Commit
·
0ca9b62
1
Parent(s):
67e60df
Upload 8 files
Browse files- all_results.json +13 -0
- config.json +34 -0
- eval_results.json +8 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- train_results.json +8 -0
- trainer_state.json +826 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"eval_accuracy": 0.7152466367713004,
|
| 4 |
+
"eval_loss": 0.8530937433242798,
|
| 5 |
+
"eval_runtime": 12.7761,
|
| 6 |
+
"eval_samples_per_second": 34.909,
|
| 7 |
+
"eval_steps_per_second": 4.383,
|
| 8 |
+
"total_flos": 7.325059066442809e+18,
|
| 9 |
+
"train_loss": 0.2753036499082569,
|
| 10 |
+
"train_runtime": 1319.5023,
|
| 11 |
+
"train_samples_per_second": 24.294,
|
| 12 |
+
"train_steps_per_second": 0.764
|
| 13 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-384",
|
| 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": "negative",
|
| 13 |
+
"1": "neutral",
|
| 14 |
+
"2": "positive"
|
| 15 |
+
},
|
| 16 |
+
"image_size": 384,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 3072,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"negative": "0",
|
| 21 |
+
"neutral": "1",
|
| 22 |
+
"positive": "2"
|
| 23 |
+
},
|
| 24 |
+
"layer_norm_eps": 1e-12,
|
| 25 |
+
"model_type": "vit",
|
| 26 |
+
"num_attention_heads": 12,
|
| 27 |
+
"num_channels": 3,
|
| 28 |
+
"num_hidden_layers": 12,
|
| 29 |
+
"patch_size": 16,
|
| 30 |
+
"problem_type": "single_label_classification",
|
| 31 |
+
"qkv_bias": true,
|
| 32 |
+
"torch_dtype": "float32",
|
| 33 |
+
"transformers_version": "4.35.2"
|
| 34 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"eval_accuracy": 0.7152466367713004,
|
| 4 |
+
"eval_loss": 0.8530937433242798,
|
| 5 |
+
"eval_runtime": 12.7761,
|
| 6 |
+
"eval_samples_per_second": 34.909,
|
| 7 |
+
"eval_steps_per_second": 4.383
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90142947438b8bb461cd65f4b72568c8d342176876668c6e19aa121cb1c23280
|
| 3 |
+
size 344394412
|
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": 384,
|
| 20 |
+
"width": 384
|
| 21 |
+
}
|
| 22 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"total_flos": 7.325059066442809e+18,
|
| 4 |
+
"train_loss": 0.2753036499082569,
|
| 5 |
+
"train_runtime": 1319.5023,
|
| 6 |
+
"train_samples_per_second": 24.294,
|
| 7 |
+
"train_steps_per_second": 0.764
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,826 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7152466367713004,
|
| 3 |
+
"best_model_checkpoint": "/content/drive/MyDrive/vit-base-valence/checkpoint-850",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 1008,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08,
|
| 13 |
+
"learning_rate": 1.9801587301587305e-05,
|
| 14 |
+
"loss": 1.1812,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.16,
|
| 19 |
+
"learning_rate": 1.9603174603174602e-05,
|
| 20 |
+
"loss": 1.138,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.24,
|
| 25 |
+
"learning_rate": 1.9404761904761906e-05,
|
| 26 |
+
"loss": 1.0099,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.32,
|
| 31 |
+
"learning_rate": 1.920634920634921e-05,
|
| 32 |
+
"loss": 1.0198,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.4,
|
| 37 |
+
"learning_rate": 1.900793650793651e-05,
|
| 38 |
+
"loss": 0.9472,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.4,
|
| 43 |
+
"eval_accuracy": 0.6031390134529148,
|
| 44 |
+
"eval_loss": 0.8905579447746277,
|
| 45 |
+
"eval_runtime": 12.7388,
|
| 46 |
+
"eval_samples_per_second": 35.011,
|
| 47 |
+
"eval_steps_per_second": 4.396,
|
| 48 |
+
"step": 50
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.48,
|
| 52 |
+
"learning_rate": 1.880952380952381e-05,
|
| 53 |
+
"loss": 0.907,
|
| 54 |
+
"step": 60
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.56,
|
| 58 |
+
"learning_rate": 1.8611111111111114e-05,
|
| 59 |
+
"loss": 0.87,
|
| 60 |
+
"step": 70
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.63,
|
| 64 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 65 |
+
"loss": 0.8141,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.71,
|
| 70 |
+
"learning_rate": 1.8214285714285715e-05,
|
| 71 |
+
"loss": 0.8191,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.79,
|
| 76 |
+
"learning_rate": 1.801587301587302e-05,
|
| 77 |
+
"loss": 0.8528,
|
| 78 |
+
"step": 100
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.79,
|
| 82 |
+
"eval_accuracy": 0.6614349775784754,
|
| 83 |
+
"eval_loss": 0.7742370367050171,
|
| 84 |
+
"eval_runtime": 12.7432,
|
| 85 |
+
"eval_samples_per_second": 34.999,
|
| 86 |
+
"eval_steps_per_second": 4.394,
|
| 87 |
+
"step": 100
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.87,
|
| 91 |
+
"learning_rate": 1.781746031746032e-05,
|
| 92 |
+
"loss": 0.7775,
|
| 93 |
+
"step": 110
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.95,
|
| 97 |
+
"learning_rate": 1.761904761904762e-05,
|
| 98 |
+
"loss": 0.7763,
|
| 99 |
+
"step": 120
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.03,
|
| 103 |
+
"learning_rate": 1.7420634920634923e-05,
|
| 104 |
+
"loss": 0.654,
|
| 105 |
+
"step": 130
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.11,
|
| 109 |
+
"learning_rate": 1.7222222222222224e-05,
|
| 110 |
+
"loss": 0.5683,
|
| 111 |
+
"step": 140
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 1.19,
|
| 115 |
+
"learning_rate": 1.7023809523809524e-05,
|
| 116 |
+
"loss": 0.544,
|
| 117 |
+
"step": 150
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 1.19,
|
| 121 |
+
"eval_accuracy": 0.6659192825112108,
|
| 122 |
+
"eval_loss": 0.7118873000144958,
|
| 123 |
+
"eval_runtime": 12.762,
|
| 124 |
+
"eval_samples_per_second": 34.947,
|
| 125 |
+
"eval_steps_per_second": 4.388,
|
| 126 |
+
"step": 150
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.27,
|
| 130 |
+
"learning_rate": 1.6825396825396828e-05,
|
| 131 |
+
"loss": 0.5242,
|
| 132 |
+
"step": 160
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 1.35,
|
| 136 |
+
"learning_rate": 1.6626984126984128e-05,
|
| 137 |
+
"loss": 0.569,
|
| 138 |
+
"step": 170
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 1.43,
|
| 142 |
+
"learning_rate": 1.642857142857143e-05,
|
| 143 |
+
"loss": 0.4828,
|
| 144 |
+
"step": 180
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.51,
|
| 148 |
+
"learning_rate": 1.6230158730158732e-05,
|
| 149 |
+
"loss": 0.4871,
|
| 150 |
+
"step": 190
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.59,
|
| 154 |
+
"learning_rate": 1.6051587301587302e-05,
|
| 155 |
+
"loss": 0.5249,
|
| 156 |
+
"step": 200
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.59,
|
| 160 |
+
"eval_accuracy": 0.6905829596412556,
|
| 161 |
+
"eval_loss": 0.6951313018798828,
|
| 162 |
+
"eval_runtime": 12.8506,
|
| 163 |
+
"eval_samples_per_second": 34.707,
|
| 164 |
+
"eval_steps_per_second": 4.358,
|
| 165 |
+
"step": 200
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.67,
|
| 169 |
+
"learning_rate": 1.5853174603174603e-05,
|
| 170 |
+
"loss": 0.5511,
|
| 171 |
+
"step": 210
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 1.75,
|
| 175 |
+
"learning_rate": 1.5654761904761906e-05,
|
| 176 |
+
"loss": 0.4835,
|
| 177 |
+
"step": 220
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.83,
|
| 181 |
+
"learning_rate": 1.5456349206349207e-05,
|
| 182 |
+
"loss": 0.4923,
|
| 183 |
+
"step": 230
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.9,
|
| 187 |
+
"learning_rate": 1.5257936507936509e-05,
|
| 188 |
+
"loss": 0.5137,
|
| 189 |
+
"step": 240
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.98,
|
| 193 |
+
"learning_rate": 1.5059523809523811e-05,
|
| 194 |
+
"loss": 0.4744,
|
| 195 |
+
"step": 250
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.98,
|
| 199 |
+
"eval_accuracy": 0.7062780269058296,
|
| 200 |
+
"eval_loss": 0.6830359697341919,
|
| 201 |
+
"eval_runtime": 12.7409,
|
| 202 |
+
"eval_samples_per_second": 35.005,
|
| 203 |
+
"eval_steps_per_second": 4.395,
|
| 204 |
+
"step": 250
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 2.06,
|
| 208 |
+
"learning_rate": 1.4861111111111113e-05,
|
| 209 |
+
"loss": 0.3409,
|
| 210 |
+
"step": 260
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 2.14,
|
| 214 |
+
"learning_rate": 1.4662698412698413e-05,
|
| 215 |
+
"loss": 0.3297,
|
| 216 |
+
"step": 270
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 2.22,
|
| 220 |
+
"learning_rate": 1.4464285714285715e-05,
|
| 221 |
+
"loss": 0.3496,
|
| 222 |
+
"step": 280
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 2.3,
|
| 226 |
+
"learning_rate": 1.4265873015873018e-05,
|
| 227 |
+
"loss": 0.3384,
|
| 228 |
+
"step": 290
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 2.38,
|
| 232 |
+
"learning_rate": 1.406746031746032e-05,
|
| 233 |
+
"loss": 0.2939,
|
| 234 |
+
"step": 300
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 2.38,
|
| 238 |
+
"eval_accuracy": 0.7130044843049327,
|
| 239 |
+
"eval_loss": 0.6794872283935547,
|
| 240 |
+
"eval_runtime": 12.7967,
|
| 241 |
+
"eval_samples_per_second": 34.853,
|
| 242 |
+
"eval_steps_per_second": 4.376,
|
| 243 |
+
"step": 300
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 2.46,
|
| 247 |
+
"learning_rate": 1.386904761904762e-05,
|
| 248 |
+
"loss": 0.3219,
|
| 249 |
+
"step": 310
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.54,
|
| 253 |
+
"learning_rate": 1.3670634920634922e-05,
|
| 254 |
+
"loss": 0.2759,
|
| 255 |
+
"step": 320
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 2.62,
|
| 259 |
+
"learning_rate": 1.3472222222222224e-05,
|
| 260 |
+
"loss": 0.252,
|
| 261 |
+
"step": 330
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 2.7,
|
| 265 |
+
"learning_rate": 1.3273809523809524e-05,
|
| 266 |
+
"loss": 0.2574,
|
| 267 |
+
"step": 340
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 2.78,
|
| 271 |
+
"learning_rate": 1.3075396825396826e-05,
|
| 272 |
+
"loss": 0.2802,
|
| 273 |
+
"step": 350
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 2.78,
|
| 277 |
+
"eval_accuracy": 0.7062780269058296,
|
| 278 |
+
"eval_loss": 0.6800991296768188,
|
| 279 |
+
"eval_runtime": 12.7516,
|
| 280 |
+
"eval_samples_per_second": 34.976,
|
| 281 |
+
"eval_steps_per_second": 4.392,
|
| 282 |
+
"step": 350
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 2.86,
|
| 286 |
+
"learning_rate": 1.2876984126984129e-05,
|
| 287 |
+
"loss": 0.3155,
|
| 288 |
+
"step": 360
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.94,
|
| 292 |
+
"learning_rate": 1.2678571428571429e-05,
|
| 293 |
+
"loss": 0.3315,
|
| 294 |
+
"step": 370
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.02,
|
| 298 |
+
"learning_rate": 1.2480158730158731e-05,
|
| 299 |
+
"loss": 0.2872,
|
| 300 |
+
"step": 380
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 3.1,
|
| 304 |
+
"learning_rate": 1.2281746031746033e-05,
|
| 305 |
+
"loss": 0.1899,
|
| 306 |
+
"step": 390
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 3.17,
|
| 310 |
+
"learning_rate": 1.2083333333333333e-05,
|
| 311 |
+
"loss": 0.1755,
|
| 312 |
+
"step": 400
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 3.17,
|
| 316 |
+
"eval_accuracy": 0.7040358744394619,
|
| 317 |
+
"eval_loss": 0.6927831172943115,
|
| 318 |
+
"eval_runtime": 12.8352,
|
| 319 |
+
"eval_samples_per_second": 34.748,
|
| 320 |
+
"eval_steps_per_second": 4.363,
|
| 321 |
+
"step": 400
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 3.25,
|
| 325 |
+
"learning_rate": 1.1884920634920635e-05,
|
| 326 |
+
"loss": 0.1763,
|
| 327 |
+
"step": 410
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 3.33,
|
| 331 |
+
"learning_rate": 1.1686507936507938e-05,
|
| 332 |
+
"loss": 0.1725,
|
| 333 |
+
"step": 420
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 3.41,
|
| 337 |
+
"learning_rate": 1.1488095238095238e-05,
|
| 338 |
+
"loss": 0.1742,
|
| 339 |
+
"step": 430
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 3.49,
|
| 343 |
+
"learning_rate": 1.128968253968254e-05,
|
| 344 |
+
"loss": 0.162,
|
| 345 |
+
"step": 440
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.57,
|
| 349 |
+
"learning_rate": 1.1091269841269842e-05,
|
| 350 |
+
"loss": 0.1921,
|
| 351 |
+
"step": 450
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 3.57,
|
| 355 |
+
"eval_accuracy": 0.7040358744394619,
|
| 356 |
+
"eval_loss": 0.7218114733695984,
|
| 357 |
+
"eval_runtime": 12.8436,
|
| 358 |
+
"eval_samples_per_second": 34.725,
|
| 359 |
+
"eval_steps_per_second": 4.36,
|
| 360 |
+
"step": 450
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 3.65,
|
| 364 |
+
"learning_rate": 1.0892857142857142e-05,
|
| 365 |
+
"loss": 0.1636,
|
| 366 |
+
"step": 460
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 3.73,
|
| 370 |
+
"learning_rate": 1.0694444444444444e-05,
|
| 371 |
+
"loss": 0.1979,
|
| 372 |
+
"step": 470
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 3.81,
|
| 376 |
+
"learning_rate": 1.0496031746031747e-05,
|
| 377 |
+
"loss": 0.1994,
|
| 378 |
+
"step": 480
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 3.89,
|
| 382 |
+
"learning_rate": 1.0297619047619047e-05,
|
| 383 |
+
"loss": 0.1562,
|
| 384 |
+
"step": 490
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 3.97,
|
| 388 |
+
"learning_rate": 1.0099206349206349e-05,
|
| 389 |
+
"loss": 0.1682,
|
| 390 |
+
"step": 500
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 3.97,
|
| 394 |
+
"eval_accuracy": 0.7062780269058296,
|
| 395 |
+
"eval_loss": 0.7192036509513855,
|
| 396 |
+
"eval_runtime": 12.7187,
|
| 397 |
+
"eval_samples_per_second": 35.066,
|
| 398 |
+
"eval_steps_per_second": 4.403,
|
| 399 |
+
"step": 500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 4.05,
|
| 403 |
+
"learning_rate": 9.900793650793653e-06,
|
| 404 |
+
"loss": 0.1473,
|
| 405 |
+
"step": 510
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 4.13,
|
| 409 |
+
"learning_rate": 9.702380952380953e-06,
|
| 410 |
+
"loss": 0.0982,
|
| 411 |
+
"step": 520
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 4.21,
|
| 415 |
+
"learning_rate": 9.503968253968255e-06,
|
| 416 |
+
"loss": 0.0967,
|
| 417 |
+
"step": 530
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 4.29,
|
| 421 |
+
"learning_rate": 9.305555555555557e-06,
|
| 422 |
+
"loss": 0.1019,
|
| 423 |
+
"step": 540
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 4.37,
|
| 427 |
+
"learning_rate": 9.107142857142858e-06,
|
| 428 |
+
"loss": 0.1369,
|
| 429 |
+
"step": 550
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 4.37,
|
| 433 |
+
"eval_accuracy": 0.7040358744394619,
|
| 434 |
+
"eval_loss": 0.7594305276870728,
|
| 435 |
+
"eval_runtime": 12.7574,
|
| 436 |
+
"eval_samples_per_second": 34.96,
|
| 437 |
+
"eval_steps_per_second": 4.39,
|
| 438 |
+
"step": 550
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 4.44,
|
| 442 |
+
"learning_rate": 8.90873015873016e-06,
|
| 443 |
+
"loss": 0.0901,
|
| 444 |
+
"step": 560
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 4.52,
|
| 448 |
+
"learning_rate": 8.710317460317462e-06,
|
| 449 |
+
"loss": 0.121,
|
| 450 |
+
"step": 570
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 4.6,
|
| 454 |
+
"learning_rate": 8.511904761904762e-06,
|
| 455 |
+
"loss": 0.0922,
|
| 456 |
+
"step": 580
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 4.68,
|
| 460 |
+
"learning_rate": 8.313492063492064e-06,
|
| 461 |
+
"loss": 0.112,
|
| 462 |
+
"step": 590
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 4.76,
|
| 466 |
+
"learning_rate": 8.115079365079366e-06,
|
| 467 |
+
"loss": 0.1177,
|
| 468 |
+
"step": 600
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 4.76,
|
| 472 |
+
"eval_accuracy": 0.7062780269058296,
|
| 473 |
+
"eval_loss": 0.7682589292526245,
|
| 474 |
+
"eval_runtime": 12.7777,
|
| 475 |
+
"eval_samples_per_second": 34.905,
|
| 476 |
+
"eval_steps_per_second": 4.383,
|
| 477 |
+
"step": 600
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 4.84,
|
| 481 |
+
"learning_rate": 7.916666666666667e-06,
|
| 482 |
+
"loss": 0.1057,
|
| 483 |
+
"step": 610
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 4.92,
|
| 487 |
+
"learning_rate": 7.718253968253969e-06,
|
| 488 |
+
"loss": 0.1122,
|
| 489 |
+
"step": 620
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 5.0,
|
| 493 |
+
"learning_rate": 7.519841269841271e-06,
|
| 494 |
+
"loss": 0.1105,
|
| 495 |
+
"step": 630
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 5.08,
|
| 499 |
+
"learning_rate": 7.321428571428572e-06,
|
| 500 |
+
"loss": 0.0735,
|
| 501 |
+
"step": 640
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 5.16,
|
| 505 |
+
"learning_rate": 7.123015873015874e-06,
|
| 506 |
+
"loss": 0.0504,
|
| 507 |
+
"step": 650
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 5.16,
|
| 511 |
+
"eval_accuracy": 0.7040358744394619,
|
| 512 |
+
"eval_loss": 0.7936971783638,
|
| 513 |
+
"eval_runtime": 12.9729,
|
| 514 |
+
"eval_samples_per_second": 34.379,
|
| 515 |
+
"eval_steps_per_second": 4.317,
|
| 516 |
+
"step": 650
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 5.24,
|
| 520 |
+
"learning_rate": 6.924603174603175e-06,
|
| 521 |
+
"loss": 0.0705,
|
| 522 |
+
"step": 660
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 5.32,
|
| 526 |
+
"learning_rate": 6.726190476190477e-06,
|
| 527 |
+
"loss": 0.0563,
|
| 528 |
+
"step": 670
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 5.4,
|
| 532 |
+
"learning_rate": 6.5277777777777784e-06,
|
| 533 |
+
"loss": 0.0683,
|
| 534 |
+
"step": 680
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 5.48,
|
| 538 |
+
"learning_rate": 6.32936507936508e-06,
|
| 539 |
+
"loss": 0.0698,
|
| 540 |
+
"step": 690
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 5.56,
|
| 544 |
+
"learning_rate": 6.130952380952382e-06,
|
| 545 |
+
"loss": 0.0797,
|
| 546 |
+
"step": 700
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 5.56,
|
| 550 |
+
"eval_accuracy": 0.7062780269058296,
|
| 551 |
+
"eval_loss": 0.8110105395317078,
|
| 552 |
+
"eval_runtime": 12.8549,
|
| 553 |
+
"eval_samples_per_second": 34.695,
|
| 554 |
+
"eval_steps_per_second": 4.356,
|
| 555 |
+
"step": 700
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 5.63,
|
| 559 |
+
"learning_rate": 5.932539682539683e-06,
|
| 560 |
+
"loss": 0.0628,
|
| 561 |
+
"step": 710
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 5.71,
|
| 565 |
+
"learning_rate": 5.734126984126984e-06,
|
| 566 |
+
"loss": 0.0794,
|
| 567 |
+
"step": 720
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 5.79,
|
| 571 |
+
"learning_rate": 5.535714285714286e-06,
|
| 572 |
+
"loss": 0.0809,
|
| 573 |
+
"step": 730
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 5.87,
|
| 577 |
+
"learning_rate": 5.337301587301587e-06,
|
| 578 |
+
"loss": 0.1132,
|
| 579 |
+
"step": 740
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 5.95,
|
| 583 |
+
"learning_rate": 5.138888888888889e-06,
|
| 584 |
+
"loss": 0.0767,
|
| 585 |
+
"step": 750
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 5.95,
|
| 589 |
+
"eval_accuracy": 0.7107623318385651,
|
| 590 |
+
"eval_loss": 0.8268046975135803,
|
| 591 |
+
"eval_runtime": 12.7986,
|
| 592 |
+
"eval_samples_per_second": 34.848,
|
| 593 |
+
"eval_steps_per_second": 4.375,
|
| 594 |
+
"step": 750
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 6.03,
|
| 598 |
+
"learning_rate": 4.940476190476191e-06,
|
| 599 |
+
"loss": 0.0694,
|
| 600 |
+
"step": 760
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 6.11,
|
| 604 |
+
"learning_rate": 4.742063492063492e-06,
|
| 605 |
+
"loss": 0.0573,
|
| 606 |
+
"step": 770
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 6.19,
|
| 610 |
+
"learning_rate": 4.543650793650794e-06,
|
| 611 |
+
"loss": 0.0548,
|
| 612 |
+
"step": 780
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 6.27,
|
| 616 |
+
"learning_rate": 4.345238095238096e-06,
|
| 617 |
+
"loss": 0.0539,
|
| 618 |
+
"step": 790
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 6.35,
|
| 622 |
+
"learning_rate": 4.146825396825397e-06,
|
| 623 |
+
"loss": 0.0597,
|
| 624 |
+
"step": 800
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 6.35,
|
| 628 |
+
"eval_accuracy": 0.7062780269058296,
|
| 629 |
+
"eval_loss": 0.8549516797065735,
|
| 630 |
+
"eval_runtime": 12.8211,
|
| 631 |
+
"eval_samples_per_second": 34.786,
|
| 632 |
+
"eval_steps_per_second": 4.368,
|
| 633 |
+
"step": 800
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 6.43,
|
| 637 |
+
"learning_rate": 3.9484126984126985e-06,
|
| 638 |
+
"loss": 0.0572,
|
| 639 |
+
"step": 810
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 6.51,
|
| 643 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 644 |
+
"loss": 0.0579,
|
| 645 |
+
"step": 820
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 6.59,
|
| 649 |
+
"learning_rate": 3.551587301587302e-06,
|
| 650 |
+
"loss": 0.0727,
|
| 651 |
+
"step": 830
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 6.67,
|
| 655 |
+
"learning_rate": 3.3531746031746034e-06,
|
| 656 |
+
"loss": 0.0522,
|
| 657 |
+
"step": 840
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 6.75,
|
| 661 |
+
"learning_rate": 3.154761904761905e-06,
|
| 662 |
+
"loss": 0.0669,
|
| 663 |
+
"step": 850
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 6.75,
|
| 667 |
+
"eval_accuracy": 0.7152466367713004,
|
| 668 |
+
"eval_loss": 0.8530937433242798,
|
| 669 |
+
"eval_runtime": 12.7041,
|
| 670 |
+
"eval_samples_per_second": 35.107,
|
| 671 |
+
"eval_steps_per_second": 4.408,
|
| 672 |
+
"step": 850
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 6.83,
|
| 676 |
+
"learning_rate": 2.9563492063492066e-06,
|
| 677 |
+
"loss": 0.0586,
|
| 678 |
+
"step": 860
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 6.9,
|
| 682 |
+
"learning_rate": 2.7579365079365083e-06,
|
| 683 |
+
"loss": 0.0445,
|
| 684 |
+
"step": 870
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 6.98,
|
| 688 |
+
"learning_rate": 2.5595238095238095e-06,
|
| 689 |
+
"loss": 0.0552,
|
| 690 |
+
"step": 880
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 7.06,
|
| 694 |
+
"learning_rate": 2.361111111111111e-06,
|
| 695 |
+
"loss": 0.0302,
|
| 696 |
+
"step": 890
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 7.14,
|
| 700 |
+
"learning_rate": 2.1626984126984128e-06,
|
| 701 |
+
"loss": 0.0483,
|
| 702 |
+
"step": 900
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 7.14,
|
| 706 |
+
"eval_accuracy": 0.7130044843049327,
|
| 707 |
+
"eval_loss": 0.8547905683517456,
|
| 708 |
+
"eval_runtime": 12.8017,
|
| 709 |
+
"eval_samples_per_second": 34.839,
|
| 710 |
+
"eval_steps_per_second": 4.374,
|
| 711 |
+
"step": 900
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 7.22,
|
| 715 |
+
"learning_rate": 1.9642857142857144e-06,
|
| 716 |
+
"loss": 0.0585,
|
| 717 |
+
"step": 910
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 7.3,
|
| 721 |
+
"learning_rate": 1.765873015873016e-06,
|
| 722 |
+
"loss": 0.0335,
|
| 723 |
+
"step": 920
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 7.38,
|
| 727 |
+
"learning_rate": 1.5674603174603175e-06,
|
| 728 |
+
"loss": 0.037,
|
| 729 |
+
"step": 930
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 7.46,
|
| 733 |
+
"learning_rate": 1.3690476190476193e-06,
|
| 734 |
+
"loss": 0.0396,
|
| 735 |
+
"step": 940
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 7.54,
|
| 739 |
+
"learning_rate": 1.1706349206349208e-06,
|
| 740 |
+
"loss": 0.0587,
|
| 741 |
+
"step": 950
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 7.54,
|
| 745 |
+
"eval_accuracy": 0.7152466367713004,
|
| 746 |
+
"eval_loss": 0.8677387237548828,
|
| 747 |
+
"eval_runtime": 12.7829,
|
| 748 |
+
"eval_samples_per_second": 34.89,
|
| 749 |
+
"eval_steps_per_second": 4.381,
|
| 750 |
+
"step": 950
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 7.62,
|
| 754 |
+
"learning_rate": 9.722222222222224e-07,
|
| 755 |
+
"loss": 0.0596,
|
| 756 |
+
"step": 960
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 7.7,
|
| 760 |
+
"learning_rate": 7.738095238095239e-07,
|
| 761 |
+
"loss": 0.0418,
|
| 762 |
+
"step": 970
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 7.78,
|
| 766 |
+
"learning_rate": 5.753968253968255e-07,
|
| 767 |
+
"loss": 0.0619,
|
| 768 |
+
"step": 980
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 7.86,
|
| 772 |
+
"learning_rate": 3.76984126984127e-07,
|
| 773 |
+
"loss": 0.0508,
|
| 774 |
+
"step": 990
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 7.94,
|
| 778 |
+
"learning_rate": 1.7857142857142858e-07,
|
| 779 |
+
"loss": 0.038,
|
| 780 |
+
"step": 1000
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 7.94,
|
| 784 |
+
"eval_accuracy": 0.7130044843049327,
|
| 785 |
+
"eval_loss": 0.8715421557426453,
|
| 786 |
+
"eval_runtime": 12.7857,
|
| 787 |
+
"eval_samples_per_second": 34.883,
|
| 788 |
+
"eval_steps_per_second": 4.38,
|
| 789 |
+
"step": 1000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 8.0,
|
| 793 |
+
"step": 1008,
|
| 794 |
+
"total_flos": 7.325059066442809e+18,
|
| 795 |
+
"train_loss": 0.2753036499082569,
|
| 796 |
+
"train_runtime": 1319.5023,
|
| 797 |
+
"train_samples_per_second": 24.294,
|
| 798 |
+
"train_steps_per_second": 0.764
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 8.0,
|
| 802 |
+
"eval_accuracy": 0.7152466367713004,
|
| 803 |
+
"eval_loss": 0.8530937433242798,
|
| 804 |
+
"eval_runtime": 12.7761,
|
| 805 |
+
"eval_samples_per_second": 34.909,
|
| 806 |
+
"eval_steps_per_second": 4.383,
|
| 807 |
+
"step": 1008
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 8.0,
|
| 811 |
+
"eval_accuracy": 0.9852757674070377,
|
| 812 |
+
"eval_loss": 0.04634059593081474,
|
| 813 |
+
"eval_runtime": 115.7634,
|
| 814 |
+
"eval_samples_per_second": 34.614,
|
| 815 |
+
"eval_steps_per_second": 4.328,
|
| 816 |
+
"step": 1008
|
| 817 |
+
}
|
| 818 |
+
],
|
| 819 |
+
"logging_steps": 10,
|
| 820 |
+
"max_steps": 1008,
|
| 821 |
+
"num_train_epochs": 8,
|
| 822 |
+
"save_steps": 50,
|
| 823 |
+
"total_flos": 7.325059066442809e+18,
|
| 824 |
+
"trial_name": null,
|
| 825 |
+
"trial_params": null
|
| 826 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8983f69236ea0c47057fa5eebd7a50ad05734d7f0a1b1fac1f1d4ed9b60b99b
|
| 3 |
+
size 4600
|