Upload folder using huggingface_hub
#5
by
Manasee27 - opened
- all_results.json +6 -6
- checkpoint-1160/config.json +40 -0
- checkpoint-1160/optimizer.pt +3 -0
- checkpoint-1160/preprocessor_config.json +22 -0
- checkpoint-1160/pytorch_model.bin +3 -0
- checkpoint-1160/rng_state.pth +3 -0
- checkpoint-1160/scheduler.pt +3 -0
- checkpoint-1160/trainer_state.json +973 -0
- checkpoint-1160/training_args.bin +3 -0
- checkpoint-5840/config.json +40 -0
- checkpoint-5840/optimizer.pt +3 -0
- checkpoint-5840/preprocessor_config.json +22 -0
- checkpoint-5840/pytorch_model.bin +3 -0
- checkpoint-5840/rng_state.pth +3 -0
- checkpoint-5840/scheduler.pt +3 -0
- checkpoint-5840/trainer_state.json +0 -0
- checkpoint-5840/training_args.bin +3 -0
- pytorch_model.bin +1 -1
- train_results.json +6 -6
- trainer_state.json +0 -0
- training_args.bin +1 -1
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second": 10.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 15.0,
|
| 3 |
+
"total_flos": 7.243048468628582e+18,
|
| 4 |
+
"train_loss": 0.1559911989340654,
|
| 5 |
+
"train_runtime": 9254.6234,
|
| 6 |
+
"train_samples_per_second": 10.114,
|
| 7 |
+
"train_steps_per_second": 0.632
|
| 8 |
}
|
checkpoint-1160/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "Brazil",
|
| 13 |
+
"1": "Canada",
|
| 14 |
+
"2": "Finland",
|
| 15 |
+
"3": "Japan",
|
| 16 |
+
"4": "United-Kingdom",
|
| 17 |
+
"5": "United_States"
|
| 18 |
+
},
|
| 19 |
+
"image_size": 224,
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"Brazil": "0",
|
| 24 |
+
"Canada": "1",
|
| 25 |
+
"Finland": "2",
|
| 26 |
+
"Japan": "3",
|
| 27 |
+
"United-Kingdom": "4",
|
| 28 |
+
"United_States": "5"
|
| 29 |
+
},
|
| 30 |
+
"layer_norm_eps": 1e-12,
|
| 31 |
+
"model_type": "vit",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_channels": 3,
|
| 34 |
+
"num_hidden_layers": 12,
|
| 35 |
+
"patch_size": 16,
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"qkv_bias": true,
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.31.0"
|
| 40 |
+
}
|
checkpoint-1160/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b662c58d7407b3c58eb022c58ceb753887abd4e03814229f9dffa4dcbde4492
|
| 3 |
+
size 686537861
|
checkpoint-1160/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": "ViTFeatureExtractor",
|
| 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 |
+
}
|
checkpoint-1160/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da704a6cb51e648486e5245ffa315bac0c7f02bb9187d11294f3945a898fb971
|
| 3 |
+
size 343278253
|
checkpoint-1160/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9f17574b458d75c9fb9942b59ba069fba448afd09ba95fb17bda6ac6d4ece4d
|
| 3 |
+
size 13617
|
checkpoint-1160/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d26209d1092214d771860d51e59eb1362f4afe3902eb12c7eddc6b4d9bf3b34e
|
| 3 |
+
size 627
|
checkpoint-1160/trainer_state.json
ADDED
|
@@ -0,0 +1,973 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6604002714157104,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-beans/checkpoint-1160",
|
| 4 |
+
"epoch": 2.9743589743589745,
|
| 5 |
+
"global_step": 1160,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 0.00019965811965811967,
|
| 13 |
+
"loss": 1.7468,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 0.00019931623931623932,
|
| 19 |
+
"loss": 1.653,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.08,
|
| 24 |
+
"learning_rate": 0.00019897435897435898,
|
| 25 |
+
"loss": 1.5026,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 0.00019863247863247864,
|
| 31 |
+
"loss": 1.4296,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.1,
|
| 36 |
+
"eval_accuracy": 0.41975308641975306,
|
| 37 |
+
"eval_loss": 1.4691457748413086,
|
| 38 |
+
"eval_runtime": 61.6215,
|
| 39 |
+
"eval_samples_per_second": 57.902,
|
| 40 |
+
"eval_steps_per_second": 7.238,
|
| 41 |
+
"step": 40
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.13,
|
| 45 |
+
"learning_rate": 0.0001982905982905983,
|
| 46 |
+
"loss": 1.369,
|
| 47 |
+
"step": 50
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.15,
|
| 51 |
+
"learning_rate": 0.00019794871794871796,
|
| 52 |
+
"loss": 1.2948,
|
| 53 |
+
"step": 60
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.18,
|
| 57 |
+
"learning_rate": 0.0001976068376068376,
|
| 58 |
+
"loss": 1.3074,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.21,
|
| 63 |
+
"learning_rate": 0.00019726495726495727,
|
| 64 |
+
"loss": 1.3074,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.21,
|
| 69 |
+
"eval_accuracy": 0.5549943883277216,
|
| 70 |
+
"eval_loss": 1.1999692916870117,
|
| 71 |
+
"eval_runtime": 33.0603,
|
| 72 |
+
"eval_samples_per_second": 107.924,
|
| 73 |
+
"eval_steps_per_second": 13.491,
|
| 74 |
+
"step": 80
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.23,
|
| 78 |
+
"learning_rate": 0.00019692307692307696,
|
| 79 |
+
"loss": 1.1725,
|
| 80 |
+
"step": 90
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.26,
|
| 84 |
+
"learning_rate": 0.00019658119658119659,
|
| 85 |
+
"loss": 1.1944,
|
| 86 |
+
"step": 100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.28,
|
| 90 |
+
"learning_rate": 0.00019623931623931624,
|
| 91 |
+
"loss": 1.18,
|
| 92 |
+
"step": 110
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.31,
|
| 96 |
+
"learning_rate": 0.0001958974358974359,
|
| 97 |
+
"loss": 1.1731,
|
| 98 |
+
"step": 120
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.31,
|
| 102 |
+
"eval_accuracy": 0.6689113355780022,
|
| 103 |
+
"eval_loss": 0.978410005569458,
|
| 104 |
+
"eval_runtime": 37.0531,
|
| 105 |
+
"eval_samples_per_second": 96.294,
|
| 106 |
+
"eval_steps_per_second": 12.037,
|
| 107 |
+
"step": 120
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.33,
|
| 111 |
+
"learning_rate": 0.00019555555555555556,
|
| 112 |
+
"loss": 0.9441,
|
| 113 |
+
"step": 130
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.36,
|
| 117 |
+
"learning_rate": 0.00019521367521367522,
|
| 118 |
+
"loss": 1.0459,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.38,
|
| 123 |
+
"learning_rate": 0.00019487179487179487,
|
| 124 |
+
"loss": 1.1109,
|
| 125 |
+
"step": 150
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.41,
|
| 129 |
+
"learning_rate": 0.00019452991452991453,
|
| 130 |
+
"loss": 1.0094,
|
| 131 |
+
"step": 160
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.41,
|
| 135 |
+
"eval_accuracy": 0.6616161616161617,
|
| 136 |
+
"eval_loss": 0.9483916759490967,
|
| 137 |
+
"eval_runtime": 32.6783,
|
| 138 |
+
"eval_samples_per_second": 109.185,
|
| 139 |
+
"eval_steps_per_second": 13.648,
|
| 140 |
+
"step": 160
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.44,
|
| 144 |
+
"learning_rate": 0.0001941880341880342,
|
| 145 |
+
"loss": 0.9165,
|
| 146 |
+
"step": 170
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.46,
|
| 150 |
+
"learning_rate": 0.00019384615384615385,
|
| 151 |
+
"loss": 1.0904,
|
| 152 |
+
"step": 180
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.49,
|
| 156 |
+
"learning_rate": 0.0001935042735042735,
|
| 157 |
+
"loss": 0.9846,
|
| 158 |
+
"step": 190
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.51,
|
| 162 |
+
"learning_rate": 0.00019316239316239316,
|
| 163 |
+
"loss": 1.0908,
|
| 164 |
+
"step": 200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.51,
|
| 168 |
+
"eval_accuracy": 0.5454545454545454,
|
| 169 |
+
"eval_loss": 1.2091434001922607,
|
| 170 |
+
"eval_runtime": 32.1465,
|
| 171 |
+
"eval_samples_per_second": 110.992,
|
| 172 |
+
"eval_steps_per_second": 13.874,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.54,
|
| 177 |
+
"learning_rate": 0.00019282051282051282,
|
| 178 |
+
"loss": 1.0193,
|
| 179 |
+
"step": 210
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.56,
|
| 183 |
+
"learning_rate": 0.0001924786324786325,
|
| 184 |
+
"loss": 0.9809,
|
| 185 |
+
"step": 220
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.59,
|
| 189 |
+
"learning_rate": 0.00019213675213675216,
|
| 190 |
+
"loss": 0.9673,
|
| 191 |
+
"step": 230
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.62,
|
| 195 |
+
"learning_rate": 0.00019179487179487182,
|
| 196 |
+
"loss": 0.9402,
|
| 197 |
+
"step": 240
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.62,
|
| 201 |
+
"eval_accuracy": 0.6481481481481481,
|
| 202 |
+
"eval_loss": 0.9721583724021912,
|
| 203 |
+
"eval_runtime": 33.9462,
|
| 204 |
+
"eval_samples_per_second": 105.107,
|
| 205 |
+
"eval_steps_per_second": 13.138,
|
| 206 |
+
"step": 240
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.64,
|
| 210 |
+
"learning_rate": 0.00019145299145299148,
|
| 211 |
+
"loss": 0.8103,
|
| 212 |
+
"step": 250
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.67,
|
| 216 |
+
"learning_rate": 0.00019111111111111114,
|
| 217 |
+
"loss": 0.8564,
|
| 218 |
+
"step": 260
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.69,
|
| 222 |
+
"learning_rate": 0.0001907692307692308,
|
| 223 |
+
"loss": 1.1074,
|
| 224 |
+
"step": 270
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.72,
|
| 228 |
+
"learning_rate": 0.00019042735042735042,
|
| 229 |
+
"loss": 1.044,
|
| 230 |
+
"step": 280
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.72,
|
| 234 |
+
"eval_accuracy": 0.5897867564534232,
|
| 235 |
+
"eval_loss": 1.1147791147232056,
|
| 236 |
+
"eval_runtime": 32.6846,
|
| 237 |
+
"eval_samples_per_second": 109.164,
|
| 238 |
+
"eval_steps_per_second": 13.646,
|
| 239 |
+
"step": 280
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.74,
|
| 243 |
+
"learning_rate": 0.00019008547008547008,
|
| 244 |
+
"loss": 1.016,
|
| 245 |
+
"step": 290
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.77,
|
| 249 |
+
"learning_rate": 0.00018974358974358974,
|
| 250 |
+
"loss": 0.8815,
|
| 251 |
+
"step": 300
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.79,
|
| 255 |
+
"learning_rate": 0.0001894017094017094,
|
| 256 |
+
"loss": 0.8718,
|
| 257 |
+
"step": 310
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.82,
|
| 261 |
+
"learning_rate": 0.00018905982905982906,
|
| 262 |
+
"loss": 0.6886,
|
| 263 |
+
"step": 320
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.82,
|
| 267 |
+
"eval_accuracy": 0.7272727272727273,
|
| 268 |
+
"eval_loss": 0.7878678441047668,
|
| 269 |
+
"eval_runtime": 31.1497,
|
| 270 |
+
"eval_samples_per_second": 114.544,
|
| 271 |
+
"eval_steps_per_second": 14.318,
|
| 272 |
+
"step": 320
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.85,
|
| 276 |
+
"learning_rate": 0.0001887179487179487,
|
| 277 |
+
"loss": 0.8122,
|
| 278 |
+
"step": 330
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.87,
|
| 282 |
+
"learning_rate": 0.00018837606837606837,
|
| 283 |
+
"loss": 0.6107,
|
| 284 |
+
"step": 340
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.9,
|
| 288 |
+
"learning_rate": 0.00018803418803418803,
|
| 289 |
+
"loss": 0.7783,
|
| 290 |
+
"step": 350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.92,
|
| 294 |
+
"learning_rate": 0.0001876923076923077,
|
| 295 |
+
"loss": 0.8432,
|
| 296 |
+
"step": 360
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.92,
|
| 300 |
+
"eval_accuracy": 0.7480359147025814,
|
| 301 |
+
"eval_loss": 0.7056980729103088,
|
| 302 |
+
"eval_runtime": 32.2533,
|
| 303 |
+
"eval_samples_per_second": 110.624,
|
| 304 |
+
"eval_steps_per_second": 13.828,
|
| 305 |
+
"step": 360
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.95,
|
| 309 |
+
"learning_rate": 0.00018735042735042737,
|
| 310 |
+
"loss": 0.8314,
|
| 311 |
+
"step": 370
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.97,
|
| 315 |
+
"learning_rate": 0.00018700854700854703,
|
| 316 |
+
"loss": 0.8488,
|
| 317 |
+
"step": 380
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.0,
|
| 321 |
+
"learning_rate": 0.0001866666666666667,
|
| 322 |
+
"loss": 0.8547,
|
| 323 |
+
"step": 390
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.03,
|
| 327 |
+
"learning_rate": 0.00018632478632478634,
|
| 328 |
+
"loss": 0.7394,
|
| 329 |
+
"step": 400
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.03,
|
| 333 |
+
"eval_accuracy": 0.6835016835016835,
|
| 334 |
+
"eval_loss": 0.9139666557312012,
|
| 335 |
+
"eval_runtime": 42.9913,
|
| 336 |
+
"eval_samples_per_second": 82.994,
|
| 337 |
+
"eval_steps_per_second": 10.374,
|
| 338 |
+
"step": 400
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.05,
|
| 342 |
+
"learning_rate": 0.000185982905982906,
|
| 343 |
+
"loss": 0.5767,
|
| 344 |
+
"step": 410
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.08,
|
| 348 |
+
"learning_rate": 0.00018564102564102566,
|
| 349 |
+
"loss": 0.5304,
|
| 350 |
+
"step": 420
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.1,
|
| 354 |
+
"learning_rate": 0.00018529914529914532,
|
| 355 |
+
"loss": 0.6296,
|
| 356 |
+
"step": 430
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.13,
|
| 360 |
+
"learning_rate": 0.00018495726495726497,
|
| 361 |
+
"loss": 0.6004,
|
| 362 |
+
"step": 440
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.13,
|
| 366 |
+
"eval_accuracy": 0.7255892255892256,
|
| 367 |
+
"eval_loss": 0.7451665997505188,
|
| 368 |
+
"eval_runtime": 31.8253,
|
| 369 |
+
"eval_samples_per_second": 112.112,
|
| 370 |
+
"eval_steps_per_second": 14.014,
|
| 371 |
+
"step": 440
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.15,
|
| 375 |
+
"learning_rate": 0.00018461538461538463,
|
| 376 |
+
"loss": 0.518,
|
| 377 |
+
"step": 450
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.18,
|
| 381 |
+
"learning_rate": 0.00018427350427350426,
|
| 382 |
+
"loss": 0.5825,
|
| 383 |
+
"step": 460
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.21,
|
| 387 |
+
"learning_rate": 0.00018393162393162392,
|
| 388 |
+
"loss": 0.5465,
|
| 389 |
+
"step": 470
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.23,
|
| 393 |
+
"learning_rate": 0.00018358974358974358,
|
| 394 |
+
"loss": 0.5201,
|
| 395 |
+
"step": 480
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.23,
|
| 399 |
+
"eval_accuracy": 0.6879910213243546,
|
| 400 |
+
"eval_loss": 0.8496339917182922,
|
| 401 |
+
"eval_runtime": 33.617,
|
| 402 |
+
"eval_samples_per_second": 106.137,
|
| 403 |
+
"eval_steps_per_second": 13.267,
|
| 404 |
+
"step": 480
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.26,
|
| 408 |
+
"learning_rate": 0.00018324786324786324,
|
| 409 |
+
"loss": 0.5016,
|
| 410 |
+
"step": 490
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.28,
|
| 414 |
+
"learning_rate": 0.00018290598290598292,
|
| 415 |
+
"loss": 0.6101,
|
| 416 |
+
"step": 500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.31,
|
| 420 |
+
"learning_rate": 0.00018256410256410258,
|
| 421 |
+
"loss": 0.5572,
|
| 422 |
+
"step": 510
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.33,
|
| 426 |
+
"learning_rate": 0.00018222222222222224,
|
| 427 |
+
"loss": 0.4039,
|
| 428 |
+
"step": 520
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.33,
|
| 432 |
+
"eval_accuracy": 0.7312008978675645,
|
| 433 |
+
"eval_loss": 0.7843908071517944,
|
| 434 |
+
"eval_runtime": 31.7575,
|
| 435 |
+
"eval_samples_per_second": 112.351,
|
| 436 |
+
"eval_steps_per_second": 14.044,
|
| 437 |
+
"step": 520
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.36,
|
| 441 |
+
"learning_rate": 0.0001818803418803419,
|
| 442 |
+
"loss": 0.4831,
|
| 443 |
+
"step": 530
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.38,
|
| 447 |
+
"learning_rate": 0.00018153846153846155,
|
| 448 |
+
"loss": 0.5493,
|
| 449 |
+
"step": 540
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.41,
|
| 453 |
+
"learning_rate": 0.0001811965811965812,
|
| 454 |
+
"loss": 0.4942,
|
| 455 |
+
"step": 550
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.44,
|
| 459 |
+
"learning_rate": 0.00018085470085470087,
|
| 460 |
+
"loss": 0.4475,
|
| 461 |
+
"step": 560
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.44,
|
| 465 |
+
"eval_accuracy": 0.6593714927048261,
|
| 466 |
+
"eval_loss": 0.9572438597679138,
|
| 467 |
+
"eval_runtime": 31.2916,
|
| 468 |
+
"eval_samples_per_second": 114.024,
|
| 469 |
+
"eval_steps_per_second": 14.253,
|
| 470 |
+
"step": 560
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.46,
|
| 474 |
+
"learning_rate": 0.00018051282051282052,
|
| 475 |
+
"loss": 0.4667,
|
| 476 |
+
"step": 570
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.49,
|
| 480 |
+
"learning_rate": 0.00018017094017094018,
|
| 481 |
+
"loss": 0.4685,
|
| 482 |
+
"step": 580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.51,
|
| 486 |
+
"learning_rate": 0.00017982905982905984,
|
| 487 |
+
"loss": 0.5028,
|
| 488 |
+
"step": 590
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.54,
|
| 492 |
+
"learning_rate": 0.0001794871794871795,
|
| 493 |
+
"loss": 0.6071,
|
| 494 |
+
"step": 600
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.54,
|
| 498 |
+
"eval_accuracy": 0.7676767676767676,
|
| 499 |
+
"eval_loss": 0.6666165590286255,
|
| 500 |
+
"eval_runtime": 32.2041,
|
| 501 |
+
"eval_samples_per_second": 110.793,
|
| 502 |
+
"eval_steps_per_second": 13.849,
|
| 503 |
+
"step": 600
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.56,
|
| 507 |
+
"learning_rate": 0.00017914529914529916,
|
| 508 |
+
"loss": 0.5134,
|
| 509 |
+
"step": 610
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.59,
|
| 513 |
+
"learning_rate": 0.0001788034188034188,
|
| 514 |
+
"loss": 0.658,
|
| 515 |
+
"step": 620
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.62,
|
| 519 |
+
"learning_rate": 0.00017846153846153847,
|
| 520 |
+
"loss": 0.606,
|
| 521 |
+
"step": 630
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.64,
|
| 525 |
+
"learning_rate": 0.00017811965811965813,
|
| 526 |
+
"loss": 0.5004,
|
| 527 |
+
"step": 640
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.64,
|
| 531 |
+
"eval_accuracy": 0.7373737373737373,
|
| 532 |
+
"eval_loss": 0.7380097508430481,
|
| 533 |
+
"eval_runtime": 32.5063,
|
| 534 |
+
"eval_samples_per_second": 109.763,
|
| 535 |
+
"eval_steps_per_second": 13.72,
|
| 536 |
+
"step": 640
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.67,
|
| 540 |
+
"learning_rate": 0.00017777777777777779,
|
| 541 |
+
"loss": 0.5404,
|
| 542 |
+
"step": 650
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.69,
|
| 546 |
+
"learning_rate": 0.00017743589743589744,
|
| 547 |
+
"loss": 0.5775,
|
| 548 |
+
"step": 660
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.72,
|
| 552 |
+
"learning_rate": 0.0001770940170940171,
|
| 553 |
+
"loss": 0.5013,
|
| 554 |
+
"step": 670
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.74,
|
| 558 |
+
"learning_rate": 0.00017675213675213676,
|
| 559 |
+
"loss": 0.6024,
|
| 560 |
+
"step": 680
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.74,
|
| 564 |
+
"eval_accuracy": 0.7407407407407407,
|
| 565 |
+
"eval_loss": 0.7545998692512512,
|
| 566 |
+
"eval_runtime": 36.6802,
|
| 567 |
+
"eval_samples_per_second": 97.273,
|
| 568 |
+
"eval_steps_per_second": 12.159,
|
| 569 |
+
"step": 680
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.77,
|
| 573 |
+
"learning_rate": 0.00017641025641025642,
|
| 574 |
+
"loss": 0.4815,
|
| 575 |
+
"step": 690
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.79,
|
| 579 |
+
"learning_rate": 0.00017606837606837607,
|
| 580 |
+
"loss": 0.4727,
|
| 581 |
+
"step": 700
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.82,
|
| 585 |
+
"learning_rate": 0.00017572649572649573,
|
| 586 |
+
"loss": 0.4698,
|
| 587 |
+
"step": 710
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.85,
|
| 591 |
+
"learning_rate": 0.0001753846153846154,
|
| 592 |
+
"loss": 0.4813,
|
| 593 |
+
"step": 720
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 1.85,
|
| 597 |
+
"eval_accuracy": 0.7609427609427609,
|
| 598 |
+
"eval_loss": 0.7190226912498474,
|
| 599 |
+
"eval_runtime": 32.2305,
|
| 600 |
+
"eval_samples_per_second": 110.702,
|
| 601 |
+
"eval_steps_per_second": 13.838,
|
| 602 |
+
"step": 720
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.87,
|
| 606 |
+
"learning_rate": 0.00017504273504273505,
|
| 607 |
+
"loss": 0.5407,
|
| 608 |
+
"step": 730
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.9,
|
| 612 |
+
"learning_rate": 0.0001747008547008547,
|
| 613 |
+
"loss": 0.4137,
|
| 614 |
+
"step": 740
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.92,
|
| 618 |
+
"learning_rate": 0.00017435897435897436,
|
| 619 |
+
"loss": 0.5252,
|
| 620 |
+
"step": 750
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.95,
|
| 624 |
+
"learning_rate": 0.00017401709401709402,
|
| 625 |
+
"loss": 0.5608,
|
| 626 |
+
"step": 760
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.95,
|
| 630 |
+
"eval_accuracy": 0.7283950617283951,
|
| 631 |
+
"eval_loss": 0.7672268748283386,
|
| 632 |
+
"eval_runtime": 34.3059,
|
| 633 |
+
"eval_samples_per_second": 104.006,
|
| 634 |
+
"eval_steps_per_second": 13.001,
|
| 635 |
+
"step": 760
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 1.97,
|
| 639 |
+
"learning_rate": 0.00017367521367521368,
|
| 640 |
+
"loss": 0.4692,
|
| 641 |
+
"step": 770
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 2.0,
|
| 645 |
+
"learning_rate": 0.00017333333333333334,
|
| 646 |
+
"loss": 0.488,
|
| 647 |
+
"step": 780
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 2.03,
|
| 651 |
+
"learning_rate": 0.000172991452991453,
|
| 652 |
+
"loss": 0.3798,
|
| 653 |
+
"step": 790
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 2.05,
|
| 657 |
+
"learning_rate": 0.00017264957264957268,
|
| 658 |
+
"loss": 0.3272,
|
| 659 |
+
"step": 800
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 2.05,
|
| 663 |
+
"eval_accuracy": 0.7497194163860831,
|
| 664 |
+
"eval_loss": 0.7055391073226929,
|
| 665 |
+
"eval_runtime": 34.4041,
|
| 666 |
+
"eval_samples_per_second": 103.709,
|
| 667 |
+
"eval_steps_per_second": 12.964,
|
| 668 |
+
"step": 800
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.08,
|
| 672 |
+
"learning_rate": 0.00017230769230769234,
|
| 673 |
+
"loss": 0.2257,
|
| 674 |
+
"step": 810
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.1,
|
| 678 |
+
"learning_rate": 0.000171965811965812,
|
| 679 |
+
"loss": 0.238,
|
| 680 |
+
"step": 820
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.13,
|
| 684 |
+
"learning_rate": 0.00017162393162393162,
|
| 685 |
+
"loss": 0.3225,
|
| 686 |
+
"step": 830
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.15,
|
| 690 |
+
"learning_rate": 0.00017128205128205128,
|
| 691 |
+
"loss": 0.2803,
|
| 692 |
+
"step": 840
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.15,
|
| 696 |
+
"eval_accuracy": 0.7514029180695847,
|
| 697 |
+
"eval_loss": 0.7669840455055237,
|
| 698 |
+
"eval_runtime": 32.953,
|
| 699 |
+
"eval_samples_per_second": 108.275,
|
| 700 |
+
"eval_steps_per_second": 13.534,
|
| 701 |
+
"step": 840
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 2.18,
|
| 705 |
+
"learning_rate": 0.00017094017094017094,
|
| 706 |
+
"loss": 0.234,
|
| 707 |
+
"step": 850
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 2.21,
|
| 711 |
+
"learning_rate": 0.0001705982905982906,
|
| 712 |
+
"loss": 0.2201,
|
| 713 |
+
"step": 860
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.23,
|
| 717 |
+
"learning_rate": 0.00017025641025641026,
|
| 718 |
+
"loss": 0.4651,
|
| 719 |
+
"step": 870
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 2.26,
|
| 723 |
+
"learning_rate": 0.0001699145299145299,
|
| 724 |
+
"loss": 0.4051,
|
| 725 |
+
"step": 880
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 2.26,
|
| 729 |
+
"eval_accuracy": 0.7620650953984287,
|
| 730 |
+
"eval_loss": 0.691470205783844,
|
| 731 |
+
"eval_runtime": 37.2444,
|
| 732 |
+
"eval_samples_per_second": 95.8,
|
| 733 |
+
"eval_steps_per_second": 11.975,
|
| 734 |
+
"step": 880
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.28,
|
| 738 |
+
"learning_rate": 0.00016957264957264957,
|
| 739 |
+
"loss": 0.3069,
|
| 740 |
+
"step": 890
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.31,
|
| 744 |
+
"learning_rate": 0.00016923076923076923,
|
| 745 |
+
"loss": 0.2787,
|
| 746 |
+
"step": 900
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.33,
|
| 750 |
+
"learning_rate": 0.00016888888888888889,
|
| 751 |
+
"loss": 0.2794,
|
| 752 |
+
"step": 910
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.36,
|
| 756 |
+
"learning_rate": 0.00016854700854700854,
|
| 757 |
+
"loss": 0.3436,
|
| 758 |
+
"step": 920
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.36,
|
| 762 |
+
"eval_accuracy": 0.7631874298540965,
|
| 763 |
+
"eval_loss": 0.6948888301849365,
|
| 764 |
+
"eval_runtime": 32.1442,
|
| 765 |
+
"eval_samples_per_second": 111.0,
|
| 766 |
+
"eval_steps_per_second": 13.875,
|
| 767 |
+
"step": 920
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.38,
|
| 771 |
+
"learning_rate": 0.00016820512820512823,
|
| 772 |
+
"loss": 0.2964,
|
| 773 |
+
"step": 930
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 2.41,
|
| 777 |
+
"learning_rate": 0.00016786324786324789,
|
| 778 |
+
"loss": 0.1973,
|
| 779 |
+
"step": 940
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.44,
|
| 783 |
+
"learning_rate": 0.00016752136752136754,
|
| 784 |
+
"loss": 0.2257,
|
| 785 |
+
"step": 950
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 2.46,
|
| 789 |
+
"learning_rate": 0.0001671794871794872,
|
| 790 |
+
"loss": 0.2759,
|
| 791 |
+
"step": 960
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 2.46,
|
| 795 |
+
"eval_accuracy": 0.7356902356902357,
|
| 796 |
+
"eval_loss": 0.8428576588630676,
|
| 797 |
+
"eval_runtime": 32.8186,
|
| 798 |
+
"eval_samples_per_second": 108.719,
|
| 799 |
+
"eval_steps_per_second": 13.59,
|
| 800 |
+
"step": 960
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 2.49,
|
| 804 |
+
"learning_rate": 0.00016683760683760686,
|
| 805 |
+
"loss": 0.272,
|
| 806 |
+
"step": 970
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.51,
|
| 810 |
+
"learning_rate": 0.00016649572649572652,
|
| 811 |
+
"loss": 0.2721,
|
| 812 |
+
"step": 980
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.54,
|
| 816 |
+
"learning_rate": 0.00016615384615384617,
|
| 817 |
+
"loss": 0.1414,
|
| 818 |
+
"step": 990
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.56,
|
| 822 |
+
"learning_rate": 0.00016581196581196583,
|
| 823 |
+
"loss": 0.5024,
|
| 824 |
+
"step": 1000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 2.56,
|
| 828 |
+
"eval_accuracy": 0.7502805836139169,
|
| 829 |
+
"eval_loss": 0.8222711086273193,
|
| 830 |
+
"eval_runtime": 32.7463,
|
| 831 |
+
"eval_samples_per_second": 108.959,
|
| 832 |
+
"eval_steps_per_second": 13.62,
|
| 833 |
+
"step": 1000
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 2.59,
|
| 837 |
+
"learning_rate": 0.00016547008547008546,
|
| 838 |
+
"loss": 0.3348,
|
| 839 |
+
"step": 1010
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 2.62,
|
| 843 |
+
"learning_rate": 0.00016512820512820512,
|
| 844 |
+
"loss": 0.3692,
|
| 845 |
+
"step": 1020
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 2.64,
|
| 849 |
+
"learning_rate": 0.00016478632478632478,
|
| 850 |
+
"loss": 0.3981,
|
| 851 |
+
"step": 1030
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 2.67,
|
| 855 |
+
"learning_rate": 0.00016444444444444444,
|
| 856 |
+
"loss": 0.3571,
|
| 857 |
+
"step": 1040
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.67,
|
| 861 |
+
"eval_accuracy": 0.7289562289562289,
|
| 862 |
+
"eval_loss": 0.8546615839004517,
|
| 863 |
+
"eval_runtime": 33.6376,
|
| 864 |
+
"eval_samples_per_second": 106.072,
|
| 865 |
+
"eval_steps_per_second": 13.259,
|
| 866 |
+
"step": 1040
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 2.69,
|
| 870 |
+
"learning_rate": 0.0001641025641025641,
|
| 871 |
+
"loss": 0.3647,
|
| 872 |
+
"step": 1050
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.72,
|
| 876 |
+
"learning_rate": 0.00016376068376068375,
|
| 877 |
+
"loss": 0.2547,
|
| 878 |
+
"step": 1060
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 2.74,
|
| 882 |
+
"learning_rate": 0.00016341880341880344,
|
| 883 |
+
"loss": 0.3708,
|
| 884 |
+
"step": 1070
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.77,
|
| 888 |
+
"learning_rate": 0.0001630769230769231,
|
| 889 |
+
"loss": 0.3197,
|
| 890 |
+
"step": 1080
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 2.77,
|
| 894 |
+
"eval_accuracy": 0.754769921436588,
|
| 895 |
+
"eval_loss": 0.7111819982528687,
|
| 896 |
+
"eval_runtime": 33.1915,
|
| 897 |
+
"eval_samples_per_second": 107.497,
|
| 898 |
+
"eval_steps_per_second": 13.437,
|
| 899 |
+
"step": 1080
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 2.79,
|
| 903 |
+
"learning_rate": 0.00016273504273504275,
|
| 904 |
+
"loss": 0.2805,
|
| 905 |
+
"step": 1090
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.82,
|
| 909 |
+
"learning_rate": 0.0001623931623931624,
|
| 910 |
+
"loss": 0.2792,
|
| 911 |
+
"step": 1100
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 2.85,
|
| 915 |
+
"learning_rate": 0.00016205128205128207,
|
| 916 |
+
"loss": 0.3781,
|
| 917 |
+
"step": 1110
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 2.87,
|
| 921 |
+
"learning_rate": 0.00016170940170940172,
|
| 922 |
+
"loss": 0.3872,
|
| 923 |
+
"step": 1120
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 2.87,
|
| 927 |
+
"eval_accuracy": 0.7485970819304153,
|
| 928 |
+
"eval_loss": 0.7666726112365723,
|
| 929 |
+
"eval_runtime": 33.3793,
|
| 930 |
+
"eval_samples_per_second": 106.892,
|
| 931 |
+
"eval_steps_per_second": 13.362,
|
| 932 |
+
"step": 1120
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.9,
|
| 936 |
+
"learning_rate": 0.00016136752136752138,
|
| 937 |
+
"loss": 0.3489,
|
| 938 |
+
"step": 1130
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.92,
|
| 942 |
+
"learning_rate": 0.00016102564102564104,
|
| 943 |
+
"loss": 0.3436,
|
| 944 |
+
"step": 1140
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.95,
|
| 948 |
+
"learning_rate": 0.0001606837606837607,
|
| 949 |
+
"loss": 0.2608,
|
| 950 |
+
"step": 1150
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.97,
|
| 954 |
+
"learning_rate": 0.00016034188034188036,
|
| 955 |
+
"loss": 0.2699,
|
| 956 |
+
"step": 1160
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.97,
|
| 960 |
+
"eval_accuracy": 0.7867564534231201,
|
| 961 |
+
"eval_loss": 0.6604002714157104,
|
| 962 |
+
"eval_runtime": 42.2756,
|
| 963 |
+
"eval_samples_per_second": 84.398,
|
| 964 |
+
"eval_steps_per_second": 10.55,
|
| 965 |
+
"step": 1160
|
| 966 |
+
}
|
| 967 |
+
],
|
| 968 |
+
"max_steps": 5850,
|
| 969 |
+
"num_train_epochs": 15,
|
| 970 |
+
"total_flos": 1.436907983794053e+18,
|
| 971 |
+
"trial_name": null,
|
| 972 |
+
"trial_params": null
|
| 973 |
+
}
|
checkpoint-1160/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:194957c9272212a5dd16465021ce223f9cefe58567f9b9ca890e75d6f7750cde
|
| 3 |
+
size 3899
|
checkpoint-5840/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "Brazil",
|
| 13 |
+
"1": "Canada",
|
| 14 |
+
"2": "Finland",
|
| 15 |
+
"3": "Japan",
|
| 16 |
+
"4": "United-Kingdom",
|
| 17 |
+
"5": "United_States"
|
| 18 |
+
},
|
| 19 |
+
"image_size": 224,
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"Brazil": "0",
|
| 24 |
+
"Canada": "1",
|
| 25 |
+
"Finland": "2",
|
| 26 |
+
"Japan": "3",
|
| 27 |
+
"United-Kingdom": "4",
|
| 28 |
+
"United_States": "5"
|
| 29 |
+
},
|
| 30 |
+
"layer_norm_eps": 1e-12,
|
| 31 |
+
"model_type": "vit",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_channels": 3,
|
| 34 |
+
"num_hidden_layers": 12,
|
| 35 |
+
"patch_size": 16,
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"qkv_bias": true,
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.31.0"
|
| 40 |
+
}
|
checkpoint-5840/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bde7a0bf4d257ae44ac48605d2ae3a0a67354ea8b037a597f7064c7cffc45e41
|
| 3 |
+
size 686537861
|
checkpoint-5840/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": "ViTFeatureExtractor",
|
| 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 |
+
}
|
checkpoint-5840/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b39254ce3321347f2fa992e7061fa635851668f8100b4d49561aa48a285de620
|
| 3 |
+
size 343278253
|
checkpoint-5840/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67beb3ce02205d9d3840c973461d205613e37446b415453d1436cf577351b657
|
| 3 |
+
size 13617
|
checkpoint-5840/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33a41b91166721500270623b46bc8ee3fa73aeadd054449d85a1e03325056e11
|
| 3 |
+
size 627
|
checkpoint-5840/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-5840/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:194957c9272212a5dd16465021ce223f9cefe58567f9b9ca890e75d6f7750cde
|
| 3 |
+
size 3899
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 343278253
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da704a6cb51e648486e5245ffa315bac0c7f02bb9187d11294f3945a898fb971
|
| 3 |
size 343278253
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second": 10.
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 15.0,
|
| 3 |
+
"total_flos": 7.243048468628582e+18,
|
| 4 |
+
"train_loss": 0.1559911989340654,
|
| 5 |
+
"train_runtime": 9254.6234,
|
| 6 |
+
"train_samples_per_second": 10.114,
|
| 7 |
+
"train_steps_per_second": 0.632
|
| 8 |
}
|
trainer_state.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3899
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:194957c9272212a5dd16465021ce223f9cefe58567f9b9ca890e75d6f7750cde
|
| 3 |
size 3899
|