Upload 8 files
Browse files- README.md +135 -3
- all_results.json +8 -0
- config.json +40 -0
- model.safetensors +3 -0
- preprocessor_config.json +25 -0
- test_results.json +8 -0
- trainer_state.json +2786 -0
- training_args.bin +3 -0
README.md
CHANGED
|
@@ -1,3 +1,135 @@
|
|
| 1 |
-
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: MCG-NJU/videomae-large-finetuned-kinetics
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
metrics:
|
| 8 |
+
- accuracy
|
| 9 |
+
model-index:
|
| 10 |
+
- name: Hibernates-MEA-R2-V0
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
# Hibernates-MEA-R2-V0
|
| 15 |
+
|
| 16 |
+
An advanced AI system for visual sequence processing, extending the capabilities of [MCG-NJU/videomae-large-finetuned-kinetics](https://huggingface.co/MCG-NJU/videomae-large-finetuned-kinetics).
|
| 17 |
+
|
| 18 |
+
Key Performance Indicators:
|
| 19 |
+
- Optimal Loss: 0.4894
|
| 20 |
+
- Peak Accuracy: 80.43%
|
| 21 |
+
|
| 22 |
+
## System Overview
|
| 23 |
+
|
| 24 |
+
Advanced AI architecture optimized for visual sequence understanding:
|
| 25 |
+
|
| 26 |
+
- Core: Deep learning transformer system
|
| 27 |
+
- Data Handling: Sequential frame processing
|
| 28 |
+
- Main Function: Visual content categorization
|
| 29 |
+
- Learning Cycles: 50 complete epochs
|
| 30 |
+
- Results Summary:
|
| 31 |
+
* Maximum Precision: 80.43% (epoch 7)
|
| 32 |
+
* Consistent Performance: 75%+ maintained
|
| 33 |
+
|
| 34 |
+
## Applications & Requirements
|
| 35 |
+
|
| 36 |
+
### Core Applications
|
| 37 |
+
- Visual sequence interpretation
|
| 38 |
+
- Dynamic content analysis
|
| 39 |
+
- Environmental context recognition
|
| 40 |
+
- Time-series visual processing
|
| 41 |
+
|
| 42 |
+
### Technical Considerations
|
| 43 |
+
- Task-specific optimization
|
| 44 |
+
- Computing needs: High-performance GPU
|
| 45 |
+
- Memory constraints: 4-sample batching
|
| 46 |
+
- Data format: Standardized input required
|
| 47 |
+
|
| 48 |
+
## Development Data
|
| 49 |
+
|
| 50 |
+
Implementation Details:
|
| 51 |
+
- Cycle Structure: 65 iterations per epoch
|
| 52 |
+
- Development Span: 3250 total iterations
|
| 53 |
+
- Assessment Methods: Dual metric system (loss/accuracy)
|
| 54 |
+
- Progress Metrics:
|
| 55 |
+
* Starting Point: 54% accuracy
|
| 56 |
+
* Final Result: 73.91%
|
| 57 |
+
* Best-case Loss: 0.4894
|
| 58 |
+
|
| 59 |
+
## Implementation Specifications
|
| 60 |
+
|
| 61 |
+
### Core Parameters
|
| 62 |
+
|
| 63 |
+
Implementation utilized the following configuration:
|
| 64 |
+
- Learning Rate: 1e-05
|
| 65 |
+
- Training Units: 4 per batch
|
| 66 |
+
- Validation Units: 4 per batch
|
| 67 |
+
- Random Seed: 42
|
| 68 |
+
- Optimization: Advanced weight management with adamw_torch
|
| 69 |
+
* Beta values: (0.9,0.999)
|
| 70 |
+
* Epsilon: 1e-08
|
| 71 |
+
- Rate Control: Linear adjustment
|
| 72 |
+
- Warmup Ratio: 0.1
|
| 73 |
+
- Total Iterations: 3250
|
| 74 |
+
|
| 75 |
+
### Development Progress
|
| 76 |
+
|
| 77 |
+
| Cycle Loss | Epoch | Step | Validation Loss | Success Rate |
|
| 78 |
+
|:----------:|:-----:|:----:|:---------------:|:------------:|
|
| 79 |
+
| 0.6186 | 0.02 | 65 | 0.7367 | 0.5435 |
|
| 80 |
+
| 0.5974 | 1.02 | 130 | 0.8185 | 0.5435 |
|
| 81 |
+
| 0.5491 | 2.02 | 195 | 0.8372 | 0.5435 |
|
| 82 |
+
| 0.6156 | 3.02 | 260 | 0.6620 | 0.5870 |
|
| 83 |
+
| 0.6255 | 4.02 | 325 | 0.6835 | 0.5435 |
|
| 84 |
+
| 0.438 | 5.02 | 390 | 1.2116 | 0.5435 |
|
| 85 |
+
| 0.4653 | 6.02 | 455 | 0.6002 | 0.5652 |
|
| 86 |
+
| 0.5876 | 7.02 | 520 | 0.4894 | 0.8043 |
|
| 87 |
+
| 0.3801 | 8.02 | 585 | 0.8324 | 0.5435 |
|
| 88 |
+
| 0.4474 | 9.02 | 650 | 1.1581 | 0.5652 |
|
| 89 |
+
| 0.694 | 10.02 | 715 | 0.5354 | 0.7174 |
|
| 90 |
+
| 0.4773 | 11.02 | 780 | 0.6181 | 0.6957 |
|
| 91 |
+
| 0.6208 | 12.02 | 845 | 0.5677 | 0.7609 |
|
| 92 |
+
| 0.344 | 13.02 | 910 | 0.7452 | 0.6087 |
|
| 93 |
+
| 0.254 | 14.02 | 975 | 0.6362 | 0.7391 |
|
| 94 |
+
| 0.4578 | 15.02 | 1040 | 0.8304 | 0.6957 |
|
| 95 |
+
| 0.3954 | 16.02 | 1105 | 0.6049 | 0.7609 |
|
| 96 |
+
| 0.248 | 17.02 | 1170 | 0.9506 | 0.6739 |
|
| 97 |
+
| 0.1334 | 18.02 | 1235 | 1.1876 | 0.6739 |
|
| 98 |
+
| 0.534 | 19.02 | 1300 | 0.6296 | 0.7391 |
|
| 99 |
+
| 0.3556 | 20.02 | 1365 | 1.3007 | 0.6957 |
|
| 100 |
+
| 0.5439 | 21.02 | 1430 | 1.5066 | 0.6739 |
|
| 101 |
+
| 0.4107 | 22.02 | 1495 | 0.9273 | 0.8043 |
|
| 102 |
+
| 0.61 | 23.02 | 1560 | 1.0008 | 0.7174 |
|
| 103 |
+
| 0.6482 | 24.02 | 1625 | 0.7548 | 0.7609 |
|
| 104 |
+
| 0.199 | 25.02 | 1690 | 0.7917 | 0.7826 |
|
| 105 |
+
| 0.1185 | 26.02 | 1755 | 0.7529 | 0.7826 |
|
| 106 |
+
| 0.3886 | 27.02 | 1820 | 0.8627 | 0.7609 |
|
| 107 |
+
| 0.0123 | 28.02 | 1885 | 1.3886 | 0.7174 |
|
| 108 |
+
| 0.5328 | 29.02 | 1950 | 1.2803 | 0.6957 |
|
| 109 |
+
| 0.2961 | 30.02 | 2015 | 1.4397 | 0.7174 |
|
| 110 |
+
| 0.1192 | 31.02 | 2080 | 2.2563 | 0.6304 |
|
| 111 |
+
| 0.145 | 32.02 | 2145 | 1.0465 | 0.7609 |
|
| 112 |
+
| 0.0924 | 33.02 | 2210 | 0.9859 | 0.7826 |
|
| 113 |
+
| 0.1016 | 34.02 | 2275 | 1.0758 | 0.7826 |
|
| 114 |
+
| 0.1894 | 35.02 | 2340 | 1.2088 | 0.7609 |
|
| 115 |
+
| 0.2657 | 36.02 | 2405 | 1.5409 | 0.7391 |
|
| 116 |
+
| 0.1235 | 37.02 | 2470 | 1.2736 | 0.7609 |
|
| 117 |
+
| 0.1539 | 38.02 | 2535 | 1.2608 | 0.7609 |
|
| 118 |
+
| 0.03 | 39.02 | 2600 | 1.2058 | 0.7609 |
|
| 119 |
+
| 0.1447 | 40.02 | 2665 | 1.1072 | 0.7609 |
|
| 120 |
+
| 0.0888 | 41.02 | 2730 | 1.1454 | 0.7826 |
|
| 121 |
+
| 0.0016 | 42.02 | 2795 | 1.1194 | 0.7826 |
|
| 122 |
+
| 0.1489 | 43.02 | 2860 | 1.2170 | 0.7609 |
|
| 123 |
+
| 0.0004 | 44.02 | 2925 | 1.1894 | 0.7609 |
|
| 124 |
+
| 0.0004 | 45.02 | 2990 | 1.3329 | 0.7391 |
|
| 125 |
+
| 0.0014 | 46.02 | 3055 | 1.1887 | 0.7609 |
|
| 126 |
+
| 0.1675 | 47.02 | 3120 | 1.2652 | 0.7391 |
|
| 127 |
+
| 0.012 | 48.02 | 3185 | 1.3228 | 0.7391 |
|
| 128 |
+
| 0.0475 | 49.02 | 3250 | 1.3507 | 0.7391 |
|
| 129 |
+
|
| 130 |
+
### System Versions
|
| 131 |
+
|
| 132 |
+
- Transformers 4.46.2
|
| 133 |
+
- Pytorch 2.0.1+cu117
|
| 134 |
+
- Datasets 3.0.1
|
| 135 |
+
- Tokenizers 0.20.0
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 49.02,
|
| 3 |
+
"eval_accuracy": 0.8043478260869565,
|
| 4 |
+
"eval_loss": 0.4894070327281952,
|
| 5 |
+
"eval_runtime": 10.4359,
|
| 6 |
+
"eval_samples_per_second": 4.408,
|
| 7 |
+
"eval_steps_per_second": 1.15
|
| 8 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "MCG-NJU/videomae-large-finetuned-kinetics",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"VideoMAEForVideoClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"decoder_hidden_size": 512,
|
| 8 |
+
"decoder_intermediate_size": 2048,
|
| 9 |
+
"decoder_num_attention_heads": 8,
|
| 10 |
+
"decoder_num_hidden_layers": 12,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.0,
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"id2label": {
|
| 15 |
+
"0": "0",
|
| 16 |
+
"1": "1"
|
| 17 |
+
},
|
| 18 |
+
"image_size": 224,
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 4096,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"0": 0,
|
| 23 |
+
"1": 1
|
| 24 |
+
},
|
| 25 |
+
"layer_norm_eps": 1e-12,
|
| 26 |
+
"model_type": "videomae",
|
| 27 |
+
"norm_pix_loss": true,
|
| 28 |
+
"num_attention_heads": 16,
|
| 29 |
+
"num_channels": 3,
|
| 30 |
+
"num_frames": 16,
|
| 31 |
+
"num_hidden_layers": 24,
|
| 32 |
+
"patch_size": 16,
|
| 33 |
+
"problem_type": "single_label_classification",
|
| 34 |
+
"qkv_bias": true,
|
| 35 |
+
"torch_dtype": "float32",
|
| 36 |
+
"transformers_version": "4.46.2",
|
| 37 |
+
"tubelet_size": 2,
|
| 38 |
+
"use_mean_pooling": true
|
| 39 |
+
}
|
| 40 |
+
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf6083214a830480318dc8cda674d9dbcee6562702308a901fda36da86087643
|
| 3 |
+
size 1215496208
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 224,
|
| 4 |
+
"width": 224
|
| 5 |
+
},
|
| 6 |
+
"do_center_crop": true,
|
| 7 |
+
"do_normalize": true,
|
| 8 |
+
"do_rescale": true,
|
| 9 |
+
"do_resize": true,
|
| 10 |
+
"image_mean": [
|
| 11 |
+
0.485,
|
| 12 |
+
0.456,
|
| 13 |
+
0.406
|
| 14 |
+
],
|
| 15 |
+
"image_processor_type": "VideoMAEImageProcessor",
|
| 16 |
+
"image_std": [
|
| 17 |
+
0.229,
|
| 18 |
+
0.224,
|
| 19 |
+
0.225
|
| 20 |
+
],
|
| 21 |
+
"resample": 2,
|
| 22 |
+
"rescale_factor": 0.00392156862745098,
|
| 23 |
+
"size": 224
|
| 24 |
+
}
|
| 25 |
+
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 49.02,
|
| 3 |
+
"eval_accuracy": 0.8043478260869565,
|
| 4 |
+
"eval_loss": 0.4894070327281952,
|
| 5 |
+
"eval_runtime": 10.4359,
|
| 6 |
+
"eval_samples_per_second": 4.408,
|
| 7 |
+
"eval_steps_per_second": 1.15
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2786 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8043478260869565,
|
| 3 |
+
"best_model_checkpoint": "CT-MAE-RIS2-Phase2-V1/checkpoint-520",
|
| 4 |
+
"epoch": 49.02,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3250,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003076923076923077,
|
| 13 |
+
"grad_norm": 2.7215452194213867,
|
| 14 |
+
"learning_rate": 3.0769230769230774e-07,
|
| 15 |
+
"loss": 0.6876,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.006153846153846154,
|
| 20 |
+
"grad_norm": 2.571739673614502,
|
| 21 |
+
"learning_rate": 6.153846153846155e-07,
|
| 22 |
+
"loss": 0.6748,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.009230769230769232,
|
| 27 |
+
"grad_norm": 4.409002780914307,
|
| 28 |
+
"learning_rate": 9.230769230769232e-07,
|
| 29 |
+
"loss": 0.6605,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.012307692307692308,
|
| 34 |
+
"grad_norm": 4.084295272827148,
|
| 35 |
+
"learning_rate": 1.230769230769231e-06,
|
| 36 |
+
"loss": 0.6389,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.015384615384615385,
|
| 41 |
+
"grad_norm": 11.590015411376953,
|
| 42 |
+
"learning_rate": 1.5384615384615387e-06,
|
| 43 |
+
"loss": 0.5934,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.018461538461538463,
|
| 48 |
+
"grad_norm": 8.831100463867188,
|
| 49 |
+
"learning_rate": 1.8461538461538465e-06,
|
| 50 |
+
"loss": 0.6186,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"eval_accuracy": 0.5434782608695652,
|
| 56 |
+
"eval_loss": 0.7367333173751831,
|
| 57 |
+
"eval_runtime": 11.5008,
|
| 58 |
+
"eval_samples_per_second": 4.0,
|
| 59 |
+
"eval_steps_per_second": 1.043,
|
| 60 |
+
"step": 65
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.0015384615384615,
|
| 64 |
+
"grad_norm": 11.279766082763672,
|
| 65 |
+
"learning_rate": 2.153846153846154e-06,
|
| 66 |
+
"loss": 0.5907,
|
| 67 |
+
"step": 70
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 1.0046153846153847,
|
| 71 |
+
"grad_norm": 5.7519145011901855,
|
| 72 |
+
"learning_rate": 2.461538461538462e-06,
|
| 73 |
+
"loss": 0.617,
|
| 74 |
+
"step": 80
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.0076923076923077,
|
| 78 |
+
"grad_norm": 4.054751873016357,
|
| 79 |
+
"learning_rate": 2.7692307692307697e-06,
|
| 80 |
+
"loss": 0.6291,
|
| 81 |
+
"step": 90
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.0107692307692309,
|
| 85 |
+
"grad_norm": 5.010909080505371,
|
| 86 |
+
"learning_rate": 3.0769230769230774e-06,
|
| 87 |
+
"loss": 0.528,
|
| 88 |
+
"step": 100
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 1.0138461538461538,
|
| 92 |
+
"grad_norm": 14.651056289672852,
|
| 93 |
+
"learning_rate": 3.384615384615385e-06,
|
| 94 |
+
"loss": 0.4573,
|
| 95 |
+
"step": 110
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.0169230769230768,
|
| 99 |
+
"grad_norm": 11.85365104675293,
|
| 100 |
+
"learning_rate": 3.692307692307693e-06,
|
| 101 |
+
"loss": 0.6294,
|
| 102 |
+
"step": 120
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.02,
|
| 106 |
+
"grad_norm": 8.004859924316406,
|
| 107 |
+
"learning_rate": 4.000000000000001e-06,
|
| 108 |
+
"loss": 0.5974,
|
| 109 |
+
"step": 130
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.02,
|
| 113 |
+
"eval_accuracy": 0.5434782608695652,
|
| 114 |
+
"eval_loss": 0.8185142874717712,
|
| 115 |
+
"eval_runtime": 10.3925,
|
| 116 |
+
"eval_samples_per_second": 4.426,
|
| 117 |
+
"eval_steps_per_second": 1.155,
|
| 118 |
+
"step": 130
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 2.003076923076923,
|
| 122 |
+
"grad_norm": 14.822097778320312,
|
| 123 |
+
"learning_rate": 4.307692307692308e-06,
|
| 124 |
+
"loss": 0.5002,
|
| 125 |
+
"step": 140
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 2.006153846153846,
|
| 129 |
+
"grad_norm": 10.74063491821289,
|
| 130 |
+
"learning_rate": 4.615384615384616e-06,
|
| 131 |
+
"loss": 0.6184,
|
| 132 |
+
"step": 150
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 2.0092307692307694,
|
| 136 |
+
"grad_norm": 9.221219062805176,
|
| 137 |
+
"learning_rate": 4.923076923076924e-06,
|
| 138 |
+
"loss": 0.5821,
|
| 139 |
+
"step": 160
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 2.0123076923076924,
|
| 143 |
+
"grad_norm": 4.616571426391602,
|
| 144 |
+
"learning_rate": 5.230769230769232e-06,
|
| 145 |
+
"loss": 0.617,
|
| 146 |
+
"step": 170
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.0153846153846153,
|
| 150 |
+
"grad_norm": 8.066702842712402,
|
| 151 |
+
"learning_rate": 5.538461538461539e-06,
|
| 152 |
+
"loss": 0.561,
|
| 153 |
+
"step": 180
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.0184615384615383,
|
| 157 |
+
"grad_norm": 7.593764781951904,
|
| 158 |
+
"learning_rate": 5.846153846153847e-06,
|
| 159 |
+
"loss": 0.5491,
|
| 160 |
+
"step": 190
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.02,
|
| 164 |
+
"eval_accuracy": 0.5434782608695652,
|
| 165 |
+
"eval_loss": 0.8371568918228149,
|
| 166 |
+
"eval_runtime": 10.653,
|
| 167 |
+
"eval_samples_per_second": 4.318,
|
| 168 |
+
"eval_steps_per_second": 1.126,
|
| 169 |
+
"step": 195
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.0015384615384617,
|
| 173 |
+
"grad_norm": 6.1916985511779785,
|
| 174 |
+
"learning_rate": 6.153846153846155e-06,
|
| 175 |
+
"loss": 0.5725,
|
| 176 |
+
"step": 200
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0046153846153847,
|
| 180 |
+
"grad_norm": 6.418662071228027,
|
| 181 |
+
"learning_rate": 6.461538461538463e-06,
|
| 182 |
+
"loss": 0.4794,
|
| 183 |
+
"step": 210
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0076923076923077,
|
| 187 |
+
"grad_norm": 9.704038619995117,
|
| 188 |
+
"learning_rate": 6.76923076923077e-06,
|
| 189 |
+
"loss": 0.4911,
|
| 190 |
+
"step": 220
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0107692307692306,
|
| 194 |
+
"grad_norm": 5.40470027923584,
|
| 195 |
+
"learning_rate": 7.076923076923078e-06,
|
| 196 |
+
"loss": 0.5041,
|
| 197 |
+
"step": 230
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.013846153846154,
|
| 201 |
+
"grad_norm": 7.060332775115967,
|
| 202 |
+
"learning_rate": 7.384615384615386e-06,
|
| 203 |
+
"loss": 0.591,
|
| 204 |
+
"step": 240
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.016923076923077,
|
| 208 |
+
"grad_norm": 34.150779724121094,
|
| 209 |
+
"learning_rate": 7.692307692307694e-06,
|
| 210 |
+
"loss": 0.5028,
|
| 211 |
+
"step": 250
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.02,
|
| 215 |
+
"grad_norm": 6.146841526031494,
|
| 216 |
+
"learning_rate": 8.000000000000001e-06,
|
| 217 |
+
"loss": 0.6156,
|
| 218 |
+
"step": 260
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.02,
|
| 222 |
+
"eval_accuracy": 0.5869565217391305,
|
| 223 |
+
"eval_loss": 0.6619947552680969,
|
| 224 |
+
"eval_runtime": 10.4311,
|
| 225 |
+
"eval_samples_per_second": 4.41,
|
| 226 |
+
"eval_steps_per_second": 1.15,
|
| 227 |
+
"step": 260
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 4.003076923076923,
|
| 231 |
+
"grad_norm": 2.565047025680542,
|
| 232 |
+
"learning_rate": 8.307692307692309e-06,
|
| 233 |
+
"loss": 0.4261,
|
| 234 |
+
"step": 270
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 4.006153846153846,
|
| 238 |
+
"grad_norm": 7.034364700317383,
|
| 239 |
+
"learning_rate": 8.615384615384617e-06,
|
| 240 |
+
"loss": 0.647,
|
| 241 |
+
"step": 280
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 4.009230769230769,
|
| 245 |
+
"grad_norm": 6.253409385681152,
|
| 246 |
+
"learning_rate": 8.923076923076925e-06,
|
| 247 |
+
"loss": 0.5691,
|
| 248 |
+
"step": 290
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.012307692307692,
|
| 252 |
+
"grad_norm": 1.1683131456375122,
|
| 253 |
+
"learning_rate": 9.230769230769232e-06,
|
| 254 |
+
"loss": 0.4258,
|
| 255 |
+
"step": 300
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 4.015384615384615,
|
| 259 |
+
"grad_norm": 18.69379234313965,
|
| 260 |
+
"learning_rate": 9.53846153846154e-06,
|
| 261 |
+
"loss": 0.9862,
|
| 262 |
+
"step": 310
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 4.018461538461539,
|
| 266 |
+
"grad_norm": 9.839452743530273,
|
| 267 |
+
"learning_rate": 9.846153846153848e-06,
|
| 268 |
+
"loss": 0.6255,
|
| 269 |
+
"step": 320
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 4.02,
|
| 273 |
+
"eval_accuracy": 0.5434782608695652,
|
| 274 |
+
"eval_loss": 0.6834591031074524,
|
| 275 |
+
"eval_runtime": 10.7712,
|
| 276 |
+
"eval_samples_per_second": 4.271,
|
| 277 |
+
"eval_steps_per_second": 1.114,
|
| 278 |
+
"step": 325
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.001538461538462,
|
| 282 |
+
"grad_norm": 11.986108779907227,
|
| 283 |
+
"learning_rate": 9.982905982905984e-06,
|
| 284 |
+
"loss": 0.5837,
|
| 285 |
+
"step": 330
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 5.004615384615384,
|
| 289 |
+
"grad_norm": 10.73773193359375,
|
| 290 |
+
"learning_rate": 9.94871794871795e-06,
|
| 291 |
+
"loss": 0.6487,
|
| 292 |
+
"step": 340
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.007692307692308,
|
| 296 |
+
"grad_norm": 2.132850170135498,
|
| 297 |
+
"learning_rate": 9.914529914529915e-06,
|
| 298 |
+
"loss": 0.5756,
|
| 299 |
+
"step": 350
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.010769230769231,
|
| 303 |
+
"grad_norm": 3.829684257507324,
|
| 304 |
+
"learning_rate": 9.880341880341882e-06,
|
| 305 |
+
"loss": 0.4838,
|
| 306 |
+
"step": 360
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.013846153846154,
|
| 310 |
+
"grad_norm": 6.597207546234131,
|
| 311 |
+
"learning_rate": 9.846153846153848e-06,
|
| 312 |
+
"loss": 0.5747,
|
| 313 |
+
"step": 370
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.016923076923077,
|
| 317 |
+
"grad_norm": 15.660137176513672,
|
| 318 |
+
"learning_rate": 9.811965811965812e-06,
|
| 319 |
+
"loss": 0.5223,
|
| 320 |
+
"step": 380
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.02,
|
| 324 |
+
"grad_norm": 13.29094123840332,
|
| 325 |
+
"learning_rate": 9.777777777777779e-06,
|
| 326 |
+
"loss": 0.438,
|
| 327 |
+
"step": 390
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.02,
|
| 331 |
+
"eval_accuracy": 0.5434782608695652,
|
| 332 |
+
"eval_loss": 1.2116292715072632,
|
| 333 |
+
"eval_runtime": 10.5835,
|
| 334 |
+
"eval_samples_per_second": 4.346,
|
| 335 |
+
"eval_steps_per_second": 1.134,
|
| 336 |
+
"step": 390
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 6.003076923076923,
|
| 340 |
+
"grad_norm": 10.139371871948242,
|
| 341 |
+
"learning_rate": 9.743589743589744e-06,
|
| 342 |
+
"loss": 0.7131,
|
| 343 |
+
"step": 400
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 6.006153846153846,
|
| 347 |
+
"grad_norm": 15.849384307861328,
|
| 348 |
+
"learning_rate": 9.70940170940171e-06,
|
| 349 |
+
"loss": 0.5023,
|
| 350 |
+
"step": 410
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 6.009230769230769,
|
| 354 |
+
"grad_norm": 5.638754844665527,
|
| 355 |
+
"learning_rate": 9.675213675213677e-06,
|
| 356 |
+
"loss": 0.4474,
|
| 357 |
+
"step": 420
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 6.012307692307692,
|
| 361 |
+
"grad_norm": 25.2630672454834,
|
| 362 |
+
"learning_rate": 9.641025641025642e-06,
|
| 363 |
+
"loss": 0.5027,
|
| 364 |
+
"step": 430
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 6.015384615384615,
|
| 368 |
+
"grad_norm": 13.194704055786133,
|
| 369 |
+
"learning_rate": 9.606837606837607e-06,
|
| 370 |
+
"loss": 0.4413,
|
| 371 |
+
"step": 440
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 6.018461538461539,
|
| 375 |
+
"grad_norm": 24.00779151916504,
|
| 376 |
+
"learning_rate": 9.572649572649575e-06,
|
| 377 |
+
"loss": 0.4653,
|
| 378 |
+
"step": 450
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 6.02,
|
| 382 |
+
"eval_accuracy": 0.5652173913043478,
|
| 383 |
+
"eval_loss": 0.6001840829849243,
|
| 384 |
+
"eval_runtime": 10.1529,
|
| 385 |
+
"eval_samples_per_second": 4.531,
|
| 386 |
+
"eval_steps_per_second": 1.182,
|
| 387 |
+
"step": 455
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 7.001538461538462,
|
| 391 |
+
"grad_norm": 8.63178539276123,
|
| 392 |
+
"learning_rate": 9.53846153846154e-06,
|
| 393 |
+
"loss": 0.573,
|
| 394 |
+
"step": 460
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 7.004615384615384,
|
| 398 |
+
"grad_norm": 26.25208282470703,
|
| 399 |
+
"learning_rate": 9.504273504273504e-06,
|
| 400 |
+
"loss": 0.5129,
|
| 401 |
+
"step": 470
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 7.007692307692308,
|
| 405 |
+
"grad_norm": 6.66032600402832,
|
| 406 |
+
"learning_rate": 9.470085470085471e-06,
|
| 407 |
+
"loss": 0.7248,
|
| 408 |
+
"step": 480
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.010769230769231,
|
| 412 |
+
"grad_norm": 3.6293530464172363,
|
| 413 |
+
"learning_rate": 9.435897435897436e-06,
|
| 414 |
+
"loss": 0.333,
|
| 415 |
+
"step": 490
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 7.013846153846154,
|
| 419 |
+
"grad_norm": 6.718216419219971,
|
| 420 |
+
"learning_rate": 9.401709401709402e-06,
|
| 421 |
+
"loss": 0.6385,
|
| 422 |
+
"step": 500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 7.016923076923077,
|
| 426 |
+
"grad_norm": 15.813761711120605,
|
| 427 |
+
"learning_rate": 9.367521367521369e-06,
|
| 428 |
+
"loss": 0.6076,
|
| 429 |
+
"step": 510
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 7.02,
|
| 433 |
+
"grad_norm": 9.687027931213379,
|
| 434 |
+
"learning_rate": 9.333333333333334e-06,
|
| 435 |
+
"loss": 0.5876,
|
| 436 |
+
"step": 520
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 7.02,
|
| 440 |
+
"eval_accuracy": 0.8043478260869565,
|
| 441 |
+
"eval_loss": 0.48940709233283997,
|
| 442 |
+
"eval_runtime": 10.0977,
|
| 443 |
+
"eval_samples_per_second": 4.555,
|
| 444 |
+
"eval_steps_per_second": 1.188,
|
| 445 |
+
"step": 520
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 8.003076923076923,
|
| 449 |
+
"grad_norm": 2.885645627975464,
|
| 450 |
+
"learning_rate": 9.2991452991453e-06,
|
| 451 |
+
"loss": 0.6155,
|
| 452 |
+
"step": 530
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 8.006153846153847,
|
| 456 |
+
"grad_norm": 41.08794021606445,
|
| 457 |
+
"learning_rate": 9.264957264957267e-06,
|
| 458 |
+
"loss": 0.4796,
|
| 459 |
+
"step": 540
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 8.009230769230768,
|
| 463 |
+
"grad_norm": 1.9967862367630005,
|
| 464 |
+
"learning_rate": 9.230769230769232e-06,
|
| 465 |
+
"loss": 0.333,
|
| 466 |
+
"step": 550
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 8.012307692307692,
|
| 470 |
+
"grad_norm": 4.82252311706543,
|
| 471 |
+
"learning_rate": 9.196581196581196e-06,
|
| 472 |
+
"loss": 0.6022,
|
| 473 |
+
"step": 560
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 8.015384615384615,
|
| 477 |
+
"grad_norm": 3.3050286769866943,
|
| 478 |
+
"learning_rate": 9.162393162393163e-06,
|
| 479 |
+
"loss": 0.4147,
|
| 480 |
+
"step": 570
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 8.018461538461539,
|
| 484 |
+
"grad_norm": 0.10243837535381317,
|
| 485 |
+
"learning_rate": 9.128205128205129e-06,
|
| 486 |
+
"loss": 0.3801,
|
| 487 |
+
"step": 580
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 8.02,
|
| 491 |
+
"eval_accuracy": 0.5434782608695652,
|
| 492 |
+
"eval_loss": 0.8324462175369263,
|
| 493 |
+
"eval_runtime": 10.0187,
|
| 494 |
+
"eval_samples_per_second": 4.591,
|
| 495 |
+
"eval_steps_per_second": 1.198,
|
| 496 |
+
"step": 585
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 9.00153846153846,
|
| 500 |
+
"grad_norm": 4.7592668533325195,
|
| 501 |
+
"learning_rate": 9.094017094017094e-06,
|
| 502 |
+
"loss": 1.0305,
|
| 503 |
+
"step": 590
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 9.004615384615384,
|
| 507 |
+
"grad_norm": 8.245183944702148,
|
| 508 |
+
"learning_rate": 9.059829059829061e-06,
|
| 509 |
+
"loss": 0.5193,
|
| 510 |
+
"step": 600
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 9.007692307692308,
|
| 514 |
+
"grad_norm": 24.77902603149414,
|
| 515 |
+
"learning_rate": 9.025641025641027e-06,
|
| 516 |
+
"loss": 0.4919,
|
| 517 |
+
"step": 610
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 9.010769230769231,
|
| 521 |
+
"grad_norm": 29.81329345703125,
|
| 522 |
+
"learning_rate": 8.991452991452992e-06,
|
| 523 |
+
"loss": 0.6428,
|
| 524 |
+
"step": 620
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 9.013846153846155,
|
| 528 |
+
"grad_norm": 8.413053512573242,
|
| 529 |
+
"learning_rate": 8.957264957264959e-06,
|
| 530 |
+
"loss": 0.3073,
|
| 531 |
+
"step": 630
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 9.016923076923076,
|
| 535 |
+
"grad_norm": 8.200304985046387,
|
| 536 |
+
"learning_rate": 8.923076923076925e-06,
|
| 537 |
+
"loss": 0.6606,
|
| 538 |
+
"step": 640
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 9.02,
|
| 542 |
+
"grad_norm": 22.091062545776367,
|
| 543 |
+
"learning_rate": 8.888888888888888e-06,
|
| 544 |
+
"loss": 0.4474,
|
| 545 |
+
"step": 650
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 9.02,
|
| 549 |
+
"eval_accuracy": 0.5652173913043478,
|
| 550 |
+
"eval_loss": 1.158095359802246,
|
| 551 |
+
"eval_runtime": 10.3694,
|
| 552 |
+
"eval_samples_per_second": 4.436,
|
| 553 |
+
"eval_steps_per_second": 1.157,
|
| 554 |
+
"step": 650
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 10.003076923076923,
|
| 558 |
+
"grad_norm": 4.3752288818359375,
|
| 559 |
+
"learning_rate": 8.854700854700855e-06,
|
| 560 |
+
"loss": 1.0286,
|
| 561 |
+
"step": 660
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 10.006153846153847,
|
| 565 |
+
"grad_norm": 5.74050235748291,
|
| 566 |
+
"learning_rate": 8.820512820512821e-06,
|
| 567 |
+
"loss": 0.4695,
|
| 568 |
+
"step": 670
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 10.009230769230768,
|
| 572 |
+
"grad_norm": 4.538609504699707,
|
| 573 |
+
"learning_rate": 8.786324786324786e-06,
|
| 574 |
+
"loss": 0.3401,
|
| 575 |
+
"step": 680
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 10.012307692307692,
|
| 579 |
+
"grad_norm": 6.925298690795898,
|
| 580 |
+
"learning_rate": 8.752136752136753e-06,
|
| 581 |
+
"loss": 0.616,
|
| 582 |
+
"step": 690
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 10.015384615384615,
|
| 586 |
+
"grad_norm": 5.259721755981445,
|
| 587 |
+
"learning_rate": 8.717948717948719e-06,
|
| 588 |
+
"loss": 0.3077,
|
| 589 |
+
"step": 700
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 10.018461538461539,
|
| 593 |
+
"grad_norm": 18.261491775512695,
|
| 594 |
+
"learning_rate": 8.683760683760684e-06,
|
| 595 |
+
"loss": 0.694,
|
| 596 |
+
"step": 710
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 10.02,
|
| 600 |
+
"eval_accuracy": 0.717391304347826,
|
| 601 |
+
"eval_loss": 0.5354219675064087,
|
| 602 |
+
"eval_runtime": 10.1729,
|
| 603 |
+
"eval_samples_per_second": 4.522,
|
| 604 |
+
"eval_steps_per_second": 1.18,
|
| 605 |
+
"step": 715
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 11.00153846153846,
|
| 609 |
+
"grad_norm": 6.640218257904053,
|
| 610 |
+
"learning_rate": 8.649572649572651e-06,
|
| 611 |
+
"loss": 0.5996,
|
| 612 |
+
"step": 720
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 11.004615384615384,
|
| 616 |
+
"grad_norm": 3.7313828468322754,
|
| 617 |
+
"learning_rate": 8.615384615384617e-06,
|
| 618 |
+
"loss": 0.2938,
|
| 619 |
+
"step": 730
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 11.007692307692308,
|
| 623 |
+
"grad_norm": 2.5741147994995117,
|
| 624 |
+
"learning_rate": 8.58119658119658e-06,
|
| 625 |
+
"loss": 0.559,
|
| 626 |
+
"step": 740
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 11.010769230769231,
|
| 630 |
+
"grad_norm": 4.06730318069458,
|
| 631 |
+
"learning_rate": 8.547008547008548e-06,
|
| 632 |
+
"loss": 0.3129,
|
| 633 |
+
"step": 750
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 11.013846153846155,
|
| 637 |
+
"grad_norm": 14.723116874694824,
|
| 638 |
+
"learning_rate": 8.512820512820513e-06,
|
| 639 |
+
"loss": 0.5298,
|
| 640 |
+
"step": 760
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 11.016923076923076,
|
| 644 |
+
"grad_norm": 11.420027732849121,
|
| 645 |
+
"learning_rate": 8.478632478632479e-06,
|
| 646 |
+
"loss": 0.4968,
|
| 647 |
+
"step": 770
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 11.02,
|
| 651 |
+
"grad_norm": 12.085827827453613,
|
| 652 |
+
"learning_rate": 8.444444444444446e-06,
|
| 653 |
+
"loss": 0.4773,
|
| 654 |
+
"step": 780
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 11.02,
|
| 658 |
+
"eval_accuracy": 0.6956521739130435,
|
| 659 |
+
"eval_loss": 0.6181442141532898,
|
| 660 |
+
"eval_runtime": 10.1957,
|
| 661 |
+
"eval_samples_per_second": 4.512,
|
| 662 |
+
"eval_steps_per_second": 1.177,
|
| 663 |
+
"step": 780
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 12.003076923076923,
|
| 667 |
+
"grad_norm": 27.175369262695312,
|
| 668 |
+
"learning_rate": 8.410256410256411e-06,
|
| 669 |
+
"loss": 0.37,
|
| 670 |
+
"step": 790
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 12.006153846153847,
|
| 674 |
+
"grad_norm": 6.941166400909424,
|
| 675 |
+
"learning_rate": 8.376068376068377e-06,
|
| 676 |
+
"loss": 0.5221,
|
| 677 |
+
"step": 800
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 12.009230769230768,
|
| 681 |
+
"grad_norm": 17.430849075317383,
|
| 682 |
+
"learning_rate": 8.341880341880344e-06,
|
| 683 |
+
"loss": 0.3823,
|
| 684 |
+
"step": 810
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 12.012307692307692,
|
| 688 |
+
"grad_norm": 0.49046531319618225,
|
| 689 |
+
"learning_rate": 8.307692307692309e-06,
|
| 690 |
+
"loss": 0.5028,
|
| 691 |
+
"step": 820
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 12.015384615384615,
|
| 695 |
+
"grad_norm": 15.024535179138184,
|
| 696 |
+
"learning_rate": 8.273504273504273e-06,
|
| 697 |
+
"loss": 0.5115,
|
| 698 |
+
"step": 830
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 12.018461538461539,
|
| 702 |
+
"grad_norm": 14.742912292480469,
|
| 703 |
+
"learning_rate": 8.23931623931624e-06,
|
| 704 |
+
"loss": 0.6208,
|
| 705 |
+
"step": 840
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 12.02,
|
| 709 |
+
"eval_accuracy": 0.7608695652173914,
|
| 710 |
+
"eval_loss": 0.567747950553894,
|
| 711 |
+
"eval_runtime": 10.3753,
|
| 712 |
+
"eval_samples_per_second": 4.434,
|
| 713 |
+
"eval_steps_per_second": 1.157,
|
| 714 |
+
"step": 845
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 13.00153846153846,
|
| 718 |
+
"grad_norm": 15.91428279876709,
|
| 719 |
+
"learning_rate": 8.205128205128205e-06,
|
| 720 |
+
"loss": 0.4844,
|
| 721 |
+
"step": 850
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 13.004615384615384,
|
| 725 |
+
"grad_norm": 0.08476120978593826,
|
| 726 |
+
"learning_rate": 8.17094017094017e-06,
|
| 727 |
+
"loss": 0.2531,
|
| 728 |
+
"step": 860
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 13.007692307692308,
|
| 732 |
+
"grad_norm": 59.409690856933594,
|
| 733 |
+
"learning_rate": 8.136752136752138e-06,
|
| 734 |
+
"loss": 0.6488,
|
| 735 |
+
"step": 870
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 13.010769230769231,
|
| 739 |
+
"grad_norm": 3.901467800140381,
|
| 740 |
+
"learning_rate": 8.102564102564103e-06,
|
| 741 |
+
"loss": 0.8015,
|
| 742 |
+
"step": 880
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 13.013846153846155,
|
| 746 |
+
"grad_norm": 6.678008556365967,
|
| 747 |
+
"learning_rate": 8.068376068376069e-06,
|
| 748 |
+
"loss": 0.5357,
|
| 749 |
+
"step": 890
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 13.016923076923076,
|
| 753 |
+
"grad_norm": 7.264644622802734,
|
| 754 |
+
"learning_rate": 8.034188034188036e-06,
|
| 755 |
+
"loss": 0.4998,
|
| 756 |
+
"step": 900
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 13.02,
|
| 760 |
+
"grad_norm": 22.3366756439209,
|
| 761 |
+
"learning_rate": 8.000000000000001e-06,
|
| 762 |
+
"loss": 0.344,
|
| 763 |
+
"step": 910
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 13.02,
|
| 767 |
+
"eval_accuracy": 0.6086956521739131,
|
| 768 |
+
"eval_loss": 0.7451769113540649,
|
| 769 |
+
"eval_runtime": 10.4561,
|
| 770 |
+
"eval_samples_per_second": 4.399,
|
| 771 |
+
"eval_steps_per_second": 1.148,
|
| 772 |
+
"step": 910
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 14.003076923076923,
|
| 776 |
+
"grad_norm": 17.2800350189209,
|
| 777 |
+
"learning_rate": 7.965811965811967e-06,
|
| 778 |
+
"loss": 0.3782,
|
| 779 |
+
"step": 920
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 14.006153846153847,
|
| 783 |
+
"grad_norm": 2.800379753112793,
|
| 784 |
+
"learning_rate": 7.931623931623932e-06,
|
| 785 |
+
"loss": 0.7975,
|
| 786 |
+
"step": 930
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 14.009230769230768,
|
| 790 |
+
"grad_norm": 14.407361030578613,
|
| 791 |
+
"learning_rate": 7.897435897435898e-06,
|
| 792 |
+
"loss": 0.4081,
|
| 793 |
+
"step": 940
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 14.012307692307692,
|
| 797 |
+
"grad_norm": 82.17062377929688,
|
| 798 |
+
"learning_rate": 7.863247863247863e-06,
|
| 799 |
+
"loss": 0.3227,
|
| 800 |
+
"step": 950
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 14.015384615384615,
|
| 804 |
+
"grad_norm": 20.13490867614746,
|
| 805 |
+
"learning_rate": 7.82905982905983e-06,
|
| 806 |
+
"loss": 0.3738,
|
| 807 |
+
"step": 960
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 14.018461538461539,
|
| 811 |
+
"grad_norm": 30.13092041015625,
|
| 812 |
+
"learning_rate": 7.794871794871796e-06,
|
| 813 |
+
"loss": 0.254,
|
| 814 |
+
"step": 970
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 14.02,
|
| 818 |
+
"eval_accuracy": 0.7391304347826086,
|
| 819 |
+
"eval_loss": 0.6362163424491882,
|
| 820 |
+
"eval_runtime": 10.2582,
|
| 821 |
+
"eval_samples_per_second": 4.484,
|
| 822 |
+
"eval_steps_per_second": 1.17,
|
| 823 |
+
"step": 975
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 15.00153846153846,
|
| 827 |
+
"grad_norm": 0.13770069181919098,
|
| 828 |
+
"learning_rate": 7.760683760683761e-06,
|
| 829 |
+
"loss": 0.4847,
|
| 830 |
+
"step": 980
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 15.004615384615384,
|
| 834 |
+
"grad_norm": 17.021846771240234,
|
| 835 |
+
"learning_rate": 7.726495726495728e-06,
|
| 836 |
+
"loss": 0.2376,
|
| 837 |
+
"step": 990
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 15.007692307692308,
|
| 841 |
+
"grad_norm": 16.193130493164062,
|
| 842 |
+
"learning_rate": 7.692307692307694e-06,
|
| 843 |
+
"loss": 0.5756,
|
| 844 |
+
"step": 1000
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 15.010769230769231,
|
| 848 |
+
"grad_norm": 17.448122024536133,
|
| 849 |
+
"learning_rate": 7.658119658119659e-06,
|
| 850 |
+
"loss": 0.5613,
|
| 851 |
+
"step": 1010
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 15.013846153846155,
|
| 855 |
+
"grad_norm": 27.816417694091797,
|
| 856 |
+
"learning_rate": 7.6239316239316244e-06,
|
| 857 |
+
"loss": 0.3936,
|
| 858 |
+
"step": 1020
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 15.016923076923076,
|
| 862 |
+
"grad_norm": 9.32864761352539,
|
| 863 |
+
"learning_rate": 7.58974358974359e-06,
|
| 864 |
+
"loss": 0.5214,
|
| 865 |
+
"step": 1030
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 15.02,
|
| 869 |
+
"grad_norm": 10.375736236572266,
|
| 870 |
+
"learning_rate": 7.555555555555556e-06,
|
| 871 |
+
"loss": 0.4578,
|
| 872 |
+
"step": 1040
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 15.02,
|
| 876 |
+
"eval_accuracy": 0.6956521739130435,
|
| 877 |
+
"eval_loss": 0.8304260969161987,
|
| 878 |
+
"eval_runtime": 10.454,
|
| 879 |
+
"eval_samples_per_second": 4.4,
|
| 880 |
+
"eval_steps_per_second": 1.148,
|
| 881 |
+
"step": 1040
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 16.00307692307692,
|
| 885 |
+
"grad_norm": 32.06553268432617,
|
| 886 |
+
"learning_rate": 7.521367521367522e-06,
|
| 887 |
+
"loss": 0.4237,
|
| 888 |
+
"step": 1050
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 16.006153846153847,
|
| 892 |
+
"grad_norm": 3.9894378185272217,
|
| 893 |
+
"learning_rate": 7.487179487179488e-06,
|
| 894 |
+
"loss": 0.3599,
|
| 895 |
+
"step": 1060
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 16.00923076923077,
|
| 899 |
+
"grad_norm": 0.25729474425315857,
|
| 900 |
+
"learning_rate": 7.452991452991454e-06,
|
| 901 |
+
"loss": 0.3732,
|
| 902 |
+
"step": 1070
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 16.012307692307694,
|
| 906 |
+
"grad_norm": 16.2904109954834,
|
| 907 |
+
"learning_rate": 7.4188034188034196e-06,
|
| 908 |
+
"loss": 0.3938,
|
| 909 |
+
"step": 1080
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 16.015384615384615,
|
| 913 |
+
"grad_norm": 10.762846946716309,
|
| 914 |
+
"learning_rate": 7.384615384615386e-06,
|
| 915 |
+
"loss": 0.312,
|
| 916 |
+
"step": 1090
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 16.018461538461537,
|
| 920 |
+
"grad_norm": 20.59312629699707,
|
| 921 |
+
"learning_rate": 7.350427350427351e-06,
|
| 922 |
+
"loss": 0.3954,
|
| 923 |
+
"step": 1100
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 16.02,
|
| 927 |
+
"eval_accuracy": 0.7608695652173914,
|
| 928 |
+
"eval_loss": 0.6048559546470642,
|
| 929 |
+
"eval_runtime": 10.7124,
|
| 930 |
+
"eval_samples_per_second": 4.294,
|
| 931 |
+
"eval_steps_per_second": 1.12,
|
| 932 |
+
"step": 1105
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 17.001538461538463,
|
| 936 |
+
"grad_norm": 11.256546020507812,
|
| 937 |
+
"learning_rate": 7.316239316239317e-06,
|
| 938 |
+
"loss": 0.4876,
|
| 939 |
+
"step": 1110
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 17.004615384615384,
|
| 943 |
+
"grad_norm": 42.64373016357422,
|
| 944 |
+
"learning_rate": 7.282051282051282e-06,
|
| 945 |
+
"loss": 0.1782,
|
| 946 |
+
"step": 1120
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 17.00769230769231,
|
| 950 |
+
"grad_norm": 34.2502326965332,
|
| 951 |
+
"learning_rate": 7.247863247863248e-06,
|
| 952 |
+
"loss": 0.4987,
|
| 953 |
+
"step": 1130
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 17.01076923076923,
|
| 957 |
+
"grad_norm": 9.817243576049805,
|
| 958 |
+
"learning_rate": 7.213675213675214e-06,
|
| 959 |
+
"loss": 0.2962,
|
| 960 |
+
"step": 1140
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 17.013846153846153,
|
| 964 |
+
"grad_norm": 39.85562515258789,
|
| 965 |
+
"learning_rate": 7.17948717948718e-06,
|
| 966 |
+
"loss": 0.4634,
|
| 967 |
+
"step": 1150
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 17.016923076923078,
|
| 971 |
+
"grad_norm": 32.21989059448242,
|
| 972 |
+
"learning_rate": 7.145299145299146e-06,
|
| 973 |
+
"loss": 0.2256,
|
| 974 |
+
"step": 1160
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 17.02,
|
| 978 |
+
"grad_norm": 32.11091232299805,
|
| 979 |
+
"learning_rate": 7.111111111111112e-06,
|
| 980 |
+
"loss": 0.248,
|
| 981 |
+
"step": 1170
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 17.02,
|
| 985 |
+
"eval_accuracy": 0.6739130434782609,
|
| 986 |
+
"eval_loss": 0.9506217241287231,
|
| 987 |
+
"eval_runtime": 10.8727,
|
| 988 |
+
"eval_samples_per_second": 4.231,
|
| 989 |
+
"eval_steps_per_second": 1.104,
|
| 990 |
+
"step": 1170
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 18.00307692307692,
|
| 994 |
+
"grad_norm": 48.11212158203125,
|
| 995 |
+
"learning_rate": 7.076923076923078e-06,
|
| 996 |
+
"loss": 0.7942,
|
| 997 |
+
"step": 1180
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 18.006153846153847,
|
| 1001 |
+
"grad_norm": 39.191593170166016,
|
| 1002 |
+
"learning_rate": 7.0427350427350435e-06,
|
| 1003 |
+
"loss": 0.4622,
|
| 1004 |
+
"step": 1190
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 18.00923076923077,
|
| 1008 |
+
"grad_norm": 27.27468490600586,
|
| 1009 |
+
"learning_rate": 7.008547008547009e-06,
|
| 1010 |
+
"loss": 0.4094,
|
| 1011 |
+
"step": 1200
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 18.012307692307694,
|
| 1015 |
+
"grad_norm": 13.529548645019531,
|
| 1016 |
+
"learning_rate": 6.974358974358974e-06,
|
| 1017 |
+
"loss": 0.137,
|
| 1018 |
+
"step": 1210
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 18.015384615384615,
|
| 1022 |
+
"grad_norm": 2.5885872840881348,
|
| 1023 |
+
"learning_rate": 6.940170940170941e-06,
|
| 1024 |
+
"loss": 0.3264,
|
| 1025 |
+
"step": 1220
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 18.018461538461537,
|
| 1029 |
+
"grad_norm": 51.945762634277344,
|
| 1030 |
+
"learning_rate": 6.905982905982906e-06,
|
| 1031 |
+
"loss": 0.1334,
|
| 1032 |
+
"step": 1230
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 18.02,
|
| 1036 |
+
"eval_accuracy": 0.6739130434782609,
|
| 1037 |
+
"eval_loss": 1.1876178979873657,
|
| 1038 |
+
"eval_runtime": 10.6484,
|
| 1039 |
+
"eval_samples_per_second": 4.32,
|
| 1040 |
+
"eval_steps_per_second": 1.127,
|
| 1041 |
+
"step": 1235
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 19.001538461538463,
|
| 1045 |
+
"grad_norm": 0.604591965675354,
|
| 1046 |
+
"learning_rate": 6.871794871794872e-06,
|
| 1047 |
+
"loss": 0.1617,
|
| 1048 |
+
"step": 1240
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 19.004615384615384,
|
| 1052 |
+
"grad_norm": 65.89838409423828,
|
| 1053 |
+
"learning_rate": 6.837606837606839e-06,
|
| 1054 |
+
"loss": 0.5149,
|
| 1055 |
+
"step": 1250
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 19.00769230769231,
|
| 1059 |
+
"grad_norm": 103.18575286865234,
|
| 1060 |
+
"learning_rate": 6.803418803418804e-06,
|
| 1061 |
+
"loss": 0.299,
|
| 1062 |
+
"step": 1260
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 19.01076923076923,
|
| 1066 |
+
"grad_norm": 0.5941734313964844,
|
| 1067 |
+
"learning_rate": 6.76923076923077e-06,
|
| 1068 |
+
"loss": 0.5991,
|
| 1069 |
+
"step": 1270
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 19.013846153846153,
|
| 1073 |
+
"grad_norm": 22.712234497070312,
|
| 1074 |
+
"learning_rate": 6.735042735042736e-06,
|
| 1075 |
+
"loss": 0.3312,
|
| 1076 |
+
"step": 1280
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 19.016923076923078,
|
| 1080 |
+
"grad_norm": 0.26104092597961426,
|
| 1081 |
+
"learning_rate": 6.700854700854701e-06,
|
| 1082 |
+
"loss": 0.3076,
|
| 1083 |
+
"step": 1290
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 19.02,
|
| 1087 |
+
"grad_norm": 58.68645477294922,
|
| 1088 |
+
"learning_rate": 6.666666666666667e-06,
|
| 1089 |
+
"loss": 0.534,
|
| 1090 |
+
"step": 1300
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 19.02,
|
| 1094 |
+
"eval_accuracy": 0.7391304347826086,
|
| 1095 |
+
"eval_loss": 0.6295628547668457,
|
| 1096 |
+
"eval_runtime": 11.1938,
|
| 1097 |
+
"eval_samples_per_second": 4.109,
|
| 1098 |
+
"eval_steps_per_second": 1.072,
|
| 1099 |
+
"step": 1300
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 20.00307692307692,
|
| 1103 |
+
"grad_norm": 0.13804925978183746,
|
| 1104 |
+
"learning_rate": 6.632478632478633e-06,
|
| 1105 |
+
"loss": 0.4236,
|
| 1106 |
+
"step": 1310
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 20.006153846153847,
|
| 1110 |
+
"grad_norm": 0.2732415497303009,
|
| 1111 |
+
"learning_rate": 6.598290598290598e-06,
|
| 1112 |
+
"loss": 0.2337,
|
| 1113 |
+
"step": 1320
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 20.00923076923077,
|
| 1117 |
+
"grad_norm": 0.22181116044521332,
|
| 1118 |
+
"learning_rate": 6.564102564102565e-06,
|
| 1119 |
+
"loss": 0.2577,
|
| 1120 |
+
"step": 1330
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 20.012307692307694,
|
| 1124 |
+
"grad_norm": 20.500429153442383,
|
| 1125 |
+
"learning_rate": 6.529914529914531e-06,
|
| 1126 |
+
"loss": 0.3988,
|
| 1127 |
+
"step": 1340
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 20.015384615384615,
|
| 1131 |
+
"grad_norm": 0.25182831287384033,
|
| 1132 |
+
"learning_rate": 6.495726495726496e-06,
|
| 1133 |
+
"loss": 0.3886,
|
| 1134 |
+
"step": 1350
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 20.018461538461537,
|
| 1138 |
+
"grad_norm": 29.723634719848633,
|
| 1139 |
+
"learning_rate": 6.461538461538463e-06,
|
| 1140 |
+
"loss": 0.3556,
|
| 1141 |
+
"step": 1360
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 20.02,
|
| 1145 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1146 |
+
"eval_loss": 1.3006523847579956,
|
| 1147 |
+
"eval_runtime": 10.3551,
|
| 1148 |
+
"eval_samples_per_second": 4.442,
|
| 1149 |
+
"eval_steps_per_second": 1.159,
|
| 1150 |
+
"step": 1365
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 21.001538461538463,
|
| 1154 |
+
"grad_norm": 30.7800350189209,
|
| 1155 |
+
"learning_rate": 6.427350427350428e-06,
|
| 1156 |
+
"loss": 0.6987,
|
| 1157 |
+
"step": 1370
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 21.004615384615384,
|
| 1161 |
+
"grad_norm": 13.209580421447754,
|
| 1162 |
+
"learning_rate": 6.3931623931623935e-06,
|
| 1163 |
+
"loss": 0.1747,
|
| 1164 |
+
"step": 1380
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 21.00769230769231,
|
| 1168 |
+
"grad_norm": 74.4799575805664,
|
| 1169 |
+
"learning_rate": 6.358974358974359e-06,
|
| 1170 |
+
"loss": 0.2258,
|
| 1171 |
+
"step": 1390
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 21.01076923076923,
|
| 1175 |
+
"grad_norm": 87.08421325683594,
|
| 1176 |
+
"learning_rate": 6.324786324786325e-06,
|
| 1177 |
+
"loss": 0.3794,
|
| 1178 |
+
"step": 1400
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 21.013846153846153,
|
| 1182 |
+
"grad_norm": 35.24946594238281,
|
| 1183 |
+
"learning_rate": 6.290598290598291e-06,
|
| 1184 |
+
"loss": 0.5507,
|
| 1185 |
+
"step": 1410
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 21.016923076923078,
|
| 1189 |
+
"grad_norm": 0.3907718360424042,
|
| 1190 |
+
"learning_rate": 6.256410256410257e-06,
|
| 1191 |
+
"loss": 0.3903,
|
| 1192 |
+
"step": 1420
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 21.02,
|
| 1196 |
+
"grad_norm": 67.88304901123047,
|
| 1197 |
+
"learning_rate": 6.222222222222223e-06,
|
| 1198 |
+
"loss": 0.5439,
|
| 1199 |
+
"step": 1430
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 21.02,
|
| 1203 |
+
"eval_accuracy": 0.6739130434782609,
|
| 1204 |
+
"eval_loss": 1.5065968036651611,
|
| 1205 |
+
"eval_runtime": 10.2377,
|
| 1206 |
+
"eval_samples_per_second": 4.493,
|
| 1207 |
+
"eval_steps_per_second": 1.172,
|
| 1208 |
+
"step": 1430
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 22.00307692307692,
|
| 1212 |
+
"grad_norm": 0.019926143810153008,
|
| 1213 |
+
"learning_rate": 6.188034188034189e-06,
|
| 1214 |
+
"loss": 0.4587,
|
| 1215 |
+
"step": 1440
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 22.006153846153847,
|
| 1219 |
+
"grad_norm": 61.04457092285156,
|
| 1220 |
+
"learning_rate": 6.153846153846155e-06,
|
| 1221 |
+
"loss": 0.0936,
|
| 1222 |
+
"step": 1450
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 22.00923076923077,
|
| 1226 |
+
"grad_norm": 69.38724517822266,
|
| 1227 |
+
"learning_rate": 6.11965811965812e-06,
|
| 1228 |
+
"loss": 0.394,
|
| 1229 |
+
"step": 1460
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 22.012307692307694,
|
| 1233 |
+
"grad_norm": 58.5753173828125,
|
| 1234 |
+
"learning_rate": 6.085470085470086e-06,
|
| 1235 |
+
"loss": 0.2639,
|
| 1236 |
+
"step": 1470
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 22.015384615384615,
|
| 1240 |
+
"grad_norm": 0.5316572785377502,
|
| 1241 |
+
"learning_rate": 6.051282051282051e-06,
|
| 1242 |
+
"loss": 0.476,
|
| 1243 |
+
"step": 1480
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 22.018461538461537,
|
| 1247 |
+
"grad_norm": 29.042409896850586,
|
| 1248 |
+
"learning_rate": 6.0170940170940174e-06,
|
| 1249 |
+
"loss": 0.4107,
|
| 1250 |
+
"step": 1490
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 22.02,
|
| 1254 |
+
"eval_accuracy": 0.8043478260869565,
|
| 1255 |
+
"eval_loss": 0.9273331761360168,
|
| 1256 |
+
"eval_runtime": 10.2871,
|
| 1257 |
+
"eval_samples_per_second": 4.472,
|
| 1258 |
+
"eval_steps_per_second": 1.167,
|
| 1259 |
+
"step": 1495
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 23.001538461538463,
|
| 1263 |
+
"grad_norm": 0.7934656143188477,
|
| 1264 |
+
"learning_rate": 5.982905982905983e-06,
|
| 1265 |
+
"loss": 0.4271,
|
| 1266 |
+
"step": 1500
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 23.004615384615384,
|
| 1270 |
+
"grad_norm": 15.743200302124023,
|
| 1271 |
+
"learning_rate": 5.948717948717949e-06,
|
| 1272 |
+
"loss": 0.2889,
|
| 1273 |
+
"step": 1510
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 23.00769230769231,
|
| 1277 |
+
"grad_norm": 0.05252040550112724,
|
| 1278 |
+
"learning_rate": 5.914529914529915e-06,
|
| 1279 |
+
"loss": 0.1282,
|
| 1280 |
+
"step": 1520
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 23.01076923076923,
|
| 1284 |
+
"grad_norm": 27.15765953063965,
|
| 1285 |
+
"learning_rate": 5.880341880341881e-06,
|
| 1286 |
+
"loss": 0.455,
|
| 1287 |
+
"step": 1530
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 23.013846153846153,
|
| 1291 |
+
"grad_norm": 154.0335235595703,
|
| 1292 |
+
"learning_rate": 5.846153846153847e-06,
|
| 1293 |
+
"loss": 0.5051,
|
| 1294 |
+
"step": 1540
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 23.016923076923078,
|
| 1298 |
+
"grad_norm": 1.5758213996887207,
|
| 1299 |
+
"learning_rate": 5.8119658119658126e-06,
|
| 1300 |
+
"loss": 0.4148,
|
| 1301 |
+
"step": 1550
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 23.02,
|
| 1305 |
+
"grad_norm": 46.614864349365234,
|
| 1306 |
+
"learning_rate": 5.777777777777778e-06,
|
| 1307 |
+
"loss": 0.61,
|
| 1308 |
+
"step": 1560
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 23.02,
|
| 1312 |
+
"eval_accuracy": 0.717391304347826,
|
| 1313 |
+
"eval_loss": 1.00082528591156,
|
| 1314 |
+
"eval_runtime": 10.3956,
|
| 1315 |
+
"eval_samples_per_second": 4.425,
|
| 1316 |
+
"eval_steps_per_second": 1.154,
|
| 1317 |
+
"step": 1560
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 24.00307692307692,
|
| 1321 |
+
"grad_norm": 1.2954469919204712,
|
| 1322 |
+
"learning_rate": 5.743589743589743e-06,
|
| 1323 |
+
"loss": 0.3196,
|
| 1324 |
+
"step": 1570
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 24.006153846153847,
|
| 1328 |
+
"grad_norm": 0.4691097140312195,
|
| 1329 |
+
"learning_rate": 5.70940170940171e-06,
|
| 1330 |
+
"loss": 0.1125,
|
| 1331 |
+
"step": 1580
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 24.00923076923077,
|
| 1335 |
+
"grad_norm": 66.216552734375,
|
| 1336 |
+
"learning_rate": 5.675213675213675e-06,
|
| 1337 |
+
"loss": 0.3891,
|
| 1338 |
+
"step": 1590
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 24.012307692307694,
|
| 1342 |
+
"grad_norm": 0.1557655781507492,
|
| 1343 |
+
"learning_rate": 5.641025641025641e-06,
|
| 1344 |
+
"loss": 0.03,
|
| 1345 |
+
"step": 1600
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 24.015384615384615,
|
| 1349 |
+
"grad_norm": 2.08573842048645,
|
| 1350 |
+
"learning_rate": 5.606837606837608e-06,
|
| 1351 |
+
"loss": 0.3544,
|
| 1352 |
+
"step": 1610
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 24.018461538461537,
|
| 1356 |
+
"grad_norm": 2.804753303527832,
|
| 1357 |
+
"learning_rate": 5.572649572649573e-06,
|
| 1358 |
+
"loss": 0.6482,
|
| 1359 |
+
"step": 1620
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 24.02,
|
| 1363 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1364 |
+
"eval_loss": 0.7548108100891113,
|
| 1365 |
+
"eval_runtime": 10.8151,
|
| 1366 |
+
"eval_samples_per_second": 4.253,
|
| 1367 |
+
"eval_steps_per_second": 1.11,
|
| 1368 |
+
"step": 1625
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 25.001538461538463,
|
| 1372 |
+
"grad_norm": 83.31999969482422,
|
| 1373 |
+
"learning_rate": 5.538461538461539e-06,
|
| 1374 |
+
"loss": 0.2905,
|
| 1375 |
+
"step": 1630
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 25.004615384615384,
|
| 1379 |
+
"grad_norm": 23.389434814453125,
|
| 1380 |
+
"learning_rate": 5.504273504273505e-06,
|
| 1381 |
+
"loss": 0.438,
|
| 1382 |
+
"step": 1640
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 25.00769230769231,
|
| 1386 |
+
"grad_norm": 34.89247512817383,
|
| 1387 |
+
"learning_rate": 5.470085470085471e-06,
|
| 1388 |
+
"loss": 0.0759,
|
| 1389 |
+
"step": 1650
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 25.01076923076923,
|
| 1393 |
+
"grad_norm": 0.49820569157600403,
|
| 1394 |
+
"learning_rate": 5.435897435897436e-06,
|
| 1395 |
+
"loss": 0.1849,
|
| 1396 |
+
"step": 1660
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 25.013846153846153,
|
| 1400 |
+
"grad_norm": 0.12784357368946075,
|
| 1401 |
+
"learning_rate": 5.401709401709402e-06,
|
| 1402 |
+
"loss": 0.259,
|
| 1403 |
+
"step": 1670
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 25.016923076923078,
|
| 1407 |
+
"grad_norm": 69.064208984375,
|
| 1408 |
+
"learning_rate": 5.367521367521367e-06,
|
| 1409 |
+
"loss": 0.3631,
|
| 1410 |
+
"step": 1680
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 25.02,
|
| 1414 |
+
"grad_norm": 0.5186673402786255,
|
| 1415 |
+
"learning_rate": 5.333333333333334e-06,
|
| 1416 |
+
"loss": 0.199,
|
| 1417 |
+
"step": 1690
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 25.02,
|
| 1421 |
+
"eval_accuracy": 0.782608695652174,
|
| 1422 |
+
"eval_loss": 0.7916727662086487,
|
| 1423 |
+
"eval_runtime": 10.4377,
|
| 1424 |
+
"eval_samples_per_second": 4.407,
|
| 1425 |
+
"eval_steps_per_second": 1.15,
|
| 1426 |
+
"step": 1690
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 26.00307692307692,
|
| 1430 |
+
"grad_norm": 0.019262276589870453,
|
| 1431 |
+
"learning_rate": 5.2991452991453e-06,
|
| 1432 |
+
"loss": 0.2937,
|
| 1433 |
+
"step": 1700
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 26.006153846153847,
|
| 1437 |
+
"grad_norm": 0.6572211384773254,
|
| 1438 |
+
"learning_rate": 5.264957264957265e-06,
|
| 1439 |
+
"loss": 0.1816,
|
| 1440 |
+
"step": 1710
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 26.00923076923077,
|
| 1444 |
+
"grad_norm": 52.68994903564453,
|
| 1445 |
+
"learning_rate": 5.230769230769232e-06,
|
| 1446 |
+
"loss": 0.4716,
|
| 1447 |
+
"step": 1720
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 26.012307692307694,
|
| 1451 |
+
"grad_norm": 0.405513733625412,
|
| 1452 |
+
"learning_rate": 5.196581196581197e-06,
|
| 1453 |
+
"loss": 0.4085,
|
| 1454 |
+
"step": 1730
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 26.015384615384615,
|
| 1458 |
+
"grad_norm": 16.66892433166504,
|
| 1459 |
+
"learning_rate": 5.162393162393163e-06,
|
| 1460 |
+
"loss": 0.4931,
|
| 1461 |
+
"step": 1740
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 26.018461538461537,
|
| 1465 |
+
"grad_norm": 0.020136963576078415,
|
| 1466 |
+
"learning_rate": 5.128205128205128e-06,
|
| 1467 |
+
"loss": 0.1185,
|
| 1468 |
+
"step": 1750
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 26.02,
|
| 1472 |
+
"eval_accuracy": 0.782608695652174,
|
| 1473 |
+
"eval_loss": 0.7529156804084778,
|
| 1474 |
+
"eval_runtime": 10.9054,
|
| 1475 |
+
"eval_samples_per_second": 4.218,
|
| 1476 |
+
"eval_steps_per_second": 1.1,
|
| 1477 |
+
"step": 1755
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 27.001538461538463,
|
| 1481 |
+
"grad_norm": 29.928010940551758,
|
| 1482 |
+
"learning_rate": 5.094017094017094e-06,
|
| 1483 |
+
"loss": 0.2469,
|
| 1484 |
+
"step": 1760
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 27.004615384615384,
|
| 1488 |
+
"grad_norm": 19.83055877685547,
|
| 1489 |
+
"learning_rate": 5.05982905982906e-06,
|
| 1490 |
+
"loss": 0.2104,
|
| 1491 |
+
"step": 1770
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 27.00769230769231,
|
| 1495 |
+
"grad_norm": 0.28511443734169006,
|
| 1496 |
+
"learning_rate": 5.025641025641026e-06,
|
| 1497 |
+
"loss": 0.2268,
|
| 1498 |
+
"step": 1780
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 27.01076923076923,
|
| 1502 |
+
"grad_norm": 181.45603942871094,
|
| 1503 |
+
"learning_rate": 4.991452991452992e-06,
|
| 1504 |
+
"loss": 0.1389,
|
| 1505 |
+
"step": 1790
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 27.013846153846153,
|
| 1509 |
+
"grad_norm": 64.7826156616211,
|
| 1510 |
+
"learning_rate": 4.957264957264958e-06,
|
| 1511 |
+
"loss": 0.1211,
|
| 1512 |
+
"step": 1800
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 27.016923076923078,
|
| 1516 |
+
"grad_norm": 4.411494255065918,
|
| 1517 |
+
"learning_rate": 4.923076923076924e-06,
|
| 1518 |
+
"loss": 0.1387,
|
| 1519 |
+
"step": 1810
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 27.02,
|
| 1523 |
+
"grad_norm": 11.255193710327148,
|
| 1524 |
+
"learning_rate": 4.888888888888889e-06,
|
| 1525 |
+
"loss": 0.3886,
|
| 1526 |
+
"step": 1820
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 27.02,
|
| 1530 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1531 |
+
"eval_loss": 0.8627105355262756,
|
| 1532 |
+
"eval_runtime": 10.4057,
|
| 1533 |
+
"eval_samples_per_second": 4.421,
|
| 1534 |
+
"eval_steps_per_second": 1.153,
|
| 1535 |
+
"step": 1820
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 28.00307692307692,
|
| 1539 |
+
"grad_norm": 92.7336196899414,
|
| 1540 |
+
"learning_rate": 4.854700854700855e-06,
|
| 1541 |
+
"loss": 0.0313,
|
| 1542 |
+
"step": 1830
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 28.006153846153847,
|
| 1546 |
+
"grad_norm": 21.0554256439209,
|
| 1547 |
+
"learning_rate": 4.820512820512821e-06,
|
| 1548 |
+
"loss": 0.0353,
|
| 1549 |
+
"step": 1840
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 28.00923076923077,
|
| 1553 |
+
"grad_norm": 141.20700073242188,
|
| 1554 |
+
"learning_rate": 4.786324786324787e-06,
|
| 1555 |
+
"loss": 0.5025,
|
| 1556 |
+
"step": 1850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 28.012307692307694,
|
| 1560 |
+
"grad_norm": 0.3742455542087555,
|
| 1561 |
+
"learning_rate": 4.752136752136752e-06,
|
| 1562 |
+
"loss": 0.1276,
|
| 1563 |
+
"step": 1860
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 28.015384615384615,
|
| 1567 |
+
"grad_norm": 0.04171011224389076,
|
| 1568 |
+
"learning_rate": 4.717948717948718e-06,
|
| 1569 |
+
"loss": 0.0059,
|
| 1570 |
+
"step": 1870
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 28.018461538461537,
|
| 1574 |
+
"grad_norm": 2.4367685317993164,
|
| 1575 |
+
"learning_rate": 4.6837606837606844e-06,
|
| 1576 |
+
"loss": 0.0123,
|
| 1577 |
+
"step": 1880
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 28.02,
|
| 1581 |
+
"eval_accuracy": 0.717391304347826,
|
| 1582 |
+
"eval_loss": 1.3886178731918335,
|
| 1583 |
+
"eval_runtime": 10.496,
|
| 1584 |
+
"eval_samples_per_second": 4.383,
|
| 1585 |
+
"eval_steps_per_second": 1.143,
|
| 1586 |
+
"step": 1885
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 29.001538461538463,
|
| 1590 |
+
"grad_norm": 223.50958251953125,
|
| 1591 |
+
"learning_rate": 4.64957264957265e-06,
|
| 1592 |
+
"loss": 0.1912,
|
| 1593 |
+
"step": 1890
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 29.004615384615384,
|
| 1597 |
+
"grad_norm": 0.6573296189308167,
|
| 1598 |
+
"learning_rate": 4.615384615384616e-06,
|
| 1599 |
+
"loss": 0.2606,
|
| 1600 |
+
"step": 1900
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 29.00769230769231,
|
| 1604 |
+
"grad_norm": 33.679935455322266,
|
| 1605 |
+
"learning_rate": 4.581196581196582e-06,
|
| 1606 |
+
"loss": 0.1504,
|
| 1607 |
+
"step": 1910
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 29.01076923076923,
|
| 1611 |
+
"grad_norm": 36.40077209472656,
|
| 1612 |
+
"learning_rate": 4.547008547008547e-06,
|
| 1613 |
+
"loss": 0.1491,
|
| 1614 |
+
"step": 1920
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 29.013846153846153,
|
| 1618 |
+
"grad_norm": 34.2191047668457,
|
| 1619 |
+
"learning_rate": 4.512820512820513e-06,
|
| 1620 |
+
"loss": 0.2882,
|
| 1621 |
+
"step": 1930
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 29.016923076923078,
|
| 1625 |
+
"grad_norm": 0.30137479305267334,
|
| 1626 |
+
"learning_rate": 4.4786324786324796e-06,
|
| 1627 |
+
"loss": 0.2504,
|
| 1628 |
+
"step": 1940
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 29.02,
|
| 1632 |
+
"grad_norm": 0.040065910667181015,
|
| 1633 |
+
"learning_rate": 4.444444444444444e-06,
|
| 1634 |
+
"loss": 0.5328,
|
| 1635 |
+
"step": 1950
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 29.02,
|
| 1639 |
+
"eval_accuracy": 0.6956521739130435,
|
| 1640 |
+
"eval_loss": 1.2802809476852417,
|
| 1641 |
+
"eval_runtime": 10.2626,
|
| 1642 |
+
"eval_samples_per_second": 4.482,
|
| 1643 |
+
"eval_steps_per_second": 1.169,
|
| 1644 |
+
"step": 1950
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 30.00307692307692,
|
| 1648 |
+
"grad_norm": 0.04048765078186989,
|
| 1649 |
+
"learning_rate": 4.4102564102564104e-06,
|
| 1650 |
+
"loss": 0.0121,
|
| 1651 |
+
"step": 1960
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 30.006153846153847,
|
| 1655 |
+
"grad_norm": 198.26219177246094,
|
| 1656 |
+
"learning_rate": 4.376068376068377e-06,
|
| 1657 |
+
"loss": 0.2916,
|
| 1658 |
+
"step": 1970
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 30.00923076923077,
|
| 1662 |
+
"grad_norm": 0.01023260410875082,
|
| 1663 |
+
"learning_rate": 4.341880341880342e-06,
|
| 1664 |
+
"loss": 0.1678,
|
| 1665 |
+
"step": 1980
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 30.012307692307694,
|
| 1669 |
+
"grad_norm": 0.014997897669672966,
|
| 1670 |
+
"learning_rate": 4.307692307692308e-06,
|
| 1671 |
+
"loss": 0.3871,
|
| 1672 |
+
"step": 1990
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 30.015384615384615,
|
| 1676 |
+
"grad_norm": 0.48902782797813416,
|
| 1677 |
+
"learning_rate": 4.273504273504274e-06,
|
| 1678 |
+
"loss": 0.1938,
|
| 1679 |
+
"step": 2000
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 30.018461538461537,
|
| 1683 |
+
"grad_norm": 0.6367323994636536,
|
| 1684 |
+
"learning_rate": 4.239316239316239e-06,
|
| 1685 |
+
"loss": 0.2961,
|
| 1686 |
+
"step": 2010
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 30.02,
|
| 1690 |
+
"eval_accuracy": 0.717391304347826,
|
| 1691 |
+
"eval_loss": 1.439742922782898,
|
| 1692 |
+
"eval_runtime": 10.3024,
|
| 1693 |
+
"eval_samples_per_second": 4.465,
|
| 1694 |
+
"eval_steps_per_second": 1.165,
|
| 1695 |
+
"step": 2015
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 31.001538461538463,
|
| 1699 |
+
"grad_norm": 1.0408298969268799,
|
| 1700 |
+
"learning_rate": 4.2051282051282055e-06,
|
| 1701 |
+
"loss": 0.0824,
|
| 1702 |
+
"step": 2020
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 31.004615384615384,
|
| 1706 |
+
"grad_norm": 7.177045822143555,
|
| 1707 |
+
"learning_rate": 4.170940170940172e-06,
|
| 1708 |
+
"loss": 0.1443,
|
| 1709 |
+
"step": 2030
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 31.00769230769231,
|
| 1713 |
+
"grad_norm": 0.5223032832145691,
|
| 1714 |
+
"learning_rate": 4.136752136752136e-06,
|
| 1715 |
+
"loss": 0.5326,
|
| 1716 |
+
"step": 2040
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 31.01076923076923,
|
| 1720 |
+
"grad_norm": 0.12590403854846954,
|
| 1721 |
+
"learning_rate": 4.102564102564103e-06,
|
| 1722 |
+
"loss": 0.2971,
|
| 1723 |
+
"step": 2050
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 31.013846153846153,
|
| 1727 |
+
"grad_norm": 0.43242764472961426,
|
| 1728 |
+
"learning_rate": 4.068376068376069e-06,
|
| 1729 |
+
"loss": 0.2301,
|
| 1730 |
+
"step": 2060
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 31.016923076923078,
|
| 1734 |
+
"grad_norm": 0.03505908325314522,
|
| 1735 |
+
"learning_rate": 4.034188034188034e-06,
|
| 1736 |
+
"loss": 0.019,
|
| 1737 |
+
"step": 2070
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 31.02,
|
| 1741 |
+
"grad_norm": 5.052464962005615,
|
| 1742 |
+
"learning_rate": 4.000000000000001e-06,
|
| 1743 |
+
"loss": 0.1192,
|
| 1744 |
+
"step": 2080
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 31.02,
|
| 1748 |
+
"eval_accuracy": 0.6304347826086957,
|
| 1749 |
+
"eval_loss": 2.2563250064849854,
|
| 1750 |
+
"eval_runtime": 10.3617,
|
| 1751 |
+
"eval_samples_per_second": 4.439,
|
| 1752 |
+
"eval_steps_per_second": 1.158,
|
| 1753 |
+
"step": 2080
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 32.003076923076925,
|
| 1757 |
+
"grad_norm": 138.01513671875,
|
| 1758 |
+
"learning_rate": 3.965811965811966e-06,
|
| 1759 |
+
"loss": 0.3484,
|
| 1760 |
+
"step": 2090
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 32.00615384615384,
|
| 1764 |
+
"grad_norm": 0.5178282260894775,
|
| 1765 |
+
"learning_rate": 3.9316239316239315e-06,
|
| 1766 |
+
"loss": 0.0038,
|
| 1767 |
+
"step": 2100
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 32.00923076923077,
|
| 1771 |
+
"grad_norm": 81.31842041015625,
|
| 1772 |
+
"learning_rate": 3.897435897435898e-06,
|
| 1773 |
+
"loss": 0.1732,
|
| 1774 |
+
"step": 2110
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 32.012307692307694,
|
| 1778 |
+
"grad_norm": 0.07390841841697693,
|
| 1779 |
+
"learning_rate": 3.863247863247864e-06,
|
| 1780 |
+
"loss": 0.2652,
|
| 1781 |
+
"step": 2120
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 32.01538461538462,
|
| 1785 |
+
"grad_norm": 71.98724365234375,
|
| 1786 |
+
"learning_rate": 3.8290598290598295e-06,
|
| 1787 |
+
"loss": 0.4317,
|
| 1788 |
+
"step": 2130
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 32.01846153846154,
|
| 1792 |
+
"grad_norm": 18.120813369750977,
|
| 1793 |
+
"learning_rate": 3.794871794871795e-06,
|
| 1794 |
+
"loss": 0.145,
|
| 1795 |
+
"step": 2140
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 32.02,
|
| 1799 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1800 |
+
"eval_loss": 1.0464895963668823,
|
| 1801 |
+
"eval_runtime": 10.3727,
|
| 1802 |
+
"eval_samples_per_second": 4.435,
|
| 1803 |
+
"eval_steps_per_second": 1.157,
|
| 1804 |
+
"step": 2145
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 33.00153846153846,
|
| 1808 |
+
"grad_norm": 0.1161544919013977,
|
| 1809 |
+
"learning_rate": 3.760683760683761e-06,
|
| 1810 |
+
"loss": 0.0155,
|
| 1811 |
+
"step": 2150
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 33.004615384615384,
|
| 1815 |
+
"grad_norm": 0.4286612868309021,
|
| 1816 |
+
"learning_rate": 3.726495726495727e-06,
|
| 1817 |
+
"loss": 0.3714,
|
| 1818 |
+
"step": 2160
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 33.00769230769231,
|
| 1822 |
+
"grad_norm": 0.14987812936306,
|
| 1823 |
+
"learning_rate": 3.692307692307693e-06,
|
| 1824 |
+
"loss": 0.2685,
|
| 1825 |
+
"step": 2170
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 33.01076923076923,
|
| 1829 |
+
"grad_norm": 0.08653810620307922,
|
| 1830 |
+
"learning_rate": 3.6581196581196584e-06,
|
| 1831 |
+
"loss": 0.0845,
|
| 1832 |
+
"step": 2180
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 33.01384615384615,
|
| 1836 |
+
"grad_norm": 143.28871154785156,
|
| 1837 |
+
"learning_rate": 3.623931623931624e-06,
|
| 1838 |
+
"loss": 0.2904,
|
| 1839 |
+
"step": 2190
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 33.01692307692308,
|
| 1843 |
+
"grad_norm": 0.008105006068944931,
|
| 1844 |
+
"learning_rate": 3.58974358974359e-06,
|
| 1845 |
+
"loss": 0.016,
|
| 1846 |
+
"step": 2200
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 33.02,
|
| 1850 |
+
"grad_norm": 0.05293996259570122,
|
| 1851 |
+
"learning_rate": 3.555555555555556e-06,
|
| 1852 |
+
"loss": 0.0924,
|
| 1853 |
+
"step": 2210
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 33.02,
|
| 1857 |
+
"eval_accuracy": 0.782608695652174,
|
| 1858 |
+
"eval_loss": 0.9858593940734863,
|
| 1859 |
+
"eval_runtime": 10.3676,
|
| 1860 |
+
"eval_samples_per_second": 4.437,
|
| 1861 |
+
"eval_steps_per_second": 1.157,
|
| 1862 |
+
"step": 2210
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 34.003076923076925,
|
| 1866 |
+
"grad_norm": 35.07648468017578,
|
| 1867 |
+
"learning_rate": 3.5213675213675218e-06,
|
| 1868 |
+
"loss": 0.0892,
|
| 1869 |
+
"step": 2220
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 34.00615384615384,
|
| 1873 |
+
"grad_norm": 0.11222043633460999,
|
| 1874 |
+
"learning_rate": 3.487179487179487e-06,
|
| 1875 |
+
"loss": 0.1803,
|
| 1876 |
+
"step": 2230
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 34.00923076923077,
|
| 1880 |
+
"grad_norm": 0.21430917084217072,
|
| 1881 |
+
"learning_rate": 3.452991452991453e-06,
|
| 1882 |
+
"loss": 0.1777,
|
| 1883 |
+
"step": 2240
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 34.012307692307694,
|
| 1887 |
+
"grad_norm": 16.7849063873291,
|
| 1888 |
+
"learning_rate": 3.4188034188034193e-06,
|
| 1889 |
+
"loss": 0.1209,
|
| 1890 |
+
"step": 2250
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 34.01538461538462,
|
| 1894 |
+
"grad_norm": 64.78038787841797,
|
| 1895 |
+
"learning_rate": 3.384615384615385e-06,
|
| 1896 |
+
"loss": 0.1789,
|
| 1897 |
+
"step": 2260
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 34.01846153846154,
|
| 1901 |
+
"grad_norm": 0.4827536642551422,
|
| 1902 |
+
"learning_rate": 3.3504273504273506e-06,
|
| 1903 |
+
"loss": 0.1016,
|
| 1904 |
+
"step": 2270
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 34.02,
|
| 1908 |
+
"eval_accuracy": 0.782608695652174,
|
| 1909 |
+
"eval_loss": 1.075831413269043,
|
| 1910 |
+
"eval_runtime": 10.3606,
|
| 1911 |
+
"eval_samples_per_second": 4.44,
|
| 1912 |
+
"eval_steps_per_second": 1.158,
|
| 1913 |
+
"step": 2275
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 35.00153846153846,
|
| 1917 |
+
"grad_norm": 19.670827865600586,
|
| 1918 |
+
"learning_rate": 3.3162393162393165e-06,
|
| 1919 |
+
"loss": 0.0053,
|
| 1920 |
+
"step": 2280
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 35.004615384615384,
|
| 1924 |
+
"grad_norm": 0.02641270123422146,
|
| 1925 |
+
"learning_rate": 3.2820512820512823e-06,
|
| 1926 |
+
"loss": 0.078,
|
| 1927 |
+
"step": 2290
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 35.00769230769231,
|
| 1931 |
+
"grad_norm": 13.935211181640625,
|
| 1932 |
+
"learning_rate": 3.247863247863248e-06,
|
| 1933 |
+
"loss": 0.163,
|
| 1934 |
+
"step": 2300
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 35.01076923076923,
|
| 1938 |
+
"grad_norm": 80.81470489501953,
|
| 1939 |
+
"learning_rate": 3.213675213675214e-06,
|
| 1940 |
+
"loss": 0.083,
|
| 1941 |
+
"step": 2310
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 35.01384615384615,
|
| 1945 |
+
"grad_norm": 0.01466295588761568,
|
| 1946 |
+
"learning_rate": 3.1794871794871795e-06,
|
| 1947 |
+
"loss": 0.183,
|
| 1948 |
+
"step": 2320
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 35.01692307692308,
|
| 1952 |
+
"grad_norm": 0.007615539710968733,
|
| 1953 |
+
"learning_rate": 3.1452991452991453e-06,
|
| 1954 |
+
"loss": 0.5267,
|
| 1955 |
+
"step": 2330
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 35.02,
|
| 1959 |
+
"grad_norm": 0.019310418516397476,
|
| 1960 |
+
"learning_rate": 3.1111111111111116e-06,
|
| 1961 |
+
"loss": 0.1894,
|
| 1962 |
+
"step": 2340
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 35.02,
|
| 1966 |
+
"eval_accuracy": 0.7608695652173914,
|
| 1967 |
+
"eval_loss": 1.208774209022522,
|
| 1968 |
+
"eval_runtime": 10.3816,
|
| 1969 |
+
"eval_samples_per_second": 4.431,
|
| 1970 |
+
"eval_steps_per_second": 1.156,
|
| 1971 |
+
"step": 2340
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 36.003076923076925,
|
| 1975 |
+
"grad_norm": 0.17081318795681,
|
| 1976 |
+
"learning_rate": 3.0769230769230774e-06,
|
| 1977 |
+
"loss": 0.1565,
|
| 1978 |
+
"step": 2350
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 36.00615384615384,
|
| 1982 |
+
"grad_norm": 3.074644088745117,
|
| 1983 |
+
"learning_rate": 3.042735042735043e-06,
|
| 1984 |
+
"loss": 0.2069,
|
| 1985 |
+
"step": 2360
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 36.00923076923077,
|
| 1989 |
+
"grad_norm": 0.039210639894008636,
|
| 1990 |
+
"learning_rate": 3.0085470085470087e-06,
|
| 1991 |
+
"loss": 0.0321,
|
| 1992 |
+
"step": 2370
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 36.012307692307694,
|
| 1996 |
+
"grad_norm": 195.3864288330078,
|
| 1997 |
+
"learning_rate": 2.9743589743589746e-06,
|
| 1998 |
+
"loss": 0.2972,
|
| 1999 |
+
"step": 2380
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 36.01538461538462,
|
| 2003 |
+
"grad_norm": 0.26288241147994995,
|
| 2004 |
+
"learning_rate": 2.9401709401709404e-06,
|
| 2005 |
+
"loss": 0.0027,
|
| 2006 |
+
"step": 2390
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 36.01846153846154,
|
| 2010 |
+
"grad_norm": 1.5095382928848267,
|
| 2011 |
+
"learning_rate": 2.9059829059829063e-06,
|
| 2012 |
+
"loss": 0.2657,
|
| 2013 |
+
"step": 2400
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 36.02,
|
| 2017 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2018 |
+
"eval_loss": 1.540911078453064,
|
| 2019 |
+
"eval_runtime": 10.5028,
|
| 2020 |
+
"eval_samples_per_second": 4.38,
|
| 2021 |
+
"eval_steps_per_second": 1.143,
|
| 2022 |
+
"step": 2405
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 37.00153846153846,
|
| 2026 |
+
"grad_norm": 0.006212185136973858,
|
| 2027 |
+
"learning_rate": 2.8717948717948717e-06,
|
| 2028 |
+
"loss": 0.0008,
|
| 2029 |
+
"step": 2410
|
| 2030 |
+
},
|
| 2031 |
+
{
|
| 2032 |
+
"epoch": 37.004615384615384,
|
| 2033 |
+
"grad_norm": 1.8746821880340576,
|
| 2034 |
+
"learning_rate": 2.8376068376068376e-06,
|
| 2035 |
+
"loss": 0.2584,
|
| 2036 |
+
"step": 2420
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 37.00769230769231,
|
| 2040 |
+
"grad_norm": 0.13942450284957886,
|
| 2041 |
+
"learning_rate": 2.803418803418804e-06,
|
| 2042 |
+
"loss": 0.0947,
|
| 2043 |
+
"step": 2430
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 37.01076923076923,
|
| 2047 |
+
"grad_norm": 0.018762849271297455,
|
| 2048 |
+
"learning_rate": 2.7692307692307697e-06,
|
| 2049 |
+
"loss": 0.3546,
|
| 2050 |
+
"step": 2440
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 37.01384615384615,
|
| 2054 |
+
"grad_norm": 0.018523642793297768,
|
| 2055 |
+
"learning_rate": 2.7350427350427355e-06,
|
| 2056 |
+
"loss": 0.1265,
|
| 2057 |
+
"step": 2450
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 37.01692307692308,
|
| 2061 |
+
"grad_norm": 0.03315011411905289,
|
| 2062 |
+
"learning_rate": 2.700854700854701e-06,
|
| 2063 |
+
"loss": 0.1602,
|
| 2064 |
+
"step": 2460
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 37.02,
|
| 2068 |
+
"grad_norm": 0.05689428746700287,
|
| 2069 |
+
"learning_rate": 2.666666666666667e-06,
|
| 2070 |
+
"loss": 0.1235,
|
| 2071 |
+
"step": 2470
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 37.02,
|
| 2075 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2076 |
+
"eval_loss": 1.2735745906829834,
|
| 2077 |
+
"eval_runtime": 10.41,
|
| 2078 |
+
"eval_samples_per_second": 4.419,
|
| 2079 |
+
"eval_steps_per_second": 1.153,
|
| 2080 |
+
"step": 2470
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 38.003076923076925,
|
| 2084 |
+
"grad_norm": 0.03720976412296295,
|
| 2085 |
+
"learning_rate": 2.6324786324786327e-06,
|
| 2086 |
+
"loss": 0.0013,
|
| 2087 |
+
"step": 2480
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 38.00615384615384,
|
| 2091 |
+
"grad_norm": 0.026006508618593216,
|
| 2092 |
+
"learning_rate": 2.5982905982905985e-06,
|
| 2093 |
+
"loss": 0.0833,
|
| 2094 |
+
"step": 2490
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 38.00923076923077,
|
| 2098 |
+
"grad_norm": 0.11773164570331573,
|
| 2099 |
+
"learning_rate": 2.564102564102564e-06,
|
| 2100 |
+
"loss": 0.0828,
|
| 2101 |
+
"step": 2500
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 38.012307692307694,
|
| 2105 |
+
"grad_norm": 0.06585920602083206,
|
| 2106 |
+
"learning_rate": 2.52991452991453e-06,
|
| 2107 |
+
"loss": 0.2328,
|
| 2108 |
+
"step": 2510
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 38.01538461538462,
|
| 2112 |
+
"grad_norm": 64.87152862548828,
|
| 2113 |
+
"learning_rate": 2.495726495726496e-06,
|
| 2114 |
+
"loss": 0.019,
|
| 2115 |
+
"step": 2520
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 38.01846153846154,
|
| 2119 |
+
"grad_norm": 30.827510833740234,
|
| 2120 |
+
"learning_rate": 2.461538461538462e-06,
|
| 2121 |
+
"loss": 0.1539,
|
| 2122 |
+
"step": 2530
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 38.02,
|
| 2126 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2127 |
+
"eval_loss": 1.2607876062393188,
|
| 2128 |
+
"eval_runtime": 10.3711,
|
| 2129 |
+
"eval_samples_per_second": 4.435,
|
| 2130 |
+
"eval_steps_per_second": 1.157,
|
| 2131 |
+
"step": 2535
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 39.00153846153846,
|
| 2135 |
+
"grad_norm": 0.02473655343055725,
|
| 2136 |
+
"learning_rate": 2.4273504273504274e-06,
|
| 2137 |
+
"loss": 0.001,
|
| 2138 |
+
"step": 2540
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 39.004615384615384,
|
| 2142 |
+
"grad_norm": 0.024143971502780914,
|
| 2143 |
+
"learning_rate": 2.3931623931623937e-06,
|
| 2144 |
+
"loss": 0.4801,
|
| 2145 |
+
"step": 2550
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 39.00769230769231,
|
| 2149 |
+
"grad_norm": 0.03832996264100075,
|
| 2150 |
+
"learning_rate": 2.358974358974359e-06,
|
| 2151 |
+
"loss": 0.332,
|
| 2152 |
+
"step": 2560
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 39.01076923076923,
|
| 2156 |
+
"grad_norm": 0.019543638452887535,
|
| 2157 |
+
"learning_rate": 2.324786324786325e-06,
|
| 2158 |
+
"loss": 0.0006,
|
| 2159 |
+
"step": 2570
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 39.01384615384615,
|
| 2163 |
+
"grad_norm": 0.01422965619713068,
|
| 2164 |
+
"learning_rate": 2.290598290598291e-06,
|
| 2165 |
+
"loss": 0.3965,
|
| 2166 |
+
"step": 2580
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 39.01692307692308,
|
| 2170 |
+
"grad_norm": 0.04189188778400421,
|
| 2171 |
+
"learning_rate": 2.2564102564102566e-06,
|
| 2172 |
+
"loss": 0.0699,
|
| 2173 |
+
"step": 2590
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 39.02,
|
| 2177 |
+
"grad_norm": 1.5431499481201172,
|
| 2178 |
+
"learning_rate": 2.222222222222222e-06,
|
| 2179 |
+
"loss": 0.03,
|
| 2180 |
+
"step": 2600
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 39.02,
|
| 2184 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2185 |
+
"eval_loss": 1.2057995796203613,
|
| 2186 |
+
"eval_runtime": 10.1887,
|
| 2187 |
+
"eval_samples_per_second": 4.515,
|
| 2188 |
+
"eval_steps_per_second": 1.178,
|
| 2189 |
+
"step": 2600
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 40.003076923076925,
|
| 2193 |
+
"grad_norm": 0.06963592022657394,
|
| 2194 |
+
"learning_rate": 2.1880341880341884e-06,
|
| 2195 |
+
"loss": 0.1184,
|
| 2196 |
+
"step": 2610
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 40.00615384615384,
|
| 2200 |
+
"grad_norm": 0.03634906932711601,
|
| 2201 |
+
"learning_rate": 2.153846153846154e-06,
|
| 2202 |
+
"loss": 0.0023,
|
| 2203 |
+
"step": 2620
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 40.00923076923077,
|
| 2207 |
+
"grad_norm": 0.026238996535539627,
|
| 2208 |
+
"learning_rate": 2.1196581196581196e-06,
|
| 2209 |
+
"loss": 0.2015,
|
| 2210 |
+
"step": 2630
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 40.012307692307694,
|
| 2214 |
+
"grad_norm": 0.017962316051125526,
|
| 2215 |
+
"learning_rate": 2.085470085470086e-06,
|
| 2216 |
+
"loss": 0.2145,
|
| 2217 |
+
"step": 2640
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 40.01538461538462,
|
| 2221 |
+
"grad_norm": 0.0143351498991251,
|
| 2222 |
+
"learning_rate": 2.0512820512820513e-06,
|
| 2223 |
+
"loss": 0.1958,
|
| 2224 |
+
"step": 2650
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 40.01846153846154,
|
| 2228 |
+
"grad_norm": 0.08577225357294083,
|
| 2229 |
+
"learning_rate": 2.017094017094017e-06,
|
| 2230 |
+
"loss": 0.1447,
|
| 2231 |
+
"step": 2660
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 40.02,
|
| 2235 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2236 |
+
"eval_loss": 1.10723876953125,
|
| 2237 |
+
"eval_runtime": 10.6531,
|
| 2238 |
+
"eval_samples_per_second": 4.318,
|
| 2239 |
+
"eval_steps_per_second": 1.126,
|
| 2240 |
+
"step": 2665
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 41.00153846153846,
|
| 2244 |
+
"grad_norm": 0.006715043913573027,
|
| 2245 |
+
"learning_rate": 1.982905982905983e-06,
|
| 2246 |
+
"loss": 0.1283,
|
| 2247 |
+
"step": 2670
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 41.004615384615384,
|
| 2251 |
+
"grad_norm": 0.07786259055137634,
|
| 2252 |
+
"learning_rate": 1.948717948717949e-06,
|
| 2253 |
+
"loss": 0.0072,
|
| 2254 |
+
"step": 2680
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 41.00769230769231,
|
| 2258 |
+
"grad_norm": 1.9331977367401123,
|
| 2259 |
+
"learning_rate": 1.9145299145299148e-06,
|
| 2260 |
+
"loss": 0.1126,
|
| 2261 |
+
"step": 2690
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 41.01076923076923,
|
| 2265 |
+
"grad_norm": 176.3519744873047,
|
| 2266 |
+
"learning_rate": 1.8803418803418804e-06,
|
| 2267 |
+
"loss": 0.1764,
|
| 2268 |
+
"step": 2700
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 41.01384615384615,
|
| 2272 |
+
"grad_norm": 1.9080088138580322,
|
| 2273 |
+
"learning_rate": 1.8461538461538465e-06,
|
| 2274 |
+
"loss": 0.099,
|
| 2275 |
+
"step": 2710
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 41.01692307692308,
|
| 2279 |
+
"grad_norm": 0.02918870560824871,
|
| 2280 |
+
"learning_rate": 1.811965811965812e-06,
|
| 2281 |
+
"loss": 0.0805,
|
| 2282 |
+
"step": 2720
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 41.02,
|
| 2286 |
+
"grad_norm": 227.27313232421875,
|
| 2287 |
+
"learning_rate": 1.777777777777778e-06,
|
| 2288 |
+
"loss": 0.0888,
|
| 2289 |
+
"step": 2730
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 41.02,
|
| 2293 |
+
"eval_accuracy": 0.782608695652174,
|
| 2294 |
+
"eval_loss": 1.1454174518585205,
|
| 2295 |
+
"eval_runtime": 10.4748,
|
| 2296 |
+
"eval_samples_per_second": 4.392,
|
| 2297 |
+
"eval_steps_per_second": 1.146,
|
| 2298 |
+
"step": 2730
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 42.003076923076925,
|
| 2302 |
+
"grad_norm": 0.015502427704632282,
|
| 2303 |
+
"learning_rate": 1.7435897435897436e-06,
|
| 2304 |
+
"loss": 0.0901,
|
| 2305 |
+
"step": 2740
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 42.00615384615384,
|
| 2309 |
+
"grad_norm": 0.02255420945584774,
|
| 2310 |
+
"learning_rate": 1.7094017094017097e-06,
|
| 2311 |
+
"loss": 0.2097,
|
| 2312 |
+
"step": 2750
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 42.00923076923077,
|
| 2316 |
+
"grad_norm": 186.2987060546875,
|
| 2317 |
+
"learning_rate": 1.6752136752136753e-06,
|
| 2318 |
+
"loss": 0.0813,
|
| 2319 |
+
"step": 2760
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 42.012307692307694,
|
| 2323 |
+
"grad_norm": 1.8150595426559448,
|
| 2324 |
+
"learning_rate": 1.6410256410256412e-06,
|
| 2325 |
+
"loss": 0.1309,
|
| 2326 |
+
"step": 2770
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 42.01538461538462,
|
| 2330 |
+
"grad_norm": 0.03459843620657921,
|
| 2331 |
+
"learning_rate": 1.606837606837607e-06,
|
| 2332 |
+
"loss": 0.1189,
|
| 2333 |
+
"step": 2780
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 42.01846153846154,
|
| 2337 |
+
"grad_norm": 0.0022768706548959017,
|
| 2338 |
+
"learning_rate": 1.5726495726495727e-06,
|
| 2339 |
+
"loss": 0.0016,
|
| 2340 |
+
"step": 2790
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 42.02,
|
| 2344 |
+
"eval_accuracy": 0.782608695652174,
|
| 2345 |
+
"eval_loss": 1.1194366216659546,
|
| 2346 |
+
"eval_runtime": 10.782,
|
| 2347 |
+
"eval_samples_per_second": 4.266,
|
| 2348 |
+
"eval_steps_per_second": 1.113,
|
| 2349 |
+
"step": 2795
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 43.00153846153846,
|
| 2353 |
+
"grad_norm": 0.005485150497406721,
|
| 2354 |
+
"learning_rate": 1.5384615384615387e-06,
|
| 2355 |
+
"loss": 0.0008,
|
| 2356 |
+
"step": 2800
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 43.004615384615384,
|
| 2360 |
+
"grad_norm": 0.7399391531944275,
|
| 2361 |
+
"learning_rate": 1.5042735042735044e-06,
|
| 2362 |
+
"loss": 0.1494,
|
| 2363 |
+
"step": 2810
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 43.00769230769231,
|
| 2367 |
+
"grad_norm": 0.02695586159825325,
|
| 2368 |
+
"learning_rate": 1.4700854700854702e-06,
|
| 2369 |
+
"loss": 0.0015,
|
| 2370 |
+
"step": 2820
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 43.01076923076923,
|
| 2374 |
+
"grad_norm": 0.013859664089977741,
|
| 2375 |
+
"learning_rate": 1.4358974358974359e-06,
|
| 2376 |
+
"loss": 0.0703,
|
| 2377 |
+
"step": 2830
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 43.01384615384615,
|
| 2381 |
+
"grad_norm": 0.02371545322239399,
|
| 2382 |
+
"learning_rate": 1.401709401709402e-06,
|
| 2383 |
+
"loss": 0.1537,
|
| 2384 |
+
"step": 2840
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 43.01692307692308,
|
| 2388 |
+
"grad_norm": 63.8938102722168,
|
| 2389 |
+
"learning_rate": 1.3675213675213678e-06,
|
| 2390 |
+
"loss": 0.1798,
|
| 2391 |
+
"step": 2850
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 43.02,
|
| 2395 |
+
"grad_norm": 0.08463682234287262,
|
| 2396 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 2397 |
+
"loss": 0.1489,
|
| 2398 |
+
"step": 2860
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 43.02,
|
| 2402 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2403 |
+
"eval_loss": 1.2169851064682007,
|
| 2404 |
+
"eval_runtime": 10.3239,
|
| 2405 |
+
"eval_samples_per_second": 4.456,
|
| 2406 |
+
"eval_steps_per_second": 1.162,
|
| 2407 |
+
"step": 2860
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 44.003076923076925,
|
| 2411 |
+
"grad_norm": 105.62195587158203,
|
| 2412 |
+
"learning_rate": 1.2991452991452993e-06,
|
| 2413 |
+
"loss": 0.036,
|
| 2414 |
+
"step": 2870
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 44.00615384615384,
|
| 2418 |
+
"grad_norm": 191.20787048339844,
|
| 2419 |
+
"learning_rate": 1.264957264957265e-06,
|
| 2420 |
+
"loss": 0.0918,
|
| 2421 |
+
"step": 2880
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 44.00923076923077,
|
| 2425 |
+
"grad_norm": 0.17883938550949097,
|
| 2426 |
+
"learning_rate": 1.230769230769231e-06,
|
| 2427 |
+
"loss": 0.1685,
|
| 2428 |
+
"step": 2890
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 44.012307692307694,
|
| 2432 |
+
"grad_norm": 0.017934106290340424,
|
| 2433 |
+
"learning_rate": 1.1965811965811968e-06,
|
| 2434 |
+
"loss": 0.0005,
|
| 2435 |
+
"step": 2900
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 44.01538461538462,
|
| 2439 |
+
"grad_norm": 0.1711297333240509,
|
| 2440 |
+
"learning_rate": 1.1623931623931625e-06,
|
| 2441 |
+
"loss": 0.3324,
|
| 2442 |
+
"step": 2910
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 44.01846153846154,
|
| 2446 |
+
"grad_norm": 0.05496324226260185,
|
| 2447 |
+
"learning_rate": 1.1282051282051283e-06,
|
| 2448 |
+
"loss": 0.0004,
|
| 2449 |
+
"step": 2920
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 44.02,
|
| 2453 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2454 |
+
"eval_loss": 1.1893503665924072,
|
| 2455 |
+
"eval_runtime": 10.4088,
|
| 2456 |
+
"eval_samples_per_second": 4.419,
|
| 2457 |
+
"eval_steps_per_second": 1.153,
|
| 2458 |
+
"step": 2925
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 45.00153846153846,
|
| 2462 |
+
"grad_norm": 0.017374416813254356,
|
| 2463 |
+
"learning_rate": 1.0940170940170942e-06,
|
| 2464 |
+
"loss": 0.3326,
|
| 2465 |
+
"step": 2930
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 45.004615384615384,
|
| 2469 |
+
"grad_norm": 0.01586207188665867,
|
| 2470 |
+
"learning_rate": 1.0598290598290598e-06,
|
| 2471 |
+
"loss": 0.1512,
|
| 2472 |
+
"step": 2940
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 45.00769230769231,
|
| 2476 |
+
"grad_norm": 0.7407118082046509,
|
| 2477 |
+
"learning_rate": 1.0256410256410257e-06,
|
| 2478 |
+
"loss": 0.1375,
|
| 2479 |
+
"step": 2950
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 45.01076923076923,
|
| 2483 |
+
"grad_norm": 0.00448651472106576,
|
| 2484 |
+
"learning_rate": 9.914529914529915e-07,
|
| 2485 |
+
"loss": 0.0243,
|
| 2486 |
+
"step": 2960
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 45.01384615384615,
|
| 2490 |
+
"grad_norm": 0.17225775122642517,
|
| 2491 |
+
"learning_rate": 9.572649572649574e-07,
|
| 2492 |
+
"loss": 0.1769,
|
| 2493 |
+
"step": 2970
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 45.01692307692308,
|
| 2497 |
+
"grad_norm": 0.10238823294639587,
|
| 2498 |
+
"learning_rate": 9.230769230769232e-07,
|
| 2499 |
+
"loss": 0.0217,
|
| 2500 |
+
"step": 2980
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 45.02,
|
| 2504 |
+
"grad_norm": 0.019600247964262962,
|
| 2505 |
+
"learning_rate": 8.88888888888889e-07,
|
| 2506 |
+
"loss": 0.0004,
|
| 2507 |
+
"step": 2990
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 45.02,
|
| 2511 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2512 |
+
"eval_loss": 1.332923173904419,
|
| 2513 |
+
"eval_runtime": 10.5547,
|
| 2514 |
+
"eval_samples_per_second": 4.358,
|
| 2515 |
+
"eval_steps_per_second": 1.137,
|
| 2516 |
+
"step": 2990
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 46.003076923076925,
|
| 2520 |
+
"grad_norm": 0.018734971061348915,
|
| 2521 |
+
"learning_rate": 8.547008547008548e-07,
|
| 2522 |
+
"loss": 0.0202,
|
| 2523 |
+
"step": 3000
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 46.00615384615384,
|
| 2527 |
+
"grad_norm": 0.0044385273940861225,
|
| 2528 |
+
"learning_rate": 8.205128205128206e-07,
|
| 2529 |
+
"loss": 0.0009,
|
| 2530 |
+
"step": 3010
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 46.00923076923077,
|
| 2534 |
+
"grad_norm": 0.06210096552968025,
|
| 2535 |
+
"learning_rate": 7.863247863247863e-07,
|
| 2536 |
+
"loss": 0.2226,
|
| 2537 |
+
"step": 3020
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 46.012307692307694,
|
| 2541 |
+
"grad_norm": 0.004177386872470379,
|
| 2542 |
+
"learning_rate": 7.521367521367522e-07,
|
| 2543 |
+
"loss": 0.0477,
|
| 2544 |
+
"step": 3030
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 46.01538461538462,
|
| 2548 |
+
"grad_norm": 0.013359845615923405,
|
| 2549 |
+
"learning_rate": 7.179487179487179e-07,
|
| 2550 |
+
"loss": 0.0004,
|
| 2551 |
+
"step": 3040
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 46.01846153846154,
|
| 2555 |
+
"grad_norm": 0.011247428134083748,
|
| 2556 |
+
"learning_rate": 6.837606837606839e-07,
|
| 2557 |
+
"loss": 0.0014,
|
| 2558 |
+
"step": 3050
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 46.02,
|
| 2562 |
+
"eval_accuracy": 0.7608695652173914,
|
| 2563 |
+
"eval_loss": 1.1886558532714844,
|
| 2564 |
+
"eval_runtime": 10.4416,
|
| 2565 |
+
"eval_samples_per_second": 4.405,
|
| 2566 |
+
"eval_steps_per_second": 1.149,
|
| 2567 |
+
"step": 3055
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 47.00153846153846,
|
| 2571 |
+
"grad_norm": 0.0468427948653698,
|
| 2572 |
+
"learning_rate": 6.495726495726496e-07,
|
| 2573 |
+
"loss": 0.1823,
|
| 2574 |
+
"step": 3060
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 47.004615384615384,
|
| 2578 |
+
"grad_norm": 0.029943831264972687,
|
| 2579 |
+
"learning_rate": 6.153846153846155e-07,
|
| 2580 |
+
"loss": 0.0144,
|
| 2581 |
+
"step": 3070
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 47.00769230769231,
|
| 2585 |
+
"grad_norm": 0.012487895786762238,
|
| 2586 |
+
"learning_rate": 5.811965811965812e-07,
|
| 2587 |
+
"loss": 0.1565,
|
| 2588 |
+
"step": 3080
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 47.01076923076923,
|
| 2592 |
+
"grad_norm": 0.0036076002288609743,
|
| 2593 |
+
"learning_rate": 5.470085470085471e-07,
|
| 2594 |
+
"loss": 0.0188,
|
| 2595 |
+
"step": 3090
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 47.01384615384615,
|
| 2599 |
+
"grad_norm": 0.043261803686618805,
|
| 2600 |
+
"learning_rate": 5.128205128205128e-07,
|
| 2601 |
+
"loss": 0.1021,
|
| 2602 |
+
"step": 3100
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 47.01692307692308,
|
| 2606 |
+
"grad_norm": 0.31245148181915283,
|
| 2607 |
+
"learning_rate": 4.786324786324787e-07,
|
| 2608 |
+
"loss": 0.0004,
|
| 2609 |
+
"step": 3110
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 47.02,
|
| 2613 |
+
"grad_norm": 0.009071093052625656,
|
| 2614 |
+
"learning_rate": 4.444444444444445e-07,
|
| 2615 |
+
"loss": 0.1675,
|
| 2616 |
+
"step": 3120
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 47.02,
|
| 2620 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2621 |
+
"eval_loss": 1.265150785446167,
|
| 2622 |
+
"eval_runtime": 10.6277,
|
| 2623 |
+
"eval_samples_per_second": 4.328,
|
| 2624 |
+
"eval_steps_per_second": 1.129,
|
| 2625 |
+
"step": 3120
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 48.003076923076925,
|
| 2629 |
+
"grad_norm": 0.9440973997116089,
|
| 2630 |
+
"learning_rate": 4.102564102564103e-07,
|
| 2631 |
+
"loss": 0.128,
|
| 2632 |
+
"step": 3130
|
| 2633 |
+
},
|
| 2634 |
+
{
|
| 2635 |
+
"epoch": 48.00615384615384,
|
| 2636 |
+
"grad_norm": 0.008040065877139568,
|
| 2637 |
+
"learning_rate": 3.760683760683761e-07,
|
| 2638 |
+
"loss": 0.0086,
|
| 2639 |
+
"step": 3140
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 48.00923076923077,
|
| 2643 |
+
"grad_norm": 0.3104320764541626,
|
| 2644 |
+
"learning_rate": 3.4188034188034194e-07,
|
| 2645 |
+
"loss": 0.0071,
|
| 2646 |
+
"step": 3150
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 48.012307692307694,
|
| 2650 |
+
"grad_norm": 0.012552163563668728,
|
| 2651 |
+
"learning_rate": 3.0769230769230774e-07,
|
| 2652 |
+
"loss": 0.0006,
|
| 2653 |
+
"step": 3160
|
| 2654 |
+
},
|
| 2655 |
+
{
|
| 2656 |
+
"epoch": 48.01538461538462,
|
| 2657 |
+
"grad_norm": 0.014428188093006611,
|
| 2658 |
+
"learning_rate": 2.7350427350427354e-07,
|
| 2659 |
+
"loss": 0.0616,
|
| 2660 |
+
"step": 3170
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 48.01846153846154,
|
| 2664 |
+
"grad_norm": 0.03152965009212494,
|
| 2665 |
+
"learning_rate": 2.3931623931623934e-07,
|
| 2666 |
+
"loss": 0.012,
|
| 2667 |
+
"step": 3180
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 48.02,
|
| 2671 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2672 |
+
"eval_loss": 1.3228085041046143,
|
| 2673 |
+
"eval_runtime": 10.4237,
|
| 2674 |
+
"eval_samples_per_second": 4.413,
|
| 2675 |
+
"eval_steps_per_second": 1.151,
|
| 2676 |
+
"step": 3185
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 49.00153846153846,
|
| 2680 |
+
"grad_norm": 0.5161852240562439,
|
| 2681 |
+
"learning_rate": 2.0512820512820514e-07,
|
| 2682 |
+
"loss": 0.0014,
|
| 2683 |
+
"step": 3190
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 49.004615384615384,
|
| 2687 |
+
"grad_norm": 0.02032632566988468,
|
| 2688 |
+
"learning_rate": 1.7094017094017097e-07,
|
| 2689 |
+
"loss": 0.0003,
|
| 2690 |
+
"step": 3200
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 49.00769230769231,
|
| 2694 |
+
"grad_norm": 0.014277896843850613,
|
| 2695 |
+
"learning_rate": 1.3675213675213677e-07,
|
| 2696 |
+
"loss": 0.0047,
|
| 2697 |
+
"step": 3210
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 49.01076923076923,
|
| 2701 |
+
"grad_norm": 0.3848668038845062,
|
| 2702 |
+
"learning_rate": 1.0256410256410257e-07,
|
| 2703 |
+
"loss": 0.001,
|
| 2704 |
+
"step": 3220
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 49.01384615384615,
|
| 2708 |
+
"grad_norm": 0.38336512446403503,
|
| 2709 |
+
"learning_rate": 6.837606837606839e-08,
|
| 2710 |
+
"loss": 0.0148,
|
| 2711 |
+
"step": 3230
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 49.01692307692308,
|
| 2715 |
+
"grad_norm": 0.012916259467601776,
|
| 2716 |
+
"learning_rate": 3.418803418803419e-08,
|
| 2717 |
+
"loss": 0.288,
|
| 2718 |
+
"step": 3240
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 49.02,
|
| 2722 |
+
"grad_norm": 0.013959445059299469,
|
| 2723 |
+
"learning_rate": 0.0,
|
| 2724 |
+
"loss": 0.0475,
|
| 2725 |
+
"step": 3250
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 49.02,
|
| 2729 |
+
"eval_accuracy": 0.7391304347826086,
|
| 2730 |
+
"eval_loss": 1.3506942987442017,
|
| 2731 |
+
"eval_runtime": 11.7124,
|
| 2732 |
+
"eval_samples_per_second": 3.927,
|
| 2733 |
+
"eval_steps_per_second": 1.025,
|
| 2734 |
+
"step": 3250
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 49.02,
|
| 2738 |
+
"step": 3250,
|
| 2739 |
+
"total_flos": 5.708334209885798e+19,
|
| 2740 |
+
"train_loss": 0.3139108933124405,
|
| 2741 |
+
"train_runtime": 6672.0929,
|
| 2742 |
+
"train_samples_per_second": 1.948,
|
| 2743 |
+
"train_steps_per_second": 0.487
|
| 2744 |
+
},
|
| 2745 |
+
{
|
| 2746 |
+
"epoch": 49.02,
|
| 2747 |
+
"eval_accuracy": 0.8043478260869565,
|
| 2748 |
+
"eval_loss": 0.48940712213516235,
|
| 2749 |
+
"eval_runtime": 10.4364,
|
| 2750 |
+
"eval_samples_per_second": 4.408,
|
| 2751 |
+
"eval_steps_per_second": 1.15,
|
| 2752 |
+
"step": 3250
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 49.02,
|
| 2756 |
+
"eval_accuracy": 0.8043478260869565,
|
| 2757 |
+
"eval_loss": 0.4894070327281952,
|
| 2758 |
+
"eval_runtime": 10.4359,
|
| 2759 |
+
"eval_samples_per_second": 4.408,
|
| 2760 |
+
"eval_steps_per_second": 1.15,
|
| 2761 |
+
"step": 3250
|
| 2762 |
+
}
|
| 2763 |
+
],
|
| 2764 |
+
"logging_steps": 10,
|
| 2765 |
+
"max_steps": 3250,
|
| 2766 |
+
"num_input_tokens_seen": 0,
|
| 2767 |
+
"num_train_epochs": 9223372036854775807,
|
| 2768 |
+
"save_steps": 500,
|
| 2769 |
+
"stateful_callbacks": {
|
| 2770 |
+
"TrainerControl": {
|
| 2771 |
+
"args": {
|
| 2772 |
+
"should_epoch_stop": false,
|
| 2773 |
+
"should_evaluate": false,
|
| 2774 |
+
"should_log": false,
|
| 2775 |
+
"should_save": true,
|
| 2776 |
+
"should_training_stop": true
|
| 2777 |
+
},
|
| 2778 |
+
"attributes": {}
|
| 2779 |
+
}
|
| 2780 |
+
},
|
| 2781 |
+
"total_flos": 5.708334209885798e+19,
|
| 2782 |
+
"train_batch_size": 4,
|
| 2783 |
+
"trial_name": null,
|
| 2784 |
+
"trial_params": null
|
| 2785 |
+
}
|
| 2786 |
+
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d181656762aa7205f51451f080f9c71c6ef2799b228356ff94ea29a8300ca45
|
| 3 |
+
size 4859
|