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 +894 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9821428571428571,
|
| 4 |
+
"eval_loss": 0.07409419864416122,
|
| 5 |
+
"eval_runtime": 9.1741,
|
| 6 |
+
"eval_samples_per_second": 24.416,
|
| 7 |
+
"eval_steps_per_second": 3.052,
|
| 8 |
+
"train_loss": 0.05678829440015107,
|
| 9 |
+
"train_runtime": 1104.709,
|
| 10 |
+
"train_samples_per_second": 18.322,
|
| 11 |
+
"train_steps_per_second": 1.145
|
| 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": "left_hand",
|
| 13 |
+
"1": "right_hand"
|
| 14 |
+
},
|
| 15 |
+
"image_size": 224,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"left_hand": "0",
|
| 20 |
+
"right_hand": "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": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9821428571428571,
|
| 4 |
+
"eval_loss": 0.07409419864416122,
|
| 5 |
+
"eval_runtime": 9.1741,
|
| 6 |
+
"eval_samples_per_second": 24.416,
|
| 7 |
+
"eval_steps_per_second": 3.052
|
| 8 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6eeead13af0a081952255c7970285fedbbbe6323ae2fa0637ed5b21ebb4504ef
|
| 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": 5.0,
|
| 3 |
+
"train_loss": 0.05678829440015107,
|
| 4 |
+
"train_runtime": 1104.709,
|
| 5 |
+
"train_samples_per_second": 18.322,
|
| 6 |
+
"train_steps_per_second": 1.145
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,894 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.07409419864416122,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-hand-classifier-9\\checkpoint-1200",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 1265,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.04,
|
| 13 |
+
"learning_rate": 4.960474308300396e-05,
|
| 14 |
+
"loss": 0.6889,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.08,
|
| 19 |
+
"learning_rate": 4.9209486166007906e-05,
|
| 20 |
+
"loss": 0.6524,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.12,
|
| 25 |
+
"learning_rate": 4.881422924901186e-05,
|
| 26 |
+
"loss": 0.5933,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.16,
|
| 31 |
+
"learning_rate": 4.841897233201581e-05,
|
| 32 |
+
"loss": 0.5607,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.2,
|
| 37 |
+
"learning_rate": 4.8023715415019764e-05,
|
| 38 |
+
"loss": 0.5662,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.24,
|
| 43 |
+
"learning_rate": 4.762845849802372e-05,
|
| 44 |
+
"loss": 0.5097,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.28,
|
| 49 |
+
"learning_rate": 4.723320158102767e-05,
|
| 50 |
+
"loss": 0.436,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.32,
|
| 55 |
+
"learning_rate": 4.683794466403162e-05,
|
| 56 |
+
"loss": 0.3855,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.36,
|
| 61 |
+
"learning_rate": 4.644268774703557e-05,
|
| 62 |
+
"loss": 0.2617,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.4,
|
| 67 |
+
"learning_rate": 4.6047430830039526e-05,
|
| 68 |
+
"loss": 0.2234,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.4,
|
| 73 |
+
"eval_accuracy": 0.9196428571428571,
|
| 74 |
+
"eval_loss": 0.20466722548007965,
|
| 75 |
+
"eval_runtime": 13.5297,
|
| 76 |
+
"eval_samples_per_second": 16.556,
|
| 77 |
+
"eval_steps_per_second": 2.07,
|
| 78 |
+
"step": 100
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.43,
|
| 82 |
+
"learning_rate": 4.565217391304348e-05,
|
| 83 |
+
"loss": 0.1498,
|
| 84 |
+
"step": 110
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.47,
|
| 88 |
+
"learning_rate": 4.525691699604743e-05,
|
| 89 |
+
"loss": 0.1202,
|
| 90 |
+
"step": 120
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.51,
|
| 94 |
+
"learning_rate": 4.4861660079051384e-05,
|
| 95 |
+
"loss": 0.0587,
|
| 96 |
+
"step": 130
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.55,
|
| 100 |
+
"learning_rate": 4.446640316205534e-05,
|
| 101 |
+
"loss": 0.1286,
|
| 102 |
+
"step": 140
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.59,
|
| 106 |
+
"learning_rate": 4.4071146245059295e-05,
|
| 107 |
+
"loss": 0.0628,
|
| 108 |
+
"step": 150
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.63,
|
| 112 |
+
"learning_rate": 4.367588932806324e-05,
|
| 113 |
+
"loss": 0.0803,
|
| 114 |
+
"step": 160
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.67,
|
| 118 |
+
"learning_rate": 4.32806324110672e-05,
|
| 119 |
+
"loss": 0.0622,
|
| 120 |
+
"step": 170
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.71,
|
| 124 |
+
"learning_rate": 4.288537549407115e-05,
|
| 125 |
+
"loss": 0.1115,
|
| 126 |
+
"step": 180
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.75,
|
| 130 |
+
"learning_rate": 4.24901185770751e-05,
|
| 131 |
+
"loss": 0.1047,
|
| 132 |
+
"step": 190
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 0.79,
|
| 136 |
+
"learning_rate": 4.2094861660079056e-05,
|
| 137 |
+
"loss": 0.0905,
|
| 138 |
+
"step": 200
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 0.79,
|
| 142 |
+
"eval_accuracy": 0.96875,
|
| 143 |
+
"eval_loss": 0.0953177735209465,
|
| 144 |
+
"eval_runtime": 9.8248,
|
| 145 |
+
"eval_samples_per_second": 22.799,
|
| 146 |
+
"eval_steps_per_second": 2.85,
|
| 147 |
+
"step": 200
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.83,
|
| 151 |
+
"learning_rate": 4.1699604743083005e-05,
|
| 152 |
+
"loss": 0.0506,
|
| 153 |
+
"step": 210
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.87,
|
| 157 |
+
"learning_rate": 4.130434782608696e-05,
|
| 158 |
+
"loss": 0.0257,
|
| 159 |
+
"step": 220
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.91,
|
| 163 |
+
"learning_rate": 4.0909090909090915e-05,
|
| 164 |
+
"loss": 0.0165,
|
| 165 |
+
"step": 230
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.95,
|
| 169 |
+
"learning_rate": 4.051383399209486e-05,
|
| 170 |
+
"loss": 0.0876,
|
| 171 |
+
"step": 240
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.99,
|
| 175 |
+
"learning_rate": 4.011857707509882e-05,
|
| 176 |
+
"loss": 0.0508,
|
| 177 |
+
"step": 250
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.03,
|
| 181 |
+
"learning_rate": 3.9723320158102766e-05,
|
| 182 |
+
"loss": 0.0192,
|
| 183 |
+
"step": 260
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.07,
|
| 187 |
+
"learning_rate": 3.932806324110672e-05,
|
| 188 |
+
"loss": 0.0144,
|
| 189 |
+
"step": 270
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.11,
|
| 193 |
+
"learning_rate": 3.893280632411067e-05,
|
| 194 |
+
"loss": 0.0206,
|
| 195 |
+
"step": 280
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.15,
|
| 199 |
+
"learning_rate": 3.8537549407114625e-05,
|
| 200 |
+
"loss": 0.0365,
|
| 201 |
+
"step": 290
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.19,
|
| 205 |
+
"learning_rate": 3.814229249011858e-05,
|
| 206 |
+
"loss": 0.0614,
|
| 207 |
+
"step": 300
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.19,
|
| 211 |
+
"eval_accuracy": 0.9732142857142857,
|
| 212 |
+
"eval_loss": 0.08801885694265366,
|
| 213 |
+
"eval_runtime": 9.6908,
|
| 214 |
+
"eval_samples_per_second": 23.115,
|
| 215 |
+
"eval_steps_per_second": 2.889,
|
| 216 |
+
"step": 300
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 1.23,
|
| 220 |
+
"learning_rate": 3.774703557312253e-05,
|
| 221 |
+
"loss": 0.0209,
|
| 222 |
+
"step": 310
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 1.26,
|
| 226 |
+
"learning_rate": 3.735177865612648e-05,
|
| 227 |
+
"loss": 0.0117,
|
| 228 |
+
"step": 320
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.3,
|
| 232 |
+
"learning_rate": 3.695652173913043e-05,
|
| 233 |
+
"loss": 0.0096,
|
| 234 |
+
"step": 330
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.34,
|
| 238 |
+
"learning_rate": 3.656126482213439e-05,
|
| 239 |
+
"loss": 0.0399,
|
| 240 |
+
"step": 340
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.38,
|
| 244 |
+
"learning_rate": 3.616600790513834e-05,
|
| 245 |
+
"loss": 0.0089,
|
| 246 |
+
"step": 350
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 1.42,
|
| 250 |
+
"learning_rate": 3.577075098814229e-05,
|
| 251 |
+
"loss": 0.0657,
|
| 252 |
+
"step": 360
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.46,
|
| 256 |
+
"learning_rate": 3.5375494071146245e-05,
|
| 257 |
+
"loss": 0.01,
|
| 258 |
+
"step": 370
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 1.5,
|
| 262 |
+
"learning_rate": 3.49802371541502e-05,
|
| 263 |
+
"loss": 0.0094,
|
| 264 |
+
"step": 380
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.54,
|
| 268 |
+
"learning_rate": 3.4584980237154155e-05,
|
| 269 |
+
"loss": 0.0336,
|
| 270 |
+
"step": 390
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.58,
|
| 274 |
+
"learning_rate": 3.418972332015811e-05,
|
| 275 |
+
"loss": 0.0166,
|
| 276 |
+
"step": 400
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 1.58,
|
| 280 |
+
"eval_accuracy": 0.9821428571428571,
|
| 281 |
+
"eval_loss": 0.08657606691122055,
|
| 282 |
+
"eval_runtime": 10.0499,
|
| 283 |
+
"eval_samples_per_second": 22.289,
|
| 284 |
+
"eval_steps_per_second": 2.786,
|
| 285 |
+
"step": 400
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 1.62,
|
| 289 |
+
"learning_rate": 3.379446640316206e-05,
|
| 290 |
+
"loss": 0.0103,
|
| 291 |
+
"step": 410
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.66,
|
| 295 |
+
"learning_rate": 3.3438735177865614e-05,
|
| 296 |
+
"loss": 0.0239,
|
| 297 |
+
"step": 420
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.7,
|
| 301 |
+
"learning_rate": 3.304347826086956e-05,
|
| 302 |
+
"loss": 0.0075,
|
| 303 |
+
"step": 430
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.74,
|
| 307 |
+
"learning_rate": 3.264822134387352e-05,
|
| 308 |
+
"loss": 0.0065,
|
| 309 |
+
"step": 440
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.78,
|
| 313 |
+
"learning_rate": 3.225296442687747e-05,
|
| 314 |
+
"loss": 0.0065,
|
| 315 |
+
"step": 450
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.82,
|
| 319 |
+
"learning_rate": 3.185770750988142e-05,
|
| 320 |
+
"loss": 0.0061,
|
| 321 |
+
"step": 460
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.86,
|
| 325 |
+
"learning_rate": 3.1462450592885376e-05,
|
| 326 |
+
"loss": 0.0711,
|
| 327 |
+
"step": 470
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.9,
|
| 331 |
+
"learning_rate": 3.106719367588933e-05,
|
| 332 |
+
"loss": 0.0062,
|
| 333 |
+
"step": 480
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.94,
|
| 337 |
+
"learning_rate": 3.0671936758893286e-05,
|
| 338 |
+
"loss": 0.0067,
|
| 339 |
+
"step": 490
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.98,
|
| 343 |
+
"learning_rate": 3.027667984189723e-05,
|
| 344 |
+
"loss": 0.0381,
|
| 345 |
+
"step": 500
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.98,
|
| 349 |
+
"eval_accuracy": 0.9821428571428571,
|
| 350 |
+
"eval_loss": 0.08877144753932953,
|
| 351 |
+
"eval_runtime": 9.8395,
|
| 352 |
+
"eval_samples_per_second": 22.765,
|
| 353 |
+
"eval_steps_per_second": 2.846,
|
| 354 |
+
"step": 500
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.02,
|
| 358 |
+
"learning_rate": 2.9881422924901186e-05,
|
| 359 |
+
"loss": 0.0391,
|
| 360 |
+
"step": 510
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 2.06,
|
| 364 |
+
"learning_rate": 2.948616600790514e-05,
|
| 365 |
+
"loss": 0.0075,
|
| 366 |
+
"step": 520
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 2.09,
|
| 370 |
+
"learning_rate": 2.909090909090909e-05,
|
| 371 |
+
"loss": 0.0058,
|
| 372 |
+
"step": 530
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 2.13,
|
| 376 |
+
"learning_rate": 2.8695652173913044e-05,
|
| 377 |
+
"loss": 0.0053,
|
| 378 |
+
"step": 540
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 2.17,
|
| 382 |
+
"learning_rate": 2.8300395256916996e-05,
|
| 383 |
+
"loss": 0.0055,
|
| 384 |
+
"step": 550
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.21,
|
| 388 |
+
"learning_rate": 2.790513833992095e-05,
|
| 389 |
+
"loss": 0.0549,
|
| 390 |
+
"step": 560
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 2.25,
|
| 394 |
+
"learning_rate": 2.7509881422924906e-05,
|
| 395 |
+
"loss": 0.0067,
|
| 396 |
+
"step": 570
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 2.29,
|
| 400 |
+
"learning_rate": 2.7114624505928854e-05,
|
| 401 |
+
"loss": 0.0176,
|
| 402 |
+
"step": 580
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 2.33,
|
| 406 |
+
"learning_rate": 2.671936758893281e-05,
|
| 407 |
+
"loss": 0.0052,
|
| 408 |
+
"step": 590
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 2.37,
|
| 412 |
+
"learning_rate": 2.6324110671936758e-05,
|
| 413 |
+
"loss": 0.0053,
|
| 414 |
+
"step": 600
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 2.37,
|
| 418 |
+
"eval_accuracy": 0.9776785714285714,
|
| 419 |
+
"eval_loss": 0.10505396872758865,
|
| 420 |
+
"eval_runtime": 9.7768,
|
| 421 |
+
"eval_samples_per_second": 22.911,
|
| 422 |
+
"eval_steps_per_second": 2.864,
|
| 423 |
+
"step": 600
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 2.41,
|
| 427 |
+
"learning_rate": 2.5928853754940713e-05,
|
| 428 |
+
"loss": 0.0049,
|
| 429 |
+
"step": 610
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.45,
|
| 433 |
+
"learning_rate": 2.5533596837944668e-05,
|
| 434 |
+
"loss": 0.0062,
|
| 435 |
+
"step": 620
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 2.49,
|
| 439 |
+
"learning_rate": 2.5138339920948616e-05,
|
| 440 |
+
"loss": 0.0053,
|
| 441 |
+
"step": 630
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.53,
|
| 445 |
+
"learning_rate": 2.4743083003952568e-05,
|
| 446 |
+
"loss": 0.0051,
|
| 447 |
+
"step": 640
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 2.57,
|
| 451 |
+
"learning_rate": 2.4347826086956523e-05,
|
| 452 |
+
"loss": 0.0215,
|
| 453 |
+
"step": 650
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 2.61,
|
| 457 |
+
"learning_rate": 2.3952569169960475e-05,
|
| 458 |
+
"loss": 0.0043,
|
| 459 |
+
"step": 660
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 2.65,
|
| 463 |
+
"learning_rate": 2.3557312252964426e-05,
|
| 464 |
+
"loss": 0.0043,
|
| 465 |
+
"step": 670
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 2.69,
|
| 469 |
+
"learning_rate": 2.316205533596838e-05,
|
| 470 |
+
"loss": 0.0042,
|
| 471 |
+
"step": 680
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.73,
|
| 475 |
+
"learning_rate": 2.2766798418972333e-05,
|
| 476 |
+
"loss": 0.0306,
|
| 477 |
+
"step": 690
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 2.77,
|
| 481 |
+
"learning_rate": 2.2371541501976288e-05,
|
| 482 |
+
"loss": 0.0043,
|
| 483 |
+
"step": 700
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 2.77,
|
| 487 |
+
"eval_accuracy": 0.9776785714285714,
|
| 488 |
+
"eval_loss": 0.08058255165815353,
|
| 489 |
+
"eval_runtime": 10.2106,
|
| 490 |
+
"eval_samples_per_second": 21.938,
|
| 491 |
+
"eval_steps_per_second": 2.742,
|
| 492 |
+
"step": 700
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 2.81,
|
| 496 |
+
"learning_rate": 2.197628458498024e-05,
|
| 497 |
+
"loss": 0.0042,
|
| 498 |
+
"step": 710
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 2.85,
|
| 502 |
+
"learning_rate": 2.158102766798419e-05,
|
| 503 |
+
"loss": 0.0041,
|
| 504 |
+
"step": 720
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 2.89,
|
| 508 |
+
"learning_rate": 2.1185770750988143e-05,
|
| 509 |
+
"loss": 0.0173,
|
| 510 |
+
"step": 730
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 2.92,
|
| 514 |
+
"learning_rate": 2.0790513833992095e-05,
|
| 515 |
+
"loss": 0.0042,
|
| 516 |
+
"step": 740
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 2.96,
|
| 520 |
+
"learning_rate": 2.0395256916996047e-05,
|
| 521 |
+
"loss": 0.0039,
|
| 522 |
+
"step": 750
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 3.0,
|
| 526 |
+
"learning_rate": 2e-05,
|
| 527 |
+
"loss": 0.0039,
|
| 528 |
+
"step": 760
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 3.04,
|
| 532 |
+
"learning_rate": 1.9604743083003953e-05,
|
| 533 |
+
"loss": 0.0036,
|
| 534 |
+
"step": 770
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 3.08,
|
| 538 |
+
"learning_rate": 1.9209486166007905e-05,
|
| 539 |
+
"loss": 0.0037,
|
| 540 |
+
"step": 780
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 3.12,
|
| 544 |
+
"learning_rate": 1.8814229249011857e-05,
|
| 545 |
+
"loss": 0.0036,
|
| 546 |
+
"step": 790
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 3.16,
|
| 550 |
+
"learning_rate": 1.8418972332015812e-05,
|
| 551 |
+
"loss": 0.0035,
|
| 552 |
+
"step": 800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 3.16,
|
| 556 |
+
"eval_accuracy": 0.9732142857142857,
|
| 557 |
+
"eval_loss": 0.08644634485244751,
|
| 558 |
+
"eval_runtime": 9.8015,
|
| 559 |
+
"eval_samples_per_second": 22.854,
|
| 560 |
+
"eval_steps_per_second": 2.857,
|
| 561 |
+
"step": 800
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 3.2,
|
| 565 |
+
"learning_rate": 1.8023715415019763e-05,
|
| 566 |
+
"loss": 0.0035,
|
| 567 |
+
"step": 810
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 3.24,
|
| 571 |
+
"learning_rate": 1.762845849802372e-05,
|
| 572 |
+
"loss": 0.0034,
|
| 573 |
+
"step": 820
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 3.28,
|
| 577 |
+
"learning_rate": 1.723320158102767e-05,
|
| 578 |
+
"loss": 0.0035,
|
| 579 |
+
"step": 830
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 3.32,
|
| 583 |
+
"learning_rate": 1.6837944664031622e-05,
|
| 584 |
+
"loss": 0.0223,
|
| 585 |
+
"step": 840
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 3.36,
|
| 589 |
+
"learning_rate": 1.6442687747035574e-05,
|
| 590 |
+
"loss": 0.0035,
|
| 591 |
+
"step": 850
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 3.4,
|
| 595 |
+
"learning_rate": 1.6047430830039525e-05,
|
| 596 |
+
"loss": 0.0035,
|
| 597 |
+
"step": 860
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 3.44,
|
| 601 |
+
"learning_rate": 1.565217391304348e-05,
|
| 602 |
+
"loss": 0.0175,
|
| 603 |
+
"step": 870
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 3.48,
|
| 607 |
+
"learning_rate": 1.5256916996047432e-05,
|
| 608 |
+
"loss": 0.0044,
|
| 609 |
+
"step": 880
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 3.52,
|
| 613 |
+
"learning_rate": 1.4861660079051384e-05,
|
| 614 |
+
"loss": 0.0032,
|
| 615 |
+
"step": 890
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 3.56,
|
| 619 |
+
"learning_rate": 1.4466403162055337e-05,
|
| 620 |
+
"loss": 0.0087,
|
| 621 |
+
"step": 900
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 3.56,
|
| 625 |
+
"eval_accuracy": 0.9821428571428571,
|
| 626 |
+
"eval_loss": 0.08384133875370026,
|
| 627 |
+
"eval_runtime": 10.4471,
|
| 628 |
+
"eval_samples_per_second": 21.441,
|
| 629 |
+
"eval_steps_per_second": 2.68,
|
| 630 |
+
"step": 900
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 3.6,
|
| 634 |
+
"learning_rate": 1.4071146245059289e-05,
|
| 635 |
+
"loss": 0.0071,
|
| 636 |
+
"step": 910
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 3.64,
|
| 640 |
+
"learning_rate": 1.367588932806324e-05,
|
| 641 |
+
"loss": 0.0032,
|
| 642 |
+
"step": 920
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 3.68,
|
| 646 |
+
"learning_rate": 1.3280632411067195e-05,
|
| 647 |
+
"loss": 0.0081,
|
| 648 |
+
"step": 930
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 3.72,
|
| 652 |
+
"learning_rate": 1.2885375494071147e-05,
|
| 653 |
+
"loss": 0.0031,
|
| 654 |
+
"step": 940
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 3.75,
|
| 658 |
+
"learning_rate": 1.2490118577075099e-05,
|
| 659 |
+
"loss": 0.0031,
|
| 660 |
+
"step": 950
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 3.79,
|
| 664 |
+
"learning_rate": 1.2094861660079052e-05,
|
| 665 |
+
"loss": 0.0031,
|
| 666 |
+
"step": 960
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 3.83,
|
| 670 |
+
"learning_rate": 1.1699604743083006e-05,
|
| 671 |
+
"loss": 0.003,
|
| 672 |
+
"step": 970
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 3.87,
|
| 676 |
+
"learning_rate": 1.1304347826086957e-05,
|
| 677 |
+
"loss": 0.003,
|
| 678 |
+
"step": 980
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 3.91,
|
| 682 |
+
"learning_rate": 1.0909090909090909e-05,
|
| 683 |
+
"loss": 0.0179,
|
| 684 |
+
"step": 990
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 3.95,
|
| 688 |
+
"learning_rate": 1.0513833992094862e-05,
|
| 689 |
+
"loss": 0.003,
|
| 690 |
+
"step": 1000
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 3.95,
|
| 694 |
+
"eval_accuracy": 0.9821428571428571,
|
| 695 |
+
"eval_loss": 0.07715751230716705,
|
| 696 |
+
"eval_runtime": 9.6117,
|
| 697 |
+
"eval_samples_per_second": 23.305,
|
| 698 |
+
"eval_steps_per_second": 2.913,
|
| 699 |
+
"step": 1000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 3.99,
|
| 703 |
+
"learning_rate": 1.0118577075098814e-05,
|
| 704 |
+
"loss": 0.003,
|
| 705 |
+
"step": 1010
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 4.03,
|
| 709 |
+
"learning_rate": 9.723320158102767e-06,
|
| 710 |
+
"loss": 0.0046,
|
| 711 |
+
"step": 1020
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 4.07,
|
| 715 |
+
"learning_rate": 9.32806324110672e-06,
|
| 716 |
+
"loss": 0.0044,
|
| 717 |
+
"step": 1030
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 4.11,
|
| 721 |
+
"learning_rate": 8.932806324110672e-06,
|
| 722 |
+
"loss": 0.0029,
|
| 723 |
+
"step": 1040
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 4.15,
|
| 727 |
+
"learning_rate": 8.537549407114626e-06,
|
| 728 |
+
"loss": 0.0029,
|
| 729 |
+
"step": 1050
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 4.19,
|
| 733 |
+
"learning_rate": 8.142292490118577e-06,
|
| 734 |
+
"loss": 0.0061,
|
| 735 |
+
"step": 1060
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 4.23,
|
| 739 |
+
"learning_rate": 7.747035573122529e-06,
|
| 740 |
+
"loss": 0.0029,
|
| 741 |
+
"step": 1070
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 4.27,
|
| 745 |
+
"learning_rate": 7.351778656126483e-06,
|
| 746 |
+
"loss": 0.0028,
|
| 747 |
+
"step": 1080
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 4.31,
|
| 751 |
+
"learning_rate": 6.956521739130435e-06,
|
| 752 |
+
"loss": 0.0029,
|
| 753 |
+
"step": 1090
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 4.35,
|
| 757 |
+
"learning_rate": 6.5612648221343875e-06,
|
| 758 |
+
"loss": 0.0147,
|
| 759 |
+
"step": 1100
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 4.35,
|
| 763 |
+
"eval_accuracy": 0.9821428571428571,
|
| 764 |
+
"eval_loss": 0.07519398629665375,
|
| 765 |
+
"eval_runtime": 10.7399,
|
| 766 |
+
"eval_samples_per_second": 20.857,
|
| 767 |
+
"eval_steps_per_second": 2.607,
|
| 768 |
+
"step": 1100
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 4.39,
|
| 772 |
+
"learning_rate": 6.16600790513834e-06,
|
| 773 |
+
"loss": 0.0029,
|
| 774 |
+
"step": 1110
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 4.43,
|
| 778 |
+
"learning_rate": 5.770750988142293e-06,
|
| 779 |
+
"loss": 0.0029,
|
| 780 |
+
"step": 1120
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 4.47,
|
| 784 |
+
"learning_rate": 5.375494071146245e-06,
|
| 785 |
+
"loss": 0.0029,
|
| 786 |
+
"step": 1130
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 4.51,
|
| 790 |
+
"learning_rate": 4.980237154150198e-06,
|
| 791 |
+
"loss": 0.0029,
|
| 792 |
+
"step": 1140
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 4.55,
|
| 796 |
+
"learning_rate": 4.58498023715415e-06,
|
| 797 |
+
"loss": 0.0029,
|
| 798 |
+
"step": 1150
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 4.58,
|
| 802 |
+
"learning_rate": 4.189723320158103e-06,
|
| 803 |
+
"loss": 0.0029,
|
| 804 |
+
"step": 1160
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 4.62,
|
| 808 |
+
"learning_rate": 3.7944664031620557e-06,
|
| 809 |
+
"loss": 0.0028,
|
| 810 |
+
"step": 1170
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 4.66,
|
| 814 |
+
"learning_rate": 3.3992094861660078e-06,
|
| 815 |
+
"loss": 0.0028,
|
| 816 |
+
"step": 1180
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 4.7,
|
| 820 |
+
"learning_rate": 3.0039525691699607e-06,
|
| 821 |
+
"loss": 0.0028,
|
| 822 |
+
"step": 1190
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 4.74,
|
| 826 |
+
"learning_rate": 2.608695652173913e-06,
|
| 827 |
+
"loss": 0.0052,
|
| 828 |
+
"step": 1200
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 4.74,
|
| 832 |
+
"eval_accuracy": 0.9821428571428571,
|
| 833 |
+
"eval_loss": 0.07409419864416122,
|
| 834 |
+
"eval_runtime": 10.6746,
|
| 835 |
+
"eval_samples_per_second": 20.984,
|
| 836 |
+
"eval_steps_per_second": 2.623,
|
| 837 |
+
"step": 1200
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 4.78,
|
| 841 |
+
"learning_rate": 2.2134387351778658e-06,
|
| 842 |
+
"loss": 0.0028,
|
| 843 |
+
"step": 1210
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 4.82,
|
| 847 |
+
"learning_rate": 1.818181818181818e-06,
|
| 848 |
+
"loss": 0.0028,
|
| 849 |
+
"step": 1220
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 4.86,
|
| 853 |
+
"learning_rate": 1.4229249011857708e-06,
|
| 854 |
+
"loss": 0.0102,
|
| 855 |
+
"step": 1230
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 4.9,
|
| 859 |
+
"learning_rate": 1.0276679841897233e-06,
|
| 860 |
+
"loss": 0.0028,
|
| 861 |
+
"step": 1240
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 4.94,
|
| 865 |
+
"learning_rate": 6.324110671936759e-07,
|
| 866 |
+
"loss": 0.0122,
|
| 867 |
+
"step": 1250
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 4.98,
|
| 871 |
+
"learning_rate": 2.3715415019762848e-07,
|
| 872 |
+
"loss": 0.0098,
|
| 873 |
+
"step": 1260
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 5.0,
|
| 877 |
+
"step": 1265,
|
| 878 |
+
"total_flos": 1.568437869799342e+18,
|
| 879 |
+
"train_loss": 0.05678829440015107,
|
| 880 |
+
"train_runtime": 1104.709,
|
| 881 |
+
"train_samples_per_second": 18.322,
|
| 882 |
+
"train_steps_per_second": 1.145
|
| 883 |
+
}
|
| 884 |
+
],
|
| 885 |
+
"logging_steps": 10,
|
| 886 |
+
"max_steps": 1265,
|
| 887 |
+
"num_input_tokens_seen": 0,
|
| 888 |
+
"num_train_epochs": 5,
|
| 889 |
+
"save_steps": 100,
|
| 890 |
+
"total_flos": 1.568437869799342e+18,
|
| 891 |
+
"train_batch_size": 16,
|
| 892 |
+
"trial_name": null,
|
| 893 |
+
"trial_params": null
|
| 894 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc6b0a66eacdb9d5b03a4a01952bedf1a5033d5c609888b70f133498f00cb363
|
| 3 |
+
size 4728
|