Upload folder using huggingface_hub
#1
by
satwikapaul
- opened
- all_results.json +8 -0
- checkpoint-640/config.json +86 -0
- checkpoint-640/optimizer.pt +3 -0
- checkpoint-640/preprocessor_config.json +22 -0
- checkpoint-640/pytorch_model.bin +3 -0
- checkpoint-640/rng_state.pth +3 -0
- checkpoint-640/scheduler.pt +3 -0
- checkpoint-640/trainer_state.json +544 -0
- checkpoint-640/training_args.bin +3 -0
- checkpoint-960/config.json +86 -0
- checkpoint-960/optimizer.pt +3 -0
- checkpoint-960/preprocessor_config.json +22 -0
- checkpoint-960/pytorch_model.bin +3 -0
- checkpoint-960/rng_state.pth +3 -0
- checkpoint-960/scheduler.pt +3 -0
- checkpoint-960/trainer_state.json +808 -0
- checkpoint-960/training_args.bin +3 -0
- config.json +86 -0
- preprocessor_config.json +22 -0
- pytorch_model.bin +3 -0
- train_results.json +8 -0
- trainer_state.json +829 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"total_flos": 1.1936654282133504e+18,
|
| 4 |
+
"train_loss": 0.5720695184201611,
|
| 5 |
+
"train_runtime": 566.3017,
|
| 6 |
+
"train_samples_per_second": 27.194,
|
| 7 |
+
"train_steps_per_second": 1.731
|
| 8 |
+
}
|
checkpoint-640/config.json
ADDED
|
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "%3F",
|
| 13 |
+
"1": "a",
|
| 14 |
+
"10": "i",
|
| 15 |
+
"11": "j",
|
| 16 |
+
"12": "k",
|
| 17 |
+
"13": "l",
|
| 18 |
+
"14": "m",
|
| 19 |
+
"15": "n",
|
| 20 |
+
"16": "o",
|
| 21 |
+
"17": "p",
|
| 22 |
+
"18": "period",
|
| 23 |
+
"19": "q",
|
| 24 |
+
"2": "b",
|
| 25 |
+
"20": "r",
|
| 26 |
+
"21": "s",
|
| 27 |
+
"22": "t",
|
| 28 |
+
"23": "u",
|
| 29 |
+
"24": "v",
|
| 30 |
+
"25": "w",
|
| 31 |
+
"26": "x",
|
| 32 |
+
"27": "y",
|
| 33 |
+
"28": "z",
|
| 34 |
+
"3": "c",
|
| 35 |
+
"4": "capital",
|
| 36 |
+
"5": "d",
|
| 37 |
+
"6": "e",
|
| 38 |
+
"7": "f",
|
| 39 |
+
"8": "g",
|
| 40 |
+
"9": "h"
|
| 41 |
+
},
|
| 42 |
+
"image_size": 224,
|
| 43 |
+
"initializer_range": 0.02,
|
| 44 |
+
"intermediate_size": 3072,
|
| 45 |
+
"label2id": {
|
| 46 |
+
"%3F": "0",
|
| 47 |
+
"a": "1",
|
| 48 |
+
"b": "2",
|
| 49 |
+
"c": "3",
|
| 50 |
+
"capital": "4",
|
| 51 |
+
"d": "5",
|
| 52 |
+
"e": "6",
|
| 53 |
+
"f": "7",
|
| 54 |
+
"g": "8",
|
| 55 |
+
"h": "9",
|
| 56 |
+
"i": "10",
|
| 57 |
+
"j": "11",
|
| 58 |
+
"k": "12",
|
| 59 |
+
"l": "13",
|
| 60 |
+
"m": "14",
|
| 61 |
+
"n": "15",
|
| 62 |
+
"o": "16",
|
| 63 |
+
"p": "17",
|
| 64 |
+
"period": "18",
|
| 65 |
+
"q": "19",
|
| 66 |
+
"r": "20",
|
| 67 |
+
"s": "21",
|
| 68 |
+
"t": "22",
|
| 69 |
+
"u": "23",
|
| 70 |
+
"v": "24",
|
| 71 |
+
"w": "25",
|
| 72 |
+
"x": "26",
|
| 73 |
+
"y": "27",
|
| 74 |
+
"z": "28"
|
| 75 |
+
},
|
| 76 |
+
"layer_norm_eps": 1e-12,
|
| 77 |
+
"model_type": "vit",
|
| 78 |
+
"num_attention_heads": 12,
|
| 79 |
+
"num_channels": 3,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"patch_size": 16,
|
| 82 |
+
"problem_type": "single_label_classification",
|
| 83 |
+
"qkv_bias": true,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.30.2"
|
| 86 |
+
}
|
checkpoint-640/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94884ed85285a2904269a82d64f7e3b80428c956802f4bb063e4b2a7762eb248
|
| 3 |
+
size 686684933
|
checkpoint-640/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-640/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64d2fae05684355f48986d0590f7aa4df86555c2b14ef2dccd8eea77cedd3bed
|
| 3 |
+
size 343351725
|
checkpoint-640/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab20ec22798cfb03ec9b0af2e45ca3a7d282a1ec1c6c852be97ba0580928637b
|
| 3 |
+
size 14575
|
checkpoint-640/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a4139437ab7f542b6850a12d0272b279528f5c8bab9a929f83e409906b8387a0
|
| 3 |
+
size 627
|
checkpoint-640/trainer_state.json
ADDED
|
@@ -0,0 +1,544 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7314792275428772,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-beans/checkpoint-640",
|
| 4 |
+
"epoch": 13.061224489795919,
|
| 5 |
+
"global_step": 640,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.2,
|
| 12 |
+
"learning_rate": 0.00019795918367346938,
|
| 13 |
+
"loss": 3.374,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.41,
|
| 18 |
+
"learning_rate": 0.0001959183673469388,
|
| 19 |
+
"loss": 3.3572,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.61,
|
| 24 |
+
"learning_rate": 0.00019387755102040816,
|
| 25 |
+
"loss": 3.3211,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.82,
|
| 30 |
+
"learning_rate": 0.00019183673469387756,
|
| 31 |
+
"loss": 3.2257,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.82,
|
| 36 |
+
"eval_accuracy": 0.03470031545741325,
|
| 37 |
+
"eval_loss": 3.4292807579040527,
|
| 38 |
+
"eval_runtime": 3.7472,
|
| 39 |
+
"eval_samples_per_second": 84.597,
|
| 40 |
+
"eval_steps_per_second": 10.675,
|
| 41 |
+
"step": 40
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.02,
|
| 45 |
+
"learning_rate": 0.00018979591836734697,
|
| 46 |
+
"loss": 3.1143,
|
| 47 |
+
"step": 50
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 1.22,
|
| 51 |
+
"learning_rate": 0.00018775510204081634,
|
| 52 |
+
"loss": 3.032,
|
| 53 |
+
"step": 60
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.43,
|
| 57 |
+
"learning_rate": 0.00018571428571428572,
|
| 58 |
+
"loss": 2.8104,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.63,
|
| 63 |
+
"learning_rate": 0.00018367346938775512,
|
| 64 |
+
"loss": 2.6674,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.63,
|
| 69 |
+
"eval_accuracy": 0.17350157728706625,
|
| 70 |
+
"eval_loss": 2.952009916305542,
|
| 71 |
+
"eval_runtime": 3.3938,
|
| 72 |
+
"eval_samples_per_second": 93.405,
|
| 73 |
+
"eval_steps_per_second": 11.786,
|
| 74 |
+
"step": 80
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.84,
|
| 78 |
+
"learning_rate": 0.0001816326530612245,
|
| 79 |
+
"loss": 2.5131,
|
| 80 |
+
"step": 90
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 2.04,
|
| 84 |
+
"learning_rate": 0.0001795918367346939,
|
| 85 |
+
"loss": 2.5287,
|
| 86 |
+
"step": 100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 2.24,
|
| 90 |
+
"learning_rate": 0.00017755102040816327,
|
| 91 |
+
"loss": 2.2121,
|
| 92 |
+
"step": 110
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 2.45,
|
| 96 |
+
"learning_rate": 0.00017551020408163265,
|
| 97 |
+
"loss": 2.0048,
|
| 98 |
+
"step": 120
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 2.45,
|
| 102 |
+
"eval_accuracy": 0.48580441640378547,
|
| 103 |
+
"eval_loss": 2.363018035888672,
|
| 104 |
+
"eval_runtime": 3.7903,
|
| 105 |
+
"eval_samples_per_second": 83.635,
|
| 106 |
+
"eval_steps_per_second": 10.553,
|
| 107 |
+
"step": 120
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 2.65,
|
| 111 |
+
"learning_rate": 0.00017346938775510205,
|
| 112 |
+
"loss": 1.8413,
|
| 113 |
+
"step": 130
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.86,
|
| 117 |
+
"learning_rate": 0.00017142857142857143,
|
| 118 |
+
"loss": 1.7897,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 3.06,
|
| 123 |
+
"learning_rate": 0.00016938775510204083,
|
| 124 |
+
"loss": 1.607,
|
| 125 |
+
"step": 150
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 3.27,
|
| 129 |
+
"learning_rate": 0.00016734693877551023,
|
| 130 |
+
"loss": 1.3493,
|
| 131 |
+
"step": 160
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 3.27,
|
| 135 |
+
"eval_accuracy": 0.5488958990536278,
|
| 136 |
+
"eval_loss": 1.8864167928695679,
|
| 137 |
+
"eval_runtime": 4.2623,
|
| 138 |
+
"eval_samples_per_second": 74.372,
|
| 139 |
+
"eval_steps_per_second": 9.385,
|
| 140 |
+
"step": 160
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.47,
|
| 144 |
+
"learning_rate": 0.0001653061224489796,
|
| 145 |
+
"loss": 1.2421,
|
| 146 |
+
"step": 170
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.67,
|
| 150 |
+
"learning_rate": 0.00016326530612244898,
|
| 151 |
+
"loss": 1.0779,
|
| 152 |
+
"step": 180
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.88,
|
| 156 |
+
"learning_rate": 0.00016122448979591838,
|
| 157 |
+
"loss": 1.0038,
|
| 158 |
+
"step": 190
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 4.08,
|
| 162 |
+
"learning_rate": 0.00015918367346938776,
|
| 163 |
+
"loss": 1.0887,
|
| 164 |
+
"step": 200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.08,
|
| 168 |
+
"eval_accuracy": 0.7350157728706624,
|
| 169 |
+
"eval_loss": 1.4275710582733154,
|
| 170 |
+
"eval_runtime": 3.3105,
|
| 171 |
+
"eval_samples_per_second": 95.757,
|
| 172 |
+
"eval_steps_per_second": 12.083,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 4.29,
|
| 177 |
+
"learning_rate": 0.00015714285714285716,
|
| 178 |
+
"loss": 0.7734,
|
| 179 |
+
"step": 210
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 4.49,
|
| 183 |
+
"learning_rate": 0.00015510204081632654,
|
| 184 |
+
"loss": 0.7969,
|
| 185 |
+
"step": 220
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.69,
|
| 189 |
+
"learning_rate": 0.0001530612244897959,
|
| 190 |
+
"loss": 0.7953,
|
| 191 |
+
"step": 230
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 4.9,
|
| 195 |
+
"learning_rate": 0.0001510204081632653,
|
| 196 |
+
"loss": 0.6649,
|
| 197 |
+
"step": 240
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 4.9,
|
| 201 |
+
"eval_accuracy": 0.7728706624605678,
|
| 202 |
+
"eval_loss": 1.2706444263458252,
|
| 203 |
+
"eval_runtime": 4.1354,
|
| 204 |
+
"eval_samples_per_second": 76.656,
|
| 205 |
+
"eval_steps_per_second": 9.673,
|
| 206 |
+
"step": 240
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 5.1,
|
| 210 |
+
"learning_rate": 0.00014897959183673472,
|
| 211 |
+
"loss": 0.5647,
|
| 212 |
+
"step": 250
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 5.31,
|
| 216 |
+
"learning_rate": 0.0001469387755102041,
|
| 217 |
+
"loss": 0.7201,
|
| 218 |
+
"step": 260
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 5.51,
|
| 222 |
+
"learning_rate": 0.0001448979591836735,
|
| 223 |
+
"loss": 0.606,
|
| 224 |
+
"step": 270
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 5.71,
|
| 228 |
+
"learning_rate": 0.00014285714285714287,
|
| 229 |
+
"loss": 0.5396,
|
| 230 |
+
"step": 280
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 5.71,
|
| 234 |
+
"eval_accuracy": 0.7981072555205048,
|
| 235 |
+
"eval_loss": 1.125301480293274,
|
| 236 |
+
"eval_runtime": 3.3532,
|
| 237 |
+
"eval_samples_per_second": 94.537,
|
| 238 |
+
"eval_steps_per_second": 11.929,
|
| 239 |
+
"step": 280
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 5.92,
|
| 243 |
+
"learning_rate": 0.00014081632653061224,
|
| 244 |
+
"loss": 0.3407,
|
| 245 |
+
"step": 290
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 6.12,
|
| 249 |
+
"learning_rate": 0.00013877551020408165,
|
| 250 |
+
"loss": 0.3539,
|
| 251 |
+
"step": 300
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 6.33,
|
| 255 |
+
"learning_rate": 0.00013673469387755102,
|
| 256 |
+
"loss": 0.3608,
|
| 257 |
+
"step": 310
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 6.53,
|
| 261 |
+
"learning_rate": 0.0001346938775510204,
|
| 262 |
+
"loss": 0.3162,
|
| 263 |
+
"step": 320
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 6.53,
|
| 267 |
+
"eval_accuracy": 0.804416403785489,
|
| 268 |
+
"eval_loss": 1.0156543254852295,
|
| 269 |
+
"eval_runtime": 4.7629,
|
| 270 |
+
"eval_samples_per_second": 66.556,
|
| 271 |
+
"eval_steps_per_second": 8.398,
|
| 272 |
+
"step": 320
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.73,
|
| 276 |
+
"learning_rate": 0.0001326530612244898,
|
| 277 |
+
"loss": 0.335,
|
| 278 |
+
"step": 330
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 6.94,
|
| 282 |
+
"learning_rate": 0.00013061224489795917,
|
| 283 |
+
"loss": 0.2333,
|
| 284 |
+
"step": 340
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 7.14,
|
| 288 |
+
"learning_rate": 0.00012857142857142858,
|
| 289 |
+
"loss": 0.1931,
|
| 290 |
+
"step": 350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 7.35,
|
| 294 |
+
"learning_rate": 0.00012653061224489798,
|
| 295 |
+
"loss": 0.1909,
|
| 296 |
+
"step": 360
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 7.35,
|
| 300 |
+
"eval_accuracy": 0.8548895899053628,
|
| 301 |
+
"eval_loss": 0.8516904711723328,
|
| 302 |
+
"eval_runtime": 3.3664,
|
| 303 |
+
"eval_samples_per_second": 94.165,
|
| 304 |
+
"eval_steps_per_second": 11.882,
|
| 305 |
+
"step": 360
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 7.55,
|
| 309 |
+
"learning_rate": 0.00012448979591836735,
|
| 310 |
+
"loss": 0.1801,
|
| 311 |
+
"step": 370
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 7.76,
|
| 315 |
+
"learning_rate": 0.00012244897959183676,
|
| 316 |
+
"loss": 0.1785,
|
| 317 |
+
"step": 380
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 7.96,
|
| 321 |
+
"learning_rate": 0.00012040816326530613,
|
| 322 |
+
"loss": 0.1627,
|
| 323 |
+
"step": 390
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 8.16,
|
| 327 |
+
"learning_rate": 0.00011836734693877552,
|
| 328 |
+
"loss": 0.154,
|
| 329 |
+
"step": 400
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 8.16,
|
| 333 |
+
"eval_accuracy": 0.8138801261829653,
|
| 334 |
+
"eval_loss": 0.945625364780426,
|
| 335 |
+
"eval_runtime": 3.3591,
|
| 336 |
+
"eval_samples_per_second": 94.37,
|
| 337 |
+
"eval_steps_per_second": 11.908,
|
| 338 |
+
"step": 400
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 8.37,
|
| 342 |
+
"learning_rate": 0.0001163265306122449,
|
| 343 |
+
"loss": 0.1268,
|
| 344 |
+
"step": 410
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 8.57,
|
| 348 |
+
"learning_rate": 0.00011428571428571428,
|
| 349 |
+
"loss": 0.1352,
|
| 350 |
+
"step": 420
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 8.78,
|
| 354 |
+
"learning_rate": 0.00011224489795918367,
|
| 355 |
+
"loss": 0.1214,
|
| 356 |
+
"step": 430
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 8.98,
|
| 360 |
+
"learning_rate": 0.00011020408163265306,
|
| 361 |
+
"loss": 0.1519,
|
| 362 |
+
"step": 440
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 8.98,
|
| 366 |
+
"eval_accuracy": 0.8454258675078864,
|
| 367 |
+
"eval_loss": 0.8139908313751221,
|
| 368 |
+
"eval_runtime": 4.5811,
|
| 369 |
+
"eval_samples_per_second": 69.197,
|
| 370 |
+
"eval_steps_per_second": 8.731,
|
| 371 |
+
"step": 440
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 9.18,
|
| 375 |
+
"learning_rate": 0.00010816326530612246,
|
| 376 |
+
"loss": 0.0773,
|
| 377 |
+
"step": 450
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 9.39,
|
| 381 |
+
"learning_rate": 0.00010612244897959185,
|
| 382 |
+
"loss": 0.1083,
|
| 383 |
+
"step": 460
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 9.59,
|
| 387 |
+
"learning_rate": 0.00010408163265306123,
|
| 388 |
+
"loss": 0.086,
|
| 389 |
+
"step": 470
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 9.8,
|
| 393 |
+
"learning_rate": 0.00010204081632653062,
|
| 394 |
+
"loss": 0.0672,
|
| 395 |
+
"step": 480
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 9.8,
|
| 399 |
+
"eval_accuracy": 0.8517350157728707,
|
| 400 |
+
"eval_loss": 0.7407823204994202,
|
| 401 |
+
"eval_runtime": 3.8145,
|
| 402 |
+
"eval_samples_per_second": 83.104,
|
| 403 |
+
"eval_steps_per_second": 10.486,
|
| 404 |
+
"step": 480
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 10.0,
|
| 408 |
+
"learning_rate": 0.0001,
|
| 409 |
+
"loss": 0.067,
|
| 410 |
+
"step": 490
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 10.2,
|
| 414 |
+
"learning_rate": 9.79591836734694e-05,
|
| 415 |
+
"loss": 0.0559,
|
| 416 |
+
"step": 500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 10.41,
|
| 420 |
+
"learning_rate": 9.591836734693878e-05,
|
| 421 |
+
"loss": 0.055,
|
| 422 |
+
"step": 510
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 10.61,
|
| 426 |
+
"learning_rate": 9.387755102040817e-05,
|
| 427 |
+
"loss": 0.0498,
|
| 428 |
+
"step": 520
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 10.61,
|
| 432 |
+
"eval_accuracy": 0.8580441640378549,
|
| 433 |
+
"eval_loss": 0.7550804615020752,
|
| 434 |
+
"eval_runtime": 3.3075,
|
| 435 |
+
"eval_samples_per_second": 95.842,
|
| 436 |
+
"eval_steps_per_second": 12.094,
|
| 437 |
+
"step": 520
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 10.82,
|
| 441 |
+
"learning_rate": 9.183673469387756e-05,
|
| 442 |
+
"loss": 0.0483,
|
| 443 |
+
"step": 530
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 11.02,
|
| 447 |
+
"learning_rate": 8.979591836734695e-05,
|
| 448 |
+
"loss": 0.0461,
|
| 449 |
+
"step": 540
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 11.22,
|
| 453 |
+
"learning_rate": 8.775510204081632e-05,
|
| 454 |
+
"loss": 0.043,
|
| 455 |
+
"step": 550
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 11.43,
|
| 459 |
+
"learning_rate": 8.571428571428571e-05,
|
| 460 |
+
"loss": 0.0618,
|
| 461 |
+
"step": 560
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 11.43,
|
| 465 |
+
"eval_accuracy": 0.8675078864353313,
|
| 466 |
+
"eval_loss": 0.7529485821723938,
|
| 467 |
+
"eval_runtime": 3.2693,
|
| 468 |
+
"eval_samples_per_second": 96.964,
|
| 469 |
+
"eval_steps_per_second": 12.235,
|
| 470 |
+
"step": 560
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 11.63,
|
| 474 |
+
"learning_rate": 8.367346938775511e-05,
|
| 475 |
+
"loss": 0.0409,
|
| 476 |
+
"step": 570
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 11.84,
|
| 480 |
+
"learning_rate": 8.163265306122449e-05,
|
| 481 |
+
"loss": 0.0433,
|
| 482 |
+
"step": 580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 12.04,
|
| 486 |
+
"learning_rate": 7.959183673469388e-05,
|
| 487 |
+
"loss": 0.0367,
|
| 488 |
+
"step": 590
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 12.24,
|
| 492 |
+
"learning_rate": 7.755102040816327e-05,
|
| 493 |
+
"loss": 0.0352,
|
| 494 |
+
"step": 600
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 12.24,
|
| 498 |
+
"eval_accuracy": 0.8643533123028391,
|
| 499 |
+
"eval_loss": 0.7547315955162048,
|
| 500 |
+
"eval_runtime": 4.666,
|
| 501 |
+
"eval_samples_per_second": 67.938,
|
| 502 |
+
"eval_steps_per_second": 8.573,
|
| 503 |
+
"step": 600
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 12.45,
|
| 507 |
+
"learning_rate": 7.551020408163266e-05,
|
| 508 |
+
"loss": 0.0377,
|
| 509 |
+
"step": 610
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 12.65,
|
| 513 |
+
"learning_rate": 7.346938775510205e-05,
|
| 514 |
+
"loss": 0.0331,
|
| 515 |
+
"step": 620
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 12.86,
|
| 519 |
+
"learning_rate": 7.142857142857143e-05,
|
| 520 |
+
"loss": 0.0678,
|
| 521 |
+
"step": 630
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 13.06,
|
| 525 |
+
"learning_rate": 6.938775510204082e-05,
|
| 526 |
+
"loss": 0.0381,
|
| 527 |
+
"step": 640
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 13.06,
|
| 531 |
+
"eval_accuracy": 0.8769716088328076,
|
| 532 |
+
"eval_loss": 0.7314792275428772,
|
| 533 |
+
"eval_runtime": 4.0948,
|
| 534 |
+
"eval_samples_per_second": 77.416,
|
| 535 |
+
"eval_steps_per_second": 9.769,
|
| 536 |
+
"step": 640
|
| 537 |
+
}
|
| 538 |
+
],
|
| 539 |
+
"max_steps": 980,
|
| 540 |
+
"num_train_epochs": 20,
|
| 541 |
+
"total_flos": 7.79603043959083e+17,
|
| 542 |
+
"trial_name": null,
|
| 543 |
+
"trial_params": null
|
| 544 |
+
}
|
checkpoint-640/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88e90154845df1cbf20cec744411d8e64005f30ae741e837b432941d036a7976
|
| 3 |
+
size 3899
|
checkpoint-960/config.json
ADDED
|
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "%3F",
|
| 13 |
+
"1": "a",
|
| 14 |
+
"10": "i",
|
| 15 |
+
"11": "j",
|
| 16 |
+
"12": "k",
|
| 17 |
+
"13": "l",
|
| 18 |
+
"14": "m",
|
| 19 |
+
"15": "n",
|
| 20 |
+
"16": "o",
|
| 21 |
+
"17": "p",
|
| 22 |
+
"18": "period",
|
| 23 |
+
"19": "q",
|
| 24 |
+
"2": "b",
|
| 25 |
+
"20": "r",
|
| 26 |
+
"21": "s",
|
| 27 |
+
"22": "t",
|
| 28 |
+
"23": "u",
|
| 29 |
+
"24": "v",
|
| 30 |
+
"25": "w",
|
| 31 |
+
"26": "x",
|
| 32 |
+
"27": "y",
|
| 33 |
+
"28": "z",
|
| 34 |
+
"3": "c",
|
| 35 |
+
"4": "capital",
|
| 36 |
+
"5": "d",
|
| 37 |
+
"6": "e",
|
| 38 |
+
"7": "f",
|
| 39 |
+
"8": "g",
|
| 40 |
+
"9": "h"
|
| 41 |
+
},
|
| 42 |
+
"image_size": 224,
|
| 43 |
+
"initializer_range": 0.02,
|
| 44 |
+
"intermediate_size": 3072,
|
| 45 |
+
"label2id": {
|
| 46 |
+
"%3F": "0",
|
| 47 |
+
"a": "1",
|
| 48 |
+
"b": "2",
|
| 49 |
+
"c": "3",
|
| 50 |
+
"capital": "4",
|
| 51 |
+
"d": "5",
|
| 52 |
+
"e": "6",
|
| 53 |
+
"f": "7",
|
| 54 |
+
"g": "8",
|
| 55 |
+
"h": "9",
|
| 56 |
+
"i": "10",
|
| 57 |
+
"j": "11",
|
| 58 |
+
"k": "12",
|
| 59 |
+
"l": "13",
|
| 60 |
+
"m": "14",
|
| 61 |
+
"n": "15",
|
| 62 |
+
"o": "16",
|
| 63 |
+
"p": "17",
|
| 64 |
+
"period": "18",
|
| 65 |
+
"q": "19",
|
| 66 |
+
"r": "20",
|
| 67 |
+
"s": "21",
|
| 68 |
+
"t": "22",
|
| 69 |
+
"u": "23",
|
| 70 |
+
"v": "24",
|
| 71 |
+
"w": "25",
|
| 72 |
+
"x": "26",
|
| 73 |
+
"y": "27",
|
| 74 |
+
"z": "28"
|
| 75 |
+
},
|
| 76 |
+
"layer_norm_eps": 1e-12,
|
| 77 |
+
"model_type": "vit",
|
| 78 |
+
"num_attention_heads": 12,
|
| 79 |
+
"num_channels": 3,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"patch_size": 16,
|
| 82 |
+
"problem_type": "single_label_classification",
|
| 83 |
+
"qkv_bias": true,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.30.2"
|
| 86 |
+
}
|
checkpoint-960/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93c2738d89b89a1abfbde3c1c2fb2e980ac40a6a34e0532026eb9a398a756896
|
| 3 |
+
size 686684933
|
checkpoint-960/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-960/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:402624e75387dcc709da998ce3c18a6c05f1a7b6429efdcd837c29e061483b3d
|
| 3 |
+
size 343351725
|
checkpoint-960/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71a159216a00b176c11c09f939dc12b1698e9ab77e7f077134102c94d7e18557
|
| 3 |
+
size 14575
|
checkpoint-960/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e6bd41fba6957f6586433ed4a0afedba9187c02bf67b3a6e89b03bd2663dc33
|
| 3 |
+
size 627
|
checkpoint-960/trainer_state.json
ADDED
|
@@ -0,0 +1,808 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7314792275428772,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-beans/checkpoint-640",
|
| 4 |
+
"epoch": 19.591836734693878,
|
| 5 |
+
"global_step": 960,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.2,
|
| 12 |
+
"learning_rate": 0.00019795918367346938,
|
| 13 |
+
"loss": 3.374,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.41,
|
| 18 |
+
"learning_rate": 0.0001959183673469388,
|
| 19 |
+
"loss": 3.3572,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.61,
|
| 24 |
+
"learning_rate": 0.00019387755102040816,
|
| 25 |
+
"loss": 3.3211,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.82,
|
| 30 |
+
"learning_rate": 0.00019183673469387756,
|
| 31 |
+
"loss": 3.2257,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.82,
|
| 36 |
+
"eval_accuracy": 0.03470031545741325,
|
| 37 |
+
"eval_loss": 3.4292807579040527,
|
| 38 |
+
"eval_runtime": 3.7472,
|
| 39 |
+
"eval_samples_per_second": 84.597,
|
| 40 |
+
"eval_steps_per_second": 10.675,
|
| 41 |
+
"step": 40
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.02,
|
| 45 |
+
"learning_rate": 0.00018979591836734697,
|
| 46 |
+
"loss": 3.1143,
|
| 47 |
+
"step": 50
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 1.22,
|
| 51 |
+
"learning_rate": 0.00018775510204081634,
|
| 52 |
+
"loss": 3.032,
|
| 53 |
+
"step": 60
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.43,
|
| 57 |
+
"learning_rate": 0.00018571428571428572,
|
| 58 |
+
"loss": 2.8104,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.63,
|
| 63 |
+
"learning_rate": 0.00018367346938775512,
|
| 64 |
+
"loss": 2.6674,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.63,
|
| 69 |
+
"eval_accuracy": 0.17350157728706625,
|
| 70 |
+
"eval_loss": 2.952009916305542,
|
| 71 |
+
"eval_runtime": 3.3938,
|
| 72 |
+
"eval_samples_per_second": 93.405,
|
| 73 |
+
"eval_steps_per_second": 11.786,
|
| 74 |
+
"step": 80
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.84,
|
| 78 |
+
"learning_rate": 0.0001816326530612245,
|
| 79 |
+
"loss": 2.5131,
|
| 80 |
+
"step": 90
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 2.04,
|
| 84 |
+
"learning_rate": 0.0001795918367346939,
|
| 85 |
+
"loss": 2.5287,
|
| 86 |
+
"step": 100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 2.24,
|
| 90 |
+
"learning_rate": 0.00017755102040816327,
|
| 91 |
+
"loss": 2.2121,
|
| 92 |
+
"step": 110
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 2.45,
|
| 96 |
+
"learning_rate": 0.00017551020408163265,
|
| 97 |
+
"loss": 2.0048,
|
| 98 |
+
"step": 120
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 2.45,
|
| 102 |
+
"eval_accuracy": 0.48580441640378547,
|
| 103 |
+
"eval_loss": 2.363018035888672,
|
| 104 |
+
"eval_runtime": 3.7903,
|
| 105 |
+
"eval_samples_per_second": 83.635,
|
| 106 |
+
"eval_steps_per_second": 10.553,
|
| 107 |
+
"step": 120
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 2.65,
|
| 111 |
+
"learning_rate": 0.00017346938775510205,
|
| 112 |
+
"loss": 1.8413,
|
| 113 |
+
"step": 130
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.86,
|
| 117 |
+
"learning_rate": 0.00017142857142857143,
|
| 118 |
+
"loss": 1.7897,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 3.06,
|
| 123 |
+
"learning_rate": 0.00016938775510204083,
|
| 124 |
+
"loss": 1.607,
|
| 125 |
+
"step": 150
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 3.27,
|
| 129 |
+
"learning_rate": 0.00016734693877551023,
|
| 130 |
+
"loss": 1.3493,
|
| 131 |
+
"step": 160
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 3.27,
|
| 135 |
+
"eval_accuracy": 0.5488958990536278,
|
| 136 |
+
"eval_loss": 1.8864167928695679,
|
| 137 |
+
"eval_runtime": 4.2623,
|
| 138 |
+
"eval_samples_per_second": 74.372,
|
| 139 |
+
"eval_steps_per_second": 9.385,
|
| 140 |
+
"step": 160
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.47,
|
| 144 |
+
"learning_rate": 0.0001653061224489796,
|
| 145 |
+
"loss": 1.2421,
|
| 146 |
+
"step": 170
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.67,
|
| 150 |
+
"learning_rate": 0.00016326530612244898,
|
| 151 |
+
"loss": 1.0779,
|
| 152 |
+
"step": 180
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.88,
|
| 156 |
+
"learning_rate": 0.00016122448979591838,
|
| 157 |
+
"loss": 1.0038,
|
| 158 |
+
"step": 190
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 4.08,
|
| 162 |
+
"learning_rate": 0.00015918367346938776,
|
| 163 |
+
"loss": 1.0887,
|
| 164 |
+
"step": 200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.08,
|
| 168 |
+
"eval_accuracy": 0.7350157728706624,
|
| 169 |
+
"eval_loss": 1.4275710582733154,
|
| 170 |
+
"eval_runtime": 3.3105,
|
| 171 |
+
"eval_samples_per_second": 95.757,
|
| 172 |
+
"eval_steps_per_second": 12.083,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 4.29,
|
| 177 |
+
"learning_rate": 0.00015714285714285716,
|
| 178 |
+
"loss": 0.7734,
|
| 179 |
+
"step": 210
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 4.49,
|
| 183 |
+
"learning_rate": 0.00015510204081632654,
|
| 184 |
+
"loss": 0.7969,
|
| 185 |
+
"step": 220
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.69,
|
| 189 |
+
"learning_rate": 0.0001530612244897959,
|
| 190 |
+
"loss": 0.7953,
|
| 191 |
+
"step": 230
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 4.9,
|
| 195 |
+
"learning_rate": 0.0001510204081632653,
|
| 196 |
+
"loss": 0.6649,
|
| 197 |
+
"step": 240
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 4.9,
|
| 201 |
+
"eval_accuracy": 0.7728706624605678,
|
| 202 |
+
"eval_loss": 1.2706444263458252,
|
| 203 |
+
"eval_runtime": 4.1354,
|
| 204 |
+
"eval_samples_per_second": 76.656,
|
| 205 |
+
"eval_steps_per_second": 9.673,
|
| 206 |
+
"step": 240
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 5.1,
|
| 210 |
+
"learning_rate": 0.00014897959183673472,
|
| 211 |
+
"loss": 0.5647,
|
| 212 |
+
"step": 250
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 5.31,
|
| 216 |
+
"learning_rate": 0.0001469387755102041,
|
| 217 |
+
"loss": 0.7201,
|
| 218 |
+
"step": 260
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 5.51,
|
| 222 |
+
"learning_rate": 0.0001448979591836735,
|
| 223 |
+
"loss": 0.606,
|
| 224 |
+
"step": 270
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 5.71,
|
| 228 |
+
"learning_rate": 0.00014285714285714287,
|
| 229 |
+
"loss": 0.5396,
|
| 230 |
+
"step": 280
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 5.71,
|
| 234 |
+
"eval_accuracy": 0.7981072555205048,
|
| 235 |
+
"eval_loss": 1.125301480293274,
|
| 236 |
+
"eval_runtime": 3.3532,
|
| 237 |
+
"eval_samples_per_second": 94.537,
|
| 238 |
+
"eval_steps_per_second": 11.929,
|
| 239 |
+
"step": 280
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 5.92,
|
| 243 |
+
"learning_rate": 0.00014081632653061224,
|
| 244 |
+
"loss": 0.3407,
|
| 245 |
+
"step": 290
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 6.12,
|
| 249 |
+
"learning_rate": 0.00013877551020408165,
|
| 250 |
+
"loss": 0.3539,
|
| 251 |
+
"step": 300
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 6.33,
|
| 255 |
+
"learning_rate": 0.00013673469387755102,
|
| 256 |
+
"loss": 0.3608,
|
| 257 |
+
"step": 310
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 6.53,
|
| 261 |
+
"learning_rate": 0.0001346938775510204,
|
| 262 |
+
"loss": 0.3162,
|
| 263 |
+
"step": 320
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 6.53,
|
| 267 |
+
"eval_accuracy": 0.804416403785489,
|
| 268 |
+
"eval_loss": 1.0156543254852295,
|
| 269 |
+
"eval_runtime": 4.7629,
|
| 270 |
+
"eval_samples_per_second": 66.556,
|
| 271 |
+
"eval_steps_per_second": 8.398,
|
| 272 |
+
"step": 320
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.73,
|
| 276 |
+
"learning_rate": 0.0001326530612244898,
|
| 277 |
+
"loss": 0.335,
|
| 278 |
+
"step": 330
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 6.94,
|
| 282 |
+
"learning_rate": 0.00013061224489795917,
|
| 283 |
+
"loss": 0.2333,
|
| 284 |
+
"step": 340
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 7.14,
|
| 288 |
+
"learning_rate": 0.00012857142857142858,
|
| 289 |
+
"loss": 0.1931,
|
| 290 |
+
"step": 350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 7.35,
|
| 294 |
+
"learning_rate": 0.00012653061224489798,
|
| 295 |
+
"loss": 0.1909,
|
| 296 |
+
"step": 360
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 7.35,
|
| 300 |
+
"eval_accuracy": 0.8548895899053628,
|
| 301 |
+
"eval_loss": 0.8516904711723328,
|
| 302 |
+
"eval_runtime": 3.3664,
|
| 303 |
+
"eval_samples_per_second": 94.165,
|
| 304 |
+
"eval_steps_per_second": 11.882,
|
| 305 |
+
"step": 360
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 7.55,
|
| 309 |
+
"learning_rate": 0.00012448979591836735,
|
| 310 |
+
"loss": 0.1801,
|
| 311 |
+
"step": 370
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 7.76,
|
| 315 |
+
"learning_rate": 0.00012244897959183676,
|
| 316 |
+
"loss": 0.1785,
|
| 317 |
+
"step": 380
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 7.96,
|
| 321 |
+
"learning_rate": 0.00012040816326530613,
|
| 322 |
+
"loss": 0.1627,
|
| 323 |
+
"step": 390
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 8.16,
|
| 327 |
+
"learning_rate": 0.00011836734693877552,
|
| 328 |
+
"loss": 0.154,
|
| 329 |
+
"step": 400
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 8.16,
|
| 333 |
+
"eval_accuracy": 0.8138801261829653,
|
| 334 |
+
"eval_loss": 0.945625364780426,
|
| 335 |
+
"eval_runtime": 3.3591,
|
| 336 |
+
"eval_samples_per_second": 94.37,
|
| 337 |
+
"eval_steps_per_second": 11.908,
|
| 338 |
+
"step": 400
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 8.37,
|
| 342 |
+
"learning_rate": 0.0001163265306122449,
|
| 343 |
+
"loss": 0.1268,
|
| 344 |
+
"step": 410
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 8.57,
|
| 348 |
+
"learning_rate": 0.00011428571428571428,
|
| 349 |
+
"loss": 0.1352,
|
| 350 |
+
"step": 420
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 8.78,
|
| 354 |
+
"learning_rate": 0.00011224489795918367,
|
| 355 |
+
"loss": 0.1214,
|
| 356 |
+
"step": 430
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 8.98,
|
| 360 |
+
"learning_rate": 0.00011020408163265306,
|
| 361 |
+
"loss": 0.1519,
|
| 362 |
+
"step": 440
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 8.98,
|
| 366 |
+
"eval_accuracy": 0.8454258675078864,
|
| 367 |
+
"eval_loss": 0.8139908313751221,
|
| 368 |
+
"eval_runtime": 4.5811,
|
| 369 |
+
"eval_samples_per_second": 69.197,
|
| 370 |
+
"eval_steps_per_second": 8.731,
|
| 371 |
+
"step": 440
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 9.18,
|
| 375 |
+
"learning_rate": 0.00010816326530612246,
|
| 376 |
+
"loss": 0.0773,
|
| 377 |
+
"step": 450
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 9.39,
|
| 381 |
+
"learning_rate": 0.00010612244897959185,
|
| 382 |
+
"loss": 0.1083,
|
| 383 |
+
"step": 460
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 9.59,
|
| 387 |
+
"learning_rate": 0.00010408163265306123,
|
| 388 |
+
"loss": 0.086,
|
| 389 |
+
"step": 470
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 9.8,
|
| 393 |
+
"learning_rate": 0.00010204081632653062,
|
| 394 |
+
"loss": 0.0672,
|
| 395 |
+
"step": 480
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 9.8,
|
| 399 |
+
"eval_accuracy": 0.8517350157728707,
|
| 400 |
+
"eval_loss": 0.7407823204994202,
|
| 401 |
+
"eval_runtime": 3.8145,
|
| 402 |
+
"eval_samples_per_second": 83.104,
|
| 403 |
+
"eval_steps_per_second": 10.486,
|
| 404 |
+
"step": 480
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 10.0,
|
| 408 |
+
"learning_rate": 0.0001,
|
| 409 |
+
"loss": 0.067,
|
| 410 |
+
"step": 490
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 10.2,
|
| 414 |
+
"learning_rate": 9.79591836734694e-05,
|
| 415 |
+
"loss": 0.0559,
|
| 416 |
+
"step": 500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 10.41,
|
| 420 |
+
"learning_rate": 9.591836734693878e-05,
|
| 421 |
+
"loss": 0.055,
|
| 422 |
+
"step": 510
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 10.61,
|
| 426 |
+
"learning_rate": 9.387755102040817e-05,
|
| 427 |
+
"loss": 0.0498,
|
| 428 |
+
"step": 520
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 10.61,
|
| 432 |
+
"eval_accuracy": 0.8580441640378549,
|
| 433 |
+
"eval_loss": 0.7550804615020752,
|
| 434 |
+
"eval_runtime": 3.3075,
|
| 435 |
+
"eval_samples_per_second": 95.842,
|
| 436 |
+
"eval_steps_per_second": 12.094,
|
| 437 |
+
"step": 520
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 10.82,
|
| 441 |
+
"learning_rate": 9.183673469387756e-05,
|
| 442 |
+
"loss": 0.0483,
|
| 443 |
+
"step": 530
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 11.02,
|
| 447 |
+
"learning_rate": 8.979591836734695e-05,
|
| 448 |
+
"loss": 0.0461,
|
| 449 |
+
"step": 540
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 11.22,
|
| 453 |
+
"learning_rate": 8.775510204081632e-05,
|
| 454 |
+
"loss": 0.043,
|
| 455 |
+
"step": 550
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 11.43,
|
| 459 |
+
"learning_rate": 8.571428571428571e-05,
|
| 460 |
+
"loss": 0.0618,
|
| 461 |
+
"step": 560
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 11.43,
|
| 465 |
+
"eval_accuracy": 0.8675078864353313,
|
| 466 |
+
"eval_loss": 0.7529485821723938,
|
| 467 |
+
"eval_runtime": 3.2693,
|
| 468 |
+
"eval_samples_per_second": 96.964,
|
| 469 |
+
"eval_steps_per_second": 12.235,
|
| 470 |
+
"step": 560
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 11.63,
|
| 474 |
+
"learning_rate": 8.367346938775511e-05,
|
| 475 |
+
"loss": 0.0409,
|
| 476 |
+
"step": 570
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 11.84,
|
| 480 |
+
"learning_rate": 8.163265306122449e-05,
|
| 481 |
+
"loss": 0.0433,
|
| 482 |
+
"step": 580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 12.04,
|
| 486 |
+
"learning_rate": 7.959183673469388e-05,
|
| 487 |
+
"loss": 0.0367,
|
| 488 |
+
"step": 590
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 12.24,
|
| 492 |
+
"learning_rate": 7.755102040816327e-05,
|
| 493 |
+
"loss": 0.0352,
|
| 494 |
+
"step": 600
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 12.24,
|
| 498 |
+
"eval_accuracy": 0.8643533123028391,
|
| 499 |
+
"eval_loss": 0.7547315955162048,
|
| 500 |
+
"eval_runtime": 4.666,
|
| 501 |
+
"eval_samples_per_second": 67.938,
|
| 502 |
+
"eval_steps_per_second": 8.573,
|
| 503 |
+
"step": 600
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 12.45,
|
| 507 |
+
"learning_rate": 7.551020408163266e-05,
|
| 508 |
+
"loss": 0.0377,
|
| 509 |
+
"step": 610
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 12.65,
|
| 513 |
+
"learning_rate": 7.346938775510205e-05,
|
| 514 |
+
"loss": 0.0331,
|
| 515 |
+
"step": 620
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 12.86,
|
| 519 |
+
"learning_rate": 7.142857142857143e-05,
|
| 520 |
+
"loss": 0.0678,
|
| 521 |
+
"step": 630
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 13.06,
|
| 525 |
+
"learning_rate": 6.938775510204082e-05,
|
| 526 |
+
"loss": 0.0381,
|
| 527 |
+
"step": 640
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 13.06,
|
| 531 |
+
"eval_accuracy": 0.8769716088328076,
|
| 532 |
+
"eval_loss": 0.7314792275428772,
|
| 533 |
+
"eval_runtime": 4.0948,
|
| 534 |
+
"eval_samples_per_second": 77.416,
|
| 535 |
+
"eval_steps_per_second": 9.769,
|
| 536 |
+
"step": 640
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 13.27,
|
| 540 |
+
"learning_rate": 6.73469387755102e-05,
|
| 541 |
+
"loss": 0.0316,
|
| 542 |
+
"step": 650
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 13.47,
|
| 546 |
+
"learning_rate": 6.530612244897959e-05,
|
| 547 |
+
"loss": 0.0302,
|
| 548 |
+
"step": 660
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 13.67,
|
| 552 |
+
"learning_rate": 6.326530612244899e-05,
|
| 553 |
+
"loss": 0.0317,
|
| 554 |
+
"step": 670
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 13.88,
|
| 558 |
+
"learning_rate": 6.122448979591838e-05,
|
| 559 |
+
"loss": 0.0288,
|
| 560 |
+
"step": 680
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 13.88,
|
| 564 |
+
"eval_accuracy": 0.8675078864353313,
|
| 565 |
+
"eval_loss": 0.7384619116783142,
|
| 566 |
+
"eval_runtime": 4.3527,
|
| 567 |
+
"eval_samples_per_second": 72.828,
|
| 568 |
+
"eval_steps_per_second": 9.19,
|
| 569 |
+
"step": 680
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 14.08,
|
| 573 |
+
"learning_rate": 5.918367346938776e-05,
|
| 574 |
+
"loss": 0.0285,
|
| 575 |
+
"step": 690
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 14.29,
|
| 579 |
+
"learning_rate": 5.714285714285714e-05,
|
| 580 |
+
"loss": 0.0271,
|
| 581 |
+
"step": 700
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 14.49,
|
| 585 |
+
"learning_rate": 5.510204081632653e-05,
|
| 586 |
+
"loss": 0.0264,
|
| 587 |
+
"step": 710
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 14.69,
|
| 591 |
+
"learning_rate": 5.3061224489795926e-05,
|
| 592 |
+
"loss": 0.0264,
|
| 593 |
+
"step": 720
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 14.69,
|
| 597 |
+
"eval_accuracy": 0.8643533123028391,
|
| 598 |
+
"eval_loss": 0.7676613926887512,
|
| 599 |
+
"eval_runtime": 3.2589,
|
| 600 |
+
"eval_samples_per_second": 97.271,
|
| 601 |
+
"eval_steps_per_second": 12.274,
|
| 602 |
+
"step": 720
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 14.9,
|
| 606 |
+
"learning_rate": 5.102040816326531e-05,
|
| 607 |
+
"loss": 0.0258,
|
| 608 |
+
"step": 730
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 15.1,
|
| 612 |
+
"learning_rate": 4.89795918367347e-05,
|
| 613 |
+
"loss": 0.0256,
|
| 614 |
+
"step": 740
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 15.31,
|
| 618 |
+
"learning_rate": 4.6938775510204086e-05,
|
| 619 |
+
"loss": 0.0242,
|
| 620 |
+
"step": 750
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 15.51,
|
| 624 |
+
"learning_rate": 4.4897959183673474e-05,
|
| 625 |
+
"loss": 0.0243,
|
| 626 |
+
"step": 760
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 15.51,
|
| 630 |
+
"eval_accuracy": 0.8580441640378549,
|
| 631 |
+
"eval_loss": 0.7726181745529175,
|
| 632 |
+
"eval_runtime": 3.8096,
|
| 633 |
+
"eval_samples_per_second": 83.211,
|
| 634 |
+
"eval_steps_per_second": 10.5,
|
| 635 |
+
"step": 760
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 15.71,
|
| 639 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 640 |
+
"loss": 0.0234,
|
| 641 |
+
"step": 770
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 15.92,
|
| 645 |
+
"learning_rate": 4.0816326530612245e-05,
|
| 646 |
+
"loss": 0.0241,
|
| 647 |
+
"step": 780
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 16.12,
|
| 651 |
+
"learning_rate": 3.8775510204081634e-05,
|
| 652 |
+
"loss": 0.0227,
|
| 653 |
+
"step": 790
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 16.33,
|
| 657 |
+
"learning_rate": 3.673469387755102e-05,
|
| 658 |
+
"loss": 0.0228,
|
| 659 |
+
"step": 800
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 16.33,
|
| 663 |
+
"eval_accuracy": 0.8580441640378549,
|
| 664 |
+
"eval_loss": 0.7750853300094604,
|
| 665 |
+
"eval_runtime": 3.2592,
|
| 666 |
+
"eval_samples_per_second": 97.263,
|
| 667 |
+
"eval_steps_per_second": 12.273,
|
| 668 |
+
"step": 800
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 16.53,
|
| 672 |
+
"learning_rate": 3.469387755102041e-05,
|
| 673 |
+
"loss": 0.0236,
|
| 674 |
+
"step": 810
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 16.73,
|
| 678 |
+
"learning_rate": 3.265306122448979e-05,
|
| 679 |
+
"loss": 0.023,
|
| 680 |
+
"step": 820
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 16.94,
|
| 684 |
+
"learning_rate": 3.061224489795919e-05,
|
| 685 |
+
"loss": 0.0219,
|
| 686 |
+
"step": 830
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 17.14,
|
| 690 |
+
"learning_rate": 2.857142857142857e-05,
|
| 691 |
+
"loss": 0.0219,
|
| 692 |
+
"step": 840
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 17.14,
|
| 696 |
+
"eval_accuracy": 0.8643533123028391,
|
| 697 |
+
"eval_loss": 0.7764984369277954,
|
| 698 |
+
"eval_runtime": 3.4162,
|
| 699 |
+
"eval_samples_per_second": 92.793,
|
| 700 |
+
"eval_steps_per_second": 11.709,
|
| 701 |
+
"step": 840
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 17.35,
|
| 705 |
+
"learning_rate": 2.6530612244897963e-05,
|
| 706 |
+
"loss": 0.0214,
|
| 707 |
+
"step": 850
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 17.55,
|
| 711 |
+
"learning_rate": 2.448979591836735e-05,
|
| 712 |
+
"loss": 0.0219,
|
| 713 |
+
"step": 860
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 17.76,
|
| 717 |
+
"learning_rate": 2.2448979591836737e-05,
|
| 718 |
+
"loss": 0.0207,
|
| 719 |
+
"step": 870
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 17.96,
|
| 723 |
+
"learning_rate": 2.0408163265306123e-05,
|
| 724 |
+
"loss": 0.0218,
|
| 725 |
+
"step": 880
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 17.96,
|
| 729 |
+
"eval_accuracy": 0.8643533123028391,
|
| 730 |
+
"eval_loss": 0.7789060473442078,
|
| 731 |
+
"eval_runtime": 3.3084,
|
| 732 |
+
"eval_samples_per_second": 95.816,
|
| 733 |
+
"eval_steps_per_second": 12.09,
|
| 734 |
+
"step": 880
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 18.16,
|
| 738 |
+
"learning_rate": 1.836734693877551e-05,
|
| 739 |
+
"loss": 0.0219,
|
| 740 |
+
"step": 890
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 18.37,
|
| 744 |
+
"learning_rate": 1.6326530612244897e-05,
|
| 745 |
+
"loss": 0.0214,
|
| 746 |
+
"step": 900
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 18.57,
|
| 750 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 751 |
+
"loss": 0.021,
|
| 752 |
+
"step": 910
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 18.78,
|
| 756 |
+
"learning_rate": 1.2244897959183674e-05,
|
| 757 |
+
"loss": 0.0209,
|
| 758 |
+
"step": 920
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 18.78,
|
| 762 |
+
"eval_accuracy": 0.8643533123028391,
|
| 763 |
+
"eval_loss": 0.779566764831543,
|
| 764 |
+
"eval_runtime": 4.6967,
|
| 765 |
+
"eval_samples_per_second": 67.494,
|
| 766 |
+
"eval_steps_per_second": 8.517,
|
| 767 |
+
"step": 920
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 18.98,
|
| 771 |
+
"learning_rate": 1.0204081632653061e-05,
|
| 772 |
+
"loss": 0.0204,
|
| 773 |
+
"step": 930
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 19.18,
|
| 777 |
+
"learning_rate": 8.163265306122448e-06,
|
| 778 |
+
"loss": 0.0209,
|
| 779 |
+
"step": 940
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 19.39,
|
| 783 |
+
"learning_rate": 6.122448979591837e-06,
|
| 784 |
+
"loss": 0.0207,
|
| 785 |
+
"step": 950
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 19.59,
|
| 789 |
+
"learning_rate": 4.081632653061224e-06,
|
| 790 |
+
"loss": 0.0206,
|
| 791 |
+
"step": 960
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 19.59,
|
| 795 |
+
"eval_accuracy": 0.8643533123028391,
|
| 796 |
+
"eval_loss": 0.7802363634109497,
|
| 797 |
+
"eval_runtime": 3.3003,
|
| 798 |
+
"eval_samples_per_second": 96.051,
|
| 799 |
+
"eval_steps_per_second": 12.12,
|
| 800 |
+
"step": 960
|
| 801 |
+
}
|
| 802 |
+
],
|
| 803 |
+
"max_steps": 980,
|
| 804 |
+
"num_train_epochs": 20,
|
| 805 |
+
"total_flos": 1.169947141133267e+18,
|
| 806 |
+
"trial_name": null,
|
| 807 |
+
"trial_params": null
|
| 808 |
+
}
|
checkpoint-960/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88e90154845df1cbf20cec744411d8e64005f30ae741e837b432941d036a7976
|
| 3 |
+
size 3899
|
config.json
ADDED
|
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "%3F",
|
| 13 |
+
"1": "a",
|
| 14 |
+
"10": "i",
|
| 15 |
+
"11": "j",
|
| 16 |
+
"12": "k",
|
| 17 |
+
"13": "l",
|
| 18 |
+
"14": "m",
|
| 19 |
+
"15": "n",
|
| 20 |
+
"16": "o",
|
| 21 |
+
"17": "p",
|
| 22 |
+
"18": "period",
|
| 23 |
+
"19": "q",
|
| 24 |
+
"2": "b",
|
| 25 |
+
"20": "r",
|
| 26 |
+
"21": "s",
|
| 27 |
+
"22": "t",
|
| 28 |
+
"23": "u",
|
| 29 |
+
"24": "v",
|
| 30 |
+
"25": "w",
|
| 31 |
+
"26": "x",
|
| 32 |
+
"27": "y",
|
| 33 |
+
"28": "z",
|
| 34 |
+
"3": "c",
|
| 35 |
+
"4": "capital",
|
| 36 |
+
"5": "d",
|
| 37 |
+
"6": "e",
|
| 38 |
+
"7": "f",
|
| 39 |
+
"8": "g",
|
| 40 |
+
"9": "h"
|
| 41 |
+
},
|
| 42 |
+
"image_size": 224,
|
| 43 |
+
"initializer_range": 0.02,
|
| 44 |
+
"intermediate_size": 3072,
|
| 45 |
+
"label2id": {
|
| 46 |
+
"%3F": "0",
|
| 47 |
+
"a": "1",
|
| 48 |
+
"b": "2",
|
| 49 |
+
"c": "3",
|
| 50 |
+
"capital": "4",
|
| 51 |
+
"d": "5",
|
| 52 |
+
"e": "6",
|
| 53 |
+
"f": "7",
|
| 54 |
+
"g": "8",
|
| 55 |
+
"h": "9",
|
| 56 |
+
"i": "10",
|
| 57 |
+
"j": "11",
|
| 58 |
+
"k": "12",
|
| 59 |
+
"l": "13",
|
| 60 |
+
"m": "14",
|
| 61 |
+
"n": "15",
|
| 62 |
+
"o": "16",
|
| 63 |
+
"p": "17",
|
| 64 |
+
"period": "18",
|
| 65 |
+
"q": "19",
|
| 66 |
+
"r": "20",
|
| 67 |
+
"s": "21",
|
| 68 |
+
"t": "22",
|
| 69 |
+
"u": "23",
|
| 70 |
+
"v": "24",
|
| 71 |
+
"w": "25",
|
| 72 |
+
"x": "26",
|
| 73 |
+
"y": "27",
|
| 74 |
+
"z": "28"
|
| 75 |
+
},
|
| 76 |
+
"layer_norm_eps": 1e-12,
|
| 77 |
+
"model_type": "vit",
|
| 78 |
+
"num_attention_heads": 12,
|
| 79 |
+
"num_channels": 3,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"patch_size": 16,
|
| 82 |
+
"problem_type": "single_label_classification",
|
| 83 |
+
"qkv_bias": true,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.30.2"
|
| 86 |
+
}
|
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 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64d2fae05684355f48986d0590f7aa4df86555c2b14ef2dccd8eea77cedd3bed
|
| 3 |
+
size 343351725
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"total_flos": 1.1936654282133504e+18,
|
| 4 |
+
"train_loss": 0.5720695184201611,
|
| 5 |
+
"train_runtime": 566.3017,
|
| 6 |
+
"train_samples_per_second": 27.194,
|
| 7 |
+
"train_steps_per_second": 1.731
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,829 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7314792275428772,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-beans/checkpoint-640",
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"global_step": 980,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.2,
|
| 12 |
+
"learning_rate": 0.00019795918367346938,
|
| 13 |
+
"loss": 3.374,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.41,
|
| 18 |
+
"learning_rate": 0.0001959183673469388,
|
| 19 |
+
"loss": 3.3572,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.61,
|
| 24 |
+
"learning_rate": 0.00019387755102040816,
|
| 25 |
+
"loss": 3.3211,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.82,
|
| 30 |
+
"learning_rate": 0.00019183673469387756,
|
| 31 |
+
"loss": 3.2257,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.82,
|
| 36 |
+
"eval_accuracy": 0.03470031545741325,
|
| 37 |
+
"eval_loss": 3.4292807579040527,
|
| 38 |
+
"eval_runtime": 3.7472,
|
| 39 |
+
"eval_samples_per_second": 84.597,
|
| 40 |
+
"eval_steps_per_second": 10.675,
|
| 41 |
+
"step": 40
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.02,
|
| 45 |
+
"learning_rate": 0.00018979591836734697,
|
| 46 |
+
"loss": 3.1143,
|
| 47 |
+
"step": 50
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 1.22,
|
| 51 |
+
"learning_rate": 0.00018775510204081634,
|
| 52 |
+
"loss": 3.032,
|
| 53 |
+
"step": 60
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.43,
|
| 57 |
+
"learning_rate": 0.00018571428571428572,
|
| 58 |
+
"loss": 2.8104,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.63,
|
| 63 |
+
"learning_rate": 0.00018367346938775512,
|
| 64 |
+
"loss": 2.6674,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.63,
|
| 69 |
+
"eval_accuracy": 0.17350157728706625,
|
| 70 |
+
"eval_loss": 2.952009916305542,
|
| 71 |
+
"eval_runtime": 3.3938,
|
| 72 |
+
"eval_samples_per_second": 93.405,
|
| 73 |
+
"eval_steps_per_second": 11.786,
|
| 74 |
+
"step": 80
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.84,
|
| 78 |
+
"learning_rate": 0.0001816326530612245,
|
| 79 |
+
"loss": 2.5131,
|
| 80 |
+
"step": 90
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 2.04,
|
| 84 |
+
"learning_rate": 0.0001795918367346939,
|
| 85 |
+
"loss": 2.5287,
|
| 86 |
+
"step": 100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 2.24,
|
| 90 |
+
"learning_rate": 0.00017755102040816327,
|
| 91 |
+
"loss": 2.2121,
|
| 92 |
+
"step": 110
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 2.45,
|
| 96 |
+
"learning_rate": 0.00017551020408163265,
|
| 97 |
+
"loss": 2.0048,
|
| 98 |
+
"step": 120
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 2.45,
|
| 102 |
+
"eval_accuracy": 0.48580441640378547,
|
| 103 |
+
"eval_loss": 2.363018035888672,
|
| 104 |
+
"eval_runtime": 3.7903,
|
| 105 |
+
"eval_samples_per_second": 83.635,
|
| 106 |
+
"eval_steps_per_second": 10.553,
|
| 107 |
+
"step": 120
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 2.65,
|
| 111 |
+
"learning_rate": 0.00017346938775510205,
|
| 112 |
+
"loss": 1.8413,
|
| 113 |
+
"step": 130
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.86,
|
| 117 |
+
"learning_rate": 0.00017142857142857143,
|
| 118 |
+
"loss": 1.7897,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 3.06,
|
| 123 |
+
"learning_rate": 0.00016938775510204083,
|
| 124 |
+
"loss": 1.607,
|
| 125 |
+
"step": 150
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 3.27,
|
| 129 |
+
"learning_rate": 0.00016734693877551023,
|
| 130 |
+
"loss": 1.3493,
|
| 131 |
+
"step": 160
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 3.27,
|
| 135 |
+
"eval_accuracy": 0.5488958990536278,
|
| 136 |
+
"eval_loss": 1.8864167928695679,
|
| 137 |
+
"eval_runtime": 4.2623,
|
| 138 |
+
"eval_samples_per_second": 74.372,
|
| 139 |
+
"eval_steps_per_second": 9.385,
|
| 140 |
+
"step": 160
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.47,
|
| 144 |
+
"learning_rate": 0.0001653061224489796,
|
| 145 |
+
"loss": 1.2421,
|
| 146 |
+
"step": 170
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.67,
|
| 150 |
+
"learning_rate": 0.00016326530612244898,
|
| 151 |
+
"loss": 1.0779,
|
| 152 |
+
"step": 180
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.88,
|
| 156 |
+
"learning_rate": 0.00016122448979591838,
|
| 157 |
+
"loss": 1.0038,
|
| 158 |
+
"step": 190
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 4.08,
|
| 162 |
+
"learning_rate": 0.00015918367346938776,
|
| 163 |
+
"loss": 1.0887,
|
| 164 |
+
"step": 200
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.08,
|
| 168 |
+
"eval_accuracy": 0.7350157728706624,
|
| 169 |
+
"eval_loss": 1.4275710582733154,
|
| 170 |
+
"eval_runtime": 3.3105,
|
| 171 |
+
"eval_samples_per_second": 95.757,
|
| 172 |
+
"eval_steps_per_second": 12.083,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 4.29,
|
| 177 |
+
"learning_rate": 0.00015714285714285716,
|
| 178 |
+
"loss": 0.7734,
|
| 179 |
+
"step": 210
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 4.49,
|
| 183 |
+
"learning_rate": 0.00015510204081632654,
|
| 184 |
+
"loss": 0.7969,
|
| 185 |
+
"step": 220
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.69,
|
| 189 |
+
"learning_rate": 0.0001530612244897959,
|
| 190 |
+
"loss": 0.7953,
|
| 191 |
+
"step": 230
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 4.9,
|
| 195 |
+
"learning_rate": 0.0001510204081632653,
|
| 196 |
+
"loss": 0.6649,
|
| 197 |
+
"step": 240
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 4.9,
|
| 201 |
+
"eval_accuracy": 0.7728706624605678,
|
| 202 |
+
"eval_loss": 1.2706444263458252,
|
| 203 |
+
"eval_runtime": 4.1354,
|
| 204 |
+
"eval_samples_per_second": 76.656,
|
| 205 |
+
"eval_steps_per_second": 9.673,
|
| 206 |
+
"step": 240
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 5.1,
|
| 210 |
+
"learning_rate": 0.00014897959183673472,
|
| 211 |
+
"loss": 0.5647,
|
| 212 |
+
"step": 250
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 5.31,
|
| 216 |
+
"learning_rate": 0.0001469387755102041,
|
| 217 |
+
"loss": 0.7201,
|
| 218 |
+
"step": 260
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 5.51,
|
| 222 |
+
"learning_rate": 0.0001448979591836735,
|
| 223 |
+
"loss": 0.606,
|
| 224 |
+
"step": 270
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 5.71,
|
| 228 |
+
"learning_rate": 0.00014285714285714287,
|
| 229 |
+
"loss": 0.5396,
|
| 230 |
+
"step": 280
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 5.71,
|
| 234 |
+
"eval_accuracy": 0.7981072555205048,
|
| 235 |
+
"eval_loss": 1.125301480293274,
|
| 236 |
+
"eval_runtime": 3.3532,
|
| 237 |
+
"eval_samples_per_second": 94.537,
|
| 238 |
+
"eval_steps_per_second": 11.929,
|
| 239 |
+
"step": 280
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 5.92,
|
| 243 |
+
"learning_rate": 0.00014081632653061224,
|
| 244 |
+
"loss": 0.3407,
|
| 245 |
+
"step": 290
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 6.12,
|
| 249 |
+
"learning_rate": 0.00013877551020408165,
|
| 250 |
+
"loss": 0.3539,
|
| 251 |
+
"step": 300
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 6.33,
|
| 255 |
+
"learning_rate": 0.00013673469387755102,
|
| 256 |
+
"loss": 0.3608,
|
| 257 |
+
"step": 310
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 6.53,
|
| 261 |
+
"learning_rate": 0.0001346938775510204,
|
| 262 |
+
"loss": 0.3162,
|
| 263 |
+
"step": 320
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 6.53,
|
| 267 |
+
"eval_accuracy": 0.804416403785489,
|
| 268 |
+
"eval_loss": 1.0156543254852295,
|
| 269 |
+
"eval_runtime": 4.7629,
|
| 270 |
+
"eval_samples_per_second": 66.556,
|
| 271 |
+
"eval_steps_per_second": 8.398,
|
| 272 |
+
"step": 320
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.73,
|
| 276 |
+
"learning_rate": 0.0001326530612244898,
|
| 277 |
+
"loss": 0.335,
|
| 278 |
+
"step": 330
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 6.94,
|
| 282 |
+
"learning_rate": 0.00013061224489795917,
|
| 283 |
+
"loss": 0.2333,
|
| 284 |
+
"step": 340
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 7.14,
|
| 288 |
+
"learning_rate": 0.00012857142857142858,
|
| 289 |
+
"loss": 0.1931,
|
| 290 |
+
"step": 350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 7.35,
|
| 294 |
+
"learning_rate": 0.00012653061224489798,
|
| 295 |
+
"loss": 0.1909,
|
| 296 |
+
"step": 360
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 7.35,
|
| 300 |
+
"eval_accuracy": 0.8548895899053628,
|
| 301 |
+
"eval_loss": 0.8516904711723328,
|
| 302 |
+
"eval_runtime": 3.3664,
|
| 303 |
+
"eval_samples_per_second": 94.165,
|
| 304 |
+
"eval_steps_per_second": 11.882,
|
| 305 |
+
"step": 360
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 7.55,
|
| 309 |
+
"learning_rate": 0.00012448979591836735,
|
| 310 |
+
"loss": 0.1801,
|
| 311 |
+
"step": 370
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 7.76,
|
| 315 |
+
"learning_rate": 0.00012244897959183676,
|
| 316 |
+
"loss": 0.1785,
|
| 317 |
+
"step": 380
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 7.96,
|
| 321 |
+
"learning_rate": 0.00012040816326530613,
|
| 322 |
+
"loss": 0.1627,
|
| 323 |
+
"step": 390
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 8.16,
|
| 327 |
+
"learning_rate": 0.00011836734693877552,
|
| 328 |
+
"loss": 0.154,
|
| 329 |
+
"step": 400
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 8.16,
|
| 333 |
+
"eval_accuracy": 0.8138801261829653,
|
| 334 |
+
"eval_loss": 0.945625364780426,
|
| 335 |
+
"eval_runtime": 3.3591,
|
| 336 |
+
"eval_samples_per_second": 94.37,
|
| 337 |
+
"eval_steps_per_second": 11.908,
|
| 338 |
+
"step": 400
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 8.37,
|
| 342 |
+
"learning_rate": 0.0001163265306122449,
|
| 343 |
+
"loss": 0.1268,
|
| 344 |
+
"step": 410
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 8.57,
|
| 348 |
+
"learning_rate": 0.00011428571428571428,
|
| 349 |
+
"loss": 0.1352,
|
| 350 |
+
"step": 420
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 8.78,
|
| 354 |
+
"learning_rate": 0.00011224489795918367,
|
| 355 |
+
"loss": 0.1214,
|
| 356 |
+
"step": 430
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 8.98,
|
| 360 |
+
"learning_rate": 0.00011020408163265306,
|
| 361 |
+
"loss": 0.1519,
|
| 362 |
+
"step": 440
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 8.98,
|
| 366 |
+
"eval_accuracy": 0.8454258675078864,
|
| 367 |
+
"eval_loss": 0.8139908313751221,
|
| 368 |
+
"eval_runtime": 4.5811,
|
| 369 |
+
"eval_samples_per_second": 69.197,
|
| 370 |
+
"eval_steps_per_second": 8.731,
|
| 371 |
+
"step": 440
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 9.18,
|
| 375 |
+
"learning_rate": 0.00010816326530612246,
|
| 376 |
+
"loss": 0.0773,
|
| 377 |
+
"step": 450
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 9.39,
|
| 381 |
+
"learning_rate": 0.00010612244897959185,
|
| 382 |
+
"loss": 0.1083,
|
| 383 |
+
"step": 460
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 9.59,
|
| 387 |
+
"learning_rate": 0.00010408163265306123,
|
| 388 |
+
"loss": 0.086,
|
| 389 |
+
"step": 470
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 9.8,
|
| 393 |
+
"learning_rate": 0.00010204081632653062,
|
| 394 |
+
"loss": 0.0672,
|
| 395 |
+
"step": 480
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 9.8,
|
| 399 |
+
"eval_accuracy": 0.8517350157728707,
|
| 400 |
+
"eval_loss": 0.7407823204994202,
|
| 401 |
+
"eval_runtime": 3.8145,
|
| 402 |
+
"eval_samples_per_second": 83.104,
|
| 403 |
+
"eval_steps_per_second": 10.486,
|
| 404 |
+
"step": 480
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 10.0,
|
| 408 |
+
"learning_rate": 0.0001,
|
| 409 |
+
"loss": 0.067,
|
| 410 |
+
"step": 490
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 10.2,
|
| 414 |
+
"learning_rate": 9.79591836734694e-05,
|
| 415 |
+
"loss": 0.0559,
|
| 416 |
+
"step": 500
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 10.41,
|
| 420 |
+
"learning_rate": 9.591836734693878e-05,
|
| 421 |
+
"loss": 0.055,
|
| 422 |
+
"step": 510
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 10.61,
|
| 426 |
+
"learning_rate": 9.387755102040817e-05,
|
| 427 |
+
"loss": 0.0498,
|
| 428 |
+
"step": 520
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 10.61,
|
| 432 |
+
"eval_accuracy": 0.8580441640378549,
|
| 433 |
+
"eval_loss": 0.7550804615020752,
|
| 434 |
+
"eval_runtime": 3.3075,
|
| 435 |
+
"eval_samples_per_second": 95.842,
|
| 436 |
+
"eval_steps_per_second": 12.094,
|
| 437 |
+
"step": 520
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 10.82,
|
| 441 |
+
"learning_rate": 9.183673469387756e-05,
|
| 442 |
+
"loss": 0.0483,
|
| 443 |
+
"step": 530
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 11.02,
|
| 447 |
+
"learning_rate": 8.979591836734695e-05,
|
| 448 |
+
"loss": 0.0461,
|
| 449 |
+
"step": 540
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 11.22,
|
| 453 |
+
"learning_rate": 8.775510204081632e-05,
|
| 454 |
+
"loss": 0.043,
|
| 455 |
+
"step": 550
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 11.43,
|
| 459 |
+
"learning_rate": 8.571428571428571e-05,
|
| 460 |
+
"loss": 0.0618,
|
| 461 |
+
"step": 560
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 11.43,
|
| 465 |
+
"eval_accuracy": 0.8675078864353313,
|
| 466 |
+
"eval_loss": 0.7529485821723938,
|
| 467 |
+
"eval_runtime": 3.2693,
|
| 468 |
+
"eval_samples_per_second": 96.964,
|
| 469 |
+
"eval_steps_per_second": 12.235,
|
| 470 |
+
"step": 560
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 11.63,
|
| 474 |
+
"learning_rate": 8.367346938775511e-05,
|
| 475 |
+
"loss": 0.0409,
|
| 476 |
+
"step": 570
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 11.84,
|
| 480 |
+
"learning_rate": 8.163265306122449e-05,
|
| 481 |
+
"loss": 0.0433,
|
| 482 |
+
"step": 580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 12.04,
|
| 486 |
+
"learning_rate": 7.959183673469388e-05,
|
| 487 |
+
"loss": 0.0367,
|
| 488 |
+
"step": 590
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 12.24,
|
| 492 |
+
"learning_rate": 7.755102040816327e-05,
|
| 493 |
+
"loss": 0.0352,
|
| 494 |
+
"step": 600
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 12.24,
|
| 498 |
+
"eval_accuracy": 0.8643533123028391,
|
| 499 |
+
"eval_loss": 0.7547315955162048,
|
| 500 |
+
"eval_runtime": 4.666,
|
| 501 |
+
"eval_samples_per_second": 67.938,
|
| 502 |
+
"eval_steps_per_second": 8.573,
|
| 503 |
+
"step": 600
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 12.45,
|
| 507 |
+
"learning_rate": 7.551020408163266e-05,
|
| 508 |
+
"loss": 0.0377,
|
| 509 |
+
"step": 610
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 12.65,
|
| 513 |
+
"learning_rate": 7.346938775510205e-05,
|
| 514 |
+
"loss": 0.0331,
|
| 515 |
+
"step": 620
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 12.86,
|
| 519 |
+
"learning_rate": 7.142857142857143e-05,
|
| 520 |
+
"loss": 0.0678,
|
| 521 |
+
"step": 630
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 13.06,
|
| 525 |
+
"learning_rate": 6.938775510204082e-05,
|
| 526 |
+
"loss": 0.0381,
|
| 527 |
+
"step": 640
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 13.06,
|
| 531 |
+
"eval_accuracy": 0.8769716088328076,
|
| 532 |
+
"eval_loss": 0.7314792275428772,
|
| 533 |
+
"eval_runtime": 4.0948,
|
| 534 |
+
"eval_samples_per_second": 77.416,
|
| 535 |
+
"eval_steps_per_second": 9.769,
|
| 536 |
+
"step": 640
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 13.27,
|
| 540 |
+
"learning_rate": 6.73469387755102e-05,
|
| 541 |
+
"loss": 0.0316,
|
| 542 |
+
"step": 650
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 13.47,
|
| 546 |
+
"learning_rate": 6.530612244897959e-05,
|
| 547 |
+
"loss": 0.0302,
|
| 548 |
+
"step": 660
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 13.67,
|
| 552 |
+
"learning_rate": 6.326530612244899e-05,
|
| 553 |
+
"loss": 0.0317,
|
| 554 |
+
"step": 670
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 13.88,
|
| 558 |
+
"learning_rate": 6.122448979591838e-05,
|
| 559 |
+
"loss": 0.0288,
|
| 560 |
+
"step": 680
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 13.88,
|
| 564 |
+
"eval_accuracy": 0.8675078864353313,
|
| 565 |
+
"eval_loss": 0.7384619116783142,
|
| 566 |
+
"eval_runtime": 4.3527,
|
| 567 |
+
"eval_samples_per_second": 72.828,
|
| 568 |
+
"eval_steps_per_second": 9.19,
|
| 569 |
+
"step": 680
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 14.08,
|
| 573 |
+
"learning_rate": 5.918367346938776e-05,
|
| 574 |
+
"loss": 0.0285,
|
| 575 |
+
"step": 690
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 14.29,
|
| 579 |
+
"learning_rate": 5.714285714285714e-05,
|
| 580 |
+
"loss": 0.0271,
|
| 581 |
+
"step": 700
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 14.49,
|
| 585 |
+
"learning_rate": 5.510204081632653e-05,
|
| 586 |
+
"loss": 0.0264,
|
| 587 |
+
"step": 710
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 14.69,
|
| 591 |
+
"learning_rate": 5.3061224489795926e-05,
|
| 592 |
+
"loss": 0.0264,
|
| 593 |
+
"step": 720
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 14.69,
|
| 597 |
+
"eval_accuracy": 0.8643533123028391,
|
| 598 |
+
"eval_loss": 0.7676613926887512,
|
| 599 |
+
"eval_runtime": 3.2589,
|
| 600 |
+
"eval_samples_per_second": 97.271,
|
| 601 |
+
"eval_steps_per_second": 12.274,
|
| 602 |
+
"step": 720
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 14.9,
|
| 606 |
+
"learning_rate": 5.102040816326531e-05,
|
| 607 |
+
"loss": 0.0258,
|
| 608 |
+
"step": 730
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 15.1,
|
| 612 |
+
"learning_rate": 4.89795918367347e-05,
|
| 613 |
+
"loss": 0.0256,
|
| 614 |
+
"step": 740
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 15.31,
|
| 618 |
+
"learning_rate": 4.6938775510204086e-05,
|
| 619 |
+
"loss": 0.0242,
|
| 620 |
+
"step": 750
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 15.51,
|
| 624 |
+
"learning_rate": 4.4897959183673474e-05,
|
| 625 |
+
"loss": 0.0243,
|
| 626 |
+
"step": 760
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 15.51,
|
| 630 |
+
"eval_accuracy": 0.8580441640378549,
|
| 631 |
+
"eval_loss": 0.7726181745529175,
|
| 632 |
+
"eval_runtime": 3.8096,
|
| 633 |
+
"eval_samples_per_second": 83.211,
|
| 634 |
+
"eval_steps_per_second": 10.5,
|
| 635 |
+
"step": 760
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 15.71,
|
| 639 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 640 |
+
"loss": 0.0234,
|
| 641 |
+
"step": 770
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 15.92,
|
| 645 |
+
"learning_rate": 4.0816326530612245e-05,
|
| 646 |
+
"loss": 0.0241,
|
| 647 |
+
"step": 780
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 16.12,
|
| 651 |
+
"learning_rate": 3.8775510204081634e-05,
|
| 652 |
+
"loss": 0.0227,
|
| 653 |
+
"step": 790
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 16.33,
|
| 657 |
+
"learning_rate": 3.673469387755102e-05,
|
| 658 |
+
"loss": 0.0228,
|
| 659 |
+
"step": 800
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 16.33,
|
| 663 |
+
"eval_accuracy": 0.8580441640378549,
|
| 664 |
+
"eval_loss": 0.7750853300094604,
|
| 665 |
+
"eval_runtime": 3.2592,
|
| 666 |
+
"eval_samples_per_second": 97.263,
|
| 667 |
+
"eval_steps_per_second": 12.273,
|
| 668 |
+
"step": 800
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 16.53,
|
| 672 |
+
"learning_rate": 3.469387755102041e-05,
|
| 673 |
+
"loss": 0.0236,
|
| 674 |
+
"step": 810
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 16.73,
|
| 678 |
+
"learning_rate": 3.265306122448979e-05,
|
| 679 |
+
"loss": 0.023,
|
| 680 |
+
"step": 820
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 16.94,
|
| 684 |
+
"learning_rate": 3.061224489795919e-05,
|
| 685 |
+
"loss": 0.0219,
|
| 686 |
+
"step": 830
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 17.14,
|
| 690 |
+
"learning_rate": 2.857142857142857e-05,
|
| 691 |
+
"loss": 0.0219,
|
| 692 |
+
"step": 840
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 17.14,
|
| 696 |
+
"eval_accuracy": 0.8643533123028391,
|
| 697 |
+
"eval_loss": 0.7764984369277954,
|
| 698 |
+
"eval_runtime": 3.4162,
|
| 699 |
+
"eval_samples_per_second": 92.793,
|
| 700 |
+
"eval_steps_per_second": 11.709,
|
| 701 |
+
"step": 840
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 17.35,
|
| 705 |
+
"learning_rate": 2.6530612244897963e-05,
|
| 706 |
+
"loss": 0.0214,
|
| 707 |
+
"step": 850
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 17.55,
|
| 711 |
+
"learning_rate": 2.448979591836735e-05,
|
| 712 |
+
"loss": 0.0219,
|
| 713 |
+
"step": 860
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 17.76,
|
| 717 |
+
"learning_rate": 2.2448979591836737e-05,
|
| 718 |
+
"loss": 0.0207,
|
| 719 |
+
"step": 870
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 17.96,
|
| 723 |
+
"learning_rate": 2.0408163265306123e-05,
|
| 724 |
+
"loss": 0.0218,
|
| 725 |
+
"step": 880
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 17.96,
|
| 729 |
+
"eval_accuracy": 0.8643533123028391,
|
| 730 |
+
"eval_loss": 0.7789060473442078,
|
| 731 |
+
"eval_runtime": 3.3084,
|
| 732 |
+
"eval_samples_per_second": 95.816,
|
| 733 |
+
"eval_steps_per_second": 12.09,
|
| 734 |
+
"step": 880
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 18.16,
|
| 738 |
+
"learning_rate": 1.836734693877551e-05,
|
| 739 |
+
"loss": 0.0219,
|
| 740 |
+
"step": 890
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 18.37,
|
| 744 |
+
"learning_rate": 1.6326530612244897e-05,
|
| 745 |
+
"loss": 0.0214,
|
| 746 |
+
"step": 900
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 18.57,
|
| 750 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 751 |
+
"loss": 0.021,
|
| 752 |
+
"step": 910
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 18.78,
|
| 756 |
+
"learning_rate": 1.2244897959183674e-05,
|
| 757 |
+
"loss": 0.0209,
|
| 758 |
+
"step": 920
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 18.78,
|
| 762 |
+
"eval_accuracy": 0.8643533123028391,
|
| 763 |
+
"eval_loss": 0.779566764831543,
|
| 764 |
+
"eval_runtime": 4.6967,
|
| 765 |
+
"eval_samples_per_second": 67.494,
|
| 766 |
+
"eval_steps_per_second": 8.517,
|
| 767 |
+
"step": 920
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 18.98,
|
| 771 |
+
"learning_rate": 1.0204081632653061e-05,
|
| 772 |
+
"loss": 0.0204,
|
| 773 |
+
"step": 930
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 19.18,
|
| 777 |
+
"learning_rate": 8.163265306122448e-06,
|
| 778 |
+
"loss": 0.0209,
|
| 779 |
+
"step": 940
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 19.39,
|
| 783 |
+
"learning_rate": 6.122448979591837e-06,
|
| 784 |
+
"loss": 0.0207,
|
| 785 |
+
"step": 950
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 19.59,
|
| 789 |
+
"learning_rate": 4.081632653061224e-06,
|
| 790 |
+
"loss": 0.0206,
|
| 791 |
+
"step": 960
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 19.59,
|
| 795 |
+
"eval_accuracy": 0.8643533123028391,
|
| 796 |
+
"eval_loss": 0.7802363634109497,
|
| 797 |
+
"eval_runtime": 3.3003,
|
| 798 |
+
"eval_samples_per_second": 96.051,
|
| 799 |
+
"eval_steps_per_second": 12.12,
|
| 800 |
+
"step": 960
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 19.8,
|
| 804 |
+
"learning_rate": 2.040816326530612e-06,
|
| 805 |
+
"loss": 0.0201,
|
| 806 |
+
"step": 970
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 20.0,
|
| 810 |
+
"learning_rate": 0.0,
|
| 811 |
+
"loss": 0.0204,
|
| 812 |
+
"step": 980
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 20.0,
|
| 816 |
+
"step": 980,
|
| 817 |
+
"total_flos": 1.1936654282133504e+18,
|
| 818 |
+
"train_loss": 0.5720695184201611,
|
| 819 |
+
"train_runtime": 566.3017,
|
| 820 |
+
"train_samples_per_second": 27.194,
|
| 821 |
+
"train_steps_per_second": 1.731
|
| 822 |
+
}
|
| 823 |
+
],
|
| 824 |
+
"max_steps": 980,
|
| 825 |
+
"num_train_epochs": 20,
|
| 826 |
+
"total_flos": 1.1936654282133504e+18,
|
| 827 |
+
"trial_name": null,
|
| 828 |
+
"trial_params": null
|
| 829 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88e90154845df1cbf20cec744411d8e64005f30ae741e837b432941d036a7976
|
| 3 |
+
size 3899
|