kiddothe2b
commited on
Commit
•
5443746
1
Parent(s):
dbd6e53
500k training steps with 128 tokens
Browse files- README.md +70 -1
- all_results.json +12 -0
- config.json +27 -0
- eval_results.json +8 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- train_results.json +7 -0
- trainer_state.json +3105 -0
- training_args.bin +3 -0
README.md
CHANGED
|
@@ -1,3 +1,72 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
tags:
|
| 3 |
+
- generated_from_trainer
|
| 4 |
+
datasets:
|
| 5 |
+
- custom_legal_danish_corpus
|
| 6 |
+
model-index:
|
| 7 |
+
- name: danish-lex-lm-base-mlm
|
| 8 |
+
results: []
|
| 9 |
---
|
| 10 |
+
|
| 11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
+
|
| 14 |
+
# danish-lex-lm-base-mlm
|
| 15 |
+
|
| 16 |
+
This model is a fine-tuned version of [data/PLMs/danish-lm/danish-lex-lm-base](https://huggingface.co/data/PLMs/danish-lm/danish-lex-lm-base) on the custom_legal_danish_corpus dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.7302
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 0.0001
|
| 38 |
+
- train_batch_size: 16
|
| 39 |
+
- eval_batch_size: 16
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: tpu
|
| 42 |
+
- num_devices: 8
|
| 43 |
+
- gradient_accumulation_steps: 2
|
| 44 |
+
- total_train_batch_size: 256
|
| 45 |
+
- total_eval_batch_size: 128
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.05
|
| 49 |
+
- training_steps: 500000
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
+
|:-------------:|:-----:|:------:|:---------------:|
|
| 55 |
+
| 1.4648 | 5.36 | 50000 | 1.2920 |
|
| 56 |
+
| 1.2165 | 10.72 | 100000 | 1.0625 |
|
| 57 |
+
| 1.0952 | 16.07 | 150000 | 0.9611 |
|
| 58 |
+
| 1.0233 | 21.43 | 200000 | 0.8931 |
|
| 59 |
+
| 0.963 | 26.79 | 250000 | 0.8477 |
|
| 60 |
+
| 0.9122 | 32.15 | 300000 | 0.8168 |
|
| 61 |
+
| 0.8697 | 37.51 | 350000 | 0.7836 |
|
| 62 |
+
| 0.8397 | 42.86 | 400000 | 0.7560 |
|
| 63 |
+
| 0.8231 | 48.22 | 450000 | 0.7476 |
|
| 64 |
+
| 0.8207 | 53.58 | 500000 | 0.7243 |
|
| 65 |
+
|
| 66 |
+
|
| 67 |
+
### Framework versions
|
| 68 |
+
|
| 69 |
+
- Transformers 4.18.0
|
| 70 |
+
- Pytorch 1.12.0+cu102
|
| 71 |
+
- Datasets 2.0.0
|
| 72 |
+
- Tokenizers 0.12.0
|
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 53.58,
|
| 3 |
+
"eval_loss": 0.7302267551422119,
|
| 4 |
+
"eval_runtime": 3.6523,
|
| 5 |
+
"eval_samples_per_second": 2738.025,
|
| 6 |
+
"eval_steps_per_second": 21.63,
|
| 7 |
+
"perplexity": 2.0755511962245876,
|
| 8 |
+
"train_loss": 1.193239736328125,
|
| 9 |
+
"train_runtime": 102111.9968,
|
| 10 |
+
"train_samples_per_second": 1253.526,
|
| 11 |
+
"train_steps_per_second": 4.897
|
| 12 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "data/PLMs/danish-lm/danish-lex-lm-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"pad_token_id": 0,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.18.0",
|
| 24 |
+
"type_vocab_size": 1,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 32000
|
| 27 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 53.58,
|
| 3 |
+
"eval_loss": 0.7302267551422119,
|
| 4 |
+
"eval_runtime": 3.6523,
|
| 5 |
+
"eval_samples_per_second": 2738.025,
|
| 6 |
+
"eval_steps_per_second": 21.63,
|
| 7 |
+
"perplexity": 2.0755511962245876
|
| 8 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:45679ead40348bea4447e1d2c33cde39cb39fd41d7f5c2571f2291863d0115cd
|
| 3 |
+
size 442675755
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": "<mask>"}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"model_max_length": 512, "bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": "<mask>", "special_tokens_map_file": "data/PLMs/danish-lm/danish-lex-lm-base/special_tokens_map.json", "name_or_path": "data/PLMs/danish-lm/danish-lex-lm-base", "tokenizer_class": "PreTrainedTokenizerFast"}
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 53.58,
|
| 3 |
+
"train_loss": 1.193239736328125,
|
| 4 |
+
"train_runtime": 102111.9968,
|
| 5 |
+
"train_samples_per_second": 1253.526,
|
| 6 |
+
"train_steps_per_second": 4.897
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 53.57908272610373,
|
| 5 |
+
"global_step": 500000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.11,
|
| 12 |
+
"learning_rate": 4.000000000000001e-06,
|
| 13 |
+
"loss": 8.8501,
|
| 14 |
+
"step": 1000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.21,
|
| 18 |
+
"learning_rate": 8.000000000000001e-06,
|
| 19 |
+
"loss": 7.0526,
|
| 20 |
+
"step": 2000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.32,
|
| 24 |
+
"learning_rate": 1.2e-05,
|
| 25 |
+
"loss": 6.6396,
|
| 26 |
+
"step": 3000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.43,
|
| 30 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 31 |
+
"loss": 6.457,
|
| 32 |
+
"step": 4000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.54,
|
| 36 |
+
"learning_rate": 2e-05,
|
| 37 |
+
"loss": 6.3073,
|
| 38 |
+
"step": 5000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.64,
|
| 42 |
+
"learning_rate": 2.4e-05,
|
| 43 |
+
"loss": 6.1934,
|
| 44 |
+
"step": 6000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.75,
|
| 48 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 49 |
+
"loss": 6.0929,
|
| 50 |
+
"step": 7000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.86,
|
| 54 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 55 |
+
"loss": 6.0097,
|
| 56 |
+
"step": 8000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.96,
|
| 60 |
+
"learning_rate": 3.6e-05,
|
| 61 |
+
"loss": 5.9222,
|
| 62 |
+
"step": 9000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.07,
|
| 66 |
+
"learning_rate": 4e-05,
|
| 67 |
+
"loss": 5.8487,
|
| 68 |
+
"step": 10000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.18,
|
| 72 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 73 |
+
"loss": 5.7752,
|
| 74 |
+
"step": 11000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.29,
|
| 78 |
+
"learning_rate": 4.8e-05,
|
| 79 |
+
"loss": 5.4387,
|
| 80 |
+
"step": 12000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.39,
|
| 84 |
+
"learning_rate": 5.2000000000000004e-05,
|
| 85 |
+
"loss": 4.7713,
|
| 86 |
+
"step": 13000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.5,
|
| 90 |
+
"learning_rate": 5.6000000000000006e-05,
|
| 91 |
+
"loss": 3.9954,
|
| 92 |
+
"step": 14000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.61,
|
| 96 |
+
"learning_rate": 6e-05,
|
| 97 |
+
"loss": 3.4334,
|
| 98 |
+
"step": 15000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.71,
|
| 102 |
+
"learning_rate": 6.400000000000001e-05,
|
| 103 |
+
"loss": 3.0948,
|
| 104 |
+
"step": 16000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.82,
|
| 108 |
+
"learning_rate": 6.800000000000001e-05,
|
| 109 |
+
"loss": 2.835,
|
| 110 |
+
"step": 17000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.93,
|
| 114 |
+
"learning_rate": 7.2e-05,
|
| 115 |
+
"loss": 2.6402,
|
| 116 |
+
"step": 18000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.04,
|
| 120 |
+
"learning_rate": 7.6e-05,
|
| 121 |
+
"loss": 2.4955,
|
| 122 |
+
"step": 19000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 2.14,
|
| 126 |
+
"learning_rate": 8e-05,
|
| 127 |
+
"loss": 2.3712,
|
| 128 |
+
"step": 20000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.25,
|
| 132 |
+
"learning_rate": 8.4e-05,
|
| 133 |
+
"loss": 2.2723,
|
| 134 |
+
"step": 21000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.36,
|
| 138 |
+
"learning_rate": 8.800000000000001e-05,
|
| 139 |
+
"loss": 2.1948,
|
| 140 |
+
"step": 22000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.46,
|
| 144 |
+
"learning_rate": 9.200000000000001e-05,
|
| 145 |
+
"loss": 2.1228,
|
| 146 |
+
"step": 23000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.57,
|
| 150 |
+
"learning_rate": 9.6e-05,
|
| 151 |
+
"loss": 2.0669,
|
| 152 |
+
"step": 24000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.68,
|
| 156 |
+
"learning_rate": 0.0001,
|
| 157 |
+
"loss": 2.0114,
|
| 158 |
+
"step": 25000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.79,
|
| 162 |
+
"learning_rate": 9.999890641901125e-05,
|
| 163 |
+
"loss": 1.9638,
|
| 164 |
+
"step": 26000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.89,
|
| 168 |
+
"learning_rate": 9.99956257238817e-05,
|
| 169 |
+
"loss": 1.9188,
|
| 170 |
+
"step": 27000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.0,
|
| 174 |
+
"learning_rate": 9.999015805811965e-05,
|
| 175 |
+
"loss": 1.8771,
|
| 176 |
+
"step": 28000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.11,
|
| 180 |
+
"learning_rate": 9.998250366089848e-05,
|
| 181 |
+
"loss": 1.8311,
|
| 182 |
+
"step": 29000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.21,
|
| 186 |
+
"learning_rate": 9.997266286704631e-05,
|
| 187 |
+
"loss": 1.8029,
|
| 188 |
+
"step": 30000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.32,
|
| 192 |
+
"learning_rate": 9.996063610703137e-05,
|
| 193 |
+
"loss": 1.7731,
|
| 194 |
+
"step": 31000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.43,
|
| 198 |
+
"learning_rate": 9.994642390694308e-05,
|
| 199 |
+
"loss": 1.7475,
|
| 200 |
+
"step": 32000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.54,
|
| 204 |
+
"learning_rate": 9.993002688846913e-05,
|
| 205 |
+
"loss": 1.7192,
|
| 206 |
+
"step": 33000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.64,
|
| 210 |
+
"learning_rate": 9.991144576886823e-05,
|
| 211 |
+
"loss": 1.6999,
|
| 212 |
+
"step": 34000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 9.989068136093873e-05,
|
| 217 |
+
"loss": 1.6824,
|
| 218 |
+
"step": 35000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.86,
|
| 222 |
+
"learning_rate": 9.986773457298311e-05,
|
| 223 |
+
"loss": 1.6575,
|
| 224 |
+
"step": 36000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.96,
|
| 228 |
+
"learning_rate": 9.984260640876821e-05,
|
| 229 |
+
"loss": 1.6386,
|
| 230 |
+
"step": 37000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.07,
|
| 234 |
+
"learning_rate": 9.981529796748134e-05,
|
| 235 |
+
"loss": 1.6187,
|
| 236 |
+
"step": 38000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.18,
|
| 240 |
+
"learning_rate": 9.97858104436822e-05,
|
| 241 |
+
"loss": 1.6023,
|
| 242 |
+
"step": 39000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 4.29,
|
| 246 |
+
"learning_rate": 9.975414512725057e-05,
|
| 247 |
+
"loss": 1.5847,
|
| 248 |
+
"step": 40000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.39,
|
| 252 |
+
"learning_rate": 9.972030340333001e-05,
|
| 253 |
+
"loss": 1.57,
|
| 254 |
+
"step": 41000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 4.5,
|
| 258 |
+
"learning_rate": 9.968428675226714e-05,
|
| 259 |
+
"loss": 1.5602,
|
| 260 |
+
"step": 42000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.61,
|
| 264 |
+
"learning_rate": 9.964609674954696e-05,
|
| 265 |
+
"loss": 1.5471,
|
| 266 |
+
"step": 43000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 4.71,
|
| 270 |
+
"learning_rate": 9.96057350657239e-05,
|
| 271 |
+
"loss": 1.536,
|
| 272 |
+
"step": 44000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 4.82,
|
| 276 |
+
"learning_rate": 9.956320346634876e-05,
|
| 277 |
+
"loss": 1.5244,
|
| 278 |
+
"step": 45000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.93,
|
| 282 |
+
"learning_rate": 9.95185038118915e-05,
|
| 283 |
+
"loss": 1.5115,
|
| 284 |
+
"step": 46000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 5.04,
|
| 288 |
+
"learning_rate": 9.94716380576598e-05,
|
| 289 |
+
"loss": 1.4992,
|
| 290 |
+
"step": 47000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 5.14,
|
| 294 |
+
"learning_rate": 9.942260825371358e-05,
|
| 295 |
+
"loss": 1.4833,
|
| 296 |
+
"step": 48000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.25,
|
| 300 |
+
"learning_rate": 9.937141654477528e-05,
|
| 301 |
+
"loss": 1.4766,
|
| 302 |
+
"step": 49000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 5.36,
|
| 306 |
+
"learning_rate": 9.931806517013612e-05,
|
| 307 |
+
"loss": 1.4648,
|
| 308 |
+
"step": 50000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 5.36,
|
| 312 |
+
"eval_loss": 1.2920496463775635,
|
| 313 |
+
"eval_runtime": 9.2098,
|
| 314 |
+
"eval_samples_per_second": 1085.8,
|
| 315 |
+
"eval_steps_per_second": 8.578,
|
| 316 |
+
"step": 50000
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.47,
|
| 320 |
+
"learning_rate": 9.926255646355804e-05,
|
| 321 |
+
"loss": 1.458,
|
| 322 |
+
"step": 51000
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.57,
|
| 326 |
+
"learning_rate": 9.92048928531717e-05,
|
| 327 |
+
"loss": 1.4494,
|
| 328 |
+
"step": 52000
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.68,
|
| 332 |
+
"learning_rate": 9.914507686137019e-05,
|
| 333 |
+
"loss": 1.446,
|
| 334 |
+
"step": 53000
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.79,
|
| 338 |
+
"learning_rate": 9.90831111046988e-05,
|
| 339 |
+
"loss": 1.4337,
|
| 340 |
+
"step": 54000
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 5.89,
|
| 344 |
+
"learning_rate": 9.901899829374047e-05,
|
| 345 |
+
"loss": 1.4313,
|
| 346 |
+
"step": 55000
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"learning_rate": 9.895274123299723e-05,
|
| 351 |
+
"loss": 1.419,
|
| 352 |
+
"step": 56000
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.11,
|
| 356 |
+
"learning_rate": 9.888434282076758e-05,
|
| 357 |
+
"loss": 1.4027,
|
| 358 |
+
"step": 57000
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 6.22,
|
| 362 |
+
"learning_rate": 9.881380604901964e-05,
|
| 363 |
+
"loss": 1.3996,
|
| 364 |
+
"step": 58000
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 6.32,
|
| 368 |
+
"learning_rate": 9.87411340032603e-05,
|
| 369 |
+
"loss": 1.395,
|
| 370 |
+
"step": 59000
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.43,
|
| 374 |
+
"learning_rate": 9.86663298624003e-05,
|
| 375 |
+
"loss": 1.3883,
|
| 376 |
+
"step": 60000
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 6.54,
|
| 380 |
+
"learning_rate": 9.858939689861506e-05,
|
| 381 |
+
"loss": 1.3806,
|
| 382 |
+
"step": 61000
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 6.64,
|
| 386 |
+
"learning_rate": 9.851033847720166e-05,
|
| 387 |
+
"loss": 1.3773,
|
| 388 |
+
"step": 62000
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 6.75,
|
| 392 |
+
"learning_rate": 9.842915805643155e-05,
|
| 393 |
+
"loss": 1.3697,
|
| 394 |
+
"step": 63000
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.86,
|
| 398 |
+
"learning_rate": 9.834585918739936e-05,
|
| 399 |
+
"loss": 1.367,
|
| 400 |
+
"step": 64000
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 6.97,
|
| 404 |
+
"learning_rate": 9.826044551386744e-05,
|
| 405 |
+
"loss": 1.3617,
|
| 406 |
+
"step": 65000
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 7.07,
|
| 410 |
+
"learning_rate": 9.817292077210659e-05,
|
| 411 |
+
"loss": 1.3513,
|
| 412 |
+
"step": 66000
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.18,
|
| 416 |
+
"learning_rate": 9.808328879073251e-05,
|
| 417 |
+
"loss": 1.3407,
|
| 418 |
+
"step": 67000
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 7.29,
|
| 422 |
+
"learning_rate": 9.799155349053851e-05,
|
| 423 |
+
"loss": 1.3358,
|
| 424 |
+
"step": 68000
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 7.39,
|
| 428 |
+
"learning_rate": 9.789771888432375e-05,
|
| 429 |
+
"loss": 1.3302,
|
| 430 |
+
"step": 69000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.5,
|
| 434 |
+
"learning_rate": 9.780178907671789e-05,
|
| 435 |
+
"loss": 1.3295,
|
| 436 |
+
"step": 70000
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 7.61,
|
| 440 |
+
"learning_rate": 9.77037682640015e-05,
|
| 441 |
+
"loss": 1.3249,
|
| 442 |
+
"step": 71000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 7.72,
|
| 446 |
+
"learning_rate": 9.760366073392246e-05,
|
| 447 |
+
"loss": 1.3204,
|
| 448 |
+
"step": 72000
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 7.82,
|
| 452 |
+
"learning_rate": 9.750147086550844e-05,
|
| 453 |
+
"loss": 1.3185,
|
| 454 |
+
"step": 73000
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.93,
|
| 458 |
+
"learning_rate": 9.739720312887535e-05,
|
| 459 |
+
"loss": 1.3126,
|
| 460 |
+
"step": 74000
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 8.04,
|
| 464 |
+
"learning_rate": 9.729086208503174e-05,
|
| 465 |
+
"loss": 1.3051,
|
| 466 |
+
"step": 75000
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 8.14,
|
| 470 |
+
"learning_rate": 9.718245238567939e-05,
|
| 471 |
+
"loss": 1.2962,
|
| 472 |
+
"step": 76000
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.25,
|
| 476 |
+
"learning_rate": 9.707197877300974e-05,
|
| 477 |
+
"loss": 1.2906,
|
| 478 |
+
"step": 77000
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.36,
|
| 482 |
+
"learning_rate": 9.695944607949649e-05,
|
| 483 |
+
"loss": 1.2888,
|
| 484 |
+
"step": 78000
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.47,
|
| 488 |
+
"learning_rate": 9.684485922768422e-05,
|
| 489 |
+
"loss": 1.2903,
|
| 490 |
+
"step": 79000
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.57,
|
| 494 |
+
"learning_rate": 9.672822322997305e-05,
|
| 495 |
+
"loss": 1.2837,
|
| 496 |
+
"step": 80000
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.68,
|
| 500 |
+
"learning_rate": 9.660954318839933e-05,
|
| 501 |
+
"loss": 1.2823,
|
| 502 |
+
"step": 81000
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.79,
|
| 506 |
+
"learning_rate": 9.648882429441257e-05,
|
| 507 |
+
"loss": 1.2746,
|
| 508 |
+
"step": 82000
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.89,
|
| 512 |
+
"learning_rate": 9.636607182864827e-05,
|
| 513 |
+
"loss": 1.2721,
|
| 514 |
+
"step": 83000
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 9.624129116069694e-05,
|
| 519 |
+
"loss": 1.2696,
|
| 520 |
+
"step": 84000
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.11,
|
| 524 |
+
"learning_rate": 9.611448774886924e-05,
|
| 525 |
+
"loss": 1.2589,
|
| 526 |
+
"step": 85000
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 9.22,
|
| 530 |
+
"learning_rate": 9.598566713995718e-05,
|
| 531 |
+
"loss": 1.255,
|
| 532 |
+
"step": 86000
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 9.32,
|
| 536 |
+
"learning_rate": 9.58548349689915e-05,
|
| 537 |
+
"loss": 1.2547,
|
| 538 |
+
"step": 87000
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 9.43,
|
| 542 |
+
"learning_rate": 9.572199695899522e-05,
|
| 543 |
+
"loss": 1.2502,
|
| 544 |
+
"step": 88000
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 9.54,
|
| 548 |
+
"learning_rate": 9.558715892073323e-05,
|
| 549 |
+
"loss": 1.2513,
|
| 550 |
+
"step": 89000
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.64,
|
| 554 |
+
"learning_rate": 9.545032675245813e-05,
|
| 555 |
+
"loss": 1.2441,
|
| 556 |
+
"step": 90000
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 9.75,
|
| 560 |
+
"learning_rate": 9.531150643965223e-05,
|
| 561 |
+
"loss": 1.2447,
|
| 562 |
+
"step": 91000
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 9.86,
|
| 566 |
+
"learning_rate": 9.517070405476575e-05,
|
| 567 |
+
"loss": 1.2386,
|
| 568 |
+
"step": 92000
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 9.97,
|
| 572 |
+
"learning_rate": 9.502792575695112e-05,
|
| 573 |
+
"loss": 1.2364,
|
| 574 |
+
"step": 93000
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 10.07,
|
| 578 |
+
"learning_rate": 9.488317779179361e-05,
|
| 579 |
+
"loss": 1.227,
|
| 580 |
+
"step": 94000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 10.18,
|
| 584 |
+
"learning_rate": 9.473646649103818e-05,
|
| 585 |
+
"loss": 1.2196,
|
| 586 |
+
"step": 95000
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 10.29,
|
| 590 |
+
"learning_rate": 9.458779827231237e-05,
|
| 591 |
+
"loss": 1.2226,
|
| 592 |
+
"step": 96000
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 10.39,
|
| 596 |
+
"learning_rate": 9.443717963884569e-05,
|
| 597 |
+
"loss": 1.2182,
|
| 598 |
+
"step": 97000
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 10.5,
|
| 602 |
+
"learning_rate": 9.428461717918511e-05,
|
| 603 |
+
"loss": 1.2235,
|
| 604 |
+
"step": 98000
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 10.61,
|
| 608 |
+
"learning_rate": 9.413011756690685e-05,
|
| 609 |
+
"loss": 1.2172,
|
| 610 |
+
"step": 99000
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 10.72,
|
| 614 |
+
"learning_rate": 9.397368756032445e-05,
|
| 615 |
+
"loss": 1.2165,
|
| 616 |
+
"step": 100000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 10.72,
|
| 620 |
+
"eval_loss": 1.0625041723251343,
|
| 621 |
+
"eval_runtime": 3.5569,
|
| 622 |
+
"eval_samples_per_second": 2811.465,
|
| 623 |
+
"eval_steps_per_second": 22.211,
|
| 624 |
+
"step": 100000
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.82,
|
| 628 |
+
"learning_rate": 9.381533400219318e-05,
|
| 629 |
+
"loss": 1.2112,
|
| 630 |
+
"step": 101000
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 10.93,
|
| 634 |
+
"learning_rate": 9.365506381941066e-05,
|
| 635 |
+
"loss": 1.2081,
|
| 636 |
+
"step": 102000
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 11.04,
|
| 640 |
+
"learning_rate": 9.349288402271388e-05,
|
| 641 |
+
"loss": 1.2042,
|
| 642 |
+
"step": 103000
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 11.14,
|
| 646 |
+
"learning_rate": 9.332880170637252e-05,
|
| 647 |
+
"loss": 1.1959,
|
| 648 |
+
"step": 104000
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 11.25,
|
| 652 |
+
"learning_rate": 9.316282404787871e-05,
|
| 653 |
+
"loss": 1.197,
|
| 654 |
+
"step": 105000
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 11.36,
|
| 658 |
+
"learning_rate": 9.299495830763286e-05,
|
| 659 |
+
"loss": 1.1942,
|
| 660 |
+
"step": 106000
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 11.47,
|
| 664 |
+
"learning_rate": 9.282521182862629e-05,
|
| 665 |
+
"loss": 1.1893,
|
| 666 |
+
"step": 107000
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 11.57,
|
| 670 |
+
"learning_rate": 9.265359203611987e-05,
|
| 671 |
+
"loss": 1.1894,
|
| 672 |
+
"step": 108000
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 11.68,
|
| 676 |
+
"learning_rate": 9.248010643731935e-05,
|
| 677 |
+
"loss": 1.1883,
|
| 678 |
+
"step": 109000
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 11.79,
|
| 682 |
+
"learning_rate": 9.230476262104677e-05,
|
| 683 |
+
"loss": 1.1897,
|
| 684 |
+
"step": 110000
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 11.89,
|
| 688 |
+
"learning_rate": 9.212756825740873e-05,
|
| 689 |
+
"loss": 1.1865,
|
| 690 |
+
"step": 111000
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 12.0,
|
| 694 |
+
"learning_rate": 9.194853109746074e-05,
|
| 695 |
+
"loss": 1.1825,
|
| 696 |
+
"step": 112000
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 12.11,
|
| 700 |
+
"learning_rate": 9.176765897286813e-05,
|
| 701 |
+
"loss": 1.1734,
|
| 702 |
+
"step": 113000
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 12.22,
|
| 706 |
+
"learning_rate": 9.158495979556358e-05,
|
| 707 |
+
"loss": 1.1681,
|
| 708 |
+
"step": 114000
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 12.32,
|
| 712 |
+
"learning_rate": 9.140044155740101e-05,
|
| 713 |
+
"loss": 1.1722,
|
| 714 |
+
"step": 115000
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 12.43,
|
| 718 |
+
"learning_rate": 9.121411232980588e-05,
|
| 719 |
+
"loss": 1.1689,
|
| 720 |
+
"step": 116000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 12.54,
|
| 724 |
+
"learning_rate": 9.102598026342222e-05,
|
| 725 |
+
"loss": 1.1681,
|
| 726 |
+
"step": 117000
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 12.64,
|
| 730 |
+
"learning_rate": 9.083605358775612e-05,
|
| 731 |
+
"loss": 1.1621,
|
| 732 |
+
"step": 118000
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 12.75,
|
| 736 |
+
"learning_rate": 9.064434061081562e-05,
|
| 737 |
+
"loss": 1.1648,
|
| 738 |
+
"step": 119000
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 12.86,
|
| 742 |
+
"learning_rate": 9.045084971874738e-05,
|
| 743 |
+
"loss": 1.1645,
|
| 744 |
+
"step": 120000
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 12.97,
|
| 748 |
+
"learning_rate": 9.025558937546988e-05,
|
| 749 |
+
"loss": 1.1624,
|
| 750 |
+
"step": 121000
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 13.07,
|
| 754 |
+
"learning_rate": 9.005856812230304e-05,
|
| 755 |
+
"loss": 1.1526,
|
| 756 |
+
"step": 122000
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 13.18,
|
| 760 |
+
"learning_rate": 8.98597945775948e-05,
|
| 761 |
+
"loss": 1.15,
|
| 762 |
+
"step": 123000
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 13.29,
|
| 766 |
+
"learning_rate": 8.965927743634391e-05,
|
| 767 |
+
"loss": 1.147,
|
| 768 |
+
"step": 124000
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 13.39,
|
| 772 |
+
"learning_rate": 8.945702546981969e-05,
|
| 773 |
+
"loss": 1.1507,
|
| 774 |
+
"step": 125000
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 13.5,
|
| 778 |
+
"learning_rate": 8.92530475251784e-05,
|
| 779 |
+
"loss": 1.1458,
|
| 780 |
+
"step": 126000
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 13.61,
|
| 784 |
+
"learning_rate": 8.90473525250761e-05,
|
| 785 |
+
"loss": 1.148,
|
| 786 |
+
"step": 127000
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 13.72,
|
| 790 |
+
"learning_rate": 8.883994946727849e-05,
|
| 791 |
+
"loss": 1.1441,
|
| 792 |
+
"step": 128000
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 13.82,
|
| 796 |
+
"learning_rate": 8.863084742426719e-05,
|
| 797 |
+
"loss": 1.1442,
|
| 798 |
+
"step": 129000
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 13.93,
|
| 802 |
+
"learning_rate": 8.842005554284296e-05,
|
| 803 |
+
"loss": 1.1413,
|
| 804 |
+
"step": 130000
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 14.04,
|
| 808 |
+
"learning_rate": 8.820758304372557e-05,
|
| 809 |
+
"loss": 1.1346,
|
| 810 |
+
"step": 131000
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 14.14,
|
| 814 |
+
"learning_rate": 8.799343922115044e-05,
|
| 815 |
+
"loss": 1.129,
|
| 816 |
+
"step": 132000
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 14.25,
|
| 820 |
+
"learning_rate": 8.77776334424621e-05,
|
| 821 |
+
"loss": 1.1293,
|
| 822 |
+
"step": 133000
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 14.36,
|
| 826 |
+
"learning_rate": 8.756017514770443e-05,
|
| 827 |
+
"loss": 1.1257,
|
| 828 |
+
"step": 134000
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 14.47,
|
| 832 |
+
"learning_rate": 8.73410738492077e-05,
|
| 833 |
+
"loss": 1.1296,
|
| 834 |
+
"step": 135000
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 14.57,
|
| 838 |
+
"learning_rate": 8.71203391311725e-05,
|
| 839 |
+
"loss": 1.1265,
|
| 840 |
+
"step": 136000
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 14.68,
|
| 844 |
+
"learning_rate": 8.689798064925049e-05,
|
| 845 |
+
"loss": 1.1283,
|
| 846 |
+
"step": 137000
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 14.79,
|
| 850 |
+
"learning_rate": 8.6674008130122e-05,
|
| 851 |
+
"loss": 1.1211,
|
| 852 |
+
"step": 138000
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 14.89,
|
| 856 |
+
"learning_rate": 8.644843137107059e-05,
|
| 857 |
+
"loss": 1.1206,
|
| 858 |
+
"step": 139000
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 15.0,
|
| 862 |
+
"learning_rate": 8.622126023955446e-05,
|
| 863 |
+
"loss": 1.1223,
|
| 864 |
+
"step": 140000
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 15.11,
|
| 868 |
+
"learning_rate": 8.599250467277483e-05,
|
| 869 |
+
"loss": 1.1146,
|
| 870 |
+
"step": 141000
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 15.22,
|
| 874 |
+
"learning_rate": 8.576217467724128e-05,
|
| 875 |
+
"loss": 1.1136,
|
| 876 |
+
"step": 142000
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 15.32,
|
| 880 |
+
"learning_rate": 8.553028032833397e-05,
|
| 881 |
+
"loss": 1.1111,
|
| 882 |
+
"step": 143000
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 15.43,
|
| 886 |
+
"learning_rate": 8.529683176986295e-05,
|
| 887 |
+
"loss": 1.1128,
|
| 888 |
+
"step": 144000
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 15.54,
|
| 892 |
+
"learning_rate": 8.506183921362443e-05,
|
| 893 |
+
"loss": 1.1126,
|
| 894 |
+
"step": 145000
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 15.65,
|
| 898 |
+
"learning_rate": 8.482531293895412e-05,
|
| 899 |
+
"loss": 1.1063,
|
| 900 |
+
"step": 146000
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 15.75,
|
| 904 |
+
"learning_rate": 8.458726329227747e-05,
|
| 905 |
+
"loss": 1.1081,
|
| 906 |
+
"step": 147000
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 15.86,
|
| 910 |
+
"learning_rate": 8.434770068665723e-05,
|
| 911 |
+
"loss": 1.1057,
|
| 912 |
+
"step": 148000
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 15.97,
|
| 916 |
+
"learning_rate": 8.410663560133784e-05,
|
| 917 |
+
"loss": 1.1041,
|
| 918 |
+
"step": 149000
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 16.07,
|
| 922 |
+
"learning_rate": 8.386407858128706e-05,
|
| 923 |
+
"loss": 1.0952,
|
| 924 |
+
"step": 150000
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 16.07,
|
| 928 |
+
"eval_loss": 0.9611279368400574,
|
| 929 |
+
"eval_runtime": 3.5721,
|
| 930 |
+
"eval_samples_per_second": 2799.479,
|
| 931 |
+
"eval_steps_per_second": 22.116,
|
| 932 |
+
"step": 150000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 16.18,
|
| 936 |
+
"learning_rate": 8.362004023673474e-05,
|
| 937 |
+
"loss": 1.0963,
|
| 938 |
+
"step": 151000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 16.29,
|
| 942 |
+
"learning_rate": 8.337453124270863e-05,
|
| 943 |
+
"loss": 1.0939,
|
| 944 |
+
"step": 152000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 16.4,
|
| 948 |
+
"learning_rate": 8.31275623385675e-05,
|
| 949 |
+
"loss": 1.0942,
|
| 950 |
+
"step": 153000
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 16.5,
|
| 954 |
+
"learning_rate": 8.287914432753123e-05,
|
| 955 |
+
"loss": 1.0929,
|
| 956 |
+
"step": 154000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 16.61,
|
| 960 |
+
"learning_rate": 8.262928807620843e-05,
|
| 961 |
+
"loss": 1.0931,
|
| 962 |
+
"step": 155000
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 16.72,
|
| 966 |
+
"learning_rate": 8.237800451412095e-05,
|
| 967 |
+
"loss": 1.0893,
|
| 968 |
+
"step": 156000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 16.82,
|
| 972 |
+
"learning_rate": 8.212530463322583e-05,
|
| 973 |
+
"loss": 1.0904,
|
| 974 |
+
"step": 157000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 16.93,
|
| 978 |
+
"learning_rate": 8.18711994874345e-05,
|
| 979 |
+
"loss": 1.0879,
|
| 980 |
+
"step": 158000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 17.04,
|
| 984 |
+
"learning_rate": 8.161570019212921e-05,
|
| 985 |
+
"loss": 1.084,
|
| 986 |
+
"step": 159000
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 17.15,
|
| 990 |
+
"learning_rate": 8.135881792367686e-05,
|
| 991 |
+
"loss": 1.081,
|
| 992 |
+
"step": 160000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 17.25,
|
| 996 |
+
"learning_rate": 8.110056391894005e-05,
|
| 997 |
+
"loss": 1.0787,
|
| 998 |
+
"step": 161000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 17.36,
|
| 1002 |
+
"learning_rate": 8.084094947478556e-05,
|
| 1003 |
+
"loss": 1.0776,
|
| 1004 |
+
"step": 162000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 17.47,
|
| 1008 |
+
"learning_rate": 8.057998594759022e-05,
|
| 1009 |
+
"loss": 1.0782,
|
| 1010 |
+
"step": 163000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 17.57,
|
| 1014 |
+
"learning_rate": 8.031768475274413e-05,
|
| 1015 |
+
"loss": 1.076,
|
| 1016 |
+
"step": 164000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 17.68,
|
| 1020 |
+
"learning_rate": 8.005405736415126e-05,
|
| 1021 |
+
"loss": 1.0744,
|
| 1022 |
+
"step": 165000
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 17.79,
|
| 1026 |
+
"learning_rate": 7.978911531372765e-05,
|
| 1027 |
+
"loss": 1.0757,
|
| 1028 |
+
"step": 166000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 17.9,
|
| 1032 |
+
"learning_rate": 7.952287019089685e-05,
|
| 1033 |
+
"loss": 1.0768,
|
| 1034 |
+
"step": 167000
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 18.0,
|
| 1038 |
+
"learning_rate": 7.925533364208309e-05,
|
| 1039 |
+
"loss": 1.0705,
|
| 1040 |
+
"step": 168000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 18.11,
|
| 1044 |
+
"learning_rate": 7.898651737020166e-05,
|
| 1045 |
+
"loss": 1.0634,
|
| 1046 |
+
"step": 169000
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 18.22,
|
| 1050 |
+
"learning_rate": 7.871643313414718e-05,
|
| 1051 |
+
"loss": 1.0654,
|
| 1052 |
+
"step": 170000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 18.32,
|
| 1056 |
+
"learning_rate": 7.844509274827907e-05,
|
| 1057 |
+
"loss": 1.0634,
|
| 1058 |
+
"step": 171000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 18.43,
|
| 1062 |
+
"learning_rate": 7.817250808190483e-05,
|
| 1063 |
+
"loss": 1.0643,
|
| 1064 |
+
"step": 172000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 18.54,
|
| 1068 |
+
"learning_rate": 7.789869105876083e-05,
|
| 1069 |
+
"loss": 1.0625,
|
| 1070 |
+
"step": 173000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 18.65,
|
| 1074 |
+
"learning_rate": 7.762365365649067e-05,
|
| 1075 |
+
"loss": 1.0637,
|
| 1076 |
+
"step": 174000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 18.75,
|
| 1080 |
+
"learning_rate": 7.734740790612136e-05,
|
| 1081 |
+
"loss": 1.0619,
|
| 1082 |
+
"step": 175000
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 18.86,
|
| 1086 |
+
"learning_rate": 7.70699658915369e-05,
|
| 1087 |
+
"loss": 1.0625,
|
| 1088 |
+
"step": 176000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 18.97,
|
| 1092 |
+
"learning_rate": 7.679133974894983e-05,
|
| 1093 |
+
"loss": 1.0582,
|
| 1094 |
+
"step": 177000
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 19.07,
|
| 1098 |
+
"learning_rate": 7.651154166637025e-05,
|
| 1099 |
+
"loss": 1.0488,
|
| 1100 |
+
"step": 178000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 19.18,
|
| 1104 |
+
"learning_rate": 7.623058388307269e-05,
|
| 1105 |
+
"loss": 1.0516,
|
| 1106 |
+
"step": 179000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 19.29,
|
| 1110 |
+
"learning_rate": 7.594847868906076e-05,
|
| 1111 |
+
"loss": 1.0497,
|
| 1112 |
+
"step": 180000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 19.4,
|
| 1116 |
+
"learning_rate": 7.566523842452958e-05,
|
| 1117 |
+
"loss": 1.0517,
|
| 1118 |
+
"step": 181000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 19.5,
|
| 1122 |
+
"learning_rate": 7.538087547932585e-05,
|
| 1123 |
+
"loss": 1.0523,
|
| 1124 |
+
"step": 182000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 19.61,
|
| 1128 |
+
"learning_rate": 7.509540229240601e-05,
|
| 1129 |
+
"loss": 1.0444,
|
| 1130 |
+
"step": 183000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 19.72,
|
| 1134 |
+
"learning_rate": 7.480883135129211e-05,
|
| 1135 |
+
"loss": 1.0482,
|
| 1136 |
+
"step": 184000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 19.82,
|
| 1140 |
+
"learning_rate": 7.452117519152542e-05,
|
| 1141 |
+
"loss": 1.0453,
|
| 1142 |
+
"step": 185000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 19.93,
|
| 1146 |
+
"learning_rate": 7.423244639611826e-05,
|
| 1147 |
+
"loss": 1.0433,
|
| 1148 |
+
"step": 186000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 20.04,
|
| 1152 |
+
"learning_rate": 7.394265759500348e-05,
|
| 1153 |
+
"loss": 1.0397,
|
| 1154 |
+
"step": 187000
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 20.15,
|
| 1158 |
+
"learning_rate": 7.365182146448205e-05,
|
| 1159 |
+
"loss": 1.0356,
|
| 1160 |
+
"step": 188000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 20.25,
|
| 1164 |
+
"learning_rate": 7.335995072666848e-05,
|
| 1165 |
+
"loss": 1.0386,
|
| 1166 |
+
"step": 189000
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 20.36,
|
| 1170 |
+
"learning_rate": 7.30670581489344e-05,
|
| 1171 |
+
"loss": 1.0348,
|
| 1172 |
+
"step": 190000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 20.47,
|
| 1176 |
+
"learning_rate": 7.277315654334997e-05,
|
| 1177 |
+
"loss": 1.0372,
|
| 1178 |
+
"step": 191000
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 20.57,
|
| 1182 |
+
"learning_rate": 7.247825876612353e-05,
|
| 1183 |
+
"loss": 1.0352,
|
| 1184 |
+
"step": 192000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 20.68,
|
| 1188 |
+
"learning_rate": 7.218237771703921e-05,
|
| 1189 |
+
"loss": 1.0351,
|
| 1190 |
+
"step": 193000
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 20.79,
|
| 1194 |
+
"learning_rate": 7.188552633889259e-05,
|
| 1195 |
+
"loss": 1.0346,
|
| 1196 |
+
"step": 194000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 20.9,
|
| 1200 |
+
"learning_rate": 7.158771761692464e-05,
|
| 1201 |
+
"loss": 1.031,
|
| 1202 |
+
"step": 195000
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 21.0,
|
| 1206 |
+
"learning_rate": 7.128896457825364e-05,
|
| 1207 |
+
"loss": 1.0327,
|
| 1208 |
+
"step": 196000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 21.11,
|
| 1212 |
+
"learning_rate": 7.09892802913053e-05,
|
| 1213 |
+
"loss": 1.0235,
|
| 1214 |
+
"step": 197000
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 21.22,
|
| 1218 |
+
"learning_rate": 7.068867786524116e-05,
|
| 1219 |
+
"loss": 1.0244,
|
| 1220 |
+
"step": 198000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 21.32,
|
| 1224 |
+
"learning_rate": 7.038717044938519e-05,
|
| 1225 |
+
"loss": 1.024,
|
| 1226 |
+
"step": 199000
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 21.43,
|
| 1230 |
+
"learning_rate": 7.008477123264848e-05,
|
| 1231 |
+
"loss": 1.0233,
|
| 1232 |
+
"step": 200000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 21.43,
|
| 1236 |
+
"eval_loss": 0.8930683732032776,
|
| 1237 |
+
"eval_runtime": 3.6639,
|
| 1238 |
+
"eval_samples_per_second": 2729.338,
|
| 1239 |
+
"eval_steps_per_second": 21.562,
|
| 1240 |
+
"step": 200000
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 21.54,
|
| 1244 |
+
"learning_rate": 6.978149344295242e-05,
|
| 1245 |
+
"loss": 1.0246,
|
| 1246 |
+
"step": 201000
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 21.65,
|
| 1250 |
+
"learning_rate": 6.947735034665002e-05,
|
| 1251 |
+
"loss": 1.024,
|
| 1252 |
+
"step": 202000
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 21.75,
|
| 1256 |
+
"learning_rate": 6.917235524794558e-05,
|
| 1257 |
+
"loss": 1.0221,
|
| 1258 |
+
"step": 203000
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 21.86,
|
| 1262 |
+
"learning_rate": 6.886652148831279e-05,
|
| 1263 |
+
"loss": 1.0196,
|
| 1264 |
+
"step": 204000
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 21.97,
|
| 1268 |
+
"learning_rate": 6.855986244591104e-05,
|
| 1269 |
+
"loss": 1.0175,
|
| 1270 |
+
"step": 205000
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 22.07,
|
| 1274 |
+
"learning_rate": 6.825239153500029e-05,
|
| 1275 |
+
"loss": 1.0148,
|
| 1276 |
+
"step": 206000
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 22.18,
|
| 1280 |
+
"learning_rate": 6.794412220535426e-05,
|
| 1281 |
+
"loss": 1.0104,
|
| 1282 |
+
"step": 207000
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 22.29,
|
| 1286 |
+
"learning_rate": 6.763506794167208e-05,
|
| 1287 |
+
"loss": 1.0129,
|
| 1288 |
+
"step": 208000
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 22.4,
|
| 1292 |
+
"learning_rate": 6.732524226298841e-05,
|
| 1293 |
+
"loss": 1.0107,
|
| 1294 |
+
"step": 209000
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 22.5,
|
| 1298 |
+
"learning_rate": 6.701465872208216e-05,
|
| 1299 |
+
"loss": 1.0067,
|
| 1300 |
+
"step": 210000
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 22.61,
|
| 1304 |
+
"learning_rate": 6.670333090488356e-05,
|
| 1305 |
+
"loss": 1.0115,
|
| 1306 |
+
"step": 211000
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 22.72,
|
| 1310 |
+
"learning_rate": 6.639127242987988e-05,
|
| 1311 |
+
"loss": 1.0115,
|
| 1312 |
+
"step": 212000
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 22.82,
|
| 1316 |
+
"learning_rate": 6.607849694751977e-05,
|
| 1317 |
+
"loss": 1.0124,
|
| 1318 |
+
"step": 213000
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 22.93,
|
| 1322 |
+
"learning_rate": 6.576501813961609e-05,
|
| 1323 |
+
"loss": 1.0047,
|
| 1324 |
+
"step": 214000
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 23.04,
|
| 1328 |
+
"learning_rate": 6.545084971874738e-05,
|
| 1329 |
+
"loss": 1.0056,
|
| 1330 |
+
"step": 215000
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 23.15,
|
| 1334 |
+
"learning_rate": 6.513600542765817e-05,
|
| 1335 |
+
"loss": 1.0023,
|
| 1336 |
+
"step": 216000
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 23.25,
|
| 1340 |
+
"learning_rate": 6.48204990386577e-05,
|
| 1341 |
+
"loss": 1.0016,
|
| 1342 |
+
"step": 217000
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 23.36,
|
| 1346 |
+
"learning_rate": 6.450434435301751e-05,
|
| 1347 |
+
"loss": 0.9974,
|
| 1348 |
+
"step": 218000
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 23.47,
|
| 1352 |
+
"learning_rate": 6.418755520036775e-05,
|
| 1353 |
+
"loss": 0.9979,
|
| 1354 |
+
"step": 219000
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 23.57,
|
| 1358 |
+
"learning_rate": 6.387014543809223e-05,
|
| 1359 |
+
"loss": 0.9999,
|
| 1360 |
+
"step": 220000
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 23.68,
|
| 1364 |
+
"learning_rate": 6.355212895072223e-05,
|
| 1365 |
+
"loss": 0.9965,
|
| 1366 |
+
"step": 221000
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 23.79,
|
| 1370 |
+
"learning_rate": 6.323351964932908e-05,
|
| 1371 |
+
"loss": 0.9979,
|
| 1372 |
+
"step": 222000
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 23.9,
|
| 1376 |
+
"learning_rate": 6.291433147091583e-05,
|
| 1377 |
+
"loss": 0.9948,
|
| 1378 |
+
"step": 223000
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 24.0,
|
| 1382 |
+
"learning_rate": 6.259457837780742e-05,
|
| 1383 |
+
"loss": 0.9938,
|
| 1384 |
+
"step": 224000
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 24.11,
|
| 1388 |
+
"learning_rate": 6.227427435703997e-05,
|
| 1389 |
+
"loss": 0.9888,
|
| 1390 |
+
"step": 225000
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 24.22,
|
| 1394 |
+
"learning_rate": 6.195343341974899e-05,
|
| 1395 |
+
"loss": 0.9886,
|
| 1396 |
+
"step": 226000
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 24.32,
|
| 1400 |
+
"learning_rate": 6.163206960055651e-05,
|
| 1401 |
+
"loss": 0.9878,
|
| 1402 |
+
"step": 227000
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 24.43,
|
| 1406 |
+
"learning_rate": 6.131019695695702e-05,
|
| 1407 |
+
"loss": 0.985,
|
| 1408 |
+
"step": 228000
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 24.54,
|
| 1412 |
+
"learning_rate": 6.0987829568702656e-05,
|
| 1413 |
+
"loss": 0.9861,
|
| 1414 |
+
"step": 229000
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 24.65,
|
| 1418 |
+
"learning_rate": 6.066498153718735e-05,
|
| 1419 |
+
"loss": 0.9883,
|
| 1420 |
+
"step": 230000
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 24.75,
|
| 1424 |
+
"learning_rate": 6.034166698482984e-05,
|
| 1425 |
+
"loss": 0.9854,
|
| 1426 |
+
"step": 231000
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 24.86,
|
| 1430 |
+
"learning_rate": 6.001790005445607e-05,
|
| 1431 |
+
"loss": 0.9864,
|
| 1432 |
+
"step": 232000
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 24.97,
|
| 1436 |
+
"learning_rate": 5.969369490868042e-05,
|
| 1437 |
+
"loss": 0.9869,
|
| 1438 |
+
"step": 233000
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 25.08,
|
| 1442 |
+
"learning_rate": 5.9369065729286245e-05,
|
| 1443 |
+
"loss": 0.9796,
|
| 1444 |
+
"step": 234000
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 25.18,
|
| 1448 |
+
"learning_rate": 5.90440267166055e-05,
|
| 1449 |
+
"loss": 0.978,
|
| 1450 |
+
"step": 235000
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 25.29,
|
| 1454 |
+
"learning_rate": 5.871859208889759e-05,
|
| 1455 |
+
"loss": 0.9777,
|
| 1456 |
+
"step": 236000
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 25.4,
|
| 1460 |
+
"learning_rate": 5.8392776081727385e-05,
|
| 1461 |
+
"loss": 0.9759,
|
| 1462 |
+
"step": 237000
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 25.5,
|
| 1466 |
+
"learning_rate": 5.8066592947342555e-05,
|
| 1467 |
+
"loss": 0.9801,
|
| 1468 |
+
"step": 238000
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 25.61,
|
| 1472 |
+
"learning_rate": 5.7740056954050084e-05,
|
| 1473 |
+
"loss": 0.9785,
|
| 1474 |
+
"step": 239000
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 25.72,
|
| 1478 |
+
"learning_rate": 5.74131823855921e-05,
|
| 1479 |
+
"loss": 0.9735,
|
| 1480 |
+
"step": 240000
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 25.83,
|
| 1484 |
+
"learning_rate": 5.7085983540521216e-05,
|
| 1485 |
+
"loss": 0.9753,
|
| 1486 |
+
"step": 241000
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 25.93,
|
| 1490 |
+
"learning_rate": 5.675847473157485e-05,
|
| 1491 |
+
"loss": 0.9762,
|
| 1492 |
+
"step": 242000
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 26.04,
|
| 1496 |
+
"learning_rate": 5.6430670285049314e-05,
|
| 1497 |
+
"loss": 0.969,
|
| 1498 |
+
"step": 243000
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 26.15,
|
| 1502 |
+
"learning_rate": 5.6102584540173006e-05,
|
| 1503 |
+
"loss": 0.9661,
|
| 1504 |
+
"step": 244000
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 26.25,
|
| 1508 |
+
"learning_rate": 5.577423184847932e-05,
|
| 1509 |
+
"loss": 0.965,
|
| 1510 |
+
"step": 245000
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 26.36,
|
| 1514 |
+
"learning_rate": 5.544562657317863e-05,
|
| 1515 |
+
"loss": 0.965,
|
| 1516 |
+
"step": 246000
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 26.47,
|
| 1520 |
+
"learning_rate": 5.511678308853026e-05,
|
| 1521 |
+
"loss": 0.9665,
|
| 1522 |
+
"step": 247000
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 26.58,
|
| 1526 |
+
"learning_rate": 5.478771577921351e-05,
|
| 1527 |
+
"loss": 0.9618,
|
| 1528 |
+
"step": 248000
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 26.68,
|
| 1532 |
+
"learning_rate": 5.445843903969854e-05,
|
| 1533 |
+
"loss": 0.968,
|
| 1534 |
+
"step": 249000
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 26.79,
|
| 1538 |
+
"learning_rate": 5.4128967273616625e-05,
|
| 1539 |
+
"loss": 0.963,
|
| 1540 |
+
"step": 250000
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 26.79,
|
| 1544 |
+
"eval_loss": 0.8477376103401184,
|
| 1545 |
+
"eval_runtime": 3.6783,
|
| 1546 |
+
"eval_samples_per_second": 2718.681,
|
| 1547 |
+
"eval_steps_per_second": 21.478,
|
| 1548 |
+
"step": 250000
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 26.9,
|
| 1552 |
+
"learning_rate": 5.379931489313016e-05,
|
| 1553 |
+
"loss": 0.9618,
|
| 1554 |
+
"step": 251000
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 27.0,
|
| 1558 |
+
"learning_rate": 5.3469496318302204e-05,
|
| 1559 |
+
"loss": 0.9646,
|
| 1560 |
+
"step": 252000
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 27.11,
|
| 1564 |
+
"learning_rate": 5.313952597646568e-05,
|
| 1565 |
+
"loss": 0.955,
|
| 1566 |
+
"step": 253000
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 27.22,
|
| 1570 |
+
"learning_rate": 5.280941830159227e-05,
|
| 1571 |
+
"loss": 0.953,
|
| 1572 |
+
"step": 254000
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 27.33,
|
| 1576 |
+
"learning_rate": 5.247918773366112e-05,
|
| 1577 |
+
"loss": 0.9562,
|
| 1578 |
+
"step": 255000
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 27.43,
|
| 1582 |
+
"learning_rate": 5.214884871802703e-05,
|
| 1583 |
+
"loss": 0.9564,
|
| 1584 |
+
"step": 256000
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 27.54,
|
| 1588 |
+
"learning_rate": 5.1818415704788725e-05,
|
| 1589 |
+
"loss": 0.9559,
|
| 1590 |
+
"step": 257000
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 27.65,
|
| 1594 |
+
"learning_rate": 5.148790314815663e-05,
|
| 1595 |
+
"loss": 0.957,
|
| 1596 |
+
"step": 258000
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 27.75,
|
| 1600 |
+
"learning_rate": 5.1157325505820694e-05,
|
| 1601 |
+
"loss": 0.9576,
|
| 1602 |
+
"step": 259000
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 27.86,
|
| 1606 |
+
"learning_rate": 5.0826697238317935e-05,
|
| 1607 |
+
"loss": 0.9539,
|
| 1608 |
+
"step": 260000
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 27.97,
|
| 1612 |
+
"learning_rate": 5.0496032808399815e-05,
|
| 1613 |
+
"loss": 0.9522,
|
| 1614 |
+
"step": 261000
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 28.08,
|
| 1618 |
+
"learning_rate": 5.016534668039976e-05,
|
| 1619 |
+
"loss": 0.9492,
|
| 1620 |
+
"step": 262000
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 28.18,
|
| 1624 |
+
"learning_rate": 4.9834653319600246e-05,
|
| 1625 |
+
"loss": 0.9456,
|
| 1626 |
+
"step": 263000
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 28.29,
|
| 1630 |
+
"learning_rate": 4.950396719160018e-05,
|
| 1631 |
+
"loss": 0.9452,
|
| 1632 |
+
"step": 264000
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 28.4,
|
| 1636 |
+
"learning_rate": 4.917330276168208e-05,
|
| 1637 |
+
"loss": 0.9462,
|
| 1638 |
+
"step": 265000
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 28.5,
|
| 1642 |
+
"learning_rate": 4.884267449417931e-05,
|
| 1643 |
+
"loss": 0.9448,
|
| 1644 |
+
"step": 266000
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 28.61,
|
| 1648 |
+
"learning_rate": 4.851209685184338e-05,
|
| 1649 |
+
"loss": 0.9447,
|
| 1650 |
+
"step": 267000
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 28.72,
|
| 1654 |
+
"learning_rate": 4.818158429521129e-05,
|
| 1655 |
+
"loss": 0.9445,
|
| 1656 |
+
"step": 268000
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 28.83,
|
| 1660 |
+
"learning_rate": 4.785115128197298e-05,
|
| 1661 |
+
"loss": 0.9431,
|
| 1662 |
+
"step": 269000
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 28.93,
|
| 1666 |
+
"learning_rate": 4.7520812266338885e-05,
|
| 1667 |
+
"loss": 0.9447,
|
| 1668 |
+
"step": 270000
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 29.04,
|
| 1672 |
+
"learning_rate": 4.7190581698407725e-05,
|
| 1673 |
+
"loss": 0.9404,
|
| 1674 |
+
"step": 271000
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 29.15,
|
| 1678 |
+
"learning_rate": 4.6860474023534335e-05,
|
| 1679 |
+
"loss": 0.937,
|
| 1680 |
+
"step": 272000
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 29.25,
|
| 1684 |
+
"learning_rate": 4.65305036816978e-05,
|
| 1685 |
+
"loss": 0.9381,
|
| 1686 |
+
"step": 273000
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 29.36,
|
| 1690 |
+
"learning_rate": 4.620068510686985e-05,
|
| 1691 |
+
"loss": 0.9389,
|
| 1692 |
+
"step": 274000
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 29.47,
|
| 1696 |
+
"learning_rate": 4.5871032726383386e-05,
|
| 1697 |
+
"loss": 0.9342,
|
| 1698 |
+
"step": 275000
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 29.58,
|
| 1702 |
+
"learning_rate": 4.554156096030149e-05,
|
| 1703 |
+
"loss": 0.9332,
|
| 1704 |
+
"step": 276000
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 29.68,
|
| 1708 |
+
"learning_rate": 4.5212284220786494e-05,
|
| 1709 |
+
"loss": 0.9361,
|
| 1710 |
+
"step": 277000
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 29.79,
|
| 1714 |
+
"learning_rate": 4.488321691146975e-05,
|
| 1715 |
+
"loss": 0.9342,
|
| 1716 |
+
"step": 278000
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 29.9,
|
| 1720 |
+
"learning_rate": 4.4554373426821374e-05,
|
| 1721 |
+
"loss": 0.9341,
|
| 1722 |
+
"step": 279000
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 30.0,
|
| 1726 |
+
"learning_rate": 4.4225768151520694e-05,
|
| 1727 |
+
"loss": 0.9331,
|
| 1728 |
+
"step": 280000
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 30.11,
|
| 1732 |
+
"learning_rate": 4.3897415459827e-05,
|
| 1733 |
+
"loss": 0.9289,
|
| 1734 |
+
"step": 281000
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 30.22,
|
| 1738 |
+
"learning_rate": 4.3569329714950704e-05,
|
| 1739 |
+
"loss": 0.9279,
|
| 1740 |
+
"step": 282000
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 30.33,
|
| 1744 |
+
"learning_rate": 4.324152526842517e-05,
|
| 1745 |
+
"loss": 0.9297,
|
| 1746 |
+
"step": 283000
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 30.43,
|
| 1750 |
+
"learning_rate": 4.291401645947879e-05,
|
| 1751 |
+
"loss": 0.9254,
|
| 1752 |
+
"step": 284000
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 30.54,
|
| 1756 |
+
"learning_rate": 4.2586817614407895e-05,
|
| 1757 |
+
"loss": 0.925,
|
| 1758 |
+
"step": 285000
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 30.65,
|
| 1762 |
+
"learning_rate": 4.2259943045949934e-05,
|
| 1763 |
+
"loss": 0.9247,
|
| 1764 |
+
"step": 286000
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 30.75,
|
| 1768 |
+
"learning_rate": 4.1933407052657456e-05,
|
| 1769 |
+
"loss": 0.9272,
|
| 1770 |
+
"step": 287000
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 30.86,
|
| 1774 |
+
"learning_rate": 4.160722391827262e-05,
|
| 1775 |
+
"loss": 0.9247,
|
| 1776 |
+
"step": 288000
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 30.97,
|
| 1780 |
+
"learning_rate": 4.1281407911102425e-05,
|
| 1781 |
+
"loss": 0.9237,
|
| 1782 |
+
"step": 289000
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 31.08,
|
| 1786 |
+
"learning_rate": 4.095597328339452e-05,
|
| 1787 |
+
"loss": 0.9201,
|
| 1788 |
+
"step": 290000
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 31.18,
|
| 1792 |
+
"learning_rate": 4.063093427071376e-05,
|
| 1793 |
+
"loss": 0.9181,
|
| 1794 |
+
"step": 291000
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 31.29,
|
| 1798 |
+
"learning_rate": 4.0306305091319595e-05,
|
| 1799 |
+
"loss": 0.917,
|
| 1800 |
+
"step": 292000
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 31.4,
|
| 1804 |
+
"learning_rate": 3.9982099945543945e-05,
|
| 1805 |
+
"loss": 0.9171,
|
| 1806 |
+
"step": 293000
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 31.5,
|
| 1810 |
+
"learning_rate": 3.965833301517017e-05,
|
| 1811 |
+
"loss": 0.9179,
|
| 1812 |
+
"step": 294000
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 31.61,
|
| 1816 |
+
"learning_rate": 3.933501846281267e-05,
|
| 1817 |
+
"loss": 0.9143,
|
| 1818 |
+
"step": 295000
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 31.72,
|
| 1822 |
+
"learning_rate": 3.901217043129735e-05,
|
| 1823 |
+
"loss": 0.9151,
|
| 1824 |
+
"step": 296000
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 31.83,
|
| 1828 |
+
"learning_rate": 3.8689803043043e-05,
|
| 1829 |
+
"loss": 0.9197,
|
| 1830 |
+
"step": 297000
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 31.93,
|
| 1834 |
+
"learning_rate": 3.836793039944349e-05,
|
| 1835 |
+
"loss": 0.9147,
|
| 1836 |
+
"step": 298000
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 32.04,
|
| 1840 |
+
"learning_rate": 3.8046566580251e-05,
|
| 1841 |
+
"loss": 0.9117,
|
| 1842 |
+
"step": 299000
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 32.15,
|
| 1846 |
+
"learning_rate": 3.772572564296005e-05,
|
| 1847 |
+
"loss": 0.9122,
|
| 1848 |
+
"step": 300000
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 32.15,
|
| 1852 |
+
"eval_loss": 0.8167890906333923,
|
| 1853 |
+
"eval_runtime": 3.704,
|
| 1854 |
+
"eval_samples_per_second": 2699.813,
|
| 1855 |
+
"eval_steps_per_second": 21.329,
|
| 1856 |
+
"step": 300000
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 32.25,
|
| 1860 |
+
"learning_rate": 3.74054216221926e-05,
|
| 1861 |
+
"loss": 0.9088,
|
| 1862 |
+
"step": 301000
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 32.36,
|
| 1866 |
+
"learning_rate": 3.7085668529084184e-05,
|
| 1867 |
+
"loss": 0.9104,
|
| 1868 |
+
"step": 302000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 32.47,
|
| 1872 |
+
"learning_rate": 3.676648035067093e-05,
|
| 1873 |
+
"loss": 0.9114,
|
| 1874 |
+
"step": 303000
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 32.58,
|
| 1878 |
+
"learning_rate": 3.6447871049277796e-05,
|
| 1879 |
+
"loss": 0.9061,
|
| 1880 |
+
"step": 304000
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 32.68,
|
| 1884 |
+
"learning_rate": 3.612985456190778e-05,
|
| 1885 |
+
"loss": 0.9081,
|
| 1886 |
+
"step": 305000
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 32.79,
|
| 1890 |
+
"learning_rate": 3.581244479963225e-05,
|
| 1891 |
+
"loss": 0.9049,
|
| 1892 |
+
"step": 306000
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 32.9,
|
| 1896 |
+
"learning_rate": 3.5495655646982505e-05,
|
| 1897 |
+
"loss": 0.9086,
|
| 1898 |
+
"step": 307000
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 33.0,
|
| 1902 |
+
"learning_rate": 3.517950096134232e-05,
|
| 1903 |
+
"loss": 0.9074,
|
| 1904 |
+
"step": 308000
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 33.11,
|
| 1908 |
+
"learning_rate": 3.4863994572341843e-05,
|
| 1909 |
+
"loss": 0.9008,
|
| 1910 |
+
"step": 309000
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 33.22,
|
| 1914 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 1915 |
+
"loss": 0.9009,
|
| 1916 |
+
"step": 310000
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 33.33,
|
| 1920 |
+
"learning_rate": 3.423498186038393e-05,
|
| 1921 |
+
"loss": 0.9014,
|
| 1922 |
+
"step": 311000
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 33.43,
|
| 1926 |
+
"learning_rate": 3.392150305248024e-05,
|
| 1927 |
+
"loss": 0.9039,
|
| 1928 |
+
"step": 312000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 33.54,
|
| 1932 |
+
"learning_rate": 3.360872757012011e-05,
|
| 1933 |
+
"loss": 0.8986,
|
| 1934 |
+
"step": 313000
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 33.65,
|
| 1938 |
+
"learning_rate": 3.329666909511645e-05,
|
| 1939 |
+
"loss": 0.8996,
|
| 1940 |
+
"step": 314000
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 33.75,
|
| 1944 |
+
"learning_rate": 3.298534127791785e-05,
|
| 1945 |
+
"loss": 0.8973,
|
| 1946 |
+
"step": 315000
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 33.86,
|
| 1950 |
+
"learning_rate": 3.267475773701161e-05,
|
| 1951 |
+
"loss": 0.8984,
|
| 1952 |
+
"step": 316000
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 33.97,
|
| 1956 |
+
"learning_rate": 3.236493205832795e-05,
|
| 1957 |
+
"loss": 0.8967,
|
| 1958 |
+
"step": 317000
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 34.08,
|
| 1962 |
+
"learning_rate": 3.205587779464576e-05,
|
| 1963 |
+
"loss": 0.8951,
|
| 1964 |
+
"step": 318000
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 34.18,
|
| 1968 |
+
"learning_rate": 3.1747608464999725e-05,
|
| 1969 |
+
"loss": 0.8916,
|
| 1970 |
+
"step": 319000
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 34.29,
|
| 1974 |
+
"learning_rate": 3.144013755408895e-05,
|
| 1975 |
+
"loss": 0.8941,
|
| 1976 |
+
"step": 320000
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 34.4,
|
| 1980 |
+
"learning_rate": 3.113347851168721e-05,
|
| 1981 |
+
"loss": 0.8918,
|
| 1982 |
+
"step": 321000
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 34.5,
|
| 1986 |
+
"learning_rate": 3.082764475205442e-05,
|
| 1987 |
+
"loss": 0.8925,
|
| 1988 |
+
"step": 322000
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 34.61,
|
| 1992 |
+
"learning_rate": 3.052264965335e-05,
|
| 1993 |
+
"loss": 0.889,
|
| 1994 |
+
"step": 323000
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 34.72,
|
| 1998 |
+
"learning_rate": 3.0218506557047598e-05,
|
| 1999 |
+
"loss": 0.8916,
|
| 2000 |
+
"step": 324000
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 34.83,
|
| 2004 |
+
"learning_rate": 2.991522876735154e-05,
|
| 2005 |
+
"loss": 0.8928,
|
| 2006 |
+
"step": 325000
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 34.93,
|
| 2010 |
+
"learning_rate": 2.9612829550614836e-05,
|
| 2011 |
+
"loss": 0.8918,
|
| 2012 |
+
"step": 326000
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 35.04,
|
| 2016 |
+
"learning_rate": 2.931132213475884e-05,
|
| 2017 |
+
"loss": 0.8902,
|
| 2018 |
+
"step": 327000
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 35.15,
|
| 2022 |
+
"learning_rate": 2.9010719708694722e-05,
|
| 2023 |
+
"loss": 0.8852,
|
| 2024 |
+
"step": 328000
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 35.26,
|
| 2028 |
+
"learning_rate": 2.8711035421746367e-05,
|
| 2029 |
+
"loss": 0.8843,
|
| 2030 |
+
"step": 329000
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 35.36,
|
| 2034 |
+
"learning_rate": 2.8412282383075363e-05,
|
| 2035 |
+
"loss": 0.886,
|
| 2036 |
+
"step": 330000
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 35.47,
|
| 2040 |
+
"learning_rate": 2.811447366110741e-05,
|
| 2041 |
+
"loss": 0.8837,
|
| 2042 |
+
"step": 331000
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 35.58,
|
| 2046 |
+
"learning_rate": 2.7817622282960815e-05,
|
| 2047 |
+
"loss": 0.885,
|
| 2048 |
+
"step": 332000
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 35.68,
|
| 2052 |
+
"learning_rate": 2.7521741233876496e-05,
|
| 2053 |
+
"loss": 0.8844,
|
| 2054 |
+
"step": 333000
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 35.79,
|
| 2058 |
+
"learning_rate": 2.7226843456650037e-05,
|
| 2059 |
+
"loss": 0.8826,
|
| 2060 |
+
"step": 334000
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 35.9,
|
| 2064 |
+
"learning_rate": 2.693294185106562e-05,
|
| 2065 |
+
"loss": 0.8828,
|
| 2066 |
+
"step": 335000
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 36.01,
|
| 2070 |
+
"learning_rate": 2.6640049273331515e-05,
|
| 2071 |
+
"loss": 0.8807,
|
| 2072 |
+
"step": 336000
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 36.11,
|
| 2076 |
+
"learning_rate": 2.6348178535517966e-05,
|
| 2077 |
+
"loss": 0.8773,
|
| 2078 |
+
"step": 337000
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 36.22,
|
| 2082 |
+
"learning_rate": 2.6057342404996522e-05,
|
| 2083 |
+
"loss": 0.8771,
|
| 2084 |
+
"step": 338000
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 36.33,
|
| 2088 |
+
"learning_rate": 2.5767553603881767e-05,
|
| 2089 |
+
"loss": 0.8762,
|
| 2090 |
+
"step": 339000
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 36.43,
|
| 2094 |
+
"learning_rate": 2.547882480847461e-05,
|
| 2095 |
+
"loss": 0.8746,
|
| 2096 |
+
"step": 340000
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 36.54,
|
| 2100 |
+
"learning_rate": 2.5191168648707887e-05,
|
| 2101 |
+
"loss": 0.8759,
|
| 2102 |
+
"step": 341000
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 36.65,
|
| 2106 |
+
"learning_rate": 2.490459770759398e-05,
|
| 2107 |
+
"loss": 0.8769,
|
| 2108 |
+
"step": 342000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 36.76,
|
| 2112 |
+
"learning_rate": 2.4619124520674146e-05,
|
| 2113 |
+
"loss": 0.8786,
|
| 2114 |
+
"step": 343000
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 36.86,
|
| 2118 |
+
"learning_rate": 2.433476157547044e-05,
|
| 2119 |
+
"loss": 0.8784,
|
| 2120 |
+
"step": 344000
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 36.97,
|
| 2124 |
+
"learning_rate": 2.405152131093926e-05,
|
| 2125 |
+
"loss": 0.877,
|
| 2126 |
+
"step": 345000
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 37.08,
|
| 2130 |
+
"learning_rate": 2.3769416116927335e-05,
|
| 2131 |
+
"loss": 0.8728,
|
| 2132 |
+
"step": 346000
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 37.18,
|
| 2136 |
+
"learning_rate": 2.3488458333629777e-05,
|
| 2137 |
+
"loss": 0.8707,
|
| 2138 |
+
"step": 347000
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 37.29,
|
| 2142 |
+
"learning_rate": 2.3208660251050158e-05,
|
| 2143 |
+
"loss": 0.8703,
|
| 2144 |
+
"step": 348000
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 37.4,
|
| 2148 |
+
"learning_rate": 2.29300341084631e-05,
|
| 2149 |
+
"loss": 0.873,
|
| 2150 |
+
"step": 349000
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 37.51,
|
| 2154 |
+
"learning_rate": 2.2652592093878666e-05,
|
| 2155 |
+
"loss": 0.8697,
|
| 2156 |
+
"step": 350000
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 37.51,
|
| 2160 |
+
"eval_loss": 0.7835835218429565,
|
| 2161 |
+
"eval_runtime": 3.6871,
|
| 2162 |
+
"eval_samples_per_second": 2712.188,
|
| 2163 |
+
"eval_steps_per_second": 21.426,
|
| 2164 |
+
"step": 350000
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 37.61,
|
| 2168 |
+
"learning_rate": 2.237634634350934e-05,
|
| 2169 |
+
"loss": 0.8737,
|
| 2170 |
+
"step": 351000
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 37.72,
|
| 2174 |
+
"learning_rate": 2.2101308941239203e-05,
|
| 2175 |
+
"loss": 0.8695,
|
| 2176 |
+
"step": 352000
|
| 2177 |
+
},
|
| 2178 |
+
{
|
| 2179 |
+
"epoch": 37.83,
|
| 2180 |
+
"learning_rate": 2.182749191809518e-05,
|
| 2181 |
+
"loss": 0.8693,
|
| 2182 |
+
"step": 353000
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 37.93,
|
| 2186 |
+
"learning_rate": 2.1554907251720945e-05,
|
| 2187 |
+
"loss": 0.8692,
|
| 2188 |
+
"step": 354000
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 38.04,
|
| 2192 |
+
"learning_rate": 2.128356686585282e-05,
|
| 2193 |
+
"loss": 0.868,
|
| 2194 |
+
"step": 355000
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 38.15,
|
| 2198 |
+
"learning_rate": 2.1013482629798333e-05,
|
| 2199 |
+
"loss": 0.8634,
|
| 2200 |
+
"step": 356000
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 38.26,
|
| 2204 |
+
"learning_rate": 2.0744666357916925e-05,
|
| 2205 |
+
"loss": 0.8643,
|
| 2206 |
+
"step": 357000
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 38.36,
|
| 2210 |
+
"learning_rate": 2.0477129809103147e-05,
|
| 2211 |
+
"loss": 0.8639,
|
| 2212 |
+
"step": 358000
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 38.47,
|
| 2216 |
+
"learning_rate": 2.0210884686272368e-05,
|
| 2217 |
+
"loss": 0.8648,
|
| 2218 |
+
"step": 359000
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 38.58,
|
| 2222 |
+
"learning_rate": 1.9945942635848748e-05,
|
| 2223 |
+
"loss": 0.8632,
|
| 2224 |
+
"step": 360000
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 38.68,
|
| 2228 |
+
"learning_rate": 1.9682315247255894e-05,
|
| 2229 |
+
"loss": 0.8643,
|
| 2230 |
+
"step": 361000
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 38.79,
|
| 2234 |
+
"learning_rate": 1.942001405240979e-05,
|
| 2235 |
+
"loss": 0.8632,
|
| 2236 |
+
"step": 362000
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 38.9,
|
| 2240 |
+
"learning_rate": 1.9159050525214452e-05,
|
| 2241 |
+
"loss": 0.862,
|
| 2242 |
+
"step": 363000
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 39.01,
|
| 2246 |
+
"learning_rate": 1.8899436081059975e-05,
|
| 2247 |
+
"loss": 0.8619,
|
| 2248 |
+
"step": 364000
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 39.11,
|
| 2252 |
+
"learning_rate": 1.8641182076323148e-05,
|
| 2253 |
+
"loss": 0.8599,
|
| 2254 |
+
"step": 365000
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 39.22,
|
| 2258 |
+
"learning_rate": 1.838429980787081e-05,
|
| 2259 |
+
"loss": 0.8579,
|
| 2260 |
+
"step": 366000
|
| 2261 |
+
},
|
| 2262 |
+
{
|
| 2263 |
+
"epoch": 39.33,
|
| 2264 |
+
"learning_rate": 1.8128800512565513e-05,
|
| 2265 |
+
"loss": 0.8569,
|
| 2266 |
+
"step": 367000
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 39.43,
|
| 2270 |
+
"learning_rate": 1.787469536677419e-05,
|
| 2271 |
+
"loss": 0.8547,
|
| 2272 |
+
"step": 368000
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 39.54,
|
| 2276 |
+
"learning_rate": 1.7621995485879062e-05,
|
| 2277 |
+
"loss": 0.8572,
|
| 2278 |
+
"step": 369000
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 39.65,
|
| 2282 |
+
"learning_rate": 1.7370711923791567e-05,
|
| 2283 |
+
"loss": 0.8573,
|
| 2284 |
+
"step": 370000
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 39.76,
|
| 2288 |
+
"learning_rate": 1.712085567246878e-05,
|
| 2289 |
+
"loss": 0.8573,
|
| 2290 |
+
"step": 371000
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 39.86,
|
| 2294 |
+
"learning_rate": 1.6872437661432517e-05,
|
| 2295 |
+
"loss": 0.8587,
|
| 2296 |
+
"step": 372000
|
| 2297 |
+
},
|
| 2298 |
+
{
|
| 2299 |
+
"epoch": 39.97,
|
| 2300 |
+
"learning_rate": 1.662546875729138e-05,
|
| 2301 |
+
"loss": 0.8558,
|
| 2302 |
+
"step": 373000
|
| 2303 |
+
},
|
| 2304 |
+
{
|
| 2305 |
+
"epoch": 40.08,
|
| 2306 |
+
"learning_rate": 1.637995976326527e-05,
|
| 2307 |
+
"loss": 0.8561,
|
| 2308 |
+
"step": 374000
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 40.18,
|
| 2312 |
+
"learning_rate": 1.6135921418712956e-05,
|
| 2313 |
+
"loss": 0.8545,
|
| 2314 |
+
"step": 375000
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 40.29,
|
| 2318 |
+
"learning_rate": 1.5893364398662176e-05,
|
| 2319 |
+
"loss": 0.8508,
|
| 2320 |
+
"step": 376000
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 40.4,
|
| 2324 |
+
"learning_rate": 1.5652299313342773e-05,
|
| 2325 |
+
"loss": 0.8536,
|
| 2326 |
+
"step": 377000
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 40.51,
|
| 2330 |
+
"learning_rate": 1.5412736707722537e-05,
|
| 2331 |
+
"loss": 0.8533,
|
| 2332 |
+
"step": 378000
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 40.61,
|
| 2336 |
+
"learning_rate": 1.517468706104589e-05,
|
| 2337 |
+
"loss": 0.851,
|
| 2338 |
+
"step": 379000
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 40.72,
|
| 2342 |
+
"learning_rate": 1.4938160786375572e-05,
|
| 2343 |
+
"loss": 0.8529,
|
| 2344 |
+
"step": 380000
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 40.83,
|
| 2348 |
+
"learning_rate": 1.470316823013707e-05,
|
| 2349 |
+
"loss": 0.8526,
|
| 2350 |
+
"step": 381000
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 40.93,
|
| 2354 |
+
"learning_rate": 1.4469719671666043e-05,
|
| 2355 |
+
"loss": 0.8523,
|
| 2356 |
+
"step": 382000
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 41.04,
|
| 2360 |
+
"learning_rate": 1.4237825322758736e-05,
|
| 2361 |
+
"loss": 0.8494,
|
| 2362 |
+
"step": 383000
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 41.15,
|
| 2366 |
+
"learning_rate": 1.4007495327225162e-05,
|
| 2367 |
+
"loss": 0.8488,
|
| 2368 |
+
"step": 384000
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 41.26,
|
| 2372 |
+
"learning_rate": 1.3778739760445552e-05,
|
| 2373 |
+
"loss": 0.8486,
|
| 2374 |
+
"step": 385000
|
| 2375 |
+
},
|
| 2376 |
+
{
|
| 2377 |
+
"epoch": 41.36,
|
| 2378 |
+
"learning_rate": 1.3551568628929434e-05,
|
| 2379 |
+
"loss": 0.8481,
|
| 2380 |
+
"step": 386000
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 41.47,
|
| 2384 |
+
"learning_rate": 1.3325991869878013e-05,
|
| 2385 |
+
"loss": 0.8509,
|
| 2386 |
+
"step": 387000
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 41.58,
|
| 2390 |
+
"learning_rate": 1.3102019350749528e-05,
|
| 2391 |
+
"loss": 0.8467,
|
| 2392 |
+
"step": 388000
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 41.68,
|
| 2396 |
+
"learning_rate": 1.2879660868827508e-05,
|
| 2397 |
+
"loss": 0.8477,
|
| 2398 |
+
"step": 389000
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 41.79,
|
| 2402 |
+
"learning_rate": 1.2658926150792322e-05,
|
| 2403 |
+
"loss": 0.8461,
|
| 2404 |
+
"step": 390000
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 41.9,
|
| 2408 |
+
"learning_rate": 1.243982485229559e-05,
|
| 2409 |
+
"loss": 0.8477,
|
| 2410 |
+
"step": 391000
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 42.01,
|
| 2414 |
+
"learning_rate": 1.2222366557537911e-05,
|
| 2415 |
+
"loss": 0.8464,
|
| 2416 |
+
"step": 392000
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 42.11,
|
| 2420 |
+
"learning_rate": 1.2006560778849578e-05,
|
| 2421 |
+
"loss": 0.8419,
|
| 2422 |
+
"step": 393000
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 42.22,
|
| 2426 |
+
"learning_rate": 1.1792416956274444e-05,
|
| 2427 |
+
"loss": 0.8407,
|
| 2428 |
+
"step": 394000
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 42.33,
|
| 2432 |
+
"learning_rate": 1.157994445715706e-05,
|
| 2433 |
+
"loss": 0.8424,
|
| 2434 |
+
"step": 395000
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 42.43,
|
| 2438 |
+
"learning_rate": 1.1369152575732822e-05,
|
| 2439 |
+
"loss": 0.843,
|
| 2440 |
+
"step": 396000
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 42.54,
|
| 2444 |
+
"learning_rate": 1.1160050532721528e-05,
|
| 2445 |
+
"loss": 0.842,
|
| 2446 |
+
"step": 397000
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 42.65,
|
| 2450 |
+
"learning_rate": 1.095264747492391e-05,
|
| 2451 |
+
"loss": 0.8443,
|
| 2452 |
+
"step": 398000
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 42.76,
|
| 2456 |
+
"learning_rate": 1.0746952474821614e-05,
|
| 2457 |
+
"loss": 0.8418,
|
| 2458 |
+
"step": 399000
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 42.86,
|
| 2462 |
+
"learning_rate": 1.0542974530180327e-05,
|
| 2463 |
+
"loss": 0.8397,
|
| 2464 |
+
"step": 400000
|
| 2465 |
+
},
|
| 2466 |
+
{
|
| 2467 |
+
"epoch": 42.86,
|
| 2468 |
+
"eval_loss": 0.7559537291526794,
|
| 2469 |
+
"eval_runtime": 3.6973,
|
| 2470 |
+
"eval_samples_per_second": 2704.676,
|
| 2471 |
+
"eval_steps_per_second": 21.367,
|
| 2472 |
+
"step": 400000
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 42.97,
|
| 2476 |
+
"learning_rate": 1.0340722563656107e-05,
|
| 2477 |
+
"loss": 0.8447,
|
| 2478 |
+
"step": 401000
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 43.08,
|
| 2482 |
+
"learning_rate": 1.0140205422405214e-05,
|
| 2483 |
+
"loss": 0.8393,
|
| 2484 |
+
"step": 402000
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 43.18,
|
| 2488 |
+
"learning_rate": 9.941431877696955e-06,
|
| 2489 |
+
"loss": 0.8377,
|
| 2490 |
+
"step": 403000
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 43.29,
|
| 2494 |
+
"learning_rate": 9.744410624530148e-06,
|
| 2495 |
+
"loss": 0.8376,
|
| 2496 |
+
"step": 404000
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 43.4,
|
| 2500 |
+
"learning_rate": 9.549150281252633e-06,
|
| 2501 |
+
"loss": 0.8408,
|
| 2502 |
+
"step": 405000
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 43.51,
|
| 2506 |
+
"learning_rate": 9.355659389184396e-06,
|
| 2507 |
+
"loss": 0.8359,
|
| 2508 |
+
"step": 406000
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 43.61,
|
| 2512 |
+
"learning_rate": 9.163946412243896e-06,
|
| 2513 |
+
"loss": 0.839,
|
| 2514 |
+
"step": 407000
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 43.72,
|
| 2518 |
+
"learning_rate": 8.974019736577777e-06,
|
| 2519 |
+
"loss": 0.8371,
|
| 2520 |
+
"step": 408000
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 43.83,
|
| 2524 |
+
"learning_rate": 8.785887670194138e-06,
|
| 2525 |
+
"loss": 0.8377,
|
| 2526 |
+
"step": 409000
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 43.93,
|
| 2530 |
+
"learning_rate": 8.599558442598998e-06,
|
| 2531 |
+
"loss": 0.8372,
|
| 2532 |
+
"step": 410000
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 44.04,
|
| 2536 |
+
"learning_rate": 8.415040204436426e-06,
|
| 2537 |
+
"loss": 0.8369,
|
| 2538 |
+
"step": 411000
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 44.15,
|
| 2542 |
+
"learning_rate": 8.232341027131885e-06,
|
| 2543 |
+
"loss": 0.8363,
|
| 2544 |
+
"step": 412000
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 44.26,
|
| 2548 |
+
"learning_rate": 8.051468902539272e-06,
|
| 2549 |
+
"loss": 0.8337,
|
| 2550 |
+
"step": 413000
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 44.36,
|
| 2554 |
+
"learning_rate": 7.872431742591268e-06,
|
| 2555 |
+
"loss": 0.8324,
|
| 2556 |
+
"step": 414000
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 44.47,
|
| 2560 |
+
"learning_rate": 7.695237378953223e-06,
|
| 2561 |
+
"loss": 0.8361,
|
| 2562 |
+
"step": 415000
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 44.58,
|
| 2566 |
+
"learning_rate": 7.519893562680663e-06,
|
| 2567 |
+
"loss": 0.834,
|
| 2568 |
+
"step": 416000
|
| 2569 |
+
},
|
| 2570 |
+
{
|
| 2571 |
+
"epoch": 44.68,
|
| 2572 |
+
"learning_rate": 7.3464079638801365e-06,
|
| 2573 |
+
"loss": 0.8323,
|
| 2574 |
+
"step": 417000
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 44.79,
|
| 2578 |
+
"learning_rate": 7.174788171373731e-06,
|
| 2579 |
+
"loss": 0.8362,
|
| 2580 |
+
"step": 418000
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 44.9,
|
| 2584 |
+
"learning_rate": 7.005041692367154e-06,
|
| 2585 |
+
"loss": 0.8325,
|
| 2586 |
+
"step": 419000
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 45.01,
|
| 2590 |
+
"learning_rate": 6.837175952121306e-06,
|
| 2591 |
+
"loss": 0.8342,
|
| 2592 |
+
"step": 420000
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 45.11,
|
| 2596 |
+
"learning_rate": 6.671198293627479e-06,
|
| 2597 |
+
"loss": 0.8323,
|
| 2598 |
+
"step": 421000
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 45.22,
|
| 2602 |
+
"learning_rate": 6.5071159772861436e-06,
|
| 2603 |
+
"loss": 0.8331,
|
| 2604 |
+
"step": 422000
|
| 2605 |
+
},
|
| 2606 |
+
{
|
| 2607 |
+
"epoch": 45.33,
|
| 2608 |
+
"learning_rate": 6.344936180589351e-06,
|
| 2609 |
+
"loss": 0.8313,
|
| 2610 |
+
"step": 423000
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 45.44,
|
| 2614 |
+
"learning_rate": 6.184665997806832e-06,
|
| 2615 |
+
"loss": 0.8296,
|
| 2616 |
+
"step": 424000
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 45.54,
|
| 2620 |
+
"learning_rate": 6.026312439675552e-06,
|
| 2621 |
+
"loss": 0.8324,
|
| 2622 |
+
"step": 425000
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 45.65,
|
| 2626 |
+
"learning_rate": 5.869882433093155e-06,
|
| 2627 |
+
"loss": 0.8315,
|
| 2628 |
+
"step": 426000
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 45.76,
|
| 2632 |
+
"learning_rate": 5.715382820814885e-06,
|
| 2633 |
+
"loss": 0.8338,
|
| 2634 |
+
"step": 427000
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 45.86,
|
| 2638 |
+
"learning_rate": 5.562820361154314e-06,
|
| 2639 |
+
"loss": 0.8319,
|
| 2640 |
+
"step": 428000
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 45.97,
|
| 2644 |
+
"learning_rate": 5.412201727687644e-06,
|
| 2645 |
+
"loss": 0.8283,
|
| 2646 |
+
"step": 429000
|
| 2647 |
+
},
|
| 2648 |
+
{
|
| 2649 |
+
"epoch": 46.08,
|
| 2650 |
+
"learning_rate": 5.263533508961827e-06,
|
| 2651 |
+
"loss": 0.8304,
|
| 2652 |
+
"step": 430000
|
| 2653 |
+
},
|
| 2654 |
+
{
|
| 2655 |
+
"epoch": 46.19,
|
| 2656 |
+
"learning_rate": 5.116822208206396e-06,
|
| 2657 |
+
"loss": 0.8271,
|
| 2658 |
+
"step": 431000
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 46.29,
|
| 2662 |
+
"learning_rate": 4.972074243048897e-06,
|
| 2663 |
+
"loss": 0.8288,
|
| 2664 |
+
"step": 432000
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 46.4,
|
| 2668 |
+
"learning_rate": 4.829295945234258e-06,
|
| 2669 |
+
"loss": 0.8289,
|
| 2670 |
+
"step": 433000
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 46.51,
|
| 2674 |
+
"learning_rate": 4.688493560347773e-06,
|
| 2675 |
+
"loss": 0.8284,
|
| 2676 |
+
"step": 434000
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 46.61,
|
| 2680 |
+
"learning_rate": 4.549673247541875e-06,
|
| 2681 |
+
"loss": 0.8272,
|
| 2682 |
+
"step": 435000
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 46.72,
|
| 2686 |
+
"learning_rate": 4.412841079266777e-06,
|
| 2687 |
+
"loss": 0.8288,
|
| 2688 |
+
"step": 436000
|
| 2689 |
+
},
|
| 2690 |
+
{
|
| 2691 |
+
"epoch": 46.83,
|
| 2692 |
+
"learning_rate": 4.27800304100478e-06,
|
| 2693 |
+
"loss": 0.8275,
|
| 2694 |
+
"step": 437000
|
| 2695 |
+
},
|
| 2696 |
+
{
|
| 2697 |
+
"epoch": 46.94,
|
| 2698 |
+
"learning_rate": 4.145165031008508e-06,
|
| 2699 |
+
"loss": 0.8273,
|
| 2700 |
+
"step": 438000
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 47.04,
|
| 2704 |
+
"learning_rate": 4.01433286004283e-06,
|
| 2705 |
+
"loss": 0.8292,
|
| 2706 |
+
"step": 439000
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 47.15,
|
| 2710 |
+
"learning_rate": 3.885512251130763e-06,
|
| 2711 |
+
"loss": 0.8222,
|
| 2712 |
+
"step": 440000
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 47.26,
|
| 2716 |
+
"learning_rate": 3.75870883930306e-06,
|
| 2717 |
+
"loss": 0.829,
|
| 2718 |
+
"step": 441000
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 47.36,
|
| 2722 |
+
"learning_rate": 3.6339281713517303e-06,
|
| 2723 |
+
"loss": 0.8273,
|
| 2724 |
+
"step": 442000
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 47.47,
|
| 2728 |
+
"learning_rate": 3.511175705587433e-06,
|
| 2729 |
+
"loss": 0.8261,
|
| 2730 |
+
"step": 443000
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 47.58,
|
| 2734 |
+
"learning_rate": 3.390456811600673e-06,
|
| 2735 |
+
"loss": 0.8272,
|
| 2736 |
+
"step": 444000
|
| 2737 |
+
},
|
| 2738 |
+
{
|
| 2739 |
+
"epoch": 47.69,
|
| 2740 |
+
"learning_rate": 3.271776770026963e-06,
|
| 2741 |
+
"loss": 0.8258,
|
| 2742 |
+
"step": 445000
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 47.79,
|
| 2746 |
+
"learning_rate": 3.155140772315773e-06,
|
| 2747 |
+
"loss": 0.8261,
|
| 2748 |
+
"step": 446000
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 47.9,
|
| 2752 |
+
"learning_rate": 3.040553920503503e-06,
|
| 2753 |
+
"loss": 0.8265,
|
| 2754 |
+
"step": 447000
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 48.01,
|
| 2758 |
+
"learning_rate": 2.928021226990263e-06,
|
| 2759 |
+
"loss": 0.8239,
|
| 2760 |
+
"step": 448000
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 48.11,
|
| 2764 |
+
"learning_rate": 2.817547614320615e-06,
|
| 2765 |
+
"loss": 0.8233,
|
| 2766 |
+
"step": 449000
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 48.22,
|
| 2770 |
+
"learning_rate": 2.7091379149682685e-06,
|
| 2771 |
+
"loss": 0.8231,
|
| 2772 |
+
"step": 450000
|
| 2773 |
+
},
|
| 2774 |
+
{
|
| 2775 |
+
"epoch": 48.22,
|
| 2776 |
+
"eval_loss": 0.7475652098655701,
|
| 2777 |
+
"eval_runtime": 3.7,
|
| 2778 |
+
"eval_samples_per_second": 2702.714,
|
| 2779 |
+
"eval_steps_per_second": 21.351,
|
| 2780 |
+
"step": 450000
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 48.33,
|
| 2784 |
+
"learning_rate": 2.602796871124663e-06,
|
| 2785 |
+
"loss": 0.824,
|
| 2786 |
+
"step": 451000
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 48.44,
|
| 2790 |
+
"learning_rate": 2.4985291344915674e-06,
|
| 2791 |
+
"loss": 0.8261,
|
| 2792 |
+
"step": 452000
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"epoch": 48.54,
|
| 2796 |
+
"learning_rate": 2.3963392660775575e-06,
|
| 2797 |
+
"loss": 0.8208,
|
| 2798 |
+
"step": 453000
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 48.65,
|
| 2802 |
+
"learning_rate": 2.296231735998511e-06,
|
| 2803 |
+
"loss": 0.8228,
|
| 2804 |
+
"step": 454000
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 48.76,
|
| 2808 |
+
"learning_rate": 2.1982109232821178e-06,
|
| 2809 |
+
"loss": 0.8241,
|
| 2810 |
+
"step": 455000
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 48.86,
|
| 2814 |
+
"learning_rate": 2.102281115676258e-06,
|
| 2815 |
+
"loss": 0.8267,
|
| 2816 |
+
"step": 456000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 48.97,
|
| 2820 |
+
"learning_rate": 2.008446509461498e-06,
|
| 2821 |
+
"loss": 0.8266,
|
| 2822 |
+
"step": 457000
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 49.08,
|
| 2826 |
+
"learning_rate": 1.91671120926748e-06,
|
| 2827 |
+
"loss": 0.8227,
|
| 2828 |
+
"step": 458000
|
| 2829 |
+
},
|
| 2830 |
+
{
|
| 2831 |
+
"epoch": 49.19,
|
| 2832 |
+
"learning_rate": 1.8270792278934302e-06,
|
| 2833 |
+
"loss": 0.8195,
|
| 2834 |
+
"step": 459000
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 49.29,
|
| 2838 |
+
"learning_rate": 1.7395544861325718e-06,
|
| 2839 |
+
"loss": 0.8222,
|
| 2840 |
+
"step": 460000
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 49.4,
|
| 2844 |
+
"learning_rate": 1.6541408126006463e-06,
|
| 2845 |
+
"loss": 0.8213,
|
| 2846 |
+
"step": 461000
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 49.51,
|
| 2850 |
+
"learning_rate": 1.5708419435684462e-06,
|
| 2851 |
+
"loss": 0.8233,
|
| 2852 |
+
"step": 462000
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 49.61,
|
| 2856 |
+
"learning_rate": 1.4896615227983468e-06,
|
| 2857 |
+
"loss": 0.8216,
|
| 2858 |
+
"step": 463000
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 49.72,
|
| 2862 |
+
"learning_rate": 1.4106031013849496e-06,
|
| 2863 |
+
"loss": 0.8221,
|
| 2864 |
+
"step": 464000
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 49.83,
|
| 2868 |
+
"learning_rate": 1.333670137599713e-06,
|
| 2869 |
+
"loss": 0.8221,
|
| 2870 |
+
"step": 465000
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 49.94,
|
| 2874 |
+
"learning_rate": 1.2588659967397e-06,
|
| 2875 |
+
"loss": 0.8223,
|
| 2876 |
+
"step": 466000
|
| 2877 |
+
},
|
| 2878 |
+
{
|
| 2879 |
+
"epoch": 50.04,
|
| 2880 |
+
"learning_rate": 1.1861939509803687e-06,
|
| 2881 |
+
"loss": 0.8229,
|
| 2882 |
+
"step": 467000
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 50.15,
|
| 2886 |
+
"learning_rate": 1.1156571792324211e-06,
|
| 2887 |
+
"loss": 0.823,
|
| 2888 |
+
"step": 468000
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 50.26,
|
| 2892 |
+
"learning_rate": 1.0472587670027678e-06,
|
| 2893 |
+
"loss": 0.8219,
|
| 2894 |
+
"step": 469000
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 50.36,
|
| 2898 |
+
"learning_rate": 9.810017062595322e-07,
|
| 2899 |
+
"loss": 0.8225,
|
| 2900 |
+
"step": 470000
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 50.47,
|
| 2904 |
+
"learning_rate": 9.168888953011989e-07,
|
| 2905 |
+
"loss": 0.8205,
|
| 2906 |
+
"step": 471000
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 50.58,
|
| 2910 |
+
"learning_rate": 8.549231386298151e-07,
|
| 2911 |
+
"loss": 0.8215,
|
| 2912 |
+
"step": 472000
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 50.69,
|
| 2916 |
+
"learning_rate": 7.951071468283167e-07,
|
| 2917 |
+
"loss": 0.8206,
|
| 2918 |
+
"step": 473000
|
| 2919 |
+
},
|
| 2920 |
+
{
|
| 2921 |
+
"epoch": 50.79,
|
| 2922 |
+
"learning_rate": 7.374435364419674e-07,
|
| 2923 |
+
"loss": 0.8203,
|
| 2924 |
+
"step": 474000
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 50.9,
|
| 2928 |
+
"learning_rate": 6.819348298638839e-07,
|
| 2929 |
+
"loss": 0.8207,
|
| 2930 |
+
"step": 475000
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 51.01,
|
| 2934 |
+
"learning_rate": 6.285834552247128e-07,
|
| 2935 |
+
"loss": 0.8208,
|
| 2936 |
+
"step": 476000
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 51.11,
|
| 2940 |
+
"learning_rate": 5.773917462864264e-07,
|
| 2941 |
+
"loss": 0.8206,
|
| 2942 |
+
"step": 477000
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 51.22,
|
| 2946 |
+
"learning_rate": 5.283619423401998e-07,
|
| 2947 |
+
"loss": 0.8198,
|
| 2948 |
+
"step": 478000
|
| 2949 |
+
},
|
| 2950 |
+
{
|
| 2951 |
+
"epoch": 51.33,
|
| 2952 |
+
"learning_rate": 4.814961881085045e-07,
|
| 2953 |
+
"loss": 0.8221,
|
| 2954 |
+
"step": 479000
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 51.44,
|
| 2958 |
+
"learning_rate": 4.367965336512403e-07,
|
| 2959 |
+
"loss": 0.8208,
|
| 2960 |
+
"step": 480000
|
| 2961 |
+
},
|
| 2962 |
+
{
|
| 2963 |
+
"epoch": 51.54,
|
| 2964 |
+
"learning_rate": 3.9426493427611177e-07,
|
| 2965 |
+
"loss": 0.8195,
|
| 2966 |
+
"step": 481000
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 51.65,
|
| 2970 |
+
"learning_rate": 3.5390325045304706e-07,
|
| 2971 |
+
"loss": 0.8217,
|
| 2972 |
+
"step": 482000
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 51.76,
|
| 2976 |
+
"learning_rate": 3.157132477328628e-07,
|
| 2977 |
+
"loss": 0.8234,
|
| 2978 |
+
"step": 483000
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 51.86,
|
| 2982 |
+
"learning_rate": 2.796965966699927e-07,
|
| 2983 |
+
"loss": 0.8223,
|
| 2984 |
+
"step": 484000
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 51.97,
|
| 2988 |
+
"learning_rate": 2.458548727494292e-07,
|
| 2989 |
+
"loss": 0.8225,
|
| 2990 |
+
"step": 485000
|
| 2991 |
+
},
|
| 2992 |
+
{
|
| 2993 |
+
"epoch": 52.08,
|
| 2994 |
+
"learning_rate": 2.1418955631781202e-07,
|
| 2995 |
+
"loss": 0.8207,
|
| 2996 |
+
"step": 486000
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 52.19,
|
| 3000 |
+
"learning_rate": 1.847020325186577e-07,
|
| 3001 |
+
"loss": 0.8193,
|
| 3002 |
+
"step": 487000
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 52.29,
|
| 3006 |
+
"learning_rate": 1.5739359123178587e-07,
|
| 3007 |
+
"loss": 0.8181,
|
| 3008 |
+
"step": 488000
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 52.4,
|
| 3012 |
+
"learning_rate": 1.3226542701689215e-07,
|
| 3013 |
+
"loss": 0.8194,
|
| 3014 |
+
"step": 489000
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 52.51,
|
| 3018 |
+
"learning_rate": 1.0931863906127327e-07,
|
| 3019 |
+
"loss": 0.8213,
|
| 3020 |
+
"step": 490000
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 52.61,
|
| 3024 |
+
"learning_rate": 8.855423113177664e-08,
|
| 3025 |
+
"loss": 0.8214,
|
| 3026 |
+
"step": 491000
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 52.72,
|
| 3030 |
+
"learning_rate": 6.997311153086883e-08,
|
| 3031 |
+
"loss": 0.8228,
|
| 3032 |
+
"step": 492000
|
| 3033 |
+
},
|
| 3034 |
+
{
|
| 3035 |
+
"epoch": 52.83,
|
| 3036 |
+
"learning_rate": 5.3576093056922906e-08,
|
| 3037 |
+
"loss": 0.8235,
|
| 3038 |
+
"step": 493000
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 52.94,
|
| 3042 |
+
"learning_rate": 3.936389296864129e-08,
|
| 3043 |
+
"loss": 0.8207,
|
| 3044 |
+
"step": 494000
|
| 3045 |
+
},
|
| 3046 |
+
{
|
| 3047 |
+
"epoch": 53.04,
|
| 3048 |
+
"learning_rate": 2.7337132953697554e-08,
|
| 3049 |
+
"loss": 0.8212,
|
| 3050 |
+
"step": 495000
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 53.15,
|
| 3054 |
+
"learning_rate": 1.749633910153592e-08,
|
| 3055 |
+
"loss": 0.8231,
|
| 3056 |
+
"step": 496000
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 53.26,
|
| 3060 |
+
"learning_rate": 9.841941880361916e-09,
|
| 3061 |
+
"loss": 0.8191,
|
| 3062 |
+
"step": 497000
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 53.36,
|
| 3066 |
+
"learning_rate": 4.3742761183018784e-09,
|
| 3067 |
+
"loss": 0.8217,
|
| 3068 |
+
"step": 498000
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 53.47,
|
| 3072 |
+
"learning_rate": 1.0935809887702154e-09,
|
| 3073 |
+
"loss": 0.8196,
|
| 3074 |
+
"step": 499000
|
| 3075 |
+
},
|
| 3076 |
+
{
|
| 3077 |
+
"epoch": 53.58,
|
| 3078 |
+
"learning_rate": 0.0,
|
| 3079 |
+
"loss": 0.8207,
|
| 3080 |
+
"step": 500000
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 53.58,
|
| 3084 |
+
"eval_loss": 0.7243289351463318,
|
| 3085 |
+
"eval_runtime": 3.5788,
|
| 3086 |
+
"eval_samples_per_second": 2794.204,
|
| 3087 |
+
"eval_steps_per_second": 22.074,
|
| 3088 |
+
"step": 500000
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"epoch": 53.58,
|
| 3092 |
+
"step": 500000,
|
| 3093 |
+
"total_flos": 1.052837412864e+18,
|
| 3094 |
+
"train_loss": 1.193239736328125,
|
| 3095 |
+
"train_runtime": 102111.9968,
|
| 3096 |
+
"train_samples_per_second": 1253.526,
|
| 3097 |
+
"train_steps_per_second": 4.897
|
| 3098 |
+
}
|
| 3099 |
+
],
|
| 3100 |
+
"max_steps": 500000,
|
| 3101 |
+
"num_train_epochs": 54,
|
| 3102 |
+
"total_flos": 1.052837412864e+18,
|
| 3103 |
+
"trial_name": null,
|
| 3104 |
+
"trial_params": null
|
| 3105 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a997fbe9f0407af3fd443a2bc6f731767935deacd1575a93145aa8a6b91adbc5
|
| 3 |
+
size 3119
|