194ad85a12a1181f
|
| 3 |
+
size 694317645
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7f06d75ab90b5e2edbd49ee5d6e1f232cde2ddb85c1cff8c9b5f93b6d14c085
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a9bace81c5fda36e81d152c6c83f858f6f523157d1641c36b5ccc654f2c8859
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1419 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 1.0,
|
| 3 |
+
"best_model_checkpoint": "ai_detector_v3/checkpoint-1709",
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2279,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.4619883040935671e-06,
|
| 14 |
+
"loss": 1.3185,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 2.9239766081871343e-06,
|
| 20 |
+
"loss": 0.9483,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 4.3859649122807014e-06,
|
| 26 |
+
"loss": 0.591,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.07,
|
| 31 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 32 |
+
"loss": 0.3762,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.09,
|
| 37 |
+
"learning_rate": 7.3099415204678366e-06,
|
| 38 |
+
"loss": 0.2428,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.11,
|
| 43 |
+
"learning_rate": 8.771929824561403e-06,
|
| 44 |
+
"loss": 0.1249,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.12,
|
| 49 |
+
"learning_rate": 1.023391812865497e-05,
|
| 50 |
+
"loss": 0.0659,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.14,
|
| 55 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 56 |
+
"loss": 0.0381,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.16,
|
| 61 |
+
"learning_rate": 1.3157894736842106e-05,
|
| 62 |
+
"loss": 0.0321,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.18,
|
| 67 |
+
"learning_rate": 1.4619883040935673e-05,
|
| 68 |
+
"loss": 0.0301,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.19,
|
| 73 |
+
"learning_rate": 1.608187134502924e-05,
|
| 74 |
+
"loss": 0.0395,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.21,
|
| 79 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 80 |
+
"loss": 0.0197,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.23,
|
| 85 |
+
"learning_rate": 1.9005847953216373e-05,
|
| 86 |
+
"loss": 0.0226,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.25,
|
| 91 |
+
"learning_rate": 2.046783625730994e-05,
|
| 92 |
+
"loss": 0.0233,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.26,
|
| 97 |
+
"learning_rate": 2.1929824561403507e-05,
|
| 98 |
+
"loss": 0.0098,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.28,
|
| 103 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 104 |
+
"loss": 0.0188,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.3,
|
| 109 |
+
"learning_rate": 2.485380116959064e-05,
|
| 110 |
+
"loss": 0.014,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.32,
|
| 115 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 116 |
+
"loss": 0.0354,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.33,
|
| 121 |
+
"learning_rate": 2.777777777777778e-05,
|
| 122 |
+
"loss": 0.0126,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.35,
|
| 127 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 128 |
+
"loss": 0.0063,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.37,
|
| 133 |
+
"learning_rate": 3.0701754385964913e-05,
|
| 134 |
+
"loss": 0.0325,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.39,
|
| 139 |
+
"learning_rate": 3.216374269005848e-05,
|
| 140 |
+
"loss": 0.0139,
|
| 141 |
+
"step": 220
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.4,
|
| 145 |
+
"learning_rate": 3.362573099415205e-05,
|
| 146 |
+
"loss": 0.0258,
|
| 147 |
+
"step": 230
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.42,
|
| 151 |
+
"learning_rate": 3.508771929824561e-05,
|
| 152 |
+
"loss": 0.032,
|
| 153 |
+
"step": 240
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.44,
|
| 157 |
+
"learning_rate": 3.654970760233918e-05,
|
| 158 |
+
"loss": 0.0188,
|
| 159 |
+
"step": 250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.46,
|
| 163 |
+
"learning_rate": 3.8011695906432746e-05,
|
| 164 |
+
"loss": 0.0005,
|
| 165 |
+
"step": 260
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.47,
|
| 169 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 170 |
+
"loss": 0.0124,
|
| 171 |
+
"step": 270
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.49,
|
| 175 |
+
"learning_rate": 4.093567251461988e-05,
|
| 176 |
+
"loss": 0.0615,
|
| 177 |
+
"step": 280
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.51,
|
| 181 |
+
"learning_rate": 4.239766081871345e-05,
|
| 182 |
+
"loss": 0.0297,
|
| 183 |
+
"step": 290
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.53,
|
| 187 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 188 |
+
"loss": 0.0387,
|
| 189 |
+
"step": 300
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.54,
|
| 193 |
+
"learning_rate": 4.5321637426900585e-05,
|
| 194 |
+
"loss": 0.0151,
|
| 195 |
+
"step": 310
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.56,
|
| 199 |
+
"learning_rate": 4.678362573099415e-05,
|
| 200 |
+
"loss": 0.0329,
|
| 201 |
+
"step": 320
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.58,
|
| 205 |
+
"learning_rate": 4.824561403508772e-05,
|
| 206 |
+
"loss": 0.0155,
|
| 207 |
+
"step": 330
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.6,
|
| 211 |
+
"learning_rate": 4.970760233918128e-05,
|
| 212 |
+
"loss": 0.019,
|
| 213 |
+
"step": 340
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.61,
|
| 217 |
+
"learning_rate": 4.986979166666667e-05,
|
| 218 |
+
"loss": 0.0359,
|
| 219 |
+
"step": 350
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.63,
|
| 223 |
+
"learning_rate": 4.970703125e-05,
|
| 224 |
+
"loss": 0.0117,
|
| 225 |
+
"step": 360
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.65,
|
| 229 |
+
"learning_rate": 4.954427083333333e-05,
|
| 230 |
+
"loss": 0.0169,
|
| 231 |
+
"step": 370
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.67,
|
| 235 |
+
"learning_rate": 4.938151041666667e-05,
|
| 236 |
+
"loss": 0.0084,
|
| 237 |
+
"step": 380
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.68,
|
| 241 |
+
"learning_rate": 4.921875e-05,
|
| 242 |
+
"loss": 0.0178,
|
| 243 |
+
"step": 390
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.7,
|
| 247 |
+
"learning_rate": 4.905598958333333e-05,
|
| 248 |
+
"loss": 0.0063,
|
| 249 |
+
"step": 400
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.72,
|
| 253 |
+
"learning_rate": 4.889322916666667e-05,
|
| 254 |
+
"loss": 0.034,
|
| 255 |
+
"step": 410
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.74,
|
| 259 |
+
"learning_rate": 4.873046875e-05,
|
| 260 |
+
"loss": 0.0216,
|
| 261 |
+
"step": 420
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.75,
|
| 265 |
+
"learning_rate": 4.856770833333333e-05,
|
| 266 |
+
"loss": 0.003,
|
| 267 |
+
"step": 430
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.77,
|
| 271 |
+
"learning_rate": 4.840494791666667e-05,
|
| 272 |
+
"loss": 0.0081,
|
| 273 |
+
"step": 440
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.79,
|
| 277 |
+
"learning_rate": 4.82421875e-05,
|
| 278 |
+
"loss": 0.009,
|
| 279 |
+
"step": 450
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.81,
|
| 283 |
+
"learning_rate": 4.807942708333333e-05,
|
| 284 |
+
"loss": 0.0015,
|
| 285 |
+
"step": 460
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.82,
|
| 289 |
+
"learning_rate": 4.791666666666667e-05,
|
| 290 |
+
"loss": 0.0056,
|
| 291 |
+
"step": 470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.84,
|
| 295 |
+
"learning_rate": 4.775390625e-05,
|
| 296 |
+
"loss": 0.0179,
|
| 297 |
+
"step": 480
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.86,
|
| 301 |
+
"learning_rate": 4.759114583333334e-05,
|
| 302 |
+
"loss": 0.0051,
|
| 303 |
+
"step": 490
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.88,
|
| 307 |
+
"learning_rate": 4.742838541666667e-05,
|
| 308 |
+
"loss": 0.0005,
|
| 309 |
+
"step": 500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.9,
|
| 313 |
+
"learning_rate": 4.7265625000000005e-05,
|
| 314 |
+
"loss": 0.0244,
|
| 315 |
+
"step": 510
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.91,
|
| 319 |
+
"learning_rate": 4.7102864583333336e-05,
|
| 320 |
+
"loss": 0.0008,
|
| 321 |
+
"step": 520
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.93,
|
| 325 |
+
"learning_rate": 4.6940104166666674e-05,
|
| 326 |
+
"loss": 0.0074,
|
| 327 |
+
"step": 530
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.95,
|
| 331 |
+
"learning_rate": 4.6777343750000005e-05,
|
| 332 |
+
"loss": 0.0083,
|
| 333 |
+
"step": 540
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.97,
|
| 337 |
+
"learning_rate": 4.6614583333333336e-05,
|
| 338 |
+
"loss": 0.0139,
|
| 339 |
+
"step": 550
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.98,
|
| 343 |
+
"learning_rate": 4.6451822916666674e-05,
|
| 344 |
+
"loss": 0.0243,
|
| 345 |
+
"step": 560
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0,
|
| 349 |
+
"eval_accuracy": 0.9985,
|
| 350 |
+
"eval_loss": 0.00362112233415246,
|
| 351 |
+
"eval_runtime": 132.255,
|
| 352 |
+
"eval_samples_per_second": 15.122,
|
| 353 |
+
"eval_steps_per_second": 0.945,
|
| 354 |
+
"step": 569
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.0,
|
| 358 |
+
"learning_rate": 4.6289062500000005e-05,
|
| 359 |
+
"loss": 0.0081,
|
| 360 |
+
"step": 570
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.02,
|
| 364 |
+
"learning_rate": 4.6126302083333335e-05,
|
| 365 |
+
"loss": 0.0097,
|
| 366 |
+
"step": 580
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.04,
|
| 370 |
+
"learning_rate": 4.596354166666667e-05,
|
| 371 |
+
"loss": 0.0084,
|
| 372 |
+
"step": 590
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 1.05,
|
| 376 |
+
"learning_rate": 4.5800781250000004e-05,
|
| 377 |
+
"loss": 0.0032,
|
| 378 |
+
"step": 600
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 1.07,
|
| 382 |
+
"learning_rate": 4.5638020833333335e-05,
|
| 383 |
+
"loss": 0.017,
|
| 384 |
+
"step": 610
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.09,
|
| 388 |
+
"learning_rate": 4.547526041666667e-05,
|
| 389 |
+
"loss": 0.0099,
|
| 390 |
+
"step": 620
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 1.11,
|
| 394 |
+
"learning_rate": 4.5312500000000004e-05,
|
| 395 |
+
"loss": 0.0126,
|
| 396 |
+
"step": 630
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.12,
|
| 400 |
+
"learning_rate": 4.5149739583333335e-05,
|
| 401 |
+
"loss": 0.0001,
|
| 402 |
+
"step": 640
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.14,
|
| 406 |
+
"learning_rate": 4.498697916666667e-05,
|
| 407 |
+
"loss": 0.0009,
|
| 408 |
+
"step": 650
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.16,
|
| 412 |
+
"learning_rate": 4.482421875e-05,
|
| 413 |
+
"loss": 0.0134,
|
| 414 |
+
"step": 660
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.18,
|
| 418 |
+
"learning_rate": 4.4661458333333334e-05,
|
| 419 |
+
"loss": 0.0115,
|
| 420 |
+
"step": 670
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.19,
|
| 424 |
+
"learning_rate": 4.449869791666667e-05,
|
| 425 |
+
"loss": 0.0019,
|
| 426 |
+
"step": 680
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.21,
|
| 430 |
+
"learning_rate": 4.43359375e-05,
|
| 431 |
+
"loss": 0.0005,
|
| 432 |
+
"step": 690
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.23,
|
| 436 |
+
"learning_rate": 4.4173177083333334e-05,
|
| 437 |
+
"loss": 0.0229,
|
| 438 |
+
"step": 700
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 1.25,
|
| 442 |
+
"learning_rate": 4.401041666666667e-05,
|
| 443 |
+
"loss": 0.0052,
|
| 444 |
+
"step": 710
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.26,
|
| 448 |
+
"learning_rate": 4.384765625e-05,
|
| 449 |
+
"loss": 0.0088,
|
| 450 |
+
"step": 720
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.28,
|
| 454 |
+
"learning_rate": 4.368489583333333e-05,
|
| 455 |
+
"loss": 0.0252,
|
| 456 |
+
"step": 730
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 1.3,
|
| 460 |
+
"learning_rate": 4.352213541666667e-05,
|
| 461 |
+
"loss": 0.0143,
|
| 462 |
+
"step": 740
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 1.32,
|
| 466 |
+
"learning_rate": 4.3359375e-05,
|
| 467 |
+
"loss": 0.0051,
|
| 468 |
+
"step": 750
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.33,
|
| 472 |
+
"learning_rate": 4.319661458333333e-05,
|
| 473 |
+
"loss": 0.0282,
|
| 474 |
+
"step": 760
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.35,
|
| 478 |
+
"learning_rate": 4.303385416666667e-05,
|
| 479 |
+
"loss": 0.0107,
|
| 480 |
+
"step": 770
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.37,
|
| 484 |
+
"learning_rate": 4.287109375e-05,
|
| 485 |
+
"loss": 0.022,
|
| 486 |
+
"step": 780
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.39,
|
| 490 |
+
"learning_rate": 4.270833333333333e-05,
|
| 491 |
+
"loss": 0.0206,
|
| 492 |
+
"step": 790
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.4,
|
| 496 |
+
"learning_rate": 4.254557291666667e-05,
|
| 497 |
+
"loss": 0.0218,
|
| 498 |
+
"step": 800
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 1.42,
|
| 502 |
+
"learning_rate": 4.23828125e-05,
|
| 503 |
+
"loss": 0.0038,
|
| 504 |
+
"step": 810
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.44,
|
| 508 |
+
"learning_rate": 4.222005208333333e-05,
|
| 509 |
+
"loss": 0.0176,
|
| 510 |
+
"step": 820
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.46,
|
| 514 |
+
"learning_rate": 4.205729166666667e-05,
|
| 515 |
+
"loss": 0.0154,
|
| 516 |
+
"step": 830
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 1.47,
|
| 520 |
+
"learning_rate": 4.189453125e-05,
|
| 521 |
+
"loss": 0.0066,
|
| 522 |
+
"step": 840
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.49,
|
| 526 |
+
"learning_rate": 4.173177083333333e-05,
|
| 527 |
+
"loss": 0.013,
|
| 528 |
+
"step": 850
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.51,
|
| 532 |
+
"learning_rate": 4.156901041666667e-05,
|
| 533 |
+
"loss": 0.0027,
|
| 534 |
+
"step": 860
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 1.53,
|
| 538 |
+
"learning_rate": 4.140625e-05,
|
| 539 |
+
"loss": 0.0085,
|
| 540 |
+
"step": 870
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 1.54,
|
| 544 |
+
"learning_rate": 4.124348958333333e-05,
|
| 545 |
+
"loss": 0.0099,
|
| 546 |
+
"step": 880
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 1.56,
|
| 550 |
+
"learning_rate": 4.108072916666667e-05,
|
| 551 |
+
"loss": 0.0059,
|
| 552 |
+
"step": 890
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.58,
|
| 556 |
+
"learning_rate": 4.091796875e-05,
|
| 557 |
+
"loss": 0.0067,
|
| 558 |
+
"step": 900
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 1.6,
|
| 562 |
+
"learning_rate": 4.075520833333333e-05,
|
| 563 |
+
"loss": 0.0073,
|
| 564 |
+
"step": 910
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.61,
|
| 568 |
+
"learning_rate": 4.059244791666667e-05,
|
| 569 |
+
"loss": 0.006,
|
| 570 |
+
"step": 920
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.63,
|
| 574 |
+
"learning_rate": 4.04296875e-05,
|
| 575 |
+
"loss": 0.0022,
|
| 576 |
+
"step": 930
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.65,
|
| 580 |
+
"learning_rate": 4.026692708333333e-05,
|
| 581 |
+
"loss": 0.0007,
|
| 582 |
+
"step": 940
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.67,
|
| 586 |
+
"learning_rate": 4.010416666666667e-05,
|
| 587 |
+
"loss": 0.0031,
|
| 588 |
+
"step": 950
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 1.68,
|
| 592 |
+
"learning_rate": 3.994140625e-05,
|
| 593 |
+
"loss": 0.0107,
|
| 594 |
+
"step": 960
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.7,
|
| 598 |
+
"learning_rate": 3.9778645833333337e-05,
|
| 599 |
+
"loss": 0.0582,
|
| 600 |
+
"step": 970
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 1.72,
|
| 604 |
+
"learning_rate": 3.961588541666667e-05,
|
| 605 |
+
"loss": 0.003,
|
| 606 |
+
"step": 980
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 1.74,
|
| 610 |
+
"learning_rate": 3.9453125000000005e-05,
|
| 611 |
+
"loss": 0.0234,
|
| 612 |
+
"step": 990
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.76,
|
| 616 |
+
"learning_rate": 3.9290364583333336e-05,
|
| 617 |
+
"loss": 0.0026,
|
| 618 |
+
"step": 1000
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 1.77,
|
| 622 |
+
"learning_rate": 3.912760416666667e-05,
|
| 623 |
+
"loss": 0.0067,
|
| 624 |
+
"step": 1010
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.79,
|
| 628 |
+
"learning_rate": 3.8964843750000005e-05,
|
| 629 |
+
"loss": 0.0079,
|
| 630 |
+
"step": 1020
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 1.81,
|
| 634 |
+
"learning_rate": 3.8802083333333336e-05,
|
| 635 |
+
"loss": 0.0045,
|
| 636 |
+
"step": 1030
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.83,
|
| 640 |
+
"learning_rate": 3.8639322916666673e-05,
|
| 641 |
+
"loss": 0.0007,
|
| 642 |
+
"step": 1040
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 1.84,
|
| 646 |
+
"learning_rate": 3.8476562500000004e-05,
|
| 647 |
+
"loss": 0.0058,
|
| 648 |
+
"step": 1050
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 1.86,
|
| 652 |
+
"learning_rate": 3.8313802083333335e-05,
|
| 653 |
+
"loss": 0.0004,
|
| 654 |
+
"step": 1060
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.88,
|
| 658 |
+
"learning_rate": 3.815104166666667e-05,
|
| 659 |
+
"loss": 0.0005,
|
| 660 |
+
"step": 1070
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 1.9,
|
| 664 |
+
"learning_rate": 3.7988281250000004e-05,
|
| 665 |
+
"loss": 0.01,
|
| 666 |
+
"step": 1080
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 1.91,
|
| 670 |
+
"learning_rate": 3.7825520833333335e-05,
|
| 671 |
+
"loss": 0.0002,
|
| 672 |
+
"step": 1090
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.93,
|
| 676 |
+
"learning_rate": 3.766276041666667e-05,
|
| 677 |
+
"loss": 0.0137,
|
| 678 |
+
"step": 1100
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.95,
|
| 682 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 683 |
+
"loss": 0.0002,
|
| 684 |
+
"step": 1110
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 1.97,
|
| 688 |
+
"learning_rate": 3.7337239583333334e-05,
|
| 689 |
+
"loss": 0.0067,
|
| 690 |
+
"step": 1120
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 1.98,
|
| 694 |
+
"learning_rate": 3.717447916666667e-05,
|
| 695 |
+
"loss": 0.0002,
|
| 696 |
+
"step": 1130
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 2.0,
|
| 700 |
+
"eval_accuracy": 0.9995,
|
| 701 |
+
"eval_loss": 0.0011123190633952618,
|
| 702 |
+
"eval_runtime": 135.6796,
|
| 703 |
+
"eval_samples_per_second": 14.741,
|
| 704 |
+
"eval_steps_per_second": 0.921,
|
| 705 |
+
"step": 1139
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 2.0,
|
| 709 |
+
"learning_rate": 3.701171875e-05,
|
| 710 |
+
"loss": 0.0006,
|
| 711 |
+
"step": 1140
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.02,
|
| 715 |
+
"learning_rate": 3.6848958333333334e-05,
|
| 716 |
+
"loss": 0.0001,
|
| 717 |
+
"step": 1150
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.04,
|
| 721 |
+
"learning_rate": 3.668619791666667e-05,
|
| 722 |
+
"loss": 0.0001,
|
| 723 |
+
"step": 1160
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.05,
|
| 727 |
+
"learning_rate": 3.65234375e-05,
|
| 728 |
+
"loss": 0.0002,
|
| 729 |
+
"step": 1170
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.07,
|
| 733 |
+
"learning_rate": 3.6360677083333334e-05,
|
| 734 |
+
"loss": 0.0001,
|
| 735 |
+
"step": 1180
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 2.09,
|
| 739 |
+
"learning_rate": 3.619791666666667e-05,
|
| 740 |
+
"loss": 0.004,
|
| 741 |
+
"step": 1190
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.11,
|
| 745 |
+
"learning_rate": 3.603515625e-05,
|
| 746 |
+
"loss": 0.0038,
|
| 747 |
+
"step": 1200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 2.12,
|
| 751 |
+
"learning_rate": 3.587239583333333e-05,
|
| 752 |
+
"loss": 0.0234,
|
| 753 |
+
"step": 1210
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.14,
|
| 757 |
+
"learning_rate": 3.570963541666667e-05,
|
| 758 |
+
"loss": 0.0009,
|
| 759 |
+
"step": 1220
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.16,
|
| 763 |
+
"learning_rate": 3.5546875e-05,
|
| 764 |
+
"loss": 0.0022,
|
| 765 |
+
"step": 1230
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.18,
|
| 769 |
+
"learning_rate": 3.538411458333333e-05,
|
| 770 |
+
"loss": 0.0001,
|
| 771 |
+
"step": 1240
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.19,
|
| 775 |
+
"learning_rate": 3.522135416666667e-05,
|
| 776 |
+
"loss": 0.0003,
|
| 777 |
+
"step": 1250
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.21,
|
| 781 |
+
"learning_rate": 3.505859375e-05,
|
| 782 |
+
"loss": 0.0071,
|
| 783 |
+
"step": 1260
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.23,
|
| 787 |
+
"learning_rate": 3.489583333333333e-05,
|
| 788 |
+
"loss": 0.0004,
|
| 789 |
+
"step": 1270
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 2.25,
|
| 793 |
+
"learning_rate": 3.473307291666667e-05,
|
| 794 |
+
"loss": 0.0008,
|
| 795 |
+
"step": 1280
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.26,
|
| 799 |
+
"learning_rate": 3.45703125e-05,
|
| 800 |
+
"loss": 0.0048,
|
| 801 |
+
"step": 1290
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 2.28,
|
| 805 |
+
"learning_rate": 3.440755208333333e-05,
|
| 806 |
+
"loss": 0.0,
|
| 807 |
+
"step": 1300
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.3,
|
| 811 |
+
"learning_rate": 3.424479166666667e-05,
|
| 812 |
+
"loss": 0.0,
|
| 813 |
+
"step": 1310
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.32,
|
| 817 |
+
"learning_rate": 3.408203125e-05,
|
| 818 |
+
"loss": 0.0,
|
| 819 |
+
"step": 1320
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.33,
|
| 823 |
+
"learning_rate": 3.391927083333333e-05,
|
| 824 |
+
"loss": 0.0,
|
| 825 |
+
"step": 1330
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.35,
|
| 829 |
+
"learning_rate": 3.375651041666667e-05,
|
| 830 |
+
"loss": 0.0,
|
| 831 |
+
"step": 1340
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.37,
|
| 835 |
+
"learning_rate": 3.359375e-05,
|
| 836 |
+
"loss": 0.0002,
|
| 837 |
+
"step": 1350
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.39,
|
| 841 |
+
"learning_rate": 3.343098958333333e-05,
|
| 842 |
+
"loss": 0.0001,
|
| 843 |
+
"step": 1360
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.4,
|
| 847 |
+
"learning_rate": 3.326822916666667e-05,
|
| 848 |
+
"loss": 0.0019,
|
| 849 |
+
"step": 1370
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.42,
|
| 853 |
+
"learning_rate": 3.310546875e-05,
|
| 854 |
+
"loss": 0.0001,
|
| 855 |
+
"step": 1380
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.44,
|
| 859 |
+
"learning_rate": 3.294270833333333e-05,
|
| 860 |
+
"loss": 0.0,
|
| 861 |
+
"step": 1390
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.46,
|
| 865 |
+
"learning_rate": 3.277994791666667e-05,
|
| 866 |
+
"loss": 0.0,
|
| 867 |
+
"step": 1400
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.47,
|
| 871 |
+
"learning_rate": 3.26171875e-05,
|
| 872 |
+
"loss": 0.0,
|
| 873 |
+
"step": 1410
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.49,
|
| 877 |
+
"learning_rate": 3.245442708333333e-05,
|
| 878 |
+
"loss": 0.0,
|
| 879 |
+
"step": 1420
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.51,
|
| 883 |
+
"learning_rate": 3.229166666666667e-05,
|
| 884 |
+
"loss": 0.0,
|
| 885 |
+
"step": 1430
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.53,
|
| 889 |
+
"learning_rate": 3.212890625e-05,
|
| 890 |
+
"loss": 0.0,
|
| 891 |
+
"step": 1440
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.54,
|
| 895 |
+
"learning_rate": 3.196614583333333e-05,
|
| 896 |
+
"loss": 0.0022,
|
| 897 |
+
"step": 1450
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.56,
|
| 901 |
+
"learning_rate": 3.180338541666667e-05,
|
| 902 |
+
"loss": 0.0152,
|
| 903 |
+
"step": 1460
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.58,
|
| 907 |
+
"learning_rate": 3.1640625e-05,
|
| 908 |
+
"loss": 0.0022,
|
| 909 |
+
"step": 1470
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.6,
|
| 913 |
+
"learning_rate": 3.1477864583333336e-05,
|
| 914 |
+
"loss": 0.0002,
|
| 915 |
+
"step": 1480
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.62,
|
| 919 |
+
"learning_rate": 3.131510416666667e-05,
|
| 920 |
+
"loss": 0.0008,
|
| 921 |
+
"step": 1490
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.63,
|
| 925 |
+
"learning_rate": 3.1152343750000005e-05,
|
| 926 |
+
"loss": 0.0003,
|
| 927 |
+
"step": 1500
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.65,
|
| 931 |
+
"learning_rate": 3.0989583333333336e-05,
|
| 932 |
+
"loss": 0.0051,
|
| 933 |
+
"step": 1510
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 2.67,
|
| 937 |
+
"learning_rate": 3.082682291666667e-05,
|
| 938 |
+
"loss": 0.0,
|
| 939 |
+
"step": 1520
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 2.69,
|
| 943 |
+
"learning_rate": 3.0664062500000004e-05,
|
| 944 |
+
"loss": 0.0,
|
| 945 |
+
"step": 1530
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 2.7,
|
| 949 |
+
"learning_rate": 3.0501302083333332e-05,
|
| 950 |
+
"loss": 0.0035,
|
| 951 |
+
"step": 1540
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 2.72,
|
| 955 |
+
"learning_rate": 3.033854166666667e-05,
|
| 956 |
+
"loss": 0.0084,
|
| 957 |
+
"step": 1550
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 2.74,
|
| 961 |
+
"learning_rate": 3.017578125e-05,
|
| 962 |
+
"loss": 0.0002,
|
| 963 |
+
"step": 1560
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 2.76,
|
| 967 |
+
"learning_rate": 3.001302083333333e-05,
|
| 968 |
+
"loss": 0.0168,
|
| 969 |
+
"step": 1570
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.77,
|
| 973 |
+
"learning_rate": 2.985026041666667e-05,
|
| 974 |
+
"loss": 0.0019,
|
| 975 |
+
"step": 1580
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 2.79,
|
| 979 |
+
"learning_rate": 2.96875e-05,
|
| 980 |
+
"loss": 0.0002,
|
| 981 |
+
"step": 1590
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 2.81,
|
| 985 |
+
"learning_rate": 2.9524739583333334e-05,
|
| 986 |
+
"loss": 0.0005,
|
| 987 |
+
"step": 1600
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 2.83,
|
| 991 |
+
"learning_rate": 2.936197916666667e-05,
|
| 992 |
+
"loss": 0.001,
|
| 993 |
+
"step": 1610
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 2.84,
|
| 997 |
+
"learning_rate": 2.9199218750000003e-05,
|
| 998 |
+
"loss": 0.0,
|
| 999 |
+
"step": 1620
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 2.86,
|
| 1003 |
+
"learning_rate": 2.9036458333333334e-05,
|
| 1004 |
+
"loss": 0.0,
|
| 1005 |
+
"step": 1630
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 2.88,
|
| 1009 |
+
"learning_rate": 2.887369791666667e-05,
|
| 1010 |
+
"loss": 0.0,
|
| 1011 |
+
"step": 1640
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 2.9,
|
| 1015 |
+
"learning_rate": 2.8710937500000002e-05,
|
| 1016 |
+
"loss": 0.002,
|
| 1017 |
+
"step": 1650
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.91,
|
| 1021 |
+
"learning_rate": 2.8548177083333333e-05,
|
| 1022 |
+
"loss": 0.0,
|
| 1023 |
+
"step": 1660
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 2.93,
|
| 1027 |
+
"learning_rate": 2.838541666666667e-05,
|
| 1028 |
+
"loss": 0.0,
|
| 1029 |
+
"step": 1670
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.95,
|
| 1033 |
+
"learning_rate": 2.8222656250000002e-05,
|
| 1034 |
+
"loss": 0.0,
|
| 1035 |
+
"step": 1680
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 2.97,
|
| 1039 |
+
"learning_rate": 2.8059895833333333e-05,
|
| 1040 |
+
"loss": 0.0,
|
| 1041 |
+
"step": 1690
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 2.98,
|
| 1045 |
+
"learning_rate": 2.789713541666667e-05,
|
| 1046 |
+
"loss": 0.0006,
|
| 1047 |
+
"step": 1700
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.0,
|
| 1051 |
+
"eval_accuracy": 1.0,
|
| 1052 |
+
"eval_loss": 9.786892042029649e-05,
|
| 1053 |
+
"eval_runtime": 133.229,
|
| 1054 |
+
"eval_samples_per_second": 15.012,
|
| 1055 |
+
"eval_steps_per_second": 0.938,
|
| 1056 |
+
"step": 1709
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 3.0,
|
| 1060 |
+
"learning_rate": 2.7734375e-05,
|
| 1061 |
+
"loss": 0.0017,
|
| 1062 |
+
"step": 1710
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 3.02,
|
| 1066 |
+
"learning_rate": 2.7571614583333333e-05,
|
| 1067 |
+
"loss": 0.001,
|
| 1068 |
+
"step": 1720
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 3.04,
|
| 1072 |
+
"learning_rate": 2.740885416666667e-05,
|
| 1073 |
+
"loss": 0.0195,
|
| 1074 |
+
"step": 1730
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 3.05,
|
| 1078 |
+
"learning_rate": 2.724609375e-05,
|
| 1079 |
+
"loss": 0.0001,
|
| 1080 |
+
"step": 1740
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 3.07,
|
| 1084 |
+
"learning_rate": 2.7083333333333332e-05,
|
| 1085 |
+
"loss": 0.0141,
|
| 1086 |
+
"step": 1750
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 3.09,
|
| 1090 |
+
"learning_rate": 2.692057291666667e-05,
|
| 1091 |
+
"loss": 0.0002,
|
| 1092 |
+
"step": 1760
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 3.11,
|
| 1096 |
+
"learning_rate": 2.67578125e-05,
|
| 1097 |
+
"loss": 0.0072,
|
| 1098 |
+
"step": 1770
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 3.12,
|
| 1102 |
+
"learning_rate": 2.659505208333333e-05,
|
| 1103 |
+
"loss": 0.0,
|
| 1104 |
+
"step": 1780
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 3.14,
|
| 1108 |
+
"learning_rate": 2.643229166666667e-05,
|
| 1109 |
+
"loss": 0.0001,
|
| 1110 |
+
"step": 1790
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 3.16,
|
| 1114 |
+
"learning_rate": 2.626953125e-05,
|
| 1115 |
+
"loss": 0.0,
|
| 1116 |
+
"step": 1800
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 3.18,
|
| 1120 |
+
"learning_rate": 2.610677083333333e-05,
|
| 1121 |
+
"loss": 0.0,
|
| 1122 |
+
"step": 1810
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 3.19,
|
| 1126 |
+
"learning_rate": 2.594401041666667e-05,
|
| 1127 |
+
"loss": 0.0001,
|
| 1128 |
+
"step": 1820
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 3.21,
|
| 1132 |
+
"learning_rate": 2.578125e-05,
|
| 1133 |
+
"loss": 0.0098,
|
| 1134 |
+
"step": 1830
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 3.23,
|
| 1138 |
+
"learning_rate": 2.5618489583333334e-05,
|
| 1139 |
+
"loss": 0.0014,
|
| 1140 |
+
"step": 1840
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 3.25,
|
| 1144 |
+
"learning_rate": 2.545572916666667e-05,
|
| 1145 |
+
"loss": 0.0003,
|
| 1146 |
+
"step": 1850
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 3.26,
|
| 1150 |
+
"learning_rate": 2.5292968750000003e-05,
|
| 1151 |
+
"loss": 0.0021,
|
| 1152 |
+
"step": 1860
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 3.28,
|
| 1156 |
+
"learning_rate": 2.5130208333333334e-05,
|
| 1157 |
+
"loss": 0.0,
|
| 1158 |
+
"step": 1870
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 3.3,
|
| 1162 |
+
"learning_rate": 2.4967447916666668e-05,
|
| 1163 |
+
"loss": 0.0005,
|
| 1164 |
+
"step": 1880
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 3.32,
|
| 1168 |
+
"learning_rate": 2.4804687500000002e-05,
|
| 1169 |
+
"loss": 0.0054,
|
| 1170 |
+
"step": 1890
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 3.33,
|
| 1174 |
+
"learning_rate": 2.4641927083333337e-05,
|
| 1175 |
+
"loss": 0.0004,
|
| 1176 |
+
"step": 1900
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 3.35,
|
| 1180 |
+
"learning_rate": 2.4479166666666668e-05,
|
| 1181 |
+
"loss": 0.0,
|
| 1182 |
+
"step": 1910
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 3.37,
|
| 1186 |
+
"learning_rate": 2.4316406250000002e-05,
|
| 1187 |
+
"loss": 0.0,
|
| 1188 |
+
"step": 1920
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 3.39,
|
| 1192 |
+
"learning_rate": 2.4153645833333336e-05,
|
| 1193 |
+
"loss": 0.0,
|
| 1194 |
+
"step": 1930
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 3.41,
|
| 1198 |
+
"learning_rate": 2.3990885416666667e-05,
|
| 1199 |
+
"loss": 0.0,
|
| 1200 |
+
"step": 1940
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 3.42,
|
| 1204 |
+
"learning_rate": 2.3828125e-05,
|
| 1205 |
+
"loss": 0.0,
|
| 1206 |
+
"step": 1950
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 3.44,
|
| 1210 |
+
"learning_rate": 2.3665364583333336e-05,
|
| 1211 |
+
"loss": 0.0053,
|
| 1212 |
+
"step": 1960
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 3.46,
|
| 1216 |
+
"learning_rate": 2.3502604166666667e-05,
|
| 1217 |
+
"loss": 0.0025,
|
| 1218 |
+
"step": 1970
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 3.48,
|
| 1222 |
+
"learning_rate": 2.333984375e-05,
|
| 1223 |
+
"loss": 0.0002,
|
| 1224 |
+
"step": 1980
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 3.49,
|
| 1228 |
+
"learning_rate": 2.3177083333333335e-05,
|
| 1229 |
+
"loss": 0.0001,
|
| 1230 |
+
"step": 1990
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 3.51,
|
| 1234 |
+
"learning_rate": 2.3014322916666666e-05,
|
| 1235 |
+
"loss": 0.0019,
|
| 1236 |
+
"step": 2000
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 3.53,
|
| 1240 |
+
"learning_rate": 2.28515625e-05,
|
| 1241 |
+
"loss": 0.0,
|
| 1242 |
+
"step": 2010
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 3.55,
|
| 1246 |
+
"learning_rate": 2.2688802083333335e-05,
|
| 1247 |
+
"loss": 0.0001,
|
| 1248 |
+
"step": 2020
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 3.56,
|
| 1252 |
+
"learning_rate": 2.2526041666666666e-05,
|
| 1253 |
+
"loss": 0.0031,
|
| 1254 |
+
"step": 2030
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 3.58,
|
| 1258 |
+
"learning_rate": 2.236328125e-05,
|
| 1259 |
+
"loss": 0.0003,
|
| 1260 |
+
"step": 2040
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 3.6,
|
| 1264 |
+
"learning_rate": 2.2200520833333335e-05,
|
| 1265 |
+
"loss": 0.0002,
|
| 1266 |
+
"step": 2050
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 3.62,
|
| 1270 |
+
"learning_rate": 2.2037760416666665e-05,
|
| 1271 |
+
"loss": 0.0109,
|
| 1272 |
+
"step": 2060
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 3.63,
|
| 1276 |
+
"learning_rate": 2.1875e-05,
|
| 1277 |
+
"loss": 0.0001,
|
| 1278 |
+
"step": 2070
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 3.65,
|
| 1282 |
+
"learning_rate": 2.1712239583333334e-05,
|
| 1283 |
+
"loss": 0.0014,
|
| 1284 |
+
"step": 2080
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 3.67,
|
| 1288 |
+
"learning_rate": 2.154947916666667e-05,
|
| 1289 |
+
"loss": 0.0001,
|
| 1290 |
+
"step": 2090
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 3.69,
|
| 1294 |
+
"learning_rate": 2.138671875e-05,
|
| 1295 |
+
"loss": 0.0013,
|
| 1296 |
+
"step": 2100
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 3.7,
|
| 1300 |
+
"learning_rate": 2.1223958333333334e-05,
|
| 1301 |
+
"loss": 0.0,
|
| 1302 |
+
"step": 2110
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 3.72,
|
| 1306 |
+
"learning_rate": 2.1061197916666668e-05,
|
| 1307 |
+
"loss": 0.0,
|
| 1308 |
+
"step": 2120
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 3.74,
|
| 1312 |
+
"learning_rate": 2.0898437500000002e-05,
|
| 1313 |
+
"loss": 0.0,
|
| 1314 |
+
"step": 2130
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.76,
|
| 1318 |
+
"learning_rate": 2.0735677083333337e-05,
|
| 1319 |
+
"loss": 0.0,
|
| 1320 |
+
"step": 2140
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 3.77,
|
| 1324 |
+
"learning_rate": 2.0572916666666668e-05,
|
| 1325 |
+
"loss": 0.0,
|
| 1326 |
+
"step": 2150
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 3.79,
|
| 1330 |
+
"learning_rate": 2.0410156250000002e-05,
|
| 1331 |
+
"loss": 0.0,
|
| 1332 |
+
"step": 2160
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 3.81,
|
| 1336 |
+
"learning_rate": 2.0247395833333336e-05,
|
| 1337 |
+
"loss": 0.0195,
|
| 1338 |
+
"step": 2170
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 3.83,
|
| 1342 |
+
"learning_rate": 2.0084635416666667e-05,
|
| 1343 |
+
"loss": 0.02,
|
| 1344 |
+
"step": 2180
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 3.84,
|
| 1348 |
+
"learning_rate": 1.9921875e-05,
|
| 1349 |
+
"loss": 0.0,
|
| 1350 |
+
"step": 2190
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 3.86,
|
| 1354 |
+
"learning_rate": 1.9759114583333336e-05,
|
| 1355 |
+
"loss": 0.0,
|
| 1356 |
+
"step": 2200
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 3.88,
|
| 1360 |
+
"learning_rate": 1.9596354166666667e-05,
|
| 1361 |
+
"loss": 0.0,
|
| 1362 |
+
"step": 2210
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 3.9,
|
| 1366 |
+
"learning_rate": 1.943359375e-05,
|
| 1367 |
+
"loss": 0.0,
|
| 1368 |
+
"step": 2220
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 3.91,
|
| 1372 |
+
"learning_rate": 1.9270833333333335e-05,
|
| 1373 |
+
"loss": 0.0008,
|
| 1374 |
+
"step": 2230
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 3.93,
|
| 1378 |
+
"learning_rate": 1.9108072916666666e-05,
|
| 1379 |
+
"loss": 0.0,
|
| 1380 |
+
"step": 2240
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 3.95,
|
| 1384 |
+
"learning_rate": 1.89453125e-05,
|
| 1385 |
+
"loss": 0.0002,
|
| 1386 |
+
"step": 2250
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 3.97,
|
| 1390 |
+
"learning_rate": 1.8782552083333335e-05,
|
| 1391 |
+
"loss": 0.0055,
|
| 1392 |
+
"step": 2260
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 3.98,
|
| 1396 |
+
"learning_rate": 1.8619791666666666e-05,
|
| 1397 |
+
"loss": 0.0002,
|
| 1398 |
+
"step": 2270
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 4.0,
|
| 1402 |
+
"eval_accuracy": 1.0,
|
| 1403 |
+
"eval_loss": 4.17689407186117e-05,
|
| 1404 |
+
"eval_runtime": 133.2569,
|
| 1405 |
+
"eval_samples_per_second": 15.009,
|
| 1406 |
+
"eval_steps_per_second": 0.938,
|
| 1407 |
+
"step": 2279
|
| 1408 |
+
}
|
| 1409 |
+
],
|
| 1410 |
+
"logging_steps": 10,
|
| 1411 |
+
"max_steps": 3414,
|
| 1412 |
+
"num_input_tokens_seen": 0,
|
| 1413 |
+
"num_train_epochs": 6,
|
| 1414 |
+
"save_steps": 500,
|
| 1415 |
+
"total_flos": 1.1427169361682235e+19,
|
| 1416 |
+
"train_batch_size": 16,
|
| 1417 |
+
"trial_name": null,
|
| 1418 |
+
"trial_params": null
|
| 1419 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:371b385f32e6c123a40eb3bc4908529aedeb055c4b738e4cb69f9c3efaa3c4cc
|
| 3 |
+
size 4664
|