Upload 8 files
Browse files- all_results.json +12 -0
- config.json +32 -0
- eval_results.json +8 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- train_results.json +7 -0
- trainer_state.json +726 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9719101123595506,
|
| 4 |
+
"eval_loss": 0.1326947957277298,
|
| 5 |
+
"eval_runtime": 7.941,
|
| 6 |
+
"eval_samples_per_second": 22.415,
|
| 7 |
+
"eval_steps_per_second": 2.896,
|
| 8 |
+
"train_loss": 0.12084352607184118,
|
| 9 |
+
"train_runtime": 814.8954,
|
| 10 |
+
"train_samples_per_second": 19.806,
|
| 11 |
+
"train_steps_per_second": 1.239
|
| 12 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "clear",
|
| 13 |
+
"1": "noise"
|
| 14 |
+
},
|
| 15 |
+
"image_size": 224,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"clear": "0",
|
| 20 |
+
"noise": "1"
|
| 21 |
+
},
|
| 22 |
+
"layer_norm_eps": 1e-12,
|
| 23 |
+
"model_type": "vit",
|
| 24 |
+
"num_attention_heads": 12,
|
| 25 |
+
"num_channels": 3,
|
| 26 |
+
"num_hidden_layers": 12,
|
| 27 |
+
"patch_size": 16,
|
| 28 |
+
"problem_type": "single_label_classification",
|
| 29 |
+
"qkv_bias": true,
|
| 30 |
+
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.36.2"
|
| 32 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9719101123595506,
|
| 4 |
+
"eval_loss": 0.1326947957277298,
|
| 5 |
+
"eval_runtime": 7.941,
|
| 6 |
+
"eval_samples_per_second": 22.415,
|
| 7 |
+
"eval_steps_per_second": 2.896
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ef30cad94bd87f6eae432dc019441f94da911e28e8d24bf6d193d052f23a25e
|
| 3 |
+
size 343223968
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 0.12084352607184118,
|
| 4 |
+
"train_runtime": 814.8954,
|
| 5 |
+
"train_samples_per_second": 19.806,
|
| 6 |
+
"train_steps_per_second": 1.239
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,726 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.1326947957277298,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-noise-classifier-merge\\checkpoint-500",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1010,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.1,
|
| 13 |
+
"learning_rate": 9.900990099009901e-06,
|
| 14 |
+
"loss": 0.6809,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.2,
|
| 19 |
+
"learning_rate": 9.801980198019802e-06,
|
| 20 |
+
"loss": 0.6748,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.3,
|
| 25 |
+
"learning_rate": 9.702970297029704e-06,
|
| 26 |
+
"loss": 0.6559,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.4,
|
| 31 |
+
"learning_rate": 9.603960396039604e-06,
|
| 32 |
+
"loss": 0.6097,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.5,
|
| 37 |
+
"learning_rate": 9.504950495049505e-06,
|
| 38 |
+
"loss": 0.6009,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.59,
|
| 43 |
+
"learning_rate": 9.405940594059405e-06,
|
| 44 |
+
"loss": 0.5522,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.69,
|
| 49 |
+
"learning_rate": 9.306930693069308e-06,
|
| 50 |
+
"loss": 0.5324,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.79,
|
| 55 |
+
"learning_rate": 9.20792079207921e-06,
|
| 56 |
+
"loss": 0.4652,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.89,
|
| 61 |
+
"learning_rate": 9.10891089108911e-06,
|
| 62 |
+
"loss": 0.4078,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.99,
|
| 67 |
+
"learning_rate": 9.009900990099011e-06,
|
| 68 |
+
"loss": 0.3836,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.99,
|
| 73 |
+
"eval_accuracy": 0.9213483146067416,
|
| 74 |
+
"eval_loss": 0.35515740513801575,
|
| 75 |
+
"eval_runtime": 7.2882,
|
| 76 |
+
"eval_samples_per_second": 24.423,
|
| 77 |
+
"eval_steps_per_second": 3.156,
|
| 78 |
+
"step": 100
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.09,
|
| 82 |
+
"learning_rate": 8.910891089108911e-06,
|
| 83 |
+
"loss": 0.3241,
|
| 84 |
+
"step": 110
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.19,
|
| 88 |
+
"learning_rate": 8.811881188118812e-06,
|
| 89 |
+
"loss": 0.3047,
|
| 90 |
+
"step": 120
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.29,
|
| 94 |
+
"learning_rate": 8.712871287128714e-06,
|
| 95 |
+
"loss": 0.2842,
|
| 96 |
+
"step": 130
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.39,
|
| 100 |
+
"learning_rate": 8.613861386138615e-06,
|
| 101 |
+
"loss": 0.2481,
|
| 102 |
+
"step": 140
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.49,
|
| 106 |
+
"learning_rate": 8.514851485148515e-06,
|
| 107 |
+
"loss": 0.2672,
|
| 108 |
+
"step": 150
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.58,
|
| 112 |
+
"learning_rate": 8.415841584158416e-06,
|
| 113 |
+
"loss": 0.2617,
|
| 114 |
+
"step": 160
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.68,
|
| 118 |
+
"learning_rate": 8.316831683168318e-06,
|
| 119 |
+
"loss": 0.1862,
|
| 120 |
+
"step": 170
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 1.78,
|
| 124 |
+
"learning_rate": 8.217821782178218e-06,
|
| 125 |
+
"loss": 0.1962,
|
| 126 |
+
"step": 180
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.88,
|
| 130 |
+
"learning_rate": 8.11881188118812e-06,
|
| 131 |
+
"loss": 0.1716,
|
| 132 |
+
"step": 190
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.98,
|
| 136 |
+
"learning_rate": 8.019801980198021e-06,
|
| 137 |
+
"loss": 0.1923,
|
| 138 |
+
"step": 200
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.98,
|
| 142 |
+
"eval_accuracy": 0.9325842696629213,
|
| 143 |
+
"eval_loss": 0.19820977747440338,
|
| 144 |
+
"eval_runtime": 7.3054,
|
| 145 |
+
"eval_samples_per_second": 24.365,
|
| 146 |
+
"eval_steps_per_second": 3.148,
|
| 147 |
+
"step": 200
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 2.08,
|
| 151 |
+
"learning_rate": 7.920792079207921e-06,
|
| 152 |
+
"loss": 0.201,
|
| 153 |
+
"step": 210
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.18,
|
| 157 |
+
"learning_rate": 7.821782178217822e-06,
|
| 158 |
+
"loss": 0.1257,
|
| 159 |
+
"step": 220
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.28,
|
| 163 |
+
"learning_rate": 7.722772277227724e-06,
|
| 164 |
+
"loss": 0.1639,
|
| 165 |
+
"step": 230
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.38,
|
| 169 |
+
"learning_rate": 7.6237623762376246e-06,
|
| 170 |
+
"loss": 0.094,
|
| 171 |
+
"step": 240
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.48,
|
| 175 |
+
"learning_rate": 7.524752475247525e-06,
|
| 176 |
+
"loss": 0.1489,
|
| 177 |
+
"step": 250
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 2.57,
|
| 181 |
+
"learning_rate": 7.425742574257426e-06,
|
| 182 |
+
"loss": 0.1405,
|
| 183 |
+
"step": 260
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 2.67,
|
| 187 |
+
"learning_rate": 7.326732673267327e-06,
|
| 188 |
+
"loss": 0.1139,
|
| 189 |
+
"step": 270
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 2.77,
|
| 193 |
+
"learning_rate": 7.227722772277228e-06,
|
| 194 |
+
"loss": 0.1405,
|
| 195 |
+
"step": 280
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 2.87,
|
| 199 |
+
"learning_rate": 7.128712871287129e-06,
|
| 200 |
+
"loss": 0.1273,
|
| 201 |
+
"step": 290
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.97,
|
| 205 |
+
"learning_rate": 7.02970297029703e-06,
|
| 206 |
+
"loss": 0.0953,
|
| 207 |
+
"step": 300
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 2.97,
|
| 211 |
+
"eval_accuracy": 0.9606741573033708,
|
| 212 |
+
"eval_loss": 0.13618314266204834,
|
| 213 |
+
"eval_runtime": 7.2985,
|
| 214 |
+
"eval_samples_per_second": 24.389,
|
| 215 |
+
"eval_steps_per_second": 3.151,
|
| 216 |
+
"step": 300
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 3.07,
|
| 220 |
+
"learning_rate": 6.930693069306931e-06,
|
| 221 |
+
"loss": 0.0708,
|
| 222 |
+
"step": 310
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 3.17,
|
| 226 |
+
"learning_rate": 6.831683168316833e-06,
|
| 227 |
+
"loss": 0.1164,
|
| 228 |
+
"step": 320
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 3.27,
|
| 232 |
+
"learning_rate": 6.732673267326733e-06,
|
| 233 |
+
"loss": 0.1114,
|
| 234 |
+
"step": 330
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 3.37,
|
| 238 |
+
"learning_rate": 6.633663366336635e-06,
|
| 239 |
+
"loss": 0.1032,
|
| 240 |
+
"step": 340
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 3.47,
|
| 244 |
+
"learning_rate": 6.534653465346535e-06,
|
| 245 |
+
"loss": 0.0625,
|
| 246 |
+
"step": 350
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.56,
|
| 250 |
+
"learning_rate": 6.4356435643564364e-06,
|
| 251 |
+
"loss": 0.0401,
|
| 252 |
+
"step": 360
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 3.66,
|
| 256 |
+
"learning_rate": 6.336633663366337e-06,
|
| 257 |
+
"loss": 0.0829,
|
| 258 |
+
"step": 370
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 3.76,
|
| 262 |
+
"learning_rate": 6.237623762376238e-06,
|
| 263 |
+
"loss": 0.1053,
|
| 264 |
+
"step": 380
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 3.86,
|
| 268 |
+
"learning_rate": 6.138613861386139e-06,
|
| 269 |
+
"loss": 0.0642,
|
| 270 |
+
"step": 390
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 3.96,
|
| 274 |
+
"learning_rate": 6.03960396039604e-06,
|
| 275 |
+
"loss": 0.0622,
|
| 276 |
+
"step": 400
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 3.96,
|
| 280 |
+
"eval_accuracy": 0.9662921348314607,
|
| 281 |
+
"eval_loss": 0.14083810150623322,
|
| 282 |
+
"eval_runtime": 8.2574,
|
| 283 |
+
"eval_samples_per_second": 21.556,
|
| 284 |
+
"eval_steps_per_second": 2.785,
|
| 285 |
+
"step": 400
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.06,
|
| 289 |
+
"learning_rate": 5.940594059405941e-06,
|
| 290 |
+
"loss": 0.0481,
|
| 291 |
+
"step": 410
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 4.16,
|
| 295 |
+
"learning_rate": 5.841584158415842e-06,
|
| 296 |
+
"loss": 0.0648,
|
| 297 |
+
"step": 420
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 4.26,
|
| 301 |
+
"learning_rate": 5.7425742574257425e-06,
|
| 302 |
+
"loss": 0.049,
|
| 303 |
+
"step": 430
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 4.36,
|
| 307 |
+
"learning_rate": 5.643564356435644e-06,
|
| 308 |
+
"loss": 0.0459,
|
| 309 |
+
"step": 440
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 4.46,
|
| 313 |
+
"learning_rate": 5.544554455445545e-06,
|
| 314 |
+
"loss": 0.0986,
|
| 315 |
+
"step": 450
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 4.55,
|
| 319 |
+
"learning_rate": 5.4455445544554465e-06,
|
| 320 |
+
"loss": 0.0899,
|
| 321 |
+
"step": 460
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 4.65,
|
| 325 |
+
"learning_rate": 5.346534653465347e-06,
|
| 326 |
+
"loss": 0.0484,
|
| 327 |
+
"step": 470
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 4.75,
|
| 331 |
+
"learning_rate": 5.247524752475248e-06,
|
| 332 |
+
"loss": 0.0699,
|
| 333 |
+
"step": 480
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 4.85,
|
| 337 |
+
"learning_rate": 5.148514851485149e-06,
|
| 338 |
+
"loss": 0.0248,
|
| 339 |
+
"step": 490
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.95,
|
| 343 |
+
"learning_rate": 5.04950495049505e-06,
|
| 344 |
+
"loss": 0.0392,
|
| 345 |
+
"step": 500
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 4.95,
|
| 349 |
+
"eval_accuracy": 0.9719101123595506,
|
| 350 |
+
"eval_loss": 0.1326947957277298,
|
| 351 |
+
"eval_runtime": 8.3115,
|
| 352 |
+
"eval_samples_per_second": 21.416,
|
| 353 |
+
"eval_steps_per_second": 2.767,
|
| 354 |
+
"step": 500
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 5.05,
|
| 358 |
+
"learning_rate": 4.950495049504951e-06,
|
| 359 |
+
"loss": 0.0226,
|
| 360 |
+
"step": 510
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 5.15,
|
| 364 |
+
"learning_rate": 4.851485148514852e-06,
|
| 365 |
+
"loss": 0.064,
|
| 366 |
+
"step": 520
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.25,
|
| 370 |
+
"learning_rate": 4.7524752475247525e-06,
|
| 371 |
+
"loss": 0.0675,
|
| 372 |
+
"step": 530
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 5.35,
|
| 376 |
+
"learning_rate": 4.653465346534654e-06,
|
| 377 |
+
"loss": 0.0452,
|
| 378 |
+
"step": 540
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 5.45,
|
| 382 |
+
"learning_rate": 4.554455445544555e-06,
|
| 383 |
+
"loss": 0.0365,
|
| 384 |
+
"step": 550
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 5.54,
|
| 388 |
+
"learning_rate": 4.455445544554456e-06,
|
| 389 |
+
"loss": 0.0394,
|
| 390 |
+
"step": 560
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 5.64,
|
| 394 |
+
"learning_rate": 4.356435643564357e-06,
|
| 395 |
+
"loss": 0.0272,
|
| 396 |
+
"step": 570
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 5.74,
|
| 400 |
+
"learning_rate": 4.2574257425742575e-06,
|
| 401 |
+
"loss": 0.0672,
|
| 402 |
+
"step": 580
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 5.84,
|
| 406 |
+
"learning_rate": 4.158415841584159e-06,
|
| 407 |
+
"loss": 0.0404,
|
| 408 |
+
"step": 590
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 5.94,
|
| 412 |
+
"learning_rate": 4.05940594059406e-06,
|
| 413 |
+
"loss": 0.0197,
|
| 414 |
+
"step": 600
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 5.94,
|
| 418 |
+
"eval_accuracy": 0.9719101123595506,
|
| 419 |
+
"eval_loss": 0.13345670700073242,
|
| 420 |
+
"eval_runtime": 8.0014,
|
| 421 |
+
"eval_samples_per_second": 22.246,
|
| 422 |
+
"eval_steps_per_second": 2.875,
|
| 423 |
+
"step": 600
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 6.04,
|
| 427 |
+
"learning_rate": 3.960396039603961e-06,
|
| 428 |
+
"loss": 0.028,
|
| 429 |
+
"step": 610
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 6.14,
|
| 433 |
+
"learning_rate": 3.861386138613862e-06,
|
| 434 |
+
"loss": 0.0381,
|
| 435 |
+
"step": 620
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 6.24,
|
| 439 |
+
"learning_rate": 3.7623762376237625e-06,
|
| 440 |
+
"loss": 0.0403,
|
| 441 |
+
"step": 630
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 6.34,
|
| 445 |
+
"learning_rate": 3.6633663366336635e-06,
|
| 446 |
+
"loss": 0.0183,
|
| 447 |
+
"step": 640
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.44,
|
| 451 |
+
"learning_rate": 3.5643564356435644e-06,
|
| 452 |
+
"loss": 0.0324,
|
| 453 |
+
"step": 650
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 6.53,
|
| 457 |
+
"learning_rate": 3.4653465346534653e-06,
|
| 458 |
+
"loss": 0.0411,
|
| 459 |
+
"step": 660
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 6.63,
|
| 463 |
+
"learning_rate": 3.3663366336633666e-06,
|
| 464 |
+
"loss": 0.0247,
|
| 465 |
+
"step": 670
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 6.73,
|
| 469 |
+
"learning_rate": 3.2673267326732676e-06,
|
| 470 |
+
"loss": 0.0381,
|
| 471 |
+
"step": 680
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 6.83,
|
| 475 |
+
"learning_rate": 3.1683168316831685e-06,
|
| 476 |
+
"loss": 0.017,
|
| 477 |
+
"step": 690
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 6.93,
|
| 481 |
+
"learning_rate": 3.0693069306930694e-06,
|
| 482 |
+
"loss": 0.0278,
|
| 483 |
+
"step": 700
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 6.93,
|
| 487 |
+
"eval_accuracy": 0.9606741573033708,
|
| 488 |
+
"eval_loss": 0.13798531889915466,
|
| 489 |
+
"eval_runtime": 8.0678,
|
| 490 |
+
"eval_samples_per_second": 22.063,
|
| 491 |
+
"eval_steps_per_second": 2.851,
|
| 492 |
+
"step": 700
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 7.03,
|
| 496 |
+
"learning_rate": 2.9702970297029703e-06,
|
| 497 |
+
"loss": 0.0349,
|
| 498 |
+
"step": 710
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 7.13,
|
| 502 |
+
"learning_rate": 2.8712871287128712e-06,
|
| 503 |
+
"loss": 0.0224,
|
| 504 |
+
"step": 720
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 7.23,
|
| 508 |
+
"learning_rate": 2.7722772277227726e-06,
|
| 509 |
+
"loss": 0.0565,
|
| 510 |
+
"step": 730
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 7.33,
|
| 514 |
+
"learning_rate": 2.6732673267326735e-06,
|
| 515 |
+
"loss": 0.0394,
|
| 516 |
+
"step": 740
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 7.43,
|
| 520 |
+
"learning_rate": 2.5742574257425744e-06,
|
| 521 |
+
"loss": 0.0265,
|
| 522 |
+
"step": 750
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 7.52,
|
| 526 |
+
"learning_rate": 2.4752475247524753e-06,
|
| 527 |
+
"loss": 0.0157,
|
| 528 |
+
"step": 760
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 7.62,
|
| 532 |
+
"learning_rate": 2.3762376237623762e-06,
|
| 533 |
+
"loss": 0.0164,
|
| 534 |
+
"step": 770
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 7.72,
|
| 538 |
+
"learning_rate": 2.2772277227722776e-06,
|
| 539 |
+
"loss": 0.0153,
|
| 540 |
+
"step": 780
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 7.82,
|
| 544 |
+
"learning_rate": 2.1782178217821785e-06,
|
| 545 |
+
"loss": 0.0153,
|
| 546 |
+
"step": 790
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 7.92,
|
| 550 |
+
"learning_rate": 2.0792079207920794e-06,
|
| 551 |
+
"loss": 0.0153,
|
| 552 |
+
"step": 800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 7.92,
|
| 556 |
+
"eval_accuracy": 0.9662921348314607,
|
| 557 |
+
"eval_loss": 0.13522128760814667,
|
| 558 |
+
"eval_runtime": 7.8876,
|
| 559 |
+
"eval_samples_per_second": 22.567,
|
| 560 |
+
"eval_steps_per_second": 2.916,
|
| 561 |
+
"step": 800
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 8.02,
|
| 565 |
+
"learning_rate": 1.9801980198019803e-06,
|
| 566 |
+
"loss": 0.0305,
|
| 567 |
+
"step": 810
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.12,
|
| 571 |
+
"learning_rate": 1.8811881188118813e-06,
|
| 572 |
+
"loss": 0.0351,
|
| 573 |
+
"step": 820
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 8.22,
|
| 577 |
+
"learning_rate": 1.7821782178217822e-06,
|
| 578 |
+
"loss": 0.0148,
|
| 579 |
+
"step": 830
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 8.32,
|
| 583 |
+
"learning_rate": 1.6831683168316833e-06,
|
| 584 |
+
"loss": 0.0381,
|
| 585 |
+
"step": 840
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 8.42,
|
| 589 |
+
"learning_rate": 1.5841584158415842e-06,
|
| 590 |
+
"loss": 0.0151,
|
| 591 |
+
"step": 850
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 8.51,
|
| 595 |
+
"learning_rate": 1.4851485148514852e-06,
|
| 596 |
+
"loss": 0.0145,
|
| 597 |
+
"step": 860
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 8.61,
|
| 601 |
+
"learning_rate": 1.3861386138613863e-06,
|
| 602 |
+
"loss": 0.0266,
|
| 603 |
+
"step": 870
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 8.71,
|
| 607 |
+
"learning_rate": 1.2871287128712872e-06,
|
| 608 |
+
"loss": 0.0146,
|
| 609 |
+
"step": 880
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 8.81,
|
| 613 |
+
"learning_rate": 1.1881188118811881e-06,
|
| 614 |
+
"loss": 0.0147,
|
| 615 |
+
"step": 890
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 8.91,
|
| 619 |
+
"learning_rate": 1.0891089108910893e-06,
|
| 620 |
+
"loss": 0.0284,
|
| 621 |
+
"step": 900
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 8.91,
|
| 625 |
+
"eval_accuracy": 0.9662921348314607,
|
| 626 |
+
"eval_loss": 0.13962529599666595,
|
| 627 |
+
"eval_runtime": 7.6158,
|
| 628 |
+
"eval_samples_per_second": 23.372,
|
| 629 |
+
"eval_steps_per_second": 3.02,
|
| 630 |
+
"step": 900
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 9.01,
|
| 634 |
+
"learning_rate": 9.900990099009902e-07,
|
| 635 |
+
"loss": 0.0221,
|
| 636 |
+
"step": 910
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 9.11,
|
| 640 |
+
"learning_rate": 8.910891089108911e-07,
|
| 641 |
+
"loss": 0.014,
|
| 642 |
+
"step": 920
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 9.21,
|
| 646 |
+
"learning_rate": 7.920792079207921e-07,
|
| 647 |
+
"loss": 0.0266,
|
| 648 |
+
"step": 930
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 9.31,
|
| 652 |
+
"learning_rate": 6.930693069306931e-07,
|
| 653 |
+
"loss": 0.0138,
|
| 654 |
+
"step": 940
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 9.41,
|
| 658 |
+
"learning_rate": 5.940594059405941e-07,
|
| 659 |
+
"loss": 0.033,
|
| 660 |
+
"step": 950
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 9.5,
|
| 664 |
+
"learning_rate": 4.950495049504951e-07,
|
| 665 |
+
"loss": 0.0371,
|
| 666 |
+
"step": 960
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 9.6,
|
| 670 |
+
"learning_rate": 3.9603960396039606e-07,
|
| 671 |
+
"loss": 0.0148,
|
| 672 |
+
"step": 970
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 9.7,
|
| 676 |
+
"learning_rate": 2.9702970297029703e-07,
|
| 677 |
+
"loss": 0.017,
|
| 678 |
+
"step": 980
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 9.8,
|
| 682 |
+
"learning_rate": 1.9801980198019803e-07,
|
| 683 |
+
"loss": 0.0138,
|
| 684 |
+
"step": 990
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 9.9,
|
| 688 |
+
"learning_rate": 9.900990099009901e-08,
|
| 689 |
+
"loss": 0.0252,
|
| 690 |
+
"step": 1000
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 9.9,
|
| 694 |
+
"eval_accuracy": 0.9662921348314607,
|
| 695 |
+
"eval_loss": 0.1401771754026413,
|
| 696 |
+
"eval_runtime": 7.8442,
|
| 697 |
+
"eval_samples_per_second": 22.692,
|
| 698 |
+
"eval_steps_per_second": 2.932,
|
| 699 |
+
"step": 1000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 10.0,
|
| 703 |
+
"learning_rate": 0.0,
|
| 704 |
+
"loss": 0.014,
|
| 705 |
+
"step": 1010
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 10.0,
|
| 709 |
+
"step": 1010,
|
| 710 |
+
"total_flos": 1.250720712379515e+18,
|
| 711 |
+
"train_loss": 0.12084352607184118,
|
| 712 |
+
"train_runtime": 814.8954,
|
| 713 |
+
"train_samples_per_second": 19.806,
|
| 714 |
+
"train_steps_per_second": 1.239
|
| 715 |
+
}
|
| 716 |
+
],
|
| 717 |
+
"logging_steps": 10,
|
| 718 |
+
"max_steps": 1010,
|
| 719 |
+
"num_input_tokens_seen": 0,
|
| 720 |
+
"num_train_epochs": 10,
|
| 721 |
+
"save_steps": 100,
|
| 722 |
+
"total_flos": 1.250720712379515e+18,
|
| 723 |
+
"train_batch_size": 16,
|
| 724 |
+
"trial_name": null,
|
| 725 |
+
"trial_params": null
|
| 726 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b6b59d92822fe8bb4974ff1ee21c177556bc3c2eb90faa4d364837fbe490142
|
| 3 |
+
size 4283
|