Upload 15 files
Browse files- .gitattributes +2 -0
- README.md +183 -3
- all_results.json +14 -0
- config.json +55 -0
- generation_config.json +15 -0
- model.safetensors +3 -0
- model_info.json +26 -0
- source.spm +3 -0
- special_tokens_map.json +5 -0
- target.spm +3 -0
- test_results.json +9 -0
- tokenizer_config.json +39 -0
- train_results.json +8 -0
- trainer_state.json +2527 -0
- training_args.bin +3 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
source.spm filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
target.spm filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -1,3 +1,183 @@
|
|
| 1 |
-
---
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- arc
|
| 4 |
+
tags:
|
| 5 |
+
- diacritization
|
| 6 |
+
- aramaic
|
| 7 |
+
- vocalization
|
| 8 |
+
- targum
|
| 9 |
+
- semitic-languages
|
| 10 |
+
- sequence-to-sequence
|
| 11 |
+
license: mit
|
| 12 |
+
base_model: Helsinki-NLP/opus-mt-afa-afa
|
| 13 |
+
library_name: transformers
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# Aramaic Diacritization Model (MarianMT)
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned MarianMT model for Aramaic text diacritization (vocalization), converting consonantal Aramaic text to fully vocalized text with nikkud (vowel points).
|
| 19 |
+
|
| 20 |
+
## Model Description
|
| 21 |
+
|
| 22 |
+
- **Model type:** MarianMT (Encoder-Decoder Transformer)
|
| 23 |
+
- **Language:** Aramaic (arc2arc)
|
| 24 |
+
- **Task:** Text diacritization/vocalization
|
| 25 |
+
- **Base model:** [Helsinki-NLP/opus-mt-afa-afa](https://huggingface.co/Helsinki-NLP/opus-mt-afa-afa)
|
| 26 |
+
- **Parameters:** 61,924,352 (61.9M)
|
| 27 |
+
|
| 28 |
+
## Model Architecture
|
| 29 |
+
|
| 30 |
+
- **Architecture:** MarianMT (Marian Machine Translation)
|
| 31 |
+
- **Encoder layers:** 6
|
| 32 |
+
- **Decoder layers:** 6
|
| 33 |
+
- **Hidden size:** 512
|
| 34 |
+
- **Attention heads:** 8
|
| 35 |
+
- **Feed-forward dimension:** 2048
|
| 36 |
+
- **Vocabulary size:** 33,714
|
| 37 |
+
- **Max sequence length:** 512 tokens
|
| 38 |
+
- **Activation function:** Swish
|
| 39 |
+
- **Position embeddings:** Static
|
| 40 |
+
|
| 41 |
+
## Training Details
|
| 42 |
+
|
| 43 |
+
### Training Configuration
|
| 44 |
+
- **Training data:** 12,110 examples
|
| 45 |
+
- **Validation data:** 1,514 examples
|
| 46 |
+
- **Batch size:** 8
|
| 47 |
+
- **Gradient accumulation steps:** 2
|
| 48 |
+
- **Effective batch size:** 16
|
| 49 |
+
- **Learning rate:** 1e-5
|
| 50 |
+
- **Warmup steps:** 1,000
|
| 51 |
+
- **Max epochs:** 100
|
| 52 |
+
- **Training completed at:** Epoch 36.33
|
| 53 |
+
- **Mixed precision:** FP16 enabled
|
| 54 |
+
|
| 55 |
+
### Training Metrics
|
| 56 |
+
- **Final training loss:** 0.283
|
| 57 |
+
- **Training runtime:** 21,727 seconds (~6 hours)
|
| 58 |
+
- **Training samples per second:** 55.7
|
| 59 |
+
- **Training steps per second:** 3.48
|
| 60 |
+
|
| 61 |
+
## Evaluation Results
|
| 62 |
+
|
| 63 |
+
### Test Set Performance
|
| 64 |
+
- **BLEU Score:** 72.90
|
| 65 |
+
- **Character Accuracy:** 63.78%
|
| 66 |
+
- **Evaluation Loss:** 0.088
|
| 67 |
+
- **Evaluation Runtime:** 311.5 seconds
|
| 68 |
+
- **Evaluation samples per second:** 4.86
|
| 69 |
+
|
| 70 |
+
## Usage
|
| 71 |
+
|
| 72 |
+
### Basic Usage
|
| 73 |
+
|
| 74 |
+
```python
|
| 75 |
+
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
| 76 |
+
|
| 77 |
+
# Load model and tokenizer
|
| 78 |
+
model_name = "johnlockejrr/aramaic-diacritization-model"
|
| 79 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 80 |
+
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
|
| 81 |
+
|
| 82 |
+
# Example input (consonantal Aramaic text)
|
| 83 |
+
consonantal_text = "בקדמין ברא יי ית שמיא וית ארעא"
|
| 84 |
+
|
| 85 |
+
# Tokenize input
|
| 86 |
+
inputs = tokenizer(consonantal_text, return_tensors="pt", max_length=512, truncation=True)
|
| 87 |
+
|
| 88 |
+
# Generate vocalized text
|
| 89 |
+
outputs = model.generate(**inputs, max_length=512, num_beams=4, early_stopping=True)
|
| 90 |
+
|
| 91 |
+
# Decode output
|
| 92 |
+
vocalized_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
| 93 |
+
print(f"Input: {consonantal_text}")
|
| 94 |
+
print(f"Output: {vocalized_text}")
|
| 95 |
+
```
|
| 96 |
+
|
| 97 |
+
### Using the Pipeline
|
| 98 |
+
|
| 99 |
+
```python
|
| 100 |
+
from transformers import pipeline
|
| 101 |
+
|
| 102 |
+
diacritizer = pipeline("text2text-generation", model="johnlockejrr/aramaic-diacritization-model")
|
| 103 |
+
|
| 104 |
+
# Process text
|
| 105 |
+
consonantal_text = "בראשית ברא אלהים את השמים ואת הארץ"
|
| 106 |
+
vocalized_text = diacritizer(consonantal_text)[0]['generated_text']
|
| 107 |
+
print(vocalized_text)
|
| 108 |
+
```
|
| 109 |
+
|
| 110 |
+
## Training Data
|
| 111 |
+
|
| 112 |
+
The model was trained on a custom Aramaic diacritization dataset with the following characteristics:
|
| 113 |
+
|
| 114 |
+
- **Source:** Consonantal Aramaic text (without vowel points)
|
| 115 |
+
- **Target:** Vocalized Aramaic text (with nikkud/vowel points)
|
| 116 |
+
- **Data format:** CSV with columns: consonantal, vocalized, book, chapter, verse
|
| 117 |
+
- **Data split:** 80% train, 10% validation, 10% test
|
| 118 |
+
- **Text cleaning:** Preserves nikkud in target text, removes punctuation from source
|
| 119 |
+
|
| 120 |
+
### Data Preprocessing
|
| 121 |
+
- **Input cleaning:** Removes punctuation and formatting while preserving letters
|
| 122 |
+
- **Target preservation:** Maintains all nikkud (vowel points) and diacritical marks
|
| 123 |
+
- **Length filtering:** Removes sequences shorter than 2 characters or longer than 1000 characters
|
| 124 |
+
- **Duplicate handling:** Removes exact duplicates to prevent data leakage
|
| 125 |
+
|
| 126 |
+
## Limitations and Bias
|
| 127 |
+
|
| 128 |
+
- **Domain specificity:** Trained primarily on religious/biblical Aramaic texts
|
| 129 |
+
- **Vocabulary coverage:** Limited to the vocabulary present in the training corpus
|
| 130 |
+
- **Length constraints:** Maximum input/output length of 512 tokens
|
| 131 |
+
- **Style consistency:** May not handle modern Aramaic dialects or contemporary usage
|
| 132 |
+
- **Performance:** Character accuracy of ~64% indicates room for improvement
|
| 133 |
+
|
| 134 |
+
## Environmental Impact
|
| 135 |
+
|
| 136 |
+
- **Hardware used:** NVIDIA GPU (GTX 3060 12GB)
|
| 137 |
+
- **Training time:** ~6 hours
|
| 138 |
+
- **Carbon emissions:** Estimated low (single GPU, moderate training time)
|
| 139 |
+
- **Energy efficiency:** FP16 mixed precision used to reduce memory usage
|
| 140 |
+
|
| 141 |
+
## Citation
|
| 142 |
+
|
| 143 |
+
If you use this model in your research, please cite:
|
| 144 |
+
|
| 145 |
+
```bibtex
|
| 146 |
+
@misc{aramaic-diacritization-2024,
|
| 147 |
+
title={Aramaic Diacritization Model},
|
| 148 |
+
author={Your Name},
|
| 149 |
+
year={2024},
|
| 150 |
+
howpublished={Hugging Face Model Hub},
|
| 151 |
+
url={https://huggingface.co/johnlockejrr/aramaic-diacritization-model}
|
| 152 |
+
}
|
| 153 |
+
```
|
| 154 |
+
|
| 155 |
+
## License
|
| 156 |
+
|
| 157 |
+
[Specify your license here]
|
| 158 |
+
|
| 159 |
+
## Acknowledgments
|
| 160 |
+
|
| 161 |
+
- Base model: [Helsinki-NLP/opus-mt-afa-afa](https://huggingface.co/Helsinki-NLP/opus-mt-afa-afa)
|
| 162 |
+
- Training framework: Hugging Face Transformers
|
| 163 |
+
- Dataset: Custom Aramaic diacritization corpus
|
| 164 |
+
|
| 165 |
+
## Model Files
|
| 166 |
+
|
| 167 |
+
- `model.safetensors` - Model weights (234MB)
|
| 168 |
+
- `config.json` - Model configuration
|
| 169 |
+
- `tokenizer_config.json` - Tokenizer configuration
|
| 170 |
+
- `source.spm` / `target.spm` - SentencePiece models
|
| 171 |
+
- `vocab.json` - Vocabulary file
|
| 172 |
+
- `generation_config.json` - Generation parameters
|
| 173 |
+
|
| 174 |
+
## Training Scripts
|
| 175 |
+
|
| 176 |
+
The model was trained using custom scripts:
|
| 177 |
+
- `train_arc2arc_improved_deep.py` - Main training script
|
| 178 |
+
- `run_arc2arc_improved_deep.sh` - Training execution script
|
| 179 |
+
- `run_resume_arc2arc_deep.sh` - Resume training script
|
| 180 |
+
|
| 181 |
+
## Contact
|
| 182 |
+
|
| 183 |
+
For questions, issues, or contributions, please open an issue on the model repository.
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 36.32760898282695,
|
| 3 |
+
"eval_bleu": 72.89736867105582,
|
| 4 |
+
"eval_char_accuracy": 63.78173613339268,
|
| 5 |
+
"eval_loss": 0.08819781988859177,
|
| 6 |
+
"eval_runtime": 311.5285,
|
| 7 |
+
"eval_samples_per_second": 4.86,
|
| 8 |
+
"eval_steps_per_second": 0.61,
|
| 9 |
+
"total_flos": 7035660725649408.0,
|
| 10 |
+
"train_loss": 0.282735899699818,
|
| 11 |
+
"train_runtime": 21727.9767,
|
| 12 |
+
"train_samples_per_second": 55.735,
|
| 13 |
+
"train_steps_per_second": 3.484
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"activation_function": "swish",
|
| 4 |
+
"add_bias_logits": false,
|
| 5 |
+
"add_final_layer_norm": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"MarianMTModel"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"bos_token_id": 0,
|
| 11 |
+
"classif_dropout": 0.0,
|
| 12 |
+
"classifier_dropout": 0.0,
|
| 13 |
+
"d_model": 512,
|
| 14 |
+
"decoder_attention_heads": 8,
|
| 15 |
+
"decoder_ffn_dim": 2048,
|
| 16 |
+
"decoder_layerdrop": 0.0,
|
| 17 |
+
"decoder_layers": 6,
|
| 18 |
+
"decoder_start_token_id": 33713,
|
| 19 |
+
"decoder_vocab_size": 33714,
|
| 20 |
+
"dropout": 0.1,
|
| 21 |
+
"encoder_attention_heads": 8,
|
| 22 |
+
"encoder_ffn_dim": 2048,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 6,
|
| 25 |
+
"eos_token_id": 0,
|
| 26 |
+
"extra_pos_embeddings": 33714,
|
| 27 |
+
"forced_eos_token_id": 0,
|
| 28 |
+
"id2label": {
|
| 29 |
+
"0": "LABEL_0",
|
| 30 |
+
"1": "LABEL_1",
|
| 31 |
+
"2": "LABEL_2"
|
| 32 |
+
},
|
| 33 |
+
"init_std": 0.02,
|
| 34 |
+
"is_encoder_decoder": true,
|
| 35 |
+
"label2id": {
|
| 36 |
+
"LABEL_0": 0,
|
| 37 |
+
"LABEL_1": 1,
|
| 38 |
+
"LABEL_2": 2
|
| 39 |
+
},
|
| 40 |
+
"max_length": null,
|
| 41 |
+
"max_position_embeddings": 512,
|
| 42 |
+
"model_type": "marian",
|
| 43 |
+
"normalize_before": false,
|
| 44 |
+
"normalize_embedding": false,
|
| 45 |
+
"num_beams": null,
|
| 46 |
+
"num_hidden_layers": 6,
|
| 47 |
+
"pad_token_id": 33713,
|
| 48 |
+
"scale_embedding": true,
|
| 49 |
+
"share_encoder_decoder_embeddings": true,
|
| 50 |
+
"static_position_embeddings": true,
|
| 51 |
+
"torch_dtype": "float32",
|
| 52 |
+
"transformers_version": "4.52.4",
|
| 53 |
+
"use_cache": true,
|
| 54 |
+
"vocab_size": 33714
|
| 55 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bad_words_ids": [
|
| 3 |
+
[
|
| 4 |
+
33713
|
| 5 |
+
]
|
| 6 |
+
],
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"decoder_start_token_id": 33713,
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"forced_eos_token_id": 0,
|
| 11 |
+
"max_length": 512,
|
| 12 |
+
"num_beams": 4,
|
| 13 |
+
"pad_token_id": 33713,
|
| 14 |
+
"transformers_version": "4.52.4"
|
| 15 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a50c1778afad506b902f0c0bef3f680042b96cd3fadd5c00f7652b97d6c7bca5
|
| 3 |
+
size 245764168
|
model_info.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_name": "Helsinki-NLP/opus-mt-afa-afa",
|
| 3 |
+
"direction": "arc2arc",
|
| 4 |
+
"vocabulary_size": 33714,
|
| 5 |
+
"model_parameters": 61924352,
|
| 6 |
+
"training_config": {
|
| 7 |
+
"dataset_path": "aramaic_diacritization_dataset_deep",
|
| 8 |
+
"output_dir": "./aramaic_diacritization_model_deep",
|
| 9 |
+
"model_name": "Helsinki-NLP/opus-mt-afa-afa",
|
| 10 |
+
"batch_size": 8,
|
| 11 |
+
"learning_rate": 1e-05,
|
| 12 |
+
"num_epochs": 100,
|
| 13 |
+
"max_input_length": 512,
|
| 14 |
+
"max_target_length": 512,
|
| 15 |
+
"eval_steps": 500,
|
| 16 |
+
"save_steps": 500,
|
| 17 |
+
"warmup_steps": 1000,
|
| 18 |
+
"gradient_accumulation_steps": 2,
|
| 19 |
+
"use_fp16": true,
|
| 20 |
+
"use_wandb": false,
|
| 21 |
+
"skip_evaluation": false,
|
| 22 |
+
"seed": 42,
|
| 23 |
+
"resume_from_checkpoint": null,
|
| 24 |
+
"strict_resume": false
|
| 25 |
+
}
|
| 26 |
+
}
|
source.spm
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:191fc6c0ae01c8044b66544b34f00c3690d052af746cda8f07f4886f98d4df32
|
| 3 |
+
size 842154
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eos_token": "</s>",
|
| 3 |
+
"pad_token": "<pad>",
|
| 4 |
+
"unk_token": "<unk>"
|
| 5 |
+
}
|
target.spm
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db1c1a7f37a9362214e7f93173e3d8e770cc7a2cf0375e5f92fcd2ef19db2b65
|
| 3 |
+
size 842006
|
test_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 36.32760898282695,
|
| 3 |
+
"eval_bleu": 72.89736867105582,
|
| 4 |
+
"eval_char_accuracy": 63.78173613339268,
|
| 5 |
+
"eval_loss": 0.08819781988859177,
|
| 6 |
+
"eval_runtime": 311.5285,
|
| 7 |
+
"eval_samples_per_second": 4.86,
|
| 8 |
+
"eval_steps_per_second": 0.61
|
| 9 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "</s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<unk>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"33713": {
|
| 20 |
+
"content": "<pad>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
}
|
| 27 |
+
},
|
| 28 |
+
"clean_up_tokenization_spaces": false,
|
| 29 |
+
"eos_token": "</s>",
|
| 30 |
+
"extra_special_tokens": {},
|
| 31 |
+
"model_max_length": 512,
|
| 32 |
+
"pad_token": "<pad>",
|
| 33 |
+
"separate_vocabs": false,
|
| 34 |
+
"source_lang": "afa",
|
| 35 |
+
"sp_model_kwargs": {},
|
| 36 |
+
"target_lang": "afa",
|
| 37 |
+
"tokenizer_class": "MarianTokenizer",
|
| 38 |
+
"unk_token": "<unk>"
|
| 39 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 36.32760898282695,
|
| 3 |
+
"total_flos": 7035660725649408.0,
|
| 4 |
+
"train_loss": 0.282735899699818,
|
| 5 |
+
"train_runtime": 21727.9767,
|
| 6 |
+
"train_samples_per_second": 55.735,
|
| 7 |
+
"train_steps_per_second": 3.484
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2527 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 26000,
|
| 3 |
+
"best_metric": 74.62046528623556,
|
| 4 |
+
"best_model_checkpoint": "./aramaic_diacritization_model_deep/checkpoint-26000",
|
| 5 |
+
"epoch": 36.32760898282695,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 27500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13210039630118892,
|
| 14 |
+
"grad_norm": 49.535423278808594,
|
| 15 |
+
"learning_rate": 9.400000000000001e-07,
|
| 16 |
+
"loss": 13.2291,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26420079260237783,
|
| 21 |
+
"grad_norm": 8.345534324645996,
|
| 22 |
+
"learning_rate": 1.94e-06,
|
| 23 |
+
"loss": 4.9005,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3963011889035667,
|
| 28 |
+
"grad_norm": 6.039630889892578,
|
| 29 |
+
"learning_rate": 2.9400000000000002e-06,
|
| 30 |
+
"loss": 3.1719,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5284015852047557,
|
| 35 |
+
"grad_norm": 5.1691412925720215,
|
| 36 |
+
"learning_rate": 3.94e-06,
|
| 37 |
+
"loss": 2.6108,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6605019815059445,
|
| 42 |
+
"grad_norm": 4.929126262664795,
|
| 43 |
+
"learning_rate": 4.94e-06,
|
| 44 |
+
"loss": 2.2213,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.6605019815059445,
|
| 49 |
+
"eval_bleu": 0.19172411762066338,
|
| 50 |
+
"eval_char_accuracy": 10.030946701760158,
|
| 51 |
+
"eval_loss": 1.71331787109375,
|
| 52 |
+
"eval_runtime": 308.7303,
|
| 53 |
+
"eval_samples_per_second": 4.904,
|
| 54 |
+
"eval_steps_per_second": 0.615,
|
| 55 |
+
"step": 500
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 0.7926023778071334,
|
| 59 |
+
"grad_norm": 5.598935604095459,
|
| 60 |
+
"learning_rate": 5.94e-06,
|
| 61 |
+
"loss": 1.9171,
|
| 62 |
+
"step": 600
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.9247027741083224,
|
| 66 |
+
"grad_norm": 5.103219032287598,
|
| 67 |
+
"learning_rate": 6.9400000000000005e-06,
|
| 68 |
+
"loss": 1.6624,
|
| 69 |
+
"step": 700
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 1.0568031704095113,
|
| 73 |
+
"grad_norm": 4.852541923522949,
|
| 74 |
+
"learning_rate": 7.94e-06,
|
| 75 |
+
"loss": 1.473,
|
| 76 |
+
"step": 800
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 1.1889035667107002,
|
| 80 |
+
"grad_norm": 4.4410624504089355,
|
| 81 |
+
"learning_rate": 8.94e-06,
|
| 82 |
+
"loss": 1.3081,
|
| 83 |
+
"step": 900
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.321003963011889,
|
| 87 |
+
"grad_norm": 3.9946470260620117,
|
| 88 |
+
"learning_rate": 9.940000000000001e-06,
|
| 89 |
+
"loss": 1.1709,
|
| 90 |
+
"step": 1000
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.321003963011889,
|
| 94 |
+
"eval_bleu": 5.2707215480174545,
|
| 95 |
+
"eval_char_accuracy": 18.391696825135714,
|
| 96 |
+
"eval_loss": 0.8503363132476807,
|
| 97 |
+
"eval_runtime": 326.5655,
|
| 98 |
+
"eval_samples_per_second": 4.636,
|
| 99 |
+
"eval_steps_per_second": 0.582,
|
| 100 |
+
"step": 1000
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 1.453104359313078,
|
| 104 |
+
"grad_norm": 5.828360557556152,
|
| 105 |
+
"learning_rate": 9.99996092907511e-06,
|
| 106 |
+
"loss": 1.0774,
|
| 107 |
+
"step": 1100
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.5852047556142668,
|
| 111 |
+
"grad_norm": 3.975123643875122,
|
| 112 |
+
"learning_rate": 9.999833582267183e-06,
|
| 113 |
+
"loss": 0.9803,
|
| 114 |
+
"step": 1200
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.7173051519154559,
|
| 118 |
+
"grad_norm": 4.110162258148193,
|
| 119 |
+
"learning_rate": 9.999617802644021e-06,
|
| 120 |
+
"loss": 0.9023,
|
| 121 |
+
"step": 1300
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.8494055482166445,
|
| 125 |
+
"grad_norm": 4.341949462890625,
|
| 126 |
+
"learning_rate": 9.999313594022158e-06,
|
| 127 |
+
"loss": 0.8494,
|
| 128 |
+
"step": 1400
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.9815059445178336,
|
| 132 |
+
"grad_norm": 3.7582991123199463,
|
| 133 |
+
"learning_rate": 9.99892096178217e-06,
|
| 134 |
+
"loss": 0.7841,
|
| 135 |
+
"step": 1500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.9815059445178336,
|
| 139 |
+
"eval_bleu": 13.790865530076871,
|
| 140 |
+
"eval_char_accuracy": 25.990602895213026,
|
| 141 |
+
"eval_loss": 0.5515339374542236,
|
| 142 |
+
"eval_runtime": 310.3546,
|
| 143 |
+
"eval_samples_per_second": 4.878,
|
| 144 |
+
"eval_steps_per_second": 0.612,
|
| 145 |
+
"step": 1500
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 2.1136063408190227,
|
| 149 |
+
"grad_norm": 4.066399574279785,
|
| 150 |
+
"learning_rate": 9.998439912868608e-06,
|
| 151 |
+
"loss": 0.7379,
|
| 152 |
+
"step": 1600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.2457067371202113,
|
| 156 |
+
"grad_norm": 3.739553928375244,
|
| 157 |
+
"learning_rate": 9.997870455789855e-06,
|
| 158 |
+
"loss": 0.6859,
|
| 159 |
+
"step": 1700
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.3778071334214004,
|
| 163 |
+
"grad_norm": 4.019917964935303,
|
| 164 |
+
"learning_rate": 9.997212600617986e-06,
|
| 165 |
+
"loss": 0.6547,
|
| 166 |
+
"step": 1800
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 2.509907529722589,
|
| 170 |
+
"grad_norm": 3.1273276805877686,
|
| 171 |
+
"learning_rate": 9.99646635898858e-06,
|
| 172 |
+
"loss": 0.6313,
|
| 173 |
+
"step": 1900
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 2.642007926023778,
|
| 177 |
+
"grad_norm": 3.0856070518493652,
|
| 178 |
+
"learning_rate": 9.995631744100536e-06,
|
| 179 |
+
"loss": 0.6058,
|
| 180 |
+
"step": 2000
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 2.642007926023778,
|
| 184 |
+
"eval_bleu": 22.141127198903924,
|
| 185 |
+
"eval_char_accuracy": 30.873190491857212,
|
| 186 |
+
"eval_loss": 0.4170660674571991,
|
| 187 |
+
"eval_runtime": 310.921,
|
| 188 |
+
"eval_samples_per_second": 4.869,
|
| 189 |
+
"eval_steps_per_second": 0.611,
|
| 190 |
+
"step": 2000
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 2.7741083223249667,
|
| 194 |
+
"grad_norm": 3.660649299621582,
|
| 195 |
+
"learning_rate": 9.994708770715807e-06,
|
| 196 |
+
"loss": 0.5758,
|
| 197 |
+
"step": 2100
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 2.906208718626156,
|
| 201 |
+
"grad_norm": 3.3534188270568848,
|
| 202 |
+
"learning_rate": 9.993697455159165e-06,
|
| 203 |
+
"loss": 0.5507,
|
| 204 |
+
"step": 2200
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.038309114927345,
|
| 208 |
+
"grad_norm": 2.831392526626587,
|
| 209 |
+
"learning_rate": 9.992597815317901e-06,
|
| 210 |
+
"loss": 0.5334,
|
| 211 |
+
"step": 2300
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.1704095112285335,
|
| 215 |
+
"grad_norm": 3.2069804668426514,
|
| 216 |
+
"learning_rate": 9.991409870641512e-06,
|
| 217 |
+
"loss": 0.508,
|
| 218 |
+
"step": 2400
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3025099075297226,
|
| 222 |
+
"grad_norm": 3.3302793502807617,
|
| 223 |
+
"learning_rate": 9.990133642141359e-06,
|
| 224 |
+
"loss": 0.4816,
|
| 225 |
+
"step": 2500
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.3025099075297226,
|
| 229 |
+
"eval_bleu": 28.8217714543364,
|
| 230 |
+
"eval_char_accuracy": 34.30457312057904,
|
| 231 |
+
"eval_loss": 0.34086647629737854,
|
| 232 |
+
"eval_runtime": 312.1625,
|
| 233 |
+
"eval_samples_per_second": 4.85,
|
| 234 |
+
"eval_steps_per_second": 0.609,
|
| 235 |
+
"step": 2500
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 3.4346103038309117,
|
| 239 |
+
"grad_norm": 3.0527114868164062,
|
| 240 |
+
"learning_rate": 9.988769152390284e-06,
|
| 241 |
+
"loss": 0.4779,
|
| 242 |
+
"step": 2600
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 3.5667107001321003,
|
| 246 |
+
"grad_norm": 2.557722568511963,
|
| 247 |
+
"learning_rate": 9.987316425522226e-06,
|
| 248 |
+
"loss": 0.4626,
|
| 249 |
+
"step": 2700
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 3.6988110964332894,
|
| 253 |
+
"grad_norm": 2.993014097213745,
|
| 254 |
+
"learning_rate": 9.985775487231788e-06,
|
| 255 |
+
"loss": 0.4452,
|
| 256 |
+
"step": 2800
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 3.830911492734478,
|
| 260 |
+
"grad_norm": 2.7321043014526367,
|
| 261 |
+
"learning_rate": 9.984146364773777e-06,
|
| 262 |
+
"loss": 0.4408,
|
| 263 |
+
"step": 2900
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 3.963011889035667,
|
| 267 |
+
"grad_norm": 2.8790836334228516,
|
| 268 |
+
"learning_rate": 9.982429086962729e-06,
|
| 269 |
+
"loss": 0.4108,
|
| 270 |
+
"step": 3000
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 3.963011889035667,
|
| 274 |
+
"eval_bleu": 33.52667574407562,
|
| 275 |
+
"eval_char_accuracy": 37.42289027800625,
|
| 276 |
+
"eval_loss": 0.29226553440093994,
|
| 277 |
+
"eval_runtime": 316.4732,
|
| 278 |
+
"eval_samples_per_second": 4.784,
|
| 279 |
+
"eval_steps_per_second": 0.6,
|
| 280 |
+
"step": 3000
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 4.095112285336856,
|
| 284 |
+
"grad_norm": 2.8862805366516113,
|
| 285 |
+
"learning_rate": 9.980623684172396e-06,
|
| 286 |
+
"loss": 0.4134,
|
| 287 |
+
"step": 3100
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 4.227212681638045,
|
| 291 |
+
"grad_norm": 2.4621527194976807,
|
| 292 |
+
"learning_rate": 9.978730188335215e-06,
|
| 293 |
+
"loss": 0.3919,
|
| 294 |
+
"step": 3200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 4.359313077939234,
|
| 298 |
+
"grad_norm": 2.822957992553711,
|
| 299 |
+
"learning_rate": 9.976748632941733e-06,
|
| 300 |
+
"loss": 0.384,
|
| 301 |
+
"step": 3300
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 4.491413474240423,
|
| 305 |
+
"grad_norm": 2.448110818862915,
|
| 306 |
+
"learning_rate": 9.974679053040018e-06,
|
| 307 |
+
"loss": 0.3735,
|
| 308 |
+
"step": 3400
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 4.623513870541611,
|
| 312 |
+
"grad_norm": 2.2914109230041504,
|
| 313 |
+
"learning_rate": 9.972521485235045e-06,
|
| 314 |
+
"loss": 0.3604,
|
| 315 |
+
"step": 3500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 4.623513870541611,
|
| 319 |
+
"eval_bleu": 37.45556816331904,
|
| 320 |
+
"eval_char_accuracy": 40.02354416844876,
|
| 321 |
+
"eval_loss": 0.25823718309402466,
|
| 322 |
+
"eval_runtime": 320.7507,
|
| 323 |
+
"eval_samples_per_second": 4.72,
|
| 324 |
+
"eval_steps_per_second": 0.592,
|
| 325 |
+
"step": 3500
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.755614266842801,
|
| 329 |
+
"grad_norm": 2.5924477577209473,
|
| 330 |
+
"learning_rate": 9.970275967688047e-06,
|
| 331 |
+
"loss": 0.3624,
|
| 332 |
+
"step": 3600
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.887714663143989,
|
| 336 |
+
"grad_norm": 2.49125075340271,
|
| 337 |
+
"learning_rate": 9.967942540115829e-06,
|
| 338 |
+
"loss": 0.3508,
|
| 339 |
+
"step": 3700
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 5.019815059445178,
|
| 343 |
+
"grad_norm": 2.464569330215454,
|
| 344 |
+
"learning_rate": 9.965521243790079e-06,
|
| 345 |
+
"loss": 0.3355,
|
| 346 |
+
"step": 3800
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 5.1519154557463676,
|
| 350 |
+
"grad_norm": 2.6026785373687744,
|
| 351 |
+
"learning_rate": 9.963012121536635e-06,
|
| 352 |
+
"loss": 0.3284,
|
| 353 |
+
"step": 3900
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 5.284015852047556,
|
| 357 |
+
"grad_norm": 2.351313591003418,
|
| 358 |
+
"learning_rate": 9.96041521773472e-06,
|
| 359 |
+
"loss": 0.328,
|
| 360 |
+
"step": 4000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 5.284015852047556,
|
| 364 |
+
"eval_bleu": 40.737027833366824,
|
| 365 |
+
"eval_char_accuracy": 42.49156933706202,
|
| 366 |
+
"eval_loss": 0.23189863562583923,
|
| 367 |
+
"eval_runtime": 320.7889,
|
| 368 |
+
"eval_samples_per_second": 4.72,
|
| 369 |
+
"eval_steps_per_second": 0.592,
|
| 370 |
+
"step": 4000
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 5.416116248348745,
|
| 374 |
+
"grad_norm": 3.0100066661834717,
|
| 375 |
+
"learning_rate": 9.95773057831617e-06,
|
| 376 |
+
"loss": 0.311,
|
| 377 |
+
"step": 4100
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 5.5482166446499335,
|
| 381 |
+
"grad_norm": 2.049722671508789,
|
| 382 |
+
"learning_rate": 9.954958250764604e-06,
|
| 383 |
+
"loss": 0.3136,
|
| 384 |
+
"step": 4200
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 5.680317040951123,
|
| 388 |
+
"grad_norm": 2.1132702827453613,
|
| 389 |
+
"learning_rate": 9.952098284114604e-06,
|
| 390 |
+
"loss": 0.3,
|
| 391 |
+
"step": 4300
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 5.812417437252312,
|
| 395 |
+
"grad_norm": 2.174574613571167,
|
| 396 |
+
"learning_rate": 9.949150728950833e-06,
|
| 397 |
+
"loss": 0.3093,
|
| 398 |
+
"step": 4400
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 5.9445178335535,
|
| 402 |
+
"grad_norm": 2.8350446224212646,
|
| 403 |
+
"learning_rate": 9.946115637407145e-06,
|
| 404 |
+
"loss": 0.2988,
|
| 405 |
+
"step": 4500
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 5.9445178335535,
|
| 409 |
+
"eval_bleu": 43.8559984358133,
|
| 410 |
+
"eval_char_accuracy": 44.56273646981411,
|
| 411 |
+
"eval_loss": 0.21132107079029083,
|
| 412 |
+
"eval_runtime": 319.4845,
|
| 413 |
+
"eval_samples_per_second": 4.739,
|
| 414 |
+
"eval_steps_per_second": 0.595,
|
| 415 |
+
"step": 4500
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 6.07661822985469,
|
| 419 |
+
"grad_norm": 2.289716958999634,
|
| 420 |
+
"learning_rate": 9.94299306316567e-06,
|
| 421 |
+
"loss": 0.2938,
|
| 422 |
+
"step": 4600
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 6.208718626155878,
|
| 426 |
+
"grad_norm": 2.9307034015655518,
|
| 427 |
+
"learning_rate": 9.939783061455845e-06,
|
| 428 |
+
"loss": 0.2814,
|
| 429 |
+
"step": 4700
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 6.340819022457067,
|
| 433 |
+
"grad_norm": 2.3431613445281982,
|
| 434 |
+
"learning_rate": 9.936485689053462e-06,
|
| 435 |
+
"loss": 0.2782,
|
| 436 |
+
"step": 4800
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 6.472919418758257,
|
| 440 |
+
"grad_norm": 2.2339768409729004,
|
| 441 |
+
"learning_rate": 9.933101004279647e-06,
|
| 442 |
+
"loss": 0.2752,
|
| 443 |
+
"step": 4900
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 6.605019815059445,
|
| 447 |
+
"grad_norm": 2.076145887374878,
|
| 448 |
+
"learning_rate": 9.92962906699983e-06,
|
| 449 |
+
"loss": 0.265,
|
| 450 |
+
"step": 5000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 6.605019815059445,
|
| 454 |
+
"eval_bleu": 47.09645003475602,
|
| 455 |
+
"eval_char_accuracy": 46.29205050172725,
|
| 456 |
+
"eval_loss": 0.1954895406961441,
|
| 457 |
+
"eval_runtime": 322.9917,
|
| 458 |
+
"eval_samples_per_second": 4.687,
|
| 459 |
+
"eval_steps_per_second": 0.588,
|
| 460 |
+
"step": 5000
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 6.737120211360634,
|
| 464 |
+
"grad_norm": 1.715720295906067,
|
| 465 |
+
"learning_rate": 9.926069938622698e-06,
|
| 466 |
+
"loss": 0.266,
|
| 467 |
+
"step": 5100
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 6.869220607661823,
|
| 471 |
+
"grad_norm": 2.501234531402588,
|
| 472 |
+
"learning_rate": 9.922423682099088e-06,
|
| 473 |
+
"loss": 0.2633,
|
| 474 |
+
"step": 5200
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 7.001321003963012,
|
| 478 |
+
"grad_norm": 1.929123044013977,
|
| 479 |
+
"learning_rate": 9.918690361920898e-06,
|
| 480 |
+
"loss": 0.2584,
|
| 481 |
+
"step": 5300
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 7.133421400264201,
|
| 485 |
+
"grad_norm": 2.0370264053344727,
|
| 486 |
+
"learning_rate": 9.914870044119924e-06,
|
| 487 |
+
"loss": 0.2451,
|
| 488 |
+
"step": 5400
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 7.265521796565389,
|
| 492 |
+
"grad_norm": 2.3562278747558594,
|
| 493 |
+
"learning_rate": 9.91096279626671e-06,
|
| 494 |
+
"loss": 0.2476,
|
| 495 |
+
"step": 5500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 7.265521796565389,
|
| 499 |
+
"eval_bleu": 48.90758910970442,
|
| 500 |
+
"eval_char_accuracy": 46.93308932390195,
|
| 501 |
+
"eval_loss": 0.18339309096336365,
|
| 502 |
+
"eval_runtime": 328.302,
|
| 503 |
+
"eval_samples_per_second": 4.612,
|
| 504 |
+
"eval_steps_per_second": 0.579,
|
| 505 |
+
"step": 5500
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 7.397622192866579,
|
| 509 |
+
"grad_norm": 2.410529613494873,
|
| 510 |
+
"learning_rate": 9.90696868746934e-06,
|
| 511 |
+
"loss": 0.2419,
|
| 512 |
+
"step": 5600
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 7.5297225891677675,
|
| 516 |
+
"grad_norm": 1.685939908027649,
|
| 517 |
+
"learning_rate": 9.902887788372223e-06,
|
| 518 |
+
"loss": 0.2448,
|
| 519 |
+
"step": 5700
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 7.661822985468956,
|
| 523 |
+
"grad_norm": 2.3180549144744873,
|
| 524 |
+
"learning_rate": 9.89872017115484e-06,
|
| 525 |
+
"loss": 0.2379,
|
| 526 |
+
"step": 5800
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 7.793923381770146,
|
| 530 |
+
"grad_norm": 2.4159021377563477,
|
| 531 |
+
"learning_rate": 9.894465909530471e-06,
|
| 532 |
+
"loss": 0.2339,
|
| 533 |
+
"step": 5900
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 7.926023778071334,
|
| 537 |
+
"grad_norm": 2.0757477283477783,
|
| 538 |
+
"learning_rate": 9.890125078744884e-06,
|
| 539 |
+
"loss": 0.2356,
|
| 540 |
+
"step": 6000
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 7.926023778071334,
|
| 544 |
+
"eval_bleu": 51.273473767746786,
|
| 545 |
+
"eval_char_accuracy": 49.088563086033886,
|
| 546 |
+
"eval_loss": 0.1728673279285431,
|
| 547 |
+
"eval_runtime": 317.183,
|
| 548 |
+
"eval_samples_per_second": 4.773,
|
| 549 |
+
"eval_steps_per_second": 0.599,
|
| 550 |
+
"step": 6000
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 8.058124174372523,
|
| 554 |
+
"grad_norm": 1.9405860900878906,
|
| 555 |
+
"learning_rate": 9.885697755575015e-06,
|
| 556 |
+
"loss": 0.2251,
|
| 557 |
+
"step": 6100
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 8.190224570673712,
|
| 561 |
+
"grad_norm": 1.7473342418670654,
|
| 562 |
+
"learning_rate": 9.881184018327597e-06,
|
| 563 |
+
"loss": 0.2195,
|
| 564 |
+
"step": 6200
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 8.3223249669749,
|
| 568 |
+
"grad_norm": 1.7633724212646484,
|
| 569 |
+
"learning_rate": 9.876583946837787e-06,
|
| 570 |
+
"loss": 0.219,
|
| 571 |
+
"step": 6300
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 8.45442536327609,
|
| 575 |
+
"grad_norm": 2.1117053031921387,
|
| 576 |
+
"learning_rate": 9.871897622467748e-06,
|
| 577 |
+
"loss": 0.2148,
|
| 578 |
+
"step": 6400
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 8.58652575957728,
|
| 582 |
+
"grad_norm": 2.114854574203491,
|
| 583 |
+
"learning_rate": 9.867125128105211e-06,
|
| 584 |
+
"loss": 0.2222,
|
| 585 |
+
"step": 6500
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 8.58652575957728,
|
| 589 |
+
"eval_bleu": 53.15738681385049,
|
| 590 |
+
"eval_char_accuracy": 49.680251686132586,
|
| 591 |
+
"eval_loss": 0.1632871925830841,
|
| 592 |
+
"eval_runtime": 325.9103,
|
| 593 |
+
"eval_samples_per_second": 4.645,
|
| 594 |
+
"eval_steps_per_second": 0.583,
|
| 595 |
+
"step": 6500
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 8.718626155878468,
|
| 599 |
+
"grad_norm": 2.6163322925567627,
|
| 600 |
+
"learning_rate": 9.862266548162008e-06,
|
| 601 |
+
"loss": 0.2141,
|
| 602 |
+
"step": 6600
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 8.850726552179657,
|
| 606 |
+
"grad_norm": 2.1705501079559326,
|
| 607 |
+
"learning_rate": 9.857321968572577e-06,
|
| 608 |
+
"loss": 0.2126,
|
| 609 |
+
"step": 6700
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 8.982826948480845,
|
| 613 |
+
"grad_norm": 2.4352259635925293,
|
| 614 |
+
"learning_rate": 9.85229147679245e-06,
|
| 615 |
+
"loss": 0.2124,
|
| 616 |
+
"step": 6800
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 9.114927344782034,
|
| 620 |
+
"grad_norm": 1.912975788116455,
|
| 621 |
+
"learning_rate": 9.847175161796696e-06,
|
| 622 |
+
"loss": 0.2032,
|
| 623 |
+
"step": 6900
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 9.247027741083222,
|
| 627 |
+
"grad_norm": 1.7575359344482422,
|
| 628 |
+
"learning_rate": 9.841973114078358e-06,
|
| 629 |
+
"loss": 0.2005,
|
| 630 |
+
"step": 7000
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 9.247027741083222,
|
| 634 |
+
"eval_bleu": 54.81436388056976,
|
| 635 |
+
"eval_char_accuracy": 50.20562592531667,
|
| 636 |
+
"eval_loss": 0.1546466052532196,
|
| 637 |
+
"eval_runtime": 319.5902,
|
| 638 |
+
"eval_samples_per_second": 4.737,
|
| 639 |
+
"eval_steps_per_second": 0.595,
|
| 640 |
+
"step": 7000
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 9.379128137384413,
|
| 644 |
+
"grad_norm": 1.6994798183441162,
|
| 645 |
+
"learning_rate": 9.836685425646842e-06,
|
| 646 |
+
"loss": 0.1929,
|
| 647 |
+
"step": 7100
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 9.511228533685602,
|
| 651 |
+
"grad_norm": 1.8375500440597534,
|
| 652 |
+
"learning_rate": 9.831312190026295e-06,
|
| 653 |
+
"loss": 0.1954,
|
| 654 |
+
"step": 7200
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 9.64332892998679,
|
| 658 |
+
"grad_norm": 2.735320568084717,
|
| 659 |
+
"learning_rate": 9.825853502253951e-06,
|
| 660 |
+
"loss": 0.1949,
|
| 661 |
+
"step": 7300
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 9.775429326287979,
|
| 665 |
+
"grad_norm": 1.9880143404006958,
|
| 666 |
+
"learning_rate": 9.820309458878447e-06,
|
| 667 |
+
"loss": 0.196,
|
| 668 |
+
"step": 7400
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 9.907529722589167,
|
| 672 |
+
"grad_norm": 3.1160881519317627,
|
| 673 |
+
"learning_rate": 9.814680157958122e-06,
|
| 674 |
+
"loss": 0.1957,
|
| 675 |
+
"step": 7500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 9.907529722589167,
|
| 679 |
+
"eval_bleu": 56.684180257446236,
|
| 680 |
+
"eval_char_accuracy": 51.796656522454356,
|
| 681 |
+
"eval_loss": 0.14744216203689575,
|
| 682 |
+
"eval_runtime": 318.7305,
|
| 683 |
+
"eval_samples_per_second": 4.75,
|
| 684 |
+
"eval_steps_per_second": 0.596,
|
| 685 |
+
"step": 7500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 10.039630118890356,
|
| 689 |
+
"grad_norm": 1.975994348526001,
|
| 690 |
+
"learning_rate": 9.808965699059276e-06,
|
| 691 |
+
"loss": 0.1964,
|
| 692 |
+
"step": 7600
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 10.171730515191545,
|
| 696 |
+
"grad_norm": 1.6857510805130005,
|
| 697 |
+
"learning_rate": 9.80316618325441e-06,
|
| 698 |
+
"loss": 0.1832,
|
| 699 |
+
"step": 7700
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 10.303830911492735,
|
| 703 |
+
"grad_norm": 1.6473827362060547,
|
| 704 |
+
"learning_rate": 9.797281713120438e-06,
|
| 705 |
+
"loss": 0.1846,
|
| 706 |
+
"step": 7800
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 10.435931307793924,
|
| 710 |
+
"grad_norm": 2.1330363750457764,
|
| 711 |
+
"learning_rate": 9.79131239273688e-06,
|
| 712 |
+
"loss": 0.1783,
|
| 713 |
+
"step": 7900
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 10.568031704095112,
|
| 717 |
+
"grad_norm": 2.0598771572113037,
|
| 718 |
+
"learning_rate": 9.785258327684007e-06,
|
| 719 |
+
"loss": 0.183,
|
| 720 |
+
"step": 8000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 10.568031704095112,
|
| 724 |
+
"eval_bleu": 57.92928420103779,
|
| 725 |
+
"eval_char_accuracy": 51.580235236058556,
|
| 726 |
+
"eval_loss": 0.1413801610469818,
|
| 727 |
+
"eval_runtime": 318.3694,
|
| 728 |
+
"eval_samples_per_second": 4.755,
|
| 729 |
+
"eval_steps_per_second": 0.597,
|
| 730 |
+
"step": 8000
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 10.700132100396301,
|
| 734 |
+
"grad_norm": 1.8669029474258423,
|
| 735 |
+
"learning_rate": 9.779119625040988e-06,
|
| 736 |
+
"loss": 0.1801,
|
| 737 |
+
"step": 8100
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 10.83223249669749,
|
| 741 |
+
"grad_norm": 1.967623233795166,
|
| 742 |
+
"learning_rate": 9.772896393383991e-06,
|
| 743 |
+
"loss": 0.1772,
|
| 744 |
+
"step": 8200
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 10.964332892998678,
|
| 748 |
+
"grad_norm": 2.0888118743896484,
|
| 749 |
+
"learning_rate": 9.766588742784255e-06,
|
| 750 |
+
"loss": 0.1741,
|
| 751 |
+
"step": 8300
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 11.096433289299869,
|
| 755 |
+
"grad_norm": 1.8615264892578125,
|
| 756 |
+
"learning_rate": 9.760196784806155e-06,
|
| 757 |
+
"loss": 0.1733,
|
| 758 |
+
"step": 8400
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 11.228533685601057,
|
| 762 |
+
"grad_norm": 1.786023736000061,
|
| 763 |
+
"learning_rate": 9.753720632505219e-06,
|
| 764 |
+
"loss": 0.171,
|
| 765 |
+
"step": 8500
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 11.228533685601057,
|
| 769 |
+
"eval_bleu": 58.97794662773473,
|
| 770 |
+
"eval_char_accuracy": 53.53111120250041,
|
| 771 |
+
"eval_loss": 0.13721999526023865,
|
| 772 |
+
"eval_runtime": 310.3442,
|
| 773 |
+
"eval_samples_per_second": 4.878,
|
| 774 |
+
"eval_steps_per_second": 0.612,
|
| 775 |
+
"step": 8500
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 11.360634081902246,
|
| 779 |
+
"grad_norm": 2.22037935256958,
|
| 780 |
+
"learning_rate": 9.74716040042614e-06,
|
| 781 |
+
"loss": 0.169,
|
| 782 |
+
"step": 8600
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 11.492734478203435,
|
| 786 |
+
"grad_norm": 2.2769389152526855,
|
| 787 |
+
"learning_rate": 9.740516204600734e-06,
|
| 788 |
+
"loss": 0.1631,
|
| 789 |
+
"step": 8700
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 11.624834874504623,
|
| 793 |
+
"grad_norm": 2.1513566970825195,
|
| 794 |
+
"learning_rate": 9.733788162545902e-06,
|
| 795 |
+
"loss": 0.1669,
|
| 796 |
+
"step": 8800
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 11.756935270805812,
|
| 800 |
+
"grad_norm": 1.640358328819275,
|
| 801 |
+
"learning_rate": 9.726976393261547e-06,
|
| 802 |
+
"loss": 0.1674,
|
| 803 |
+
"step": 8900
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 11.889035667107,
|
| 807 |
+
"grad_norm": 1.6976934671401978,
|
| 808 |
+
"learning_rate": 9.720081017228462e-06,
|
| 809 |
+
"loss": 0.1646,
|
| 810 |
+
"step": 9000
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 11.889035667107,
|
| 814 |
+
"eval_bleu": 60.38328284328657,
|
| 815 |
+
"eval_char_accuracy": 54.70215084717881,
|
| 816 |
+
"eval_loss": 0.13088105618953705,
|
| 817 |
+
"eval_runtime": 316.5706,
|
| 818 |
+
"eval_samples_per_second": 4.783,
|
| 819 |
+
"eval_steps_per_second": 0.6,
|
| 820 |
+
"step": 9000
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 12.021136063408191,
|
| 824 |
+
"grad_norm": 1.7420779466629028,
|
| 825 |
+
"learning_rate": 9.713102156406213e-06,
|
| 826 |
+
"loss": 0.1629,
|
| 827 |
+
"step": 9100
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 12.15323645970938,
|
| 831 |
+
"grad_norm": 1.9723796844482422,
|
| 832 |
+
"learning_rate": 9.706039934230967e-06,
|
| 833 |
+
"loss": 0.1578,
|
| 834 |
+
"step": 9200
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 12.285336856010568,
|
| 838 |
+
"grad_norm": 2.3517324924468994,
|
| 839 |
+
"learning_rate": 9.698894475613323e-06,
|
| 840 |
+
"loss": 0.1561,
|
| 841 |
+
"step": 9300
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 12.417437252311757,
|
| 845 |
+
"grad_norm": 1.5132865905761719,
|
| 846 |
+
"learning_rate": 9.691665906936088e-06,
|
| 847 |
+
"loss": 0.157,
|
| 848 |
+
"step": 9400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 12.549537648612946,
|
| 852 |
+
"grad_norm": 2.435624599456787,
|
| 853 |
+
"learning_rate": 9.684354356052055e-06,
|
| 854 |
+
"loss": 0.1538,
|
| 855 |
+
"step": 9500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 12.549537648612946,
|
| 859 |
+
"eval_bleu": 61.08870144349316,
|
| 860 |
+
"eval_char_accuracy": 55.10055107747986,
|
| 861 |
+
"eval_loss": 0.12675440311431885,
|
| 862 |
+
"eval_runtime": 314.3152,
|
| 863 |
+
"eval_samples_per_second": 4.817,
|
| 864 |
+
"eval_steps_per_second": 0.604,
|
| 865 |
+
"step": 9500
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 12.681638044914134,
|
| 869 |
+
"grad_norm": 2.092256784439087,
|
| 870 |
+
"learning_rate": 9.676959952281733e-06,
|
| 871 |
+
"loss": 0.1518,
|
| 872 |
+
"step": 9600
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 12.813738441215325,
|
| 876 |
+
"grad_norm": 1.7985179424285889,
|
| 877 |
+
"learning_rate": 9.669482826411065e-06,
|
| 878 |
+
"loss": 0.158,
|
| 879 |
+
"step": 9700
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 12.945838837516513,
|
| 883 |
+
"grad_norm": 1.7571889162063599,
|
| 884 |
+
"learning_rate": 9.66192311068911e-06,
|
| 885 |
+
"loss": 0.152,
|
| 886 |
+
"step": 9800
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 13.077939233817702,
|
| 890 |
+
"grad_norm": 1.3526843786239624,
|
| 891 |
+
"learning_rate": 9.654280938825705e-06,
|
| 892 |
+
"loss": 0.1426,
|
| 893 |
+
"step": 9900
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 13.21003963011889,
|
| 897 |
+
"grad_norm": 1.6651784181594849,
|
| 898 |
+
"learning_rate": 9.646556445989106e-06,
|
| 899 |
+
"loss": 0.1476,
|
| 900 |
+
"step": 10000
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 13.21003963011889,
|
| 904 |
+
"eval_bleu": 61.91608356246592,
|
| 905 |
+
"eval_char_accuracy": 55.006477216647475,
|
| 906 |
+
"eval_loss": 0.12389995902776718,
|
| 907 |
+
"eval_runtime": 315.2745,
|
| 908 |
+
"eval_samples_per_second": 4.802,
|
| 909 |
+
"eval_steps_per_second": 0.603,
|
| 910 |
+
"step": 10000
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 13.34214002642008,
|
| 914 |
+
"grad_norm": 1.5739635229110718,
|
| 915 |
+
"learning_rate": 9.63874976880359e-06,
|
| 916 |
+
"loss": 0.148,
|
| 917 |
+
"step": 10100
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 13.474240422721268,
|
| 921 |
+
"grad_norm": 1.779477834701538,
|
| 922 |
+
"learning_rate": 9.63086104534704e-06,
|
| 923 |
+
"loss": 0.1469,
|
| 924 |
+
"step": 10200
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 13.606340819022456,
|
| 928 |
+
"grad_norm": 1.5422449111938477,
|
| 929 |
+
"learning_rate": 9.622890415148505e-06,
|
| 930 |
+
"loss": 0.143,
|
| 931 |
+
"step": 10300
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 13.738441215323647,
|
| 935 |
+
"grad_norm": 1.80446457862854,
|
| 936 |
+
"learning_rate": 9.61483801918573e-06,
|
| 937 |
+
"loss": 0.1424,
|
| 938 |
+
"step": 10400
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 13.870541611624835,
|
| 942 |
+
"grad_norm": 1.7641818523406982,
|
| 943 |
+
"learning_rate": 9.606703999882667e-06,
|
| 944 |
+
"loss": 0.1406,
|
| 945 |
+
"step": 10500
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 13.870541611624835,
|
| 949 |
+
"eval_bleu": 62.97532147351367,
|
| 950 |
+
"eval_char_accuracy": 56.364636453364035,
|
| 951 |
+
"eval_loss": 0.12048687040805817,
|
| 952 |
+
"eval_runtime": 315.4459,
|
| 953 |
+
"eval_samples_per_second": 4.8,
|
| 954 |
+
"eval_steps_per_second": 0.602,
|
| 955 |
+
"step": 10500
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 14.002642007926024,
|
| 959 |
+
"grad_norm": 1.5506337881088257,
|
| 960 |
+
"learning_rate": 9.598488501106947e-06,
|
| 961 |
+
"loss": 0.1436,
|
| 962 |
+
"step": 10600
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 14.134742404227213,
|
| 966 |
+
"grad_norm": 1.338537335395813,
|
| 967 |
+
"learning_rate": 9.590191668167343e-06,
|
| 968 |
+
"loss": 0.1396,
|
| 969 |
+
"step": 10700
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 14.266842800528401,
|
| 973 |
+
"grad_norm": 1.9432475566864014,
|
| 974 |
+
"learning_rate": 9.581813647811199e-06,
|
| 975 |
+
"loss": 0.1427,
|
| 976 |
+
"step": 10800
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 14.39894319682959,
|
| 980 |
+
"grad_norm": 1.8222265243530273,
|
| 981 |
+
"learning_rate": 9.573354588221833e-06,
|
| 982 |
+
"loss": 0.1352,
|
| 983 |
+
"step": 10900
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 14.531043593130779,
|
| 987 |
+
"grad_norm": 1.7141766548156738,
|
| 988 |
+
"learning_rate": 9.564814639015915e-06,
|
| 989 |
+
"loss": 0.1361,
|
| 990 |
+
"step": 11000
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 14.531043593130779,
|
| 994 |
+
"eval_bleu": 63.45481969831487,
|
| 995 |
+
"eval_char_accuracy": 56.61755634150354,
|
| 996 |
+
"eval_loss": 0.11702162027359009,
|
| 997 |
+
"eval_runtime": 318.0259,
|
| 998 |
+
"eval_samples_per_second": 4.761,
|
| 999 |
+
"eval_steps_per_second": 0.597,
|
| 1000 |
+
"step": 11000
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 14.663143989431969,
|
| 1004 |
+
"grad_norm": 1.473866581916809,
|
| 1005 |
+
"learning_rate": 9.556193951240821e-06,
|
| 1006 |
+
"loss": 0.1302,
|
| 1007 |
+
"step": 11100
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 14.795244385733158,
|
| 1011 |
+
"grad_norm": 1.3840774297714233,
|
| 1012 |
+
"learning_rate": 9.547492677371968e-06,
|
| 1013 |
+
"loss": 0.1355,
|
| 1014 |
+
"step": 11200
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 14.927344782034346,
|
| 1018 |
+
"grad_norm": 1.8060030937194824,
|
| 1019 |
+
"learning_rate": 9.538710971310104e-06,
|
| 1020 |
+
"loss": 0.1332,
|
| 1021 |
+
"step": 11300
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 15.059445178335535,
|
| 1025 |
+
"grad_norm": 1.8242031335830688,
|
| 1026 |
+
"learning_rate": 9.529848988378597e-06,
|
| 1027 |
+
"loss": 0.1247,
|
| 1028 |
+
"step": 11400
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 15.191545574636724,
|
| 1032 |
+
"grad_norm": 1.3062950372695923,
|
| 1033 |
+
"learning_rate": 9.520906885320682e-06,
|
| 1034 |
+
"loss": 0.1295,
|
| 1035 |
+
"step": 11500
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 15.191545574636724,
|
| 1039 |
+
"eval_bleu": 64.22910948855701,
|
| 1040 |
+
"eval_char_accuracy": 57.40099111696003,
|
| 1041 |
+
"eval_loss": 0.11363621801137924,
|
| 1042 |
+
"eval_runtime": 315.2122,
|
| 1043 |
+
"eval_samples_per_second": 4.803,
|
| 1044 |
+
"eval_steps_per_second": 0.603,
|
| 1045 |
+
"step": 11500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 15.323645970937912,
|
| 1049 |
+
"grad_norm": 1.5021114349365234,
|
| 1050 |
+
"learning_rate": 9.511884820296695e-06,
|
| 1051 |
+
"loss": 0.1292,
|
| 1052 |
+
"step": 11600
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 15.455746367239101,
|
| 1056 |
+
"grad_norm": 1.8947999477386475,
|
| 1057 |
+
"learning_rate": 9.502782952881268e-06,
|
| 1058 |
+
"loss": 0.128,
|
| 1059 |
+
"step": 11700
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 15.587846763540291,
|
| 1063 |
+
"grad_norm": 1.5116643905639648,
|
| 1064 |
+
"learning_rate": 9.493601444060514e-06,
|
| 1065 |
+
"loss": 0.1276,
|
| 1066 |
+
"step": 11800
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 15.71994715984148,
|
| 1070 |
+
"grad_norm": 1.5134871006011963,
|
| 1071 |
+
"learning_rate": 9.48434045622917e-06,
|
| 1072 |
+
"loss": 0.132,
|
| 1073 |
+
"step": 11900
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 15.852047556142669,
|
| 1077 |
+
"grad_norm": 1.7438267469406128,
|
| 1078 |
+
"learning_rate": 9.475000153187733e-06,
|
| 1079 |
+
"loss": 0.1243,
|
| 1080 |
+
"step": 12000
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 15.852047556142669,
|
| 1084 |
+
"eval_bleu": 65.05289431113529,
|
| 1085 |
+
"eval_char_accuracy": 57.989081263365684,
|
| 1086 |
+
"eval_loss": 0.1115257665514946,
|
| 1087 |
+
"eval_runtime": 309.8437,
|
| 1088 |
+
"eval_samples_per_second": 4.886,
|
| 1089 |
+
"eval_steps_per_second": 0.613,
|
| 1090 |
+
"step": 12000
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 15.984147952443857,
|
| 1094 |
+
"grad_norm": 1.913855791091919,
|
| 1095 |
+
"learning_rate": 9.46558070013956e-06,
|
| 1096 |
+
"loss": 0.1261,
|
| 1097 |
+
"step": 12100
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 16.116248348745046,
|
| 1101 |
+
"grad_norm": 1.4257742166519165,
|
| 1102 |
+
"learning_rate": 9.456082263687946e-06,
|
| 1103 |
+
"loss": 0.117,
|
| 1104 |
+
"step": 12200
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 16.248348745046236,
|
| 1108 |
+
"grad_norm": 1.49489164352417,
|
| 1109 |
+
"learning_rate": 9.44650501183318e-06,
|
| 1110 |
+
"loss": 0.1232,
|
| 1111 |
+
"step": 12300
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 16.380449141347423,
|
| 1115 |
+
"grad_norm": 1.7977708578109741,
|
| 1116 |
+
"learning_rate": 9.436849113969567e-06,
|
| 1117 |
+
"loss": 0.1212,
|
| 1118 |
+
"step": 12400
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 16.512549537648614,
|
| 1122 |
+
"grad_norm": 1.2978798151016235,
|
| 1123 |
+
"learning_rate": 9.427212472501483e-06,
|
| 1124 |
+
"loss": 0.122,
|
| 1125 |
+
"step": 12500
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 16.512549537648614,
|
| 1129 |
+
"eval_bleu": 65.59844085659223,
|
| 1130 |
+
"eval_char_accuracy": 58.32784997532489,
|
| 1131 |
+
"eval_loss": 0.10947112739086151,
|
| 1132 |
+
"eval_runtime": 332.8226,
|
| 1133 |
+
"eval_samples_per_second": 4.549,
|
| 1134 |
+
"eval_steps_per_second": 0.571,
|
| 1135 |
+
"step": 12500
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 16.6446499339498,
|
| 1139 |
+
"grad_norm": 1.5628472566604614,
|
| 1140 |
+
"learning_rate": 9.417400578537868e-06,
|
| 1141 |
+
"loss": 0.1219,
|
| 1142 |
+
"step": 12600
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 16.77675033025099,
|
| 1146 |
+
"grad_norm": 1.5304769277572632,
|
| 1147 |
+
"learning_rate": 9.407510553339931e-06,
|
| 1148 |
+
"loss": 0.1192,
|
| 1149 |
+
"step": 12700
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 16.90885072655218,
|
| 1153 |
+
"grad_norm": 1.6370787620544434,
|
| 1154 |
+
"learning_rate": 9.397542571834054e-06,
|
| 1155 |
+
"loss": 0.1181,
|
| 1156 |
+
"step": 12800
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 17.040951122853368,
|
| 1160 |
+
"grad_norm": 1.5388678312301636,
|
| 1161 |
+
"learning_rate": 9.387496810325436e-06,
|
| 1162 |
+
"loss": 0.1137,
|
| 1163 |
+
"step": 12900
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 17.17305151915456,
|
| 1167 |
+
"grad_norm": 1.8251888751983643,
|
| 1168 |
+
"learning_rate": 9.377373446494984e-06,
|
| 1169 |
+
"loss": 0.1122,
|
| 1170 |
+
"step": 13000
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 17.17305151915456,
|
| 1174 |
+
"eval_bleu": 66.79852257149867,
|
| 1175 |
+
"eval_char_accuracy": 58.47692877117947,
|
| 1176 |
+
"eval_loss": 0.1069813147187233,
|
| 1177 |
+
"eval_runtime": 350.9901,
|
| 1178 |
+
"eval_samples_per_second": 4.314,
|
| 1179 |
+
"eval_steps_per_second": 0.541,
|
| 1180 |
+
"step": 13000
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 17.305151915455745,
|
| 1184 |
+
"grad_norm": 1.6730031967163086,
|
| 1185 |
+
"learning_rate": 9.367172659396172e-06,
|
| 1186 |
+
"loss": 0.1123,
|
| 1187 |
+
"step": 13100
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 17.437252311756936,
|
| 1191 |
+
"grad_norm": 1.2402830123901367,
|
| 1192 |
+
"learning_rate": 9.35689462945187e-06,
|
| 1193 |
+
"loss": 0.1125,
|
| 1194 |
+
"step": 13200
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 17.569352708058123,
|
| 1198 |
+
"grad_norm": 1.9425466060638428,
|
| 1199 |
+
"learning_rate": 9.34653953845115e-06,
|
| 1200 |
+
"loss": 0.1109,
|
| 1201 |
+
"step": 13300
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 17.701453104359313,
|
| 1205 |
+
"grad_norm": 3.105457067489624,
|
| 1206 |
+
"learning_rate": 9.33610756954608e-06,
|
| 1207 |
+
"loss": 0.1161,
|
| 1208 |
+
"step": 13400
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 17.833553500660503,
|
| 1212 |
+
"grad_norm": 1.679442048072815,
|
| 1213 |
+
"learning_rate": 9.325598907248478e-06,
|
| 1214 |
+
"loss": 0.1131,
|
| 1215 |
+
"step": 13500
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 17.833553500660503,
|
| 1219 |
+
"eval_bleu": 66.72469802657548,
|
| 1220 |
+
"eval_char_accuracy": 59.26910264846191,
|
| 1221 |
+
"eval_loss": 0.10426344722509384,
|
| 1222 |
+
"eval_runtime": 353.0152,
|
| 1223 |
+
"eval_samples_per_second": 4.289,
|
| 1224 |
+
"eval_steps_per_second": 0.538,
|
| 1225 |
+
"step": 13500
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 17.96565389696169,
|
| 1229 |
+
"grad_norm": 1.1806349754333496,
|
| 1230 |
+
"learning_rate": 9.315013737426645e-06,
|
| 1231 |
+
"loss": 0.115,
|
| 1232 |
+
"step": 13600
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 18.09775429326288,
|
| 1236 |
+
"grad_norm": 1.5437259674072266,
|
| 1237 |
+
"learning_rate": 9.304352247302091e-06,
|
| 1238 |
+
"loss": 0.1071,
|
| 1239 |
+
"step": 13700
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 18.229854689564068,
|
| 1243 |
+
"grad_norm": 3.9584083557128906,
|
| 1244 |
+
"learning_rate": 9.293614625446205e-06,
|
| 1245 |
+
"loss": 0.11,
|
| 1246 |
+
"step": 13800
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 18.361955085865258,
|
| 1250 |
+
"grad_norm": 1.4071292877197266,
|
| 1251 |
+
"learning_rate": 9.282801061776937e-06,
|
| 1252 |
+
"loss": 0.1093,
|
| 1253 |
+
"step": 13900
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 18.494055482166445,
|
| 1257 |
+
"grad_norm": 1.599787950515747,
|
| 1258 |
+
"learning_rate": 9.271911747555425e-06,
|
| 1259 |
+
"loss": 0.1057,
|
| 1260 |
+
"step": 14000
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 18.494055482166445,
|
| 1264 |
+
"eval_bleu": 67.26763967365643,
|
| 1265 |
+
"eval_char_accuracy": 58.966318473433134,
|
| 1266 |
+
"eval_loss": 0.10266197472810745,
|
| 1267 |
+
"eval_runtime": 318.6046,
|
| 1268 |
+
"eval_samples_per_second": 4.752,
|
| 1269 |
+
"eval_steps_per_second": 0.596,
|
| 1270 |
+
"step": 14000
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 18.626155878467635,
|
| 1274 |
+
"grad_norm": 1.3243365287780762,
|
| 1275 |
+
"learning_rate": 9.260946875382624e-06,
|
| 1276 |
+
"loss": 0.1054,
|
| 1277 |
+
"step": 14100
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 18.758256274768826,
|
| 1281 |
+
"grad_norm": 1.8453656435012817,
|
| 1282 |
+
"learning_rate": 9.249906639195894e-06,
|
| 1283 |
+
"loss": 0.1096,
|
| 1284 |
+
"step": 14200
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 18.890356671070013,
|
| 1288 |
+
"grad_norm": 1.2308754920959473,
|
| 1289 |
+
"learning_rate": 9.238791234265565e-06,
|
| 1290 |
+
"loss": 0.1045,
|
| 1291 |
+
"step": 14300
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 19.022457067371203,
|
| 1295 |
+
"grad_norm": 1.3015666007995605,
|
| 1296 |
+
"learning_rate": 9.22760085719149e-06,
|
| 1297 |
+
"loss": 0.1065,
|
| 1298 |
+
"step": 14400
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 19.15455746367239,
|
| 1302 |
+
"grad_norm": 1.3568576574325562,
|
| 1303 |
+
"learning_rate": 9.21633570589957e-06,
|
| 1304 |
+
"loss": 0.1024,
|
| 1305 |
+
"step": 14500
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 19.15455746367239,
|
| 1309 |
+
"eval_bleu": 67.65790078495372,
|
| 1310 |
+
"eval_char_accuracy": 59.99547622964303,
|
| 1311 |
+
"eval_loss": 0.10274580866098404,
|
| 1312 |
+
"eval_runtime": 317.6719,
|
| 1313 |
+
"eval_samples_per_second": 4.766,
|
| 1314 |
+
"eval_steps_per_second": 0.598,
|
| 1315 |
+
"step": 14500
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 19.28665785997358,
|
| 1319 |
+
"grad_norm": 1.6260790824890137,
|
| 1320 |
+
"learning_rate": 9.204995979638241e-06,
|
| 1321 |
+
"loss": 0.1025,
|
| 1322 |
+
"step": 14600
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 19.418758256274767,
|
| 1326 |
+
"grad_norm": 1.1852062940597534,
|
| 1327 |
+
"learning_rate": 9.193581878974964e-06,
|
| 1328 |
+
"loss": 0.101,
|
| 1329 |
+
"step": 14700
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 19.550858652575958,
|
| 1333 |
+
"grad_norm": 2.5521767139434814,
|
| 1334 |
+
"learning_rate": 9.18209360579267e-06,
|
| 1335 |
+
"loss": 0.0999,
|
| 1336 |
+
"step": 14800
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 19.682959048877148,
|
| 1340 |
+
"grad_norm": 1.0231155157089233,
|
| 1341 |
+
"learning_rate": 9.17053136328619e-06,
|
| 1342 |
+
"loss": 0.1055,
|
| 1343 |
+
"step": 14900
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 19.815059445178335,
|
| 1347 |
+
"grad_norm": 1.3518396615982056,
|
| 1348 |
+
"learning_rate": 9.15889535595866e-06,
|
| 1349 |
+
"loss": 0.1014,
|
| 1350 |
+
"step": 15000
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 19.815059445178335,
|
| 1354 |
+
"eval_bleu": 68.43528395628573,
|
| 1355 |
+
"eval_char_accuracy": 61.64870866918901,
|
| 1356 |
+
"eval_loss": 0.09970895200967789,
|
| 1357 |
+
"eval_runtime": 314.0763,
|
| 1358 |
+
"eval_samples_per_second": 4.82,
|
| 1359 |
+
"eval_steps_per_second": 0.605,
|
| 1360 |
+
"step": 15000
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 19.947159841479525,
|
| 1364 |
+
"grad_norm": 1.2678903341293335,
|
| 1365 |
+
"learning_rate": 9.147185789617907e-06,
|
| 1366 |
+
"loss": 0.1005,
|
| 1367 |
+
"step": 15100
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 20.079260237780712,
|
| 1371 |
+
"grad_norm": 1.2439242601394653,
|
| 1372 |
+
"learning_rate": 9.13540287137281e-06,
|
| 1373 |
+
"loss": 0.0989,
|
| 1374 |
+
"step": 15200
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 20.211360634081903,
|
| 1378 |
+
"grad_norm": 1.3169798851013184,
|
| 1379 |
+
"learning_rate": 9.123546809629632e-06,
|
| 1380 |
+
"loss": 0.1006,
|
| 1381 |
+
"step": 15300
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 20.34346103038309,
|
| 1385 |
+
"grad_norm": 1.3063526153564453,
|
| 1386 |
+
"learning_rate": 9.111617814088332e-06,
|
| 1387 |
+
"loss": 0.0966,
|
| 1388 |
+
"step": 15400
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 20.47556142668428,
|
| 1392 |
+
"grad_norm": 1.8397212028503418,
|
| 1393 |
+
"learning_rate": 9.099616095738867e-06,
|
| 1394 |
+
"loss": 0.0965,
|
| 1395 |
+
"step": 15500
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 20.47556142668428,
|
| 1399 |
+
"eval_bleu": 68.40618966106221,
|
| 1400 |
+
"eval_char_accuracy": 60.30751357131107,
|
| 1401 |
+
"eval_loss": 0.09965521842241287,
|
| 1402 |
+
"eval_runtime": 315.776,
|
| 1403 |
+
"eval_samples_per_second": 4.795,
|
| 1404 |
+
"eval_steps_per_second": 0.602,
|
| 1405 |
+
"step": 15500
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 20.60766182298547,
|
| 1409 |
+
"grad_norm": 1.544118046760559,
|
| 1410 |
+
"learning_rate": 9.087541866857453e-06,
|
| 1411 |
+
"loss": 0.0954,
|
| 1412 |
+
"step": 15600
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 20.739762219286657,
|
| 1416 |
+
"grad_norm": 1.5378237962722778,
|
| 1417 |
+
"learning_rate": 9.075395341002804e-06,
|
| 1418 |
+
"loss": 0.0975,
|
| 1419 |
+
"step": 15700
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 20.871862615587848,
|
| 1423 |
+
"grad_norm": 1.0197510719299316,
|
| 1424 |
+
"learning_rate": 9.06317673301237e-06,
|
| 1425 |
+
"loss": 0.0964,
|
| 1426 |
+
"step": 15800
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 21.003963011889034,
|
| 1430 |
+
"grad_norm": 0.9621543884277344,
|
| 1431 |
+
"learning_rate": 9.05088625899852e-06,
|
| 1432 |
+
"loss": 0.0925,
|
| 1433 |
+
"step": 15900
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 21.136063408190225,
|
| 1437 |
+
"grad_norm": 1.356550931930542,
|
| 1438 |
+
"learning_rate": 9.038524136344736e-06,
|
| 1439 |
+
"loss": 0.0917,
|
| 1440 |
+
"step": 16000
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 21.136063408190225,
|
| 1444 |
+
"eval_bleu": 69.05043210174456,
|
| 1445 |
+
"eval_char_accuracy": 60.879153643691396,
|
| 1446 |
+
"eval_loss": 0.0960288867354393,
|
| 1447 |
+
"eval_runtime": 317.6216,
|
| 1448 |
+
"eval_samples_per_second": 4.767,
|
| 1449 |
+
"eval_steps_per_second": 0.598,
|
| 1450 |
+
"step": 16000
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 21.268163804491415,
|
| 1454 |
+
"grad_norm": 2.0014472007751465,
|
| 1455 |
+
"learning_rate": 9.026090583701755e-06,
|
| 1456 |
+
"loss": 0.0962,
|
| 1457 |
+
"step": 16100
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 21.400264200792602,
|
| 1461 |
+
"grad_norm": 1.4762715101242065,
|
| 1462 |
+
"learning_rate": 9.013585820983713e-06,
|
| 1463 |
+
"loss": 0.0917,
|
| 1464 |
+
"step": 16200
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 21.532364597093792,
|
| 1468 |
+
"grad_norm": 1.242245078086853,
|
| 1469 |
+
"learning_rate": 9.001010069364241e-06,
|
| 1470 |
+
"loss": 0.0907,
|
| 1471 |
+
"step": 16300
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 21.66446499339498,
|
| 1475 |
+
"grad_norm": 1.9390721321105957,
|
| 1476 |
+
"learning_rate": 8.98836355127257e-06,
|
| 1477 |
+
"loss": 0.0918,
|
| 1478 |
+
"step": 16400
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 21.79656538969617,
|
| 1482 |
+
"grad_norm": 1.070357322692871,
|
| 1483 |
+
"learning_rate": 8.975646490389581e-06,
|
| 1484 |
+
"loss": 0.0903,
|
| 1485 |
+
"step": 16500
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 21.79656538969617,
|
| 1489 |
+
"eval_bleu": 69.68531857632678,
|
| 1490 |
+
"eval_char_accuracy": 62.48920463892087,
|
| 1491 |
+
"eval_loss": 0.09554192423820496,
|
| 1492 |
+
"eval_runtime": 335.0637,
|
| 1493 |
+
"eval_samples_per_second": 4.519,
|
| 1494 |
+
"eval_steps_per_second": 0.567,
|
| 1495 |
+
"step": 16500
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 21.928665785997357,
|
| 1499 |
+
"grad_norm": 1.8990777730941772,
|
| 1500 |
+
"learning_rate": 8.962859111643862e-06,
|
| 1501 |
+
"loss": 0.0946,
|
| 1502 |
+
"step": 16600
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 22.060766182298547,
|
| 1506 |
+
"grad_norm": 1.6244333982467651,
|
| 1507 |
+
"learning_rate": 8.950001641207719e-06,
|
| 1508 |
+
"loss": 0.0895,
|
| 1509 |
+
"step": 16700
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 22.192866578599737,
|
| 1513 |
+
"grad_norm": 1.6511205434799194,
|
| 1514 |
+
"learning_rate": 8.937074306493187e-06,
|
| 1515 |
+
"loss": 0.0907,
|
| 1516 |
+
"step": 16800
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 22.324966974900924,
|
| 1520 |
+
"grad_norm": 1.421342372894287,
|
| 1521 |
+
"learning_rate": 8.924077336147992e-06,
|
| 1522 |
+
"loss": 0.0864,
|
| 1523 |
+
"step": 16900
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 22.457067371202115,
|
| 1527 |
+
"grad_norm": 1.572800874710083,
|
| 1528 |
+
"learning_rate": 8.911010960051522e-06,
|
| 1529 |
+
"loss": 0.088,
|
| 1530 |
+
"step": 17000
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 22.457067371202115,
|
| 1534 |
+
"eval_bleu": 69.94746089002493,
|
| 1535 |
+
"eval_char_accuracy": 61.5572051324231,
|
| 1536 |
+
"eval_loss": 0.09443064033985138,
|
| 1537 |
+
"eval_runtime": 345.5738,
|
| 1538 |
+
"eval_samples_per_second": 4.381,
|
| 1539 |
+
"eval_steps_per_second": 0.55,
|
| 1540 |
+
"step": 17000
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 22.5891677675033,
|
| 1544 |
+
"grad_norm": 1.2412383556365967,
|
| 1545 |
+
"learning_rate": 8.897875409310755e-06,
|
| 1546 |
+
"loss": 0.085,
|
| 1547 |
+
"step": 17100
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 22.721268163804492,
|
| 1551 |
+
"grad_norm": 1.4429903030395508,
|
| 1552 |
+
"learning_rate": 8.884803301685314e-06,
|
| 1553 |
+
"loss": 0.0908,
|
| 1554 |
+
"step": 17200
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 22.85336856010568,
|
| 1558 |
+
"grad_norm": 1.1933690309524536,
|
| 1559 |
+
"learning_rate": 8.871530785794356e-06,
|
| 1560 |
+
"loss": 0.092,
|
| 1561 |
+
"step": 17300
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 22.98546895640687,
|
| 1565 |
+
"grad_norm": 1.399186611175537,
|
| 1566 |
+
"learning_rate": 8.85818979355093e-06,
|
| 1567 |
+
"loss": 0.0837,
|
| 1568 |
+
"step": 17400
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 23.11756935270806,
|
| 1572 |
+
"grad_norm": 1.2939783334732056,
|
| 1573 |
+
"learning_rate": 8.844780560919194e-06,
|
| 1574 |
+
"loss": 0.0871,
|
| 1575 |
+
"step": 17500
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 23.11756935270806,
|
| 1579 |
+
"eval_bleu": 70.43896205976631,
|
| 1580 |
+
"eval_char_accuracy": 61.825032900148045,
|
| 1581 |
+
"eval_loss": 0.0947960913181305,
|
| 1582 |
+
"eval_runtime": 324.1925,
|
| 1583 |
+
"eval_samples_per_second": 4.67,
|
| 1584 |
+
"eval_steps_per_second": 0.586,
|
| 1585 |
+
"step": 17500
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 23.249669749009247,
|
| 1589 |
+
"grad_norm": 1.2098222970962524,
|
| 1590 |
+
"learning_rate": 8.831303325070279e-06,
|
| 1591 |
+
"loss": 0.0827,
|
| 1592 |
+
"step": 17600
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 23.381770145310437,
|
| 1596 |
+
"grad_norm": 1.5045851469039917,
|
| 1597 |
+
"learning_rate": 8.8177583243781e-06,
|
| 1598 |
+
"loss": 0.0838,
|
| 1599 |
+
"step": 17700
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 23.513870541611624,
|
| 1603 |
+
"grad_norm": 1.5295897722244263,
|
| 1604 |
+
"learning_rate": 8.80414579841514e-06,
|
| 1605 |
+
"loss": 0.0858,
|
| 1606 |
+
"step": 17800
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 23.645970937912814,
|
| 1610 |
+
"grad_norm": 1.4860461950302124,
|
| 1611 |
+
"learning_rate": 8.790465987948212e-06,
|
| 1612 |
+
"loss": 0.0875,
|
| 1613 |
+
"step": 17900
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 23.778071334214,
|
| 1617 |
+
"grad_norm": 1.4711731672286987,
|
| 1618 |
+
"learning_rate": 8.776719134934199e-06,
|
| 1619 |
+
"loss": 0.0828,
|
| 1620 |
+
"step": 18000
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 23.778071334214,
|
| 1624 |
+
"eval_bleu": 70.5623767627479,
|
| 1625 |
+
"eval_char_accuracy": 62.505140648132915,
|
| 1626 |
+
"eval_loss": 0.0924154594540596,
|
| 1627 |
+
"eval_runtime": 342.2412,
|
| 1628 |
+
"eval_samples_per_second": 4.424,
|
| 1629 |
+
"eval_steps_per_second": 0.555,
|
| 1630 |
+
"step": 18000
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 23.91017173051519,
|
| 1634 |
+
"grad_norm": 1.4447731971740723,
|
| 1635 |
+
"learning_rate": 8.762905482515775e-06,
|
| 1636 |
+
"loss": 0.0814,
|
| 1637 |
+
"step": 18100
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 24.042272126816382,
|
| 1641 |
+
"grad_norm": 1.350907802581787,
|
| 1642 |
+
"learning_rate": 8.749025275017107e-06,
|
| 1643 |
+
"loss": 0.0806,
|
| 1644 |
+
"step": 18200
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 24.17437252311757,
|
| 1648 |
+
"grad_norm": 1.7207551002502441,
|
| 1649 |
+
"learning_rate": 8.735078757939532e-06,
|
| 1650 |
+
"loss": 0.08,
|
| 1651 |
+
"step": 18300
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 24.30647291941876,
|
| 1655 |
+
"grad_norm": 1.0851505994796753,
|
| 1656 |
+
"learning_rate": 8.721066177957213e-06,
|
| 1657 |
+
"loss": 0.0779,
|
| 1658 |
+
"step": 18400
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 24.438573315719946,
|
| 1662 |
+
"grad_norm": 1.2182328701019287,
|
| 1663 |
+
"learning_rate": 8.70698778291278e-06,
|
| 1664 |
+
"loss": 0.0814,
|
| 1665 |
+
"step": 18500
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 24.438573315719946,
|
| 1669 |
+
"eval_bleu": 70.99257164437572,
|
| 1670 |
+
"eval_char_accuracy": 62.22189093600922,
|
| 1671 |
+
"eval_loss": 0.09152651578187943,
|
| 1672 |
+
"eval_runtime": 323.7174,
|
| 1673 |
+
"eval_samples_per_second": 4.677,
|
| 1674 |
+
"eval_steps_per_second": 0.587,
|
| 1675 |
+
"step": 18500
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 24.570673712021136,
|
| 1679 |
+
"grad_norm": 2.0363285541534424,
|
| 1680 |
+
"learning_rate": 8.69284382181294e-06,
|
| 1681 |
+
"loss": 0.0821,
|
| 1682 |
+
"step": 18600
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 24.702774108322323,
|
| 1686 |
+
"grad_norm": 1.3864432573318481,
|
| 1687 |
+
"learning_rate": 8.67863454482408e-06,
|
| 1688 |
+
"loss": 0.0809,
|
| 1689 |
+
"step": 18700
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 24.834874504623514,
|
| 1693 |
+
"grad_norm": 2.032351493835449,
|
| 1694 |
+
"learning_rate": 8.664360203267838e-06,
|
| 1695 |
+
"loss": 0.0819,
|
| 1696 |
+
"step": 18800
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 24.966974900924704,
|
| 1700 |
+
"grad_norm": 1.2007182836532593,
|
| 1701 |
+
"learning_rate": 8.65002104961666e-06,
|
| 1702 |
+
"loss": 0.0819,
|
| 1703 |
+
"step": 18900
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 25.09907529722589,
|
| 1707 |
+
"grad_norm": 1.167693853378296,
|
| 1708 |
+
"learning_rate": 8.635617337489331e-06,
|
| 1709 |
+
"loss": 0.0778,
|
| 1710 |
+
"step": 19000
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 25.09907529722589,
|
| 1714 |
+
"eval_bleu": 70.99070971451881,
|
| 1715 |
+
"eval_char_accuracy": 63.15440450732028,
|
| 1716 |
+
"eval_loss": 0.09109245985746384,
|
| 1717 |
+
"eval_runtime": 331.1976,
|
| 1718 |
+
"eval_samples_per_second": 4.571,
|
| 1719 |
+
"eval_steps_per_second": 0.574,
|
| 1720 |
+
"step": 19000
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 25.23117569352708,
|
| 1724 |
+
"grad_norm": 1.9939295053482056,
|
| 1725 |
+
"learning_rate": 8.621149321646495e-06,
|
| 1726 |
+
"loss": 0.076,
|
| 1727 |
+
"step": 19100
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 25.36327608982827,
|
| 1731 |
+
"grad_norm": 1.148555874824524,
|
| 1732 |
+
"learning_rate": 8.60661725798614e-06,
|
| 1733 |
+
"loss": 0.078,
|
| 1734 |
+
"step": 19200
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 25.49537648612946,
|
| 1738 |
+
"grad_norm": 1.159621238708496,
|
| 1739 |
+
"learning_rate": 8.592167677001219e-06,
|
| 1740 |
+
"loss": 0.0823,
|
| 1741 |
+
"step": 19300
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 25.627476882430646,
|
| 1745 |
+
"grad_norm": 1.7136179208755493,
|
| 1746 |
+
"learning_rate": 8.57750892397125e-06,
|
| 1747 |
+
"loss": 0.0755,
|
| 1748 |
+
"step": 19400
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 25.759577278731836,
|
| 1752 |
+
"grad_norm": 1.105714201927185,
|
| 1753 |
+
"learning_rate": 8.5627868949981e-06,
|
| 1754 |
+
"loss": 0.0756,
|
| 1755 |
+
"step": 19500
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 25.759577278731836,
|
| 1759 |
+
"eval_bleu": 71.15813151741806,
|
| 1760 |
+
"eval_char_accuracy": 63.43405576575094,
|
| 1761 |
+
"eval_loss": 0.09030098468065262,
|
| 1762 |
+
"eval_runtime": 316.3492,
|
| 1763 |
+
"eval_samples_per_second": 4.786,
|
| 1764 |
+
"eval_steps_per_second": 0.601,
|
| 1765 |
+
"step": 19500
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 25.891677675033026,
|
| 1769 |
+
"grad_norm": 1.6091099977493286,
|
| 1770 |
+
"learning_rate": 8.548001850472529e-06,
|
| 1771 |
+
"loss": 0.0778,
|
| 1772 |
+
"step": 19600
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 26.023778071334213,
|
| 1776 |
+
"grad_norm": 1.1730881929397583,
|
| 1777 |
+
"learning_rate": 8.533154051899864e-06,
|
| 1778 |
+
"loss": 0.0787,
|
| 1779 |
+
"step": 19700
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 26.155878467635404,
|
| 1783 |
+
"grad_norm": 1.2703460454940796,
|
| 1784 |
+
"learning_rate": 8.518243761895369e-06,
|
| 1785 |
+
"loss": 0.0711,
|
| 1786 |
+
"step": 19800
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 26.28797886393659,
|
| 1790 |
+
"grad_norm": 1.3379662036895752,
|
| 1791 |
+
"learning_rate": 8.503271244179608e-06,
|
| 1792 |
+
"loss": 0.075,
|
| 1793 |
+
"step": 19900
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 26.42007926023778,
|
| 1797 |
+
"grad_norm": 1.370871901512146,
|
| 1798 |
+
"learning_rate": 8.488236763573772e-06,
|
| 1799 |
+
"loss": 0.0717,
|
| 1800 |
+
"step": 20000
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 26.42007926023778,
|
| 1804 |
+
"eval_bleu": 72.10017178272712,
|
| 1805 |
+
"eval_char_accuracy": 63.853018588583645,
|
| 1806 |
+
"eval_loss": 0.08871379494667053,
|
| 1807 |
+
"eval_runtime": 316.1038,
|
| 1808 |
+
"eval_samples_per_second": 4.79,
|
| 1809 |
+
"eval_steps_per_second": 0.601,
|
| 1810 |
+
"step": 20000
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 26.552179656538968,
|
| 1814 |
+
"grad_norm": 1.3396387100219727,
|
| 1815 |
+
"learning_rate": 8.473140585995004e-06,
|
| 1816 |
+
"loss": 0.0726,
|
| 1817 |
+
"step": 20100
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 26.68428005284016,
|
| 1821 |
+
"grad_norm": 1.1219794750213623,
|
| 1822 |
+
"learning_rate": 8.457982978451683e-06,
|
| 1823 |
+
"loss": 0.0754,
|
| 1824 |
+
"step": 20200
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 26.81638044914135,
|
| 1828 |
+
"grad_norm": 1.0815324783325195,
|
| 1829 |
+
"learning_rate": 8.442764209038717e-06,
|
| 1830 |
+
"loss": 0.0745,
|
| 1831 |
+
"step": 20300
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 26.948480845442536,
|
| 1835 |
+
"grad_norm": 1.4396206140518188,
|
| 1836 |
+
"learning_rate": 8.427484546932789e-06,
|
| 1837 |
+
"loss": 0.0749,
|
| 1838 |
+
"step": 20400
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 27.080581241743726,
|
| 1842 |
+
"grad_norm": 1.2644987106323242,
|
| 1843 |
+
"learning_rate": 8.4121442623876e-06,
|
| 1844 |
+
"loss": 0.0731,
|
| 1845 |
+
"step": 20500
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 27.080581241743726,
|
| 1849 |
+
"eval_bleu": 71.61514434619416,
|
| 1850 |
+
"eval_char_accuracy": 63.06290097055437,
|
| 1851 |
+
"eval_loss": 0.08903466165065765,
|
| 1852 |
+
"eval_runtime": 316.4287,
|
| 1853 |
+
"eval_samples_per_second": 4.785,
|
| 1854 |
+
"eval_steps_per_second": 0.6,
|
| 1855 |
+
"step": 20500
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 27.212681638044913,
|
| 1859 |
+
"grad_norm": 1.3456913232803345,
|
| 1860 |
+
"learning_rate": 8.396743626729093e-06,
|
| 1861 |
+
"loss": 0.0728,
|
| 1862 |
+
"step": 20600
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 27.344782034346103,
|
| 1866 |
+
"grad_norm": 1.1268248558044434,
|
| 1867 |
+
"learning_rate": 8.381282912350646e-06,
|
| 1868 |
+
"loss": 0.072,
|
| 1869 |
+
"step": 20700
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 27.476882430647294,
|
| 1873 |
+
"grad_norm": 0.964856743812561,
|
| 1874 |
+
"learning_rate": 8.365762392708259e-06,
|
| 1875 |
+
"loss": 0.0711,
|
| 1876 |
+
"step": 20800
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 27.60898282694848,
|
| 1880 |
+
"grad_norm": 1.2877197265625,
|
| 1881 |
+
"learning_rate": 8.350182342315719e-06,
|
| 1882 |
+
"loss": 0.0681,
|
| 1883 |
+
"step": 20900
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 27.74108322324967,
|
| 1887 |
+
"grad_norm": 3.0796854496002197,
|
| 1888 |
+
"learning_rate": 8.334543036739743e-06,
|
| 1889 |
+
"loss": 0.0681,
|
| 1890 |
+
"step": 21000
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 27.74108322324967,
|
| 1894 |
+
"eval_bleu": 72.44362762449254,
|
| 1895 |
+
"eval_char_accuracy": 64.56139990129955,
|
| 1896 |
+
"eval_loss": 0.08809462934732437,
|
| 1897 |
+
"eval_runtime": 319.538,
|
| 1898 |
+
"eval_samples_per_second": 4.738,
|
| 1899 |
+
"eval_steps_per_second": 0.595,
|
| 1900 |
+
"step": 21000
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 27.873183619550858,
|
| 1904 |
+
"grad_norm": 1.445993185043335,
|
| 1905 |
+
"learning_rate": 8.3188447525951e-06,
|
| 1906 |
+
"loss": 0.0701,
|
| 1907 |
+
"step": 21100
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 28.005284015852048,
|
| 1911 |
+
"grad_norm": 0.9414767622947693,
|
| 1912 |
+
"learning_rate": 8.303087767539723e-06,
|
| 1913 |
+
"loss": 0.0698,
|
| 1914 |
+
"step": 21200
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 28.137384412153235,
|
| 1918 |
+
"grad_norm": 1.0644490718841553,
|
| 1919 |
+
"learning_rate": 8.28758923914531e-06,
|
| 1920 |
+
"loss": 0.0674,
|
| 1921 |
+
"step": 21300
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 28.269484808454425,
|
| 1925 |
+
"grad_norm": 1.2708711624145508,
|
| 1926 |
+
"learning_rate": 8.27171684949204e-06,
|
| 1927 |
+
"loss": 0.0689,
|
| 1928 |
+
"step": 21400
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 28.401585204755616,
|
| 1932 |
+
"grad_norm": 1.2177033424377441,
|
| 1933 |
+
"learning_rate": 8.25578659248641e-06,
|
| 1934 |
+
"loss": 0.0677,
|
| 1935 |
+
"step": 21500
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 28.401585204755616,
|
| 1939 |
+
"eval_bleu": 72.7907604984095,
|
| 1940 |
+
"eval_char_accuracy": 64.78141964138838,
|
| 1941 |
+
"eval_loss": 0.0869230180978775,
|
| 1942 |
+
"eval_runtime": 315.3901,
|
| 1943 |
+
"eval_samples_per_second": 4.8,
|
| 1944 |
+
"eval_steps_per_second": 0.602,
|
| 1945 |
+
"step": 21500
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 28.533685601056803,
|
| 1949 |
+
"grad_norm": 1.1179392337799072,
|
| 1950 |
+
"learning_rate": 8.239798749889293e-06,
|
| 1951 |
+
"loss": 0.0673,
|
| 1952 |
+
"step": 21600
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 28.665785997357993,
|
| 1956 |
+
"grad_norm": 1.2472251653671265,
|
| 1957 |
+
"learning_rate": 8.223753604480086e-06,
|
| 1958 |
+
"loss": 0.0682,
|
| 1959 |
+
"step": 21700
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 28.79788639365918,
|
| 1963 |
+
"grad_norm": 0.8336161375045776,
|
| 1964 |
+
"learning_rate": 8.207651440051714e-06,
|
| 1965 |
+
"loss": 0.0689,
|
| 1966 |
+
"step": 21800
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 28.92998678996037,
|
| 1970 |
+
"grad_norm": 1.2652006149291992,
|
| 1971 |
+
"learning_rate": 8.1914925414056e-06,
|
| 1972 |
+
"loss": 0.0688,
|
| 1973 |
+
"step": 21900
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 29.062087186261557,
|
| 1977 |
+
"grad_norm": 1.1424204111099243,
|
| 1978 |
+
"learning_rate": 8.175277194346636e-06,
|
| 1979 |
+
"loss": 0.0677,
|
| 1980 |
+
"step": 22000
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 29.062087186261557,
|
| 1984 |
+
"eval_bleu": 72.84923297683062,
|
| 1985 |
+
"eval_char_accuracy": 64.63285491034709,
|
| 1986 |
+
"eval_loss": 0.08696427941322327,
|
| 1987 |
+
"eval_runtime": 316.7435,
|
| 1988 |
+
"eval_samples_per_second": 4.78,
|
| 1989 |
+
"eval_steps_per_second": 0.6,
|
| 1990 |
+
"step": 22000
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 29.194187582562748,
|
| 1994 |
+
"grad_norm": 1.617885708808899,
|
| 1995 |
+
"learning_rate": 8.159005685678126e-06,
|
| 1996 |
+
"loss": 0.0638,
|
| 1997 |
+
"step": 22100
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 29.326287978863938,
|
| 2001 |
+
"grad_norm": 1.2440978288650513,
|
| 2002 |
+
"learning_rate": 8.142678303196715e-06,
|
| 2003 |
+
"loss": 0.0606,
|
| 2004 |
+
"step": 22200
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 29.458388375165125,
|
| 2008 |
+
"grad_norm": 1.1825751066207886,
|
| 2009 |
+
"learning_rate": 8.12629533568729e-06,
|
| 2010 |
+
"loss": 0.0661,
|
| 2011 |
+
"step": 22300
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 29.590488771466315,
|
| 2015 |
+
"grad_norm": 1.5328364372253418,
|
| 2016 |
+
"learning_rate": 8.109857072917887e-06,
|
| 2017 |
+
"loss": 0.0647,
|
| 2018 |
+
"step": 22400
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 29.722589167767502,
|
| 2022 |
+
"grad_norm": 1.1308438777923584,
|
| 2023 |
+
"learning_rate": 8.093363805634556e-06,
|
| 2024 |
+
"loss": 0.0666,
|
| 2025 |
+
"step": 22500
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 29.722589167767502,
|
| 2029 |
+
"eval_bleu": 73.15661772633777,
|
| 2030 |
+
"eval_char_accuracy": 63.81086527389373,
|
| 2031 |
+
"eval_loss": 0.08648520708084106,
|
| 2032 |
+
"eval_runtime": 315.6708,
|
| 2033 |
+
"eval_samples_per_second": 4.796,
|
| 2034 |
+
"eval_steps_per_second": 0.602,
|
| 2035 |
+
"step": 22500
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 29.854689564068693,
|
| 2039 |
+
"grad_norm": 1.0442135334014893,
|
| 2040 |
+
"learning_rate": 8.076815825556213e-06,
|
| 2041 |
+
"loss": 0.0648,
|
| 2042 |
+
"step": 22600
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 29.98678996036988,
|
| 2046 |
+
"grad_norm": 1.363897681236267,
|
| 2047 |
+
"learning_rate": 8.060213425369492e-06,
|
| 2048 |
+
"loss": 0.0654,
|
| 2049 |
+
"step": 22700
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 30.11889035667107,
|
| 2053 |
+
"grad_norm": 0.7751464247703552,
|
| 2054 |
+
"learning_rate": 8.043556898723568e-06,
|
| 2055 |
+
"loss": 0.0628,
|
| 2056 |
+
"step": 22800
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 30.25099075297226,
|
| 2060 |
+
"grad_norm": 0.8685150742530823,
|
| 2061 |
+
"learning_rate": 8.026846540224956e-06,
|
| 2062 |
+
"loss": 0.0584,
|
| 2063 |
+
"step": 22900
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 30.383091149273447,
|
| 2067 |
+
"grad_norm": 1.1484973430633545,
|
| 2068 |
+
"learning_rate": 8.0100826454323e-06,
|
| 2069 |
+
"loss": 0.0604,
|
| 2070 |
+
"step": 23000
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 30.383091149273447,
|
| 2074 |
+
"eval_bleu": 73.25107285034876,
|
| 2075 |
+
"eval_char_accuracy": 65.01943164994243,
|
| 2076 |
+
"eval_loss": 0.08666232973337173,
|
| 2077 |
+
"eval_runtime": 323.7755,
|
| 2078 |
+
"eval_samples_per_second": 4.676,
|
| 2079 |
+
"eval_steps_per_second": 0.587,
|
| 2080 |
+
"step": 23000
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 30.515191545574638,
|
| 2084 |
+
"grad_norm": 0.986289918422699,
|
| 2085 |
+
"learning_rate": 7.993265510851148e-06,
|
| 2086 |
+
"loss": 0.0688,
|
| 2087 |
+
"step": 23100
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 30.647291941875825,
|
| 2091 |
+
"grad_norm": 1.0403423309326172,
|
| 2092 |
+
"learning_rate": 7.97639543392872e-06,
|
| 2093 |
+
"loss": 0.0638,
|
| 2094 |
+
"step": 23200
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 30.779392338177015,
|
| 2098 |
+
"grad_norm": 1.517040729522705,
|
| 2099 |
+
"learning_rate": 7.959472713048617e-06,
|
| 2100 |
+
"loss": 0.0653,
|
| 2101 |
+
"step": 23300
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 30.911492734478202,
|
| 2105 |
+
"grad_norm": 1.1347965002059937,
|
| 2106 |
+
"learning_rate": 7.942497647525576e-06,
|
| 2107 |
+
"loss": 0.0642,
|
| 2108 |
+
"step": 23400
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 31.043593130779392,
|
| 2112 |
+
"grad_norm": 1.0789778232574463,
|
| 2113 |
+
"learning_rate": 7.925470537600155e-06,
|
| 2114 |
+
"loss": 0.0614,
|
| 2115 |
+
"step": 23500
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 31.043593130779392,
|
| 2119 |
+
"eval_bleu": 73.23277401857816,
|
| 2120 |
+
"eval_char_accuracy": 65.0646693535121,
|
| 2121 |
+
"eval_loss": 0.08618722856044769,
|
| 2122 |
+
"eval_runtime": 313.4951,
|
| 2123 |
+
"eval_samples_per_second": 4.829,
|
| 2124 |
+
"eval_steps_per_second": 0.606,
|
| 2125 |
+
"step": 23500
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 31.175693527080583,
|
| 2129 |
+
"grad_norm": 1.4853187799453735,
|
| 2130 |
+
"learning_rate": 7.908391684433432e-06,
|
| 2131 |
+
"loss": 0.0585,
|
| 2132 |
+
"step": 23600
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 31.30779392338177,
|
| 2136 |
+
"grad_norm": 0.9656835794448853,
|
| 2137 |
+
"learning_rate": 7.891261390101675e-06,
|
| 2138 |
+
"loss": 0.0578,
|
| 2139 |
+
"step": 23700
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 31.43989431968296,
|
| 2143 |
+
"grad_norm": 1.1521549224853516,
|
| 2144 |
+
"learning_rate": 7.874079957590997e-06,
|
| 2145 |
+
"loss": 0.0622,
|
| 2146 |
+
"step": 23800
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 31.571994715984147,
|
| 2150 |
+
"grad_norm": 1.0636780261993408,
|
| 2151 |
+
"learning_rate": 7.856847690792002e-06,
|
| 2152 |
+
"loss": 0.0604,
|
| 2153 |
+
"step": 23900
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 31.704095112285337,
|
| 2157 |
+
"grad_norm": 1.0833789110183716,
|
| 2158 |
+
"learning_rate": 7.839564894494409e-06,
|
| 2159 |
+
"loss": 0.0633,
|
| 2160 |
+
"step": 24000
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 31.704095112285337,
|
| 2164 |
+
"eval_bleu": 73.62536575895216,
|
| 2165 |
+
"eval_char_accuracy": 65.15565882546471,
|
| 2166 |
+
"eval_loss": 0.08546082675457001,
|
| 2167 |
+
"eval_runtime": 318.0064,
|
| 2168 |
+
"eval_samples_per_second": 4.761,
|
| 2169 |
+
"eval_steps_per_second": 0.597,
|
| 2170 |
+
"step": 24000
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 31.836195508586528,
|
| 2174 |
+
"grad_norm": 1.1620845794677734,
|
| 2175 |
+
"learning_rate": 7.822231874381658e-06,
|
| 2176 |
+
"loss": 0.0604,
|
| 2177 |
+
"step": 24100
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 31.968295904887714,
|
| 2181 |
+
"grad_norm": 1.315012812614441,
|
| 2182 |
+
"learning_rate": 7.804848937025507e-06,
|
| 2183 |
+
"loss": 0.0593,
|
| 2184 |
+
"step": 24200
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 32.1003963011889,
|
| 2188 |
+
"grad_norm": 0.8739562034606934,
|
| 2189 |
+
"learning_rate": 7.787416389880605e-06,
|
| 2190 |
+
"loss": 0.0608,
|
| 2191 |
+
"step": 24300
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 32.23249669749009,
|
| 2195 |
+
"grad_norm": 0.9168538451194763,
|
| 2196 |
+
"learning_rate": 7.769934541279059e-06,
|
| 2197 |
+
"loss": 0.0577,
|
| 2198 |
+
"step": 24400
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 32.36459709379128,
|
| 2202 |
+
"grad_norm": 0.9820032715797424,
|
| 2203 |
+
"learning_rate": 7.752403700424978e-06,
|
| 2204 |
+
"loss": 0.0569,
|
| 2205 |
+
"step": 24500
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 32.36459709379128,
|
| 2209 |
+
"eval_bleu": 73.72149088930817,
|
| 2210 |
+
"eval_char_accuracy": 65.55457312057904,
|
| 2211 |
+
"eval_loss": 0.08627723157405853,
|
| 2212 |
+
"eval_runtime": 314.5801,
|
| 2213 |
+
"eval_samples_per_second": 4.813,
|
| 2214 |
+
"eval_steps_per_second": 0.604,
|
| 2215 |
+
"step": 24500
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 32.49669749009247,
|
| 2219 |
+
"grad_norm": 1.0042686462402344,
|
| 2220 |
+
"learning_rate": 7.734824177389006e-06,
|
| 2221 |
+
"loss": 0.0582,
|
| 2222 |
+
"step": 24600
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 32.628797886393656,
|
| 2226 |
+
"grad_norm": 1.251654863357544,
|
| 2227 |
+
"learning_rate": 7.71719628310283e-06,
|
| 2228 |
+
"loss": 0.0589,
|
| 2229 |
+
"step": 24700
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 32.760898282694846,
|
| 2233 |
+
"grad_norm": 1.3150684833526611,
|
| 2234 |
+
"learning_rate": 7.699520329353694e-06,
|
| 2235 |
+
"loss": 0.0585,
|
| 2236 |
+
"step": 24800
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 32.89299867899604,
|
| 2240 |
+
"grad_norm": 1.318556547164917,
|
| 2241 |
+
"learning_rate": 7.681796628778876e-06,
|
| 2242 |
+
"loss": 0.0588,
|
| 2243 |
+
"step": 24900
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 33.02509907529723,
|
| 2247 |
+
"grad_norm": 1.2693874835968018,
|
| 2248 |
+
"learning_rate": 7.664025494860155e-06,
|
| 2249 |
+
"loss": 0.0605,
|
| 2250 |
+
"step": 25000
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 33.02509907529723,
|
| 2254 |
+
"eval_bleu": 73.95631775899457,
|
| 2255 |
+
"eval_char_accuracy": 65.17056670505016,
|
| 2256 |
+
"eval_loss": 0.08447689563035965,
|
| 2257 |
+
"eval_runtime": 317.601,
|
| 2258 |
+
"eval_samples_per_second": 4.767,
|
| 2259 |
+
"eval_steps_per_second": 0.598,
|
| 2260 |
+
"step": 25000
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 33.15719947159842,
|
| 2264 |
+
"grad_norm": 0.7866926193237305,
|
| 2265 |
+
"learning_rate": 7.646207241918272e-06,
|
| 2266 |
+
"loss": 0.055,
|
| 2267 |
+
"step": 25100
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 33.2892998678996,
|
| 2271 |
+
"grad_norm": 1.0340533256530762,
|
| 2272 |
+
"learning_rate": 7.628342185107373e-06,
|
| 2273 |
+
"loss": 0.0563,
|
| 2274 |
+
"step": 25200
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 33.42140026420079,
|
| 2278 |
+
"grad_norm": 1.6704190969467163,
|
| 2279 |
+
"learning_rate": 7.610430640409427e-06,
|
| 2280 |
+
"loss": 0.0568,
|
| 2281 |
+
"step": 25300
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 33.55350066050198,
|
| 2285 |
+
"grad_norm": 1.4271676540374756,
|
| 2286 |
+
"learning_rate": 7.592472924628642e-06,
|
| 2287 |
+
"loss": 0.056,
|
| 2288 |
+
"step": 25400
|
| 2289 |
+
},
|
| 2290 |
+
{
|
| 2291 |
+
"epoch": 33.68560105680317,
|
| 2292 |
+
"grad_norm": 1.3886315822601318,
|
| 2293 |
+
"learning_rate": 7.574469355385865e-06,
|
| 2294 |
+
"loss": 0.0552,
|
| 2295 |
+
"step": 25500
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 33.68560105680317,
|
| 2299 |
+
"eval_bleu": 73.67062952266826,
|
| 2300 |
+
"eval_char_accuracy": 65.04410676098043,
|
| 2301 |
+
"eval_loss": 0.08498267084360123,
|
| 2302 |
+
"eval_runtime": 314.7186,
|
| 2303 |
+
"eval_samples_per_second": 4.811,
|
| 2304 |
+
"eval_steps_per_second": 0.604,
|
| 2305 |
+
"step": 25500
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 33.81770145310436,
|
| 2309 |
+
"grad_norm": 1.1037873029708862,
|
| 2310 |
+
"learning_rate": 7.556420251112956e-06,
|
| 2311 |
+
"loss": 0.0551,
|
| 2312 |
+
"step": 25600
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 33.949801849405546,
|
| 2316 |
+
"grad_norm": 2.125624418258667,
|
| 2317 |
+
"learning_rate": 7.538325931047159e-06,
|
| 2318 |
+
"loss": 0.0591,
|
| 2319 |
+
"step": 25700
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 34.081902245706736,
|
| 2323 |
+
"grad_norm": 1.674501895904541,
|
| 2324 |
+
"learning_rate": 7.52018671522546e-06,
|
| 2325 |
+
"loss": 0.0561,
|
| 2326 |
+
"step": 25800
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 34.21400264200793,
|
| 2330 |
+
"grad_norm": 1.386206030845642,
|
| 2331 |
+
"learning_rate": 7.502002924478924e-06,
|
| 2332 |
+
"loss": 0.0509,
|
| 2333 |
+
"step": 25900
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 34.34610303830912,
|
| 2337 |
+
"grad_norm": 1.0778214931488037,
|
| 2338 |
+
"learning_rate": 7.48377488042701e-06,
|
| 2339 |
+
"loss": 0.0544,
|
| 2340 |
+
"step": 26000
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 34.34610303830912,
|
| 2344 |
+
"eval_bleu": 74.62046528623556,
|
| 2345 |
+
"eval_char_accuracy": 66.02442835992763,
|
| 2346 |
+
"eval_loss": 0.08464069664478302,
|
| 2347 |
+
"eval_runtime": 316.1374,
|
| 2348 |
+
"eval_samples_per_second": 4.789,
|
| 2349 |
+
"eval_steps_per_second": 0.601,
|
| 2350 |
+
"step": 26000
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 34.4782034346103,
|
| 2354 |
+
"grad_norm": 0.8076276779174805,
|
| 2355 |
+
"learning_rate": 7.465502905471907e-06,
|
| 2356 |
+
"loss": 0.055,
|
| 2357 |
+
"step": 26100
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 34.61030383091149,
|
| 2361 |
+
"grad_norm": 1.1508395671844482,
|
| 2362 |
+
"learning_rate": 7.447187322792806e-06,
|
| 2363 |
+
"loss": 0.057,
|
| 2364 |
+
"step": 26200
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 34.74240422721268,
|
| 2368 |
+
"grad_norm": 1.2695698738098145,
|
| 2369 |
+
"learning_rate": 7.4288284563401945e-06,
|
| 2370 |
+
"loss": 0.055,
|
| 2371 |
+
"step": 26300
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 34.87450462351387,
|
| 2375 |
+
"grad_norm": 1.166051983833313,
|
| 2376 |
+
"learning_rate": 7.410426630830131e-06,
|
| 2377 |
+
"loss": 0.0552,
|
| 2378 |
+
"step": 26400
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 35.00660501981506,
|
| 2382 |
+
"grad_norm": 0.9517413973808289,
|
| 2383 |
+
"learning_rate": 7.391982171738496e-06,
|
| 2384 |
+
"loss": 0.0555,
|
| 2385 |
+
"step": 26500
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 35.00660501981506,
|
| 2389 |
+
"eval_bleu": 74.18793581426324,
|
| 2390 |
+
"eval_char_accuracy": 65.63373910182597,
|
| 2391 |
+
"eval_loss": 0.08454510569572449,
|
| 2392 |
+
"eval_runtime": 313.1895,
|
| 2393 |
+
"eval_samples_per_second": 4.834,
|
| 2394 |
+
"eval_steps_per_second": 0.607,
|
| 2395 |
+
"step": 26500
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 35.138705416116245,
|
| 2399 |
+
"grad_norm": 1.1265789270401,
|
| 2400 |
+
"learning_rate": 7.373495405295236e-06,
|
| 2401 |
+
"loss": 0.0529,
|
| 2402 |
+
"step": 26600
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 35.270805812417436,
|
| 2406 |
+
"grad_norm": 1.0067466497421265,
|
| 2407 |
+
"learning_rate": 7.354966658478594e-06,
|
| 2408 |
+
"loss": 0.0502,
|
| 2409 |
+
"step": 26700
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 35.402906208718626,
|
| 2413 |
+
"grad_norm": 0.9871610999107361,
|
| 2414 |
+
"learning_rate": 7.336396259009325e-06,
|
| 2415 |
+
"loss": 0.0508,
|
| 2416 |
+
"step": 26800
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 35.53500660501982,
|
| 2420 |
+
"grad_norm": 1.4898390769958496,
|
| 2421 |
+
"learning_rate": 7.317784535344905e-06,
|
| 2422 |
+
"loss": 0.0544,
|
| 2423 |
+
"step": 26900
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 35.66710700132101,
|
| 2427 |
+
"grad_norm": 1.1168763637542725,
|
| 2428 |
+
"learning_rate": 7.2991318166737126e-06,
|
| 2429 |
+
"loss": 0.0535,
|
| 2430 |
+
"step": 27000
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 35.66710700132101,
|
| 2434 |
+
"eval_bleu": 74.4767569683976,
|
| 2435 |
+
"eval_char_accuracy": 65.44353512090805,
|
| 2436 |
+
"eval_loss": 0.08464961498975754,
|
| 2437 |
+
"eval_runtime": 313.4254,
|
| 2438 |
+
"eval_samples_per_second": 4.83,
|
| 2439 |
+
"eval_steps_per_second": 0.606,
|
| 2440 |
+
"step": 27000
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 35.79920739762219,
|
| 2444 |
+
"grad_norm": 1.197704553604126,
|
| 2445 |
+
"learning_rate": 7.280625566954032e-06,
|
| 2446 |
+
"loss": 0.0547,
|
| 2447 |
+
"step": 27100
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 35.93130779392338,
|
| 2451 |
+
"grad_norm": 1.6144860982894897,
|
| 2452 |
+
"learning_rate": 7.261892250434568e-06,
|
| 2453 |
+
"loss": 0.0516,
|
| 2454 |
+
"step": 27200
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 36.06340819022457,
|
| 2458 |
+
"grad_norm": 1.1599304676055908,
|
| 2459 |
+
"learning_rate": 7.243118927483657e-06,
|
| 2460 |
+
"loss": 0.0502,
|
| 2461 |
+
"step": 27300
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 36.19550858652576,
|
| 2465 |
+
"grad_norm": 1.043449878692627,
|
| 2466 |
+
"learning_rate": 7.22430593014791e-06,
|
| 2467 |
+
"loss": 0.0472,
|
| 2468 |
+
"step": 27400
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 36.32760898282695,
|
| 2472 |
+
"grad_norm": 1.1489434242248535,
|
| 2473 |
+
"learning_rate": 7.205453591175666e-06,
|
| 2474 |
+
"loss": 0.0558,
|
| 2475 |
+
"step": 27500
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 36.32760898282695,
|
| 2479 |
+
"eval_bleu": 74.20946250489693,
|
| 2480 |
+
"eval_char_accuracy": 65.89436996216483,
|
| 2481 |
+
"eval_loss": 0.08468983322381973,
|
| 2482 |
+
"eval_runtime": 314.7938,
|
| 2483 |
+
"eval_samples_per_second": 4.809,
|
| 2484 |
+
"eval_steps_per_second": 0.604,
|
| 2485 |
+
"step": 27500
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 36.32760898282695,
|
| 2489 |
+
"step": 27500,
|
| 2490 |
+
"total_flos": 7035660725649408.0,
|
| 2491 |
+
"train_loss": 0.282735899699818,
|
| 2492 |
+
"train_runtime": 21727.9767,
|
| 2493 |
+
"train_samples_per_second": 55.735,
|
| 2494 |
+
"train_steps_per_second": 3.484
|
| 2495 |
+
}
|
| 2496 |
+
],
|
| 2497 |
+
"logging_steps": 100,
|
| 2498 |
+
"max_steps": 75700,
|
| 2499 |
+
"num_input_tokens_seen": 0,
|
| 2500 |
+
"num_train_epochs": 100,
|
| 2501 |
+
"save_steps": 500,
|
| 2502 |
+
"stateful_callbacks": {
|
| 2503 |
+
"EarlyStoppingCallback": {
|
| 2504 |
+
"args": {
|
| 2505 |
+
"early_stopping_patience": 3,
|
| 2506 |
+
"early_stopping_threshold": 0.0
|
| 2507 |
+
},
|
| 2508 |
+
"attributes": {
|
| 2509 |
+
"early_stopping_patience_counter": 3
|
| 2510 |
+
}
|
| 2511 |
+
},
|
| 2512 |
+
"TrainerControl": {
|
| 2513 |
+
"args": {
|
| 2514 |
+
"should_epoch_stop": false,
|
| 2515 |
+
"should_evaluate": false,
|
| 2516 |
+
"should_log": false,
|
| 2517 |
+
"should_save": true,
|
| 2518 |
+
"should_training_stop": true
|
| 2519 |
+
},
|
| 2520 |
+
"attributes": {}
|
| 2521 |
+
}
|
| 2522 |
+
},
|
| 2523 |
+
"total_flos": 7035660725649408.0,
|
| 2524 |
+
"train_batch_size": 8,
|
| 2525 |
+
"trial_name": null,
|
| 2526 |
+
"trial_params": null
|
| 2527 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5952b6a5e7d8317da73453941ab91679dc14918147c96729a68d3d232bb1ebd7
|
| 3 |
+
size 5841
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|