tebak-gambar-mobilevit
Browse files- README.md +70 -0
- all_results.json +13 -0
- config.json +741 -0
- model.safetensors +3 -0
- preprocessor_config.json +17 -0
- test_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +3821 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
base_model: apple/mobilevit-x-small
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
metrics:
|
| 7 |
+
- accuracy
|
| 8 |
+
model-index:
|
| 9 |
+
- name: tebak-gambar-mobilevit
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# tebak-gambar-mobilevit
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [apple/mobilevit-x-small](https://huggingface.co/apple/mobilevit-x-small) on an unknown dataset.
|
| 19 |
+
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 1.0799
|
| 21 |
+
- Accuracy: 0.7289
|
| 22 |
+
|
| 23 |
+
## Model description
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Intended uses & limitations
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training and evaluation data
|
| 32 |
+
|
| 33 |
+
More information needed
|
| 34 |
+
|
| 35 |
+
## Training procedure
|
| 36 |
+
|
| 37 |
+
### Training hyperparameters
|
| 38 |
+
|
| 39 |
+
The following hyperparameters were used during training:
|
| 40 |
+
- learning_rate: 0.0008
|
| 41 |
+
- train_batch_size: 256
|
| 42 |
+
- eval_batch_size: 256
|
| 43 |
+
- seed: 42
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: linear
|
| 46 |
+
- num_epochs: 3
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 52 |
+
|:-------------:|:------:|:-----:|:---------------:|:--------:|
|
| 53 |
+
| 1.5084 | 0.2844 | 5000 | 1.4700 | 0.6364 |
|
| 54 |
+
| 1.3684 | 0.5689 | 10000 | 1.3353 | 0.6674 |
|
| 55 |
+
| 1.3568 | 0.8533 | 15000 | 1.2764 | 0.6804 |
|
| 56 |
+
| 1.226 | 1.1377 | 20000 | 1.2323 | 0.6924 |
|
| 57 |
+
| 1.2125 | 1.4222 | 25000 | 1.1850 | 0.7031 |
|
| 58 |
+
| 1.1912 | 1.7066 | 30000 | 1.1567 | 0.7092 |
|
| 59 |
+
| 1.1902 | 1.9910 | 35000 | 1.1297 | 0.7165 |
|
| 60 |
+
| 1.131 | 2.2754 | 40000 | 1.1106 | 0.7213 |
|
| 61 |
+
| 1.124 | 2.5599 | 45000 | 1.0916 | 0.7258 |
|
| 62 |
+
| 1.1245 | 2.8443 | 50000 | 1.0782 | 0.7300 |
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
### Framework versions
|
| 66 |
+
|
| 67 |
+
- Transformers 4.42.4
|
| 68 |
+
- Pytorch 2.3.1+cu121
|
| 69 |
+
- Datasets 2.20.0
|
| 70 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.728948,
|
| 4 |
+
"eval_loss": 1.0799410343170166,
|
| 5 |
+
"eval_runtime": 79.2396,
|
| 6 |
+
"eval_samples_per_second": 3154.988,
|
| 7 |
+
"eval_steps_per_second": 12.33,
|
| 8 |
+
"total_flos": 1.3116020904e+17,
|
| 9 |
+
"train_loss": 1.269093582550002,
|
| 10 |
+
"train_runtime": 7252.1306,
|
| 11 |
+
"train_samples_per_second": 1861.522,
|
| 12 |
+
"train_steps_per_second": 7.272
|
| 13 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,741 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "apple/mobilevit-x-small",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MobileViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"aspp_dropout_prob": 0.1,
|
| 7 |
+
"aspp_out_channels": 256,
|
| 8 |
+
"atrous_rates": [
|
| 9 |
+
6,
|
| 10 |
+
12,
|
| 11 |
+
18
|
| 12 |
+
],
|
| 13 |
+
"attention_probs_dropout_prob": 0.0,
|
| 14 |
+
"classifier_dropout_prob": 0.1,
|
| 15 |
+
"conv_kernel_size": 3,
|
| 16 |
+
"expand_ratio": 4.0,
|
| 17 |
+
"hidden_act": "silu",
|
| 18 |
+
"hidden_dropout_prob": 0.1,
|
| 19 |
+
"hidden_sizes": [
|
| 20 |
+
96,
|
| 21 |
+
120,
|
| 22 |
+
144
|
| 23 |
+
],
|
| 24 |
+
"id2label": {
|
| 25 |
+
"0": "aircraft carrier",
|
| 26 |
+
"1": "airplane",
|
| 27 |
+
"10": "asparagus",
|
| 28 |
+
"100": "dumbbell",
|
| 29 |
+
"101": "ear",
|
| 30 |
+
"102": "elbow",
|
| 31 |
+
"103": "elephant",
|
| 32 |
+
"104": "envelope",
|
| 33 |
+
"105": "eraser",
|
| 34 |
+
"106": "eye",
|
| 35 |
+
"107": "eyeglasses",
|
| 36 |
+
"108": "face",
|
| 37 |
+
"109": "fan",
|
| 38 |
+
"11": "axe",
|
| 39 |
+
"110": "feather",
|
| 40 |
+
"111": "fence",
|
| 41 |
+
"112": "finger",
|
| 42 |
+
"113": "fire hydrant",
|
| 43 |
+
"114": "fireplace",
|
| 44 |
+
"115": "firetruck",
|
| 45 |
+
"116": "fish",
|
| 46 |
+
"117": "flamingo",
|
| 47 |
+
"118": "flashlight",
|
| 48 |
+
"119": "flip flops",
|
| 49 |
+
"12": "backpack",
|
| 50 |
+
"120": "floor lamp",
|
| 51 |
+
"121": "flower",
|
| 52 |
+
"122": "flying saucer",
|
| 53 |
+
"123": "foot",
|
| 54 |
+
"124": "fork",
|
| 55 |
+
"125": "frog",
|
| 56 |
+
"126": "frying pan",
|
| 57 |
+
"127": "garden hose",
|
| 58 |
+
"128": "garden",
|
| 59 |
+
"129": "giraffe",
|
| 60 |
+
"13": "banana",
|
| 61 |
+
"130": "goatee",
|
| 62 |
+
"131": "golf club",
|
| 63 |
+
"132": "grapes",
|
| 64 |
+
"133": "grass",
|
| 65 |
+
"134": "guitar",
|
| 66 |
+
"135": "hamburger",
|
| 67 |
+
"136": "hammer",
|
| 68 |
+
"137": "hand",
|
| 69 |
+
"138": "harp",
|
| 70 |
+
"139": "hat",
|
| 71 |
+
"14": "bandage",
|
| 72 |
+
"140": "headphones",
|
| 73 |
+
"141": "hedgehog",
|
| 74 |
+
"142": "helicopter",
|
| 75 |
+
"143": "helmet",
|
| 76 |
+
"144": "hexagon",
|
| 77 |
+
"145": "hockey puck",
|
| 78 |
+
"146": "hockey stick",
|
| 79 |
+
"147": "horse",
|
| 80 |
+
"148": "hospital",
|
| 81 |
+
"149": "hot air balloon",
|
| 82 |
+
"15": "barn",
|
| 83 |
+
"150": "hot dog",
|
| 84 |
+
"151": "hot tub",
|
| 85 |
+
"152": "hourglass",
|
| 86 |
+
"153": "house plant",
|
| 87 |
+
"154": "house",
|
| 88 |
+
"155": "hurricane",
|
| 89 |
+
"156": "ice cream",
|
| 90 |
+
"157": "jacket",
|
| 91 |
+
"158": "jail",
|
| 92 |
+
"159": "kangaroo",
|
| 93 |
+
"16": "baseball bat",
|
| 94 |
+
"160": "key",
|
| 95 |
+
"161": "keyboard",
|
| 96 |
+
"162": "knee",
|
| 97 |
+
"163": "knife",
|
| 98 |
+
"164": "ladder",
|
| 99 |
+
"165": "lantern",
|
| 100 |
+
"166": "laptop",
|
| 101 |
+
"167": "leaf",
|
| 102 |
+
"168": "leg",
|
| 103 |
+
"169": "light bulb",
|
| 104 |
+
"17": "baseball",
|
| 105 |
+
"170": "lighter",
|
| 106 |
+
"171": "lighthouse",
|
| 107 |
+
"172": "lightning",
|
| 108 |
+
"173": "line",
|
| 109 |
+
"174": "lion",
|
| 110 |
+
"175": "lipstick",
|
| 111 |
+
"176": "lobster",
|
| 112 |
+
"177": "lollipop",
|
| 113 |
+
"178": "mailbox",
|
| 114 |
+
"179": "map",
|
| 115 |
+
"18": "basket",
|
| 116 |
+
"180": "marker",
|
| 117 |
+
"181": "matches",
|
| 118 |
+
"182": "megaphone",
|
| 119 |
+
"183": "mermaid",
|
| 120 |
+
"184": "microphone",
|
| 121 |
+
"185": "microwave",
|
| 122 |
+
"186": "monkey",
|
| 123 |
+
"187": "moon",
|
| 124 |
+
"188": "mosquito",
|
| 125 |
+
"189": "motorbike",
|
| 126 |
+
"19": "basketball",
|
| 127 |
+
"190": "mountain",
|
| 128 |
+
"191": "mouse",
|
| 129 |
+
"192": "moustache",
|
| 130 |
+
"193": "mouth",
|
| 131 |
+
"194": "mug",
|
| 132 |
+
"195": "mushroom",
|
| 133 |
+
"196": "nail",
|
| 134 |
+
"197": "necklace",
|
| 135 |
+
"198": "nose",
|
| 136 |
+
"199": "ocean",
|
| 137 |
+
"2": "alarm clock",
|
| 138 |
+
"20": "bat",
|
| 139 |
+
"200": "octagon",
|
| 140 |
+
"201": "octopus",
|
| 141 |
+
"202": "onion",
|
| 142 |
+
"203": "oven",
|
| 143 |
+
"204": "owl",
|
| 144 |
+
"205": "paint can",
|
| 145 |
+
"206": "paintbrush",
|
| 146 |
+
"207": "palm tree",
|
| 147 |
+
"208": "panda",
|
| 148 |
+
"209": "pants",
|
| 149 |
+
"21": "bathtub",
|
| 150 |
+
"210": "paper clip",
|
| 151 |
+
"211": "parachute",
|
| 152 |
+
"212": "parrot",
|
| 153 |
+
"213": "passport",
|
| 154 |
+
"214": "peanut",
|
| 155 |
+
"215": "pear",
|
| 156 |
+
"216": "peas",
|
| 157 |
+
"217": "pencil",
|
| 158 |
+
"218": "penguin",
|
| 159 |
+
"219": "piano",
|
| 160 |
+
"22": "beach",
|
| 161 |
+
"220": "pickup truck",
|
| 162 |
+
"221": "picture frame",
|
| 163 |
+
"222": "pig",
|
| 164 |
+
"223": "pillow",
|
| 165 |
+
"224": "pineapple",
|
| 166 |
+
"225": "pizza",
|
| 167 |
+
"226": "pliers",
|
| 168 |
+
"227": "police car",
|
| 169 |
+
"228": "pond",
|
| 170 |
+
"229": "pool",
|
| 171 |
+
"23": "bear",
|
| 172 |
+
"230": "popsicle",
|
| 173 |
+
"231": "postcard",
|
| 174 |
+
"232": "potato",
|
| 175 |
+
"233": "power outlet",
|
| 176 |
+
"234": "purse",
|
| 177 |
+
"235": "rabbit",
|
| 178 |
+
"236": "raccoon",
|
| 179 |
+
"237": "radio",
|
| 180 |
+
"238": "rain",
|
| 181 |
+
"239": "rainbow",
|
| 182 |
+
"24": "beard",
|
| 183 |
+
"240": "rake",
|
| 184 |
+
"241": "remote control",
|
| 185 |
+
"242": "rhinoceros",
|
| 186 |
+
"243": "rifle",
|
| 187 |
+
"244": "river",
|
| 188 |
+
"245": "roller coaster",
|
| 189 |
+
"246": "rollerskates",
|
| 190 |
+
"247": "sailboat",
|
| 191 |
+
"248": "sandwich",
|
| 192 |
+
"249": "saw",
|
| 193 |
+
"25": "bed",
|
| 194 |
+
"250": "saxophone",
|
| 195 |
+
"251": "school bus",
|
| 196 |
+
"252": "scissors",
|
| 197 |
+
"253": "scorpion",
|
| 198 |
+
"254": "screwdriver",
|
| 199 |
+
"255": "sea turtle",
|
| 200 |
+
"256": "see saw",
|
| 201 |
+
"257": "shark",
|
| 202 |
+
"258": "sheep",
|
| 203 |
+
"259": "shoe",
|
| 204 |
+
"26": "bee",
|
| 205 |
+
"260": "shorts",
|
| 206 |
+
"261": "shovel",
|
| 207 |
+
"262": "sink",
|
| 208 |
+
"263": "skateboard",
|
| 209 |
+
"264": "skull",
|
| 210 |
+
"265": "skyscraper",
|
| 211 |
+
"266": "sleeping bag",
|
| 212 |
+
"267": "smiley face",
|
| 213 |
+
"268": "snail",
|
| 214 |
+
"269": "snake",
|
| 215 |
+
"27": "belt",
|
| 216 |
+
"270": "snorkel",
|
| 217 |
+
"271": "snowflake",
|
| 218 |
+
"272": "snowman",
|
| 219 |
+
"273": "soccer ball",
|
| 220 |
+
"274": "sock",
|
| 221 |
+
"275": "speedboat",
|
| 222 |
+
"276": "spider",
|
| 223 |
+
"277": "spoon",
|
| 224 |
+
"278": "spreadsheet",
|
| 225 |
+
"279": "square",
|
| 226 |
+
"28": "bench",
|
| 227 |
+
"280": "squiggle",
|
| 228 |
+
"281": "squirrel",
|
| 229 |
+
"282": "stairs",
|
| 230 |
+
"283": "star",
|
| 231 |
+
"284": "steak",
|
| 232 |
+
"285": "stereo",
|
| 233 |
+
"286": "stethoscope",
|
| 234 |
+
"287": "stitches",
|
| 235 |
+
"288": "stop sign",
|
| 236 |
+
"289": "stove",
|
| 237 |
+
"29": "bicycle",
|
| 238 |
+
"290": "strawberry",
|
| 239 |
+
"291": "streetlight",
|
| 240 |
+
"292": "string bean",
|
| 241 |
+
"293": "submarine",
|
| 242 |
+
"294": "suitcase",
|
| 243 |
+
"295": "sun",
|
| 244 |
+
"296": "swan",
|
| 245 |
+
"297": "sweater",
|
| 246 |
+
"298": "swing set",
|
| 247 |
+
"299": "sword",
|
| 248 |
+
"3": "ambulance",
|
| 249 |
+
"30": "binoculars",
|
| 250 |
+
"300": "syringe",
|
| 251 |
+
"301": "t-shirt",
|
| 252 |
+
"302": "table",
|
| 253 |
+
"303": "teapot",
|
| 254 |
+
"304": "teddy-bear",
|
| 255 |
+
"305": "telephone",
|
| 256 |
+
"306": "television",
|
| 257 |
+
"307": "tennis racquet",
|
| 258 |
+
"308": "tent",
|
| 259 |
+
"309": "The Eiffel Tower",
|
| 260 |
+
"31": "bird",
|
| 261 |
+
"310": "The Great Wall of China",
|
| 262 |
+
"311": "The Mona Lisa",
|
| 263 |
+
"312": "tiger",
|
| 264 |
+
"313": "toaster",
|
| 265 |
+
"314": "toe",
|
| 266 |
+
"315": "toilet",
|
| 267 |
+
"316": "tooth",
|
| 268 |
+
"317": "toothbrush",
|
| 269 |
+
"318": "toothpaste",
|
| 270 |
+
"319": "tornado",
|
| 271 |
+
"32": "birthday cake",
|
| 272 |
+
"320": "tractor",
|
| 273 |
+
"321": "traffic light",
|
| 274 |
+
"322": "train",
|
| 275 |
+
"323": "tree",
|
| 276 |
+
"324": "triangle",
|
| 277 |
+
"325": "trombone",
|
| 278 |
+
"326": "truck",
|
| 279 |
+
"327": "trumpet",
|
| 280 |
+
"328": "umbrella",
|
| 281 |
+
"329": "underwear",
|
| 282 |
+
"33": "blackberry",
|
| 283 |
+
"330": "van",
|
| 284 |
+
"331": "vase",
|
| 285 |
+
"332": "violin",
|
| 286 |
+
"333": "washing machine",
|
| 287 |
+
"334": "watermelon",
|
| 288 |
+
"335": "waterslide",
|
| 289 |
+
"336": "whale",
|
| 290 |
+
"337": "wheel",
|
| 291 |
+
"338": "windmill",
|
| 292 |
+
"339": "wine bottle",
|
| 293 |
+
"34": "blueberry",
|
| 294 |
+
"340": "wine glass",
|
| 295 |
+
"341": "wristwatch",
|
| 296 |
+
"342": "yoga",
|
| 297 |
+
"343": "zebra",
|
| 298 |
+
"344": "zigzag",
|
| 299 |
+
"35": "book",
|
| 300 |
+
"36": "boomerang",
|
| 301 |
+
"37": "bottlecap",
|
| 302 |
+
"38": "bowtie",
|
| 303 |
+
"39": "bracelet",
|
| 304 |
+
"4": "angel",
|
| 305 |
+
"40": "brain",
|
| 306 |
+
"41": "bread",
|
| 307 |
+
"42": "bridge",
|
| 308 |
+
"43": "broccoli",
|
| 309 |
+
"44": "broom",
|
| 310 |
+
"45": "bucket",
|
| 311 |
+
"46": "bulldozer",
|
| 312 |
+
"47": "bus",
|
| 313 |
+
"48": "bush",
|
| 314 |
+
"49": "butterfly",
|
| 315 |
+
"5": "animal migration",
|
| 316 |
+
"50": "cactus",
|
| 317 |
+
"51": "cake",
|
| 318 |
+
"52": "calculator",
|
| 319 |
+
"53": "calendar",
|
| 320 |
+
"54": "camel",
|
| 321 |
+
"55": "camera",
|
| 322 |
+
"56": "camouflage",
|
| 323 |
+
"57": "campfire",
|
| 324 |
+
"58": "candle",
|
| 325 |
+
"59": "cannon",
|
| 326 |
+
"6": "ant",
|
| 327 |
+
"60": "canoe",
|
| 328 |
+
"61": "car",
|
| 329 |
+
"62": "carrot",
|
| 330 |
+
"63": "castle",
|
| 331 |
+
"64": "cat",
|
| 332 |
+
"65": "ceiling fan",
|
| 333 |
+
"66": "cell phone",
|
| 334 |
+
"67": "cello",
|
| 335 |
+
"68": "chair",
|
| 336 |
+
"69": "chandelier",
|
| 337 |
+
"7": "anvil",
|
| 338 |
+
"70": "church",
|
| 339 |
+
"71": "circle",
|
| 340 |
+
"72": "clarinet",
|
| 341 |
+
"73": "clock",
|
| 342 |
+
"74": "cloud",
|
| 343 |
+
"75": "coffee cup",
|
| 344 |
+
"76": "compass",
|
| 345 |
+
"77": "computer",
|
| 346 |
+
"78": "cookie",
|
| 347 |
+
"79": "cooler",
|
| 348 |
+
"8": "apple",
|
| 349 |
+
"80": "couch",
|
| 350 |
+
"81": "cow",
|
| 351 |
+
"82": "crab",
|
| 352 |
+
"83": "crayon",
|
| 353 |
+
"84": "crocodile",
|
| 354 |
+
"85": "crown",
|
| 355 |
+
"86": "cruise ship",
|
| 356 |
+
"87": "cup",
|
| 357 |
+
"88": "diamond",
|
| 358 |
+
"89": "dishwasher",
|
| 359 |
+
"9": "arm",
|
| 360 |
+
"90": "diving board",
|
| 361 |
+
"91": "dog",
|
| 362 |
+
"92": "dolphin",
|
| 363 |
+
"93": "donut",
|
| 364 |
+
"94": "door",
|
| 365 |
+
"95": "dragon",
|
| 366 |
+
"96": "dresser",
|
| 367 |
+
"97": "drill",
|
| 368 |
+
"98": "drums",
|
| 369 |
+
"99": "duck"
|
| 370 |
+
},
|
| 371 |
+
"image_size": 28,
|
| 372 |
+
"initializer_range": 0.02,
|
| 373 |
+
"label2id": {
|
| 374 |
+
"The Eiffel Tower": "309",
|
| 375 |
+
"The Great Wall of China": "310",
|
| 376 |
+
"The Mona Lisa": "311",
|
| 377 |
+
"aircraft carrier": "0",
|
| 378 |
+
"airplane": "1",
|
| 379 |
+
"alarm clock": "2",
|
| 380 |
+
"ambulance": "3",
|
| 381 |
+
"angel": "4",
|
| 382 |
+
"animal migration": "5",
|
| 383 |
+
"ant": "6",
|
| 384 |
+
"anvil": "7",
|
| 385 |
+
"apple": "8",
|
| 386 |
+
"arm": "9",
|
| 387 |
+
"asparagus": "10",
|
| 388 |
+
"axe": "11",
|
| 389 |
+
"backpack": "12",
|
| 390 |
+
"banana": "13",
|
| 391 |
+
"bandage": "14",
|
| 392 |
+
"barn": "15",
|
| 393 |
+
"baseball": "17",
|
| 394 |
+
"baseball bat": "16",
|
| 395 |
+
"basket": "18",
|
| 396 |
+
"basketball": "19",
|
| 397 |
+
"bat": "20",
|
| 398 |
+
"bathtub": "21",
|
| 399 |
+
"beach": "22",
|
| 400 |
+
"bear": "23",
|
| 401 |
+
"beard": "24",
|
| 402 |
+
"bed": "25",
|
| 403 |
+
"bee": "26",
|
| 404 |
+
"belt": "27",
|
| 405 |
+
"bench": "28",
|
| 406 |
+
"bicycle": "29",
|
| 407 |
+
"binoculars": "30",
|
| 408 |
+
"bird": "31",
|
| 409 |
+
"birthday cake": "32",
|
| 410 |
+
"blackberry": "33",
|
| 411 |
+
"blueberry": "34",
|
| 412 |
+
"book": "35",
|
| 413 |
+
"boomerang": "36",
|
| 414 |
+
"bottlecap": "37",
|
| 415 |
+
"bowtie": "38",
|
| 416 |
+
"bracelet": "39",
|
| 417 |
+
"brain": "40",
|
| 418 |
+
"bread": "41",
|
| 419 |
+
"bridge": "42",
|
| 420 |
+
"broccoli": "43",
|
| 421 |
+
"broom": "44",
|
| 422 |
+
"bucket": "45",
|
| 423 |
+
"bulldozer": "46",
|
| 424 |
+
"bus": "47",
|
| 425 |
+
"bush": "48",
|
| 426 |
+
"butterfly": "49",
|
| 427 |
+
"cactus": "50",
|
| 428 |
+
"cake": "51",
|
| 429 |
+
"calculator": "52",
|
| 430 |
+
"calendar": "53",
|
| 431 |
+
"camel": "54",
|
| 432 |
+
"camera": "55",
|
| 433 |
+
"camouflage": "56",
|
| 434 |
+
"campfire": "57",
|
| 435 |
+
"candle": "58",
|
| 436 |
+
"cannon": "59",
|
| 437 |
+
"canoe": "60",
|
| 438 |
+
"car": "61",
|
| 439 |
+
"carrot": "62",
|
| 440 |
+
"castle": "63",
|
| 441 |
+
"cat": "64",
|
| 442 |
+
"ceiling fan": "65",
|
| 443 |
+
"cell phone": "66",
|
| 444 |
+
"cello": "67",
|
| 445 |
+
"chair": "68",
|
| 446 |
+
"chandelier": "69",
|
| 447 |
+
"church": "70",
|
| 448 |
+
"circle": "71",
|
| 449 |
+
"clarinet": "72",
|
| 450 |
+
"clock": "73",
|
| 451 |
+
"cloud": "74",
|
| 452 |
+
"coffee cup": "75",
|
| 453 |
+
"compass": "76",
|
| 454 |
+
"computer": "77",
|
| 455 |
+
"cookie": "78",
|
| 456 |
+
"cooler": "79",
|
| 457 |
+
"couch": "80",
|
| 458 |
+
"cow": "81",
|
| 459 |
+
"crab": "82",
|
| 460 |
+
"crayon": "83",
|
| 461 |
+
"crocodile": "84",
|
| 462 |
+
"crown": "85",
|
| 463 |
+
"cruise ship": "86",
|
| 464 |
+
"cup": "87",
|
| 465 |
+
"diamond": "88",
|
| 466 |
+
"dishwasher": "89",
|
| 467 |
+
"diving board": "90",
|
| 468 |
+
"dog": "91",
|
| 469 |
+
"dolphin": "92",
|
| 470 |
+
"donut": "93",
|
| 471 |
+
"door": "94",
|
| 472 |
+
"dragon": "95",
|
| 473 |
+
"dresser": "96",
|
| 474 |
+
"drill": "97",
|
| 475 |
+
"drums": "98",
|
| 476 |
+
"duck": "99",
|
| 477 |
+
"dumbbell": "100",
|
| 478 |
+
"ear": "101",
|
| 479 |
+
"elbow": "102",
|
| 480 |
+
"elephant": "103",
|
| 481 |
+
"envelope": "104",
|
| 482 |
+
"eraser": "105",
|
| 483 |
+
"eye": "106",
|
| 484 |
+
"eyeglasses": "107",
|
| 485 |
+
"face": "108",
|
| 486 |
+
"fan": "109",
|
| 487 |
+
"feather": "110",
|
| 488 |
+
"fence": "111",
|
| 489 |
+
"finger": "112",
|
| 490 |
+
"fire hydrant": "113",
|
| 491 |
+
"fireplace": "114",
|
| 492 |
+
"firetruck": "115",
|
| 493 |
+
"fish": "116",
|
| 494 |
+
"flamingo": "117",
|
| 495 |
+
"flashlight": "118",
|
| 496 |
+
"flip flops": "119",
|
| 497 |
+
"floor lamp": "120",
|
| 498 |
+
"flower": "121",
|
| 499 |
+
"flying saucer": "122",
|
| 500 |
+
"foot": "123",
|
| 501 |
+
"fork": "124",
|
| 502 |
+
"frog": "125",
|
| 503 |
+
"frying pan": "126",
|
| 504 |
+
"garden": "128",
|
| 505 |
+
"garden hose": "127",
|
| 506 |
+
"giraffe": "129",
|
| 507 |
+
"goatee": "130",
|
| 508 |
+
"golf club": "131",
|
| 509 |
+
"grapes": "132",
|
| 510 |
+
"grass": "133",
|
| 511 |
+
"guitar": "134",
|
| 512 |
+
"hamburger": "135",
|
| 513 |
+
"hammer": "136",
|
| 514 |
+
"hand": "137",
|
| 515 |
+
"harp": "138",
|
| 516 |
+
"hat": "139",
|
| 517 |
+
"headphones": "140",
|
| 518 |
+
"hedgehog": "141",
|
| 519 |
+
"helicopter": "142",
|
| 520 |
+
"helmet": "143",
|
| 521 |
+
"hexagon": "144",
|
| 522 |
+
"hockey puck": "145",
|
| 523 |
+
"hockey stick": "146",
|
| 524 |
+
"horse": "147",
|
| 525 |
+
"hospital": "148",
|
| 526 |
+
"hot air balloon": "149",
|
| 527 |
+
"hot dog": "150",
|
| 528 |
+
"hot tub": "151",
|
| 529 |
+
"hourglass": "152",
|
| 530 |
+
"house": "154",
|
| 531 |
+
"house plant": "153",
|
| 532 |
+
"hurricane": "155",
|
| 533 |
+
"ice cream": "156",
|
| 534 |
+
"jacket": "157",
|
| 535 |
+
"jail": "158",
|
| 536 |
+
"kangaroo": "159",
|
| 537 |
+
"key": "160",
|
| 538 |
+
"keyboard": "161",
|
| 539 |
+
"knee": "162",
|
| 540 |
+
"knife": "163",
|
| 541 |
+
"ladder": "164",
|
| 542 |
+
"lantern": "165",
|
| 543 |
+
"laptop": "166",
|
| 544 |
+
"leaf": "167",
|
| 545 |
+
"leg": "168",
|
| 546 |
+
"light bulb": "169",
|
| 547 |
+
"lighter": "170",
|
| 548 |
+
"lighthouse": "171",
|
| 549 |
+
"lightning": "172",
|
| 550 |
+
"line": "173",
|
| 551 |
+
"lion": "174",
|
| 552 |
+
"lipstick": "175",
|
| 553 |
+
"lobster": "176",
|
| 554 |
+
"lollipop": "177",
|
| 555 |
+
"mailbox": "178",
|
| 556 |
+
"map": "179",
|
| 557 |
+
"marker": "180",
|
| 558 |
+
"matches": "181",
|
| 559 |
+
"megaphone": "182",
|
| 560 |
+
"mermaid": "183",
|
| 561 |
+
"microphone": "184",
|
| 562 |
+
"microwave": "185",
|
| 563 |
+
"monkey": "186",
|
| 564 |
+
"moon": "187",
|
| 565 |
+
"mosquito": "188",
|
| 566 |
+
"motorbike": "189",
|
| 567 |
+
"mountain": "190",
|
| 568 |
+
"mouse": "191",
|
| 569 |
+
"moustache": "192",
|
| 570 |
+
"mouth": "193",
|
| 571 |
+
"mug": "194",
|
| 572 |
+
"mushroom": "195",
|
| 573 |
+
"nail": "196",
|
| 574 |
+
"necklace": "197",
|
| 575 |
+
"nose": "198",
|
| 576 |
+
"ocean": "199",
|
| 577 |
+
"octagon": "200",
|
| 578 |
+
"octopus": "201",
|
| 579 |
+
"onion": "202",
|
| 580 |
+
"oven": "203",
|
| 581 |
+
"owl": "204",
|
| 582 |
+
"paint can": "205",
|
| 583 |
+
"paintbrush": "206",
|
| 584 |
+
"palm tree": "207",
|
| 585 |
+
"panda": "208",
|
| 586 |
+
"pants": "209",
|
| 587 |
+
"paper clip": "210",
|
| 588 |
+
"parachute": "211",
|
| 589 |
+
"parrot": "212",
|
| 590 |
+
"passport": "213",
|
| 591 |
+
"peanut": "214",
|
| 592 |
+
"pear": "215",
|
| 593 |
+
"peas": "216",
|
| 594 |
+
"pencil": "217",
|
| 595 |
+
"penguin": "218",
|
| 596 |
+
"piano": "219",
|
| 597 |
+
"pickup truck": "220",
|
| 598 |
+
"picture frame": "221",
|
| 599 |
+
"pig": "222",
|
| 600 |
+
"pillow": "223",
|
| 601 |
+
"pineapple": "224",
|
| 602 |
+
"pizza": "225",
|
| 603 |
+
"pliers": "226",
|
| 604 |
+
"police car": "227",
|
| 605 |
+
"pond": "228",
|
| 606 |
+
"pool": "229",
|
| 607 |
+
"popsicle": "230",
|
| 608 |
+
"postcard": "231",
|
| 609 |
+
"potato": "232",
|
| 610 |
+
"power outlet": "233",
|
| 611 |
+
"purse": "234",
|
| 612 |
+
"rabbit": "235",
|
| 613 |
+
"raccoon": "236",
|
| 614 |
+
"radio": "237",
|
| 615 |
+
"rain": "238",
|
| 616 |
+
"rainbow": "239",
|
| 617 |
+
"rake": "240",
|
| 618 |
+
"remote control": "241",
|
| 619 |
+
"rhinoceros": "242",
|
| 620 |
+
"rifle": "243",
|
| 621 |
+
"river": "244",
|
| 622 |
+
"roller coaster": "245",
|
| 623 |
+
"rollerskates": "246",
|
| 624 |
+
"sailboat": "247",
|
| 625 |
+
"sandwich": "248",
|
| 626 |
+
"saw": "249",
|
| 627 |
+
"saxophone": "250",
|
| 628 |
+
"school bus": "251",
|
| 629 |
+
"scissors": "252",
|
| 630 |
+
"scorpion": "253",
|
| 631 |
+
"screwdriver": "254",
|
| 632 |
+
"sea turtle": "255",
|
| 633 |
+
"see saw": "256",
|
| 634 |
+
"shark": "257",
|
| 635 |
+
"sheep": "258",
|
| 636 |
+
"shoe": "259",
|
| 637 |
+
"shorts": "260",
|
| 638 |
+
"shovel": "261",
|
| 639 |
+
"sink": "262",
|
| 640 |
+
"skateboard": "263",
|
| 641 |
+
"skull": "264",
|
| 642 |
+
"skyscraper": "265",
|
| 643 |
+
"sleeping bag": "266",
|
| 644 |
+
"smiley face": "267",
|
| 645 |
+
"snail": "268",
|
| 646 |
+
"snake": "269",
|
| 647 |
+
"snorkel": "270",
|
| 648 |
+
"snowflake": "271",
|
| 649 |
+
"snowman": "272",
|
| 650 |
+
"soccer ball": "273",
|
| 651 |
+
"sock": "274",
|
| 652 |
+
"speedboat": "275",
|
| 653 |
+
"spider": "276",
|
| 654 |
+
"spoon": "277",
|
| 655 |
+
"spreadsheet": "278",
|
| 656 |
+
"square": "279",
|
| 657 |
+
"squiggle": "280",
|
| 658 |
+
"squirrel": "281",
|
| 659 |
+
"stairs": "282",
|
| 660 |
+
"star": "283",
|
| 661 |
+
"steak": "284",
|
| 662 |
+
"stereo": "285",
|
| 663 |
+
"stethoscope": "286",
|
| 664 |
+
"stitches": "287",
|
| 665 |
+
"stop sign": "288",
|
| 666 |
+
"stove": "289",
|
| 667 |
+
"strawberry": "290",
|
| 668 |
+
"streetlight": "291",
|
| 669 |
+
"string bean": "292",
|
| 670 |
+
"submarine": "293",
|
| 671 |
+
"suitcase": "294",
|
| 672 |
+
"sun": "295",
|
| 673 |
+
"swan": "296",
|
| 674 |
+
"sweater": "297",
|
| 675 |
+
"swing set": "298",
|
| 676 |
+
"sword": "299",
|
| 677 |
+
"syringe": "300",
|
| 678 |
+
"t-shirt": "301",
|
| 679 |
+
"table": "302",
|
| 680 |
+
"teapot": "303",
|
| 681 |
+
"teddy-bear": "304",
|
| 682 |
+
"telephone": "305",
|
| 683 |
+
"television": "306",
|
| 684 |
+
"tennis racquet": "307",
|
| 685 |
+
"tent": "308",
|
| 686 |
+
"tiger": "312",
|
| 687 |
+
"toaster": "313",
|
| 688 |
+
"toe": "314",
|
| 689 |
+
"toilet": "315",
|
| 690 |
+
"tooth": "316",
|
| 691 |
+
"toothbrush": "317",
|
| 692 |
+
"toothpaste": "318",
|
| 693 |
+
"tornado": "319",
|
| 694 |
+
"tractor": "320",
|
| 695 |
+
"traffic light": "321",
|
| 696 |
+
"train": "322",
|
| 697 |
+
"tree": "323",
|
| 698 |
+
"triangle": "324",
|
| 699 |
+
"trombone": "325",
|
| 700 |
+
"truck": "326",
|
| 701 |
+
"trumpet": "327",
|
| 702 |
+
"umbrella": "328",
|
| 703 |
+
"underwear": "329",
|
| 704 |
+
"van": "330",
|
| 705 |
+
"vase": "331",
|
| 706 |
+
"violin": "332",
|
| 707 |
+
"washing machine": "333",
|
| 708 |
+
"watermelon": "334",
|
| 709 |
+
"waterslide": "335",
|
| 710 |
+
"whale": "336",
|
| 711 |
+
"wheel": "337",
|
| 712 |
+
"windmill": "338",
|
| 713 |
+
"wine bottle": "339",
|
| 714 |
+
"wine glass": "340",
|
| 715 |
+
"wristwatch": "341",
|
| 716 |
+
"yoga": "342",
|
| 717 |
+
"zebra": "343",
|
| 718 |
+
"zigzag": "344"
|
| 719 |
+
},
|
| 720 |
+
"layer_norm_eps": 1e-05,
|
| 721 |
+
"mlp_ratio": 2.0,
|
| 722 |
+
"model_type": "mobilevit",
|
| 723 |
+
"neck_hidden_sizes": [
|
| 724 |
+
16,
|
| 725 |
+
32,
|
| 726 |
+
48,
|
| 727 |
+
64,
|
| 728 |
+
80,
|
| 729 |
+
96,
|
| 730 |
+
384
|
| 731 |
+
],
|
| 732 |
+
"num_attention_heads": 4,
|
| 733 |
+
"num_channels": 1,
|
| 734 |
+
"output_stride": 32,
|
| 735 |
+
"patch_size": 2,
|
| 736 |
+
"problem_type": "single_label_classification",
|
| 737 |
+
"qkv_bias": true,
|
| 738 |
+
"semantic_loss_ignore_index": 255,
|
| 739 |
+
"torch_dtype": "float32",
|
| 740 |
+
"transformers_version": "4.42.4"
|
| 741 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:515a6ad114e024e5fd13dc70624fc54dc109be5014b295778bb5bde059f6e93f
|
| 3 |
+
size 8341964
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 28,
|
| 4 |
+
"width": 28
|
| 5 |
+
},
|
| 6 |
+
"do_center_crop": true,
|
| 7 |
+
"do_convert_rgb": false,
|
| 8 |
+
"do_flip_channel_order": false,
|
| 9 |
+
"do_rescale": true,
|
| 10 |
+
"do_resize": true,
|
| 11 |
+
"image_processor_type": "MobileViTImageProcessor",
|
| 12 |
+
"resample": 2,
|
| 13 |
+
"rescale_factor": 0.00392156862745098,
|
| 14 |
+
"size": {
|
| 15 |
+
"shortest_edge": 28
|
| 16 |
+
}
|
| 17 |
+
}
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.728948,
|
| 4 |
+
"eval_loss": 1.0799410343170166,
|
| 5 |
+
"eval_runtime": 79.2396,
|
| 6 |
+
"eval_samples_per_second": 3154.988,
|
| 7 |
+
"eval_steps_per_second": 12.33
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 1.3116020904e+17,
|
| 4 |
+
"train_loss": 1.269093582550002,
|
| 5 |
+
"train_runtime": 7252.1306,
|
| 6 |
+
"train_samples_per_second": 1861.522,
|
| 7 |
+
"train_steps_per_second": 7.272
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3821 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 5000,
|
| 6 |
+
"global_step": 52737,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.005688605722737357,
|
| 13 |
+
"grad_norm": 2.727705717086792,
|
| 14 |
+
"learning_rate": 0.0007984982080891974,
|
| 15 |
+
"loss": 2.3682,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.011377211445474714,
|
| 20 |
+
"grad_norm": 2.5978145599365234,
|
| 21 |
+
"learning_rate": 0.0007969812465631342,
|
| 22 |
+
"loss": 2.2098,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01706581716821207,
|
| 27 |
+
"grad_norm": 2.488832473754883,
|
| 28 |
+
"learning_rate": 0.0007954642850370708,
|
| 29 |
+
"loss": 2.1041,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02275442289094943,
|
| 34 |
+
"grad_norm": 2.1368465423583984,
|
| 35 |
+
"learning_rate": 0.0007939473235110074,
|
| 36 |
+
"loss": 2.0233,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.028443028613686784,
|
| 41 |
+
"grad_norm": 2.7280213832855225,
|
| 42 |
+
"learning_rate": 0.0007924303619849442,
|
| 43 |
+
"loss": 1.9639,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03413163433642414,
|
| 48 |
+
"grad_norm": 2.5257463455200195,
|
| 49 |
+
"learning_rate": 0.0007909134004588809,
|
| 50 |
+
"loss": 2.0229,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0398202400591615,
|
| 55 |
+
"grad_norm": 2.748051404953003,
|
| 56 |
+
"learning_rate": 0.0007893964389328175,
|
| 57 |
+
"loss": 1.9867,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.04550884578189886,
|
| 62 |
+
"grad_norm": 2.212047815322876,
|
| 63 |
+
"learning_rate": 0.0007878794774067543,
|
| 64 |
+
"loss": 1.9354,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.05119745150463621,
|
| 69 |
+
"grad_norm": 2.423400640487671,
|
| 70 |
+
"learning_rate": 0.000786362515880691,
|
| 71 |
+
"loss": 1.9127,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.05688605722737357,
|
| 76 |
+
"grad_norm": 2.379678726196289,
|
| 77 |
+
"learning_rate": 0.0007848455543546277,
|
| 78 |
+
"loss": 1.8927,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.06257466295011092,
|
| 83 |
+
"grad_norm": 2.5806541442871094,
|
| 84 |
+
"learning_rate": 0.0007833285928285645,
|
| 85 |
+
"loss": 1.8471,
|
| 86 |
+
"step": 1100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06826326867284828,
|
| 90 |
+
"grad_norm": 2.4539499282836914,
|
| 91 |
+
"learning_rate": 0.0007818116313025011,
|
| 92 |
+
"loss": 1.8296,
|
| 93 |
+
"step": 1200
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07395187439558565,
|
| 97 |
+
"grad_norm": 2.7818546295166016,
|
| 98 |
+
"learning_rate": 0.0007802946697764378,
|
| 99 |
+
"loss": 1.8136,
|
| 100 |
+
"step": 1300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.079640480118323,
|
| 104 |
+
"grad_norm": 2.979959487915039,
|
| 105 |
+
"learning_rate": 0.0007787777082503746,
|
| 106 |
+
"loss": 1.7844,
|
| 107 |
+
"step": 1400
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.08532908584106036,
|
| 111 |
+
"grad_norm": 2.3885879516601562,
|
| 112 |
+
"learning_rate": 0.0007772607467243113,
|
| 113 |
+
"loss": 1.7627,
|
| 114 |
+
"step": 1500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.09101769156379771,
|
| 118 |
+
"grad_norm": 2.2447025775909424,
|
| 119 |
+
"learning_rate": 0.0007757437851982479,
|
| 120 |
+
"loss": 1.7375,
|
| 121 |
+
"step": 1600
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.09670629728653507,
|
| 125 |
+
"grad_norm": 2.184580087661743,
|
| 126 |
+
"learning_rate": 0.0007742268236721846,
|
| 127 |
+
"loss": 1.7212,
|
| 128 |
+
"step": 1700
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.10239490300927243,
|
| 132 |
+
"grad_norm": 2.2506866455078125,
|
| 133 |
+
"learning_rate": 0.0007727098621461213,
|
| 134 |
+
"loss": 1.7264,
|
| 135 |
+
"step": 1800
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.10808350873200978,
|
| 139 |
+
"grad_norm": 2.059812068939209,
|
| 140 |
+
"learning_rate": 0.000771192900620058,
|
| 141 |
+
"loss": 1.7123,
|
| 142 |
+
"step": 1900
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.11377211445474714,
|
| 146 |
+
"grad_norm": 2.3013007640838623,
|
| 147 |
+
"learning_rate": 0.0007696759390939948,
|
| 148 |
+
"loss": 1.7122,
|
| 149 |
+
"step": 2000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.11946072017748449,
|
| 153 |
+
"grad_norm": 2.7073047161102295,
|
| 154 |
+
"learning_rate": 0.0007681741471831921,
|
| 155 |
+
"loss": 1.6847,
|
| 156 |
+
"step": 2100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.12514932590022185,
|
| 160 |
+
"grad_norm": 2.023949384689331,
|
| 161 |
+
"learning_rate": 0.0007666571856571288,
|
| 162 |
+
"loss": 1.6941,
|
| 163 |
+
"step": 2200
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1308379316229592,
|
| 167 |
+
"grad_norm": 1.9444501399993896,
|
| 168 |
+
"learning_rate": 0.0007651402241310656,
|
| 169 |
+
"loss": 1.6682,
|
| 170 |
+
"step": 2300
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.13652653734569656,
|
| 174 |
+
"grad_norm": 2.691826105117798,
|
| 175 |
+
"learning_rate": 0.0007636232626050023,
|
| 176 |
+
"loss": 1.6409,
|
| 177 |
+
"step": 2400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1422151430684339,
|
| 181 |
+
"grad_norm": 2.483386993408203,
|
| 182 |
+
"learning_rate": 0.0007621063010789389,
|
| 183 |
+
"loss": 1.662,
|
| 184 |
+
"step": 2500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.1479037487911713,
|
| 188 |
+
"grad_norm": 2.1850545406341553,
|
| 189 |
+
"learning_rate": 0.0007605893395528756,
|
| 190 |
+
"loss": 1.6532,
|
| 191 |
+
"step": 2600
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.15359235451390865,
|
| 195 |
+
"grad_norm": 1.989560842514038,
|
| 196 |
+
"learning_rate": 0.0007590723780268123,
|
| 197 |
+
"loss": 1.6231,
|
| 198 |
+
"step": 2700
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.159280960236646,
|
| 202 |
+
"grad_norm": 2.1362531185150146,
|
| 203 |
+
"learning_rate": 0.000757555416500749,
|
| 204 |
+
"loss": 1.6019,
|
| 205 |
+
"step": 2800
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.16496956595938336,
|
| 209 |
+
"grad_norm": 2.3262641429901123,
|
| 210 |
+
"learning_rate": 0.0007560384549746857,
|
| 211 |
+
"loss": 1.6103,
|
| 212 |
+
"step": 2900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.17065817168212072,
|
| 216 |
+
"grad_norm": 2.297419309616089,
|
| 217 |
+
"learning_rate": 0.0007545214934486224,
|
| 218 |
+
"loss": 1.6314,
|
| 219 |
+
"step": 3000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17634677740485807,
|
| 223 |
+
"grad_norm": 2.1368629932403564,
|
| 224 |
+
"learning_rate": 0.0007530045319225591,
|
| 225 |
+
"loss": 1.5838,
|
| 226 |
+
"step": 3100
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.18203538312759543,
|
| 230 |
+
"grad_norm": 2.3383195400238037,
|
| 231 |
+
"learning_rate": 0.0007514875703964959,
|
| 232 |
+
"loss": 1.5857,
|
| 233 |
+
"step": 3200
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.18772398885033278,
|
| 237 |
+
"grad_norm": 2.149740219116211,
|
| 238 |
+
"learning_rate": 0.0007499706088704326,
|
| 239 |
+
"loss": 1.6016,
|
| 240 |
+
"step": 3300
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.19341259457307014,
|
| 244 |
+
"grad_norm": 2.096703290939331,
|
| 245 |
+
"learning_rate": 0.0007484536473443692,
|
| 246 |
+
"loss": 1.5904,
|
| 247 |
+
"step": 3400
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.1991012002958075,
|
| 251 |
+
"grad_norm": 2.2043957710266113,
|
| 252 |
+
"learning_rate": 0.000746936685818306,
|
| 253 |
+
"loss": 1.5787,
|
| 254 |
+
"step": 3500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.20478980601854485,
|
| 258 |
+
"grad_norm": 2.6369898319244385,
|
| 259 |
+
"learning_rate": 0.0007454197242922427,
|
| 260 |
+
"loss": 1.5539,
|
| 261 |
+
"step": 3600
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2104784117412822,
|
| 265 |
+
"grad_norm": 1.9776628017425537,
|
| 266 |
+
"learning_rate": 0.0007439027627661794,
|
| 267 |
+
"loss": 1.5815,
|
| 268 |
+
"step": 3700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.21616701746401956,
|
| 272 |
+
"grad_norm": 2.2001795768737793,
|
| 273 |
+
"learning_rate": 0.0007423858012401161,
|
| 274 |
+
"loss": 1.5583,
|
| 275 |
+
"step": 3800
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.22185562318675692,
|
| 279 |
+
"grad_norm": 2.2252562046051025,
|
| 280 |
+
"learning_rate": 0.0007408688397140527,
|
| 281 |
+
"loss": 1.5539,
|
| 282 |
+
"step": 3900
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.22754422890949427,
|
| 286 |
+
"grad_norm": 2.15871262550354,
|
| 287 |
+
"learning_rate": 0.0007393518781879895,
|
| 288 |
+
"loss": 1.5786,
|
| 289 |
+
"step": 4000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.23323283463223163,
|
| 293 |
+
"grad_norm": 2.026066303253174,
|
| 294 |
+
"learning_rate": 0.0007378500862771869,
|
| 295 |
+
"loss": 1.5452,
|
| 296 |
+
"step": 4100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.23892144035496898,
|
| 300 |
+
"grad_norm": 2.116511583328247,
|
| 301 |
+
"learning_rate": 0.0007363331247511235,
|
| 302 |
+
"loss": 1.5381,
|
| 303 |
+
"step": 4200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24461004607770637,
|
| 307 |
+
"grad_norm": 1.9454152584075928,
|
| 308 |
+
"learning_rate": 0.0007348161632250602,
|
| 309 |
+
"loss": 1.557,
|
| 310 |
+
"step": 4300
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2502986518004437,
|
| 314 |
+
"grad_norm": 1.8668495416641235,
|
| 315 |
+
"learning_rate": 0.000733299201698997,
|
| 316 |
+
"loss": 1.5406,
|
| 317 |
+
"step": 4400
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.2559872575231811,
|
| 321 |
+
"grad_norm": 2.0886125564575195,
|
| 322 |
+
"learning_rate": 0.0007317822401729337,
|
| 323 |
+
"loss": 1.519,
|
| 324 |
+
"step": 4500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2616758632459184,
|
| 328 |
+
"grad_norm": 2.5768446922302246,
|
| 329 |
+
"learning_rate": 0.0007302652786468704,
|
| 330 |
+
"loss": 1.5178,
|
| 331 |
+
"step": 4600
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.2673644689686558,
|
| 335 |
+
"grad_norm": 2.4169631004333496,
|
| 336 |
+
"learning_rate": 0.000728748317120807,
|
| 337 |
+
"loss": 1.5283,
|
| 338 |
+
"step": 4700
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.2730530746913931,
|
| 342 |
+
"grad_norm": 2.7676291465759277,
|
| 343 |
+
"learning_rate": 0.0007272313555947437,
|
| 344 |
+
"loss": 1.5265,
|
| 345 |
+
"step": 4800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.2787416804141305,
|
| 349 |
+
"grad_norm": 1.9152452945709229,
|
| 350 |
+
"learning_rate": 0.0007257143940686805,
|
| 351 |
+
"loss": 1.5152,
|
| 352 |
+
"step": 4900
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.2844302861368678,
|
| 356 |
+
"grad_norm": 2.56608510017395,
|
| 357 |
+
"learning_rate": 0.0007241974325426172,
|
| 358 |
+
"loss": 1.5084,
|
| 359 |
+
"step": 5000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.2844302861368678,
|
| 363 |
+
"eval_accuracy": 0.636444,
|
| 364 |
+
"eval_loss": 1.469992995262146,
|
| 365 |
+
"eval_runtime": 85.5907,
|
| 366 |
+
"eval_samples_per_second": 2920.879,
|
| 367 |
+
"eval_steps_per_second": 11.415,
|
| 368 |
+
"step": 5000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.2901188918596052,
|
| 372 |
+
"grad_norm": 2.2403135299682617,
|
| 373 |
+
"learning_rate": 0.0007226804710165538,
|
| 374 |
+
"loss": 1.5076,
|
| 375 |
+
"step": 5100
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.2958074975823426,
|
| 379 |
+
"grad_norm": 2.058535099029541,
|
| 380 |
+
"learning_rate": 0.0007211635094904906,
|
| 381 |
+
"loss": 1.4973,
|
| 382 |
+
"step": 5200
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.3014961033050799,
|
| 386 |
+
"grad_norm": 1.9374159574508667,
|
| 387 |
+
"learning_rate": 0.0007196465479644273,
|
| 388 |
+
"loss": 1.5055,
|
| 389 |
+
"step": 5300
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.3071847090278173,
|
| 393 |
+
"grad_norm": 1.8894695043563843,
|
| 394 |
+
"learning_rate": 0.000718129586438364,
|
| 395 |
+
"loss": 1.4996,
|
| 396 |
+
"step": 5400
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.31287331475055463,
|
| 400 |
+
"grad_norm": 2.5466501712799072,
|
| 401 |
+
"learning_rate": 0.0007166126249123008,
|
| 402 |
+
"loss": 1.5006,
|
| 403 |
+
"step": 5500
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.318561920473292,
|
| 407 |
+
"grad_norm": 1.9721605777740479,
|
| 408 |
+
"learning_rate": 0.0007150956633862374,
|
| 409 |
+
"loss": 1.4932,
|
| 410 |
+
"step": 5600
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.32425052619602934,
|
| 414 |
+
"grad_norm": 1.8921763896942139,
|
| 415 |
+
"learning_rate": 0.0007135787018601741,
|
| 416 |
+
"loss": 1.4762,
|
| 417 |
+
"step": 5700
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.3299391319187667,
|
| 421 |
+
"grad_norm": 2.49052357673645,
|
| 422 |
+
"learning_rate": 0.0007120617403341108,
|
| 423 |
+
"loss": 1.4636,
|
| 424 |
+
"step": 5800
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.33562773764150405,
|
| 428 |
+
"grad_norm": 1.8825891017913818,
|
| 429 |
+
"learning_rate": 0.0007105447788080475,
|
| 430 |
+
"loss": 1.4903,
|
| 431 |
+
"step": 5900
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.34131634336424144,
|
| 435 |
+
"grad_norm": 1.9227776527404785,
|
| 436 |
+
"learning_rate": 0.0007090278172819842,
|
| 437 |
+
"loss": 1.4602,
|
| 438 |
+
"step": 6000
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 0.34700494908697876,
|
| 442 |
+
"grad_norm": 2.173774242401123,
|
| 443 |
+
"learning_rate": 0.0007075108557559209,
|
| 444 |
+
"loss": 1.4581,
|
| 445 |
+
"step": 6100
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.35269355480971615,
|
| 449 |
+
"grad_norm": 1.9840656518936157,
|
| 450 |
+
"learning_rate": 0.0007059938942298576,
|
| 451 |
+
"loss": 1.4504,
|
| 452 |
+
"step": 6200
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.3583821605324535,
|
| 456 |
+
"grad_norm": 2.368171453475952,
|
| 457 |
+
"learning_rate": 0.000704492102319055,
|
| 458 |
+
"loss": 1.4659,
|
| 459 |
+
"step": 6300
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.36407076625519086,
|
| 463 |
+
"grad_norm": 2.005125045776367,
|
| 464 |
+
"learning_rate": 0.0007029751407929917,
|
| 465 |
+
"loss": 1.4698,
|
| 466 |
+
"step": 6400
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.3697593719779282,
|
| 470 |
+
"grad_norm": 1.8724095821380615,
|
| 471 |
+
"learning_rate": 0.0007014581792669284,
|
| 472 |
+
"loss": 1.4429,
|
| 473 |
+
"step": 6500
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.37544797770066557,
|
| 477 |
+
"grad_norm": 1.8412431478500366,
|
| 478 |
+
"learning_rate": 0.0006999412177408651,
|
| 479 |
+
"loss": 1.4368,
|
| 480 |
+
"step": 6600
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.3811365834234029,
|
| 484 |
+
"grad_norm": 1.9016755819320679,
|
| 485 |
+
"learning_rate": 0.0006984242562148018,
|
| 486 |
+
"loss": 1.4351,
|
| 487 |
+
"step": 6700
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.3868251891461403,
|
| 491 |
+
"grad_norm": 1.9896953105926514,
|
| 492 |
+
"learning_rate": 0.0006969072946887384,
|
| 493 |
+
"loss": 1.4563,
|
| 494 |
+
"step": 6800
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.39251379486887766,
|
| 498 |
+
"grad_norm": 2.3341548442840576,
|
| 499 |
+
"learning_rate": 0.0006953903331626751,
|
| 500 |
+
"loss": 1.4457,
|
| 501 |
+
"step": 6900
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.398202400591615,
|
| 505 |
+
"grad_norm": 1.95259690284729,
|
| 506 |
+
"learning_rate": 0.0006938733716366119,
|
| 507 |
+
"loss": 1.4636,
|
| 508 |
+
"step": 7000
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.4038910063143524,
|
| 512 |
+
"grad_norm": 1.8444461822509766,
|
| 513 |
+
"learning_rate": 0.0006923564101105486,
|
| 514 |
+
"loss": 1.4418,
|
| 515 |
+
"step": 7100
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.4095796120370897,
|
| 519 |
+
"grad_norm": 1.9170624017715454,
|
| 520 |
+
"learning_rate": 0.0006908394485844853,
|
| 521 |
+
"loss": 1.4267,
|
| 522 |
+
"step": 7200
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.4152682177598271,
|
| 526 |
+
"grad_norm": 1.6293827295303345,
|
| 527 |
+
"learning_rate": 0.000689322487058422,
|
| 528 |
+
"loss": 1.4474,
|
| 529 |
+
"step": 7300
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.4209568234825644,
|
| 533 |
+
"grad_norm": 2.2202467918395996,
|
| 534 |
+
"learning_rate": 0.0006878055255323587,
|
| 535 |
+
"loss": 1.4166,
|
| 536 |
+
"step": 7400
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.4266454292053018,
|
| 540 |
+
"grad_norm": 1.9069397449493408,
|
| 541 |
+
"learning_rate": 0.0006862885640062954,
|
| 542 |
+
"loss": 1.4194,
|
| 543 |
+
"step": 7500
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.4323340349280391,
|
| 547 |
+
"grad_norm": 2.0205297470092773,
|
| 548 |
+
"learning_rate": 0.0006847716024802322,
|
| 549 |
+
"loss": 1.4328,
|
| 550 |
+
"step": 7600
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.4380226406507765,
|
| 554 |
+
"grad_norm": 1.6736252307891846,
|
| 555 |
+
"learning_rate": 0.0006832546409541689,
|
| 556 |
+
"loss": 1.4309,
|
| 557 |
+
"step": 7700
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.44371124637351383,
|
| 561 |
+
"grad_norm": 1.7010937929153442,
|
| 562 |
+
"learning_rate": 0.0006817376794281055,
|
| 563 |
+
"loss": 1.412,
|
| 564 |
+
"step": 7800
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.4493998520962512,
|
| 568 |
+
"grad_norm": 2.748424768447876,
|
| 569 |
+
"learning_rate": 0.0006802207179020422,
|
| 570 |
+
"loss": 1.421,
|
| 571 |
+
"step": 7900
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.45508845781898855,
|
| 575 |
+
"grad_norm": 1.908728837966919,
|
| 576 |
+
"learning_rate": 0.0006787037563759789,
|
| 577 |
+
"loss": 1.4172,
|
| 578 |
+
"step": 8000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.46077706354172593,
|
| 582 |
+
"grad_norm": 1.8672014474868774,
|
| 583 |
+
"learning_rate": 0.0006771867948499157,
|
| 584 |
+
"loss": 1.4448,
|
| 585 |
+
"step": 8100
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.46646566926446326,
|
| 589 |
+
"grad_norm": 2.128519058227539,
|
| 590 |
+
"learning_rate": 0.0006756698333238524,
|
| 591 |
+
"loss": 1.4158,
|
| 592 |
+
"step": 8200
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.47215427498720064,
|
| 596 |
+
"grad_norm": 1.7498713731765747,
|
| 597 |
+
"learning_rate": 0.000674152871797789,
|
| 598 |
+
"loss": 1.412,
|
| 599 |
+
"step": 8300
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.47784288070993797,
|
| 603 |
+
"grad_norm": 1.7801289558410645,
|
| 604 |
+
"learning_rate": 0.0006726510798869864,
|
| 605 |
+
"loss": 1.4146,
|
| 606 |
+
"step": 8400
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.48353148643267535,
|
| 610 |
+
"grad_norm": 1.9360538721084595,
|
| 611 |
+
"learning_rate": 0.0006711341183609232,
|
| 612 |
+
"loss": 1.4252,
|
| 613 |
+
"step": 8500
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.48922009215541273,
|
| 617 |
+
"grad_norm": 2.3669304847717285,
|
| 618 |
+
"learning_rate": 0.0006696171568348598,
|
| 619 |
+
"loss": 1.4057,
|
| 620 |
+
"step": 8600
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.49490869787815006,
|
| 624 |
+
"grad_norm": 1.7751379013061523,
|
| 625 |
+
"learning_rate": 0.0006681001953087965,
|
| 626 |
+
"loss": 1.4049,
|
| 627 |
+
"step": 8700
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.5005973036008874,
|
| 631 |
+
"grad_norm": 2.5389885902404785,
|
| 632 |
+
"learning_rate": 0.0006665832337827332,
|
| 633 |
+
"loss": 1.3837,
|
| 634 |
+
"step": 8800
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.5062859093236248,
|
| 638 |
+
"grad_norm": 2.5082690715789795,
|
| 639 |
+
"learning_rate": 0.0006650662722566699,
|
| 640 |
+
"loss": 1.3924,
|
| 641 |
+
"step": 8900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.5119745150463622,
|
| 645 |
+
"grad_norm": 2.011589527130127,
|
| 646 |
+
"learning_rate": 0.0006635493107306066,
|
| 647 |
+
"loss": 1.3956,
|
| 648 |
+
"step": 9000
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.5176631207690995,
|
| 652 |
+
"grad_norm": 1.819793939590454,
|
| 653 |
+
"learning_rate": 0.0006620323492045433,
|
| 654 |
+
"loss": 1.4063,
|
| 655 |
+
"step": 9100
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.5233517264918368,
|
| 659 |
+
"grad_norm": 2.081247568130493,
|
| 660 |
+
"learning_rate": 0.00066051538767848,
|
| 661 |
+
"loss": 1.4145,
|
| 662 |
+
"step": 9200
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.5290403322145742,
|
| 666 |
+
"grad_norm": 2.151563882827759,
|
| 667 |
+
"learning_rate": 0.0006589984261524168,
|
| 668 |
+
"loss": 1.4002,
|
| 669 |
+
"step": 9300
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.5347289379373116,
|
| 673 |
+
"grad_norm": 1.9170759916305542,
|
| 674 |
+
"learning_rate": 0.0006574814646263535,
|
| 675 |
+
"loss": 1.3863,
|
| 676 |
+
"step": 9400
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.5404175436600489,
|
| 680 |
+
"grad_norm": 1.6435186862945557,
|
| 681 |
+
"learning_rate": 0.0006559645031002901,
|
| 682 |
+
"loss": 1.3888,
|
| 683 |
+
"step": 9500
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.5461061493827862,
|
| 687 |
+
"grad_norm": 1.8130972385406494,
|
| 688 |
+
"learning_rate": 0.0006544475415742268,
|
| 689 |
+
"loss": 1.3904,
|
| 690 |
+
"step": 9600
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.5517947551055237,
|
| 694 |
+
"grad_norm": 1.8200345039367676,
|
| 695 |
+
"learning_rate": 0.0006529305800481636,
|
| 696 |
+
"loss": 1.3647,
|
| 697 |
+
"step": 9700
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.557483360828261,
|
| 701 |
+
"grad_norm": 1.7286423444747925,
|
| 702 |
+
"learning_rate": 0.0006514136185221003,
|
| 703 |
+
"loss": 1.3815,
|
| 704 |
+
"step": 9800
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.5631719665509983,
|
| 708 |
+
"grad_norm": 2.345879554748535,
|
| 709 |
+
"learning_rate": 0.000649896656996037,
|
| 710 |
+
"loss": 1.3919,
|
| 711 |
+
"step": 9900
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.5688605722737357,
|
| 715 |
+
"grad_norm": 1.8189209699630737,
|
| 716 |
+
"learning_rate": 0.0006483796954699736,
|
| 717 |
+
"loss": 1.3684,
|
| 718 |
+
"step": 10000
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.5688605722737357,
|
| 722 |
+
"eval_accuracy": 0.667392,
|
| 723 |
+
"eval_loss": 1.3353288173675537,
|
| 724 |
+
"eval_runtime": 85.0475,
|
| 725 |
+
"eval_samples_per_second": 2939.533,
|
| 726 |
+
"eval_steps_per_second": 11.488,
|
| 727 |
+
"step": 10000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.5745491779964731,
|
| 731 |
+
"grad_norm": 1.7264429330825806,
|
| 732 |
+
"learning_rate": 0.0006468627339439103,
|
| 733 |
+
"loss": 1.4044,
|
| 734 |
+
"step": 10100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.5802377837192104,
|
| 738 |
+
"grad_norm": 1.8806540966033936,
|
| 739 |
+
"learning_rate": 0.0006453457724178471,
|
| 740 |
+
"loss": 1.3746,
|
| 741 |
+
"step": 10200
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.5859263894419477,
|
| 745 |
+
"grad_norm": 1.7714815139770508,
|
| 746 |
+
"learning_rate": 0.0006438288108917838,
|
| 747 |
+
"loss": 1.3837,
|
| 748 |
+
"step": 10300
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.5916149951646852,
|
| 752 |
+
"grad_norm": 1.713157057762146,
|
| 753 |
+
"learning_rate": 0.0006423118493657205,
|
| 754 |
+
"loss": 1.3939,
|
| 755 |
+
"step": 10400
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 0.5973036008874225,
|
| 759 |
+
"grad_norm": 2.169168472290039,
|
| 760 |
+
"learning_rate": 0.0006408100574549179,
|
| 761 |
+
"loss": 1.3658,
|
| 762 |
+
"step": 10500
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.6029922066101598,
|
| 766 |
+
"grad_norm": 1.727501630783081,
|
| 767 |
+
"learning_rate": 0.0006392930959288546,
|
| 768 |
+
"loss": 1.3907,
|
| 769 |
+
"step": 10600
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.6086808123328972,
|
| 773 |
+
"grad_norm": 2.0120322704315186,
|
| 774 |
+
"learning_rate": 0.0006377761344027913,
|
| 775 |
+
"loss": 1.3757,
|
| 776 |
+
"step": 10700
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.6143694180556346,
|
| 780 |
+
"grad_norm": 1.799139142036438,
|
| 781 |
+
"learning_rate": 0.0006362591728767279,
|
| 782 |
+
"loss": 1.3803,
|
| 783 |
+
"step": 10800
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.6200580237783719,
|
| 787 |
+
"grad_norm": 1.8817808628082275,
|
| 788 |
+
"learning_rate": 0.0006347422113506646,
|
| 789 |
+
"loss": 1.3702,
|
| 790 |
+
"step": 10900
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.6257466295011093,
|
| 794 |
+
"grad_norm": 2.1144518852233887,
|
| 795 |
+
"learning_rate": 0.0006332252498246013,
|
| 796 |
+
"loss": 1.3832,
|
| 797 |
+
"step": 11000
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.6314352352238466,
|
| 801 |
+
"grad_norm": 2.1396071910858154,
|
| 802 |
+
"learning_rate": 0.0006317082882985381,
|
| 803 |
+
"loss": 1.3611,
|
| 804 |
+
"step": 11100
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.637123840946584,
|
| 808 |
+
"grad_norm": 1.6794757843017578,
|
| 809 |
+
"learning_rate": 0.0006301913267724747,
|
| 810 |
+
"loss": 1.368,
|
| 811 |
+
"step": 11200
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.6428124466693214,
|
| 815 |
+
"grad_norm": 2.268433094024658,
|
| 816 |
+
"learning_rate": 0.0006286743652464114,
|
| 817 |
+
"loss": 1.3498,
|
| 818 |
+
"step": 11300
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.6485010523920587,
|
| 822 |
+
"grad_norm": 1.8515706062316895,
|
| 823 |
+
"learning_rate": 0.0006271574037203482,
|
| 824 |
+
"loss": 1.3489,
|
| 825 |
+
"step": 11400
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.654189658114796,
|
| 829 |
+
"grad_norm": 2.482171058654785,
|
| 830 |
+
"learning_rate": 0.0006256404421942849,
|
| 831 |
+
"loss": 1.3501,
|
| 832 |
+
"step": 11500
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.6598782638375335,
|
| 836 |
+
"grad_norm": 1.9485667943954468,
|
| 837 |
+
"learning_rate": 0.0006241234806682216,
|
| 838 |
+
"loss": 1.3483,
|
| 839 |
+
"step": 11600
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 0.6655668695602708,
|
| 843 |
+
"grad_norm": 1.8601367473602295,
|
| 844 |
+
"learning_rate": 0.0006226065191421583,
|
| 845 |
+
"loss": 1.3392,
|
| 846 |
+
"step": 11700
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 0.6712554752830081,
|
| 850 |
+
"grad_norm": 1.870851993560791,
|
| 851 |
+
"learning_rate": 0.000621089557616095,
|
| 852 |
+
"loss": 1.352,
|
| 853 |
+
"step": 11800
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.6769440810057455,
|
| 857 |
+
"grad_norm": 1.9454014301300049,
|
| 858 |
+
"learning_rate": 0.0006195725960900317,
|
| 859 |
+
"loss": 1.3537,
|
| 860 |
+
"step": 11900
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.6826326867284829,
|
| 864 |
+
"grad_norm": 1.9180669784545898,
|
| 865 |
+
"learning_rate": 0.0006180556345639685,
|
| 866 |
+
"loss": 1.3541,
|
| 867 |
+
"step": 12000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.6883212924512202,
|
| 871 |
+
"grad_norm": 1.7796809673309326,
|
| 872 |
+
"learning_rate": 0.0006165386730379051,
|
| 873 |
+
"loss": 1.331,
|
| 874 |
+
"step": 12100
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.6940098981739575,
|
| 878 |
+
"grad_norm": 2.040998935699463,
|
| 879 |
+
"learning_rate": 0.0006150217115118417,
|
| 880 |
+
"loss": 1.3214,
|
| 881 |
+
"step": 12200
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 0.699698503896695,
|
| 885 |
+
"grad_norm": 1.7188791036605835,
|
| 886 |
+
"learning_rate": 0.0006135047499857785,
|
| 887 |
+
"loss": 1.3577,
|
| 888 |
+
"step": 12300
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.7053871096194323,
|
| 892 |
+
"grad_norm": 1.9152625799179077,
|
| 893 |
+
"learning_rate": 0.0006119877884597152,
|
| 894 |
+
"loss": 1.3682,
|
| 895 |
+
"step": 12400
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.7110757153421696,
|
| 899 |
+
"grad_norm": 2.150810718536377,
|
| 900 |
+
"learning_rate": 0.0006104708269336519,
|
| 901 |
+
"loss": 1.3388,
|
| 902 |
+
"step": 12500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.716764321064907,
|
| 906 |
+
"grad_norm": 1.97470223903656,
|
| 907 |
+
"learning_rate": 0.0006089538654075887,
|
| 908 |
+
"loss": 1.3319,
|
| 909 |
+
"step": 12600
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.7224529267876444,
|
| 913 |
+
"grad_norm": 1.663122296333313,
|
| 914 |
+
"learning_rate": 0.0006074369038815253,
|
| 915 |
+
"loss": 1.3593,
|
| 916 |
+
"step": 12700
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.7281415325103817,
|
| 920 |
+
"grad_norm": 1.6453677415847778,
|
| 921 |
+
"learning_rate": 0.0006059351119707227,
|
| 922 |
+
"loss": 1.3592,
|
| 923 |
+
"step": 12800
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.733830138233119,
|
| 927 |
+
"grad_norm": 1.6896419525146484,
|
| 928 |
+
"learning_rate": 0.0006044181504446595,
|
| 929 |
+
"loss": 1.3183,
|
| 930 |
+
"step": 12900
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 0.7395187439558564,
|
| 934 |
+
"grad_norm": 1.7903008460998535,
|
| 935 |
+
"learning_rate": 0.000602901188918596,
|
| 936 |
+
"loss": 1.3373,
|
| 937 |
+
"step": 13000
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.7452073496785938,
|
| 941 |
+
"grad_norm": 2.2026655673980713,
|
| 942 |
+
"learning_rate": 0.0006013842273925327,
|
| 943 |
+
"loss": 1.3403,
|
| 944 |
+
"step": 13100
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.7508959554013311,
|
| 948 |
+
"grad_norm": 1.9204201698303223,
|
| 949 |
+
"learning_rate": 0.0005998672658664695,
|
| 950 |
+
"loss": 1.3199,
|
| 951 |
+
"step": 13200
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.7565845611240685,
|
| 955 |
+
"grad_norm": 1.946899652481079,
|
| 956 |
+
"learning_rate": 0.0005983503043404062,
|
| 957 |
+
"loss": 1.3298,
|
| 958 |
+
"step": 13300
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.7622731668468058,
|
| 962 |
+
"grad_norm": 2.019131898880005,
|
| 963 |
+
"learning_rate": 0.0005968333428143428,
|
| 964 |
+
"loss": 1.3449,
|
| 965 |
+
"step": 13400
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 0.7679617725695432,
|
| 969 |
+
"grad_norm": 1.848008155822754,
|
| 970 |
+
"learning_rate": 0.0005953163812882796,
|
| 971 |
+
"loss": 1.3206,
|
| 972 |
+
"step": 13500
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 0.7736503782922806,
|
| 976 |
+
"grad_norm": 2.373288631439209,
|
| 977 |
+
"learning_rate": 0.0005937994197622163,
|
| 978 |
+
"loss": 1.3564,
|
| 979 |
+
"step": 13600
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.7793389840150179,
|
| 983 |
+
"grad_norm": 2.556985855102539,
|
| 984 |
+
"learning_rate": 0.000592282458236153,
|
| 985 |
+
"loss": 1.3254,
|
| 986 |
+
"step": 13700
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.7850275897377553,
|
| 990 |
+
"grad_norm": 1.8957433700561523,
|
| 991 |
+
"learning_rate": 0.0005907654967100898,
|
| 992 |
+
"loss": 1.3498,
|
| 993 |
+
"step": 13800
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.7907161954604927,
|
| 997 |
+
"grad_norm": 1.7315127849578857,
|
| 998 |
+
"learning_rate": 0.0005892485351840264,
|
| 999 |
+
"loss": 1.3249,
|
| 1000 |
+
"step": 13900
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 0.79640480118323,
|
| 1004 |
+
"grad_norm": 1.973764419555664,
|
| 1005 |
+
"learning_rate": 0.0005877315736579631,
|
| 1006 |
+
"loss": 1.3305,
|
| 1007 |
+
"step": 14000
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 0.8020934069059673,
|
| 1011 |
+
"grad_norm": 1.711145281791687,
|
| 1012 |
+
"learning_rate": 0.0005862146121318999,
|
| 1013 |
+
"loss": 1.3011,
|
| 1014 |
+
"step": 14100
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 0.8077820126287047,
|
| 1018 |
+
"grad_norm": 1.8515042066574097,
|
| 1019 |
+
"learning_rate": 0.0005846976506058365,
|
| 1020 |
+
"loss": 1.3195,
|
| 1021 |
+
"step": 14200
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.8134706183514421,
|
| 1025 |
+
"grad_norm": 1.6278700828552246,
|
| 1026 |
+
"learning_rate": 0.0005831806890797733,
|
| 1027 |
+
"loss": 1.3308,
|
| 1028 |
+
"step": 14300
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.8191592240741794,
|
| 1032 |
+
"grad_norm": 1.444455623626709,
|
| 1033 |
+
"learning_rate": 0.0005816637275537099,
|
| 1034 |
+
"loss": 1.3119,
|
| 1035 |
+
"step": 14400
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.8248478297969167,
|
| 1039 |
+
"grad_norm": 1.6277796030044556,
|
| 1040 |
+
"learning_rate": 0.0005801467660276466,
|
| 1041 |
+
"loss": 1.3227,
|
| 1042 |
+
"step": 14500
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.8305364355196542,
|
| 1046 |
+
"grad_norm": 1.8428665399551392,
|
| 1047 |
+
"learning_rate": 0.0005786298045015834,
|
| 1048 |
+
"loss": 1.3336,
|
| 1049 |
+
"step": 14600
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 0.8362250412423915,
|
| 1053 |
+
"grad_norm": 1.6377763748168945,
|
| 1054 |
+
"learning_rate": 0.0005771128429755201,
|
| 1055 |
+
"loss": 1.3141,
|
| 1056 |
+
"step": 14700
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 0.8419136469651288,
|
| 1060 |
+
"grad_norm": 1.7305645942687988,
|
| 1061 |
+
"learning_rate": 0.0005755958814494568,
|
| 1062 |
+
"loss": 1.3062,
|
| 1063 |
+
"step": 14800
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 0.8476022526878662,
|
| 1067 |
+
"grad_norm": 2.469701051712036,
|
| 1068 |
+
"learning_rate": 0.0005740940895386541,
|
| 1069 |
+
"loss": 1.3074,
|
| 1070 |
+
"step": 14900
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.8532908584106036,
|
| 1074 |
+
"grad_norm": 1.952755331993103,
|
| 1075 |
+
"learning_rate": 0.0005725771280125909,
|
| 1076 |
+
"loss": 1.3568,
|
| 1077 |
+
"step": 15000
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.8532908584106036,
|
| 1081 |
+
"eval_accuracy": 0.68038,
|
| 1082 |
+
"eval_loss": 1.2764052152633667,
|
| 1083 |
+
"eval_runtime": 82.5452,
|
| 1084 |
+
"eval_samples_per_second": 3028.643,
|
| 1085 |
+
"eval_steps_per_second": 11.836,
|
| 1086 |
+
"step": 15000
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 0.8589794641333409,
|
| 1090 |
+
"grad_norm": 3.221471071243286,
|
| 1091 |
+
"learning_rate": 0.0005710601664865274,
|
| 1092 |
+
"loss": 1.3341,
|
| 1093 |
+
"step": 15100
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 0.8646680698560782,
|
| 1097 |
+
"grad_norm": 2.2455317974090576,
|
| 1098 |
+
"learning_rate": 0.0005695432049604642,
|
| 1099 |
+
"loss": 1.3276,
|
| 1100 |
+
"step": 15200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.8703566755788157,
|
| 1104 |
+
"grad_norm": 1.8076684474945068,
|
| 1105 |
+
"learning_rate": 0.0005680262434344009,
|
| 1106 |
+
"loss": 1.2922,
|
| 1107 |
+
"step": 15300
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.876045281301553,
|
| 1111 |
+
"grad_norm": 1.701774001121521,
|
| 1112 |
+
"learning_rate": 0.0005665092819083376,
|
| 1113 |
+
"loss": 1.3003,
|
| 1114 |
+
"step": 15400
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.8817338870242903,
|
| 1118 |
+
"grad_norm": 1.5403673648834229,
|
| 1119 |
+
"learning_rate": 0.0005649923203822744,
|
| 1120 |
+
"loss": 1.3207,
|
| 1121 |
+
"step": 15500
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 0.8874224927470277,
|
| 1125 |
+
"grad_norm": 1.9462639093399048,
|
| 1126 |
+
"learning_rate": 0.000563475358856211,
|
| 1127 |
+
"loss": 1.3098,
|
| 1128 |
+
"step": 15600
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 0.8931110984697651,
|
| 1132 |
+
"grad_norm": 1.6688456535339355,
|
| 1133 |
+
"learning_rate": 0.0005619583973301477,
|
| 1134 |
+
"loss": 1.2993,
|
| 1135 |
+
"step": 15700
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 0.8987997041925024,
|
| 1139 |
+
"grad_norm": 1.6060837507247925,
|
| 1140 |
+
"learning_rate": 0.0005604414358040845,
|
| 1141 |
+
"loss": 1.3145,
|
| 1142 |
+
"step": 15800
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.9044883099152398,
|
| 1146 |
+
"grad_norm": 1.8593111038208008,
|
| 1147 |
+
"learning_rate": 0.0005589244742780212,
|
| 1148 |
+
"loss": 1.2836,
|
| 1149 |
+
"step": 15900
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.9101769156379771,
|
| 1153 |
+
"grad_norm": 2.035261869430542,
|
| 1154 |
+
"learning_rate": 0.0005574075127519579,
|
| 1155 |
+
"loss": 1.3125,
|
| 1156 |
+
"step": 16000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.9158655213607145,
|
| 1160 |
+
"grad_norm": 1.6091046333312988,
|
| 1161 |
+
"learning_rate": 0.0005558905512258946,
|
| 1162 |
+
"loss": 1.2868,
|
| 1163 |
+
"step": 16100
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 0.9215541270834519,
|
| 1167 |
+
"grad_norm": 1.656204104423523,
|
| 1168 |
+
"learning_rate": 0.0005543735896998313,
|
| 1169 |
+
"loss": 1.3075,
|
| 1170 |
+
"step": 16200
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 0.9272427328061892,
|
| 1174 |
+
"grad_norm": 1.5555946826934814,
|
| 1175 |
+
"learning_rate": 0.0005528566281737679,
|
| 1176 |
+
"loss": 1.2963,
|
| 1177 |
+
"step": 16300
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 0.9329313385289265,
|
| 1181 |
+
"grad_norm": 1.7379626035690308,
|
| 1182 |
+
"learning_rate": 0.0005513396666477047,
|
| 1183 |
+
"loss": 1.2905,
|
| 1184 |
+
"step": 16400
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.938619944251664,
|
| 1188 |
+
"grad_norm": 1.5103166103363037,
|
| 1189 |
+
"learning_rate": 0.0005498227051216414,
|
| 1190 |
+
"loss": 1.2848,
|
| 1191 |
+
"step": 16500
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.9443085499744013,
|
| 1195 |
+
"grad_norm": 1.5895978212356567,
|
| 1196 |
+
"learning_rate": 0.000548305743595578,
|
| 1197 |
+
"loss": 1.293,
|
| 1198 |
+
"step": 16600
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 0.9499971556971386,
|
| 1202 |
+
"grad_norm": 1.6526978015899658,
|
| 1203 |
+
"learning_rate": 0.0005467887820695148,
|
| 1204 |
+
"loss": 1.288,
|
| 1205 |
+
"step": 16700
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 0.9556857614198759,
|
| 1209 |
+
"grad_norm": 1.7471717596054077,
|
| 1210 |
+
"learning_rate": 0.0005452718205434515,
|
| 1211 |
+
"loss": 1.3099,
|
| 1212 |
+
"step": 16800
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 0.9613743671426134,
|
| 1216 |
+
"grad_norm": 1.5995450019836426,
|
| 1217 |
+
"learning_rate": 0.0005437700286326488,
|
| 1218 |
+
"loss": 1.2886,
|
| 1219 |
+
"step": 16900
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.9670629728653507,
|
| 1223 |
+
"grad_norm": 1.7462047338485718,
|
| 1224 |
+
"learning_rate": 0.0005422530671065856,
|
| 1225 |
+
"loss": 1.317,
|
| 1226 |
+
"step": 17000
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.972751578588088,
|
| 1230 |
+
"grad_norm": 1.5739308595657349,
|
| 1231 |
+
"learning_rate": 0.0005407361055805223,
|
| 1232 |
+
"loss": 1.2997,
|
| 1233 |
+
"step": 17100
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.9784401843108255,
|
| 1237 |
+
"grad_norm": 1.6608139276504517,
|
| 1238 |
+
"learning_rate": 0.0005392191440544589,
|
| 1239 |
+
"loss": 1.3037,
|
| 1240 |
+
"step": 17200
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 0.9841287900335628,
|
| 1244 |
+
"grad_norm": 1.7515637874603271,
|
| 1245 |
+
"learning_rate": 0.0005377021825283956,
|
| 1246 |
+
"loss": 1.302,
|
| 1247 |
+
"step": 17300
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 0.9898173957563001,
|
| 1251 |
+
"grad_norm": 1.572986364364624,
|
| 1252 |
+
"learning_rate": 0.0005361852210023323,
|
| 1253 |
+
"loss": 1.2945,
|
| 1254 |
+
"step": 17400
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 0.9955060014790375,
|
| 1258 |
+
"grad_norm": 1.9207016229629517,
|
| 1259 |
+
"learning_rate": 0.000534668259476269,
|
| 1260 |
+
"loss": 1.2747,
|
| 1261 |
+
"step": 17500
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.0011946072017748,
|
| 1265 |
+
"grad_norm": 1.9010945558547974,
|
| 1266 |
+
"learning_rate": 0.0005331512979502058,
|
| 1267 |
+
"loss": 1.263,
|
| 1268 |
+
"step": 17600
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.0068832129245122,
|
| 1272 |
+
"grad_norm": 2.4259393215179443,
|
| 1273 |
+
"learning_rate": 0.0005316343364241425,
|
| 1274 |
+
"loss": 1.2741,
|
| 1275 |
+
"step": 17700
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.0125718186472497,
|
| 1279 |
+
"grad_norm": 2.5002028942108154,
|
| 1280 |
+
"learning_rate": 0.0005301173748980791,
|
| 1281 |
+
"loss": 1.2686,
|
| 1282 |
+
"step": 17800
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.0182604243699869,
|
| 1286 |
+
"grad_norm": 1.7075704336166382,
|
| 1287 |
+
"learning_rate": 0.0005286004133720159,
|
| 1288 |
+
"loss": 1.2661,
|
| 1289 |
+
"step": 17900
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.0239490300927243,
|
| 1293 |
+
"grad_norm": 1.7390458583831787,
|
| 1294 |
+
"learning_rate": 0.0005270834518459526,
|
| 1295 |
+
"loss": 1.2698,
|
| 1296 |
+
"step": 18000
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 1.0296376358154615,
|
| 1300 |
+
"grad_norm": 1.980185627937317,
|
| 1301 |
+
"learning_rate": 0.0005255664903198893,
|
| 1302 |
+
"loss": 1.2569,
|
| 1303 |
+
"step": 18100
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.035326241538199,
|
| 1307 |
+
"grad_norm": 1.79970383644104,
|
| 1308 |
+
"learning_rate": 0.0005240495287938261,
|
| 1309 |
+
"loss": 1.2738,
|
| 1310 |
+
"step": 18200
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.0410148472609364,
|
| 1314 |
+
"grad_norm": 1.6184749603271484,
|
| 1315 |
+
"learning_rate": 0.0005225325672677627,
|
| 1316 |
+
"loss": 1.2637,
|
| 1317 |
+
"step": 18300
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 1.0467034529836736,
|
| 1321 |
+
"grad_norm": 2.3463358879089355,
|
| 1322 |
+
"learning_rate": 0.0005210156057416993,
|
| 1323 |
+
"loss": 1.2665,
|
| 1324 |
+
"step": 18400
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 1.052392058706411,
|
| 1328 |
+
"grad_norm": 1.8550745248794556,
|
| 1329 |
+
"learning_rate": 0.0005194986442156361,
|
| 1330 |
+
"loss": 1.2664,
|
| 1331 |
+
"step": 18500
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.0580806644291485,
|
| 1335 |
+
"grad_norm": 1.8582580089569092,
|
| 1336 |
+
"learning_rate": 0.0005179816826895728,
|
| 1337 |
+
"loss": 1.2442,
|
| 1338 |
+
"step": 18600
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.0637692701518857,
|
| 1342 |
+
"grad_norm": 1.88007390499115,
|
| 1343 |
+
"learning_rate": 0.0005164647211635095,
|
| 1344 |
+
"loss": 1.2536,
|
| 1345 |
+
"step": 18700
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.0694578758746232,
|
| 1349 |
+
"grad_norm": 1.804671287536621,
|
| 1350 |
+
"learning_rate": 0.0005149477596374462,
|
| 1351 |
+
"loss": 1.2459,
|
| 1352 |
+
"step": 18800
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.0751464815973604,
|
| 1356 |
+
"grad_norm": 1.7329107522964478,
|
| 1357 |
+
"learning_rate": 0.0005134307981113829,
|
| 1358 |
+
"loss": 1.2499,
|
| 1359 |
+
"step": 18900
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.0808350873200978,
|
| 1363 |
+
"grad_norm": 1.693323016166687,
|
| 1364 |
+
"learning_rate": 0.0005119290062005802,
|
| 1365 |
+
"loss": 1.25,
|
| 1366 |
+
"step": 19000
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 1.0865236930428352,
|
| 1370 |
+
"grad_norm": 1.600060224533081,
|
| 1371 |
+
"learning_rate": 0.000510412044674517,
|
| 1372 |
+
"loss": 1.2515,
|
| 1373 |
+
"step": 19100
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.0922122987655725,
|
| 1377 |
+
"grad_norm": 1.8084614276885986,
|
| 1378 |
+
"learning_rate": 0.0005088950831484537,
|
| 1379 |
+
"loss": 1.246,
|
| 1380 |
+
"step": 19200
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 1.09790090448831,
|
| 1384 |
+
"grad_norm": 1.8022205829620361,
|
| 1385 |
+
"learning_rate": 0.0005073781216223904,
|
| 1386 |
+
"loss": 1.2597,
|
| 1387 |
+
"step": 19300
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.1035895102110473,
|
| 1391 |
+
"grad_norm": 1.6137562990188599,
|
| 1392 |
+
"learning_rate": 0.0005058611600963271,
|
| 1393 |
+
"loss": 1.2685,
|
| 1394 |
+
"step": 19400
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.1092781159337846,
|
| 1398 |
+
"grad_norm": 1.7756201028823853,
|
| 1399 |
+
"learning_rate": 0.0005043441985702637,
|
| 1400 |
+
"loss": 1.2606,
|
| 1401 |
+
"step": 19500
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 1.114966721656522,
|
| 1405 |
+
"grad_norm": 1.8828805685043335,
|
| 1406 |
+
"learning_rate": 0.0005028272370442004,
|
| 1407 |
+
"loss": 1.2582,
|
| 1408 |
+
"step": 19600
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.1206553273792594,
|
| 1412 |
+
"grad_norm": 1.6829185485839844,
|
| 1413 |
+
"learning_rate": 0.0005013102755181372,
|
| 1414 |
+
"loss": 1.2563,
|
| 1415 |
+
"step": 19700
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 1.1263439331019967,
|
| 1419 |
+
"grad_norm": 1.6716195344924927,
|
| 1420 |
+
"learning_rate": 0.0004997933139920739,
|
| 1421 |
+
"loss": 1.2405,
|
| 1422 |
+
"step": 19800
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 1.132032538824734,
|
| 1426 |
+
"grad_norm": 1.7629872560501099,
|
| 1427 |
+
"learning_rate": 0.0004982763524660106,
|
| 1428 |
+
"loss": 1.2649,
|
| 1429 |
+
"step": 19900
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.1377211445474713,
|
| 1433 |
+
"grad_norm": 1.704967737197876,
|
| 1434 |
+
"learning_rate": 0.0004967593909399473,
|
| 1435 |
+
"loss": 1.226,
|
| 1436 |
+
"step": 20000
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.1377211445474713,
|
| 1440 |
+
"eval_accuracy": 0.692396,
|
| 1441 |
+
"eval_loss": 1.2322564125061035,
|
| 1442 |
+
"eval_runtime": 82.0826,
|
| 1443 |
+
"eval_samples_per_second": 3045.712,
|
| 1444 |
+
"eval_steps_per_second": 11.903,
|
| 1445 |
+
"step": 20000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.1434097502702087,
|
| 1449 |
+
"grad_norm": 1.5182781219482422,
|
| 1450 |
+
"learning_rate": 0.000495242429413884,
|
| 1451 |
+
"loss": 1.2343,
|
| 1452 |
+
"step": 20100
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.1490983559929462,
|
| 1456 |
+
"grad_norm": 2.637796640396118,
|
| 1457 |
+
"learning_rate": 0.0004937254678878207,
|
| 1458 |
+
"loss": 1.2506,
|
| 1459 |
+
"step": 20200
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.1547869617156834,
|
| 1463 |
+
"grad_norm": 1.8955748081207275,
|
| 1464 |
+
"learning_rate": 0.0004922085063617575,
|
| 1465 |
+
"loss": 1.2625,
|
| 1466 |
+
"step": 20300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.1604755674384208,
|
| 1470 |
+
"grad_norm": 2.0370640754699707,
|
| 1471 |
+
"learning_rate": 0.0004906915448356942,
|
| 1472 |
+
"loss": 1.2551,
|
| 1473 |
+
"step": 20400
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.1661641731611583,
|
| 1477 |
+
"grad_norm": 1.8047020435333252,
|
| 1478 |
+
"learning_rate": 0.0004891745833096308,
|
| 1479 |
+
"loss": 1.2489,
|
| 1480 |
+
"step": 20500
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.1718527788838955,
|
| 1484 |
+
"grad_norm": 1.5440089702606201,
|
| 1485 |
+
"learning_rate": 0.0004876576217835675,
|
| 1486 |
+
"loss": 1.2646,
|
| 1487 |
+
"step": 20600
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.177541384606633,
|
| 1491 |
+
"grad_norm": 1.5029830932617188,
|
| 1492 |
+
"learning_rate": 0.00048614066025750425,
|
| 1493 |
+
"loss": 1.2536,
|
| 1494 |
+
"step": 20700
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.1832299903293704,
|
| 1498 |
+
"grad_norm": 1.4674205780029297,
|
| 1499 |
+
"learning_rate": 0.0004846236987314409,
|
| 1500 |
+
"loss": 1.2457,
|
| 1501 |
+
"step": 20800
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.1889185960521076,
|
| 1505 |
+
"grad_norm": 1.5259037017822266,
|
| 1506 |
+
"learning_rate": 0.00048310673720537765,
|
| 1507 |
+
"loss": 1.2449,
|
| 1508 |
+
"step": 20900
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.194607201774845,
|
| 1512 |
+
"grad_norm": 1.6339012384414673,
|
| 1513 |
+
"learning_rate": 0.0004815897756793144,
|
| 1514 |
+
"loss": 1.2163,
|
| 1515 |
+
"step": 21000
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.2002958074975822,
|
| 1519 |
+
"grad_norm": 1.5565885305404663,
|
| 1520 |
+
"learning_rate": 0.00048007281415325106,
|
| 1521 |
+
"loss": 1.2461,
|
| 1522 |
+
"step": 21100
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.2059844132203197,
|
| 1526 |
+
"grad_norm": 1.676540493965149,
|
| 1527 |
+
"learning_rate": 0.0004785558526271878,
|
| 1528 |
+
"loss": 1.2555,
|
| 1529 |
+
"step": 21200
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.2116730189430571,
|
| 1533 |
+
"grad_norm": 1.6003342866897583,
|
| 1534 |
+
"learning_rate": 0.00047705406071638513,
|
| 1535 |
+
"loss": 1.2507,
|
| 1536 |
+
"step": 21300
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.2173616246657943,
|
| 1540 |
+
"grad_norm": 2.2655630111694336,
|
| 1541 |
+
"learning_rate": 0.00047553709919032175,
|
| 1542 |
+
"loss": 1.2144,
|
| 1543 |
+
"step": 21400
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.2230502303885318,
|
| 1547 |
+
"grad_norm": 1.695094108581543,
|
| 1548 |
+
"learning_rate": 0.0004740201376642585,
|
| 1549 |
+
"loss": 1.2415,
|
| 1550 |
+
"step": 21500
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.2287388361112692,
|
| 1554 |
+
"grad_norm": 1.8387731313705444,
|
| 1555 |
+
"learning_rate": 0.0004725031761381952,
|
| 1556 |
+
"loss": 1.2406,
|
| 1557 |
+
"step": 21600
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.2344274418340064,
|
| 1561 |
+
"grad_norm": 1.6776598691940308,
|
| 1562 |
+
"learning_rate": 0.0004709862146121319,
|
| 1563 |
+
"loss": 1.2673,
|
| 1564 |
+
"step": 21700
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.2401160475567439,
|
| 1568 |
+
"grad_norm": 1.6573506593704224,
|
| 1569 |
+
"learning_rate": 0.0004694692530860686,
|
| 1570 |
+
"loss": 1.2587,
|
| 1571 |
+
"step": 21800
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.2458046532794813,
|
| 1575 |
+
"grad_norm": 1.6786317825317383,
|
| 1576 |
+
"learning_rate": 0.00046795229156000535,
|
| 1577 |
+
"loss": 1.2464,
|
| 1578 |
+
"step": 21900
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.2514932590022185,
|
| 1582 |
+
"grad_norm": 1.887971043586731,
|
| 1583 |
+
"learning_rate": 0.00046643533003394203,
|
| 1584 |
+
"loss": 1.2501,
|
| 1585 |
+
"step": 22000
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.257181864724956,
|
| 1589 |
+
"grad_norm": 1.7499343156814575,
|
| 1590 |
+
"learning_rate": 0.00046491836850787876,
|
| 1591 |
+
"loss": 1.2296,
|
| 1592 |
+
"step": 22100
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.2628704704476932,
|
| 1596 |
+
"grad_norm": 2.057670831680298,
|
| 1597 |
+
"learning_rate": 0.00046340140698181544,
|
| 1598 |
+
"loss": 1.2346,
|
| 1599 |
+
"step": 22200
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.2685590761704306,
|
| 1603 |
+
"grad_norm": 1.7353135347366333,
|
| 1604 |
+
"learning_rate": 0.00046188444545575217,
|
| 1605 |
+
"loss": 1.2512,
|
| 1606 |
+
"step": 22300
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.274247681893168,
|
| 1610 |
+
"grad_norm": 2.0662734508514404,
|
| 1611 |
+
"learning_rate": 0.0004603674839296889,
|
| 1612 |
+
"loss": 1.2493,
|
| 1613 |
+
"step": 22400
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.2799362876159053,
|
| 1617 |
+
"grad_norm": 1.5519914627075195,
|
| 1618 |
+
"learning_rate": 0.0004588505224036255,
|
| 1619 |
+
"loss": 1.2404,
|
| 1620 |
+
"step": 22500
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.2856248933386427,
|
| 1624 |
+
"grad_norm": 1.8667906522750854,
|
| 1625 |
+
"learning_rate": 0.0004573335608775622,
|
| 1626 |
+
"loss": 1.247,
|
| 1627 |
+
"step": 22600
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.29131349906138,
|
| 1631 |
+
"grad_norm": 1.8621453046798706,
|
| 1632 |
+
"learning_rate": 0.00045581659935149893,
|
| 1633 |
+
"loss": 1.2428,
|
| 1634 |
+
"step": 22700
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.2970021047841174,
|
| 1638 |
+
"grad_norm": 1.7203937768936157,
|
| 1639 |
+
"learning_rate": 0.00045429963782543566,
|
| 1640 |
+
"loss": 1.2273,
|
| 1641 |
+
"step": 22800
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.3026907105068548,
|
| 1645 |
+
"grad_norm": 1.7497667074203491,
|
| 1646 |
+
"learning_rate": 0.00045278267629937234,
|
| 1647 |
+
"loss": 1.2458,
|
| 1648 |
+
"step": 22900
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.3083793162295922,
|
| 1652 |
+
"grad_norm": 2.057507276535034,
|
| 1653 |
+
"learning_rate": 0.0004512657147733091,
|
| 1654 |
+
"loss": 1.2325,
|
| 1655 |
+
"step": 23000
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.3140679219523295,
|
| 1659 |
+
"grad_norm": 1.4594337940216064,
|
| 1660 |
+
"learning_rate": 0.0004497487532472458,
|
| 1661 |
+
"loss": 1.2319,
|
| 1662 |
+
"step": 23100
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.319756527675067,
|
| 1666 |
+
"grad_norm": 2.1696736812591553,
|
| 1667 |
+
"learning_rate": 0.0004482317917211825,
|
| 1668 |
+
"loss": 1.234,
|
| 1669 |
+
"step": 23200
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.3254451333978041,
|
| 1673 |
+
"grad_norm": 1.8165019750595093,
|
| 1674 |
+
"learning_rate": 0.0004467148301951192,
|
| 1675 |
+
"loss": 1.2256,
|
| 1676 |
+
"step": 23300
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.3311337391205416,
|
| 1680 |
+
"grad_norm": 1.5531728267669678,
|
| 1681 |
+
"learning_rate": 0.00044519786866905594,
|
| 1682 |
+
"loss": 1.2518,
|
| 1683 |
+
"step": 23400
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.336822344843279,
|
| 1687 |
+
"grad_norm": 1.4592831134796143,
|
| 1688 |
+
"learning_rate": 0.0004436809071429926,
|
| 1689 |
+
"loss": 1.2192,
|
| 1690 |
+
"step": 23500
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.3425109505660162,
|
| 1694 |
+
"grad_norm": 1.74478280544281,
|
| 1695 |
+
"learning_rate": 0.00044216394561692935,
|
| 1696 |
+
"loss": 1.2427,
|
| 1697 |
+
"step": 23600
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.3481995562887537,
|
| 1701 |
+
"grad_norm": 1.8685113191604614,
|
| 1702 |
+
"learning_rate": 0.000440646984090866,
|
| 1703 |
+
"loss": 1.2581,
|
| 1704 |
+
"step": 23700
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.3538881620114909,
|
| 1708 |
+
"grad_norm": 1.7366535663604736,
|
| 1709 |
+
"learning_rate": 0.0004391300225648027,
|
| 1710 |
+
"loss": 1.2471,
|
| 1711 |
+
"step": 23800
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.3595767677342283,
|
| 1715 |
+
"grad_norm": 1.6585444211959839,
|
| 1716 |
+
"learning_rate": 0.0004376130610387394,
|
| 1717 |
+
"loss": 1.2198,
|
| 1718 |
+
"step": 23900
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 1.3652653734569657,
|
| 1722 |
+
"grad_norm": 1.9299806356430054,
|
| 1723 |
+
"learning_rate": 0.0004360960995126761,
|
| 1724 |
+
"loss": 1.2314,
|
| 1725 |
+
"step": 24000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 1.3709539791797032,
|
| 1729 |
+
"grad_norm": 1.8172481060028076,
|
| 1730 |
+
"learning_rate": 0.00043457913798661285,
|
| 1731 |
+
"loss": 1.2098,
|
| 1732 |
+
"step": 24100
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 1.3766425849024404,
|
| 1736 |
+
"grad_norm": 1.5579493045806885,
|
| 1737 |
+
"learning_rate": 0.0004330621764605495,
|
| 1738 |
+
"loss": 1.2043,
|
| 1739 |
+
"step": 24200
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.3823311906251778,
|
| 1743 |
+
"grad_norm": 1.8178203105926514,
|
| 1744 |
+
"learning_rate": 0.00043154521493448625,
|
| 1745 |
+
"loss": 1.2235,
|
| 1746 |
+
"step": 24300
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.388019796347915,
|
| 1750 |
+
"grad_norm": 1.676126480102539,
|
| 1751 |
+
"learning_rate": 0.000430028253408423,
|
| 1752 |
+
"loss": 1.2388,
|
| 1753 |
+
"step": 24400
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.3937084020706525,
|
| 1757 |
+
"grad_norm": 1.863893985748291,
|
| 1758 |
+
"learning_rate": 0.00042851129188235966,
|
| 1759 |
+
"loss": 1.2206,
|
| 1760 |
+
"step": 24500
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.39939700779339,
|
| 1764 |
+
"grad_norm": 1.5618318319320679,
|
| 1765 |
+
"learning_rate": 0.0004269943303562964,
|
| 1766 |
+
"loss": 1.2359,
|
| 1767 |
+
"step": 24600
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 1.4050856135161272,
|
| 1771 |
+
"grad_norm": 1.3972681760787964,
|
| 1772 |
+
"learning_rate": 0.0004254773688302331,
|
| 1773 |
+
"loss": 1.2152,
|
| 1774 |
+
"step": 24700
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 1.4107742192388646,
|
| 1778 |
+
"grad_norm": 1.584274411201477,
|
| 1779 |
+
"learning_rate": 0.00042396040730416975,
|
| 1780 |
+
"loss": 1.211,
|
| 1781 |
+
"step": 24800
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.4164628249616018,
|
| 1785 |
+
"grad_norm": 1.7282276153564453,
|
| 1786 |
+
"learning_rate": 0.0004224434457781064,
|
| 1787 |
+
"loss": 1.2034,
|
| 1788 |
+
"step": 24900
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 1.4221514306843392,
|
| 1792 |
+
"grad_norm": 2.2420654296875,
|
| 1793 |
+
"learning_rate": 0.00042092648425204316,
|
| 1794 |
+
"loss": 1.2125,
|
| 1795 |
+
"step": 25000
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.4221514306843392,
|
| 1799 |
+
"eval_accuracy": 0.703072,
|
| 1800 |
+
"eval_loss": 1.185011863708496,
|
| 1801 |
+
"eval_runtime": 81.7158,
|
| 1802 |
+
"eval_samples_per_second": 3059.385,
|
| 1803 |
+
"eval_steps_per_second": 11.956,
|
| 1804 |
+
"step": 25000
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 1.4278400364070767,
|
| 1808 |
+
"grad_norm": 1.5998648405075073,
|
| 1809 |
+
"learning_rate": 0.0004194095227259799,
|
| 1810 |
+
"loss": 1.2145,
|
| 1811 |
+
"step": 25100
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 1.433528642129814,
|
| 1815 |
+
"grad_norm": 1.8546173572540283,
|
| 1816 |
+
"learning_rate": 0.00041789256119991656,
|
| 1817 |
+
"loss": 1.2293,
|
| 1818 |
+
"step": 25200
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 1.4392172478525513,
|
| 1822 |
+
"grad_norm": 1.6815022230148315,
|
| 1823 |
+
"learning_rate": 0.00041639076928911395,
|
| 1824 |
+
"loss": 1.2165,
|
| 1825 |
+
"step": 25300
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 1.4449058535752888,
|
| 1829 |
+
"grad_norm": 1.5567988157272339,
|
| 1830 |
+
"learning_rate": 0.0004148889773783113,
|
| 1831 |
+
"loss": 1.2231,
|
| 1832 |
+
"step": 25400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 1.450594459298026,
|
| 1836 |
+
"grad_norm": 1.9424728155136108,
|
| 1837 |
+
"learning_rate": 0.000413372015852248,
|
| 1838 |
+
"loss": 1.2302,
|
| 1839 |
+
"step": 25500
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 1.4562830650207634,
|
| 1843 |
+
"grad_norm": 1.9052510261535645,
|
| 1844 |
+
"learning_rate": 0.00041185505432618464,
|
| 1845 |
+
"loss": 1.2174,
|
| 1846 |
+
"step": 25600
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 1.4619716707435009,
|
| 1850 |
+
"grad_norm": 1.470513939857483,
|
| 1851 |
+
"learning_rate": 0.0004103380928001213,
|
| 1852 |
+
"loss": 1.2167,
|
| 1853 |
+
"step": 25700
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 1.467660276466238,
|
| 1857 |
+
"grad_norm": 1.5082899332046509,
|
| 1858 |
+
"learning_rate": 0.00040882113127405805,
|
| 1859 |
+
"loss": 1.2098,
|
| 1860 |
+
"step": 25800
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 1.4733488821889755,
|
| 1864 |
+
"grad_norm": 1.7309447526931763,
|
| 1865 |
+
"learning_rate": 0.0004073041697479948,
|
| 1866 |
+
"loss": 1.2111,
|
| 1867 |
+
"step": 25900
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 1.4790374879117127,
|
| 1871 |
+
"grad_norm": 1.7894470691680908,
|
| 1872 |
+
"learning_rate": 0.00040578720822193146,
|
| 1873 |
+
"loss": 1.2237,
|
| 1874 |
+
"step": 26000
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.4847260936344502,
|
| 1878 |
+
"grad_norm": 1.712557077407837,
|
| 1879 |
+
"learning_rate": 0.0004042702466958682,
|
| 1880 |
+
"loss": 1.2314,
|
| 1881 |
+
"step": 26100
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 1.4904146993571876,
|
| 1885 |
+
"grad_norm": 1.794565200805664,
|
| 1886 |
+
"learning_rate": 0.0004027532851698049,
|
| 1887 |
+
"loss": 1.2072,
|
| 1888 |
+
"step": 26200
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 1.4961033050799248,
|
| 1892 |
+
"grad_norm": 1.655179500579834,
|
| 1893 |
+
"learning_rate": 0.0004012363236437416,
|
| 1894 |
+
"loss": 1.218,
|
| 1895 |
+
"step": 26300
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 1.5017919108026623,
|
| 1899 |
+
"grad_norm": 1.8749343156814575,
|
| 1900 |
+
"learning_rate": 0.00039971936211767833,
|
| 1901 |
+
"loss": 1.2037,
|
| 1902 |
+
"step": 26400
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 1.5074805165253995,
|
| 1906 |
+
"grad_norm": 1.5337320566177368,
|
| 1907 |
+
"learning_rate": 0.000398202400591615,
|
| 1908 |
+
"loss": 1.2179,
|
| 1909 |
+
"step": 26500
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 1.513169122248137,
|
| 1913 |
+
"grad_norm": 1.5731686353683472,
|
| 1914 |
+
"learning_rate": 0.0003966854390655517,
|
| 1915 |
+
"loss": 1.2037,
|
| 1916 |
+
"step": 26600
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 1.5188577279708744,
|
| 1920 |
+
"grad_norm": 1.5700329542160034,
|
| 1921 |
+
"learning_rate": 0.0003951684775394884,
|
| 1922 |
+
"loss": 1.2189,
|
| 1923 |
+
"step": 26700
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 1.5245463336936118,
|
| 1927 |
+
"grad_norm": 1.9315118789672852,
|
| 1928 |
+
"learning_rate": 0.00039365151601342515,
|
| 1929 |
+
"loss": 1.2314,
|
| 1930 |
+
"step": 26800
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 1.530234939416349,
|
| 1934 |
+
"grad_norm": 1.6017844676971436,
|
| 1935 |
+
"learning_rate": 0.0003921345544873618,
|
| 1936 |
+
"loss": 1.211,
|
| 1937 |
+
"step": 26900
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 1.5359235451390862,
|
| 1941 |
+
"grad_norm": 1.586595058441162,
|
| 1942 |
+
"learning_rate": 0.0003906175929612985,
|
| 1943 |
+
"loss": 1.2079,
|
| 1944 |
+
"step": 27000
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 1.5416121508618237,
|
| 1948 |
+
"grad_norm": 1.8215593099594116,
|
| 1949 |
+
"learning_rate": 0.00038910063143523523,
|
| 1950 |
+
"loss": 1.2022,
|
| 1951 |
+
"step": 27100
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 1.5473007565845611,
|
| 1955 |
+
"grad_norm": 1.7390124797821045,
|
| 1956 |
+
"learning_rate": 0.00038758366990917196,
|
| 1957 |
+
"loss": 1.2143,
|
| 1958 |
+
"step": 27200
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 1.5529893623072986,
|
| 1962 |
+
"grad_norm": 1.792608618736267,
|
| 1963 |
+
"learning_rate": 0.00038606670838310864,
|
| 1964 |
+
"loss": 1.2104,
|
| 1965 |
+
"step": 27300
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.558677968030036,
|
| 1969 |
+
"grad_norm": 1.802167296409607,
|
| 1970 |
+
"learning_rate": 0.00038454974685704537,
|
| 1971 |
+
"loss": 1.1924,
|
| 1972 |
+
"step": 27400
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 1.5643665737527732,
|
| 1976 |
+
"grad_norm": 1.7943332195281982,
|
| 1977 |
+
"learning_rate": 0.0003830327853309821,
|
| 1978 |
+
"loss": 1.2096,
|
| 1979 |
+
"step": 27500
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 1.5700551794755104,
|
| 1983 |
+
"grad_norm": 1.745893120765686,
|
| 1984 |
+
"learning_rate": 0.0003815158238049187,
|
| 1985 |
+
"loss": 1.1941,
|
| 1986 |
+
"step": 27600
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 1.5757437851982479,
|
| 1990 |
+
"grad_norm": 1.6740118265151978,
|
| 1991 |
+
"learning_rate": 0.00037999886227885546,
|
| 1992 |
+
"loss": 1.2355,
|
| 1993 |
+
"step": 27700
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 1.5814323909209853,
|
| 1997 |
+
"grad_norm": 1.681840419769287,
|
| 1998 |
+
"learning_rate": 0.0003784970703680528,
|
| 1999 |
+
"loss": 1.2034,
|
| 2000 |
+
"step": 27800
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 1.5871209966437227,
|
| 2004 |
+
"grad_norm": 1.6897751092910767,
|
| 2005 |
+
"learning_rate": 0.0003769801088419895,
|
| 2006 |
+
"loss": 1.2169,
|
| 2007 |
+
"step": 27900
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 1.59280960236646,
|
| 2011 |
+
"grad_norm": 1.686784267425537,
|
| 2012 |
+
"learning_rate": 0.0003754631473159262,
|
| 2013 |
+
"loss": 1.2093,
|
| 2014 |
+
"step": 28000
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 1.5984982080891972,
|
| 2018 |
+
"grad_norm": 1.6020421981811523,
|
| 2019 |
+
"learning_rate": 0.00037394618578986293,
|
| 2020 |
+
"loss": 1.2131,
|
| 2021 |
+
"step": 28100
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 1.6041868138119346,
|
| 2025 |
+
"grad_norm": 1.478246808052063,
|
| 2026 |
+
"learning_rate": 0.0003724292242637996,
|
| 2027 |
+
"loss": 1.2048,
|
| 2028 |
+
"step": 28200
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 1.609875419534672,
|
| 2032 |
+
"grad_norm": 1.4912410974502563,
|
| 2033 |
+
"learning_rate": 0.00037091226273773634,
|
| 2034 |
+
"loss": 1.18,
|
| 2035 |
+
"step": 28300
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 1.6155640252574095,
|
| 2039 |
+
"grad_norm": 1.6362539529800415,
|
| 2040 |
+
"learning_rate": 0.00036939530121167307,
|
| 2041 |
+
"loss": 1.2022,
|
| 2042 |
+
"step": 28400
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 1.6212526309801467,
|
| 2046 |
+
"grad_norm": 1.5238479375839233,
|
| 2047 |
+
"learning_rate": 0.00036787833968560975,
|
| 2048 |
+
"loss": 1.2105,
|
| 2049 |
+
"step": 28500
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 1.6269412367028842,
|
| 2053 |
+
"grad_norm": 1.6359635591506958,
|
| 2054 |
+
"learning_rate": 0.0003663613781595464,
|
| 2055 |
+
"loss": 1.1833,
|
| 2056 |
+
"step": 28600
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.6326298424256214,
|
| 2060 |
+
"grad_norm": 1.6206257343292236,
|
| 2061 |
+
"learning_rate": 0.00036484441663348316,
|
| 2062 |
+
"loss": 1.1945,
|
| 2063 |
+
"step": 28700
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 1.6383184481483588,
|
| 2067 |
+
"grad_norm": 1.7032015323638916,
|
| 2068 |
+
"learning_rate": 0.00036332745510741983,
|
| 2069 |
+
"loss": 1.2065,
|
| 2070 |
+
"step": 28800
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 1.6440070538710962,
|
| 2074 |
+
"grad_norm": 1.7177228927612305,
|
| 2075 |
+
"learning_rate": 0.00036181049358135657,
|
| 2076 |
+
"loss": 1.2035,
|
| 2077 |
+
"step": 28900
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 1.6496956595938337,
|
| 2081 |
+
"grad_norm": 1.5967752933502197,
|
| 2082 |
+
"learning_rate": 0.0003602935320552933,
|
| 2083 |
+
"loss": 1.2036,
|
| 2084 |
+
"step": 29000
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 1.655384265316571,
|
| 2088 |
+
"grad_norm": 1.6632803678512573,
|
| 2089 |
+
"learning_rate": 0.00035877657052923,
|
| 2090 |
+
"loss": 1.226,
|
| 2091 |
+
"step": 29100
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 1.6610728710393081,
|
| 2095 |
+
"grad_norm": 1.5134357213974,
|
| 2096 |
+
"learning_rate": 0.00035725960900316665,
|
| 2097 |
+
"loss": 1.1947,
|
| 2098 |
+
"step": 29200
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 1.6667614767620456,
|
| 2102 |
+
"grad_norm": 1.5506322383880615,
|
| 2103 |
+
"learning_rate": 0.0003557426474771034,
|
| 2104 |
+
"loss": 1.1963,
|
| 2105 |
+
"step": 29300
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 1.672450082484783,
|
| 2109 |
+
"grad_norm": 1.4821183681488037,
|
| 2110 |
+
"learning_rate": 0.0003542256859510401,
|
| 2111 |
+
"loss": 1.2039,
|
| 2112 |
+
"step": 29400
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 1.6781386882075204,
|
| 2116 |
+
"grad_norm": 2.278379440307617,
|
| 2117 |
+
"learning_rate": 0.0003527087244249768,
|
| 2118 |
+
"loss": 1.205,
|
| 2119 |
+
"step": 29500
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 1.6838272939302577,
|
| 2123 |
+
"grad_norm": 1.5077921152114868,
|
| 2124 |
+
"learning_rate": 0.0003511917628989135,
|
| 2125 |
+
"loss": 1.1984,
|
| 2126 |
+
"step": 29600
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 1.689515899652995,
|
| 2130 |
+
"grad_norm": 1.629607915878296,
|
| 2131 |
+
"learning_rate": 0.0003496748013728502,
|
| 2132 |
+
"loss": 1.2023,
|
| 2133 |
+
"step": 29700
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 1.6952045053757323,
|
| 2137 |
+
"grad_norm": 1.5007668733596802,
|
| 2138 |
+
"learning_rate": 0.0003481578398467869,
|
| 2139 |
+
"loss": 1.1907,
|
| 2140 |
+
"step": 29800
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 1.7008931110984697,
|
| 2144 |
+
"grad_norm": 1.7543882131576538,
|
| 2145 |
+
"learning_rate": 0.0003466408783207236,
|
| 2146 |
+
"loss": 1.1949,
|
| 2147 |
+
"step": 29900
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.7065817168212072,
|
| 2151 |
+
"grad_norm": 1.6254594326019287,
|
| 2152 |
+
"learning_rate": 0.00034512391679466034,
|
| 2153 |
+
"loss": 1.1912,
|
| 2154 |
+
"step": 30000
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 1.7065817168212072,
|
| 2158 |
+
"eval_accuracy": 0.709248,
|
| 2159 |
+
"eval_loss": 1.1566522121429443,
|
| 2160 |
+
"eval_runtime": 79.5399,
|
| 2161 |
+
"eval_samples_per_second": 3143.077,
|
| 2162 |
+
"eval_steps_per_second": 12.283,
|
| 2163 |
+
"step": 30000
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 1.7122703225439446,
|
| 2167 |
+
"grad_norm": 1.873049020767212,
|
| 2168 |
+
"learning_rate": 0.000343606955268597,
|
| 2169 |
+
"loss": 1.2063,
|
| 2170 |
+
"step": 30100
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 1.7179589282666818,
|
| 2174 |
+
"grad_norm": 1.5862141847610474,
|
| 2175 |
+
"learning_rate": 0.00034208999374253375,
|
| 2176 |
+
"loss": 1.1926,
|
| 2177 |
+
"step": 30200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.723647533989419,
|
| 2181 |
+
"grad_norm": 1.9915696382522583,
|
| 2182 |
+
"learning_rate": 0.0003405730322164704,
|
| 2183 |
+
"loss": 1.1952,
|
| 2184 |
+
"step": 30300
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 1.7293361397121565,
|
| 2188 |
+
"grad_norm": 1.856048822402954,
|
| 2189 |
+
"learning_rate": 0.0003390560706904071,
|
| 2190 |
+
"loss": 1.1953,
|
| 2191 |
+
"step": 30400
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 1.735024745434894,
|
| 2195 |
+
"grad_norm": 1.6758267879486084,
|
| 2196 |
+
"learning_rate": 0.00033753910916434383,
|
| 2197 |
+
"loss": 1.1906,
|
| 2198 |
+
"step": 30500
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 1.7407133511576314,
|
| 2202 |
+
"grad_norm": 1.8683140277862549,
|
| 2203 |
+
"learning_rate": 0.00033602214763828056,
|
| 2204 |
+
"loss": 1.2025,
|
| 2205 |
+
"step": 30600
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 1.7464019568803686,
|
| 2209 |
+
"grad_norm": 1.452721118927002,
|
| 2210 |
+
"learning_rate": 0.00033450518611221724,
|
| 2211 |
+
"loss": 1.1866,
|
| 2212 |
+
"step": 30700
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 1.752090562603106,
|
| 2216 |
+
"grad_norm": 1.5711089372634888,
|
| 2217 |
+
"learning_rate": 0.0003329882245861539,
|
| 2218 |
+
"loss": 1.1856,
|
| 2219 |
+
"step": 30800
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 1.7577791683258432,
|
| 2223 |
+
"grad_norm": 2.0584185123443604,
|
| 2224 |
+
"learning_rate": 0.00033147126306009065,
|
| 2225 |
+
"loss": 1.1873,
|
| 2226 |
+
"step": 30900
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 1.7634677740485807,
|
| 2230 |
+
"grad_norm": 1.5743275880813599,
|
| 2231 |
+
"learning_rate": 0.0003299543015340274,
|
| 2232 |
+
"loss": 1.1988,
|
| 2233 |
+
"step": 31000
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 1.7691563797713181,
|
| 2237 |
+
"grad_norm": 1.5788936614990234,
|
| 2238 |
+
"learning_rate": 0.00032843734000796406,
|
| 2239 |
+
"loss": 1.1932,
|
| 2240 |
+
"step": 31100
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 1.7748449854940556,
|
| 2244 |
+
"grad_norm": 1.6406651735305786,
|
| 2245 |
+
"learning_rate": 0.0003269203784819008,
|
| 2246 |
+
"loss": 1.1876,
|
| 2247 |
+
"step": 31200
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 1.7805335912167928,
|
| 2251 |
+
"grad_norm": 1.6410019397735596,
|
| 2252 |
+
"learning_rate": 0.00032540341695583747,
|
| 2253 |
+
"loss": 1.1859,
|
| 2254 |
+
"step": 31300
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 1.78622219693953,
|
| 2258 |
+
"grad_norm": 1.548140287399292,
|
| 2259 |
+
"learning_rate": 0.00032388645542977414,
|
| 2260 |
+
"loss": 1.2032,
|
| 2261 |
+
"step": 31400
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 1.7919108026622674,
|
| 2265 |
+
"grad_norm": 1.9242947101593018,
|
| 2266 |
+
"learning_rate": 0.0003223694939037109,
|
| 2267 |
+
"loss": 1.199,
|
| 2268 |
+
"step": 31500
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 1.7975994083850049,
|
| 2272 |
+
"grad_norm": 2.0189428329467773,
|
| 2273 |
+
"learning_rate": 0.0003208525323776476,
|
| 2274 |
+
"loss": 1.1832,
|
| 2275 |
+
"step": 31600
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 1.8032880141077423,
|
| 2279 |
+
"grad_norm": 1.740432620048523,
|
| 2280 |
+
"learning_rate": 0.0003193355708515843,
|
| 2281 |
+
"loss": 1.183,
|
| 2282 |
+
"step": 31700
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 1.8089766198304795,
|
| 2286 |
+
"grad_norm": 1.743503451347351,
|
| 2287 |
+
"learning_rate": 0.0003178337789407816,
|
| 2288 |
+
"loss": 1.1991,
|
| 2289 |
+
"step": 31800
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 1.8146652255532167,
|
| 2293 |
+
"grad_norm": 1.7166736125946045,
|
| 2294 |
+
"learning_rate": 0.00031631681741471835,
|
| 2295 |
+
"loss": 1.2031,
|
| 2296 |
+
"step": 31900
|
| 2297 |
+
},
|
| 2298 |
+
{
|
| 2299 |
+
"epoch": 1.8203538312759542,
|
| 2300 |
+
"grad_norm": 1.626386046409607,
|
| 2301 |
+
"learning_rate": 0.000314799855888655,
|
| 2302 |
+
"loss": 1.1987,
|
| 2303 |
+
"step": 32000
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 1.8260424369986916,
|
| 2307 |
+
"grad_norm": 1.5402131080627441,
|
| 2308 |
+
"learning_rate": 0.00031328289436259176,
|
| 2309 |
+
"loss": 1.172,
|
| 2310 |
+
"step": 32100
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 1.831731042721429,
|
| 2314 |
+
"grad_norm": 1.6522256135940552,
|
| 2315 |
+
"learning_rate": 0.0003117659328365285,
|
| 2316 |
+
"loss": 1.179,
|
| 2317 |
+
"step": 32200
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 1.8374196484441665,
|
| 2321 |
+
"grad_norm": 1.482009768486023,
|
| 2322 |
+
"learning_rate": 0.00031024897131046517,
|
| 2323 |
+
"loss": 1.1903,
|
| 2324 |
+
"step": 32300
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 1.8431082541669037,
|
| 2328 |
+
"grad_norm": 1.6417380571365356,
|
| 2329 |
+
"learning_rate": 0.00030873200978440184,
|
| 2330 |
+
"loss": 1.2024,
|
| 2331 |
+
"step": 32400
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 1.848796859889641,
|
| 2335 |
+
"grad_norm": 1.532333493232727,
|
| 2336 |
+
"learning_rate": 0.0003072150482583386,
|
| 2337 |
+
"loss": 1.1843,
|
| 2338 |
+
"step": 32500
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 1.8544854656123784,
|
| 2342 |
+
"grad_norm": 2.004293441772461,
|
| 2343 |
+
"learning_rate": 0.00030569808673227525,
|
| 2344 |
+
"loss": 1.192,
|
| 2345 |
+
"step": 32600
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 1.8601740713351158,
|
| 2349 |
+
"grad_norm": 1.7226125001907349,
|
| 2350 |
+
"learning_rate": 0.000304181125206212,
|
| 2351 |
+
"loss": 1.1902,
|
| 2352 |
+
"step": 32700
|
| 2353 |
+
},
|
| 2354 |
+
{
|
| 2355 |
+
"epoch": 1.8658626770578532,
|
| 2356 |
+
"grad_norm": 1.7714165449142456,
|
| 2357 |
+
"learning_rate": 0.0003026641636801487,
|
| 2358 |
+
"loss": 1.1908,
|
| 2359 |
+
"step": 32800
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 1.8715512827805905,
|
| 2363 |
+
"grad_norm": 1.5100337266921997,
|
| 2364 |
+
"learning_rate": 0.00030114720215408534,
|
| 2365 |
+
"loss": 1.1735,
|
| 2366 |
+
"step": 32900
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 1.8772398885033277,
|
| 2370 |
+
"grad_norm": 1.6792744398117065,
|
| 2371 |
+
"learning_rate": 0.00029963024062802207,
|
| 2372 |
+
"loss": 1.191,
|
| 2373 |
+
"step": 33000
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 1.8829284942260651,
|
| 2377 |
+
"grad_norm": 1.705554723739624,
|
| 2378 |
+
"learning_rate": 0.0002981132791019588,
|
| 2379 |
+
"loss": 1.1878,
|
| 2380 |
+
"step": 33100
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 1.8886170999488026,
|
| 2384 |
+
"grad_norm": 1.4528917074203491,
|
| 2385 |
+
"learning_rate": 0.0002965963175758955,
|
| 2386 |
+
"loss": 1.1685,
|
| 2387 |
+
"step": 33200
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 1.89430570567154,
|
| 2391 |
+
"grad_norm": 1.7752711772918701,
|
| 2392 |
+
"learning_rate": 0.0002950793560498322,
|
| 2393 |
+
"loss": 1.1743,
|
| 2394 |
+
"step": 33300
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 1.8999943113942772,
|
| 2398 |
+
"grad_norm": 1.762074589729309,
|
| 2399 |
+
"learning_rate": 0.00029356239452376894,
|
| 2400 |
+
"loss": 1.1775,
|
| 2401 |
+
"step": 33400
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 1.9056829171170147,
|
| 2405 |
+
"grad_norm": 1.6388828754425049,
|
| 2406 |
+
"learning_rate": 0.0002920454329977056,
|
| 2407 |
+
"loss": 1.1762,
|
| 2408 |
+
"step": 33500
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 1.9113715228397519,
|
| 2412 |
+
"grad_norm": 1.5171791315078735,
|
| 2413 |
+
"learning_rate": 0.0002905284714716423,
|
| 2414 |
+
"loss": 1.1649,
|
| 2415 |
+
"step": 33600
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 1.9170601285624893,
|
| 2419 |
+
"grad_norm": 1.6547460556030273,
|
| 2420 |
+
"learning_rate": 0.000289011509945579,
|
| 2421 |
+
"loss": 1.1904,
|
| 2422 |
+
"step": 33700
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 1.9227487342852267,
|
| 2426 |
+
"grad_norm": 1.705083966255188,
|
| 2427 |
+
"learning_rate": 0.00028750971803477636,
|
| 2428 |
+
"loss": 1.1667,
|
| 2429 |
+
"step": 33800
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 1.9284373400079642,
|
| 2433 |
+
"grad_norm": 1.731803059577942,
|
| 2434 |
+
"learning_rate": 0.00028599275650871304,
|
| 2435 |
+
"loss": 1.1788,
|
| 2436 |
+
"step": 33900
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 1.9341259457307014,
|
| 2440 |
+
"grad_norm": 2.056766986846924,
|
| 2441 |
+
"learning_rate": 0.00028447579498264977,
|
| 2442 |
+
"loss": 1.1878,
|
| 2443 |
+
"step": 34000
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 1.9398145514534386,
|
| 2447 |
+
"grad_norm": 1.8016914129257202,
|
| 2448 |
+
"learning_rate": 0.00028295883345658644,
|
| 2449 |
+
"loss": 1.1632,
|
| 2450 |
+
"step": 34100
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 1.945503157176176,
|
| 2454 |
+
"grad_norm": 1.7706475257873535,
|
| 2455 |
+
"learning_rate": 0.0002814418719305232,
|
| 2456 |
+
"loss": 1.1658,
|
| 2457 |
+
"step": 34200
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 1.9511917628989135,
|
| 2461 |
+
"grad_norm": 1.8184970617294312,
|
| 2462 |
+
"learning_rate": 0.0002799249104044599,
|
| 2463 |
+
"loss": 1.1666,
|
| 2464 |
+
"step": 34300
|
| 2465 |
+
},
|
| 2466 |
+
{
|
| 2467 |
+
"epoch": 1.956880368621651,
|
| 2468 |
+
"grad_norm": 1.6529743671417236,
|
| 2469 |
+
"learning_rate": 0.0002784079488783966,
|
| 2470 |
+
"loss": 1.1846,
|
| 2471 |
+
"step": 34400
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 1.9625689743443882,
|
| 2475 |
+
"grad_norm": 1.5860931873321533,
|
| 2476 |
+
"learning_rate": 0.00027689098735233326,
|
| 2477 |
+
"loss": 1.1917,
|
| 2478 |
+
"step": 34500
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 1.9682575800671256,
|
| 2482 |
+
"grad_norm": 1.672756552696228,
|
| 2483 |
+
"learning_rate": 0.00027537402582627,
|
| 2484 |
+
"loss": 1.1654,
|
| 2485 |
+
"step": 34600
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 1.9739461857898628,
|
| 2489 |
+
"grad_norm": 1.7606583833694458,
|
| 2490 |
+
"learning_rate": 0.0002738570643002067,
|
| 2491 |
+
"loss": 1.176,
|
| 2492 |
+
"step": 34700
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 1.9796347915126002,
|
| 2496 |
+
"grad_norm": 1.912277340888977,
|
| 2497 |
+
"learning_rate": 0.0002723401027741434,
|
| 2498 |
+
"loss": 1.1695,
|
| 2499 |
+
"step": 34800
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 1.9853233972353377,
|
| 2503 |
+
"grad_norm": 1.7096484899520874,
|
| 2504 |
+
"learning_rate": 0.00027082314124808013,
|
| 2505 |
+
"loss": 1.1669,
|
| 2506 |
+
"step": 34900
|
| 2507 |
+
},
|
| 2508 |
+
{
|
| 2509 |
+
"epoch": 1.9910120029580751,
|
| 2510 |
+
"grad_norm": 1.7793241739273071,
|
| 2511 |
+
"learning_rate": 0.0002693061797220168,
|
| 2512 |
+
"loss": 1.1902,
|
| 2513 |
+
"step": 35000
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 1.9910120029580751,
|
| 2517 |
+
"eval_accuracy": 0.71648,
|
| 2518 |
+
"eval_loss": 1.1297262907028198,
|
| 2519 |
+
"eval_runtime": 79.5966,
|
| 2520 |
+
"eval_samples_per_second": 3140.839,
|
| 2521 |
+
"eval_steps_per_second": 12.274,
|
| 2522 |
+
"step": 35000
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 1.9967006086808123,
|
| 2526 |
+
"grad_norm": 1.4913907051086426,
|
| 2527 |
+
"learning_rate": 0.0002677892181959535,
|
| 2528 |
+
"loss": 1.1607,
|
| 2529 |
+
"step": 35100
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 2.0023892144035496,
|
| 2533 |
+
"grad_norm": 1.639985203742981,
|
| 2534 |
+
"learning_rate": 0.0002662874262851509,
|
| 2535 |
+
"loss": 1.1727,
|
| 2536 |
+
"step": 35200
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 2.008077820126287,
|
| 2540 |
+
"grad_norm": 1.6419970989227295,
|
| 2541 |
+
"learning_rate": 0.00026477046475908755,
|
| 2542 |
+
"loss": 1.1392,
|
| 2543 |
+
"step": 35300
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.0137664258490244,
|
| 2547 |
+
"grad_norm": 1.8132672309875488,
|
| 2548 |
+
"learning_rate": 0.00026325350323302423,
|
| 2549 |
+
"loss": 1.1503,
|
| 2550 |
+
"step": 35400
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 2.019455031571762,
|
| 2554 |
+
"grad_norm": 1.4656819105148315,
|
| 2555 |
+
"learning_rate": 0.00026173654170696096,
|
| 2556 |
+
"loss": 1.1565,
|
| 2557 |
+
"step": 35500
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 2.0251436372944993,
|
| 2561 |
+
"grad_norm": 1.3595716953277588,
|
| 2562 |
+
"learning_rate": 0.0002602195801808977,
|
| 2563 |
+
"loss": 1.1526,
|
| 2564 |
+
"step": 35600
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 2.0308322430172363,
|
| 2568 |
+
"grad_norm": 1.6904360055923462,
|
| 2569 |
+
"learning_rate": 0.00025870261865483437,
|
| 2570 |
+
"loss": 1.1448,
|
| 2571 |
+
"step": 35700
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 2.0365208487399737,
|
| 2575 |
+
"grad_norm": 1.7240209579467773,
|
| 2576 |
+
"learning_rate": 0.0002571856571287711,
|
| 2577 |
+
"loss": 1.1424,
|
| 2578 |
+
"step": 35800
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 2.042209454462711,
|
| 2582 |
+
"grad_norm": 1.5376731157302856,
|
| 2583 |
+
"learning_rate": 0.00025566869560270783,
|
| 2584 |
+
"loss": 1.143,
|
| 2585 |
+
"step": 35900
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 2.0478980601854486,
|
| 2589 |
+
"grad_norm": 1.893202781677246,
|
| 2590 |
+
"learning_rate": 0.00025415173407664445,
|
| 2591 |
+
"loss": 1.1519,
|
| 2592 |
+
"step": 36000
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 2.053586665908186,
|
| 2596 |
+
"grad_norm": 1.9057211875915527,
|
| 2597 |
+
"learning_rate": 0.0002526347725505812,
|
| 2598 |
+
"loss": 1.1375,
|
| 2599 |
+
"step": 36100
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 2.059275271630923,
|
| 2603 |
+
"grad_norm": 1.7818187475204468,
|
| 2604 |
+
"learning_rate": 0.0002511178110245179,
|
| 2605 |
+
"loss": 1.1424,
|
| 2606 |
+
"step": 36200
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 2.0649638773536605,
|
| 2610 |
+
"grad_norm": 1.825323462486267,
|
| 2611 |
+
"learning_rate": 0.0002496008494984546,
|
| 2612 |
+
"loss": 1.1196,
|
| 2613 |
+
"step": 36300
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 2.070652483076398,
|
| 2617 |
+
"grad_norm": 2.0049736499786377,
|
| 2618 |
+
"learning_rate": 0.0002480838879723913,
|
| 2619 |
+
"loss": 1.1317,
|
| 2620 |
+
"step": 36400
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 2.0763410887991354,
|
| 2624 |
+
"grad_norm": 1.599846363067627,
|
| 2625 |
+
"learning_rate": 0.000246566926446328,
|
| 2626 |
+
"loss": 1.1573,
|
| 2627 |
+
"step": 36500
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 2.082029694521873,
|
| 2631 |
+
"grad_norm": 1.5434855222702026,
|
| 2632 |
+
"learning_rate": 0.00024504996492026473,
|
| 2633 |
+
"loss": 1.1493,
|
| 2634 |
+
"step": 36600
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 2.0877183002446102,
|
| 2638 |
+
"grad_norm": 1.6306787729263306,
|
| 2639 |
+
"learning_rate": 0.0002435330033942014,
|
| 2640 |
+
"loss": 1.1564,
|
| 2641 |
+
"step": 36700
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 2.0934069059673472,
|
| 2645 |
+
"grad_norm": 1.6914353370666504,
|
| 2646 |
+
"learning_rate": 0.00024201604186813814,
|
| 2647 |
+
"loss": 1.1395,
|
| 2648 |
+
"step": 36800
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 2.0990955116900847,
|
| 2652 |
+
"grad_norm": 1.6444432735443115,
|
| 2653 |
+
"learning_rate": 0.00024049908034207485,
|
| 2654 |
+
"loss": 1.1615,
|
| 2655 |
+
"step": 36900
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 2.104784117412822,
|
| 2659 |
+
"grad_norm": 1.821244239807129,
|
| 2660 |
+
"learning_rate": 0.00023898211881601155,
|
| 2661 |
+
"loss": 1.1429,
|
| 2662 |
+
"step": 37000
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 2.1104727231355596,
|
| 2666 |
+
"grad_norm": 1.6050491333007812,
|
| 2667 |
+
"learning_rate": 0.00023746515728994823,
|
| 2668 |
+
"loss": 1.1376,
|
| 2669 |
+
"step": 37100
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 2.116161328858297,
|
| 2673 |
+
"grad_norm": 1.7375249862670898,
|
| 2674 |
+
"learning_rate": 0.00023594819576388493,
|
| 2675 |
+
"loss": 1.1253,
|
| 2676 |
+
"step": 37200
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 2.121849934581034,
|
| 2680 |
+
"grad_norm": 2.0717177391052246,
|
| 2681 |
+
"learning_rate": 0.00023443123423782166,
|
| 2682 |
+
"loss": 1.1527,
|
| 2683 |
+
"step": 37300
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 2.1275385403037714,
|
| 2687 |
+
"grad_norm": 1.43324875831604,
|
| 2688 |
+
"learning_rate": 0.00023291427271175837,
|
| 2689 |
+
"loss": 1.1475,
|
| 2690 |
+
"step": 37400
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 2.133227146026509,
|
| 2694 |
+
"grad_norm": 1.448669195175171,
|
| 2695 |
+
"learning_rate": 0.00023139731118569507,
|
| 2696 |
+
"loss": 1.1133,
|
| 2697 |
+
"step": 37500
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 2.1389157517492463,
|
| 2701 |
+
"grad_norm": 1.521912932395935,
|
| 2702 |
+
"learning_rate": 0.00022988034965963178,
|
| 2703 |
+
"loss": 1.1292,
|
| 2704 |
+
"step": 37600
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 2.1446043574719837,
|
| 2708 |
+
"grad_norm": 1.6070728302001953,
|
| 2709 |
+
"learning_rate": 0.00022836338813356845,
|
| 2710 |
+
"loss": 1.1384,
|
| 2711 |
+
"step": 37700
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 2.1502929631947207,
|
| 2715 |
+
"grad_norm": 1.3853884935379028,
|
| 2716 |
+
"learning_rate": 0.00022684642660750516,
|
| 2717 |
+
"loss": 1.1344,
|
| 2718 |
+
"step": 37800
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 2.155981568917458,
|
| 2722 |
+
"grad_norm": 1.569415807723999,
|
| 2723 |
+
"learning_rate": 0.0002253294650814419,
|
| 2724 |
+
"loss": 1.1572,
|
| 2725 |
+
"step": 37900
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 2.1616701746401956,
|
| 2729 |
+
"grad_norm": 1.544966220855713,
|
| 2730 |
+
"learning_rate": 0.0002238125035553786,
|
| 2731 |
+
"loss": 1.1378,
|
| 2732 |
+
"step": 38000
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 2.167358780362933,
|
| 2736 |
+
"grad_norm": 1.6090420484542847,
|
| 2737 |
+
"learning_rate": 0.0002222955420293153,
|
| 2738 |
+
"loss": 1.1331,
|
| 2739 |
+
"step": 38100
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 2.1730473860856705,
|
| 2743 |
+
"grad_norm": 1.542605996131897,
|
| 2744 |
+
"learning_rate": 0.00022077858050325197,
|
| 2745 |
+
"loss": 1.1302,
|
| 2746 |
+
"step": 38200
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.178735991808408,
|
| 2750 |
+
"grad_norm": 1.744084119796753,
|
| 2751 |
+
"learning_rate": 0.00021926161897718868,
|
| 2752 |
+
"loss": 1.1305,
|
| 2753 |
+
"step": 38300
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 2.184424597531145,
|
| 2757 |
+
"grad_norm": 1.630118489265442,
|
| 2758 |
+
"learning_rate": 0.0002177446574511254,
|
| 2759 |
+
"loss": 1.1294,
|
| 2760 |
+
"step": 38400
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 2.1901132032538824,
|
| 2764 |
+
"grad_norm": 1.6920104026794434,
|
| 2765 |
+
"learning_rate": 0.0002162276959250621,
|
| 2766 |
+
"loss": 1.1337,
|
| 2767 |
+
"step": 38500
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 2.19580180897662,
|
| 2771 |
+
"grad_norm": 1.654189944267273,
|
| 2772 |
+
"learning_rate": 0.00021471073439899882,
|
| 2773 |
+
"loss": 1.1182,
|
| 2774 |
+
"step": 38600
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 2.2014904146993572,
|
| 2778 |
+
"grad_norm": 1.8575996160507202,
|
| 2779 |
+
"learning_rate": 0.00021319377287293555,
|
| 2780 |
+
"loss": 1.1261,
|
| 2781 |
+
"step": 38700
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.2071790204220947,
|
| 2785 |
+
"grad_norm": 1.5796535015106201,
|
| 2786 |
+
"learning_rate": 0.0002116768113468722,
|
| 2787 |
+
"loss": 1.1389,
|
| 2788 |
+
"step": 38800
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 2.212867626144832,
|
| 2792 |
+
"grad_norm": 1.6893657445907593,
|
| 2793 |
+
"learning_rate": 0.00021015984982080893,
|
| 2794 |
+
"loss": 1.122,
|
| 2795 |
+
"step": 38900
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.218556231867569,
|
| 2799 |
+
"grad_norm": 1.5983092784881592,
|
| 2800 |
+
"learning_rate": 0.00020864288829474563,
|
| 2801 |
+
"loss": 1.1487,
|
| 2802 |
+
"step": 39000
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 2.2242448375903066,
|
| 2806 |
+
"grad_norm": 1.632049798965454,
|
| 2807 |
+
"learning_rate": 0.00020712592676868234,
|
| 2808 |
+
"loss": 1.1476,
|
| 2809 |
+
"step": 39100
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 2.229933443313044,
|
| 2813 |
+
"grad_norm": 2.039854049682617,
|
| 2814 |
+
"learning_rate": 0.00020562413485787965,
|
| 2815 |
+
"loss": 1.1443,
|
| 2816 |
+
"step": 39200
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 2.2356220490357814,
|
| 2820 |
+
"grad_norm": 1.5673627853393555,
|
| 2821 |
+
"learning_rate": 0.00020410717333181638,
|
| 2822 |
+
"loss": 1.1259,
|
| 2823 |
+
"step": 39300
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 2.241310654758519,
|
| 2827 |
+
"grad_norm": 1.6900497674942017,
|
| 2828 |
+
"learning_rate": 0.00020259021180575308,
|
| 2829 |
+
"loss": 1.1356,
|
| 2830 |
+
"step": 39400
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 2.246999260481256,
|
| 2834 |
+
"grad_norm": 1.8306878805160522,
|
| 2835 |
+
"learning_rate": 0.00020107325027968979,
|
| 2836 |
+
"loss": 1.1349,
|
| 2837 |
+
"step": 39500
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.2526878662039933,
|
| 2841 |
+
"grad_norm": 1.620490550994873,
|
| 2842 |
+
"learning_rate": 0.0001995562887536265,
|
| 2843 |
+
"loss": 1.1417,
|
| 2844 |
+
"step": 39600
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 2.2583764719267307,
|
| 2848 |
+
"grad_norm": 1.828751802444458,
|
| 2849 |
+
"learning_rate": 0.0001980393272275632,
|
| 2850 |
+
"loss": 1.1302,
|
| 2851 |
+
"step": 39700
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 2.264065077649468,
|
| 2855 |
+
"grad_norm": 1.4963942766189575,
|
| 2856 |
+
"learning_rate": 0.0001965223657014999,
|
| 2857 |
+
"loss": 1.152,
|
| 2858 |
+
"step": 39800
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 2.2697536833722056,
|
| 2862 |
+
"grad_norm": 2.081669807434082,
|
| 2863 |
+
"learning_rate": 0.0001950054041754366,
|
| 2864 |
+
"loss": 1.1385,
|
| 2865 |
+
"step": 39900
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 2.2754422890949426,
|
| 2869 |
+
"grad_norm": 1.6873656511306763,
|
| 2870 |
+
"learning_rate": 0.0001934884426493733,
|
| 2871 |
+
"loss": 1.131,
|
| 2872 |
+
"step": 40000
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 2.2754422890949426,
|
| 2876 |
+
"eval_accuracy": 0.721316,
|
| 2877 |
+
"eval_loss": 1.1105972528457642,
|
| 2878 |
+
"eval_runtime": 80.6985,
|
| 2879 |
+
"eval_samples_per_second": 3097.949,
|
| 2880 |
+
"eval_steps_per_second": 12.107,
|
| 2881 |
+
"step": 40000
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 2.28113089481768,
|
| 2885 |
+
"grad_norm": 1.5599457025527954,
|
| 2886 |
+
"learning_rate": 0.00019197148112331004,
|
| 2887 |
+
"loss": 1.1525,
|
| 2888 |
+
"step": 40100
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 2.2868195005404175,
|
| 2892 |
+
"grad_norm": 1.816628098487854,
|
| 2893 |
+
"learning_rate": 0.00019045451959724672,
|
| 2894 |
+
"loss": 1.1295,
|
| 2895 |
+
"step": 40200
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 2.292508106263155,
|
| 2899 |
+
"grad_norm": 1.5481749773025513,
|
| 2900 |
+
"learning_rate": 0.00018893755807118342,
|
| 2901 |
+
"loss": 1.124,
|
| 2902 |
+
"step": 40300
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 2.2981967119858924,
|
| 2906 |
+
"grad_norm": 1.632873296737671,
|
| 2907 |
+
"learning_rate": 0.00018742059654512015,
|
| 2908 |
+
"loss": 1.1217,
|
| 2909 |
+
"step": 40400
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.3038853177086294,
|
| 2913 |
+
"grad_norm": 1.4403363466262817,
|
| 2914 |
+
"learning_rate": 0.00018590363501905683,
|
| 2915 |
+
"loss": 1.1315,
|
| 2916 |
+
"step": 40500
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 2.309573923431367,
|
| 2920 |
+
"grad_norm": 1.6744205951690674,
|
| 2921 |
+
"learning_rate": 0.00018438667349299353,
|
| 2922 |
+
"loss": 1.1473,
|
| 2923 |
+
"step": 40600
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 2.3152625291541042,
|
| 2927 |
+
"grad_norm": 1.5021002292633057,
|
| 2928 |
+
"learning_rate": 0.00018286971196693026,
|
| 2929 |
+
"loss": 1.1127,
|
| 2930 |
+
"step": 40700
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.3209511348768417,
|
| 2934 |
+
"grad_norm": 1.689931869506836,
|
| 2935 |
+
"learning_rate": 0.00018135275044086694,
|
| 2936 |
+
"loss": 1.1394,
|
| 2937 |
+
"step": 40800
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 2.326639740599579,
|
| 2941 |
+
"grad_norm": 2.1370577812194824,
|
| 2942 |
+
"learning_rate": 0.00017983578891480367,
|
| 2943 |
+
"loss": 1.148,
|
| 2944 |
+
"step": 40900
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 2.3323283463223166,
|
| 2948 |
+
"grad_norm": 1.9048566818237305,
|
| 2949 |
+
"learning_rate": 0.00017831882738874038,
|
| 2950 |
+
"loss": 1.1181,
|
| 2951 |
+
"step": 41000
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.338016952045054,
|
| 2955 |
+
"grad_norm": 1.8328748941421509,
|
| 2956 |
+
"learning_rate": 0.00017680186586267705,
|
| 2957 |
+
"loss": 1.1302,
|
| 2958 |
+
"step": 41100
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 2.343705557767791,
|
| 2962 |
+
"grad_norm": 1.7709869146347046,
|
| 2963 |
+
"learning_rate": 0.0001753000739518744,
|
| 2964 |
+
"loss": 1.1369,
|
| 2965 |
+
"step": 41200
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 2.3493941634905284,
|
| 2969 |
+
"grad_norm": 1.6296570301055908,
|
| 2970 |
+
"learning_rate": 0.00017378311242581112,
|
| 2971 |
+
"loss": 1.1302,
|
| 2972 |
+
"step": 41300
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 2.355082769213266,
|
| 2976 |
+
"grad_norm": 1.6044236421585083,
|
| 2977 |
+
"learning_rate": 0.0001722661508997478,
|
| 2978 |
+
"loss": 1.1313,
|
| 2979 |
+
"step": 41400
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 2.3607713749360033,
|
| 2983 |
+
"grad_norm": 1.4571659564971924,
|
| 2984 |
+
"learning_rate": 0.00017074918937368453,
|
| 2985 |
+
"loss": 1.1249,
|
| 2986 |
+
"step": 41500
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.3664599806587407,
|
| 2990 |
+
"grad_norm": 1.7237457036972046,
|
| 2991 |
+
"learning_rate": 0.00016923222784762123,
|
| 2992 |
+
"loss": 1.1312,
|
| 2993 |
+
"step": 41600
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.3721485863814777,
|
| 2997 |
+
"grad_norm": 1.552881121635437,
|
| 2998 |
+
"learning_rate": 0.0001677152663215579,
|
| 2999 |
+
"loss": 1.1282,
|
| 3000 |
+
"step": 41700
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 2.377837192104215,
|
| 3004 |
+
"grad_norm": 1.6091784238815308,
|
| 3005 |
+
"learning_rate": 0.00016619830479549464,
|
| 3006 |
+
"loss": 1.1236,
|
| 3007 |
+
"step": 41800
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 2.3835257978269526,
|
| 3011 |
+
"grad_norm": 1.8620885610580444,
|
| 3012 |
+
"learning_rate": 0.00016468134326943134,
|
| 3013 |
+
"loss": 1.1469,
|
| 3014 |
+
"step": 41900
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 2.38921440354969,
|
| 3018 |
+
"grad_norm": 1.717551827430725,
|
| 3019 |
+
"learning_rate": 0.00016316438174336802,
|
| 3020 |
+
"loss": 1.121,
|
| 3021 |
+
"step": 42000
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 2.3949030092724275,
|
| 3025 |
+
"grad_norm": 1.6212184429168701,
|
| 3026 |
+
"learning_rate": 0.00016164742021730475,
|
| 3027 |
+
"loss": 1.0997,
|
| 3028 |
+
"step": 42100
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 2.4005916149951645,
|
| 3032 |
+
"grad_norm": 1.3878498077392578,
|
| 3033 |
+
"learning_rate": 0.00016013045869124146,
|
| 3034 |
+
"loss": 1.1362,
|
| 3035 |
+
"step": 42200
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.406280220717902,
|
| 3039 |
+
"grad_norm": 1.6336196660995483,
|
| 3040 |
+
"learning_rate": 0.00015861349716517816,
|
| 3041 |
+
"loss": 1.1256,
|
| 3042 |
+
"step": 42300
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.4119688264406394,
|
| 3046 |
+
"grad_norm": 1.7155201435089111,
|
| 3047 |
+
"learning_rate": 0.00015709653563911486,
|
| 3048 |
+
"loss": 1.1133,
|
| 3049 |
+
"step": 42400
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.417657432163377,
|
| 3053 |
+
"grad_norm": 1.7675564289093018,
|
| 3054 |
+
"learning_rate": 0.00015557957411305157,
|
| 3055 |
+
"loss": 1.1416,
|
| 3056 |
+
"step": 42500
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.4233460378861142,
|
| 3060 |
+
"grad_norm": 1.676527976989746,
|
| 3061 |
+
"learning_rate": 0.00015406261258698827,
|
| 3062 |
+
"loss": 1.1378,
|
| 3063 |
+
"step": 42600
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 2.4290346436088512,
|
| 3067 |
+
"grad_norm": 1.6293052434921265,
|
| 3068 |
+
"learning_rate": 0.00015254565106092498,
|
| 3069 |
+
"loss": 1.1177,
|
| 3070 |
+
"step": 42700
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.4347232493315887,
|
| 3074 |
+
"grad_norm": 1.5264780521392822,
|
| 3075 |
+
"learning_rate": 0.00015102868953486168,
|
| 3076 |
+
"loss": 1.1063,
|
| 3077 |
+
"step": 42800
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 2.440411855054326,
|
| 3081 |
+
"grad_norm": 1.6453486680984497,
|
| 3082 |
+
"learning_rate": 0.00014951172800879839,
|
| 3083 |
+
"loss": 1.1375,
|
| 3084 |
+
"step": 42900
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 2.4461004607770636,
|
| 3088 |
+
"grad_norm": 1.692336082458496,
|
| 3089 |
+
"learning_rate": 0.0001479947664827351,
|
| 3090 |
+
"loss": 1.1004,
|
| 3091 |
+
"step": 43000
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 2.451789066499801,
|
| 3095 |
+
"grad_norm": 1.868812084197998,
|
| 3096 |
+
"learning_rate": 0.0001464778049566718,
|
| 3097 |
+
"loss": 1.1288,
|
| 3098 |
+
"step": 43100
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.4574776722225384,
|
| 3102 |
+
"grad_norm": 1.7713991403579712,
|
| 3103 |
+
"learning_rate": 0.0001449608434306085,
|
| 3104 |
+
"loss": 1.1229,
|
| 3105 |
+
"step": 43200
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 2.4631662779452754,
|
| 3109 |
+
"grad_norm": 1.6394290924072266,
|
| 3110 |
+
"learning_rate": 0.00014345905151980583,
|
| 3111 |
+
"loss": 1.0968,
|
| 3112 |
+
"step": 43300
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.468854883668013,
|
| 3116 |
+
"grad_norm": 1.7240723371505737,
|
| 3117 |
+
"learning_rate": 0.00014194208999374254,
|
| 3118 |
+
"loss": 1.1151,
|
| 3119 |
+
"step": 43400
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 2.4745434893907503,
|
| 3123 |
+
"grad_norm": 1.9284464120864868,
|
| 3124 |
+
"learning_rate": 0.00014042512846767924,
|
| 3125 |
+
"loss": 1.1302,
|
| 3126 |
+
"step": 43500
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 2.4802320951134877,
|
| 3130 |
+
"grad_norm": 1.6855792999267578,
|
| 3131 |
+
"learning_rate": 0.00013890816694161595,
|
| 3132 |
+
"loss": 1.1163,
|
| 3133 |
+
"step": 43600
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 2.485920700836225,
|
| 3137 |
+
"grad_norm": 1.8182587623596191,
|
| 3138 |
+
"learning_rate": 0.00013739120541555265,
|
| 3139 |
+
"loss": 1.1172,
|
| 3140 |
+
"step": 43700
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 2.4916093065589626,
|
| 3144 |
+
"grad_norm": 1.5971157550811768,
|
| 3145 |
+
"learning_rate": 0.00013587424388948935,
|
| 3146 |
+
"loss": 1.1071,
|
| 3147 |
+
"step": 43800
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 2.4972979122816996,
|
| 3151 |
+
"grad_norm": 1.7139756679534912,
|
| 3152 |
+
"learning_rate": 0.00013435728236342606,
|
| 3153 |
+
"loss": 1.1239,
|
| 3154 |
+
"step": 43900
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.502986518004437,
|
| 3158 |
+
"grad_norm": 1.7199363708496094,
|
| 3159 |
+
"learning_rate": 0.00013284032083736276,
|
| 3160 |
+
"loss": 1.1444,
|
| 3161 |
+
"step": 44000
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 2.5086751237271745,
|
| 3165 |
+
"grad_norm": 1.7295994758605957,
|
| 3166 |
+
"learning_rate": 0.00013132335931129947,
|
| 3167 |
+
"loss": 1.122,
|
| 3168 |
+
"step": 44100
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 2.514363729449912,
|
| 3172 |
+
"grad_norm": 1.9433492422103882,
|
| 3173 |
+
"learning_rate": 0.00012980639778523617,
|
| 3174 |
+
"loss": 1.1209,
|
| 3175 |
+
"step": 44200
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 2.5200523351726494,
|
| 3179 |
+
"grad_norm": 1.5811411142349243,
|
| 3180 |
+
"learning_rate": 0.0001282894362591729,
|
| 3181 |
+
"loss": 1.1084,
|
| 3182 |
+
"step": 44300
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 2.5257409408953864,
|
| 3186 |
+
"grad_norm": 1.5232020616531372,
|
| 3187 |
+
"learning_rate": 0.00012677247473310958,
|
| 3188 |
+
"loss": 1.1372,
|
| 3189 |
+
"step": 44400
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 2.531429546618124,
|
| 3193 |
+
"grad_norm": 2.6212551593780518,
|
| 3194 |
+
"learning_rate": 0.00012525551320704628,
|
| 3195 |
+
"loss": 1.1246,
|
| 3196 |
+
"step": 44500
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.5371181523408612,
|
| 3200 |
+
"grad_norm": 1.4962718486785889,
|
| 3201 |
+
"learning_rate": 0.00012373855168098301,
|
| 3202 |
+
"loss": 1.1386,
|
| 3203 |
+
"step": 44600
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 2.5428067580635987,
|
| 3207 |
+
"grad_norm": 1.7713087797164917,
|
| 3208 |
+
"learning_rate": 0.0001222215901549197,
|
| 3209 |
+
"loss": 1.1314,
|
| 3210 |
+
"step": 44700
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.548495363786336,
|
| 3214 |
+
"grad_norm": 1.5493218898773193,
|
| 3215 |
+
"learning_rate": 0.00012070462862885641,
|
| 3216 |
+
"loss": 1.1204,
|
| 3217 |
+
"step": 44800
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 2.554183969509073,
|
| 3221 |
+
"grad_norm": 1.6126313209533691,
|
| 3222 |
+
"learning_rate": 0.00011918766710279313,
|
| 3223 |
+
"loss": 1.1283,
|
| 3224 |
+
"step": 44900
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 2.5598725752318106,
|
| 3228 |
+
"grad_norm": 1.5327433347702026,
|
| 3229 |
+
"learning_rate": 0.00011767070557672982,
|
| 3230 |
+
"loss": 1.124,
|
| 3231 |
+
"step": 45000
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 2.5598725752318106,
|
| 3235 |
+
"eval_accuracy": 0.725824,
|
| 3236 |
+
"eval_loss": 1.0916061401367188,
|
| 3237 |
+
"eval_runtime": 80.2149,
|
| 3238 |
+
"eval_samples_per_second": 3116.626,
|
| 3239 |
+
"eval_steps_per_second": 12.18,
|
| 3240 |
+
"step": 45000
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 2.565561180954548,
|
| 3244 |
+
"grad_norm": 1.5026576519012451,
|
| 3245 |
+
"learning_rate": 0.00011615374405066652,
|
| 3246 |
+
"loss": 1.1197,
|
| 3247 |
+
"step": 45100
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 2.5712497866772854,
|
| 3251 |
+
"grad_norm": 1.6989002227783203,
|
| 3252 |
+
"learning_rate": 0.00011463678252460321,
|
| 3253 |
+
"loss": 1.1247,
|
| 3254 |
+
"step": 45200
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 2.576938392400023,
|
| 3258 |
+
"grad_norm": 1.5901920795440674,
|
| 3259 |
+
"learning_rate": 0.00011313499061380057,
|
| 3260 |
+
"loss": 1.1352,
|
| 3261 |
+
"step": 45300
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 2.58262699812276,
|
| 3265 |
+
"grad_norm": 1.4382330179214478,
|
| 3266 |
+
"learning_rate": 0.00011161802908773727,
|
| 3267 |
+
"loss": 1.1093,
|
| 3268 |
+
"step": 45400
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 2.5883156038454973,
|
| 3272 |
+
"grad_norm": 1.8520530462265015,
|
| 3273 |
+
"learning_rate": 0.00011010106756167397,
|
| 3274 |
+
"loss": 1.1081,
|
| 3275 |
+
"step": 45500
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 2.5940042095682347,
|
| 3279 |
+
"grad_norm": 1.8772435188293457,
|
| 3280 |
+
"learning_rate": 0.00010858410603561066,
|
| 3281 |
+
"loss": 1.1157,
|
| 3282 |
+
"step": 45600
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 2.599692815290972,
|
| 3286 |
+
"grad_norm": 1.6013365983963013,
|
| 3287 |
+
"learning_rate": 0.00010706714450954738,
|
| 3288 |
+
"loss": 1.1463,
|
| 3289 |
+
"step": 45700
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 2.6053814210137096,
|
| 3293 |
+
"grad_norm": 1.582515835762024,
|
| 3294 |
+
"learning_rate": 0.0001055501829834841,
|
| 3295 |
+
"loss": 1.1135,
|
| 3296 |
+
"step": 45800
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 2.611070026736447,
|
| 3300 |
+
"grad_norm": 1.3782535791397095,
|
| 3301 |
+
"learning_rate": 0.00010403322145742079,
|
| 3302 |
+
"loss": 1.1101,
|
| 3303 |
+
"step": 45900
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 2.6167586324591845,
|
| 3307 |
+
"grad_norm": 1.465584397315979,
|
| 3308 |
+
"learning_rate": 0.00010251625993135749,
|
| 3309 |
+
"loss": 1.1354,
|
| 3310 |
+
"step": 46000
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 2.6224472381819215,
|
| 3314 |
+
"grad_norm": 1.4038536548614502,
|
| 3315 |
+
"learning_rate": 0.00010099929840529421,
|
| 3316 |
+
"loss": 1.1078,
|
| 3317 |
+
"step": 46100
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 2.628135843904659,
|
| 3321 |
+
"grad_norm": 1.9926286935806274,
|
| 3322 |
+
"learning_rate": 9.948233687923091e-05,
|
| 3323 |
+
"loss": 1.1044,
|
| 3324 |
+
"step": 46200
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 2.6338244496273964,
|
| 3328 |
+
"grad_norm": 1.6215740442276,
|
| 3329 |
+
"learning_rate": 9.796537535316762e-05,
|
| 3330 |
+
"loss": 1.1188,
|
| 3331 |
+
"step": 46300
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 2.639513055350134,
|
| 3335 |
+
"grad_norm": 1.5623165369033813,
|
| 3336 |
+
"learning_rate": 9.644841382710431e-05,
|
| 3337 |
+
"loss": 1.1155,
|
| 3338 |
+
"step": 46400
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 2.6452016610728712,
|
| 3342 |
+
"grad_norm": 1.491926670074463,
|
| 3343 |
+
"learning_rate": 9.493145230104102e-05,
|
| 3344 |
+
"loss": 1.1211,
|
| 3345 |
+
"step": 46500
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 2.6508902667956082,
|
| 3349 |
+
"grad_norm": 1.7084381580352783,
|
| 3350 |
+
"learning_rate": 9.341449077497773e-05,
|
| 3351 |
+
"loss": 1.1091,
|
| 3352 |
+
"step": 46600
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 2.6565788725183457,
|
| 3356 |
+
"grad_norm": 1.5060371160507202,
|
| 3357 |
+
"learning_rate": 9.189752924891443e-05,
|
| 3358 |
+
"loss": 1.1198,
|
| 3359 |
+
"step": 46700
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 2.662267478241083,
|
| 3363 |
+
"grad_norm": 1.7321504354476929,
|
| 3364 |
+
"learning_rate": 9.038056772285112e-05,
|
| 3365 |
+
"loss": 1.1157,
|
| 3366 |
+
"step": 46800
|
| 3367 |
+
},
|
| 3368 |
+
{
|
| 3369 |
+
"epoch": 2.6679560839638206,
|
| 3370 |
+
"grad_norm": 1.559877634048462,
|
| 3371 |
+
"learning_rate": 8.886360619678784e-05,
|
| 3372 |
+
"loss": 1.1035,
|
| 3373 |
+
"step": 46900
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"epoch": 2.673644689686558,
|
| 3377 |
+
"grad_norm": 1.8588401079177856,
|
| 3378 |
+
"learning_rate": 8.734664467072455e-05,
|
| 3379 |
+
"loss": 1.1288,
|
| 3380 |
+
"step": 47000
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 2.679333295409295,
|
| 3384 |
+
"grad_norm": 1.751246452331543,
|
| 3385 |
+
"learning_rate": 8.582968314466125e-05,
|
| 3386 |
+
"loss": 1.1206,
|
| 3387 |
+
"step": 47100
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 2.6850219011320324,
|
| 3391 |
+
"grad_norm": 1.7309458255767822,
|
| 3392 |
+
"learning_rate": 8.431272161859795e-05,
|
| 3393 |
+
"loss": 1.1089,
|
| 3394 |
+
"step": 47200
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 2.69071050685477,
|
| 3398 |
+
"grad_norm": 1.8057925701141357,
|
| 3399 |
+
"learning_rate": 8.281092970779529e-05,
|
| 3400 |
+
"loss": 1.1244,
|
| 3401 |
+
"step": 47300
|
| 3402 |
+
},
|
| 3403 |
+
{
|
| 3404 |
+
"epoch": 2.6963991125775073,
|
| 3405 |
+
"grad_norm": 1.7594059705734253,
|
| 3406 |
+
"learning_rate": 8.1293968181732e-05,
|
| 3407 |
+
"loss": 1.1188,
|
| 3408 |
+
"step": 47400
|
| 3409 |
+
},
|
| 3410 |
+
{
|
| 3411 |
+
"epoch": 2.7020877183002447,
|
| 3412 |
+
"grad_norm": 1.686438798904419,
|
| 3413 |
+
"learning_rate": 7.97770066556687e-05,
|
| 3414 |
+
"loss": 1.1068,
|
| 3415 |
+
"step": 47500
|
| 3416 |
+
},
|
| 3417 |
+
{
|
| 3418 |
+
"epoch": 2.7077763240229817,
|
| 3419 |
+
"grad_norm": 1.6962246894836426,
|
| 3420 |
+
"learning_rate": 7.82600451296054e-05,
|
| 3421 |
+
"loss": 1.1043,
|
| 3422 |
+
"step": 47600
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 2.713464929745719,
|
| 3426 |
+
"grad_norm": 1.5946807861328125,
|
| 3427 |
+
"learning_rate": 7.67430836035421e-05,
|
| 3428 |
+
"loss": 1.1109,
|
| 3429 |
+
"step": 47700
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 2.7191535354684566,
|
| 3433 |
+
"grad_norm": 1.4834094047546387,
|
| 3434 |
+
"learning_rate": 7.522612207747881e-05,
|
| 3435 |
+
"loss": 1.114,
|
| 3436 |
+
"step": 47800
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 2.724842141191194,
|
| 3440 |
+
"grad_norm": 1.763058066368103,
|
| 3441 |
+
"learning_rate": 7.370916055141553e-05,
|
| 3442 |
+
"loss": 1.1091,
|
| 3443 |
+
"step": 47900
|
| 3444 |
+
},
|
| 3445 |
+
{
|
| 3446 |
+
"epoch": 2.7305307469139315,
|
| 3447 |
+
"grad_norm": 1.9240601062774658,
|
| 3448 |
+
"learning_rate": 7.219219902535223e-05,
|
| 3449 |
+
"loss": 1.0936,
|
| 3450 |
+
"step": 48000
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 2.7362193526366685,
|
| 3454 |
+
"grad_norm": 1.4768198728561401,
|
| 3455 |
+
"learning_rate": 7.067523749928892e-05,
|
| 3456 |
+
"loss": 1.1158,
|
| 3457 |
+
"step": 48100
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 2.7419079583594064,
|
| 3461 |
+
"grad_norm": 1.9692409038543701,
|
| 3462 |
+
"learning_rate": 6.915827597322563e-05,
|
| 3463 |
+
"loss": 1.1201,
|
| 3464 |
+
"step": 48200
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 2.7475965640821434,
|
| 3468 |
+
"grad_norm": 1.636785864830017,
|
| 3469 |
+
"learning_rate": 6.764131444716234e-05,
|
| 3470 |
+
"loss": 1.1092,
|
| 3471 |
+
"step": 48300
|
| 3472 |
+
},
|
| 3473 |
+
{
|
| 3474 |
+
"epoch": 2.753285169804881,
|
| 3475 |
+
"grad_norm": 1.5599926710128784,
|
| 3476 |
+
"learning_rate": 6.612435292109905e-05,
|
| 3477 |
+
"loss": 1.0932,
|
| 3478 |
+
"step": 48400
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 2.7589737755276182,
|
| 3482 |
+
"grad_norm": 1.695862054824829,
|
| 3483 |
+
"learning_rate": 6.460739139503574e-05,
|
| 3484 |
+
"loss": 1.1227,
|
| 3485 |
+
"step": 48500
|
| 3486 |
+
},
|
| 3487 |
+
{
|
| 3488 |
+
"epoch": 2.7646623812503557,
|
| 3489 |
+
"grad_norm": 1.8806819915771484,
|
| 3490 |
+
"learning_rate": 6.309042986897246e-05,
|
| 3491 |
+
"loss": 1.1049,
|
| 3492 |
+
"step": 48600
|
| 3493 |
+
},
|
| 3494 |
+
{
|
| 3495 |
+
"epoch": 2.770350986973093,
|
| 3496 |
+
"grad_norm": 1.814792513847351,
|
| 3497 |
+
"learning_rate": 6.157346834290916e-05,
|
| 3498 |
+
"loss": 1.1149,
|
| 3499 |
+
"step": 48700
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 2.77603959269583,
|
| 3503 |
+
"grad_norm": 2.068614959716797,
|
| 3504 |
+
"learning_rate": 6.005650681684586e-05,
|
| 3505 |
+
"loss": 1.1181,
|
| 3506 |
+
"step": 48800
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 2.7817281984185676,
|
| 3510 |
+
"grad_norm": 1.5576444864273071,
|
| 3511 |
+
"learning_rate": 5.853954529078256e-05,
|
| 3512 |
+
"loss": 1.1223,
|
| 3513 |
+
"step": 48900
|
| 3514 |
+
},
|
| 3515 |
+
{
|
| 3516 |
+
"epoch": 2.787416804141305,
|
| 3517 |
+
"grad_norm": 1.8175384998321533,
|
| 3518 |
+
"learning_rate": 5.7022583764719273e-05,
|
| 3519 |
+
"loss": 1.1113,
|
| 3520 |
+
"step": 49000
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"epoch": 2.7931054098640424,
|
| 3524 |
+
"grad_norm": 1.570915937423706,
|
| 3525 |
+
"learning_rate": 5.550562223865598e-05,
|
| 3526 |
+
"loss": 1.1123,
|
| 3527 |
+
"step": 49100
|
| 3528 |
+
},
|
| 3529 |
+
{
|
| 3530 |
+
"epoch": 2.79879401558678,
|
| 3531 |
+
"grad_norm": 1.9663364887237549,
|
| 3532 |
+
"learning_rate": 5.3988660712592675e-05,
|
| 3533 |
+
"loss": 1.1065,
|
| 3534 |
+
"step": 49200
|
| 3535 |
+
},
|
| 3536 |
+
{
|
| 3537 |
+
"epoch": 2.804482621309517,
|
| 3538 |
+
"grad_norm": 2.2906079292297363,
|
| 3539 |
+
"learning_rate": 5.248686880179001e-05,
|
| 3540 |
+
"loss": 1.0993,
|
| 3541 |
+
"step": 49300
|
| 3542 |
+
},
|
| 3543 |
+
{
|
| 3544 |
+
"epoch": 2.8101712270322543,
|
| 3545 |
+
"grad_norm": 1.566801905632019,
|
| 3546 |
+
"learning_rate": 5.096990727572673e-05,
|
| 3547 |
+
"loss": 1.0964,
|
| 3548 |
+
"step": 49400
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 2.8158598327549917,
|
| 3552 |
+
"grad_norm": 1.7769867181777954,
|
| 3553 |
+
"learning_rate": 4.9452945749663425e-05,
|
| 3554 |
+
"loss": 1.0978,
|
| 3555 |
+
"step": 49500
|
| 3556 |
+
},
|
| 3557 |
+
{
|
| 3558 |
+
"epoch": 2.821548438477729,
|
| 3559 |
+
"grad_norm": 1.9856287240982056,
|
| 3560 |
+
"learning_rate": 4.7935984223600136e-05,
|
| 3561 |
+
"loss": 1.0875,
|
| 3562 |
+
"step": 49600
|
| 3563 |
+
},
|
| 3564 |
+
{
|
| 3565 |
+
"epoch": 2.8272370442004666,
|
| 3566 |
+
"grad_norm": 1.7836079597473145,
|
| 3567 |
+
"learning_rate": 4.6419022697536834e-05,
|
| 3568 |
+
"loss": 1.1056,
|
| 3569 |
+
"step": 49700
|
| 3570 |
+
},
|
| 3571 |
+
{
|
| 3572 |
+
"epoch": 2.8329256499232036,
|
| 3573 |
+
"grad_norm": 1.9246402978897095,
|
| 3574 |
+
"learning_rate": 4.4902061171473545e-05,
|
| 3575 |
+
"loss": 1.1074,
|
| 3576 |
+
"step": 49800
|
| 3577 |
+
},
|
| 3578 |
+
{
|
| 3579 |
+
"epoch": 2.838614255645941,
|
| 3580 |
+
"grad_norm": 1.3988184928894043,
|
| 3581 |
+
"learning_rate": 4.338509964541024e-05,
|
| 3582 |
+
"loss": 1.1206,
|
| 3583 |
+
"step": 49900
|
| 3584 |
+
},
|
| 3585 |
+
{
|
| 3586 |
+
"epoch": 2.8443028613686785,
|
| 3587 |
+
"grad_norm": 1.7193849086761475,
|
| 3588 |
+
"learning_rate": 4.186813811934695e-05,
|
| 3589 |
+
"loss": 1.1245,
|
| 3590 |
+
"step": 50000
|
| 3591 |
+
},
|
| 3592 |
+
{
|
| 3593 |
+
"epoch": 2.8443028613686785,
|
| 3594 |
+
"eval_accuracy": 0.729988,
|
| 3595 |
+
"eval_loss": 1.0782374143600464,
|
| 3596 |
+
"eval_runtime": 82.4205,
|
| 3597 |
+
"eval_samples_per_second": 3033.226,
|
| 3598 |
+
"eval_steps_per_second": 11.854,
|
| 3599 |
+
"step": 50000
|
| 3600 |
+
},
|
| 3601 |
+
{
|
| 3602 |
+
"epoch": 2.849991467091416,
|
| 3603 |
+
"grad_norm": 1.7059062719345093,
|
| 3604 |
+
"learning_rate": 4.035117659328366e-05,
|
| 3605 |
+
"loss": 1.1009,
|
| 3606 |
+
"step": 50100
|
| 3607 |
+
},
|
| 3608 |
+
{
|
| 3609 |
+
"epoch": 2.8556800728141534,
|
| 3610 |
+
"grad_norm": 1.4554681777954102,
|
| 3611 |
+
"learning_rate": 3.883421506722036e-05,
|
| 3612 |
+
"loss": 1.1108,
|
| 3613 |
+
"step": 50200
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 2.8613686785368904,
|
| 3617 |
+
"grad_norm": 1.7067590951919556,
|
| 3618 |
+
"learning_rate": 3.7317253541157065e-05,
|
| 3619 |
+
"loss": 1.0956,
|
| 3620 |
+
"step": 50300
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 2.867057284259628,
|
| 3624 |
+
"grad_norm": 1.7176940441131592,
|
| 3625 |
+
"learning_rate": 3.580029201509377e-05,
|
| 3626 |
+
"loss": 1.0841,
|
| 3627 |
+
"step": 50400
|
| 3628 |
+
},
|
| 3629 |
+
{
|
| 3630 |
+
"epoch": 2.8727458899823652,
|
| 3631 |
+
"grad_norm": 1.7251313924789429,
|
| 3632 |
+
"learning_rate": 3.4283330489030474e-05,
|
| 3633 |
+
"loss": 1.0908,
|
| 3634 |
+
"step": 50500
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 2.8784344957051027,
|
| 3638 |
+
"grad_norm": 1.668372631072998,
|
| 3639 |
+
"learning_rate": 3.276636896296718e-05,
|
| 3640 |
+
"loss": 1.0954,
|
| 3641 |
+
"step": 50600
|
| 3642 |
+
},
|
| 3643 |
+
{
|
| 3644 |
+
"epoch": 2.88412310142784,
|
| 3645 |
+
"grad_norm": 1.889109492301941,
|
| 3646 |
+
"learning_rate": 3.124940743690388e-05,
|
| 3647 |
+
"loss": 1.1096,
|
| 3648 |
+
"step": 50700
|
| 3649 |
+
},
|
| 3650 |
+
{
|
| 3651 |
+
"epoch": 2.8898117071505776,
|
| 3652 |
+
"grad_norm": 1.509391188621521,
|
| 3653 |
+
"learning_rate": 2.973244591084059e-05,
|
| 3654 |
+
"loss": 1.09,
|
| 3655 |
+
"step": 50800
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 2.895500312873315,
|
| 3659 |
+
"grad_norm": 2.024489402770996,
|
| 3660 |
+
"learning_rate": 2.821548438477729e-05,
|
| 3661 |
+
"loss": 1.0993,
|
| 3662 |
+
"step": 50900
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 2.901188918596052,
|
| 3666 |
+
"grad_norm": 2.007756471633911,
|
| 3667 |
+
"learning_rate": 2.6698522858713998e-05,
|
| 3668 |
+
"loss": 1.1029,
|
| 3669 |
+
"step": 51000
|
| 3670 |
+
},
|
| 3671 |
+
{
|
| 3672 |
+
"epoch": 2.9068775243187894,
|
| 3673 |
+
"grad_norm": 1.5296841859817505,
|
| 3674 |
+
"learning_rate": 2.51815613326507e-05,
|
| 3675 |
+
"loss": 1.095,
|
| 3676 |
+
"step": 51100
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"epoch": 2.912566130041527,
|
| 3680 |
+
"grad_norm": 1.6109613180160522,
|
| 3681 |
+
"learning_rate": 2.3664599806587406e-05,
|
| 3682 |
+
"loss": 1.1044,
|
| 3683 |
+
"step": 51200
|
| 3684 |
+
},
|
| 3685 |
+
{
|
| 3686 |
+
"epoch": 2.9182547357642643,
|
| 3687 |
+
"grad_norm": 1.8067957162857056,
|
| 3688 |
+
"learning_rate": 2.216280789578474e-05,
|
| 3689 |
+
"loss": 1.1074,
|
| 3690 |
+
"step": 51300
|
| 3691 |
+
},
|
| 3692 |
+
{
|
| 3693 |
+
"epoch": 2.9239433414870017,
|
| 3694 |
+
"grad_norm": 1.6997472047805786,
|
| 3695 |
+
"learning_rate": 2.064584636972145e-05,
|
| 3696 |
+
"loss": 1.1036,
|
| 3697 |
+
"step": 51400
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 2.9296319472097387,
|
| 3701 |
+
"grad_norm": 1.6443182229995728,
|
| 3702 |
+
"learning_rate": 1.9128884843658153e-05,
|
| 3703 |
+
"loss": 1.1258,
|
| 3704 |
+
"step": 51500
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 2.935320552932476,
|
| 3708 |
+
"grad_norm": 1.868161916732788,
|
| 3709 |
+
"learning_rate": 1.7611923317594857e-05,
|
| 3710 |
+
"loss": 1.115,
|
| 3711 |
+
"step": 51600
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 2.9410091586552136,
|
| 3715 |
+
"grad_norm": 1.5620206594467163,
|
| 3716 |
+
"learning_rate": 1.609496179153156e-05,
|
| 3717 |
+
"loss": 1.1075,
|
| 3718 |
+
"step": 51700
|
| 3719 |
+
},
|
| 3720 |
+
{
|
| 3721 |
+
"epoch": 2.946697764377951,
|
| 3722 |
+
"grad_norm": 1.6326332092285156,
|
| 3723 |
+
"learning_rate": 1.4578000265468267e-05,
|
| 3724 |
+
"loss": 1.1113,
|
| 3725 |
+
"step": 51800
|
| 3726 |
+
},
|
| 3727 |
+
{
|
| 3728 |
+
"epoch": 2.9523863701006885,
|
| 3729 |
+
"grad_norm": 1.805126428604126,
|
| 3730 |
+
"learning_rate": 1.3061038739404973e-05,
|
| 3731 |
+
"loss": 1.0937,
|
| 3732 |
+
"step": 51900
|
| 3733 |
+
},
|
| 3734 |
+
{
|
| 3735 |
+
"epoch": 2.9580749758234255,
|
| 3736 |
+
"grad_norm": 1.5693707466125488,
|
| 3737 |
+
"learning_rate": 1.1544077213341677e-05,
|
| 3738 |
+
"loss": 1.1053,
|
| 3739 |
+
"step": 52000
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 2.963763581546163,
|
| 3743 |
+
"grad_norm": 1.4851309061050415,
|
| 3744 |
+
"learning_rate": 1.0027115687278382e-05,
|
| 3745 |
+
"loss": 1.1101,
|
| 3746 |
+
"step": 52100
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 2.9694521872689004,
|
| 3750 |
+
"grad_norm": 1.8946778774261475,
|
| 3751 |
+
"learning_rate": 8.510154161215086e-06,
|
| 3752 |
+
"loss": 1.0934,
|
| 3753 |
+
"step": 52200
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 2.975140792991638,
|
| 3757 |
+
"grad_norm": 1.5624499320983887,
|
| 3758 |
+
"learning_rate": 6.993192635151791e-06,
|
| 3759 |
+
"loss": 1.0994,
|
| 3760 |
+
"step": 52300
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"epoch": 2.9808293987143752,
|
| 3764 |
+
"grad_norm": 1.5662641525268555,
|
| 3765 |
+
"learning_rate": 5.476231109088497e-06,
|
| 3766 |
+
"loss": 1.1058,
|
| 3767 |
+
"step": 52400
|
| 3768 |
+
},
|
| 3769 |
+
{
|
| 3770 |
+
"epoch": 2.9865180044371122,
|
| 3771 |
+
"grad_norm": 1.514809250831604,
|
| 3772 |
+
"learning_rate": 3.959269583025201e-06,
|
| 3773 |
+
"loss": 1.0967,
|
| 3774 |
+
"step": 52500
|
| 3775 |
+
},
|
| 3776 |
+
{
|
| 3777 |
+
"epoch": 2.9922066101598497,
|
| 3778 |
+
"grad_norm": 1.8442556858062744,
|
| 3779 |
+
"learning_rate": 2.4423080569619053e-06,
|
| 3780 |
+
"loss": 1.1041,
|
| 3781 |
+
"step": 52600
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 2.997895215882587,
|
| 3785 |
+
"grad_norm": 1.7445664405822754,
|
| 3786 |
+
"learning_rate": 9.253465308986101e-07,
|
| 3787 |
+
"loss": 1.0786,
|
| 3788 |
+
"step": 52700
|
| 3789 |
+
},
|
| 3790 |
+
{
|
| 3791 |
+
"epoch": 3.0,
|
| 3792 |
+
"step": 52737,
|
| 3793 |
+
"total_flos": 1.3116020904e+17,
|
| 3794 |
+
"train_loss": 1.269093582550002,
|
| 3795 |
+
"train_runtime": 7252.1306,
|
| 3796 |
+
"train_samples_per_second": 1861.522,
|
| 3797 |
+
"train_steps_per_second": 7.272
|
| 3798 |
+
}
|
| 3799 |
+
],
|
| 3800 |
+
"logging_steps": 100,
|
| 3801 |
+
"max_steps": 52737,
|
| 3802 |
+
"num_input_tokens_seen": 0,
|
| 3803 |
+
"num_train_epochs": 3,
|
| 3804 |
+
"save_steps": 5000,
|
| 3805 |
+
"stateful_callbacks": {
|
| 3806 |
+
"TrainerControl": {
|
| 3807 |
+
"args": {
|
| 3808 |
+
"should_epoch_stop": false,
|
| 3809 |
+
"should_evaluate": false,
|
| 3810 |
+
"should_log": false,
|
| 3811 |
+
"should_save": true,
|
| 3812 |
+
"should_training_stop": true
|
| 3813 |
+
},
|
| 3814 |
+
"attributes": {}
|
| 3815 |
+
}
|
| 3816 |
+
},
|
| 3817 |
+
"total_flos": 1.3116020904e+17,
|
| 3818 |
+
"train_batch_size": 256,
|
| 3819 |
+
"trial_name": null,
|
| 3820 |
+
"trial_params": null
|
| 3821 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4741f9929f86e0a0cc5ef0c6799bcd5cfd09259484bf459d7901ff7158af6501
|
| 3 |
+
size 5112
|