Joeran Bosma
commited on
Commit
·
2739a45
1
Parent(s):
25a9f1f
Initial release
Browse files- README.md +106 -3
- all_results.json +15 -0
- config.json +27 -0
- eval_results.json +10 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +15 -0
- tokenizer.json +0 -0
- tokenizer_config.json +22 -0
- train_results.json +8 -0
- trainer_state.json +1327 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
CHANGED
|
@@ -1,3 +1,106 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-nc-sa-4.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-sa-4.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# DRAGON RoBERTa large domain-specific
|
| 6 |
+
|
| 7 |
+
Pretrained model on Dutch clinical reports using a masked language modeling (MLM) objective. It was introduced in [this](#pending) paper. The model was pretrained using domain-specific data (i.e., clinical reports) from scratch. The architecture is the same as [`xlm-roberta-large`](https://huggingface.co/xlm-roberta-large) from HuggingFace. The tokenizer was fitted to the dataset of Dutch medical reports, using the same settings for the tokenizer as [`roberta-base`](https://huggingface.co/FacebookAI/roberta-base).
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
## Model description
|
| 12 |
+
RoBERTa is a transformers model that was pretrained on a large corpus of Dutch clinical reports in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labeling them in any way with an automatic process to generate inputs and labels from those texts.
|
| 13 |
+
|
| 14 |
+
This way, the model learns an inner representation of the Dutch medical language that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled reports, for instance, you can train a standard classifier using the features produced by the BERT model as inputs.
|
| 15 |
+
|
| 16 |
+
## Model variations
|
| 17 |
+
Multiple architectures were pretrained for the DRAGON challenge.
|
| 18 |
+
|
| 19 |
+
| Model | #params | Language |
|
| 20 |
+
|------------------------|--------------------------------|-------|
|
| 21 |
+
| [`joeranbosma/dragon-bert-base-mixed-domain`](https://huggingface.co/joeranbosma/dragon-bert-base-mixed-domain) | 109M | Dutch → Dutch |
|
| 22 |
+
| [`joeranbosma/dragon-roberta-base-mixed-domain`](https://huggingface.co/joeranbosma/dragon-roberta-base-mixed-domain) | 278M | Multiple → Dutch |
|
| 23 |
+
| [`joeranbosma/dragon-roberta-large-mixed-domain`](https://huggingface.co/joeranbosma/dragon-roberta-large-mixed-domain) | 560M | Multiple → Dutch |
|
| 24 |
+
| [`joeranbosma/dragon-longformer-base-mixed-domain`](https://huggingface.co/joeranbosma/dragon-longformer-base-mixed-domain) | 149M | English → Dutch |
|
| 25 |
+
| [`joeranbosma/dragon-longformer-large-mixed-domain`](https://huggingface.co/joeranbosma/dragon-longformer-large-mixed-domain) | 435M | English → Dutch |
|
| 26 |
+
| [`joeranbosma/dragon-bert-base-domain-specific`](https://huggingface.co/joeranbosma/dragon-bert-base-domain-specific) | 109M | Dutch |
|
| 27 |
+
| [`joeranbosma/dragon-roberta-base-domain-specific`](https://huggingface.co/joeranbosma/dragon-roberta-base-domain-specific) | 278M | Dutch |
|
| 28 |
+
| [`joeranbosma/dragon-roberta-large-domain-specific`](https://huggingface.co/joeranbosma/dragon-roberta-large-domain-specific) | 560M | Dutch |
|
| 29 |
+
| [`joeranbosma/dragon-longformer-base-domain-specific`](https://huggingface.co/joeranbosma/dragon-longformer-base-domain-specific) | 149M | Dutch |
|
| 30 |
+
| [`joeranbosma/dragon-longformer-large-domain-specific`](https://huggingface.co/joeranbosma/dragon-longformer-large-domain-specific) | 435M | Dutch |
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
## Intended uses & limitations
|
| 34 |
+
You can use the raw model for masked language modeling, but it's mostly intended to be fine-tuned on a downstream task.
|
| 35 |
+
|
| 36 |
+
Note that this model is primarily aimed at being fine-tuned on tasks that use the whole text (e.g., a clinical report) to make decisions, such as sequence classification, token classification or question answering. For tasks such as text generation you should look at model like GPT2.
|
| 37 |
+
|
| 38 |
+
## How to use
|
| 39 |
+
You can use this model directly with a pipeline for masked language modeling:
|
| 40 |
+
|
| 41 |
+
```python
|
| 42 |
+
from transformers import pipeline
|
| 43 |
+
unmasker = pipeline("fill-mask", model="joeranbosma/dragon-roberta-large-domain-specific")
|
| 44 |
+
unmasker("Dit onderzoek geen aanwijzingen voor significant carcinoom. PIRADS <mask>.")
|
| 45 |
+
```
|
| 46 |
+
|
| 47 |
+
Here is how to use this model to get the features of a given text in PyTorch:
|
| 48 |
+
|
| 49 |
+
```python
|
| 50 |
+
from transformers import AutoTokenizer, AutoModel
|
| 51 |
+
tokenizer = AutoTokenizer.from_pretrained("joeranbosma/dragon-roberta-large-domain-specific")
|
| 52 |
+
model = AutoModel.from_pretrained("joeranbosma/dragon-roberta-large-domain-specific")
|
| 53 |
+
text = "Replace me by any text you'd like."
|
| 54 |
+
encoded_input = tokenizer(text, return_tensors="pt")
|
| 55 |
+
output = model(**encoded_input)
|
| 56 |
+
```
|
| 57 |
+
|
| 58 |
+
## Limitations and bias
|
| 59 |
+
Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions. This bias will also affect all fine-tuned versions of this model.
|
| 60 |
+
|
| 61 |
+
## Training data
|
| 62 |
+
For pretraining, 4,333,201 clinical reports (466,351 consecutive patients) were selected from Ziekenhuisgroep Twente from patients with a diagnostic or interventional visit between 13 July 2000 and 25 April 2023. 180,439 duplicate clinical reports (179,808 patients) were excluded, resulting in 4,152,762 included reports (463,692 patients). These reports were split into training (80%, 3,322,209 reports), validation (10%, 415,276 reports), and testing (10%, 415,277 reports). The testing reports were set aside for future analysis and are not used for pretraining.
|
| 63 |
+
|
| 64 |
+
## Training procedure
|
| 65 |
+
|
| 66 |
+
### Pretraining
|
| 67 |
+
The model was pretrained using masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then runs the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally masks the future tokens. It allows the model to learn a bidirectional representation of the sentence.
|
| 68 |
+
|
| 69 |
+
The details of the masking procedure for each sentence are the following:
|
| 70 |
+
- 15% of the tokens are masked.
|
| 71 |
+
- In 80% of the cases, the masked tokens are replaced by `[MASK]`.
|
| 72 |
+
- In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace.
|
| 73 |
+
- In the 10% remaining cases, the masked tokens are left as is.
|
| 74 |
+
|
| 75 |
+
The HuggingFace implementation was used for pretraining: [`run_mlm.py`](https://github.com/huggingface/transformers/blob/7c6ec195adbfcd22cb6baeee64dd3c24a4b80c74/examples/pytorch/language-modeling/run_mlm.py).
|
| 76 |
+
|
| 77 |
+
### Pretraining hyperparameters
|
| 78 |
+
|
| 79 |
+
The following hyperparameters were used during pretraining:
|
| 80 |
+
- `learning_rate`: 1e-4
|
| 81 |
+
- `train_batch_size`: 8
|
| 82 |
+
- `eval_batch_size`: 8
|
| 83 |
+
- `seed`: 42
|
| 84 |
+
- `gradient_accumulation_steps`: 32
|
| 85 |
+
- `total_train_batch_size`: 256
|
| 86 |
+
- `optimizer`: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 87 |
+
- `lr_scheduler_type`: linear
|
| 88 |
+
- `num_epochs`: 10.0
|
| 89 |
+
- `max_seq_length`: 512
|
| 90 |
+
|
| 91 |
+
### Framework versions
|
| 92 |
+
|
| 93 |
+
- Transformers 4.29.0.dev0
|
| 94 |
+
- Pytorch 2.0.0+cu117
|
| 95 |
+
- Datasets 2.11.0
|
| 96 |
+
- Tokenizers 0.13.3
|
| 97 |
+
|
| 98 |
+
## Evaluation results
|
| 99 |
+
|
| 100 |
+
Pending evaluation on the DRAGON benchmark.
|
| 101 |
+
|
| 102 |
+
### BibTeX entry and citation info
|
| 103 |
+
|
| 104 |
+
```bibtex
|
| 105 |
+
@article{PENDING}
|
| 106 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.8176872903958383,
|
| 4 |
+
"eval_loss": 0.7863156199455261,
|
| 5 |
+
"eval_runtime": 4180.1874,
|
| 6 |
+
"eval_samples": 103187,
|
| 7 |
+
"eval_samples_per_second": 24.685,
|
| 8 |
+
"eval_steps_per_second": 3.086,
|
| 9 |
+
"perplexity": 2.19529321323568,
|
| 10 |
+
"train_loss": 1.915723304985473,
|
| 11 |
+
"train_runtime": 1372706.2896,
|
| 12 |
+
"train_samples": 824387,
|
| 13 |
+
"train_samples_per_second": 6.006,
|
| 14 |
+
"train_steps_per_second": 0.023
|
| 15 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "roberta-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 1024,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 4096,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 16,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"pad_token_id": 1,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.29.0.dev0",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 50265
|
| 27 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.8176872903958383,
|
| 4 |
+
"eval_loss": 0.7863156199455261,
|
| 5 |
+
"eval_runtime": 4180.1874,
|
| 6 |
+
"eval_samples": 103187,
|
| 7 |
+
"eval_samples_per_second": 24.685,
|
| 8 |
+
"eval_steps_per_second": 3.086,
|
| 9 |
+
"perplexity": 2.19529321323568
|
| 10 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fed4ee103e6fb78e381436d1582c8dcab0299d5522f37fe7b58338bfc8638a83
|
| 3 |
+
size 1421788537
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"content": "<mask>",
|
| 7 |
+
"lstrip": true,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"sep_token": "</s>",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": "<s>",
|
| 4 |
+
"clean_up_tokenization_spaces": true,
|
| 5 |
+
"cls_token": "<s>",
|
| 6 |
+
"eos_token": "</s>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"mask_token": {
|
| 9 |
+
"__type": "AddedToken",
|
| 10 |
+
"content": "<mask>",
|
| 11 |
+
"lstrip": true,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"model_max_length": 512,
|
| 17 |
+
"pad_token": "<pad>",
|
| 18 |
+
"sep_token": "</s>",
|
| 19 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 20 |
+
"trim_offsets": true,
|
| 21 |
+
"unk_token": "<unk>"
|
| 22 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 1.915723304985473,
|
| 4 |
+
"train_runtime": 1372706.2896,
|
| 5 |
+
"train_samples": 824387,
|
| 6 |
+
"train_samples_per_second": 6.006,
|
| 7 |
+
"train_steps_per_second": 0.023
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1327 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7846992611885071,
|
| 3 |
+
"best_model_checkpoint": "/output/zgt-roberta-large-finetuned-Gu21schedule-BS256-10ep/checkpoint-31926",
|
| 4 |
+
"epoch": 9.99912662907937,
|
| 5 |
+
"global_step": 32200,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.1,
|
| 12 |
+
"eval_accuracy": 0.12095168799057326,
|
| 13 |
+
"eval_loss": 6.969939231872559,
|
| 14 |
+
"eval_runtime": 4175.606,
|
| 15 |
+
"eval_samples_per_second": 24.712,
|
| 16 |
+
"eval_steps_per_second": 3.089,
|
| 17 |
+
"step": 313
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.16,
|
| 21 |
+
"learning_rate": 1.5527950310559007e-05,
|
| 22 |
+
"loss": 7.8141,
|
| 23 |
+
"step": 500
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.19,
|
| 27 |
+
"eval_accuracy": 0.1524057089518688,
|
| 28 |
+
"eval_loss": 6.233978271484375,
|
| 29 |
+
"eval_runtime": 4177.4061,
|
| 30 |
+
"eval_samples_per_second": 24.701,
|
| 31 |
+
"eval_steps_per_second": 3.088,
|
| 32 |
+
"step": 626
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.29,
|
| 36 |
+
"eval_accuracy": 0.16394060788435189,
|
| 37 |
+
"eval_loss": 6.122518539428711,
|
| 38 |
+
"eval_runtime": 4179.7801,
|
| 39 |
+
"eval_samples_per_second": 24.687,
|
| 40 |
+
"eval_steps_per_second": 3.086,
|
| 41 |
+
"step": 939
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.31,
|
| 45 |
+
"learning_rate": 3.1055900621118014e-05,
|
| 46 |
+
"loss": 6.1501,
|
| 47 |
+
"step": 1000
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.39,
|
| 51 |
+
"eval_accuracy": 0.16842972332423103,
|
| 52 |
+
"eval_loss": 6.045305252075195,
|
| 53 |
+
"eval_runtime": 4182.1677,
|
| 54 |
+
"eval_samples_per_second": 24.673,
|
| 55 |
+
"eval_steps_per_second": 3.084,
|
| 56 |
+
"step": 1252
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.47,
|
| 60 |
+
"learning_rate": 4.658385093167702e-05,
|
| 61 |
+
"loss": 6.0737,
|
| 62 |
+
"step": 1500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.49,
|
| 66 |
+
"eval_accuracy": 0.16725142510331475,
|
| 67 |
+
"eval_loss": 5.987020969390869,
|
| 68 |
+
"eval_runtime": 4177.226,
|
| 69 |
+
"eval_samples_per_second": 24.702,
|
| 70 |
+
"eval_steps_per_second": 3.088,
|
| 71 |
+
"step": 1565
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.58,
|
| 75 |
+
"eval_accuracy": 0.17195286451581493,
|
| 76 |
+
"eval_loss": 5.9314141273498535,
|
| 77 |
+
"eval_runtime": 4165.2923,
|
| 78 |
+
"eval_samples_per_second": 24.773,
|
| 79 |
+
"eval_steps_per_second": 3.097,
|
| 80 |
+
"step": 1878
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.62,
|
| 84 |
+
"learning_rate": 6.211180124223603e-05,
|
| 85 |
+
"loss": 5.9753,
|
| 86 |
+
"step": 2000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.68,
|
| 90 |
+
"eval_accuracy": 0.17336381929439706,
|
| 91 |
+
"eval_loss": 5.887312412261963,
|
| 92 |
+
"eval_runtime": 4179.8742,
|
| 93 |
+
"eval_samples_per_second": 24.687,
|
| 94 |
+
"eval_steps_per_second": 3.086,
|
| 95 |
+
"step": 2191
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.78,
|
| 99 |
+
"learning_rate": 7.763975155279503e-05,
|
| 100 |
+
"loss": 5.9008,
|
| 101 |
+
"step": 2500
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.78,
|
| 105 |
+
"eval_accuracy": 0.1737213537890283,
|
| 106 |
+
"eval_loss": 5.851128101348877,
|
| 107 |
+
"eval_runtime": 4181.0014,
|
| 108 |
+
"eval_samples_per_second": 24.68,
|
| 109 |
+
"eval_steps_per_second": 3.085,
|
| 110 |
+
"step": 2504
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.87,
|
| 114 |
+
"eval_accuracy": 0.1730111369884377,
|
| 115 |
+
"eval_loss": 5.819331645965576,
|
| 116 |
+
"eval_runtime": 4182.8162,
|
| 117 |
+
"eval_samples_per_second": 24.669,
|
| 118 |
+
"eval_steps_per_second": 3.084,
|
| 119 |
+
"step": 2817
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.93,
|
| 123 |
+
"learning_rate": 9.316770186335404e-05,
|
| 124 |
+
"loss": 5.8379,
|
| 125 |
+
"step": 3000
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.97,
|
| 129 |
+
"eval_accuracy": 0.17577215201939525,
|
| 130 |
+
"eval_loss": 5.785292625427246,
|
| 131 |
+
"eval_runtime": 4184.6612,
|
| 132 |
+
"eval_samples_per_second": 24.658,
|
| 133 |
+
"eval_steps_per_second": 3.082,
|
| 134 |
+
"step": 3130
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.07,
|
| 138 |
+
"eval_accuracy": 0.1759501659892642,
|
| 139 |
+
"eval_loss": 5.762609481811523,
|
| 140 |
+
"eval_runtime": 4182.8465,
|
| 141 |
+
"eval_samples_per_second": 24.669,
|
| 142 |
+
"eval_steps_per_second": 3.084,
|
| 143 |
+
"step": 3443
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.09,
|
| 147 |
+
"learning_rate": 9.903381642512077e-05,
|
| 148 |
+
"loss": 5.7885,
|
| 149 |
+
"step": 3500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.17,
|
| 153 |
+
"eval_accuracy": 0.17622262356212584,
|
| 154 |
+
"eval_loss": 5.739773273468018,
|
| 155 |
+
"eval_runtime": 4170.2068,
|
| 156 |
+
"eval_samples_per_second": 24.744,
|
| 157 |
+
"eval_steps_per_second": 3.093,
|
| 158 |
+
"step": 3756
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.24,
|
| 162 |
+
"learning_rate": 9.730848861283644e-05,
|
| 163 |
+
"loss": 5.7464,
|
| 164 |
+
"step": 4000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.26,
|
| 168 |
+
"eval_accuracy": 0.1766708743216783,
|
| 169 |
+
"eval_loss": 5.720291614532471,
|
| 170 |
+
"eval_runtime": 4172.2987,
|
| 171 |
+
"eval_samples_per_second": 24.731,
|
| 172 |
+
"eval_steps_per_second": 3.092,
|
| 173 |
+
"step": 4069
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.36,
|
| 177 |
+
"eval_accuracy": 0.17642257605364067,
|
| 178 |
+
"eval_loss": 5.696093559265137,
|
| 179 |
+
"eval_runtime": 4172.9563,
|
| 180 |
+
"eval_samples_per_second": 24.728,
|
| 181 |
+
"eval_steps_per_second": 3.091,
|
| 182 |
+
"step": 4382
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.4,
|
| 186 |
+
"learning_rate": 9.558316080055211e-05,
|
| 187 |
+
"loss": 5.7149,
|
| 188 |
+
"step": 4500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.46,
|
| 192 |
+
"eval_accuracy": 0.17737878845804067,
|
| 193 |
+
"eval_loss": 5.683297634124756,
|
| 194 |
+
"eval_runtime": 4185.0558,
|
| 195 |
+
"eval_samples_per_second": 24.656,
|
| 196 |
+
"eval_steps_per_second": 3.082,
|
| 197 |
+
"step": 4695
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.55,
|
| 201 |
+
"learning_rate": 9.385783298826778e-05,
|
| 202 |
+
"loss": 5.6885,
|
| 203 |
+
"step": 5000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.56,
|
| 207 |
+
"eval_accuracy": 0.1773234355905989,
|
| 208 |
+
"eval_loss": 5.670944690704346,
|
| 209 |
+
"eval_runtime": 4184.0795,
|
| 210 |
+
"eval_samples_per_second": 24.662,
|
| 211 |
+
"eval_steps_per_second": 3.083,
|
| 212 |
+
"step": 5008
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.65,
|
| 216 |
+
"eval_accuracy": 0.1774909217337557,
|
| 217 |
+
"eval_loss": 5.6566619873046875,
|
| 218 |
+
"eval_runtime": 4185.5299,
|
| 219 |
+
"eval_samples_per_second": 24.653,
|
| 220 |
+
"eval_steps_per_second": 3.082,
|
| 221 |
+
"step": 5321
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.71,
|
| 225 |
+
"learning_rate": 9.213250517598345e-05,
|
| 226 |
+
"loss": 5.666,
|
| 227 |
+
"step": 5500
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.75,
|
| 231 |
+
"eval_accuracy": 0.17712345558640866,
|
| 232 |
+
"eval_loss": 5.647720813751221,
|
| 233 |
+
"eval_runtime": 4178.6309,
|
| 234 |
+
"eval_samples_per_second": 24.694,
|
| 235 |
+
"eval_steps_per_second": 3.087,
|
| 236 |
+
"step": 5634
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.85,
|
| 240 |
+
"eval_accuracy": 0.1778983102903151,
|
| 241 |
+
"eval_loss": 5.632648944854736,
|
| 242 |
+
"eval_runtime": 4179.5279,
|
| 243 |
+
"eval_samples_per_second": 24.689,
|
| 244 |
+
"eval_steps_per_second": 3.086,
|
| 245 |
+
"step": 5947
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.86,
|
| 249 |
+
"learning_rate": 9.04071773636991e-05,
|
| 250 |
+
"loss": 5.6458,
|
| 251 |
+
"step": 6000
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.94,
|
| 255 |
+
"eval_accuracy": 0.1767171879623295,
|
| 256 |
+
"eval_loss": 5.626438140869141,
|
| 257 |
+
"eval_runtime": 4166.358,
|
| 258 |
+
"eval_samples_per_second": 24.767,
|
| 259 |
+
"eval_steps_per_second": 3.096,
|
| 260 |
+
"step": 6260
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.02,
|
| 264 |
+
"learning_rate": 8.868184955141477e-05,
|
| 265 |
+
"loss": 5.6293,
|
| 266 |
+
"step": 6500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.04,
|
| 270 |
+
"eval_accuracy": 0.18033812690505616,
|
| 271 |
+
"eval_loss": 5.371770858764648,
|
| 272 |
+
"eval_runtime": 4187.0608,
|
| 273 |
+
"eval_samples_per_second": 24.644,
|
| 274 |
+
"eval_steps_per_second": 3.081,
|
| 275 |
+
"step": 6573
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 2.14,
|
| 279 |
+
"eval_accuracy": 0.22889979190190532,
|
| 280 |
+
"eval_loss": 4.608296871185303,
|
| 281 |
+
"eval_runtime": 4185.2213,
|
| 282 |
+
"eval_samples_per_second": 24.655,
|
| 283 |
+
"eval_steps_per_second": 3.082,
|
| 284 |
+
"step": 6886
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.17,
|
| 288 |
+
"learning_rate": 8.695652173913044e-05,
|
| 289 |
+
"loss": 4.9883,
|
| 290 |
+
"step": 7000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.24,
|
| 294 |
+
"eval_accuracy": 0.4614915191453084,
|
| 295 |
+
"eval_loss": 3.1735970973968506,
|
| 296 |
+
"eval_runtime": 4184.0355,
|
| 297 |
+
"eval_samples_per_second": 24.662,
|
| 298 |
+
"eval_steps_per_second": 3.083,
|
| 299 |
+
"step": 7199
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 2.33,
|
| 303 |
+
"learning_rate": 8.523119392684611e-05,
|
| 304 |
+
"loss": 3.2514,
|
| 305 |
+
"step": 7500
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.33,
|
| 309 |
+
"eval_accuracy": 0.5874539351791631,
|
| 310 |
+
"eval_loss": 2.3033511638641357,
|
| 311 |
+
"eval_runtime": 4180.5231,
|
| 312 |
+
"eval_samples_per_second": 24.683,
|
| 313 |
+
"eval_steps_per_second": 3.085,
|
| 314 |
+
"step": 7512
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.43,
|
| 318 |
+
"eval_accuracy": 0.6407011451986726,
|
| 319 |
+
"eval_loss": 1.915069818496704,
|
| 320 |
+
"eval_runtime": 4169.2257,
|
| 321 |
+
"eval_samples_per_second": 24.75,
|
| 322 |
+
"eval_steps_per_second": 3.094,
|
| 323 |
+
"step": 7825
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 2.48,
|
| 327 |
+
"learning_rate": 8.350586611456177e-05,
|
| 328 |
+
"loss": 2.1211,
|
| 329 |
+
"step": 8000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 2.53,
|
| 333 |
+
"eval_accuracy": 0.6671589447070213,
|
| 334 |
+
"eval_loss": 1.7218379974365234,
|
| 335 |
+
"eval_runtime": 4169.0733,
|
| 336 |
+
"eval_samples_per_second": 24.751,
|
| 337 |
+
"eval_steps_per_second": 3.094,
|
| 338 |
+
"step": 8138
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.62,
|
| 342 |
+
"eval_accuracy": 0.6849309492119486,
|
| 343 |
+
"eval_loss": 1.5953431129455566,
|
| 344 |
+
"eval_runtime": 4186.0587,
|
| 345 |
+
"eval_samples_per_second": 24.65,
|
| 346 |
+
"eval_steps_per_second": 3.081,
|
| 347 |
+
"step": 8451
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.64,
|
| 351 |
+
"learning_rate": 8.178053830227743e-05,
|
| 352 |
+
"loss": 1.7698,
|
| 353 |
+
"step": 8500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.72,
|
| 357 |
+
"eval_accuracy": 0.6977890826334601,
|
| 358 |
+
"eval_loss": 1.5040490627288818,
|
| 359 |
+
"eval_runtime": 4187.545,
|
| 360 |
+
"eval_samples_per_second": 24.641,
|
| 361 |
+
"eval_steps_per_second": 3.08,
|
| 362 |
+
"step": 8764
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.79,
|
| 366 |
+
"learning_rate": 8.00552104899931e-05,
|
| 367 |
+
"loss": 1.5907,
|
| 368 |
+
"step": 9000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.82,
|
| 372 |
+
"eval_accuracy": 0.7089186346832345,
|
| 373 |
+
"eval_loss": 1.4300199747085571,
|
| 374 |
+
"eval_runtime": 4186.4554,
|
| 375 |
+
"eval_samples_per_second": 24.648,
|
| 376 |
+
"eval_steps_per_second": 3.081,
|
| 377 |
+
"step": 9077
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.92,
|
| 381 |
+
"eval_accuracy": 0.716823811317197,
|
| 382 |
+
"eval_loss": 1.3781534433364868,
|
| 383 |
+
"eval_runtime": 4180.8901,
|
| 384 |
+
"eval_samples_per_second": 24.681,
|
| 385 |
+
"eval_steps_per_second": 3.085,
|
| 386 |
+
"step": 9390
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.95,
|
| 390 |
+
"learning_rate": 7.832988267770877e-05,
|
| 391 |
+
"loss": 1.4757,
|
| 392 |
+
"step": 9500
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 3.01,
|
| 396 |
+
"eval_accuracy": 0.7239590766169688,
|
| 397 |
+
"eval_loss": 1.3298745155334473,
|
| 398 |
+
"eval_runtime": 4169.3417,
|
| 399 |
+
"eval_samples_per_second": 24.749,
|
| 400 |
+
"eval_steps_per_second": 3.094,
|
| 401 |
+
"step": 9703
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 3.11,
|
| 405 |
+
"learning_rate": 7.660455486542444e-05,
|
| 406 |
+
"loss": 1.3919,
|
| 407 |
+
"step": 10000
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.11,
|
| 411 |
+
"eval_accuracy": 0.7302072364465025,
|
| 412 |
+
"eval_loss": 1.2895771265029907,
|
| 413 |
+
"eval_runtime": 4181.4693,
|
| 414 |
+
"eval_samples_per_second": 24.677,
|
| 415 |
+
"eval_steps_per_second": 3.085,
|
| 416 |
+
"step": 10016
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 3.21,
|
| 420 |
+
"eval_accuracy": 0.7356055265010495,
|
| 421 |
+
"eval_loss": 1.25456964969635,
|
| 422 |
+
"eval_runtime": 4169.8649,
|
| 423 |
+
"eval_samples_per_second": 24.746,
|
| 424 |
+
"eval_steps_per_second": 3.093,
|
| 425 |
+
"step": 10329
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 3.26,
|
| 429 |
+
"learning_rate": 7.48792270531401e-05,
|
| 430 |
+
"loss": 1.328,
|
| 431 |
+
"step": 10500
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 3.3,
|
| 435 |
+
"eval_accuracy": 0.7406319445625766,
|
| 436 |
+
"eval_loss": 1.2240813970565796,
|
| 437 |
+
"eval_runtime": 4170.5091,
|
| 438 |
+
"eval_samples_per_second": 24.742,
|
| 439 |
+
"eval_steps_per_second": 3.093,
|
| 440 |
+
"step": 10642
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 3.4,
|
| 444 |
+
"eval_accuracy": 0.745277920243707,
|
| 445 |
+
"eval_loss": 1.1944962739944458,
|
| 446 |
+
"eval_runtime": 4174.7942,
|
| 447 |
+
"eval_samples_per_second": 24.717,
|
| 448 |
+
"eval_steps_per_second": 3.09,
|
| 449 |
+
"step": 10955
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 3.42,
|
| 453 |
+
"learning_rate": 7.315389924085577e-05,
|
| 454 |
+
"loss": 1.2782,
|
| 455 |
+
"step": 11000
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 3.5,
|
| 459 |
+
"eval_accuracy": 0.7494278825432492,
|
| 460 |
+
"eval_loss": 1.1713906526565552,
|
| 461 |
+
"eval_runtime": 4174.964,
|
| 462 |
+
"eval_samples_per_second": 24.716,
|
| 463 |
+
"eval_steps_per_second": 3.09,
|
| 464 |
+
"step": 11268
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 3.57,
|
| 468 |
+
"learning_rate": 7.142857142857143e-05,
|
| 469 |
+
"loss": 1.2357,
|
| 470 |
+
"step": 11500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 3.6,
|
| 474 |
+
"eval_accuracy": 0.7529051534363226,
|
| 475 |
+
"eval_loss": 1.1493370532989502,
|
| 476 |
+
"eval_runtime": 4169.6928,
|
| 477 |
+
"eval_samples_per_second": 24.747,
|
| 478 |
+
"eval_steps_per_second": 3.094,
|
| 479 |
+
"step": 11581
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 3.69,
|
| 483 |
+
"eval_accuracy": 0.7561643316896146,
|
| 484 |
+
"eval_loss": 1.1291333436965942,
|
| 485 |
+
"eval_runtime": 4170.8848,
|
| 486 |
+
"eval_samples_per_second": 24.74,
|
| 487 |
+
"eval_steps_per_second": 3.093,
|
| 488 |
+
"step": 11894
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 3.73,
|
| 492 |
+
"learning_rate": 6.970324361628709e-05,
|
| 493 |
+
"loss": 1.1986,
|
| 494 |
+
"step": 12000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 3.79,
|
| 498 |
+
"eval_accuracy": 0.7596194578598549,
|
| 499 |
+
"eval_loss": 1.1110583543777466,
|
| 500 |
+
"eval_runtime": 4174.4741,
|
| 501 |
+
"eval_samples_per_second": 24.719,
|
| 502 |
+
"eval_steps_per_second": 3.09,
|
| 503 |
+
"step": 12207
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 3.88,
|
| 507 |
+
"learning_rate": 6.797791580400277e-05,
|
| 508 |
+
"loss": 1.1673,
|
| 509 |
+
"step": 12500
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 3.89,
|
| 513 |
+
"eval_accuracy": 0.7632696453538697,
|
| 514 |
+
"eval_loss": 1.091315507888794,
|
| 515 |
+
"eval_runtime": 4175.4221,
|
| 516 |
+
"eval_samples_per_second": 24.713,
|
| 517 |
+
"eval_steps_per_second": 3.089,
|
| 518 |
+
"step": 12520
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 3.99,
|
| 522 |
+
"eval_accuracy": 0.7654154836189226,
|
| 523 |
+
"eval_loss": 1.0766664743423462,
|
| 524 |
+
"eval_runtime": 4173.9192,
|
| 525 |
+
"eval_samples_per_second": 24.722,
|
| 526 |
+
"eval_steps_per_second": 3.09,
|
| 527 |
+
"step": 12833
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 4.04,
|
| 531 |
+
"learning_rate": 6.625258799171843e-05,
|
| 532 |
+
"loss": 1.1387,
|
| 533 |
+
"step": 13000
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 4.08,
|
| 537 |
+
"eval_accuracy": 0.7681167597878354,
|
| 538 |
+
"eval_loss": 1.0629887580871582,
|
| 539 |
+
"eval_runtime": 4175.5718,
|
| 540 |
+
"eval_samples_per_second": 24.712,
|
| 541 |
+
"eval_steps_per_second": 3.089,
|
| 542 |
+
"step": 13146
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 4.18,
|
| 546 |
+
"eval_accuracy": 0.7707155612474973,
|
| 547 |
+
"eval_loss": 1.046230673789978,
|
| 548 |
+
"eval_runtime": 4175.7461,
|
| 549 |
+
"eval_samples_per_second": 24.711,
|
| 550 |
+
"eval_steps_per_second": 3.089,
|
| 551 |
+
"step": 13459
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 4.19,
|
| 555 |
+
"learning_rate": 6.45272601794341e-05,
|
| 556 |
+
"loss": 1.1074,
|
| 557 |
+
"step": 13500
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 4.28,
|
| 561 |
+
"eval_accuracy": 0.7731418474891986,
|
| 562 |
+
"eval_loss": 1.0337274074554443,
|
| 563 |
+
"eval_runtime": 4174.5,
|
| 564 |
+
"eval_samples_per_second": 24.718,
|
| 565 |
+
"eval_steps_per_second": 3.09,
|
| 566 |
+
"step": 13772
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 4.35,
|
| 570 |
+
"learning_rate": 6.280193236714976e-05,
|
| 571 |
+
"loss": 1.0893,
|
| 572 |
+
"step": 14000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 4.37,
|
| 576 |
+
"eval_accuracy": 0.7748741121676294,
|
| 577 |
+
"eval_loss": 1.0239219665527344,
|
| 578 |
+
"eval_runtime": 4175.3624,
|
| 579 |
+
"eval_samples_per_second": 24.713,
|
| 580 |
+
"eval_steps_per_second": 3.089,
|
| 581 |
+
"step": 14085
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 4.47,
|
| 585 |
+
"eval_accuracy": 0.7766543779252701,
|
| 586 |
+
"eval_loss": 1.014146089553833,
|
| 587 |
+
"eval_runtime": 4181.6475,
|
| 588 |
+
"eval_samples_per_second": 24.676,
|
| 589 |
+
"eval_steps_per_second": 3.085,
|
| 590 |
+
"step": 14398
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 4.5,
|
| 594 |
+
"learning_rate": 6.107660455486542e-05,
|
| 595 |
+
"loss": 1.0682,
|
| 596 |
+
"step": 14500
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 4.57,
|
| 600 |
+
"eval_accuracy": 0.7785186060053325,
|
| 601 |
+
"eval_loss": 1.0032246112823486,
|
| 602 |
+
"eval_runtime": 4180.7422,
|
| 603 |
+
"eval_samples_per_second": 24.682,
|
| 604 |
+
"eval_steps_per_second": 3.085,
|
| 605 |
+
"step": 14711
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 4.66,
|
| 609 |
+
"learning_rate": 5.9351276742581096e-05,
|
| 610 |
+
"loss": 1.0524,
|
| 611 |
+
"step": 15000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 4.67,
|
| 615 |
+
"eval_accuracy": 0.7806327231861181,
|
| 616 |
+
"eval_loss": 0.9907068014144897,
|
| 617 |
+
"eval_runtime": 4180.8503,
|
| 618 |
+
"eval_samples_per_second": 24.681,
|
| 619 |
+
"eval_steps_per_second": 3.085,
|
| 620 |
+
"step": 15024
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 4.76,
|
| 624 |
+
"eval_accuracy": 0.7826726152932884,
|
| 625 |
+
"eval_loss": 0.9796966910362244,
|
| 626 |
+
"eval_runtime": 4176.9966,
|
| 627 |
+
"eval_samples_per_second": 24.704,
|
| 628 |
+
"eval_steps_per_second": 3.088,
|
| 629 |
+
"step": 15337
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 4.81,
|
| 633 |
+
"learning_rate": 5.762594893029676e-05,
|
| 634 |
+
"loss": 1.0338,
|
| 635 |
+
"step": 15500
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 4.86,
|
| 639 |
+
"eval_accuracy": 0.7838565333062765,
|
| 640 |
+
"eval_loss": 0.9712271690368652,
|
| 641 |
+
"eval_runtime": 4181.1547,
|
| 642 |
+
"eval_samples_per_second": 24.679,
|
| 643 |
+
"eval_steps_per_second": 3.085,
|
| 644 |
+
"step": 15650
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 4.96,
|
| 648 |
+
"eval_accuracy": 0.7854880736072989,
|
| 649 |
+
"eval_loss": 0.9626355767250061,
|
| 650 |
+
"eval_runtime": 4176.3487,
|
| 651 |
+
"eval_samples_per_second": 24.707,
|
| 652 |
+
"eval_steps_per_second": 3.089,
|
| 653 |
+
"step": 15963
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 4.97,
|
| 657 |
+
"learning_rate": 5.590062111801242e-05,
|
| 658 |
+
"loss": 1.0188,
|
| 659 |
+
"step": 16000
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 5.05,
|
| 663 |
+
"eval_accuracy": 0.7867994298371893,
|
| 664 |
+
"eval_loss": 0.9558107256889343,
|
| 665 |
+
"eval_runtime": 4179.0241,
|
| 666 |
+
"eval_samples_per_second": 24.692,
|
| 667 |
+
"eval_steps_per_second": 3.087,
|
| 668 |
+
"step": 16276
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 5.12,
|
| 672 |
+
"learning_rate": 5.417529330572809e-05,
|
| 673 |
+
"loss": 1.003,
|
| 674 |
+
"step": 16500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 5.15,
|
| 678 |
+
"eval_accuracy": 0.7884956846420655,
|
| 679 |
+
"eval_loss": 0.9470182657241821,
|
| 680 |
+
"eval_runtime": 4176.2829,
|
| 681 |
+
"eval_samples_per_second": 24.708,
|
| 682 |
+
"eval_steps_per_second": 3.089,
|
| 683 |
+
"step": 16589
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 5.25,
|
| 687 |
+
"eval_accuracy": 0.789320227965163,
|
| 688 |
+
"eval_loss": 0.9419927597045898,
|
| 689 |
+
"eval_runtime": 4176.4362,
|
| 690 |
+
"eval_samples_per_second": 24.707,
|
| 691 |
+
"eval_steps_per_second": 3.089,
|
| 692 |
+
"step": 16902
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 5.28,
|
| 696 |
+
"learning_rate": 5.244996549344375e-05,
|
| 697 |
+
"loss": 0.989,
|
| 698 |
+
"step": 17000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 5.35,
|
| 702 |
+
"eval_accuracy": 0.7911368944368907,
|
| 703 |
+
"eval_loss": 0.9319880604743958,
|
| 704 |
+
"eval_runtime": 4173.211,
|
| 705 |
+
"eval_samples_per_second": 24.726,
|
| 706 |
+
"eval_steps_per_second": 3.091,
|
| 707 |
+
"step": 17215
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 5.43,
|
| 711 |
+
"learning_rate": 5.072463768115943e-05,
|
| 712 |
+
"loss": 0.9786,
|
| 713 |
+
"step": 17500
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 5.44,
|
| 717 |
+
"eval_accuracy": 0.7913904624057875,
|
| 718 |
+
"eval_loss": 0.9292972683906555,
|
| 719 |
+
"eval_runtime": 4172.1427,
|
| 720 |
+
"eval_samples_per_second": 24.732,
|
| 721 |
+
"eval_steps_per_second": 3.092,
|
| 722 |
+
"step": 17528
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 5.54,
|
| 726 |
+
"eval_accuracy": 0.7932201930680667,
|
| 727 |
+
"eval_loss": 0.9203895330429077,
|
| 728 |
+
"eval_runtime": 4172.1933,
|
| 729 |
+
"eval_samples_per_second": 24.732,
|
| 730 |
+
"eval_steps_per_second": 3.092,
|
| 731 |
+
"step": 17841
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 5.59,
|
| 735 |
+
"learning_rate": 4.899930986887509e-05,
|
| 736 |
+
"loss": 0.9688,
|
| 737 |
+
"step": 18000
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 5.64,
|
| 741 |
+
"eval_accuracy": 0.7938784128863696,
|
| 742 |
+
"eval_loss": 0.9155610799789429,
|
| 743 |
+
"eval_runtime": 4171.4989,
|
| 744 |
+
"eval_samples_per_second": 24.736,
|
| 745 |
+
"eval_steps_per_second": 3.092,
|
| 746 |
+
"step": 18154
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 5.73,
|
| 750 |
+
"eval_accuracy": 0.7952068414816341,
|
| 751 |
+
"eval_loss": 0.9091127514839172,
|
| 752 |
+
"eval_runtime": 4177.1927,
|
| 753 |
+
"eval_samples_per_second": 24.702,
|
| 754 |
+
"eval_steps_per_second": 3.088,
|
| 755 |
+
"step": 18467
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 5.74,
|
| 759 |
+
"learning_rate": 4.727398205659075e-05,
|
| 760 |
+
"loss": 0.9576,
|
| 761 |
+
"step": 18500
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 5.83,
|
| 765 |
+
"eval_accuracy": 0.7959312809475213,
|
| 766 |
+
"eval_loss": 0.9035018682479858,
|
| 767 |
+
"eval_runtime": 4178.199,
|
| 768 |
+
"eval_samples_per_second": 24.697,
|
| 769 |
+
"eval_steps_per_second": 3.087,
|
| 770 |
+
"step": 18780
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 5.9,
|
| 774 |
+
"learning_rate": 4.554865424430642e-05,
|
| 775 |
+
"loss": 0.9489,
|
| 776 |
+
"step": 19000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 5.93,
|
| 780 |
+
"eval_accuracy": 0.7968186229313537,
|
| 781 |
+
"eval_loss": 0.8974489569664001,
|
| 782 |
+
"eval_runtime": 4175.3753,
|
| 783 |
+
"eval_samples_per_second": 24.713,
|
| 784 |
+
"eval_steps_per_second": 3.089,
|
| 785 |
+
"step": 19093
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 6.03,
|
| 789 |
+
"eval_accuracy": 0.7980068894657456,
|
| 790 |
+
"eval_loss": 0.8927856087684631,
|
| 791 |
+
"eval_runtime": 4174.5551,
|
| 792 |
+
"eval_samples_per_second": 24.718,
|
| 793 |
+
"eval_steps_per_second": 3.09,
|
| 794 |
+
"step": 19406
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 6.06,
|
| 798 |
+
"learning_rate": 4.382332643202209e-05,
|
| 799 |
+
"loss": 0.9384,
|
| 800 |
+
"step": 19500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 6.12,
|
| 804 |
+
"eval_accuracy": 0.7989207834455775,
|
| 805 |
+
"eval_loss": 0.8878790140151978,
|
| 806 |
+
"eval_runtime": 4176.0286,
|
| 807 |
+
"eval_samples_per_second": 24.709,
|
| 808 |
+
"eval_steps_per_second": 3.089,
|
| 809 |
+
"step": 19719
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 6.21,
|
| 813 |
+
"learning_rate": 4.209799861973775e-05,
|
| 814 |
+
"loss": 0.9292,
|
| 815 |
+
"step": 20000
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 6.22,
|
| 819 |
+
"eval_accuracy": 0.7995861067424526,
|
| 820 |
+
"eval_loss": 0.8841228485107422,
|
| 821 |
+
"eval_runtime": 4174.1874,
|
| 822 |
+
"eval_samples_per_second": 24.72,
|
| 823 |
+
"eval_steps_per_second": 3.09,
|
| 824 |
+
"step": 20032
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 6.32,
|
| 828 |
+
"eval_accuracy": 0.8005012964274263,
|
| 829 |
+
"eval_loss": 0.8798208236694336,
|
| 830 |
+
"eval_runtime": 4177.291,
|
| 831 |
+
"eval_samples_per_second": 24.702,
|
| 832 |
+
"eval_steps_per_second": 3.088,
|
| 833 |
+
"step": 20345
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 6.37,
|
| 837 |
+
"learning_rate": 4.0372670807453414e-05,
|
| 838 |
+
"loss": 0.921,
|
| 839 |
+
"step": 20500
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 6.41,
|
| 843 |
+
"eval_accuracy": 0.8013221542890863,
|
| 844 |
+
"eval_loss": 0.8738238215446472,
|
| 845 |
+
"eval_runtime": 4179.3089,
|
| 846 |
+
"eval_samples_per_second": 24.69,
|
| 847 |
+
"eval_steps_per_second": 3.086,
|
| 848 |
+
"step": 20658
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 6.51,
|
| 852 |
+
"eval_accuracy": 0.8020292962858104,
|
| 853 |
+
"eval_loss": 0.8702828288078308,
|
| 854 |
+
"eval_runtime": 4183.644,
|
| 855 |
+
"eval_samples_per_second": 24.664,
|
| 856 |
+
"eval_steps_per_second": 3.083,
|
| 857 |
+
"step": 20971
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 6.52,
|
| 861 |
+
"learning_rate": 3.864734299516908e-05,
|
| 862 |
+
"loss": 0.9142,
|
| 863 |
+
"step": 21000
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 6.61,
|
| 867 |
+
"eval_accuracy": 0.8031108659339165,
|
| 868 |
+
"eval_loss": 0.8653974533081055,
|
| 869 |
+
"eval_runtime": 4177.1307,
|
| 870 |
+
"eval_samples_per_second": 24.703,
|
| 871 |
+
"eval_steps_per_second": 3.088,
|
| 872 |
+
"step": 21284
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 6.68,
|
| 876 |
+
"learning_rate": 3.692201518288475e-05,
|
| 877 |
+
"loss": 0.9059,
|
| 878 |
+
"step": 21500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 6.71,
|
| 882 |
+
"eval_accuracy": 0.803630800811672,
|
| 883 |
+
"eval_loss": 0.8622693419456482,
|
| 884 |
+
"eval_runtime": 4177.9531,
|
| 885 |
+
"eval_samples_per_second": 24.698,
|
| 886 |
+
"eval_steps_per_second": 3.087,
|
| 887 |
+
"step": 21597
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 6.8,
|
| 891 |
+
"eval_accuracy": 0.8044488106784644,
|
| 892 |
+
"eval_loss": 0.8577073812484741,
|
| 893 |
+
"eval_runtime": 4179.042,
|
| 894 |
+
"eval_samples_per_second": 24.692,
|
| 895 |
+
"eval_steps_per_second": 3.087,
|
| 896 |
+
"step": 21910
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 6.83,
|
| 900 |
+
"learning_rate": 3.519668737060042e-05,
|
| 901 |
+
"loss": 0.9,
|
| 902 |
+
"step": 22000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 6.9,
|
| 906 |
+
"eval_accuracy": 0.8051737023740322,
|
| 907 |
+
"eval_loss": 0.8543989062309265,
|
| 908 |
+
"eval_runtime": 4184.397,
|
| 909 |
+
"eval_samples_per_second": 24.66,
|
| 910 |
+
"eval_steps_per_second": 3.083,
|
| 911 |
+
"step": 22223
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 6.99,
|
| 915 |
+
"learning_rate": 3.347135955831608e-05,
|
| 916 |
+
"loss": 0.8938,
|
| 917 |
+
"step": 22500
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 7.0,
|
| 921 |
+
"eval_accuracy": 0.8057043385627138,
|
| 922 |
+
"eval_loss": 0.8503552079200745,
|
| 923 |
+
"eval_runtime": 4183.5003,
|
| 924 |
+
"eval_samples_per_second": 24.665,
|
| 925 |
+
"eval_steps_per_second": 3.083,
|
| 926 |
+
"step": 22536
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 7.1,
|
| 930 |
+
"eval_accuracy": 0.8066939882795919,
|
| 931 |
+
"eval_loss": 0.8455684781074524,
|
| 932 |
+
"eval_runtime": 4184.5066,
|
| 933 |
+
"eval_samples_per_second": 24.659,
|
| 934 |
+
"eval_steps_per_second": 3.083,
|
| 935 |
+
"step": 22849
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 7.14,
|
| 939 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 940 |
+
"loss": 0.8863,
|
| 941 |
+
"step": 23000
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 7.19,
|
| 945 |
+
"eval_accuracy": 0.8068218962399752,
|
| 946 |
+
"eval_loss": 0.8439931273460388,
|
| 947 |
+
"eval_runtime": 4178.7482,
|
| 948 |
+
"eval_samples_per_second": 24.693,
|
| 949 |
+
"eval_steps_per_second": 3.087,
|
| 950 |
+
"step": 23162
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 7.29,
|
| 954 |
+
"eval_accuracy": 0.8075830543867027,
|
| 955 |
+
"eval_loss": 0.8405274152755737,
|
| 956 |
+
"eval_runtime": 4175.082,
|
| 957 |
+
"eval_samples_per_second": 24.715,
|
| 958 |
+
"eval_steps_per_second": 3.09,
|
| 959 |
+
"step": 23475
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 7.3,
|
| 963 |
+
"learning_rate": 3.0020703933747414e-05,
|
| 964 |
+
"loss": 0.8804,
|
| 965 |
+
"step": 23500
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 7.39,
|
| 969 |
+
"eval_accuracy": 0.8083236702044754,
|
| 970 |
+
"eval_loss": 0.8358407616615295,
|
| 971 |
+
"eval_runtime": 4173.8003,
|
| 972 |
+
"eval_samples_per_second": 24.723,
|
| 973 |
+
"eval_steps_per_second": 3.09,
|
| 974 |
+
"step": 23788
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 7.45,
|
| 978 |
+
"learning_rate": 2.829537612146308e-05,
|
| 979 |
+
"loss": 0.8743,
|
| 980 |
+
"step": 24000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 7.48,
|
| 984 |
+
"eval_accuracy": 0.8087474534271941,
|
| 985 |
+
"eval_loss": 0.8344011306762695,
|
| 986 |
+
"eval_runtime": 4174.341,
|
| 987 |
+
"eval_samples_per_second": 24.719,
|
| 988 |
+
"eval_steps_per_second": 3.09,
|
| 989 |
+
"step": 24101
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 7.58,
|
| 993 |
+
"eval_accuracy": 0.8092334825727566,
|
| 994 |
+
"eval_loss": 0.8311472535133362,
|
| 995 |
+
"eval_runtime": 4174.6403,
|
| 996 |
+
"eval_samples_per_second": 24.718,
|
| 997 |
+
"eval_steps_per_second": 3.09,
|
| 998 |
+
"step": 24414
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 7.61,
|
| 1002 |
+
"learning_rate": 2.6570048309178748e-05,
|
| 1003 |
+
"loss": 0.8706,
|
| 1004 |
+
"step": 24500
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 7.68,
|
| 1008 |
+
"eval_accuracy": 0.8097197803323306,
|
| 1009 |
+
"eval_loss": 0.8285520672798157,
|
| 1010 |
+
"eval_runtime": 4174.1958,
|
| 1011 |
+
"eval_samples_per_second": 24.72,
|
| 1012 |
+
"eval_steps_per_second": 3.09,
|
| 1013 |
+
"step": 24727
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 7.76,
|
| 1017 |
+
"learning_rate": 2.484472049689441e-05,
|
| 1018 |
+
"loss": 0.8645,
|
| 1019 |
+
"step": 25000
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 7.78,
|
| 1023 |
+
"eval_accuracy": 0.810216617587347,
|
| 1024 |
+
"eval_loss": 0.8253086805343628,
|
| 1025 |
+
"eval_runtime": 4174.8315,
|
| 1026 |
+
"eval_samples_per_second": 24.716,
|
| 1027 |
+
"eval_steps_per_second": 3.09,
|
| 1028 |
+
"step": 25040
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 7.87,
|
| 1032 |
+
"eval_accuracy": 0.8111007973238641,
|
| 1033 |
+
"eval_loss": 0.8211511373519897,
|
| 1034 |
+
"eval_runtime": 4173.759,
|
| 1035 |
+
"eval_samples_per_second": 24.723,
|
| 1036 |
+
"eval_steps_per_second": 3.09,
|
| 1037 |
+
"step": 25353
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 7.92,
|
| 1041 |
+
"learning_rate": 2.311939268461008e-05,
|
| 1042 |
+
"loss": 0.8602,
|
| 1043 |
+
"step": 25500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 7.97,
|
| 1047 |
+
"eval_accuracy": 0.8111680475751303,
|
| 1048 |
+
"eval_loss": 0.8205570578575134,
|
| 1049 |
+
"eval_runtime": 4174.1596,
|
| 1050 |
+
"eval_samples_per_second": 24.72,
|
| 1051 |
+
"eval_steps_per_second": 3.09,
|
| 1052 |
+
"step": 25666
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 8.07,
|
| 1056 |
+
"eval_accuracy": 0.8120529685373788,
|
| 1057 |
+
"eval_loss": 0.8158560991287231,
|
| 1058 |
+
"eval_runtime": 4175.0585,
|
| 1059 |
+
"eval_samples_per_second": 24.715,
|
| 1060 |
+
"eval_steps_per_second": 3.09,
|
| 1061 |
+
"step": 25979
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 8.07,
|
| 1065 |
+
"learning_rate": 2.139406487232574e-05,
|
| 1066 |
+
"loss": 0.8538,
|
| 1067 |
+
"step": 26000
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 8.16,
|
| 1071 |
+
"eval_accuracy": 0.8122099938117417,
|
| 1072 |
+
"eval_loss": 0.814548134803772,
|
| 1073 |
+
"eval_runtime": 4175.6451,
|
| 1074 |
+
"eval_samples_per_second": 24.712,
|
| 1075 |
+
"eval_steps_per_second": 3.089,
|
| 1076 |
+
"step": 26292
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 8.23,
|
| 1080 |
+
"learning_rate": 1.966873706004141e-05,
|
| 1081 |
+
"loss": 0.8482,
|
| 1082 |
+
"step": 26500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 8.26,
|
| 1086 |
+
"eval_accuracy": 0.8130589470835723,
|
| 1087 |
+
"eval_loss": 0.8115074634552002,
|
| 1088 |
+
"eval_runtime": 4173.7083,
|
| 1089 |
+
"eval_samples_per_second": 24.723,
|
| 1090 |
+
"eval_steps_per_second": 3.091,
|
| 1091 |
+
"step": 26605
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 8.36,
|
| 1095 |
+
"eval_accuracy": 0.8134356257072347,
|
| 1096 |
+
"eval_loss": 0.8089998364448547,
|
| 1097 |
+
"eval_runtime": 4174.6745,
|
| 1098 |
+
"eval_samples_per_second": 24.717,
|
| 1099 |
+
"eval_steps_per_second": 3.09,
|
| 1100 |
+
"step": 26918
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 8.38,
|
| 1104 |
+
"learning_rate": 1.7943409247757076e-05,
|
| 1105 |
+
"loss": 0.8488,
|
| 1106 |
+
"step": 27000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 8.46,
|
| 1110 |
+
"eval_accuracy": 0.8134433401076823,
|
| 1111 |
+
"eval_loss": 0.8088431358337402,
|
| 1112 |
+
"eval_runtime": 4174.8389,
|
| 1113 |
+
"eval_samples_per_second": 24.716,
|
| 1114 |
+
"eval_steps_per_second": 3.09,
|
| 1115 |
+
"step": 27231
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 8.54,
|
| 1119 |
+
"learning_rate": 1.621808143547274e-05,
|
| 1120 |
+
"loss": 0.8423,
|
| 1121 |
+
"step": 27500
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 8.55,
|
| 1125 |
+
"eval_accuracy": 0.8137619275792424,
|
| 1126 |
+
"eval_loss": 0.8057170510292053,
|
| 1127 |
+
"eval_runtime": 4176.6965,
|
| 1128 |
+
"eval_samples_per_second": 24.705,
|
| 1129 |
+
"eval_steps_per_second": 3.088,
|
| 1130 |
+
"step": 27544
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 8.65,
|
| 1134 |
+
"eval_accuracy": 0.8144356879652452,
|
| 1135 |
+
"eval_loss": 0.8039630651473999,
|
| 1136 |
+
"eval_runtime": 4175.7093,
|
| 1137 |
+
"eval_samples_per_second": 24.711,
|
| 1138 |
+
"eval_steps_per_second": 3.089,
|
| 1139 |
+
"step": 27857
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 8.69,
|
| 1143 |
+
"learning_rate": 1.4492753623188407e-05,
|
| 1144 |
+
"loss": 0.8396,
|
| 1145 |
+
"step": 28000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 8.75,
|
| 1149 |
+
"eval_accuracy": 0.8145382455874637,
|
| 1150 |
+
"eval_loss": 0.8026472926139832,
|
| 1151 |
+
"eval_runtime": 4180.1354,
|
| 1152 |
+
"eval_samples_per_second": 24.685,
|
| 1153 |
+
"eval_steps_per_second": 3.086,
|
| 1154 |
+
"step": 28170
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 8.84,
|
| 1158 |
+
"eval_accuracy": 0.8152832629079341,
|
| 1159 |
+
"eval_loss": 0.7988529205322266,
|
| 1160 |
+
"eval_runtime": 4179.9334,
|
| 1161 |
+
"eval_samples_per_second": 24.686,
|
| 1162 |
+
"eval_steps_per_second": 3.086,
|
| 1163 |
+
"step": 28483
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 8.85,
|
| 1167 |
+
"learning_rate": 1.276742581090407e-05,
|
| 1168 |
+
"loss": 0.8377,
|
| 1169 |
+
"step": 28500
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 8.94,
|
| 1173 |
+
"eval_accuracy": 0.8155252212471398,
|
| 1174 |
+
"eval_loss": 0.7979427576065063,
|
| 1175 |
+
"eval_runtime": 4178.4294,
|
| 1176 |
+
"eval_samples_per_second": 24.695,
|
| 1177 |
+
"eval_steps_per_second": 3.087,
|
| 1178 |
+
"step": 28796
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 9.01,
|
| 1182 |
+
"learning_rate": 1.1042097998619738e-05,
|
| 1183 |
+
"loss": 0.8319,
|
| 1184 |
+
"step": 29000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 9.04,
|
| 1188 |
+
"eval_accuracy": 0.8159056442908116,
|
| 1189 |
+
"eval_loss": 0.7969533801078796,
|
| 1190 |
+
"eval_runtime": 4174.229,
|
| 1191 |
+
"eval_samples_per_second": 24.72,
|
| 1192 |
+
"eval_steps_per_second": 3.09,
|
| 1193 |
+
"step": 29109
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 9.14,
|
| 1197 |
+
"eval_accuracy": 0.816157387607754,
|
| 1198 |
+
"eval_loss": 0.7945725917816162,
|
| 1199 |
+
"eval_runtime": 4172.4325,
|
| 1200 |
+
"eval_samples_per_second": 24.731,
|
| 1201 |
+
"eval_steps_per_second": 3.091,
|
| 1202 |
+
"step": 29422
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 9.16,
|
| 1206 |
+
"learning_rate": 9.316770186335403e-06,
|
| 1207 |
+
"loss": 0.8262,
|
| 1208 |
+
"step": 29500
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 9.23,
|
| 1212 |
+
"eval_accuracy": 0.8161445368180659,
|
| 1213 |
+
"eval_loss": 0.7940195202827454,
|
| 1214 |
+
"eval_runtime": 4171.791,
|
| 1215 |
+
"eval_samples_per_second": 24.734,
|
| 1216 |
+
"eval_steps_per_second": 3.092,
|
| 1217 |
+
"step": 29735
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 9.32,
|
| 1221 |
+
"learning_rate": 7.591442374051071e-06,
|
| 1222 |
+
"loss": 0.8255,
|
| 1223 |
+
"step": 30000
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 9.33,
|
| 1227 |
+
"eval_accuracy": 0.8167732822776922,
|
| 1228 |
+
"eval_loss": 0.7918646931648254,
|
| 1229 |
+
"eval_runtime": 4178.9618,
|
| 1230 |
+
"eval_samples_per_second": 24.692,
|
| 1231 |
+
"eval_steps_per_second": 3.087,
|
| 1232 |
+
"step": 30048
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 9.43,
|
| 1236 |
+
"eval_accuracy": 0.8168878384273093,
|
| 1237 |
+
"eval_loss": 0.7914555072784424,
|
| 1238 |
+
"eval_runtime": 4176.3937,
|
| 1239 |
+
"eval_samples_per_second": 24.707,
|
| 1240 |
+
"eval_steps_per_second": 3.089,
|
| 1241 |
+
"step": 30361
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 9.47,
|
| 1245 |
+
"learning_rate": 5.866114561766736e-06,
|
| 1246 |
+
"loss": 0.824,
|
| 1247 |
+
"step": 30500
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 9.53,
|
| 1251 |
+
"eval_accuracy": 0.817306118390898,
|
| 1252 |
+
"eval_loss": 0.7896197438240051,
|
| 1253 |
+
"eval_runtime": 4176.9927,
|
| 1254 |
+
"eval_samples_per_second": 24.704,
|
| 1255 |
+
"eval_steps_per_second": 3.088,
|
| 1256 |
+
"step": 30674
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 9.62,
|
| 1260 |
+
"eval_accuracy": 0.8175610442353789,
|
| 1261 |
+
"eval_loss": 0.7872260212898254,
|
| 1262 |
+
"eval_runtime": 4173.6592,
|
| 1263 |
+
"eval_samples_per_second": 24.723,
|
| 1264 |
+
"eval_steps_per_second": 3.091,
|
| 1265 |
+
"step": 30987
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 9.63,
|
| 1269 |
+
"learning_rate": 4.140786749482402e-06,
|
| 1270 |
+
"loss": 0.8218,
|
| 1271 |
+
"step": 31000
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 9.72,
|
| 1275 |
+
"eval_accuracy": 0.8176242146063594,
|
| 1276 |
+
"eval_loss": 0.7876725792884827,
|
| 1277 |
+
"eval_runtime": 4174.4186,
|
| 1278 |
+
"eval_samples_per_second": 24.719,
|
| 1279 |
+
"eval_steps_per_second": 3.09,
|
| 1280 |
+
"step": 31300
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 9.78,
|
| 1284 |
+
"learning_rate": 2.4154589371980677e-06,
|
| 1285 |
+
"loss": 0.8204,
|
| 1286 |
+
"step": 31500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 9.82,
|
| 1290 |
+
"eval_accuracy": 0.8176433057924909,
|
| 1291 |
+
"eval_loss": 0.7873775959014893,
|
| 1292 |
+
"eval_runtime": 4173.4522,
|
| 1293 |
+
"eval_samples_per_second": 24.725,
|
| 1294 |
+
"eval_steps_per_second": 3.091,
|
| 1295 |
+
"step": 31613
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 9.91,
|
| 1299 |
+
"eval_accuracy": 0.8180999846708549,
|
| 1300 |
+
"eval_loss": 0.7846992611885071,
|
| 1301 |
+
"eval_runtime": 4174.7361,
|
| 1302 |
+
"eval_samples_per_second": 24.717,
|
| 1303 |
+
"eval_steps_per_second": 3.09,
|
| 1304 |
+
"step": 31926
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 9.94,
|
| 1308 |
+
"learning_rate": 6.901311249137336e-07,
|
| 1309 |
+
"loss": 0.8177,
|
| 1310 |
+
"step": 32000
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 10.0,
|
| 1314 |
+
"step": 32200,
|
| 1315 |
+
"total_flos": 7.683327485823698e+18,
|
| 1316 |
+
"train_loss": 1.915723304985473,
|
| 1317 |
+
"train_runtime": 1372706.2896,
|
| 1318 |
+
"train_samples_per_second": 6.006,
|
| 1319 |
+
"train_steps_per_second": 0.023
|
| 1320 |
+
}
|
| 1321 |
+
],
|
| 1322 |
+
"max_steps": 32200,
|
| 1323 |
+
"num_train_epochs": 10,
|
| 1324 |
+
"total_flos": 7.683327485823698e+18,
|
| 1325 |
+
"trial_name": null,
|
| 1326 |
+
"trial_params": null
|
| 1327 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d707fe5411bc278612fcbbab1a8bd257aa8d713dbd43dd10985ce0d0c74c44c5
|
| 3 |
+
size 3963
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|