Upload folder using huggingface_hub
Browse files- README.md +71 -0
- config.json +27 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +58 -0
- trainer_state.json +2974 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: en
|
| 3 |
+
tags:
|
| 4 |
+
- ner
|
| 5 |
+
- bert
|
| 6 |
+
- mountain-ner
|
| 7 |
+
- named-entity-recognition
|
| 8 |
+
license: mit
|
| 9 |
+
datasets:
|
| 10 |
+
- NERetrieve
|
| 11 |
+
- Few-NERD
|
| 12 |
+
- mountain-ner-dataset
|
| 13 |
+
metrics:
|
| 14 |
+
- accuracy
|
| 15 |
+
- f1
|
| 16 |
+
- precision
|
| 17 |
+
- recall
|
| 18 |
+
pipeline_tag: token-classification
|
| 19 |
+
model-index:
|
| 20 |
+
- name: mountain-ner-bert-base
|
| 21 |
+
results:
|
| 22 |
+
- task:
|
| 23 |
+
type: token-classification
|
| 24 |
+
name: Named Entity Recognition
|
| 25 |
+
dataset:
|
| 26 |
+
name: mountain-ner-dataset
|
| 27 |
+
type: Gepe55o/mountain-ner-dataset
|
| 28 |
+
metrics:
|
| 29 |
+
- type: accuracy
|
| 30 |
+
value: 0.9919
|
| 31 |
+
- type: f1
|
| 32 |
+
value: 0.9048
|
| 33 |
+
- type: precision
|
| 34 |
+
value: 0.8899
|
| 35 |
+
- type: recall
|
| 36 |
+
value: 0.9202
|
| 37 |
+
---
|
| 38 |
+
## Model Description
|
| 39 |
+
mountain-ner-bert-base is a fine-tuned model based on the BERT base architecture for mountain names Entity Recognition tasks. The model is trained on the merging of two datasets: [NERetrieve](https://arxiv.org/pdf/2310.14282), [Few-NERD](https://arxiv.org/pdf/2105.07464v6), [Mountain-ner-dataset](https://huggingface.co/datasets/Gepe55o/mountain-ner-dataset). The model is trained to recognize two types of entities: `LABEL_0` (other), `LABEL_1` (mountain names).
|
| 40 |
+
|
| 41 |
+
- Model Architecture: BERT base
|
| 42 |
+
- Task: mountain names entity recognition
|
| 43 |
+
- Training Data: [mountain-ner-dataset](https://huggingface.co/datasets/Gepe55o/mountain-ner-dataset)
|
| 44 |
+
|
| 45 |
+
## Performance
|
| 46 |
+
Metrics:
|
| 47 |
+
| Epoch | Training Loss | Validation Loss | Accuracy | Precision | Recall | F1 |
|
| 48 |
+
|-------|---------------|----------------|----------|-----------|----------|----------|
|
| 49 |
+
| 1 | 0.027400 | 0.030793 | 0.988144 | 0.815692 | 0.924621 | 0.866748 |
|
| 50 |
+
| 2 | 0.020600 | 0.024568 | 0.991119 | 0.872988 | 0.921036 | 0.896369 |
|
| 51 |
+
| 3 | 0.012900 | 0.024072 | 0.991923 | 0.889878 | 0.920171 | 0.904771 |
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
Best model performance achieved at epoch 3 with:
|
| 55 |
+
- F1 Score: 0.9048
|
| 56 |
+
- Accuracy: 0.9919
|
| 57 |
+
- Precision: 0.8899
|
| 58 |
+
- Recall: 0.9202
|
| 59 |
+
|
| 60 |
+
## How to use
|
| 61 |
+
```python
|
| 62 |
+
from transformers import AutoModel, AutoTokenizer, pipeline
|
| 63 |
+
|
| 64 |
+
model = AutoModel.from_pretrained("Gepe55o/mountain-ner-bert-base")
|
| 65 |
+
tokenizer = AutoTokenizer.from_pretrained("Gepe55o/mountain-ner-bert-base")
|
| 66 |
+
|
| 67 |
+
text = "Mount Everest is the highest mountain in the world."
|
| 68 |
+
|
| 69 |
+
nlp = pipeline("ner", model=model, tokenizer=tokenizer)
|
| 70 |
+
result = nlp(text)
|
| 71 |
+
```
|
config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "dslim/bert-base-NER",
|
| 3 |
+
"_num_labels": 9,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForTokenClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 512,
|
| 16 |
+
"model_type": "bert",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 12,
|
| 19 |
+
"output_past": true,
|
| 20 |
+
"pad_token_id": 0,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.45.1",
|
| 24 |
+
"type_vocab_size": 2,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 28996
|
| 27 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4eb10816d72fe799243c2c124e62f2b2c52f942319eef79337ae99966459f76d
|
| 3 |
+
size 430908208
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:416403a07c30b16340f79f20432a8fb2a7044940a2a35f680832022c0adc4cf7
|
| 3 |
+
size 861936186
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b50ca2417b448558371bd73437bf35821e632924107103e7638ee561426d3241
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b5df567e00e789d95255f08e3b6085df32039b0e0c3e6ef0aa09a5a469b92194
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"max_len": 512,
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"never_split": null,
|
| 52 |
+
"pad_token": "[PAD]",
|
| 53 |
+
"sep_token": "[SEP]",
|
| 54 |
+
"strip_accents": null,
|
| 55 |
+
"tokenize_chinese_chars": true,
|
| 56 |
+
"tokenizer_class": "BertTokenizer",
|
| 57 |
+
"unk_token": "[UNK]"
|
| 58 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2974 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4155,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.007220216606498195,
|
| 13 |
+
"grad_norm": 15.138864517211914,
|
| 14 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 15 |
+
"loss": 0.9322,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.01444043321299639,
|
| 20 |
+
"grad_norm": 11.833885192871094,
|
| 21 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 22 |
+
"loss": 0.7888,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.021660649819494584,
|
| 27 |
+
"grad_norm": 7.410687446594238,
|
| 28 |
+
"learning_rate": 3e-06,
|
| 29 |
+
"loss": 0.5394,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02888086642599278,
|
| 34 |
+
"grad_norm": 3.045423746109009,
|
| 35 |
+
"learning_rate": 4.000000000000001e-06,
|
| 36 |
+
"loss": 0.3,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.036101083032490974,
|
| 41 |
+
"grad_norm": 0.5977299809455872,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 0.158,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.04332129963898917,
|
| 48 |
+
"grad_norm": 0.23663052916526794,
|
| 49 |
+
"learning_rate": 6e-06,
|
| 50 |
+
"loss": 0.1049,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05054151624548736,
|
| 55 |
+
"grad_norm": 0.24775876104831696,
|
| 56 |
+
"learning_rate": 7.000000000000001e-06,
|
| 57 |
+
"loss": 0.0929,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05776173285198556,
|
| 62 |
+
"grad_norm": 0.250676691532135,
|
| 63 |
+
"learning_rate": 8.000000000000001e-06,
|
| 64 |
+
"loss": 0.0818,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06498194945848375,
|
| 69 |
+
"grad_norm": 0.2527087926864624,
|
| 70 |
+
"learning_rate": 9e-06,
|
| 71 |
+
"loss": 0.0812,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.07220216606498195,
|
| 76 |
+
"grad_norm": 0.24259184300899506,
|
| 77 |
+
"learning_rate": 1e-05,
|
| 78 |
+
"loss": 0.0726,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07942238267148015,
|
| 83 |
+
"grad_norm": 0.19805486500263214,
|
| 84 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 85 |
+
"loss": 0.0657,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.08664259927797834,
|
| 90 |
+
"grad_norm": 0.21339212357997894,
|
| 91 |
+
"learning_rate": 1.2e-05,
|
| 92 |
+
"loss": 0.0649,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09386281588447654,
|
| 97 |
+
"grad_norm": 0.27472904324531555,
|
| 98 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 99 |
+
"loss": 0.0608,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.10108303249097472,
|
| 104 |
+
"grad_norm": 0.3953518867492676,
|
| 105 |
+
"learning_rate": 1.4000000000000001e-05,
|
| 106 |
+
"loss": 0.055,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.10830324909747292,
|
| 111 |
+
"grad_norm": 0.19267311692237854,
|
| 112 |
+
"learning_rate": 1.5e-05,
|
| 113 |
+
"loss": 0.0546,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.11552346570397112,
|
| 118 |
+
"grad_norm": 0.16826550662517548,
|
| 119 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 120 |
+
"loss": 0.0519,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.12274368231046931,
|
| 125 |
+
"grad_norm": 0.2655981183052063,
|
| 126 |
+
"learning_rate": 1.7000000000000003e-05,
|
| 127 |
+
"loss": 0.0521,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1299638989169675,
|
| 132 |
+
"grad_norm": 0.25509196519851685,
|
| 133 |
+
"learning_rate": 1.8e-05,
|
| 134 |
+
"loss": 0.0535,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1371841155234657,
|
| 139 |
+
"grad_norm": 0.38085201382637024,
|
| 140 |
+
"learning_rate": 1.9e-05,
|
| 141 |
+
"loss": 0.0502,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.1444043321299639,
|
| 146 |
+
"grad_norm": 0.2515423893928528,
|
| 147 |
+
"learning_rate": 2e-05,
|
| 148 |
+
"loss": 0.0502,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15162454873646208,
|
| 153 |
+
"grad_norm": 0.2517891824245453,
|
| 154 |
+
"learning_rate": 2.1e-05,
|
| 155 |
+
"loss": 0.0514,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1588447653429603,
|
| 160 |
+
"grad_norm": 0.21332941949367523,
|
| 161 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 162 |
+
"loss": 0.0463,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.16606498194945848,
|
| 167 |
+
"grad_norm": 0.2846091687679291,
|
| 168 |
+
"learning_rate": 2.3000000000000003e-05,
|
| 169 |
+
"loss": 0.0526,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.17328519855595667,
|
| 174 |
+
"grad_norm": 0.2511647045612335,
|
| 175 |
+
"learning_rate": 2.4e-05,
|
| 176 |
+
"loss": 0.0487,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18050541516245489,
|
| 181 |
+
"grad_norm": 0.20305795967578888,
|
| 182 |
+
"learning_rate": 2.5e-05,
|
| 183 |
+
"loss": 0.0437,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.18772563176895307,
|
| 188 |
+
"grad_norm": 0.27594617009162903,
|
| 189 |
+
"learning_rate": 2.6000000000000002e-05,
|
| 190 |
+
"loss": 0.0455,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.19494584837545126,
|
| 195 |
+
"grad_norm": 0.4420526921749115,
|
| 196 |
+
"learning_rate": 2.7000000000000002e-05,
|
| 197 |
+
"loss": 0.0507,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.20216606498194944,
|
| 202 |
+
"grad_norm": 0.3529132604598999,
|
| 203 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 204 |
+
"loss": 0.0483,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.20938628158844766,
|
| 209 |
+
"grad_norm": 0.18353258073329926,
|
| 210 |
+
"learning_rate": 2.9e-05,
|
| 211 |
+
"loss": 0.0435,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.21660649819494585,
|
| 216 |
+
"grad_norm": 0.2105967402458191,
|
| 217 |
+
"learning_rate": 3e-05,
|
| 218 |
+
"loss": 0.0489,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.22382671480144403,
|
| 223 |
+
"grad_norm": 0.21252156794071198,
|
| 224 |
+
"learning_rate": 3.1e-05,
|
| 225 |
+
"loss": 0.0434,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.23104693140794225,
|
| 230 |
+
"grad_norm": 0.189384326338768,
|
| 231 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 232 |
+
"loss": 0.0391,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.23826714801444043,
|
| 237 |
+
"grad_norm": 0.22159768640995026,
|
| 238 |
+
"learning_rate": 3.3e-05,
|
| 239 |
+
"loss": 0.0438,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.24548736462093862,
|
| 244 |
+
"grad_norm": 0.2974785268306732,
|
| 245 |
+
"learning_rate": 3.4000000000000007e-05,
|
| 246 |
+
"loss": 0.0421,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.2527075812274368,
|
| 251 |
+
"grad_norm": 0.21233530342578888,
|
| 252 |
+
"learning_rate": 3.5e-05,
|
| 253 |
+
"loss": 0.0419,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.259927797833935,
|
| 258 |
+
"grad_norm": 0.198137104511261,
|
| 259 |
+
"learning_rate": 3.6e-05,
|
| 260 |
+
"loss": 0.0411,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.26714801444043323,
|
| 265 |
+
"grad_norm": 0.29586824774742126,
|
| 266 |
+
"learning_rate": 3.7e-05,
|
| 267 |
+
"loss": 0.0451,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.2743682310469314,
|
| 272 |
+
"grad_norm": 0.1930268108844757,
|
| 273 |
+
"learning_rate": 3.8e-05,
|
| 274 |
+
"loss": 0.0401,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2815884476534296,
|
| 279 |
+
"grad_norm": 0.21896091103553772,
|
| 280 |
+
"learning_rate": 3.9000000000000006e-05,
|
| 281 |
+
"loss": 0.0393,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2888086642599278,
|
| 286 |
+
"grad_norm": 0.21269367635250092,
|
| 287 |
+
"learning_rate": 4e-05,
|
| 288 |
+
"loss": 0.0415,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.296028880866426,
|
| 293 |
+
"grad_norm": 0.2300368845462799,
|
| 294 |
+
"learning_rate": 4.1e-05,
|
| 295 |
+
"loss": 0.0436,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.30324909747292417,
|
| 300 |
+
"grad_norm": 0.17586210370063782,
|
| 301 |
+
"learning_rate": 4.2e-05,
|
| 302 |
+
"loss": 0.0417,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3104693140794224,
|
| 307 |
+
"grad_norm": 0.20255301892757416,
|
| 308 |
+
"learning_rate": 4.3e-05,
|
| 309 |
+
"loss": 0.0373,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3176895306859206,
|
| 314 |
+
"grad_norm": 0.29133227467536926,
|
| 315 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 316 |
+
"loss": 0.0427,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.3249097472924188,
|
| 321 |
+
"grad_norm": 0.14938658475875854,
|
| 322 |
+
"learning_rate": 4.5e-05,
|
| 323 |
+
"loss": 0.0402,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.33212996389891697,
|
| 328 |
+
"grad_norm": 0.20491056144237518,
|
| 329 |
+
"learning_rate": 4.600000000000001e-05,
|
| 330 |
+
"loss": 0.0407,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.33935018050541516,
|
| 335 |
+
"grad_norm": 0.38727957010269165,
|
| 336 |
+
"learning_rate": 4.7e-05,
|
| 337 |
+
"loss": 0.0393,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.34657039711191334,
|
| 342 |
+
"grad_norm": 0.2549976408481598,
|
| 343 |
+
"learning_rate": 4.8e-05,
|
| 344 |
+
"loss": 0.0381,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.35379061371841153,
|
| 349 |
+
"grad_norm": 0.21853362023830414,
|
| 350 |
+
"learning_rate": 4.9e-05,
|
| 351 |
+
"loss": 0.0407,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.36101083032490977,
|
| 356 |
+
"grad_norm": 0.2548893094062805,
|
| 357 |
+
"learning_rate": 5e-05,
|
| 358 |
+
"loss": 0.0366,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.36823104693140796,
|
| 363 |
+
"grad_norm": 0.2966635525226593,
|
| 364 |
+
"learning_rate": 4.986320109439125e-05,
|
| 365 |
+
"loss": 0.0388,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.37545126353790614,
|
| 370 |
+
"grad_norm": 0.2587813436985016,
|
| 371 |
+
"learning_rate": 4.9726402188782486e-05,
|
| 372 |
+
"loss": 0.0422,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.38267148014440433,
|
| 377 |
+
"grad_norm": 0.20156943798065186,
|
| 378 |
+
"learning_rate": 4.958960328317374e-05,
|
| 379 |
+
"loss": 0.0379,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.3898916967509025,
|
| 384 |
+
"grad_norm": 0.23318979144096375,
|
| 385 |
+
"learning_rate": 4.945280437756498e-05,
|
| 386 |
+
"loss": 0.0387,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3971119133574007,
|
| 391 |
+
"grad_norm": 0.21295292675495148,
|
| 392 |
+
"learning_rate": 4.931600547195623e-05,
|
| 393 |
+
"loss": 0.0363,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.4043321299638989,
|
| 398 |
+
"grad_norm": 0.23067612946033478,
|
| 399 |
+
"learning_rate": 4.917920656634747e-05,
|
| 400 |
+
"loss": 0.0396,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.41155234657039713,
|
| 405 |
+
"grad_norm": 0.1751265823841095,
|
| 406 |
+
"learning_rate": 4.904240766073871e-05,
|
| 407 |
+
"loss": 0.0386,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.4187725631768953,
|
| 412 |
+
"grad_norm": 0.32408931851387024,
|
| 413 |
+
"learning_rate": 4.8905608755129964e-05,
|
| 414 |
+
"loss": 0.0369,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.4259927797833935,
|
| 419 |
+
"grad_norm": 0.1949748396873474,
|
| 420 |
+
"learning_rate": 4.876880984952121e-05,
|
| 421 |
+
"loss": 0.0387,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.4332129963898917,
|
| 426 |
+
"grad_norm": 0.21588054299354553,
|
| 427 |
+
"learning_rate": 4.863201094391245e-05,
|
| 428 |
+
"loss": 0.0365,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.4404332129963899,
|
| 433 |
+
"grad_norm": 0.19144290685653687,
|
| 434 |
+
"learning_rate": 4.849521203830369e-05,
|
| 435 |
+
"loss": 0.035,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.44765342960288806,
|
| 440 |
+
"grad_norm": 0.25373053550720215,
|
| 441 |
+
"learning_rate": 4.835841313269494e-05,
|
| 442 |
+
"loss": 0.0391,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4548736462093863,
|
| 447 |
+
"grad_norm": 0.2404279261827469,
|
| 448 |
+
"learning_rate": 4.822161422708619e-05,
|
| 449 |
+
"loss": 0.0364,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.4620938628158845,
|
| 454 |
+
"grad_norm": 0.1697109341621399,
|
| 455 |
+
"learning_rate": 4.808481532147743e-05,
|
| 456 |
+
"loss": 0.0322,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.4693140794223827,
|
| 461 |
+
"grad_norm": 0.17318709194660187,
|
| 462 |
+
"learning_rate": 4.7948016415868674e-05,
|
| 463 |
+
"loss": 0.0365,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.47653429602888087,
|
| 468 |
+
"grad_norm": 0.18541967868804932,
|
| 469 |
+
"learning_rate": 4.781121751025992e-05,
|
| 470 |
+
"loss": 0.0394,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.48375451263537905,
|
| 475 |
+
"grad_norm": 0.21878574788570404,
|
| 476 |
+
"learning_rate": 4.7674418604651164e-05,
|
| 477 |
+
"loss": 0.0336,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.49097472924187724,
|
| 482 |
+
"grad_norm": 0.2381717413663864,
|
| 483 |
+
"learning_rate": 4.753761969904241e-05,
|
| 484 |
+
"loss": 0.0343,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.4981949458483754,
|
| 489 |
+
"grad_norm": 0.2559703290462494,
|
| 490 |
+
"learning_rate": 4.7400820793433654e-05,
|
| 491 |
+
"loss": 0.0356,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.5054151624548736,
|
| 496 |
+
"grad_norm": 0.25602829456329346,
|
| 497 |
+
"learning_rate": 4.72640218878249e-05,
|
| 498 |
+
"loss": 0.0341,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.5126353790613718,
|
| 503 |
+
"grad_norm": 0.30914345383644104,
|
| 504 |
+
"learning_rate": 4.7127222982216145e-05,
|
| 505 |
+
"loss": 0.0378,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.51985559566787,
|
| 510 |
+
"grad_norm": 0.24148257076740265,
|
| 511 |
+
"learning_rate": 4.699042407660739e-05,
|
| 512 |
+
"loss": 0.0329,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5270758122743683,
|
| 517 |
+
"grad_norm": 0.16448011994361877,
|
| 518 |
+
"learning_rate": 4.6853625170998635e-05,
|
| 519 |
+
"loss": 0.0373,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.5342960288808665,
|
| 524 |
+
"grad_norm": 0.20992177724838257,
|
| 525 |
+
"learning_rate": 4.671682626538988e-05,
|
| 526 |
+
"loss": 0.0358,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.5415162454873647,
|
| 531 |
+
"grad_norm": 0.1955261081457138,
|
| 532 |
+
"learning_rate": 4.6580027359781126e-05,
|
| 533 |
+
"loss": 0.0373,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5487364620938628,
|
| 538 |
+
"grad_norm": 0.1674635410308838,
|
| 539 |
+
"learning_rate": 4.6443228454172364e-05,
|
| 540 |
+
"loss": 0.0332,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.555956678700361,
|
| 545 |
+
"grad_norm": 0.22619013488292694,
|
| 546 |
+
"learning_rate": 4.630642954856361e-05,
|
| 547 |
+
"loss": 0.0375,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5631768953068592,
|
| 552 |
+
"grad_norm": 0.20930354297161102,
|
| 553 |
+
"learning_rate": 4.616963064295486e-05,
|
| 554 |
+
"loss": 0.0385,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5703971119133574,
|
| 559 |
+
"grad_norm": 0.23155862092971802,
|
| 560 |
+
"learning_rate": 4.6032831737346106e-05,
|
| 561 |
+
"loss": 0.0325,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5776173285198556,
|
| 566 |
+
"grad_norm": 0.25036847591400146,
|
| 567 |
+
"learning_rate": 4.5896032831737345e-05,
|
| 568 |
+
"loss": 0.0326,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5848375451263538,
|
| 573 |
+
"grad_norm": 0.17637784779071808,
|
| 574 |
+
"learning_rate": 4.575923392612859e-05,
|
| 575 |
+
"loss": 0.0352,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.592057761732852,
|
| 580 |
+
"grad_norm": 0.2410997748374939,
|
| 581 |
+
"learning_rate": 4.5622435020519835e-05,
|
| 582 |
+
"loss": 0.0343,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.5992779783393501,
|
| 587 |
+
"grad_norm": 0.18689364194869995,
|
| 588 |
+
"learning_rate": 4.548563611491109e-05,
|
| 589 |
+
"loss": 0.0337,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.6064981949458483,
|
| 594 |
+
"grad_norm": 0.1366495043039322,
|
| 595 |
+
"learning_rate": 4.5348837209302326e-05,
|
| 596 |
+
"loss": 0.0332,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.6137184115523465,
|
| 601 |
+
"grad_norm": 0.1845828890800476,
|
| 602 |
+
"learning_rate": 4.521203830369357e-05,
|
| 603 |
+
"loss": 0.0337,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.6209386281588448,
|
| 608 |
+
"grad_norm": 0.17139053344726562,
|
| 609 |
+
"learning_rate": 4.5075239398084816e-05,
|
| 610 |
+
"loss": 0.0373,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.628158844765343,
|
| 615 |
+
"grad_norm": 0.25084948539733887,
|
| 616 |
+
"learning_rate": 4.493844049247606e-05,
|
| 617 |
+
"loss": 0.0347,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.6353790613718412,
|
| 622 |
+
"grad_norm": 0.23243854939937592,
|
| 623 |
+
"learning_rate": 4.4801641586867306e-05,
|
| 624 |
+
"loss": 0.0298,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.6425992779783394,
|
| 629 |
+
"grad_norm": 0.16921032965183258,
|
| 630 |
+
"learning_rate": 4.466484268125855e-05,
|
| 631 |
+
"loss": 0.0335,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.6498194945848376,
|
| 636 |
+
"grad_norm": 0.21709440648555756,
|
| 637 |
+
"learning_rate": 4.45280437756498e-05,
|
| 638 |
+
"loss": 0.0294,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.6570397111913358,
|
| 643 |
+
"grad_norm": 0.25876304507255554,
|
| 644 |
+
"learning_rate": 4.439124487004104e-05,
|
| 645 |
+
"loss": 0.0355,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6642599277978339,
|
| 650 |
+
"grad_norm": 0.23916779458522797,
|
| 651 |
+
"learning_rate": 4.425444596443229e-05,
|
| 652 |
+
"loss": 0.0301,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6714801444043321,
|
| 657 |
+
"grad_norm": 0.19482851028442383,
|
| 658 |
+
"learning_rate": 4.411764705882353e-05,
|
| 659 |
+
"loss": 0.0331,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6787003610108303,
|
| 664 |
+
"grad_norm": 0.1518426537513733,
|
| 665 |
+
"learning_rate": 4.398084815321478e-05,
|
| 666 |
+
"loss": 0.0313,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6859205776173285,
|
| 671 |
+
"grad_norm": 0.17477025091648102,
|
| 672 |
+
"learning_rate": 4.384404924760602e-05,
|
| 673 |
+
"loss": 0.0284,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6931407942238267,
|
| 678 |
+
"grad_norm": 0.2821789085865021,
|
| 679 |
+
"learning_rate": 4.370725034199726e-05,
|
| 680 |
+
"loss": 0.0294,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.7003610108303249,
|
| 685 |
+
"grad_norm": 0.19530260562896729,
|
| 686 |
+
"learning_rate": 4.3570451436388506e-05,
|
| 687 |
+
"loss": 0.0307,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.7075812274368231,
|
| 692 |
+
"grad_norm": 0.2189357876777649,
|
| 693 |
+
"learning_rate": 4.343365253077976e-05,
|
| 694 |
+
"loss": 0.0325,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.7148014440433214,
|
| 699 |
+
"grad_norm": 0.18680965900421143,
|
| 700 |
+
"learning_rate": 4.3296853625171004e-05,
|
| 701 |
+
"loss": 0.0298,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.7220216606498195,
|
| 706 |
+
"grad_norm": 0.20822104811668396,
|
| 707 |
+
"learning_rate": 4.316005471956224e-05,
|
| 708 |
+
"loss": 0.0318,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.7292418772563177,
|
| 713 |
+
"grad_norm": 0.17637012898921967,
|
| 714 |
+
"learning_rate": 4.302325581395349e-05,
|
| 715 |
+
"loss": 0.0267,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.7364620938628159,
|
| 720 |
+
"grad_norm": 0.16878972947597504,
|
| 721 |
+
"learning_rate": 4.288645690834473e-05,
|
| 722 |
+
"loss": 0.0285,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.7436823104693141,
|
| 727 |
+
"grad_norm": 0.3033148944377899,
|
| 728 |
+
"learning_rate": 4.2749658002735984e-05,
|
| 729 |
+
"loss": 0.0315,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.7509025270758123,
|
| 734 |
+
"grad_norm": 0.21088653802871704,
|
| 735 |
+
"learning_rate": 4.261285909712722e-05,
|
| 736 |
+
"loss": 0.0329,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.7581227436823105,
|
| 741 |
+
"grad_norm": 0.20485380291938782,
|
| 742 |
+
"learning_rate": 4.247606019151847e-05,
|
| 743 |
+
"loss": 0.0321,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.7653429602888087,
|
| 748 |
+
"grad_norm": 0.21797245740890503,
|
| 749 |
+
"learning_rate": 4.233926128590971e-05,
|
| 750 |
+
"loss": 0.0288,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7725631768953068,
|
| 755 |
+
"grad_norm": 0.1585124433040619,
|
| 756 |
+
"learning_rate": 4.220246238030096e-05,
|
| 757 |
+
"loss": 0.0332,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.779783393501805,
|
| 762 |
+
"grad_norm": 0.18445636332035065,
|
| 763 |
+
"learning_rate": 4.2065663474692204e-05,
|
| 764 |
+
"loss": 0.0317,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7870036101083032,
|
| 769 |
+
"grad_norm": 0.2596589922904968,
|
| 770 |
+
"learning_rate": 4.192886456908345e-05,
|
| 771 |
+
"loss": 0.0283,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7942238267148014,
|
| 776 |
+
"grad_norm": 0.15991349518299103,
|
| 777 |
+
"learning_rate": 4.1792065663474694e-05,
|
| 778 |
+
"loss": 0.027,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.8014440433212996,
|
| 783 |
+
"grad_norm": 0.2013750970363617,
|
| 784 |
+
"learning_rate": 4.165526675786594e-05,
|
| 785 |
+
"loss": 0.0285,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.8086642599277978,
|
| 790 |
+
"grad_norm": 0.16817732155323029,
|
| 791 |
+
"learning_rate": 4.1518467852257184e-05,
|
| 792 |
+
"loss": 0.0304,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.8158844765342961,
|
| 797 |
+
"grad_norm": 0.20662030577659607,
|
| 798 |
+
"learning_rate": 4.138166894664843e-05,
|
| 799 |
+
"loss": 0.0309,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.8231046931407943,
|
| 804 |
+
"grad_norm": 0.24165768921375275,
|
| 805 |
+
"learning_rate": 4.1244870041039675e-05,
|
| 806 |
+
"loss": 0.0305,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.8303249097472925,
|
| 811 |
+
"grad_norm": 0.15945856273174286,
|
| 812 |
+
"learning_rate": 4.110807113543092e-05,
|
| 813 |
+
"loss": 0.0331,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.8375451263537906,
|
| 818 |
+
"grad_norm": 0.23878367245197296,
|
| 819 |
+
"learning_rate": 4.097127222982216e-05,
|
| 820 |
+
"loss": 0.0287,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.8447653429602888,
|
| 825 |
+
"grad_norm": 0.20223484933376312,
|
| 826 |
+
"learning_rate": 4.083447332421341e-05,
|
| 827 |
+
"loss": 0.0283,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.851985559566787,
|
| 832 |
+
"grad_norm": 0.23693232238292694,
|
| 833 |
+
"learning_rate": 4.0697674418604655e-05,
|
| 834 |
+
"loss": 0.0283,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.8592057761732852,
|
| 839 |
+
"grad_norm": 0.21208103001117706,
|
| 840 |
+
"learning_rate": 4.05608755129959e-05,
|
| 841 |
+
"loss": 0.0291,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.8664259927797834,
|
| 846 |
+
"grad_norm": 0.21978527307510376,
|
| 847 |
+
"learning_rate": 4.042407660738714e-05,
|
| 848 |
+
"loss": 0.031,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.8736462093862816,
|
| 853 |
+
"grad_norm": 0.18041111528873444,
|
| 854 |
+
"learning_rate": 4.0287277701778384e-05,
|
| 855 |
+
"loss": 0.0266,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.8808664259927798,
|
| 860 |
+
"grad_norm": 0.15536178648471832,
|
| 861 |
+
"learning_rate": 4.015047879616963e-05,
|
| 862 |
+
"loss": 0.0311,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.8880866425992779,
|
| 867 |
+
"grad_norm": 0.17821329832077026,
|
| 868 |
+
"learning_rate": 4.001367989056088e-05,
|
| 869 |
+
"loss": 0.0303,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.8953068592057761,
|
| 874 |
+
"grad_norm": 0.22445450723171234,
|
| 875 |
+
"learning_rate": 3.987688098495212e-05,
|
| 876 |
+
"loss": 0.0249,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.9025270758122743,
|
| 881 |
+
"grad_norm": 0.22271278500556946,
|
| 882 |
+
"learning_rate": 3.9740082079343365e-05,
|
| 883 |
+
"loss": 0.0287,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.9097472924187726,
|
| 888 |
+
"grad_norm": 0.23373781144618988,
|
| 889 |
+
"learning_rate": 3.960328317373461e-05,
|
| 890 |
+
"loss": 0.0301,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.9169675090252708,
|
| 895 |
+
"grad_norm": 0.1760292649269104,
|
| 896 |
+
"learning_rate": 3.9466484268125855e-05,
|
| 897 |
+
"loss": 0.0266,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.924187725631769,
|
| 902 |
+
"grad_norm": 0.19504322111606598,
|
| 903 |
+
"learning_rate": 3.93296853625171e-05,
|
| 904 |
+
"loss": 0.027,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.9314079422382672,
|
| 909 |
+
"grad_norm": 0.23191095888614655,
|
| 910 |
+
"learning_rate": 3.9192886456908346e-05,
|
| 911 |
+
"loss": 0.0275,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.9386281588447654,
|
| 916 |
+
"grad_norm": 0.28094080090522766,
|
| 917 |
+
"learning_rate": 3.905608755129959e-05,
|
| 918 |
+
"loss": 0.0299,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.9458483754512635,
|
| 923 |
+
"grad_norm": 0.15457668900489807,
|
| 924 |
+
"learning_rate": 3.8919288645690836e-05,
|
| 925 |
+
"loss": 0.0281,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.9530685920577617,
|
| 930 |
+
"grad_norm": 0.20441775023937225,
|
| 931 |
+
"learning_rate": 3.878248974008208e-05,
|
| 932 |
+
"loss": 0.0314,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.9602888086642599,
|
| 937 |
+
"grad_norm": 0.2961263954639435,
|
| 938 |
+
"learning_rate": 3.864569083447333e-05,
|
| 939 |
+
"loss": 0.0285,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.9675090252707581,
|
| 944 |
+
"grad_norm": 0.16792945563793182,
|
| 945 |
+
"learning_rate": 3.850889192886457e-05,
|
| 946 |
+
"loss": 0.0277,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.9747292418772563,
|
| 951 |
+
"grad_norm": 0.3934960961341858,
|
| 952 |
+
"learning_rate": 3.837209302325582e-05,
|
| 953 |
+
"loss": 0.0303,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.9819494584837545,
|
| 958 |
+
"grad_norm": 0.24637605249881744,
|
| 959 |
+
"learning_rate": 3.8235294117647055e-05,
|
| 960 |
+
"loss": 0.0272,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.9891696750902527,
|
| 965 |
+
"grad_norm": 0.21808254718780518,
|
| 966 |
+
"learning_rate": 3.809849521203831e-05,
|
| 967 |
+
"loss": 0.0266,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.9963898916967509,
|
| 972 |
+
"grad_norm": 0.2921408712863922,
|
| 973 |
+
"learning_rate": 3.796169630642955e-05,
|
| 974 |
+
"loss": 0.0274,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 1.0,
|
| 979 |
+
"eval_accuracy": 0.9881444816395965,
|
| 980 |
+
"eval_f1": 0.8667477505754343,
|
| 981 |
+
"eval_loss": 0.030792543664574623,
|
| 982 |
+
"eval_precision": 0.8156923076923077,
|
| 983 |
+
"eval_recall": 0.9246212226903653,
|
| 984 |
+
"eval_runtime": 271.409,
|
| 985 |
+
"eval_samples_per_second": 81.464,
|
| 986 |
+
"eval_steps_per_second": 1.275,
|
| 987 |
+
"step": 1385
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.0036101083032491,
|
| 991 |
+
"grad_norm": 0.18204669654369354,
|
| 992 |
+
"learning_rate": 3.78248974008208e-05,
|
| 993 |
+
"loss": 0.0265,
|
| 994 |
+
"step": 1390
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.0108303249097472,
|
| 998 |
+
"grad_norm": 0.13540150225162506,
|
| 999 |
+
"learning_rate": 3.7688098495212036e-05,
|
| 1000 |
+
"loss": 0.0239,
|
| 1001 |
+
"step": 1400
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.0180505415162455,
|
| 1005 |
+
"grad_norm": 0.2114001214504242,
|
| 1006 |
+
"learning_rate": 3.755129958960328e-05,
|
| 1007 |
+
"loss": 0.0208,
|
| 1008 |
+
"step": 1410
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 1.0252707581227436,
|
| 1012 |
+
"grad_norm": 0.20667189359664917,
|
| 1013 |
+
"learning_rate": 3.741450068399453e-05,
|
| 1014 |
+
"loss": 0.0282,
|
| 1015 |
+
"step": 1420
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.032490974729242,
|
| 1019 |
+
"grad_norm": 0.16794054210186005,
|
| 1020 |
+
"learning_rate": 3.727770177838578e-05,
|
| 1021 |
+
"loss": 0.0216,
|
| 1022 |
+
"step": 1430
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 1.03971119133574,
|
| 1026 |
+
"grad_norm": 0.19682972133159637,
|
| 1027 |
+
"learning_rate": 3.714090287277702e-05,
|
| 1028 |
+
"loss": 0.0209,
|
| 1029 |
+
"step": 1440
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 1.0469314079422383,
|
| 1033 |
+
"grad_norm": 0.18295951187610626,
|
| 1034 |
+
"learning_rate": 3.700410396716826e-05,
|
| 1035 |
+
"loss": 0.0251,
|
| 1036 |
+
"step": 1450
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 1.0541516245487366,
|
| 1040 |
+
"grad_norm": 0.1491762399673462,
|
| 1041 |
+
"learning_rate": 3.686730506155951e-05,
|
| 1042 |
+
"loss": 0.0215,
|
| 1043 |
+
"step": 1460
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 1.0613718411552346,
|
| 1047 |
+
"grad_norm": 0.3141365647315979,
|
| 1048 |
+
"learning_rate": 3.673050615595075e-05,
|
| 1049 |
+
"loss": 0.0235,
|
| 1050 |
+
"step": 1470
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 1.068592057761733,
|
| 1054 |
+
"grad_norm": 0.18586914241313934,
|
| 1055 |
+
"learning_rate": 3.6593707250342e-05,
|
| 1056 |
+
"loss": 0.0226,
|
| 1057 |
+
"step": 1480
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.075812274368231,
|
| 1061 |
+
"grad_norm": 0.26788514852523804,
|
| 1062 |
+
"learning_rate": 3.645690834473324e-05,
|
| 1063 |
+
"loss": 0.0214,
|
| 1064 |
+
"step": 1490
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 1.0830324909747293,
|
| 1068 |
+
"grad_norm": 0.18625912070274353,
|
| 1069 |
+
"learning_rate": 3.632010943912449e-05,
|
| 1070 |
+
"loss": 0.0223,
|
| 1071 |
+
"step": 1500
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 1.0902527075812274,
|
| 1075 |
+
"grad_norm": 0.23848237097263336,
|
| 1076 |
+
"learning_rate": 3.6183310533515733e-05,
|
| 1077 |
+
"loss": 0.0222,
|
| 1078 |
+
"step": 1510
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 1.0974729241877257,
|
| 1082 |
+
"grad_norm": 0.24550049006938934,
|
| 1083 |
+
"learning_rate": 3.604651162790698e-05,
|
| 1084 |
+
"loss": 0.0223,
|
| 1085 |
+
"step": 1520
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 1.1046931407942238,
|
| 1089 |
+
"grad_norm": 0.33519259095191956,
|
| 1090 |
+
"learning_rate": 3.5909712722298224e-05,
|
| 1091 |
+
"loss": 0.0207,
|
| 1092 |
+
"step": 1530
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 1.111913357400722,
|
| 1096 |
+
"grad_norm": 0.19318382441997528,
|
| 1097 |
+
"learning_rate": 3.577291381668947e-05,
|
| 1098 |
+
"loss": 0.0255,
|
| 1099 |
+
"step": 1540
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.1191335740072201,
|
| 1103 |
+
"grad_norm": 0.23220829665660858,
|
| 1104 |
+
"learning_rate": 3.5636114911080714e-05,
|
| 1105 |
+
"loss": 0.0218,
|
| 1106 |
+
"step": 1550
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 1.1263537906137184,
|
| 1110 |
+
"grad_norm": 0.1848268061876297,
|
| 1111 |
+
"learning_rate": 3.549931600547195e-05,
|
| 1112 |
+
"loss": 0.0252,
|
| 1113 |
+
"step": 1560
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 1.1335740072202167,
|
| 1117 |
+
"grad_norm": 0.20493029057979584,
|
| 1118 |
+
"learning_rate": 3.5362517099863205e-05,
|
| 1119 |
+
"loss": 0.0233,
|
| 1120 |
+
"step": 1570
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 1.1407942238267148,
|
| 1124 |
+
"grad_norm": 0.19206510484218597,
|
| 1125 |
+
"learning_rate": 3.522571819425445e-05,
|
| 1126 |
+
"loss": 0.0227,
|
| 1127 |
+
"step": 1580
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 1.1480144404332129,
|
| 1131 |
+
"grad_norm": 0.25989973545074463,
|
| 1132 |
+
"learning_rate": 3.5088919288645695e-05,
|
| 1133 |
+
"loss": 0.0205,
|
| 1134 |
+
"step": 1590
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 1.1552346570397112,
|
| 1138 |
+
"grad_norm": 0.20777900516986847,
|
| 1139 |
+
"learning_rate": 3.4952120383036933e-05,
|
| 1140 |
+
"loss": 0.026,
|
| 1141 |
+
"step": 1600
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.1624548736462095,
|
| 1145 |
+
"grad_norm": 0.25129932165145874,
|
| 1146 |
+
"learning_rate": 3.481532147742818e-05,
|
| 1147 |
+
"loss": 0.0246,
|
| 1148 |
+
"step": 1610
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 1.1696750902527075,
|
| 1152 |
+
"grad_norm": 0.245501309633255,
|
| 1153 |
+
"learning_rate": 3.467852257181943e-05,
|
| 1154 |
+
"loss": 0.0232,
|
| 1155 |
+
"step": 1620
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.1768953068592058,
|
| 1159 |
+
"grad_norm": 0.19470812380313873,
|
| 1160 |
+
"learning_rate": 3.4541723666210676e-05,
|
| 1161 |
+
"loss": 0.02,
|
| 1162 |
+
"step": 1630
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 1.184115523465704,
|
| 1166 |
+
"grad_norm": 0.23183578252792358,
|
| 1167 |
+
"learning_rate": 3.4404924760601914e-05,
|
| 1168 |
+
"loss": 0.0257,
|
| 1169 |
+
"step": 1640
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 1.1913357400722022,
|
| 1173 |
+
"grad_norm": 0.2162778079509735,
|
| 1174 |
+
"learning_rate": 3.426812585499316e-05,
|
| 1175 |
+
"loss": 0.0242,
|
| 1176 |
+
"step": 1650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.1985559566787003,
|
| 1180 |
+
"grad_norm": 0.15980581939220428,
|
| 1181 |
+
"learning_rate": 3.4131326949384405e-05,
|
| 1182 |
+
"loss": 0.0242,
|
| 1183 |
+
"step": 1660
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.2057761732851986,
|
| 1187 |
+
"grad_norm": 0.1984899491071701,
|
| 1188 |
+
"learning_rate": 3.399452804377565e-05,
|
| 1189 |
+
"loss": 0.0235,
|
| 1190 |
+
"step": 1670
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 1.2129963898916967,
|
| 1194 |
+
"grad_norm": 0.19146302342414856,
|
| 1195 |
+
"learning_rate": 3.3857729138166895e-05,
|
| 1196 |
+
"loss": 0.0212,
|
| 1197 |
+
"step": 1680
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 1.220216606498195,
|
| 1201 |
+
"grad_norm": 0.19686651229858398,
|
| 1202 |
+
"learning_rate": 3.372093023255814e-05,
|
| 1203 |
+
"loss": 0.0242,
|
| 1204 |
+
"step": 1690
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 1.2274368231046933,
|
| 1208 |
+
"grad_norm": 0.17915378510951996,
|
| 1209 |
+
"learning_rate": 3.3584131326949385e-05,
|
| 1210 |
+
"loss": 0.0264,
|
| 1211 |
+
"step": 1700
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 1.2346570397111913,
|
| 1215 |
+
"grad_norm": 0.18991689383983612,
|
| 1216 |
+
"learning_rate": 3.344733242134063e-05,
|
| 1217 |
+
"loss": 0.0227,
|
| 1218 |
+
"step": 1710
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 1.2418772563176894,
|
| 1222 |
+
"grad_norm": 0.16964809596538544,
|
| 1223 |
+
"learning_rate": 3.3310533515731876e-05,
|
| 1224 |
+
"loss": 0.0229,
|
| 1225 |
+
"step": 1720
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.2490974729241877,
|
| 1229 |
+
"grad_norm": 0.33044272661209106,
|
| 1230 |
+
"learning_rate": 3.317373461012312e-05,
|
| 1231 |
+
"loss": 0.0233,
|
| 1232 |
+
"step": 1730
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 1.256317689530686,
|
| 1236 |
+
"grad_norm": 0.27375954389572144,
|
| 1237 |
+
"learning_rate": 3.3036935704514366e-05,
|
| 1238 |
+
"loss": 0.023,
|
| 1239 |
+
"step": 1740
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.263537906137184,
|
| 1243 |
+
"grad_norm": 0.19099849462509155,
|
| 1244 |
+
"learning_rate": 3.290013679890561e-05,
|
| 1245 |
+
"loss": 0.0199,
|
| 1246 |
+
"step": 1750
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 1.2707581227436824,
|
| 1250 |
+
"grad_norm": 0.21204090118408203,
|
| 1251 |
+
"learning_rate": 3.276333789329685e-05,
|
| 1252 |
+
"loss": 0.0228,
|
| 1253 |
+
"step": 1760
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 1.2779783393501805,
|
| 1257 |
+
"grad_norm": 0.17759856581687927,
|
| 1258 |
+
"learning_rate": 3.26265389876881e-05,
|
| 1259 |
+
"loss": 0.0209,
|
| 1260 |
+
"step": 1770
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 1.2851985559566788,
|
| 1264 |
+
"grad_norm": 0.19070599973201752,
|
| 1265 |
+
"learning_rate": 3.248974008207935e-05,
|
| 1266 |
+
"loss": 0.0223,
|
| 1267 |
+
"step": 1780
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.2924187725631768,
|
| 1271 |
+
"grad_norm": 0.18185602128505707,
|
| 1272 |
+
"learning_rate": 3.235294117647059e-05,
|
| 1273 |
+
"loss": 0.0201,
|
| 1274 |
+
"step": 1790
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.2996389891696751,
|
| 1278 |
+
"grad_norm": 0.20534738898277283,
|
| 1279 |
+
"learning_rate": 3.221614227086183e-05,
|
| 1280 |
+
"loss": 0.0202,
|
| 1281 |
+
"step": 1800
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.3068592057761732,
|
| 1285 |
+
"grad_norm": 0.2479701191186905,
|
| 1286 |
+
"learning_rate": 3.2079343365253076e-05,
|
| 1287 |
+
"loss": 0.019,
|
| 1288 |
+
"step": 1810
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 1.3140794223826715,
|
| 1292 |
+
"grad_norm": 0.2699389159679413,
|
| 1293 |
+
"learning_rate": 3.194254445964433e-05,
|
| 1294 |
+
"loss": 0.023,
|
| 1295 |
+
"step": 1820
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 1.3212996389891698,
|
| 1299 |
+
"grad_norm": 0.22228173911571503,
|
| 1300 |
+
"learning_rate": 3.180574555403557e-05,
|
| 1301 |
+
"loss": 0.0199,
|
| 1302 |
+
"step": 1830
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 1.3285198555956679,
|
| 1306 |
+
"grad_norm": 0.1938995122909546,
|
| 1307 |
+
"learning_rate": 3.166894664842681e-05,
|
| 1308 |
+
"loss": 0.0235,
|
| 1309 |
+
"step": 1840
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.335740072202166,
|
| 1313 |
+
"grad_norm": 0.17550387978553772,
|
| 1314 |
+
"learning_rate": 3.1532147742818057e-05,
|
| 1315 |
+
"loss": 0.02,
|
| 1316 |
+
"step": 1850
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 1.3429602888086642,
|
| 1320 |
+
"grad_norm": 0.13713954389095306,
|
| 1321 |
+
"learning_rate": 3.13953488372093e-05,
|
| 1322 |
+
"loss": 0.0239,
|
| 1323 |
+
"step": 1860
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 1.3501805054151625,
|
| 1327 |
+
"grad_norm": 0.19476300477981567,
|
| 1328 |
+
"learning_rate": 3.125854993160055e-05,
|
| 1329 |
+
"loss": 0.0197,
|
| 1330 |
+
"step": 1870
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 1.3574007220216606,
|
| 1334 |
+
"grad_norm": 0.23414915800094604,
|
| 1335 |
+
"learning_rate": 3.112175102599179e-05,
|
| 1336 |
+
"loss": 0.0179,
|
| 1337 |
+
"step": 1880
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 1.364620938628159,
|
| 1341 |
+
"grad_norm": 0.20327487587928772,
|
| 1342 |
+
"learning_rate": 3.098495212038304e-05,
|
| 1343 |
+
"loss": 0.0215,
|
| 1344 |
+
"step": 1890
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 1.371841155234657,
|
| 1348 |
+
"grad_norm": 0.2250749170780182,
|
| 1349 |
+
"learning_rate": 3.084815321477428e-05,
|
| 1350 |
+
"loss": 0.0213,
|
| 1351 |
+
"step": 1900
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.3790613718411553,
|
| 1355 |
+
"grad_norm": 0.17825260758399963,
|
| 1356 |
+
"learning_rate": 3.071135430916553e-05,
|
| 1357 |
+
"loss": 0.0199,
|
| 1358 |
+
"step": 1910
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.3862815884476534,
|
| 1362 |
+
"grad_norm": 0.22727660834789276,
|
| 1363 |
+
"learning_rate": 3.057455540355677e-05,
|
| 1364 |
+
"loss": 0.0223,
|
| 1365 |
+
"step": 1920
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 1.3935018050541517,
|
| 1369 |
+
"grad_norm": 0.18458124995231628,
|
| 1370 |
+
"learning_rate": 3.0437756497948018e-05,
|
| 1371 |
+
"loss": 0.0223,
|
| 1372 |
+
"step": 1930
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 1.4007220216606497,
|
| 1376 |
+
"grad_norm": 0.2251572608947754,
|
| 1377 |
+
"learning_rate": 3.0300957592339263e-05,
|
| 1378 |
+
"loss": 0.0227,
|
| 1379 |
+
"step": 1940
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 1.407942238267148,
|
| 1383 |
+
"grad_norm": 0.2070099413394928,
|
| 1384 |
+
"learning_rate": 3.016415868673051e-05,
|
| 1385 |
+
"loss": 0.023,
|
| 1386 |
+
"step": 1950
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 1.4151624548736463,
|
| 1390 |
+
"grad_norm": 0.1247980073094368,
|
| 1391 |
+
"learning_rate": 3.002735978112175e-05,
|
| 1392 |
+
"loss": 0.0201,
|
| 1393 |
+
"step": 1960
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.4223826714801444,
|
| 1397 |
+
"grad_norm": 0.19043412804603577,
|
| 1398 |
+
"learning_rate": 2.9890560875512996e-05,
|
| 1399 |
+
"loss": 0.0186,
|
| 1400 |
+
"step": 1970
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 1.4296028880866425,
|
| 1404 |
+
"grad_norm": 0.22574257850646973,
|
| 1405 |
+
"learning_rate": 2.9753761969904244e-05,
|
| 1406 |
+
"loss": 0.0189,
|
| 1407 |
+
"step": 1980
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 1.4368231046931408,
|
| 1411 |
+
"grad_norm": 0.17224276065826416,
|
| 1412 |
+
"learning_rate": 2.961696306429549e-05,
|
| 1413 |
+
"loss": 0.0205,
|
| 1414 |
+
"step": 1990
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 1.444043321299639,
|
| 1418 |
+
"grad_norm": 0.18050329387187958,
|
| 1419 |
+
"learning_rate": 2.9480164158686728e-05,
|
| 1420 |
+
"loss": 0.0215,
|
| 1421 |
+
"step": 2000
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 1.4512635379061372,
|
| 1425 |
+
"grad_norm": 0.23059330880641937,
|
| 1426 |
+
"learning_rate": 2.9343365253077976e-05,
|
| 1427 |
+
"loss": 0.0225,
|
| 1428 |
+
"step": 2010
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 1.4584837545126355,
|
| 1432 |
+
"grad_norm": 0.24907854199409485,
|
| 1433 |
+
"learning_rate": 2.920656634746922e-05,
|
| 1434 |
+
"loss": 0.0204,
|
| 1435 |
+
"step": 2020
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.4657039711191335,
|
| 1439 |
+
"grad_norm": 0.1699414849281311,
|
| 1440 |
+
"learning_rate": 2.9069767441860467e-05,
|
| 1441 |
+
"loss": 0.024,
|
| 1442 |
+
"step": 2030
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.4729241877256318,
|
| 1446 |
+
"grad_norm": 0.16529108583927155,
|
| 1447 |
+
"learning_rate": 2.893296853625171e-05,
|
| 1448 |
+
"loss": 0.0196,
|
| 1449 |
+
"step": 2040
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.48014440433213,
|
| 1453 |
+
"grad_norm": 0.18618591129779816,
|
| 1454 |
+
"learning_rate": 2.8796169630642954e-05,
|
| 1455 |
+
"loss": 0.0235,
|
| 1456 |
+
"step": 2050
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 1.4873646209386282,
|
| 1460 |
+
"grad_norm": 0.15030409395694733,
|
| 1461 |
+
"learning_rate": 2.8659370725034202e-05,
|
| 1462 |
+
"loss": 0.0217,
|
| 1463 |
+
"step": 2060
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 1.4945848375451263,
|
| 1467 |
+
"grad_norm": 0.28873658180236816,
|
| 1468 |
+
"learning_rate": 2.8522571819425448e-05,
|
| 1469 |
+
"loss": 0.0249,
|
| 1470 |
+
"step": 2070
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 1.5018050541516246,
|
| 1474 |
+
"grad_norm": 0.19894906878471375,
|
| 1475 |
+
"learning_rate": 2.838577291381669e-05,
|
| 1476 |
+
"loss": 0.0217,
|
| 1477 |
+
"step": 2080
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.5090252707581229,
|
| 1481 |
+
"grad_norm": 0.20735248923301697,
|
| 1482 |
+
"learning_rate": 2.8248974008207935e-05,
|
| 1483 |
+
"loss": 0.0223,
|
| 1484 |
+
"step": 2090
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 1.516245487364621,
|
| 1488 |
+
"grad_norm": 0.13647274672985077,
|
| 1489 |
+
"learning_rate": 2.811217510259918e-05,
|
| 1490 |
+
"loss": 0.02,
|
| 1491 |
+
"step": 2100
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 1.523465703971119,
|
| 1495 |
+
"grad_norm": 0.16693101823329926,
|
| 1496 |
+
"learning_rate": 2.797537619699043e-05,
|
| 1497 |
+
"loss": 0.0243,
|
| 1498 |
+
"step": 2110
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.5306859205776173,
|
| 1502 |
+
"grad_norm": 0.21320907771587372,
|
| 1503 |
+
"learning_rate": 2.7838577291381667e-05,
|
| 1504 |
+
"loss": 0.0222,
|
| 1505 |
+
"step": 2120
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 1.5379061371841156,
|
| 1509 |
+
"grad_norm": 0.2132214903831482,
|
| 1510 |
+
"learning_rate": 2.7701778385772915e-05,
|
| 1511 |
+
"loss": 0.0219,
|
| 1512 |
+
"step": 2130
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 1.5451263537906137,
|
| 1516 |
+
"grad_norm": 0.1506602168083191,
|
| 1517 |
+
"learning_rate": 2.756497948016416e-05,
|
| 1518 |
+
"loss": 0.0224,
|
| 1519 |
+
"step": 2140
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.5523465703971118,
|
| 1523 |
+
"grad_norm": 0.18722060322761536,
|
| 1524 |
+
"learning_rate": 2.7428180574555406e-05,
|
| 1525 |
+
"loss": 0.0222,
|
| 1526 |
+
"step": 2150
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.55956678700361,
|
| 1530 |
+
"grad_norm": 0.22277644276618958,
|
| 1531 |
+
"learning_rate": 2.7291381668946648e-05,
|
| 1532 |
+
"loss": 0.0248,
|
| 1533 |
+
"step": 2160
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 1.5667870036101084,
|
| 1537 |
+
"grad_norm": 0.16420182585716248,
|
| 1538 |
+
"learning_rate": 2.7154582763337893e-05,
|
| 1539 |
+
"loss": 0.0197,
|
| 1540 |
+
"step": 2170
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.5740072202166067,
|
| 1544 |
+
"grad_norm": 0.23195837438106537,
|
| 1545 |
+
"learning_rate": 2.701778385772914e-05,
|
| 1546 |
+
"loss": 0.0213,
|
| 1547 |
+
"step": 2180
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 1.5812274368231047,
|
| 1551 |
+
"grad_norm": 0.19835099577903748,
|
| 1552 |
+
"learning_rate": 2.6880984952120387e-05,
|
| 1553 |
+
"loss": 0.0224,
|
| 1554 |
+
"step": 2190
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 1.5884476534296028,
|
| 1558 |
+
"grad_norm": 0.1974717378616333,
|
| 1559 |
+
"learning_rate": 2.674418604651163e-05,
|
| 1560 |
+
"loss": 0.0248,
|
| 1561 |
+
"step": 2200
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.595667870036101,
|
| 1565 |
+
"grad_norm": 0.23695822060108185,
|
| 1566 |
+
"learning_rate": 2.6607387140902874e-05,
|
| 1567 |
+
"loss": 0.0199,
|
| 1568 |
+
"step": 2210
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 1.6028880866425994,
|
| 1572 |
+
"grad_norm": 0.2396339625120163,
|
| 1573 |
+
"learning_rate": 2.647058823529412e-05,
|
| 1574 |
+
"loss": 0.0224,
|
| 1575 |
+
"step": 2220
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 1.6101083032490975,
|
| 1579 |
+
"grad_norm": 0.20308814942836761,
|
| 1580 |
+
"learning_rate": 2.6333789329685364e-05,
|
| 1581 |
+
"loss": 0.0224,
|
| 1582 |
+
"step": 2230
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 1.6173285198555956,
|
| 1586 |
+
"grad_norm": 0.2676025927066803,
|
| 1587 |
+
"learning_rate": 2.6196990424076606e-05,
|
| 1588 |
+
"loss": 0.0192,
|
| 1589 |
+
"step": 2240
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 1.6245487364620939,
|
| 1593 |
+
"grad_norm": 0.1903916597366333,
|
| 1594 |
+
"learning_rate": 2.606019151846785e-05,
|
| 1595 |
+
"loss": 0.0234,
|
| 1596 |
+
"step": 2250
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 1.6317689530685922,
|
| 1600 |
+
"grad_norm": 0.194184347987175,
|
| 1601 |
+
"learning_rate": 2.59233926128591e-05,
|
| 1602 |
+
"loss": 0.021,
|
| 1603 |
+
"step": 2260
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.6389891696750902,
|
| 1607 |
+
"grad_norm": 0.22250430285930634,
|
| 1608 |
+
"learning_rate": 2.5786593707250345e-05,
|
| 1609 |
+
"loss": 0.0211,
|
| 1610 |
+
"step": 2270
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 1.6462093862815883,
|
| 1614 |
+
"grad_norm": 0.18064111471176147,
|
| 1615 |
+
"learning_rate": 2.5649794801641587e-05,
|
| 1616 |
+
"loss": 0.0176,
|
| 1617 |
+
"step": 2280
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 1.6534296028880866,
|
| 1621 |
+
"grad_norm": 0.22021658718585968,
|
| 1622 |
+
"learning_rate": 2.5512995896032832e-05,
|
| 1623 |
+
"loss": 0.0205,
|
| 1624 |
+
"step": 2290
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 1.660649819494585,
|
| 1628 |
+
"grad_norm": 0.18112881481647491,
|
| 1629 |
+
"learning_rate": 2.5376196990424077e-05,
|
| 1630 |
+
"loss": 0.0198,
|
| 1631 |
+
"step": 2300
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.6678700361010832,
|
| 1635 |
+
"grad_norm": 0.25381046533584595,
|
| 1636 |
+
"learning_rate": 2.5239398084815325e-05,
|
| 1637 |
+
"loss": 0.0209,
|
| 1638 |
+
"step": 2310
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 1.6750902527075813,
|
| 1642 |
+
"grad_norm": 0.19742076098918915,
|
| 1643 |
+
"learning_rate": 2.5102599179206564e-05,
|
| 1644 |
+
"loss": 0.0214,
|
| 1645 |
+
"step": 2320
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.6823104693140793,
|
| 1649 |
+
"grad_norm": 0.20728257298469543,
|
| 1650 |
+
"learning_rate": 2.4965800273597812e-05,
|
| 1651 |
+
"loss": 0.0218,
|
| 1652 |
+
"step": 2330
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 1.6895306859205776,
|
| 1656 |
+
"grad_norm": 0.14151689410209656,
|
| 1657 |
+
"learning_rate": 2.4829001367989058e-05,
|
| 1658 |
+
"loss": 0.0211,
|
| 1659 |
+
"step": 2340
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 1.696750902527076,
|
| 1663 |
+
"grad_norm": 0.2515600621700287,
|
| 1664 |
+
"learning_rate": 2.46922024623803e-05,
|
| 1665 |
+
"loss": 0.0215,
|
| 1666 |
+
"step": 2350
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 1.703971119133574,
|
| 1670 |
+
"grad_norm": 0.21007981896400452,
|
| 1671 |
+
"learning_rate": 2.4555403556771548e-05,
|
| 1672 |
+
"loss": 0.0207,
|
| 1673 |
+
"step": 2360
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.711191335740072,
|
| 1677 |
+
"grad_norm": 0.20441214740276337,
|
| 1678 |
+
"learning_rate": 2.441860465116279e-05,
|
| 1679 |
+
"loss": 0.0204,
|
| 1680 |
+
"step": 2370
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 1.7184115523465704,
|
| 1684 |
+
"grad_norm": 0.21464481949806213,
|
| 1685 |
+
"learning_rate": 2.428180574555404e-05,
|
| 1686 |
+
"loss": 0.0187,
|
| 1687 |
+
"step": 2380
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.7256317689530687,
|
| 1691 |
+
"grad_norm": 0.18914371728897095,
|
| 1692 |
+
"learning_rate": 2.414500683994528e-05,
|
| 1693 |
+
"loss": 0.0214,
|
| 1694 |
+
"step": 2390
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 1.7328519855595668,
|
| 1698 |
+
"grad_norm": 0.16250531375408173,
|
| 1699 |
+
"learning_rate": 2.4008207934336525e-05,
|
| 1700 |
+
"loss": 0.0229,
|
| 1701 |
+
"step": 2400
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 1.7400722021660648,
|
| 1705 |
+
"grad_norm": 0.14336325228214264,
|
| 1706 |
+
"learning_rate": 2.387140902872777e-05,
|
| 1707 |
+
"loss": 0.0211,
|
| 1708 |
+
"step": 2410
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 1.7472924187725631,
|
| 1712 |
+
"grad_norm": 0.2290533185005188,
|
| 1713 |
+
"learning_rate": 2.3734610123119016e-05,
|
| 1714 |
+
"loss": 0.0216,
|
| 1715 |
+
"step": 2420
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.7545126353790614,
|
| 1719 |
+
"grad_norm": 0.21229666471481323,
|
| 1720 |
+
"learning_rate": 2.359781121751026e-05,
|
| 1721 |
+
"loss": 0.023,
|
| 1722 |
+
"step": 2430
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.7617328519855595,
|
| 1726 |
+
"grad_norm": 0.17029692232608795,
|
| 1727 |
+
"learning_rate": 2.3461012311901506e-05,
|
| 1728 |
+
"loss": 0.0226,
|
| 1729 |
+
"step": 2440
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 1.7689530685920578,
|
| 1733 |
+
"grad_norm": 0.1786644607782364,
|
| 1734 |
+
"learning_rate": 2.3324213406292748e-05,
|
| 1735 |
+
"loss": 0.0201,
|
| 1736 |
+
"step": 2450
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 1.7761732851985559,
|
| 1740 |
+
"grad_norm": 0.1429545283317566,
|
| 1741 |
+
"learning_rate": 2.3187414500683997e-05,
|
| 1742 |
+
"loss": 0.0201,
|
| 1743 |
+
"step": 2460
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 1.7833935018050542,
|
| 1747 |
+
"grad_norm": 0.18863752484321594,
|
| 1748 |
+
"learning_rate": 2.305061559507524e-05,
|
| 1749 |
+
"loss": 0.0205,
|
| 1750 |
+
"step": 2470
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 1.7906137184115525,
|
| 1754 |
+
"grad_norm": 0.16493657231330872,
|
| 1755 |
+
"learning_rate": 2.2913816689466487e-05,
|
| 1756 |
+
"loss": 0.0214,
|
| 1757 |
+
"step": 2480
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.7978339350180506,
|
| 1761 |
+
"grad_norm": 0.23174484074115753,
|
| 1762 |
+
"learning_rate": 2.277701778385773e-05,
|
| 1763 |
+
"loss": 0.0178,
|
| 1764 |
+
"step": 2490
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 1.8050541516245486,
|
| 1768 |
+
"grad_norm": 0.2224528193473816,
|
| 1769 |
+
"learning_rate": 2.2640218878248974e-05,
|
| 1770 |
+
"loss": 0.0217,
|
| 1771 |
+
"step": 2500
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 1.812274368231047,
|
| 1775 |
+
"grad_norm": 0.19585539400577545,
|
| 1776 |
+
"learning_rate": 2.250341997264022e-05,
|
| 1777 |
+
"loss": 0.0218,
|
| 1778 |
+
"step": 2510
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 1.8194945848375452,
|
| 1782 |
+
"grad_norm": 0.18590857088565826,
|
| 1783 |
+
"learning_rate": 2.2366621067031464e-05,
|
| 1784 |
+
"loss": 0.0211,
|
| 1785 |
+
"step": 2520
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 1.8267148014440433,
|
| 1789 |
+
"grad_norm": 0.17392487823963165,
|
| 1790 |
+
"learning_rate": 2.222982216142271e-05,
|
| 1791 |
+
"loss": 0.0208,
|
| 1792 |
+
"step": 2530
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 1.8339350180505414,
|
| 1796 |
+
"grad_norm": 0.33551010489463806,
|
| 1797 |
+
"learning_rate": 2.2093023255813955e-05,
|
| 1798 |
+
"loss": 0.0213,
|
| 1799 |
+
"step": 2540
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 1.8411552346570397,
|
| 1803 |
+
"grad_norm": 0.16854000091552734,
|
| 1804 |
+
"learning_rate": 2.19562243502052e-05,
|
| 1805 |
+
"loss": 0.0189,
|
| 1806 |
+
"step": 2550
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 1.848375451263538,
|
| 1810 |
+
"grad_norm": 0.20123836398124695,
|
| 1811 |
+
"learning_rate": 2.1819425444596445e-05,
|
| 1812 |
+
"loss": 0.0216,
|
| 1813 |
+
"step": 2560
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.855595667870036,
|
| 1817 |
+
"grad_norm": 0.1919688880443573,
|
| 1818 |
+
"learning_rate": 2.1682626538987687e-05,
|
| 1819 |
+
"loss": 0.0208,
|
| 1820 |
+
"step": 2570
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 1.8628158844765343,
|
| 1824 |
+
"grad_norm": 0.20486849546432495,
|
| 1825 |
+
"learning_rate": 2.1545827633378936e-05,
|
| 1826 |
+
"loss": 0.0204,
|
| 1827 |
+
"step": 2580
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 1.8700361010830324,
|
| 1831 |
+
"grad_norm": 0.15359218418598175,
|
| 1832 |
+
"learning_rate": 2.1409028727770177e-05,
|
| 1833 |
+
"loss": 0.0187,
|
| 1834 |
+
"step": 2590
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 1.8772563176895307,
|
| 1838 |
+
"grad_norm": 0.17525233328342438,
|
| 1839 |
+
"learning_rate": 2.1272229822161423e-05,
|
| 1840 |
+
"loss": 0.0256,
|
| 1841 |
+
"step": 2600
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 1.884476534296029,
|
| 1845 |
+
"grad_norm": 0.2669832408428192,
|
| 1846 |
+
"learning_rate": 2.1135430916552668e-05,
|
| 1847 |
+
"loss": 0.02,
|
| 1848 |
+
"step": 2610
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 1.891696750902527,
|
| 1852 |
+
"grad_norm": 0.2607351839542389,
|
| 1853 |
+
"learning_rate": 2.0998632010943913e-05,
|
| 1854 |
+
"loss": 0.021,
|
| 1855 |
+
"step": 2620
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 1.8989169675090252,
|
| 1859 |
+
"grad_norm": 0.1423048973083496,
|
| 1860 |
+
"learning_rate": 2.0861833105335158e-05,
|
| 1861 |
+
"loss": 0.0186,
|
| 1862 |
+
"step": 2630
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 1.9061371841155235,
|
| 1866 |
+
"grad_norm": 0.2874491810798645,
|
| 1867 |
+
"learning_rate": 2.0725034199726403e-05,
|
| 1868 |
+
"loss": 0.0225,
|
| 1869 |
+
"step": 2640
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 1.9133574007220218,
|
| 1873 |
+
"grad_norm": 0.19905777275562286,
|
| 1874 |
+
"learning_rate": 2.058823529411765e-05,
|
| 1875 |
+
"loss": 0.0205,
|
| 1876 |
+
"step": 2650
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 1.9205776173285198,
|
| 1880 |
+
"grad_norm": 0.18077979981899261,
|
| 1881 |
+
"learning_rate": 2.0451436388508894e-05,
|
| 1882 |
+
"loss": 0.0172,
|
| 1883 |
+
"step": 2660
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 1.927797833935018,
|
| 1887 |
+
"grad_norm": 0.21182967722415924,
|
| 1888 |
+
"learning_rate": 2.0314637482900136e-05,
|
| 1889 |
+
"loss": 0.0208,
|
| 1890 |
+
"step": 2670
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 1.9350180505415162,
|
| 1894 |
+
"grad_norm": 0.23693156242370605,
|
| 1895 |
+
"learning_rate": 2.0177838577291384e-05,
|
| 1896 |
+
"loss": 0.0208,
|
| 1897 |
+
"step": 2680
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 1.9422382671480145,
|
| 1901 |
+
"grad_norm": 0.1655312329530716,
|
| 1902 |
+
"learning_rate": 2.0041039671682626e-05,
|
| 1903 |
+
"loss": 0.0194,
|
| 1904 |
+
"step": 2690
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 1.9494584837545126,
|
| 1908 |
+
"grad_norm": 0.18168915808200836,
|
| 1909 |
+
"learning_rate": 1.990424076607387e-05,
|
| 1910 |
+
"loss": 0.0202,
|
| 1911 |
+
"step": 2700
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 1.9566787003610109,
|
| 1915 |
+
"grad_norm": 0.16807179152965546,
|
| 1916 |
+
"learning_rate": 1.9767441860465116e-05,
|
| 1917 |
+
"loss": 0.0225,
|
| 1918 |
+
"step": 2710
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 1.963898916967509,
|
| 1922 |
+
"grad_norm": 0.15267817676067352,
|
| 1923 |
+
"learning_rate": 1.963064295485636e-05,
|
| 1924 |
+
"loss": 0.0189,
|
| 1925 |
+
"step": 2720
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 1.9711191335740073,
|
| 1929 |
+
"grad_norm": 0.35743266344070435,
|
| 1930 |
+
"learning_rate": 1.9493844049247607e-05,
|
| 1931 |
+
"loss": 0.0172,
|
| 1932 |
+
"step": 2730
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 1.9783393501805056,
|
| 1936 |
+
"grad_norm": 0.14603158831596375,
|
| 1937 |
+
"learning_rate": 1.9357045143638852e-05,
|
| 1938 |
+
"loss": 0.0175,
|
| 1939 |
+
"step": 2740
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 1.9855595667870036,
|
| 1943 |
+
"grad_norm": 0.1915268450975418,
|
| 1944 |
+
"learning_rate": 1.9220246238030097e-05,
|
| 1945 |
+
"loss": 0.0202,
|
| 1946 |
+
"step": 2750
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 1.9927797833935017,
|
| 1950 |
+
"grad_norm": 0.2167324721813202,
|
| 1951 |
+
"learning_rate": 1.9083447332421342e-05,
|
| 1952 |
+
"loss": 0.0191,
|
| 1953 |
+
"step": 2760
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 2.0,
|
| 1957 |
+
"grad_norm": 0.22105805575847626,
|
| 1958 |
+
"learning_rate": 1.8946648426812584e-05,
|
| 1959 |
+
"loss": 0.0206,
|
| 1960 |
+
"step": 2770
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 2.0,
|
| 1964 |
+
"eval_accuracy": 0.9911191241647089,
|
| 1965 |
+
"eval_f1": 0.8963686600905629,
|
| 1966 |
+
"eval_loss": 0.024568406865000725,
|
| 1967 |
+
"eval_precision": 0.8729883765521071,
|
| 1968 |
+
"eval_recall": 0.9210357431847986,
|
| 1969 |
+
"eval_runtime": 271.5143,
|
| 1970 |
+
"eval_samples_per_second": 81.432,
|
| 1971 |
+
"eval_steps_per_second": 1.274,
|
| 1972 |
+
"step": 2770
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 2.0072202166064983,
|
| 1976 |
+
"grad_norm": 0.15133358538150787,
|
| 1977 |
+
"learning_rate": 1.8809849521203833e-05,
|
| 1978 |
+
"loss": 0.0135,
|
| 1979 |
+
"step": 2780
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 2.0144404332129966,
|
| 1983 |
+
"grad_norm": 0.23863369226455688,
|
| 1984 |
+
"learning_rate": 1.8673050615595075e-05,
|
| 1985 |
+
"loss": 0.0153,
|
| 1986 |
+
"step": 2790
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 2.0216606498194944,
|
| 1990 |
+
"grad_norm": 0.16403010487556458,
|
| 1991 |
+
"learning_rate": 1.853625170998632e-05,
|
| 1992 |
+
"loss": 0.0161,
|
| 1993 |
+
"step": 2800
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 2.0288808664259927,
|
| 1997 |
+
"grad_norm": 0.19030123949050903,
|
| 1998 |
+
"learning_rate": 1.8399452804377565e-05,
|
| 1999 |
+
"loss": 0.0137,
|
| 2000 |
+
"step": 2810
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 2.036101083032491,
|
| 2004 |
+
"grad_norm": 0.17739665508270264,
|
| 2005 |
+
"learning_rate": 1.826265389876881e-05,
|
| 2006 |
+
"loss": 0.0149,
|
| 2007 |
+
"step": 2820
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 2.0433212996389893,
|
| 2011 |
+
"grad_norm": 0.18450963497161865,
|
| 2012 |
+
"learning_rate": 1.8125854993160055e-05,
|
| 2013 |
+
"loss": 0.0151,
|
| 2014 |
+
"step": 2830
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 2.050541516245487,
|
| 2018 |
+
"grad_norm": 0.1654953956604004,
|
| 2019 |
+
"learning_rate": 1.79890560875513e-05,
|
| 2020 |
+
"loss": 0.0152,
|
| 2021 |
+
"step": 2840
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 2.0577617328519855,
|
| 2025 |
+
"grad_norm": 0.18726739287376404,
|
| 2026 |
+
"learning_rate": 1.7852257181942546e-05,
|
| 2027 |
+
"loss": 0.0136,
|
| 2028 |
+
"step": 2850
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 2.064981949458484,
|
| 2032 |
+
"grad_norm": 0.15999707579612732,
|
| 2033 |
+
"learning_rate": 1.771545827633379e-05,
|
| 2034 |
+
"loss": 0.0167,
|
| 2035 |
+
"step": 2860
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 2.072202166064982,
|
| 2039 |
+
"grad_norm": 0.20449534058570862,
|
| 2040 |
+
"learning_rate": 1.7578659370725033e-05,
|
| 2041 |
+
"loss": 0.0138,
|
| 2042 |
+
"step": 2870
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 2.07942238267148,
|
| 2046 |
+
"grad_norm": 0.18105897307395935,
|
| 2047 |
+
"learning_rate": 1.744186046511628e-05,
|
| 2048 |
+
"loss": 0.0137,
|
| 2049 |
+
"step": 2880
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 2.0866425992779782,
|
| 2053 |
+
"grad_norm": 0.2451828122138977,
|
| 2054 |
+
"learning_rate": 1.7305061559507523e-05,
|
| 2055 |
+
"loss": 0.0183,
|
| 2056 |
+
"step": 2890
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 2.0938628158844765,
|
| 2060 |
+
"grad_norm": 0.24325129389762878,
|
| 2061 |
+
"learning_rate": 1.716826265389877e-05,
|
| 2062 |
+
"loss": 0.0142,
|
| 2063 |
+
"step": 2900
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 2.101083032490975,
|
| 2067 |
+
"grad_norm": 0.21412310004234314,
|
| 2068 |
+
"learning_rate": 1.7031463748290014e-05,
|
| 2069 |
+
"loss": 0.0158,
|
| 2070 |
+
"step": 2910
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 2.108303249097473,
|
| 2074 |
+
"grad_norm": 0.2084820568561554,
|
| 2075 |
+
"learning_rate": 1.689466484268126e-05,
|
| 2076 |
+
"loss": 0.0146,
|
| 2077 |
+
"step": 2920
|
| 2078 |
+
},
|
| 2079 |
+
{
|
| 2080 |
+
"epoch": 2.115523465703971,
|
| 2081 |
+
"grad_norm": 0.1751072257757187,
|
| 2082 |
+
"learning_rate": 1.6757865937072504e-05,
|
| 2083 |
+
"loss": 0.0153,
|
| 2084 |
+
"step": 2930
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 2.1227436823104693,
|
| 2088 |
+
"grad_norm": 0.1661452054977417,
|
| 2089 |
+
"learning_rate": 1.662106703146375e-05,
|
| 2090 |
+
"loss": 0.0142,
|
| 2091 |
+
"step": 2940
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 2.1299638989169676,
|
| 2095 |
+
"grad_norm": 0.15145418047904968,
|
| 2096 |
+
"learning_rate": 1.6484268125854994e-05,
|
| 2097 |
+
"loss": 0.0163,
|
| 2098 |
+
"step": 2950
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 2.137184115523466,
|
| 2102 |
+
"grad_norm": 0.22948749363422394,
|
| 2103 |
+
"learning_rate": 1.634746922024624e-05,
|
| 2104 |
+
"loss": 0.0154,
|
| 2105 |
+
"step": 2960
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 2.1444043321299637,
|
| 2109 |
+
"grad_norm": 0.20568566024303436,
|
| 2110 |
+
"learning_rate": 1.621067031463748e-05,
|
| 2111 |
+
"loss": 0.0152,
|
| 2112 |
+
"step": 2970
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 2.151624548736462,
|
| 2116 |
+
"grad_norm": 0.1933920830488205,
|
| 2117 |
+
"learning_rate": 1.607387140902873e-05,
|
| 2118 |
+
"loss": 0.0163,
|
| 2119 |
+
"step": 2980
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 2.1588447653429603,
|
| 2123 |
+
"grad_norm": 0.25813716650009155,
|
| 2124 |
+
"learning_rate": 1.5937072503419972e-05,
|
| 2125 |
+
"loss": 0.0128,
|
| 2126 |
+
"step": 2990
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 2.1660649819494586,
|
| 2130 |
+
"grad_norm": 0.17636476457118988,
|
| 2131 |
+
"learning_rate": 1.580027359781122e-05,
|
| 2132 |
+
"loss": 0.0142,
|
| 2133 |
+
"step": 3000
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 2.1732851985559565,
|
| 2137 |
+
"grad_norm": 0.2212180197238922,
|
| 2138 |
+
"learning_rate": 1.5663474692202462e-05,
|
| 2139 |
+
"loss": 0.0133,
|
| 2140 |
+
"step": 3010
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 2.1805054151624548,
|
| 2144 |
+
"grad_norm": 0.23074433207511902,
|
| 2145 |
+
"learning_rate": 1.5526675786593707e-05,
|
| 2146 |
+
"loss": 0.0151,
|
| 2147 |
+
"step": 3020
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 2.187725631768953,
|
| 2151 |
+
"grad_norm": 0.2936890423297882,
|
| 2152 |
+
"learning_rate": 1.5389876880984953e-05,
|
| 2153 |
+
"loss": 0.0132,
|
| 2154 |
+
"step": 3030
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 2.1949458483754514,
|
| 2158 |
+
"grad_norm": 0.10483139753341675,
|
| 2159 |
+
"learning_rate": 1.5253077975376198e-05,
|
| 2160 |
+
"loss": 0.0146,
|
| 2161 |
+
"step": 3040
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 2.2021660649819497,
|
| 2165 |
+
"grad_norm": 0.13914550840854645,
|
| 2166 |
+
"learning_rate": 1.5116279069767441e-05,
|
| 2167 |
+
"loss": 0.0178,
|
| 2168 |
+
"step": 3050
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 2.2093862815884475,
|
| 2172 |
+
"grad_norm": 0.15463340282440186,
|
| 2173 |
+
"learning_rate": 1.4979480164158688e-05,
|
| 2174 |
+
"loss": 0.0126,
|
| 2175 |
+
"step": 3060
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 2.216606498194946,
|
| 2179 |
+
"grad_norm": 0.18321174383163452,
|
| 2180 |
+
"learning_rate": 1.4842681258549932e-05,
|
| 2181 |
+
"loss": 0.0144,
|
| 2182 |
+
"step": 3070
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 2.223826714801444,
|
| 2186 |
+
"grad_norm": 0.2047768235206604,
|
| 2187 |
+
"learning_rate": 1.4705882352941177e-05,
|
| 2188 |
+
"loss": 0.0149,
|
| 2189 |
+
"step": 3080
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.2310469314079424,
|
| 2193 |
+
"grad_norm": 0.15067626535892487,
|
| 2194 |
+
"learning_rate": 1.456908344733242e-05,
|
| 2195 |
+
"loss": 0.0134,
|
| 2196 |
+
"step": 3090
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 2.2382671480144403,
|
| 2200 |
+
"grad_norm": 0.1758282482624054,
|
| 2201 |
+
"learning_rate": 1.4432284541723667e-05,
|
| 2202 |
+
"loss": 0.0133,
|
| 2203 |
+
"step": 3100
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 2.2454873646209386,
|
| 2207 |
+
"grad_norm": 0.20649059116840363,
|
| 2208 |
+
"learning_rate": 1.429548563611491e-05,
|
| 2209 |
+
"loss": 0.0161,
|
| 2210 |
+
"step": 3110
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 2.252707581227437,
|
| 2214 |
+
"grad_norm": 0.17980459332466125,
|
| 2215 |
+
"learning_rate": 1.4158686730506158e-05,
|
| 2216 |
+
"loss": 0.0135,
|
| 2217 |
+
"step": 3120
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 2.259927797833935,
|
| 2221 |
+
"grad_norm": 0.22073961794376373,
|
| 2222 |
+
"learning_rate": 1.4021887824897401e-05,
|
| 2223 |
+
"loss": 0.0146,
|
| 2224 |
+
"step": 3130
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 2.2671480144404335,
|
| 2228 |
+
"grad_norm": 0.20626509189605713,
|
| 2229 |
+
"learning_rate": 1.3885088919288646e-05,
|
| 2230 |
+
"loss": 0.0138,
|
| 2231 |
+
"step": 3140
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.2743682310469313,
|
| 2235 |
+
"grad_norm": 0.16285601258277893,
|
| 2236 |
+
"learning_rate": 1.374829001367989e-05,
|
| 2237 |
+
"loss": 0.0136,
|
| 2238 |
+
"step": 3150
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 2.2815884476534296,
|
| 2242 |
+
"grad_norm": 0.17455926537513733,
|
| 2243 |
+
"learning_rate": 1.3611491108071137e-05,
|
| 2244 |
+
"loss": 0.0138,
|
| 2245 |
+
"step": 3160
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 2.288808664259928,
|
| 2249 |
+
"grad_norm": 0.18891878426074982,
|
| 2250 |
+
"learning_rate": 1.347469220246238e-05,
|
| 2251 |
+
"loss": 0.0151,
|
| 2252 |
+
"step": 3170
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 2.2960288808664258,
|
| 2256 |
+
"grad_norm": 0.13757237792015076,
|
| 2257 |
+
"learning_rate": 1.3337893296853627e-05,
|
| 2258 |
+
"loss": 0.017,
|
| 2259 |
+
"step": 3180
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 2.303249097472924,
|
| 2263 |
+
"grad_norm": 0.18987725675106049,
|
| 2264 |
+
"learning_rate": 1.3201094391244869e-05,
|
| 2265 |
+
"loss": 0.0147,
|
| 2266 |
+
"step": 3190
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 2.3104693140794224,
|
| 2270 |
+
"grad_norm": 0.13427282869815826,
|
| 2271 |
+
"learning_rate": 1.3064295485636116e-05,
|
| 2272 |
+
"loss": 0.0142,
|
| 2273 |
+
"step": 3200
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.3176895306859207,
|
| 2277 |
+
"grad_norm": 0.2868107557296753,
|
| 2278 |
+
"learning_rate": 1.292749658002736e-05,
|
| 2279 |
+
"loss": 0.0141,
|
| 2280 |
+
"step": 3210
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 2.324909747292419,
|
| 2284 |
+
"grad_norm": 0.19995689392089844,
|
| 2285 |
+
"learning_rate": 1.2790697674418606e-05,
|
| 2286 |
+
"loss": 0.0145,
|
| 2287 |
+
"step": 3220
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 2.332129963898917,
|
| 2291 |
+
"grad_norm": 0.15799254179000854,
|
| 2292 |
+
"learning_rate": 1.265389876880985e-05,
|
| 2293 |
+
"loss": 0.0149,
|
| 2294 |
+
"step": 3230
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 2.339350180505415,
|
| 2298 |
+
"grad_norm": 0.2354062795639038,
|
| 2299 |
+
"learning_rate": 1.2517099863201095e-05,
|
| 2300 |
+
"loss": 0.0173,
|
| 2301 |
+
"step": 3240
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 2.3465703971119134,
|
| 2305 |
+
"grad_norm": 0.2461709827184677,
|
| 2306 |
+
"learning_rate": 1.238030095759234e-05,
|
| 2307 |
+
"loss": 0.0165,
|
| 2308 |
+
"step": 3250
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 2.3537906137184117,
|
| 2312 |
+
"grad_norm": 0.17912667989730835,
|
| 2313 |
+
"learning_rate": 1.2243502051983585e-05,
|
| 2314 |
+
"loss": 0.0168,
|
| 2315 |
+
"step": 3260
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.3610108303249095,
|
| 2319 |
+
"grad_norm": 0.2090194672346115,
|
| 2320 |
+
"learning_rate": 1.210670314637483e-05,
|
| 2321 |
+
"loss": 0.0133,
|
| 2322 |
+
"step": 3270
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 2.368231046931408,
|
| 2326 |
+
"grad_norm": 0.16298013925552368,
|
| 2327 |
+
"learning_rate": 1.1969904240766076e-05,
|
| 2328 |
+
"loss": 0.0164,
|
| 2329 |
+
"step": 3280
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 2.375451263537906,
|
| 2333 |
+
"grad_norm": 0.14699536561965942,
|
| 2334 |
+
"learning_rate": 1.183310533515732e-05,
|
| 2335 |
+
"loss": 0.0165,
|
| 2336 |
+
"step": 3290
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 2.3826714801444044,
|
| 2340 |
+
"grad_norm": 0.1909872591495514,
|
| 2341 |
+
"learning_rate": 1.1696306429548564e-05,
|
| 2342 |
+
"loss": 0.0146,
|
| 2343 |
+
"step": 3300
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 2.3898916967509027,
|
| 2347 |
+
"grad_norm": 0.149457648396492,
|
| 2348 |
+
"learning_rate": 1.155950752393981e-05,
|
| 2349 |
+
"loss": 0.0154,
|
| 2350 |
+
"step": 3310
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 2.3971119133574006,
|
| 2354 |
+
"grad_norm": 0.18534165620803833,
|
| 2355 |
+
"learning_rate": 1.1422708618331055e-05,
|
| 2356 |
+
"loss": 0.0147,
|
| 2357 |
+
"step": 3320
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.404332129963899,
|
| 2361 |
+
"grad_norm": 0.17388072609901428,
|
| 2362 |
+
"learning_rate": 1.12859097127223e-05,
|
| 2363 |
+
"loss": 0.0132,
|
| 2364 |
+
"step": 3330
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 2.411552346570397,
|
| 2368 |
+
"grad_norm": 0.2680976390838623,
|
| 2369 |
+
"learning_rate": 1.1149110807113544e-05,
|
| 2370 |
+
"loss": 0.0148,
|
| 2371 |
+
"step": 3340
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 2.4187725631768955,
|
| 2375 |
+
"grad_norm": 0.2023518681526184,
|
| 2376 |
+
"learning_rate": 1.1012311901504789e-05,
|
| 2377 |
+
"loss": 0.0133,
|
| 2378 |
+
"step": 3350
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 2.4259927797833933,
|
| 2382 |
+
"grad_norm": 0.19814668595790863,
|
| 2383 |
+
"learning_rate": 1.0875512995896034e-05,
|
| 2384 |
+
"loss": 0.0139,
|
| 2385 |
+
"step": 3360
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 2.4332129963898916,
|
| 2389 |
+
"grad_norm": 0.1586819589138031,
|
| 2390 |
+
"learning_rate": 1.0738714090287279e-05,
|
| 2391 |
+
"loss": 0.0139,
|
| 2392 |
+
"step": 3370
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 2.44043321299639,
|
| 2396 |
+
"grad_norm": 0.12945076823234558,
|
| 2397 |
+
"learning_rate": 1.0601915184678524e-05,
|
| 2398 |
+
"loss": 0.0136,
|
| 2399 |
+
"step": 3380
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.4476534296028882,
|
| 2403 |
+
"grad_norm": 0.22156469523906708,
|
| 2404 |
+
"learning_rate": 1.0465116279069768e-05,
|
| 2405 |
+
"loss": 0.017,
|
| 2406 |
+
"step": 3390
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 2.4548736462093865,
|
| 2410 |
+
"grad_norm": 0.24720166623592377,
|
| 2411 |
+
"learning_rate": 1.0328317373461013e-05,
|
| 2412 |
+
"loss": 0.0131,
|
| 2413 |
+
"step": 3400
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 2.4620938628158844,
|
| 2417 |
+
"grad_norm": 0.15608453750610352,
|
| 2418 |
+
"learning_rate": 1.0191518467852258e-05,
|
| 2419 |
+
"loss": 0.0155,
|
| 2420 |
+
"step": 3410
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 2.4693140794223827,
|
| 2424 |
+
"grad_norm": 0.34953010082244873,
|
| 2425 |
+
"learning_rate": 1.0054719562243503e-05,
|
| 2426 |
+
"loss": 0.017,
|
| 2427 |
+
"step": 3420
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 2.476534296028881,
|
| 2431 |
+
"grad_norm": 0.20051924884319305,
|
| 2432 |
+
"learning_rate": 9.917920656634749e-06,
|
| 2433 |
+
"loss": 0.0136,
|
| 2434 |
+
"step": 3430
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 2.483754512635379,
|
| 2438 |
+
"grad_norm": 0.20402652025222778,
|
| 2439 |
+
"learning_rate": 9.781121751025992e-06,
|
| 2440 |
+
"loss": 0.0133,
|
| 2441 |
+
"step": 3440
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.490974729241877,
|
| 2445 |
+
"grad_norm": 0.1991117298603058,
|
| 2446 |
+
"learning_rate": 9.644322845417237e-06,
|
| 2447 |
+
"loss": 0.0153,
|
| 2448 |
+
"step": 3450
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 2.4981949458483754,
|
| 2452 |
+
"grad_norm": 0.1154688373208046,
|
| 2453 |
+
"learning_rate": 9.507523939808483e-06,
|
| 2454 |
+
"loss": 0.0126,
|
| 2455 |
+
"step": 3460
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 2.5054151624548737,
|
| 2459 |
+
"grad_norm": 0.1297905594110489,
|
| 2460 |
+
"learning_rate": 9.370725034199728e-06,
|
| 2461 |
+
"loss": 0.0146,
|
| 2462 |
+
"step": 3470
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 2.512635379061372,
|
| 2466 |
+
"grad_norm": 0.23860248923301697,
|
| 2467 |
+
"learning_rate": 9.233926128590973e-06,
|
| 2468 |
+
"loss": 0.0159,
|
| 2469 |
+
"step": 3480
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 2.51985559566787,
|
| 2473 |
+
"grad_norm": 0.20767924189567566,
|
| 2474 |
+
"learning_rate": 9.097127222982216e-06,
|
| 2475 |
+
"loss": 0.0133,
|
| 2476 |
+
"step": 3490
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 2.527075812274368,
|
| 2480 |
+
"grad_norm": 0.22992978990077972,
|
| 2481 |
+
"learning_rate": 8.960328317373462e-06,
|
| 2482 |
+
"loss": 0.016,
|
| 2483 |
+
"step": 3500
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 2.5342960288808665,
|
| 2487 |
+
"grad_norm": 0.13888780772686005,
|
| 2488 |
+
"learning_rate": 8.823529411764707e-06,
|
| 2489 |
+
"loss": 0.0129,
|
| 2490 |
+
"step": 3510
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 2.5415162454873648,
|
| 2494 |
+
"grad_norm": 0.17588818073272705,
|
| 2495 |
+
"learning_rate": 8.686730506155952e-06,
|
| 2496 |
+
"loss": 0.0155,
|
| 2497 |
+
"step": 3520
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 2.5487364620938626,
|
| 2501 |
+
"grad_norm": 0.17710047960281372,
|
| 2502 |
+
"learning_rate": 8.549931600547197e-06,
|
| 2503 |
+
"loss": 0.0147,
|
| 2504 |
+
"step": 3530
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 2.555956678700361,
|
| 2508 |
+
"grad_norm": 0.2465488165616989,
|
| 2509 |
+
"learning_rate": 8.41313269493844e-06,
|
| 2510 |
+
"loss": 0.0159,
|
| 2511 |
+
"step": 3540
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 2.563176895306859,
|
| 2515 |
+
"grad_norm": 0.211311474442482,
|
| 2516 |
+
"learning_rate": 8.276333789329686e-06,
|
| 2517 |
+
"loss": 0.0142,
|
| 2518 |
+
"step": 3550
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 2.5703971119133575,
|
| 2522 |
+
"grad_norm": 0.19003704190254211,
|
| 2523 |
+
"learning_rate": 8.139534883720931e-06,
|
| 2524 |
+
"loss": 0.0156,
|
| 2525 |
+
"step": 3560
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 2.577617328519856,
|
| 2529 |
+
"grad_norm": 0.20325423777103424,
|
| 2530 |
+
"learning_rate": 8.002735978112176e-06,
|
| 2531 |
+
"loss": 0.0138,
|
| 2532 |
+
"step": 3570
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 2.5848375451263537,
|
| 2536 |
+
"grad_norm": 0.20516729354858398,
|
| 2537 |
+
"learning_rate": 7.865937072503421e-06,
|
| 2538 |
+
"loss": 0.0146,
|
| 2539 |
+
"step": 3580
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 2.592057761732852,
|
| 2543 |
+
"grad_norm": 0.15101560950279236,
|
| 2544 |
+
"learning_rate": 7.729138166894665e-06,
|
| 2545 |
+
"loss": 0.0144,
|
| 2546 |
+
"step": 3590
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 2.5992779783393503,
|
| 2550 |
+
"grad_norm": 0.20914088189601898,
|
| 2551 |
+
"learning_rate": 7.592339261285911e-06,
|
| 2552 |
+
"loss": 0.0147,
|
| 2553 |
+
"step": 3600
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 2.606498194945848,
|
| 2557 |
+
"grad_norm": 0.21813435852527618,
|
| 2558 |
+
"learning_rate": 7.455540355677155e-06,
|
| 2559 |
+
"loss": 0.0148,
|
| 2560 |
+
"step": 3610
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 2.6137184115523464,
|
| 2564 |
+
"grad_norm": 0.23971576988697052,
|
| 2565 |
+
"learning_rate": 7.318741450068401e-06,
|
| 2566 |
+
"loss": 0.0131,
|
| 2567 |
+
"step": 3620
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 2.6209386281588447,
|
| 2571 |
+
"grad_norm": 0.17804710566997528,
|
| 2572 |
+
"learning_rate": 7.181942544459645e-06,
|
| 2573 |
+
"loss": 0.0136,
|
| 2574 |
+
"step": 3630
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 2.628158844765343,
|
| 2578 |
+
"grad_norm": 0.15980184078216553,
|
| 2579 |
+
"learning_rate": 7.04514363885089e-06,
|
| 2580 |
+
"loss": 0.0118,
|
| 2581 |
+
"step": 3640
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 2.6353790613718413,
|
| 2585 |
+
"grad_norm": 0.23447448015213013,
|
| 2586 |
+
"learning_rate": 6.908344733242135e-06,
|
| 2587 |
+
"loss": 0.0119,
|
| 2588 |
+
"step": 3650
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 2.6425992779783396,
|
| 2592 |
+
"grad_norm": 0.2265952080488205,
|
| 2593 |
+
"learning_rate": 6.77154582763338e-06,
|
| 2594 |
+
"loss": 0.0135,
|
| 2595 |
+
"step": 3660
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 2.6498194945848375,
|
| 2599 |
+
"grad_norm": 0.27709242701530457,
|
| 2600 |
+
"learning_rate": 6.634746922024625e-06,
|
| 2601 |
+
"loss": 0.0139,
|
| 2602 |
+
"step": 3670
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 2.6570397111913358,
|
| 2606 |
+
"grad_norm": 0.189510315656662,
|
| 2607 |
+
"learning_rate": 6.497948016415869e-06,
|
| 2608 |
+
"loss": 0.0128,
|
| 2609 |
+
"step": 3680
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 2.664259927797834,
|
| 2613 |
+
"grad_norm": 0.09412135928869247,
|
| 2614 |
+
"learning_rate": 6.3611491108071144e-06,
|
| 2615 |
+
"loss": 0.0155,
|
| 2616 |
+
"step": 3690
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 2.671480144404332,
|
| 2620 |
+
"grad_norm": 0.21313029527664185,
|
| 2621 |
+
"learning_rate": 6.224350205198359e-06,
|
| 2622 |
+
"loss": 0.0144,
|
| 2623 |
+
"step": 3700
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 2.67870036101083,
|
| 2627 |
+
"grad_norm": 0.35182449221611023,
|
| 2628 |
+
"learning_rate": 6.087551299589603e-06,
|
| 2629 |
+
"loss": 0.0158,
|
| 2630 |
+
"step": 3710
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 2.6859205776173285,
|
| 2634 |
+
"grad_norm": 0.20150898396968842,
|
| 2635 |
+
"learning_rate": 5.950752393980848e-06,
|
| 2636 |
+
"loss": 0.0139,
|
| 2637 |
+
"step": 3720
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 2.693140794223827,
|
| 2641 |
+
"grad_norm": 0.1621181070804596,
|
| 2642 |
+
"learning_rate": 5.8139534883720935e-06,
|
| 2643 |
+
"loss": 0.0131,
|
| 2644 |
+
"step": 3730
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 2.700361010830325,
|
| 2648 |
+
"grad_norm": 0.20502717792987823,
|
| 2649 |
+
"learning_rate": 5.677154582763338e-06,
|
| 2650 |
+
"loss": 0.0132,
|
| 2651 |
+
"step": 3740
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 2.707581227436823,
|
| 2655 |
+
"grad_norm": 0.14961372315883636,
|
| 2656 |
+
"learning_rate": 5.540355677154583e-06,
|
| 2657 |
+
"loss": 0.013,
|
| 2658 |
+
"step": 3750
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 2.7148014440433212,
|
| 2662 |
+
"grad_norm": 0.2430783361196518,
|
| 2663 |
+
"learning_rate": 5.403556771545827e-06,
|
| 2664 |
+
"loss": 0.014,
|
| 2665 |
+
"step": 3760
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 2.7220216606498195,
|
| 2669 |
+
"grad_norm": 0.2312983125448227,
|
| 2670 |
+
"learning_rate": 5.266757865937073e-06,
|
| 2671 |
+
"loss": 0.0151,
|
| 2672 |
+
"step": 3770
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 2.729241877256318,
|
| 2676 |
+
"grad_norm": 0.23572318255901337,
|
| 2677 |
+
"learning_rate": 5.129958960328318e-06,
|
| 2678 |
+
"loss": 0.0146,
|
| 2679 |
+
"step": 3780
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 2.7364620938628157,
|
| 2683 |
+
"grad_norm": 0.16730424761772156,
|
| 2684 |
+
"learning_rate": 4.993160054719562e-06,
|
| 2685 |
+
"loss": 0.014,
|
| 2686 |
+
"step": 3790
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 2.743682310469314,
|
| 2690 |
+
"grad_norm": 0.26095303893089294,
|
| 2691 |
+
"learning_rate": 4.856361149110807e-06,
|
| 2692 |
+
"loss": 0.0153,
|
| 2693 |
+
"step": 3800
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 2.7509025270758123,
|
| 2697 |
+
"grad_norm": 0.18118229508399963,
|
| 2698 |
+
"learning_rate": 4.7195622435020526e-06,
|
| 2699 |
+
"loss": 0.0148,
|
| 2700 |
+
"step": 3810
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 2.7581227436823106,
|
| 2704 |
+
"grad_norm": 0.19273220002651215,
|
| 2705 |
+
"learning_rate": 4.582763337893297e-06,
|
| 2706 |
+
"loss": 0.0132,
|
| 2707 |
+
"step": 3820
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 2.765342960288809,
|
| 2711 |
+
"grad_norm": 0.17820940911769867,
|
| 2712 |
+
"learning_rate": 4.445964432284542e-06,
|
| 2713 |
+
"loss": 0.0129,
|
| 2714 |
+
"step": 3830
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 2.7725631768953067,
|
| 2718 |
+
"grad_norm": 0.255801260471344,
|
| 2719 |
+
"learning_rate": 4.3091655266757865e-06,
|
| 2720 |
+
"loss": 0.0138,
|
| 2721 |
+
"step": 3840
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 2.779783393501805,
|
| 2725 |
+
"grad_norm": 0.2841118276119232,
|
| 2726 |
+
"learning_rate": 4.172366621067032e-06,
|
| 2727 |
+
"loss": 0.0165,
|
| 2728 |
+
"step": 3850
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 2.7870036101083033,
|
| 2732 |
+
"grad_norm": 0.19371943175792694,
|
| 2733 |
+
"learning_rate": 4.035567715458277e-06,
|
| 2734 |
+
"loss": 0.0121,
|
| 2735 |
+
"step": 3860
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 2.794223826714801,
|
| 2739 |
+
"grad_norm": 0.2262773960828781,
|
| 2740 |
+
"learning_rate": 3.898768809849521e-06,
|
| 2741 |
+
"loss": 0.0136,
|
| 2742 |
+
"step": 3870
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 2.8014440433212995,
|
| 2746 |
+
"grad_norm": 0.17181913554668427,
|
| 2747 |
+
"learning_rate": 3.7619699042407664e-06,
|
| 2748 |
+
"loss": 0.0133,
|
| 2749 |
+
"step": 3880
|
| 2750 |
+
},
|
| 2751 |
+
{
|
| 2752 |
+
"epoch": 2.808664259927798,
|
| 2753 |
+
"grad_norm": 0.1269243359565735,
|
| 2754 |
+
"learning_rate": 3.625170998632011e-06,
|
| 2755 |
+
"loss": 0.0129,
|
| 2756 |
+
"step": 3890
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 2.815884476534296,
|
| 2760 |
+
"grad_norm": 0.24962696433067322,
|
| 2761 |
+
"learning_rate": 3.488372093023256e-06,
|
| 2762 |
+
"loss": 0.0156,
|
| 2763 |
+
"step": 3900
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 2.8231046931407944,
|
| 2767 |
+
"grad_norm": 0.21303677558898926,
|
| 2768 |
+
"learning_rate": 3.3515731874145007e-06,
|
| 2769 |
+
"loss": 0.016,
|
| 2770 |
+
"step": 3910
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 2.8303249097472927,
|
| 2774 |
+
"grad_norm": 0.2242385894060135,
|
| 2775 |
+
"learning_rate": 3.2147742818057455e-06,
|
| 2776 |
+
"loss": 0.0147,
|
| 2777 |
+
"step": 3920
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 2.8375451263537905,
|
| 2781 |
+
"grad_norm": 0.1730586290359497,
|
| 2782 |
+
"learning_rate": 3.0779753761969907e-06,
|
| 2783 |
+
"loss": 0.0125,
|
| 2784 |
+
"step": 3930
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 2.844765342960289,
|
| 2788 |
+
"grad_norm": 0.17158852517604828,
|
| 2789 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 2790 |
+
"loss": 0.0141,
|
| 2791 |
+
"step": 3940
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"epoch": 2.851985559566787,
|
| 2795 |
+
"grad_norm": 0.21147002279758453,
|
| 2796 |
+
"learning_rate": 2.8043775649794802e-06,
|
| 2797 |
+
"loss": 0.0142,
|
| 2798 |
+
"step": 3950
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 2.859205776173285,
|
| 2802 |
+
"grad_norm": 0.22795933485031128,
|
| 2803 |
+
"learning_rate": 2.667578659370725e-06,
|
| 2804 |
+
"loss": 0.0133,
|
| 2805 |
+
"step": 3960
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 2.8664259927797833,
|
| 2809 |
+
"grad_norm": 0.19155286252498627,
|
| 2810 |
+
"learning_rate": 2.5307797537619698e-06,
|
| 2811 |
+
"loss": 0.014,
|
| 2812 |
+
"step": 3970
|
| 2813 |
+
},
|
| 2814 |
+
{
|
| 2815 |
+
"epoch": 2.8736462093862816,
|
| 2816 |
+
"grad_norm": 0.1995808482170105,
|
| 2817 |
+
"learning_rate": 2.393980848153215e-06,
|
| 2818 |
+
"loss": 0.0131,
|
| 2819 |
+
"step": 3980
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 2.88086642599278,
|
| 2823 |
+
"grad_norm": 0.20883041620254517,
|
| 2824 |
+
"learning_rate": 2.2571819425444598e-06,
|
| 2825 |
+
"loss": 0.0125,
|
| 2826 |
+
"step": 3990
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 2.888086642599278,
|
| 2830 |
+
"grad_norm": 0.22262053191661835,
|
| 2831 |
+
"learning_rate": 2.1203830369357045e-06,
|
| 2832 |
+
"loss": 0.0151,
|
| 2833 |
+
"step": 4000
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 2.895306859205776,
|
| 2837 |
+
"grad_norm": 0.17650048434734344,
|
| 2838 |
+
"learning_rate": 1.9835841313269493e-06,
|
| 2839 |
+
"loss": 0.0116,
|
| 2840 |
+
"step": 4010
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 2.9025270758122743,
|
| 2844 |
+
"grad_norm": 0.2514309883117676,
|
| 2845 |
+
"learning_rate": 1.8467852257181943e-06,
|
| 2846 |
+
"loss": 0.0149,
|
| 2847 |
+
"step": 4020
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 2.9097472924187726,
|
| 2851 |
+
"grad_norm": 0.13858270645141602,
|
| 2852 |
+
"learning_rate": 1.7099863201094393e-06,
|
| 2853 |
+
"loss": 0.0122,
|
| 2854 |
+
"step": 4030
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 2.916967509025271,
|
| 2858 |
+
"grad_norm": 0.20119492709636688,
|
| 2859 |
+
"learning_rate": 1.573187414500684e-06,
|
| 2860 |
+
"loss": 0.0148,
|
| 2861 |
+
"step": 4040
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 2.9241877256317688,
|
| 2865 |
+
"grad_norm": 0.2429434359073639,
|
| 2866 |
+
"learning_rate": 1.4363885088919288e-06,
|
| 2867 |
+
"loss": 0.0138,
|
| 2868 |
+
"step": 4050
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 2.931407942238267,
|
| 2872 |
+
"grad_norm": 0.158556267619133,
|
| 2873 |
+
"learning_rate": 1.2995896032831738e-06,
|
| 2874 |
+
"loss": 0.0131,
|
| 2875 |
+
"step": 4060
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 2.9386281588447654,
|
| 2879 |
+
"grad_norm": 0.15908589959144592,
|
| 2880 |
+
"learning_rate": 1.1627906976744186e-06,
|
| 2881 |
+
"loss": 0.0122,
|
| 2882 |
+
"step": 4070
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 2.9458483754512637,
|
| 2886 |
+
"grad_norm": 0.20639561116695404,
|
| 2887 |
+
"learning_rate": 1.0259917920656636e-06,
|
| 2888 |
+
"loss": 0.0145,
|
| 2889 |
+
"step": 4080
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 2.953068592057762,
|
| 2893 |
+
"grad_norm": 0.18711957335472107,
|
| 2894 |
+
"learning_rate": 8.891928864569083e-07,
|
| 2895 |
+
"loss": 0.0138,
|
| 2896 |
+
"step": 4090
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 2.96028880866426,
|
| 2900 |
+
"grad_norm": 0.2037838101387024,
|
| 2901 |
+
"learning_rate": 7.523939808481532e-07,
|
| 2902 |
+
"loss": 0.0133,
|
| 2903 |
+
"step": 4100
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 2.967509025270758,
|
| 2907 |
+
"grad_norm": 0.18255527317523956,
|
| 2908 |
+
"learning_rate": 6.155950752393981e-07,
|
| 2909 |
+
"loss": 0.0119,
|
| 2910 |
+
"step": 4110
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 2.9747292418772564,
|
| 2914 |
+
"grad_norm": 0.21082955598831177,
|
| 2915 |
+
"learning_rate": 4.78796169630643e-07,
|
| 2916 |
+
"loss": 0.0145,
|
| 2917 |
+
"step": 4120
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 2.9819494584837543,
|
| 2921 |
+
"grad_norm": 0.16639681160449982,
|
| 2922 |
+
"learning_rate": 3.4199726402188785e-07,
|
| 2923 |
+
"loss": 0.0116,
|
| 2924 |
+
"step": 4130
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 2.9891696750902526,
|
| 2928 |
+
"grad_norm": 0.19572538137435913,
|
| 2929 |
+
"learning_rate": 2.051983584131327e-07,
|
| 2930 |
+
"loss": 0.0137,
|
| 2931 |
+
"step": 4140
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 2.996389891696751,
|
| 2935 |
+
"grad_norm": 0.2220822423696518,
|
| 2936 |
+
"learning_rate": 6.839945280437757e-08,
|
| 2937 |
+
"loss": 0.0129,
|
| 2938 |
+
"step": 4150
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 3.0,
|
| 2942 |
+
"eval_accuracy": 0.9919225627188221,
|
| 2943 |
+
"eval_f1": 0.9047710499602184,
|
| 2944 |
+
"eval_loss": 0.024072397500276566,
|
| 2945 |
+
"eval_precision": 0.8898783021668151,
|
| 2946 |
+
"eval_recall": 0.9201707636931834,
|
| 2947 |
+
"eval_runtime": 271.2138,
|
| 2948 |
+
"eval_samples_per_second": 81.522,
|
| 2949 |
+
"eval_steps_per_second": 1.276,
|
| 2950 |
+
"step": 4155
|
| 2951 |
+
}
|
| 2952 |
+
],
|
| 2953 |
+
"logging_steps": 10,
|
| 2954 |
+
"max_steps": 4155,
|
| 2955 |
+
"num_input_tokens_seen": 0,
|
| 2956 |
+
"num_train_epochs": 3,
|
| 2957 |
+
"save_steps": 500,
|
| 2958 |
+
"stateful_callbacks": {
|
| 2959 |
+
"TrainerControl": {
|
| 2960 |
+
"args": {
|
| 2961 |
+
"should_epoch_stop": false,
|
| 2962 |
+
"should_evaluate": false,
|
| 2963 |
+
"should_log": false,
|
| 2964 |
+
"should_save": true,
|
| 2965 |
+
"should_training_stop": true
|
| 2966 |
+
},
|
| 2967 |
+
"attributes": {}
|
| 2968 |
+
}
|
| 2969 |
+
},
|
| 2970 |
+
"total_flos": 4.9767875544127224e+16,
|
| 2971 |
+
"train_batch_size": 64,
|
| 2972 |
+
"trial_name": null,
|
| 2973 |
+
"trial_params": null
|
| 2974 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d09b6ad6be4eda4e1f6036f4608d291c45d54830b18ad80c5a7a6617402792ec
|
| 3 |
+
size 5176
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|