Hugging Face's logo

pain
/
text2svg_summarization-words-test

Text2Text Generation
Transformers
PyTorch
mt5
Generated from Trainer
Inference Endpoints
Model card Files Community
pain commited on
Commit
3e49467
•
1 Parent(s): df43a5e

Model save

Browse files
Files changed (14) hide show
  1. .gitattributes +1 -0
  2. README.md +52 -0
  3. added_tokens.json +0 -0
  4. all_results.json +8 -0
  5. config.json +32 -0
  6. generation_config.json +7 -0
  7. pytorch_model.bin +3 -0
  8. special_tokens_map.json +5 -0
  9. spiece.model +3 -0
  10. tokenizer.json +3 -0
  11. tokenizer_config.json +0 -0
  12. train_results.json +8 -0
  13. trainer_state.json +3028 -0
  14. training_args.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: google/mt5-small
4
+ tags:
5
+ - generated_from_trainer
6
+ model-index:
7
+ - name: text2svg_summarization-words-test
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
+ # text2svg_summarization-words-test
15
+
16
+ This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on an unknown dataset.
17
+
18
+ ## Model description
19
+
20
+ More information needed
21
+
22
+ ## Intended uses & limitations
23
+
24
+ More information needed
25
+
26
+ ## Training and evaluation data
27
+
28
+ More information needed
29
+
30
+ ## Training procedure
31
+
32
+ ### Training hyperparameters
33
+
34
+ The following hyperparameters were used during training:
35
+ - learning_rate: 0.004
36
+ - train_batch_size: 8
37
+ - eval_batch_size: 8
38
+ - seed: 42
39
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
40
+ - lr_scheduler_type: linear
41
+ - num_epochs: 500.0
42
+
43
+ ### Training results
44
+
45
+
46
+
47
+ ### Framework versions
48
+
49
+ - Transformers 4.35.0.dev0
50
+ - Pytorch 2.1.0+cu121
51
+ - Datasets 2.14.6
52
+ - Tokenizers 0.14.1
added_tokens.json ADDED
The diff for this file is too large to render. See raw diff
 
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 500.0,
3
+ "train_loss": 2.748476623535156,
4
+ "train_runtime": 71445.8185,
5
+ "train_samples": 100,
6
+ "train_samples_per_second": 0.7,
7
+ "train_steps_per_second": 0.091
8
+ }
config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/mt5-small",
3
+ "architectures": [
4
+ "MT5ForConditionalGeneration"
5
+ ],
6
+ "classifier_dropout": 0.0,
7
+ "d_ff": 1024,
8
+ "d_kv": 64,
9
+ "d_model": 512,
10
+ "decoder_start_token_id": 0,
11
+ "dense_act_fn": "gelu_new",
12
+ "dropout_rate": 0.1,
13
+ "eos_token_id": 1,
14
+ "feed_forward_proj": "gated-gelu",
15
+ "initializer_factor": 1.0,
16
+ "is_encoder_decoder": true,
17
+ "is_gated_act": true,
18
+ "layer_norm_epsilon": 1e-06,
19
+ "model_type": "mt5",
20
+ "num_decoder_layers": 8,
21
+ "num_heads": 6,
22
+ "num_layers": 8,
23
+ "pad_token_id": 0,
24
+ "relative_attention_max_distance": 128,
25
+ "relative_attention_num_buckets": 32,
26
+ "tie_word_embeddings": false,
27
+ "tokenizer_class": "T5Tokenizer",
28
+ "torch_dtype": "float32",
29
+ "transformers_version": "4.35.0.dev0",
30
+ "use_cache": true,
31
+ "vocab_size": 279140
32
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "decoder_start_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.35.0.dev0"
7
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98defe36273e59336d7fde474807443622a5db03c3e8872607c0dcd7253629ed
3
+ size 1319669058
special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "eos_token": "</s>",
3
+ "pad_token": "<pad>",
4
+ "unk_token": "<unk>"
5
+ }
spiece.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6
3
+ size 4309802
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad13d641611b787f6783076c5393a3bdbedcf14eaa29892cd9f616278e8ea8db
3
+ size 22110577
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 500.0,
3
+ "train_loss": 2.748476623535156,
4
+ "train_runtime": 71445.8185,
5
+ "train_samples": 100,
6
+ "train_samples_per_second": 0.7,
7
+ "train_steps_per_second": 0.091
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,3028 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 500.0,
5
+ "eval_steps": 500,
6
+ "global_step": 6500,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "learning_rate": 0.003992,
14
+ "loss": 10.6341,
15
+ "step": 13
16
+ },
17
+ {
18
+ "epoch": 2.0,
19
+ "learning_rate": 0.003984,
20
+ "loss": 8.0261,
21
+ "step": 26
22
+ },
23
+ {
24
+ "epoch": 3.0,
25
+ "learning_rate": 0.003976,
26
+ "loss": 7.6356,
27
+ "step": 39
28
+ },
29
+ {
30
+ "epoch": 4.0,
31
+ "learning_rate": 0.003968,
32
+ "loss": 7.489,
33
+ "step": 52
34
+ },
35
+ {
36
+ "epoch": 5.0,
37
+ "learning_rate": 0.00396,
38
+ "loss": 7.3955,
39
+ "step": 65
40
+ },
41
+ {
42
+ "epoch": 6.0,
43
+ "learning_rate": 0.003952,
44
+ "loss": 7.3814,
45
+ "step": 78
46
+ },
47
+ {
48
+ "epoch": 7.0,
49
+ "learning_rate": 0.0039440000000000005,
50
+ "loss": 7.3919,
51
+ "step": 91
52
+ },
53
+ {
54
+ "epoch": 8.0,
55
+ "learning_rate": 0.003936,
56
+ "loss": 7.2877,
57
+ "step": 104
58
+ },
59
+ {
60
+ "epoch": 9.0,
61
+ "learning_rate": 0.003928,
62
+ "loss": 7.0588,
63
+ "step": 117
64
+ },
65
+ {
66
+ "epoch": 10.0,
67
+ "learning_rate": 0.00392,
68
+ "loss": 6.9853,
69
+ "step": 130
70
+ },
71
+ {
72
+ "epoch": 11.0,
73
+ "learning_rate": 0.003912,
74
+ "loss": 6.9981,
75
+ "step": 143
76
+ },
77
+ {
78
+ "epoch": 12.0,
79
+ "learning_rate": 0.003904,
80
+ "loss": 6.8759,
81
+ "step": 156
82
+ },
83
+ {
84
+ "epoch": 13.0,
85
+ "learning_rate": 0.003896,
86
+ "loss": 6.8897,
87
+ "step": 169
88
+ },
89
+ {
90
+ "epoch": 14.0,
91
+ "learning_rate": 0.003888,
92
+ "loss": 7.1851,
93
+ "step": 182
94
+ },
95
+ {
96
+ "epoch": 15.0,
97
+ "learning_rate": 0.0038799999999999998,
98
+ "loss": 7.3121,
99
+ "step": 195
100
+ },
101
+ {
102
+ "epoch": 16.0,
103
+ "learning_rate": 0.003872,
104
+ "loss": 7.2602,
105
+ "step": 208
106
+ },
107
+ {
108
+ "epoch": 17.0,
109
+ "learning_rate": 0.003864,
110
+ "loss": 7.2026,
111
+ "step": 221
112
+ },
113
+ {
114
+ "epoch": 18.0,
115
+ "learning_rate": 0.003856,
116
+ "loss": 7.2713,
117
+ "step": 234
118
+ },
119
+ {
120
+ "epoch": 19.0,
121
+ "learning_rate": 0.003848,
122
+ "loss": 7.1885,
123
+ "step": 247
124
+ },
125
+ {
126
+ "epoch": 20.0,
127
+ "learning_rate": 0.00384,
128
+ "loss": 7.2042,
129
+ "step": 260
130
+ },
131
+ {
132
+ "epoch": 21.0,
133
+ "learning_rate": 0.003832,
134
+ "loss": 7.1744,
135
+ "step": 273
136
+ },
137
+ {
138
+ "epoch": 22.0,
139
+ "learning_rate": 0.0038239999999999997,
140
+ "loss": 7.0481,
141
+ "step": 286
142
+ },
143
+ {
144
+ "epoch": 23.0,
145
+ "learning_rate": 0.003816,
146
+ "loss": 6.8698,
147
+ "step": 299
148
+ },
149
+ {
150
+ "epoch": 24.0,
151
+ "learning_rate": 0.0038079999999999998,
152
+ "loss": 6.7722,
153
+ "step": 312
154
+ },
155
+ {
156
+ "epoch": 25.0,
157
+ "learning_rate": 0.0038,
158
+ "loss": 6.7018,
159
+ "step": 325
160
+ },
161
+ {
162
+ "epoch": 26.0,
163
+ "learning_rate": 0.003792,
164
+ "loss": 6.6881,
165
+ "step": 338
166
+ },
167
+ {
168
+ "epoch": 27.0,
169
+ "learning_rate": 0.003784,
170
+ "loss": 6.7485,
171
+ "step": 351
172
+ },
173
+ {
174
+ "epoch": 28.0,
175
+ "learning_rate": 0.003776,
176
+ "loss": 6.5876,
177
+ "step": 364
178
+ },
179
+ {
180
+ "epoch": 29.0,
181
+ "learning_rate": 0.003768,
182
+ "loss": 6.5597,
183
+ "step": 377
184
+ },
185
+ {
186
+ "epoch": 30.0,
187
+ "learning_rate": 0.00376,
188
+ "loss": 6.5379,
189
+ "step": 390
190
+ },
191
+ {
192
+ "epoch": 31.0,
193
+ "learning_rate": 0.0037519999999999997,
194
+ "loss": 6.3772,
195
+ "step": 403
196
+ },
197
+ {
198
+ "epoch": 32.0,
199
+ "learning_rate": 0.0037440000000000004,
200
+ "loss": 6.3651,
201
+ "step": 416
202
+ },
203
+ {
204
+ "epoch": 33.0,
205
+ "learning_rate": 0.003736,
206
+ "loss": 6.305,
207
+ "step": 429
208
+ },
209
+ {
210
+ "epoch": 34.0,
211
+ "learning_rate": 0.0037280000000000004,
212
+ "loss": 6.2724,
213
+ "step": 442
214
+ },
215
+ {
216
+ "epoch": 35.0,
217
+ "learning_rate": 0.00372,
218
+ "loss": 6.183,
219
+ "step": 455
220
+ },
221
+ {
222
+ "epoch": 36.0,
223
+ "learning_rate": 0.0037120000000000005,
224
+ "loss": 6.2141,
225
+ "step": 468
226
+ },
227
+ {
228
+ "epoch": 37.0,
229
+ "learning_rate": 0.0037040000000000003,
230
+ "loss": 6.1447,
231
+ "step": 481
232
+ },
233
+ {
234
+ "epoch": 38.0,
235
+ "learning_rate": 0.003696,
236
+ "loss": 6.3683,
237
+ "step": 494
238
+ },
239
+ {
240
+ "epoch": 39.0,
241
+ "learning_rate": 0.0036880000000000003,
242
+ "loss": 6.2738,
243
+ "step": 507
244
+ },
245
+ {
246
+ "epoch": 40.0,
247
+ "learning_rate": 0.00368,
248
+ "loss": 6.0499,
249
+ "step": 520
250
+ },
251
+ {
252
+ "epoch": 41.0,
253
+ "learning_rate": 0.0036720000000000004,
254
+ "loss": 5.9005,
255
+ "step": 533
256
+ },
257
+ {
258
+ "epoch": 42.0,
259
+ "learning_rate": 0.003664,
260
+ "loss": 5.8533,
261
+ "step": 546
262
+ },
263
+ {
264
+ "epoch": 43.0,
265
+ "learning_rate": 0.0036560000000000004,
266
+ "loss": 5.8199,
267
+ "step": 559
268
+ },
269
+ {
270
+ "epoch": 44.0,
271
+ "learning_rate": 0.003648,
272
+ "loss": 6.051,
273
+ "step": 572
274
+ },
275
+ {
276
+ "epoch": 45.0,
277
+ "learning_rate": 0.00364,
278
+ "loss": 5.8496,
279
+ "step": 585
280
+ },
281
+ {
282
+ "epoch": 46.0,
283
+ "learning_rate": 0.0036320000000000002,
284
+ "loss": 5.7252,
285
+ "step": 598
286
+ },
287
+ {
288
+ "epoch": 47.0,
289
+ "learning_rate": 0.003624,
290
+ "loss": 5.6958,
291
+ "step": 611
292
+ },
293
+ {
294
+ "epoch": 48.0,
295
+ "learning_rate": 0.0036160000000000003,
296
+ "loss": 5.7218,
297
+ "step": 624
298
+ },
299
+ {
300
+ "epoch": 49.0,
301
+ "learning_rate": 0.003608,
302
+ "loss": 5.6656,
303
+ "step": 637
304
+ },
305
+ {
306
+ "epoch": 50.0,
307
+ "learning_rate": 0.0036000000000000003,
308
+ "loss": 5.612,
309
+ "step": 650
310
+ },
311
+ {
312
+ "epoch": 51.0,
313
+ "learning_rate": 0.003592,
314
+ "loss": 5.5532,
315
+ "step": 663
316
+ },
317
+ {
318
+ "epoch": 52.0,
319
+ "learning_rate": 0.003584,
320
+ "loss": 5.4327,
321
+ "step": 676
322
+ },
323
+ {
324
+ "epoch": 53.0,
325
+ "learning_rate": 0.003576,
326
+ "loss": 5.3979,
327
+ "step": 689
328
+ },
329
+ {
330
+ "epoch": 54.0,
331
+ "learning_rate": 0.003568,
332
+ "loss": 5.2903,
333
+ "step": 702
334
+ },
335
+ {
336
+ "epoch": 55.0,
337
+ "learning_rate": 0.0035600000000000002,
338
+ "loss": 5.4521,
339
+ "step": 715
340
+ },
341
+ {
342
+ "epoch": 56.0,
343
+ "learning_rate": 0.003552,
344
+ "loss": 5.6021,
345
+ "step": 728
346
+ },
347
+ {
348
+ "epoch": 57.0,
349
+ "learning_rate": 0.0035440000000000003,
350
+ "loss": 5.5058,
351
+ "step": 741
352
+ },
353
+ {
354
+ "epoch": 58.0,
355
+ "learning_rate": 0.003536,
356
+ "loss": 5.2167,
357
+ "step": 754
358
+ },
359
+ {
360
+ "epoch": 59.0,
361
+ "learning_rate": 0.003528,
362
+ "loss": 5.2102,
363
+ "step": 767
364
+ },
365
+ {
366
+ "epoch": 60.0,
367
+ "learning_rate": 0.00352,
368
+ "loss": 5.2617,
369
+ "step": 780
370
+ },
371
+ {
372
+ "epoch": 61.0,
373
+ "learning_rate": 0.003512,
374
+ "loss": 5.3012,
375
+ "step": 793
376
+ },
377
+ {
378
+ "epoch": 62.0,
379
+ "learning_rate": 0.003504,
380
+ "loss": 5.2158,
381
+ "step": 806
382
+ },
383
+ {
384
+ "epoch": 63.0,
385
+ "learning_rate": 0.003496,
386
+ "loss": 5.1959,
387
+ "step": 819
388
+ },
389
+ {
390
+ "epoch": 64.0,
391
+ "learning_rate": 0.003488,
392
+ "loss": 5.1716,
393
+ "step": 832
394
+ },
395
+ {
396
+ "epoch": 65.0,
397
+ "learning_rate": 0.00348,
398
+ "loss": 5.0796,
399
+ "step": 845
400
+ },
401
+ {
402
+ "epoch": 66.0,
403
+ "learning_rate": 0.0034720000000000003,
404
+ "loss": 4.9764,
405
+ "step": 858
406
+ },
407
+ {
408
+ "epoch": 67.0,
409
+ "learning_rate": 0.003464,
410
+ "loss": 4.974,
411
+ "step": 871
412
+ },
413
+ {
414
+ "epoch": 68.0,
415
+ "learning_rate": 0.003456,
416
+ "loss": 4.876,
417
+ "step": 884
418
+ },
419
+ {
420
+ "epoch": 69.0,
421
+ "learning_rate": 0.003448,
422
+ "loss": 4.8596,
423
+ "step": 897
424
+ },
425
+ {
426
+ "epoch": 70.0,
427
+ "learning_rate": 0.00344,
428
+ "loss": 4.7792,
429
+ "step": 910
430
+ },
431
+ {
432
+ "epoch": 71.0,
433
+ "learning_rate": 0.003432,
434
+ "loss": 4.765,
435
+ "step": 923
436
+ },
437
+ {
438
+ "epoch": 72.0,
439
+ "learning_rate": 0.003424,
440
+ "loss": 4.7933,
441
+ "step": 936
442
+ },
443
+ {
444
+ "epoch": 73.0,
445
+ "learning_rate": 0.003416,
446
+ "loss": 4.7636,
447
+ "step": 949
448
+ },
449
+ {
450
+ "epoch": 74.0,
451
+ "learning_rate": 0.003408,
452
+ "loss": 4.7114,
453
+ "step": 962
454
+ },
455
+ {
456
+ "epoch": 75.0,
457
+ "learning_rate": 0.0034,
458
+ "loss": 4.7079,
459
+ "step": 975
460
+ },
461
+ {
462
+ "epoch": 76.0,
463
+ "learning_rate": 0.003392,
464
+ "loss": 4.6745,
465
+ "step": 988
466
+ },
467
+ {
468
+ "epoch": 77.0,
469
+ "learning_rate": 0.003384,
470
+ "loss": 4.6765,
471
+ "step": 1001
472
+ },
473
+ {
474
+ "epoch": 78.0,
475
+ "learning_rate": 0.003376,
476
+ "loss": 4.5913,
477
+ "step": 1014
478
+ },
479
+ {
480
+ "epoch": 79.0,
481
+ "learning_rate": 0.003368,
482
+ "loss": 4.7949,
483
+ "step": 1027
484
+ },
485
+ {
486
+ "epoch": 80.0,
487
+ "learning_rate": 0.00336,
488
+ "loss": 4.6311,
489
+ "step": 1040
490
+ },
491
+ {
492
+ "epoch": 81.0,
493
+ "learning_rate": 0.003352,
494
+ "loss": 4.4818,
495
+ "step": 1053
496
+ },
497
+ {
498
+ "epoch": 82.0,
499
+ "learning_rate": 0.0033439999999999998,
500
+ "loss": 4.4462,
501
+ "step": 1066
502
+ },
503
+ {
504
+ "epoch": 83.0,
505
+ "learning_rate": 0.003336,
506
+ "loss": 4.5129,
507
+ "step": 1079
508
+ },
509
+ {
510
+ "epoch": 84.0,
511
+ "learning_rate": 0.003328,
512
+ "loss": 4.4626,
513
+ "step": 1092
514
+ },
515
+ {
516
+ "epoch": 85.0,
517
+ "learning_rate": 0.00332,
518
+ "loss": 4.3505,
519
+ "step": 1105
520
+ },
521
+ {
522
+ "epoch": 86.0,
523
+ "learning_rate": 0.003312,
524
+ "loss": 4.3377,
525
+ "step": 1118
526
+ },
527
+ {
528
+ "epoch": 87.0,
529
+ "learning_rate": 0.003304,
530
+ "loss": 4.4076,
531
+ "step": 1131
532
+ },
533
+ {
534
+ "epoch": 88.0,
535
+ "learning_rate": 0.003296,
536
+ "loss": 4.3765,
537
+ "step": 1144
538
+ },
539
+ {
540
+ "epoch": 89.0,
541
+ "learning_rate": 0.0032879999999999997,
542
+ "loss": 4.2473,
543
+ "step": 1157
544
+ },
545
+ {
546
+ "epoch": 90.0,
547
+ "learning_rate": 0.00328,
548
+ "loss": 4.2142,
549
+ "step": 1170
550
+ },
551
+ {
552
+ "epoch": 91.0,
553
+ "learning_rate": 0.0032719999999999997,
554
+ "loss": 4.1567,
555
+ "step": 1183
556
+ },
557
+ {
558
+ "epoch": 92.0,
559
+ "learning_rate": 0.003264,
560
+ "loss": 4.1569,
561
+ "step": 1196
562
+ },
563
+ {
564
+ "epoch": 93.0,
565
+ "learning_rate": 0.0032559999999999998,
566
+ "loss": 4.1347,
567
+ "step": 1209
568
+ },
569
+ {
570
+ "epoch": 94.0,
571
+ "learning_rate": 0.0032480000000000005,
572
+ "loss": 4.0786,
573
+ "step": 1222
574
+ },
575
+ {
576
+ "epoch": 95.0,
577
+ "learning_rate": 0.0032400000000000003,
578
+ "loss": 4.0796,
579
+ "step": 1235
580
+ },
581
+ {
582
+ "epoch": 96.0,
583
+ "learning_rate": 0.003232,
584
+ "loss": 4.0432,
585
+ "step": 1248
586
+ },
587
+ {
588
+ "epoch": 97.0,
589
+ "learning_rate": 0.0032240000000000003,
590
+ "loss": 4.033,
591
+ "step": 1261
592
+ },
593
+ {
594
+ "epoch": 98.0,
595
+ "learning_rate": 0.003216,
596
+ "loss": 3.952,
597
+ "step": 1274
598
+ },
599
+ {
600
+ "epoch": 99.0,
601
+ "learning_rate": 0.0032080000000000003,
602
+ "loss": 4.0043,
603
+ "step": 1287
604
+ },
605
+ {
606
+ "epoch": 100.0,
607
+ "learning_rate": 0.0032,
608
+ "loss": 4.2161,
609
+ "step": 1300
610
+ },
611
+ {
612
+ "epoch": 101.0,
613
+ "learning_rate": 0.0031920000000000004,
614
+ "loss": 4.1006,
615
+ "step": 1313
616
+ },
617
+ {
618
+ "epoch": 102.0,
619
+ "learning_rate": 0.003184,
620
+ "loss": 4.1527,
621
+ "step": 1326
622
+ },
623
+ {
624
+ "epoch": 103.0,
625
+ "learning_rate": 0.0031760000000000004,
626
+ "loss": 3.9791,
627
+ "step": 1339
628
+ },
629
+ {
630
+ "epoch": 104.0,
631
+ "learning_rate": 0.0031680000000000002,
632
+ "loss": 3.9599,
633
+ "step": 1352
634
+ },
635
+ {
636
+ "epoch": 105.0,
637
+ "learning_rate": 0.00316,
638
+ "loss": 3.9998,
639
+ "step": 1365
640
+ },
641
+ {
642
+ "epoch": 106.0,
643
+ "learning_rate": 0.0031520000000000003,
644
+ "loss": 3.973,
645
+ "step": 1378
646
+ },
647
+ {
648
+ "epoch": 107.0,
649
+ "learning_rate": 0.003144,
650
+ "loss": 3.9976,
651
+ "step": 1391
652
+ },
653
+ {
654
+ "epoch": 108.0,
655
+ "learning_rate": 0.0031360000000000003,
656
+ "loss": 3.9862,
657
+ "step": 1404
658
+ },
659
+ {
660
+ "epoch": 109.0,
661
+ "learning_rate": 0.003128,
662
+ "loss": 3.8562,
663
+ "step": 1417
664
+ },
665
+ {
666
+ "epoch": 110.0,
667
+ "learning_rate": 0.0031200000000000004,
668
+ "loss": 3.8322,
669
+ "step": 1430
670
+ },
671
+ {
672
+ "epoch": 111.0,
673
+ "learning_rate": 0.003112,
674
+ "loss": 3.8451,
675
+ "step": 1443
676
+ },
677
+ {
678
+ "epoch": 112.0,
679
+ "learning_rate": 0.003104,
680
+ "loss": 3.8274,
681
+ "step": 1456
682
+ },
683
+ {
684
+ "epoch": 113.0,
685
+ "learning_rate": 0.0030960000000000002,
686
+ "loss": 3.8483,
687
+ "step": 1469
688
+ },
689
+ {
690
+ "epoch": 114.0,
691
+ "learning_rate": 0.003088,
692
+ "loss": 3.7911,
693
+ "step": 1482
694
+ },
695
+ {
696
+ "epoch": 115.0,
697
+ "learning_rate": 0.0030800000000000003,
698
+ "loss": 3.8203,
699
+ "step": 1495
700
+ },
701
+ {
702
+ "epoch": 116.0,
703
+ "learning_rate": 0.003072,
704
+ "loss": 3.7111,
705
+ "step": 1508
706
+ },
707
+ {
708
+ "epoch": 117.0,
709
+ "learning_rate": 0.0030640000000000003,
710
+ "loss": 3.7186,
711
+ "step": 1521
712
+ },
713
+ {
714
+ "epoch": 118.0,
715
+ "learning_rate": 0.003056,
716
+ "loss": 3.6357,
717
+ "step": 1534
718
+ },
719
+ {
720
+ "epoch": 119.0,
721
+ "learning_rate": 0.003048,
722
+ "loss": 3.6484,
723
+ "step": 1547
724
+ },
725
+ {
726
+ "epoch": 120.0,
727
+ "learning_rate": 0.00304,
728
+ "loss": 3.7188,
729
+ "step": 1560
730
+ },
731
+ {
732
+ "epoch": 121.0,
733
+ "learning_rate": 0.003032,
734
+ "loss": 3.6217,
735
+ "step": 1573
736
+ },
737
+ {
738
+ "epoch": 122.0,
739
+ "learning_rate": 0.003024,
740
+ "loss": 3.5853,
741
+ "step": 1586
742
+ },
743
+ {
744
+ "epoch": 123.0,
745
+ "learning_rate": 0.003016,
746
+ "loss": 3.6381,
747
+ "step": 1599
748
+ },
749
+ {
750
+ "epoch": 124.0,
751
+ "learning_rate": 0.0030080000000000003,
752
+ "loss": 3.6051,
753
+ "step": 1612
754
+ },
755
+ {
756
+ "epoch": 125.0,
757
+ "learning_rate": 0.003,
758
+ "loss": 3.6293,
759
+ "step": 1625
760
+ },
761
+ {
762
+ "epoch": 126.0,
763
+ "learning_rate": 0.002992,
764
+ "loss": 3.626,
765
+ "step": 1638
766
+ },
767
+ {
768
+ "epoch": 127.0,
769
+ "learning_rate": 0.002984,
770
+ "loss": 3.6121,
771
+ "step": 1651
772
+ },
773
+ {
774
+ "epoch": 128.0,
775
+ "learning_rate": 0.002976,
776
+ "loss": 3.5777,
777
+ "step": 1664
778
+ },
779
+ {
780
+ "epoch": 129.0,
781
+ "learning_rate": 0.002968,
782
+ "loss": 3.551,
783
+ "step": 1677
784
+ },
785
+ {
786
+ "epoch": 130.0,
787
+ "learning_rate": 0.00296,
788
+ "loss": 3.534,
789
+ "step": 1690
790
+ },
791
+ {
792
+ "epoch": 131.0,
793
+ "learning_rate": 0.002952,
794
+ "loss": 3.5946,
795
+ "step": 1703
796
+ },
797
+ {
798
+ "epoch": 132.0,
799
+ "learning_rate": 0.002944,
800
+ "loss": 3.6511,
801
+ "step": 1716
802
+ },
803
+ {
804
+ "epoch": 133.0,
805
+ "learning_rate": 0.002936,
806
+ "loss": 3.5556,
807
+ "step": 1729
808
+ },
809
+ {
810
+ "epoch": 134.0,
811
+ "learning_rate": 0.002928,
812
+ "loss": 3.5453,
813
+ "step": 1742
814
+ },
815
+ {
816
+ "epoch": 135.0,
817
+ "learning_rate": 0.00292,
818
+ "loss": 3.5641,
819
+ "step": 1755
820
+ },
821
+ {
822
+ "epoch": 136.0,
823
+ "learning_rate": 0.002912,
824
+ "loss": 3.5357,
825
+ "step": 1768
826
+ },
827
+ {
828
+ "epoch": 137.0,
829
+ "learning_rate": 0.002904,
830
+ "loss": 3.5738,
831
+ "step": 1781
832
+ },
833
+ {
834
+ "epoch": 138.0,
835
+ "learning_rate": 0.002896,
836
+ "loss": 3.4697,
837
+ "step": 1794
838
+ },
839
+ {
840
+ "epoch": 139.0,
841
+ "learning_rate": 0.002888,
842
+ "loss": 3.4405,
843
+ "step": 1807
844
+ },
845
+ {
846
+ "epoch": 140.0,
847
+ "learning_rate": 0.0028799999999999997,
848
+ "loss": 3.3998,
849
+ "step": 1820
850
+ },
851
+ {
852
+ "epoch": 141.0,
853
+ "learning_rate": 0.002872,
854
+ "loss": 3.4035,
855
+ "step": 1833
856
+ },
857
+ {
858
+ "epoch": 142.0,
859
+ "learning_rate": 0.002864,
860
+ "loss": 3.4335,
861
+ "step": 1846
862
+ },
863
+ {
864
+ "epoch": 143.0,
865
+ "learning_rate": 0.002856,
866
+ "loss": 3.4105,
867
+ "step": 1859
868
+ },
869
+ {
870
+ "epoch": 144.0,
871
+ "learning_rate": 0.002848,
872
+ "loss": 3.3161,
873
+ "step": 1872
874
+ },
875
+ {
876
+ "epoch": 145.0,
877
+ "learning_rate": 0.00284,
878
+ "loss": 3.2802,
879
+ "step": 1885
880
+ },
881
+ {
882
+ "epoch": 146.0,
883
+ "learning_rate": 0.002832,
884
+ "loss": 3.2573,
885
+ "step": 1898
886
+ },
887
+ {
888
+ "epoch": 147.0,
889
+ "learning_rate": 0.0028239999999999997,
890
+ "loss": 3.265,
891
+ "step": 1911
892
+ },
893
+ {
894
+ "epoch": 148.0,
895
+ "learning_rate": 0.002816,
896
+ "loss": 3.3362,
897
+ "step": 1924
898
+ },
899
+ {
900
+ "epoch": 149.0,
901
+ "learning_rate": 0.0028079999999999997,
902
+ "loss": 3.2085,
903
+ "step": 1937
904
+ },
905
+ {
906
+ "epoch": 150.0,
907
+ "learning_rate": 0.0028,
908
+ "loss": 3.2445,
909
+ "step": 1950
910
+ },
911
+ {
912
+ "epoch": 151.0,
913
+ "learning_rate": 0.0027919999999999998,
914
+ "loss": 3.2212,
915
+ "step": 1963
916
+ },
917
+ {
918
+ "epoch": 152.0,
919
+ "learning_rate": 0.002784,
920
+ "loss": 3.2135,
921
+ "step": 1976
922
+ },
923
+ {
924
+ "epoch": 153.0,
925
+ "learning_rate": 0.002776,
926
+ "loss": 3.173,
927
+ "step": 1989
928
+ },
929
+ {
930
+ "epoch": 154.0,
931
+ "learning_rate": 0.002768,
932
+ "loss": 3.1946,
933
+ "step": 2002
934
+ },
935
+ {
936
+ "epoch": 155.0,
937
+ "learning_rate": 0.00276,
938
+ "loss": 3.1739,
939
+ "step": 2015
940
+ },
941
+ {
942
+ "epoch": 156.0,
943
+ "learning_rate": 0.0027519999999999997,
944
+ "loss": 3.1975,
945
+ "step": 2028
946
+ },
947
+ {
948
+ "epoch": 157.0,
949
+ "learning_rate": 0.0027440000000000003,
950
+ "loss": 3.148,
951
+ "step": 2041
952
+ },
953
+ {
954
+ "epoch": 158.0,
955
+ "learning_rate": 0.002736,
956
+ "loss": 3.1124,
957
+ "step": 2054
958
+ },
959
+ {
960
+ "epoch": 159.0,
961
+ "learning_rate": 0.0027280000000000004,
962
+ "loss": 3.1101,
963
+ "step": 2067
964
+ },
965
+ {
966
+ "epoch": 160.0,
967
+ "learning_rate": 0.00272,
968
+ "loss": 3.155,
969
+ "step": 2080
970
+ },
971
+ {
972
+ "epoch": 161.0,
973
+ "learning_rate": 0.0027120000000000004,
974
+ "loss": 3.091,
975
+ "step": 2093
976
+ },
977
+ {
978
+ "epoch": 162.0,
979
+ "learning_rate": 0.0027040000000000002,
980
+ "loss": 3.0156,
981
+ "step": 2106
982
+ },
983
+ {
984
+ "epoch": 163.0,
985
+ "learning_rate": 0.002696,
986
+ "loss": 3.031,
987
+ "step": 2119
988
+ },
989
+ {
990
+ "epoch": 164.0,
991
+ "learning_rate": 0.0026880000000000003,
992
+ "loss": 3.0426,
993
+ "step": 2132
994
+ },
995
+ {
996
+ "epoch": 165.0,
997
+ "learning_rate": 0.00268,
998
+ "loss": 2.9667,
999
+ "step": 2145
1000
+ },
1001
+ {
1002
+ "epoch": 166.0,
1003
+ "learning_rate": 0.0026720000000000003,
1004
+ "loss": 2.9496,
1005
+ "step": 2158
1006
+ },
1007
+ {
1008
+ "epoch": 167.0,
1009
+ "learning_rate": 0.002664,
1010
+ "loss": 3.0151,
1011
+ "step": 2171
1012
+ },
1013
+ {
1014
+ "epoch": 168.0,
1015
+ "learning_rate": 0.0026560000000000004,
1016
+ "loss": 3.0202,
1017
+ "step": 2184
1018
+ },
1019
+ {
1020
+ "epoch": 169.0,
1021
+ "learning_rate": 0.002648,
1022
+ "loss": 3.1202,
1023
+ "step": 2197
1024
+ },
1025
+ {
1026
+ "epoch": 170.0,
1027
+ "learning_rate": 0.00264,
1028
+ "loss": 3.0814,
1029
+ "step": 2210
1030
+ },
1031
+ {
1032
+ "epoch": 171.0,
1033
+ "learning_rate": 0.0026320000000000002,
1034
+ "loss": 2.9501,
1035
+ "step": 2223
1036
+ },
1037
+ {
1038
+ "epoch": 172.0,
1039
+ "learning_rate": 0.002624,
1040
+ "loss": 2.8994,
1041
+ "step": 2236
1042
+ },
1043
+ {
1044
+ "epoch": 173.0,
1045
+ "learning_rate": 0.0026160000000000003,
1046
+ "loss": 2.8437,
1047
+ "step": 2249
1048
+ },
1049
+ {
1050
+ "epoch": 174.0,
1051
+ "learning_rate": 0.002608,
1052
+ "loss": 2.8867,
1053
+ "step": 2262
1054
+ },
1055
+ {
1056
+ "epoch": 175.0,
1057
+ "learning_rate": 0.0026000000000000003,
1058
+ "loss": 2.8977,
1059
+ "step": 2275
1060
+ },
1061
+ {
1062
+ "epoch": 176.0,
1063
+ "learning_rate": 0.002592,
1064
+ "loss": 2.8601,
1065
+ "step": 2288
1066
+ },
1067
+ {
1068
+ "epoch": 177.0,
1069
+ "learning_rate": 0.002584,
1070
+ "loss": 2.9511,
1071
+ "step": 2301
1072
+ },
1073
+ {
1074
+ "epoch": 178.0,
1075
+ "learning_rate": 0.002576,
1076
+ "loss": 2.8396,
1077
+ "step": 2314
1078
+ },
1079
+ {
1080
+ "epoch": 179.0,
1081
+ "learning_rate": 0.002568,
1082
+ "loss": 2.8238,
1083
+ "step": 2327
1084
+ },
1085
+ {
1086
+ "epoch": 180.0,
1087
+ "learning_rate": 0.00256,
1088
+ "loss": 2.8048,
1089
+ "step": 2340
1090
+ },
1091
+ {
1092
+ "epoch": 181.0,
1093
+ "learning_rate": 0.002552,
1094
+ "loss": 2.7583,
1095
+ "step": 2353
1096
+ },
1097
+ {
1098
+ "epoch": 182.0,
1099
+ "learning_rate": 0.0025440000000000003,
1100
+ "loss": 2.7443,
1101
+ "step": 2366
1102
+ },
1103
+ {
1104
+ "epoch": 183.0,
1105
+ "learning_rate": 0.002536,
1106
+ "loss": 2.7362,
1107
+ "step": 2379
1108
+ },
1109
+ {
1110
+ "epoch": 184.0,
1111
+ "learning_rate": 0.002528,
1112
+ "loss": 2.7878,
1113
+ "step": 2392
1114
+ },
1115
+ {
1116
+ "epoch": 185.0,
1117
+ "learning_rate": 0.00252,
1118
+ "loss": 2.7811,
1119
+ "step": 2405
1120
+ },
1121
+ {
1122
+ "epoch": 186.0,
1123
+ "learning_rate": 0.002512,
1124
+ "loss": 2.7213,
1125
+ "step": 2418
1126
+ },
1127
+ {
1128
+ "epoch": 187.0,
1129
+ "learning_rate": 0.002504,
1130
+ "loss": 2.7716,
1131
+ "step": 2431
1132
+ },
1133
+ {
1134
+ "epoch": 188.0,
1135
+ "learning_rate": 0.002496,
1136
+ "loss": 2.7761,
1137
+ "step": 2444
1138
+ },
1139
+ {
1140
+ "epoch": 189.0,
1141
+ "learning_rate": 0.002488,
1142
+ "loss": 2.7456,
1143
+ "step": 2457
1144
+ },
1145
+ {
1146
+ "epoch": 190.0,
1147
+ "learning_rate": 0.00248,
1148
+ "loss": 2.9211,
1149
+ "step": 2470
1150
+ },
1151
+ {
1152
+ "epoch": 191.0,
1153
+ "learning_rate": 0.0024720000000000002,
1154
+ "loss": 2.9644,
1155
+ "step": 2483
1156
+ },
1157
+ {
1158
+ "epoch": 192.0,
1159
+ "learning_rate": 0.002464,
1160
+ "loss": 2.7444,
1161
+ "step": 2496
1162
+ },
1163
+ {
1164
+ "epoch": 193.0,
1165
+ "learning_rate": 0.002456,
1166
+ "loss": 2.7094,
1167
+ "step": 2509
1168
+ },
1169
+ {
1170
+ "epoch": 194.0,
1171
+ "learning_rate": 0.002448,
1172
+ "loss": 2.6593,
1173
+ "step": 2522
1174
+ },
1175
+ {
1176
+ "epoch": 195.0,
1177
+ "learning_rate": 0.00244,
1178
+ "loss": 2.6424,
1179
+ "step": 2535
1180
+ },
1181
+ {
1182
+ "epoch": 196.0,
1183
+ "learning_rate": 0.002432,
1184
+ "loss": 2.5913,
1185
+ "step": 2548
1186
+ },
1187
+ {
1188
+ "epoch": 197.0,
1189
+ "learning_rate": 0.002424,
1190
+ "loss": 2.6003,
1191
+ "step": 2561
1192
+ },
1193
+ {
1194
+ "epoch": 198.0,
1195
+ "learning_rate": 0.002416,
1196
+ "loss": 2.6317,
1197
+ "step": 2574
1198
+ },
1199
+ {
1200
+ "epoch": 199.0,
1201
+ "learning_rate": 0.002408,
1202
+ "loss": 2.6468,
1203
+ "step": 2587
1204
+ },
1205
+ {
1206
+ "epoch": 200.0,
1207
+ "learning_rate": 0.0024,
1208
+ "loss": 2.5951,
1209
+ "step": 2600
1210
+ },
1211
+ {
1212
+ "epoch": 201.0,
1213
+ "learning_rate": 0.002392,
1214
+ "loss": 2.5915,
1215
+ "step": 2613
1216
+ },
1217
+ {
1218
+ "epoch": 202.0,
1219
+ "learning_rate": 0.002384,
1220
+ "loss": 2.568,
1221
+ "step": 2626
1222
+ },
1223
+ {
1224
+ "epoch": 203.0,
1225
+ "learning_rate": 0.002376,
1226
+ "loss": 2.5466,
1227
+ "step": 2639
1228
+ },
1229
+ {
1230
+ "epoch": 204.0,
1231
+ "learning_rate": 0.002368,
1232
+ "loss": 2.6858,
1233
+ "step": 2652
1234
+ },
1235
+ {
1236
+ "epoch": 205.0,
1237
+ "learning_rate": 0.00236,
1238
+ "loss": 2.5551,
1239
+ "step": 2665
1240
+ },
1241
+ {
1242
+ "epoch": 206.0,
1243
+ "learning_rate": 0.002352,
1244
+ "loss": 2.5618,
1245
+ "step": 2678
1246
+ },
1247
+ {
1248
+ "epoch": 207.0,
1249
+ "learning_rate": 0.0023439999999999997,
1250
+ "loss": 2.5309,
1251
+ "step": 2691
1252
+ },
1253
+ {
1254
+ "epoch": 208.0,
1255
+ "learning_rate": 0.002336,
1256
+ "loss": 2.5307,
1257
+ "step": 2704
1258
+ },
1259
+ {
1260
+ "epoch": 209.0,
1261
+ "learning_rate": 0.0023279999999999998,
1262
+ "loss": 2.5008,
1263
+ "step": 2717
1264
+ },
1265
+ {
1266
+ "epoch": 210.0,
1267
+ "learning_rate": 0.00232,
1268
+ "loss": 2.5485,
1269
+ "step": 2730
1270
+ },
1271
+ {
1272
+ "epoch": 211.0,
1273
+ "learning_rate": 0.002312,
1274
+ "loss": 2.547,
1275
+ "step": 2743
1276
+ },
1277
+ {
1278
+ "epoch": 212.0,
1279
+ "learning_rate": 0.002304,
1280
+ "loss": 2.461,
1281
+ "step": 2756
1282
+ },
1283
+ {
1284
+ "epoch": 213.0,
1285
+ "learning_rate": 0.002296,
1286
+ "loss": 2.4375,
1287
+ "step": 2769
1288
+ },
1289
+ {
1290
+ "epoch": 214.0,
1291
+ "learning_rate": 0.0022879999999999997,
1292
+ "loss": 2.4417,
1293
+ "step": 2782
1294
+ },
1295
+ {
1296
+ "epoch": 215.0,
1297
+ "learning_rate": 0.00228,
1298
+ "loss": 2.4427,
1299
+ "step": 2795
1300
+ },
1301
+ {
1302
+ "epoch": 216.0,
1303
+ "learning_rate": 0.0022719999999999997,
1304
+ "loss": 2.4756,
1305
+ "step": 2808
1306
+ },
1307
+ {
1308
+ "epoch": 217.0,
1309
+ "learning_rate": 0.002264,
1310
+ "loss": 2.4662,
1311
+ "step": 2821
1312
+ },
1313
+ {
1314
+ "epoch": 218.0,
1315
+ "learning_rate": 0.0022559999999999998,
1316
+ "loss": 2.4931,
1317
+ "step": 2834
1318
+ },
1319
+ {
1320
+ "epoch": 219.0,
1321
+ "learning_rate": 0.0022480000000000004,
1322
+ "loss": 2.4438,
1323
+ "step": 2847
1324
+ },
1325
+ {
1326
+ "epoch": 220.0,
1327
+ "learning_rate": 0.0022400000000000002,
1328
+ "loss": 2.3834,
1329
+ "step": 2860
1330
+ },
1331
+ {
1332
+ "epoch": 221.0,
1333
+ "learning_rate": 0.002232,
1334
+ "loss": 2.4078,
1335
+ "step": 2873
1336
+ },
1337
+ {
1338
+ "epoch": 222.0,
1339
+ "learning_rate": 0.0022240000000000003,
1340
+ "loss": 2.3813,
1341
+ "step": 2886
1342
+ },
1343
+ {
1344
+ "epoch": 223.0,
1345
+ "learning_rate": 0.002216,
1346
+ "loss": 2.382,
1347
+ "step": 2899
1348
+ },
1349
+ {
1350
+ "epoch": 224.0,
1351
+ "learning_rate": 0.0022080000000000003,
1352
+ "loss": 2.361,
1353
+ "step": 2912
1354
+ },
1355
+ {
1356
+ "epoch": 225.0,
1357
+ "learning_rate": 0.0022,
1358
+ "loss": 2.3106,
1359
+ "step": 2925
1360
+ },
1361
+ {
1362
+ "epoch": 226.0,
1363
+ "learning_rate": 0.0021920000000000004,
1364
+ "loss": 2.2991,
1365
+ "step": 2938
1366
+ },
1367
+ {
1368
+ "epoch": 227.0,
1369
+ "learning_rate": 0.002184,
1370
+ "loss": 2.231,
1371
+ "step": 2951
1372
+ },
1373
+ {
1374
+ "epoch": 228.0,
1375
+ "learning_rate": 0.0021760000000000004,
1376
+ "loss": 2.2748,
1377
+ "step": 2964
1378
+ },
1379
+ {
1380
+ "epoch": 229.0,
1381
+ "learning_rate": 0.0021680000000000002,
1382
+ "loss": 2.2974,
1383
+ "step": 2977
1384
+ },
1385
+ {
1386
+ "epoch": 230.0,
1387
+ "learning_rate": 0.00216,
1388
+ "loss": 2.2974,
1389
+ "step": 2990
1390
+ },
1391
+ {
1392
+ "epoch": 231.0,
1393
+ "learning_rate": 0.0021520000000000003,
1394
+ "loss": 2.2755,
1395
+ "step": 3003
1396
+ },
1397
+ {
1398
+ "epoch": 232.0,
1399
+ "learning_rate": 0.002144,
1400
+ "loss": 2.287,
1401
+ "step": 3016
1402
+ },
1403
+ {
1404
+ "epoch": 233.0,
1405
+ "learning_rate": 0.0021360000000000003,
1406
+ "loss": 2.2462,
1407
+ "step": 3029
1408
+ },
1409
+ {
1410
+ "epoch": 234.0,
1411
+ "learning_rate": 0.002128,
1412
+ "loss": 2.2528,
1413
+ "step": 3042
1414
+ },
1415
+ {
1416
+ "epoch": 235.0,
1417
+ "learning_rate": 0.0021200000000000004,
1418
+ "loss": 2.2052,
1419
+ "step": 3055
1420
+ },
1421
+ {
1422
+ "epoch": 236.0,
1423
+ "learning_rate": 0.002112,
1424
+ "loss": 2.2461,
1425
+ "step": 3068
1426
+ },
1427
+ {
1428
+ "epoch": 237.0,
1429
+ "learning_rate": 0.002104,
1430
+ "loss": 2.2099,
1431
+ "step": 3081
1432
+ },
1433
+ {
1434
+ "epoch": 238.0,
1435
+ "learning_rate": 0.002096,
1436
+ "loss": 2.1273,
1437
+ "step": 3094
1438
+ },
1439
+ {
1440
+ "epoch": 239.0,
1441
+ "learning_rate": 0.002088,
1442
+ "loss": 2.1668,
1443
+ "step": 3107
1444
+ },
1445
+ {
1446
+ "epoch": 240.0,
1447
+ "learning_rate": 0.0020800000000000003,
1448
+ "loss": 2.1719,
1449
+ "step": 3120
1450
+ },
1451
+ {
1452
+ "epoch": 241.0,
1453
+ "learning_rate": 0.002072,
1454
+ "loss": 2.171,
1455
+ "step": 3133
1456
+ },
1457
+ {
1458
+ "epoch": 242.0,
1459
+ "learning_rate": 0.0020640000000000003,
1460
+ "loss": 2.1436,
1461
+ "step": 3146
1462
+ },
1463
+ {
1464
+ "epoch": 243.0,
1465
+ "learning_rate": 0.002056,
1466
+ "loss": 2.1698,
1467
+ "step": 3159
1468
+ },
1469
+ {
1470
+ "epoch": 244.0,
1471
+ "learning_rate": 0.002048,
1472
+ "loss": 2.1576,
1473
+ "step": 3172
1474
+ },
1475
+ {
1476
+ "epoch": 245.0,
1477
+ "learning_rate": 0.00204,
1478
+ "loss": 2.1641,
1479
+ "step": 3185
1480
+ },
1481
+ {
1482
+ "epoch": 246.0,
1483
+ "learning_rate": 0.002032,
1484
+ "loss": 2.1721,
1485
+ "step": 3198
1486
+ },
1487
+ {
1488
+ "epoch": 247.0,
1489
+ "learning_rate": 0.002024,
1490
+ "loss": 2.1615,
1491
+ "step": 3211
1492
+ },
1493
+ {
1494
+ "epoch": 248.0,
1495
+ "learning_rate": 0.002016,
1496
+ "loss": 2.0983,
1497
+ "step": 3224
1498
+ },
1499
+ {
1500
+ "epoch": 249.0,
1501
+ "learning_rate": 0.0020080000000000002,
1502
+ "loss": 2.108,
1503
+ "step": 3237
1504
+ },
1505
+ {
1506
+ "epoch": 250.0,
1507
+ "learning_rate": 0.002,
1508
+ "loss": 2.1167,
1509
+ "step": 3250
1510
+ },
1511
+ {
1512
+ "epoch": 251.0,
1513
+ "learning_rate": 0.001992,
1514
+ "loss": 2.0951,
1515
+ "step": 3263
1516
+ },
1517
+ {
1518
+ "epoch": 252.0,
1519
+ "learning_rate": 0.001984,
1520
+ "loss": 2.0415,
1521
+ "step": 3276
1522
+ },
1523
+ {
1524
+ "epoch": 253.0,
1525
+ "learning_rate": 0.001976,
1526
+ "loss": 2.101,
1527
+ "step": 3289
1528
+ },
1529
+ {
1530
+ "epoch": 254.0,
1531
+ "learning_rate": 0.001968,
1532
+ "loss": 2.1233,
1533
+ "step": 3302
1534
+ },
1535
+ {
1536
+ "epoch": 255.0,
1537
+ "learning_rate": 0.00196,
1538
+ "loss": 2.0782,
1539
+ "step": 3315
1540
+ },
1541
+ {
1542
+ "epoch": 256.0,
1543
+ "learning_rate": 0.001952,
1544
+ "loss": 2.0033,
1545
+ "step": 3328
1546
+ },
1547
+ {
1548
+ "epoch": 257.0,
1549
+ "learning_rate": 0.001944,
1550
+ "loss": 2.051,
1551
+ "step": 3341
1552
+ },
1553
+ {
1554
+ "epoch": 258.0,
1555
+ "learning_rate": 0.001936,
1556
+ "loss": 2.0587,
1557
+ "step": 3354
1558
+ },
1559
+ {
1560
+ "epoch": 259.0,
1561
+ "learning_rate": 0.001928,
1562
+ "loss": 1.9981,
1563
+ "step": 3367
1564
+ },
1565
+ {
1566
+ "epoch": 260.0,
1567
+ "learning_rate": 0.00192,
1568
+ "loss": 2.0506,
1569
+ "step": 3380
1570
+ },
1571
+ {
1572
+ "epoch": 261.0,
1573
+ "learning_rate": 0.0019119999999999999,
1574
+ "loss": 2.0815,
1575
+ "step": 3393
1576
+ },
1577
+ {
1578
+ "epoch": 262.0,
1579
+ "learning_rate": 0.0019039999999999999,
1580
+ "loss": 2.0054,
1581
+ "step": 3406
1582
+ },
1583
+ {
1584
+ "epoch": 263.0,
1585
+ "learning_rate": 0.001896,
1586
+ "loss": 1.9923,
1587
+ "step": 3419
1588
+ },
1589
+ {
1590
+ "epoch": 264.0,
1591
+ "learning_rate": 0.001888,
1592
+ "loss": 1.9892,
1593
+ "step": 3432
1594
+ },
1595
+ {
1596
+ "epoch": 265.0,
1597
+ "learning_rate": 0.00188,
1598
+ "loss": 1.9406,
1599
+ "step": 3445
1600
+ },
1601
+ {
1602
+ "epoch": 266.0,
1603
+ "learning_rate": 0.0018720000000000002,
1604
+ "loss": 1.9295,
1605
+ "step": 3458
1606
+ },
1607
+ {
1608
+ "epoch": 267.0,
1609
+ "learning_rate": 0.0018640000000000002,
1610
+ "loss": 1.9791,
1611
+ "step": 3471
1612
+ },
1613
+ {
1614
+ "epoch": 268.0,
1615
+ "learning_rate": 0.0018560000000000002,
1616
+ "loss": 1.9413,
1617
+ "step": 3484
1618
+ },
1619
+ {
1620
+ "epoch": 269.0,
1621
+ "learning_rate": 0.001848,
1622
+ "loss": 1.9363,
1623
+ "step": 3497
1624
+ },
1625
+ {
1626
+ "epoch": 270.0,
1627
+ "learning_rate": 0.00184,
1628
+ "loss": 2.0056,
1629
+ "step": 3510
1630
+ },
1631
+ {
1632
+ "epoch": 271.0,
1633
+ "learning_rate": 0.001832,
1634
+ "loss": 1.9298,
1635
+ "step": 3523
1636
+ },
1637
+ {
1638
+ "epoch": 272.0,
1639
+ "learning_rate": 0.001824,
1640
+ "loss": 1.9045,
1641
+ "step": 3536
1642
+ },
1643
+ {
1644
+ "epoch": 273.0,
1645
+ "learning_rate": 0.0018160000000000001,
1646
+ "loss": 1.9165,
1647
+ "step": 3549
1648
+ },
1649
+ {
1650
+ "epoch": 274.0,
1651
+ "learning_rate": 0.0018080000000000001,
1652
+ "loss": 1.9214,
1653
+ "step": 3562
1654
+ },
1655
+ {
1656
+ "epoch": 275.0,
1657
+ "learning_rate": 0.0018000000000000002,
1658
+ "loss": 1.9063,
1659
+ "step": 3575
1660
+ },
1661
+ {
1662
+ "epoch": 276.0,
1663
+ "learning_rate": 0.001792,
1664
+ "loss": 1.9016,
1665
+ "step": 3588
1666
+ },
1667
+ {
1668
+ "epoch": 277.0,
1669
+ "learning_rate": 0.001784,
1670
+ "loss": 1.8091,
1671
+ "step": 3601
1672
+ },
1673
+ {
1674
+ "epoch": 278.0,
1675
+ "learning_rate": 0.001776,
1676
+ "loss": 1.8626,
1677
+ "step": 3614
1678
+ },
1679
+ {
1680
+ "epoch": 279.0,
1681
+ "learning_rate": 0.001768,
1682
+ "loss": 1.8663,
1683
+ "step": 3627
1684
+ },
1685
+ {
1686
+ "epoch": 280.0,
1687
+ "learning_rate": 0.00176,
1688
+ "loss": 1.9432,
1689
+ "step": 3640
1690
+ },
1691
+ {
1692
+ "epoch": 281.0,
1693
+ "learning_rate": 0.001752,
1694
+ "loss": 1.8664,
1695
+ "step": 3653
1696
+ },
1697
+ {
1698
+ "epoch": 282.0,
1699
+ "learning_rate": 0.001744,
1700
+ "loss": 1.8603,
1701
+ "step": 3666
1702
+ },
1703
+ {
1704
+ "epoch": 283.0,
1705
+ "learning_rate": 0.0017360000000000001,
1706
+ "loss": 1.8335,
1707
+ "step": 3679
1708
+ },
1709
+ {
1710
+ "epoch": 284.0,
1711
+ "learning_rate": 0.001728,
1712
+ "loss": 1.8625,
1713
+ "step": 3692
1714
+ },
1715
+ {
1716
+ "epoch": 285.0,
1717
+ "learning_rate": 0.00172,
1718
+ "loss": 1.8043,
1719
+ "step": 3705
1720
+ },
1721
+ {
1722
+ "epoch": 286.0,
1723
+ "learning_rate": 0.001712,
1724
+ "loss": 1.8061,
1725
+ "step": 3718
1726
+ },
1727
+ {
1728
+ "epoch": 287.0,
1729
+ "learning_rate": 0.001704,
1730
+ "loss": 1.835,
1731
+ "step": 3731
1732
+ },
1733
+ {
1734
+ "epoch": 288.0,
1735
+ "learning_rate": 0.001696,
1736
+ "loss": 1.7944,
1737
+ "step": 3744
1738
+ },
1739
+ {
1740
+ "epoch": 289.0,
1741
+ "learning_rate": 0.001688,
1742
+ "loss": 1.8492,
1743
+ "step": 3757
1744
+ },
1745
+ {
1746
+ "epoch": 290.0,
1747
+ "learning_rate": 0.00168,
1748
+ "loss": 1.812,
1749
+ "step": 3770
1750
+ },
1751
+ {
1752
+ "epoch": 291.0,
1753
+ "learning_rate": 0.0016719999999999999,
1754
+ "loss": 1.8175,
1755
+ "step": 3783
1756
+ },
1757
+ {
1758
+ "epoch": 292.0,
1759
+ "learning_rate": 0.001664,
1760
+ "loss": 1.7943,
1761
+ "step": 3796
1762
+ },
1763
+ {
1764
+ "epoch": 293.0,
1765
+ "learning_rate": 0.001656,
1766
+ "loss": 1.8063,
1767
+ "step": 3809
1768
+ },
1769
+ {
1770
+ "epoch": 294.0,
1771
+ "learning_rate": 0.001648,
1772
+ "loss": 1.7992,
1773
+ "step": 3822
1774
+ },
1775
+ {
1776
+ "epoch": 295.0,
1777
+ "learning_rate": 0.00164,
1778
+ "loss": 1.7959,
1779
+ "step": 3835
1780
+ },
1781
+ {
1782
+ "epoch": 296.0,
1783
+ "learning_rate": 0.001632,
1784
+ "loss": 1.7256,
1785
+ "step": 3848
1786
+ },
1787
+ {
1788
+ "epoch": 297.0,
1789
+ "learning_rate": 0.0016240000000000002,
1790
+ "loss": 1.7673,
1791
+ "step": 3861
1792
+ },
1793
+ {
1794
+ "epoch": 298.0,
1795
+ "learning_rate": 0.001616,
1796
+ "loss": 1.8299,
1797
+ "step": 3874
1798
+ },
1799
+ {
1800
+ "epoch": 299.0,
1801
+ "learning_rate": 0.001608,
1802
+ "loss": 1.8147,
1803
+ "step": 3887
1804
+ },
1805
+ {
1806
+ "epoch": 300.0,
1807
+ "learning_rate": 0.0016,
1808
+ "loss": 1.7495,
1809
+ "step": 3900
1810
+ },
1811
+ {
1812
+ "epoch": 301.0,
1813
+ "learning_rate": 0.001592,
1814
+ "loss": 1.8001,
1815
+ "step": 3913
1816
+ },
1817
+ {
1818
+ "epoch": 302.0,
1819
+ "learning_rate": 0.0015840000000000001,
1820
+ "loss": 1.7707,
1821
+ "step": 3926
1822
+ },
1823
+ {
1824
+ "epoch": 303.0,
1825
+ "learning_rate": 0.0015760000000000001,
1826
+ "loss": 1.7283,
1827
+ "step": 3939
1828
+ },
1829
+ {
1830
+ "epoch": 304.0,
1831
+ "learning_rate": 0.0015680000000000002,
1832
+ "loss": 1.7133,
1833
+ "step": 3952
1834
+ },
1835
+ {
1836
+ "epoch": 305.0,
1837
+ "learning_rate": 0.0015600000000000002,
1838
+ "loss": 1.71,
1839
+ "step": 3965
1840
+ },
1841
+ {
1842
+ "epoch": 306.0,
1843
+ "learning_rate": 0.001552,
1844
+ "loss": 1.6685,
1845
+ "step": 3978
1846
+ },
1847
+ {
1848
+ "epoch": 307.0,
1849
+ "learning_rate": 0.001544,
1850
+ "loss": 1.6526,
1851
+ "step": 3991
1852
+ },
1853
+ {
1854
+ "epoch": 308.0,
1855
+ "learning_rate": 0.001536,
1856
+ "loss": 1.6433,
1857
+ "step": 4004
1858
+ },
1859
+ {
1860
+ "epoch": 309.0,
1861
+ "learning_rate": 0.001528,
1862
+ "loss": 1.6823,
1863
+ "step": 4017
1864
+ },
1865
+ {
1866
+ "epoch": 310.0,
1867
+ "learning_rate": 0.00152,
1868
+ "loss": 1.6843,
1869
+ "step": 4030
1870
+ },
1871
+ {
1872
+ "epoch": 311.0,
1873
+ "learning_rate": 0.001512,
1874
+ "loss": 1.7029,
1875
+ "step": 4043
1876
+ },
1877
+ {
1878
+ "epoch": 312.0,
1879
+ "learning_rate": 0.0015040000000000001,
1880
+ "loss": 1.6362,
1881
+ "step": 4056
1882
+ },
1883
+ {
1884
+ "epoch": 313.0,
1885
+ "learning_rate": 0.001496,
1886
+ "loss": 1.6648,
1887
+ "step": 4069
1888
+ },
1889
+ {
1890
+ "epoch": 314.0,
1891
+ "learning_rate": 0.001488,
1892
+ "loss": 1.7202,
1893
+ "step": 4082
1894
+ },
1895
+ {
1896
+ "epoch": 315.0,
1897
+ "learning_rate": 0.00148,
1898
+ "loss": 1.677,
1899
+ "step": 4095
1900
+ },
1901
+ {
1902
+ "epoch": 316.0,
1903
+ "learning_rate": 0.001472,
1904
+ "loss": 1.6187,
1905
+ "step": 4108
1906
+ },
1907
+ {
1908
+ "epoch": 317.0,
1909
+ "learning_rate": 0.001464,
1910
+ "loss": 1.6398,
1911
+ "step": 4121
1912
+ },
1913
+ {
1914
+ "epoch": 318.0,
1915
+ "learning_rate": 0.001456,
1916
+ "loss": 1.6371,
1917
+ "step": 4134
1918
+ },
1919
+ {
1920
+ "epoch": 319.0,
1921
+ "learning_rate": 0.001448,
1922
+ "loss": 1.6081,
1923
+ "step": 4147
1924
+ },
1925
+ {
1926
+ "epoch": 320.0,
1927
+ "learning_rate": 0.0014399999999999999,
1928
+ "loss": 1.5936,
1929
+ "step": 4160
1930
+ },
1931
+ {
1932
+ "epoch": 321.0,
1933
+ "learning_rate": 0.001432,
1934
+ "loss": 1.6336,
1935
+ "step": 4173
1936
+ },
1937
+ {
1938
+ "epoch": 322.0,
1939
+ "learning_rate": 0.001424,
1940
+ "loss": 1.6022,
1941
+ "step": 4186
1942
+ },
1943
+ {
1944
+ "epoch": 323.0,
1945
+ "learning_rate": 0.001416,
1946
+ "loss": 1.6336,
1947
+ "step": 4199
1948
+ },
1949
+ {
1950
+ "epoch": 324.0,
1951
+ "learning_rate": 0.001408,
1952
+ "loss": 1.5898,
1953
+ "step": 4212
1954
+ },
1955
+ {
1956
+ "epoch": 325.0,
1957
+ "learning_rate": 0.0014,
1958
+ "loss": 1.5528,
1959
+ "step": 4225
1960
+ },
1961
+ {
1962
+ "epoch": 326.0,
1963
+ "learning_rate": 0.001392,
1964
+ "loss": 1.5734,
1965
+ "step": 4238
1966
+ },
1967
+ {
1968
+ "epoch": 327.0,
1969
+ "learning_rate": 0.001384,
1970
+ "loss": 1.618,
1971
+ "step": 4251
1972
+ },
1973
+ {
1974
+ "epoch": 328.0,
1975
+ "learning_rate": 0.0013759999999999998,
1976
+ "loss": 1.6529,
1977
+ "step": 4264
1978
+ },
1979
+ {
1980
+ "epoch": 329.0,
1981
+ "learning_rate": 0.001368,
1982
+ "loss": 1.5824,
1983
+ "step": 4277
1984
+ },
1985
+ {
1986
+ "epoch": 330.0,
1987
+ "learning_rate": 0.00136,
1988
+ "loss": 1.609,
1989
+ "step": 4290
1990
+ },
1991
+ {
1992
+ "epoch": 331.0,
1993
+ "learning_rate": 0.0013520000000000001,
1994
+ "loss": 1.5796,
1995
+ "step": 4303
1996
+ },
1997
+ {
1998
+ "epoch": 332.0,
1999
+ "learning_rate": 0.0013440000000000001,
2000
+ "loss": 1.5924,
2001
+ "step": 4316
2002
+ },
2003
+ {
2004
+ "epoch": 333.0,
2005
+ "learning_rate": 0.0013360000000000002,
2006
+ "loss": 1.5841,
2007
+ "step": 4329
2008
+ },
2009
+ {
2010
+ "epoch": 334.0,
2011
+ "learning_rate": 0.0013280000000000002,
2012
+ "loss": 1.5487,
2013
+ "step": 4342
2014
+ },
2015
+ {
2016
+ "epoch": 335.0,
2017
+ "learning_rate": 0.00132,
2018
+ "loss": 1.4625,
2019
+ "step": 4355
2020
+ },
2021
+ {
2022
+ "epoch": 336.0,
2023
+ "learning_rate": 0.001312,
2024
+ "loss": 1.5241,
2025
+ "step": 4368
2026
+ },
2027
+ {
2028
+ "epoch": 337.0,
2029
+ "learning_rate": 0.001304,
2030
+ "loss": 1.4823,
2031
+ "step": 4381
2032
+ },
2033
+ {
2034
+ "epoch": 338.0,
2035
+ "learning_rate": 0.001296,
2036
+ "loss": 1.5027,
2037
+ "step": 4394
2038
+ },
2039
+ {
2040
+ "epoch": 339.0,
2041
+ "learning_rate": 0.001288,
2042
+ "loss": 1.5211,
2043
+ "step": 4407
2044
+ },
2045
+ {
2046
+ "epoch": 340.0,
2047
+ "learning_rate": 0.00128,
2048
+ "loss": 1.4912,
2049
+ "step": 4420
2050
+ },
2051
+ {
2052
+ "epoch": 341.0,
2053
+ "learning_rate": 0.0012720000000000001,
2054
+ "loss": 1.4792,
2055
+ "step": 4433
2056
+ },
2057
+ {
2058
+ "epoch": 342.0,
2059
+ "learning_rate": 0.001264,
2060
+ "loss": 1.4932,
2061
+ "step": 4446
2062
+ },
2063
+ {
2064
+ "epoch": 343.0,
2065
+ "learning_rate": 0.001256,
2066
+ "loss": 1.4861,
2067
+ "step": 4459
2068
+ },
2069
+ {
2070
+ "epoch": 344.0,
2071
+ "learning_rate": 0.001248,
2072
+ "loss": 1.5171,
2073
+ "step": 4472
2074
+ },
2075
+ {
2076
+ "epoch": 345.0,
2077
+ "learning_rate": 0.00124,
2078
+ "loss": 1.494,
2079
+ "step": 4485
2080
+ },
2081
+ {
2082
+ "epoch": 346.0,
2083
+ "learning_rate": 0.001232,
2084
+ "loss": 1.4992,
2085
+ "step": 4498
2086
+ },
2087
+ {
2088
+ "epoch": 347.0,
2089
+ "learning_rate": 0.001224,
2090
+ "loss": 1.5033,
2091
+ "step": 4511
2092
+ },
2093
+ {
2094
+ "epoch": 348.0,
2095
+ "learning_rate": 0.001216,
2096
+ "loss": 1.5039,
2097
+ "step": 4524
2098
+ },
2099
+ {
2100
+ "epoch": 349.0,
2101
+ "learning_rate": 0.001208,
2102
+ "loss": 1.5341,
2103
+ "step": 4537
2104
+ },
2105
+ {
2106
+ "epoch": 350.0,
2107
+ "learning_rate": 0.0012,
2108
+ "loss": 1.5049,
2109
+ "step": 4550
2110
+ },
2111
+ {
2112
+ "epoch": 351.0,
2113
+ "learning_rate": 0.001192,
2114
+ "loss": 1.5104,
2115
+ "step": 4563
2116
+ },
2117
+ {
2118
+ "epoch": 352.0,
2119
+ "learning_rate": 0.001184,
2120
+ "loss": 1.4569,
2121
+ "step": 4576
2122
+ },
2123
+ {
2124
+ "epoch": 353.0,
2125
+ "learning_rate": 0.001176,
2126
+ "loss": 1.3996,
2127
+ "step": 4589
2128
+ },
2129
+ {
2130
+ "epoch": 354.0,
2131
+ "learning_rate": 0.001168,
2132
+ "loss": 1.4337,
2133
+ "step": 4602
2134
+ },
2135
+ {
2136
+ "epoch": 355.0,
2137
+ "learning_rate": 0.00116,
2138
+ "loss": 1.4572,
2139
+ "step": 4615
2140
+ },
2141
+ {
2142
+ "epoch": 356.0,
2143
+ "learning_rate": 0.001152,
2144
+ "loss": 1.4668,
2145
+ "step": 4628
2146
+ },
2147
+ {
2148
+ "epoch": 357.0,
2149
+ "learning_rate": 0.0011439999999999998,
2150
+ "loss": 1.4298,
2151
+ "step": 4641
2152
+ },
2153
+ {
2154
+ "epoch": 358.0,
2155
+ "learning_rate": 0.0011359999999999999,
2156
+ "loss": 1.4187,
2157
+ "step": 4654
2158
+ },
2159
+ {
2160
+ "epoch": 359.0,
2161
+ "learning_rate": 0.0011279999999999999,
2162
+ "loss": 1.4026,
2163
+ "step": 4667
2164
+ },
2165
+ {
2166
+ "epoch": 360.0,
2167
+ "learning_rate": 0.0011200000000000001,
2168
+ "loss": 1.4461,
2169
+ "step": 4680
2170
+ },
2171
+ {
2172
+ "epoch": 361.0,
2173
+ "learning_rate": 0.0011120000000000001,
2174
+ "loss": 1.4497,
2175
+ "step": 4693
2176
+ },
2177
+ {
2178
+ "epoch": 362.0,
2179
+ "learning_rate": 0.0011040000000000002,
2180
+ "loss": 1.3667,
2181
+ "step": 4706
2182
+ },
2183
+ {
2184
+ "epoch": 363.0,
2185
+ "learning_rate": 0.0010960000000000002,
2186
+ "loss": 1.4237,
2187
+ "step": 4719
2188
+ },
2189
+ {
2190
+ "epoch": 364.0,
2191
+ "learning_rate": 0.0010880000000000002,
2192
+ "loss": 1.485,
2193
+ "step": 4732
2194
+ },
2195
+ {
2196
+ "epoch": 365.0,
2197
+ "learning_rate": 0.00108,
2198
+ "loss": 1.4271,
2199
+ "step": 4745
2200
+ },
2201
+ {
2202
+ "epoch": 366.0,
2203
+ "learning_rate": 0.001072,
2204
+ "loss": 1.4046,
2205
+ "step": 4758
2206
+ },
2207
+ {
2208
+ "epoch": 367.0,
2209
+ "learning_rate": 0.001064,
2210
+ "loss": 1.3771,
2211
+ "step": 4771
2212
+ },
2213
+ {
2214
+ "epoch": 368.0,
2215
+ "learning_rate": 0.001056,
2216
+ "loss": 1.4054,
2217
+ "step": 4784
2218
+ },
2219
+ {
2220
+ "epoch": 369.0,
2221
+ "learning_rate": 0.001048,
2222
+ "loss": 1.3886,
2223
+ "step": 4797
2224
+ },
2225
+ {
2226
+ "epoch": 370.0,
2227
+ "learning_rate": 0.0010400000000000001,
2228
+ "loss": 1.3583,
2229
+ "step": 4810
2230
+ },
2231
+ {
2232
+ "epoch": 371.0,
2233
+ "learning_rate": 0.0010320000000000001,
2234
+ "loss": 1.3606,
2235
+ "step": 4823
2236
+ },
2237
+ {
2238
+ "epoch": 372.0,
2239
+ "learning_rate": 0.001024,
2240
+ "loss": 1.3619,
2241
+ "step": 4836
2242
+ },
2243
+ {
2244
+ "epoch": 373.0,
2245
+ "learning_rate": 0.001016,
2246
+ "loss": 1.3723,
2247
+ "step": 4849
2248
+ },
2249
+ {
2250
+ "epoch": 374.0,
2251
+ "learning_rate": 0.001008,
2252
+ "loss": 1.3604,
2253
+ "step": 4862
2254
+ },
2255
+ {
2256
+ "epoch": 375.0,
2257
+ "learning_rate": 0.001,
2258
+ "loss": 1.3745,
2259
+ "step": 4875
2260
+ },
2261
+ {
2262
+ "epoch": 376.0,
2263
+ "learning_rate": 0.000992,
2264
+ "loss": 1.393,
2265
+ "step": 4888
2266
+ },
2267
+ {
2268
+ "epoch": 377.0,
2269
+ "learning_rate": 0.000984,
2270
+ "loss": 1.3846,
2271
+ "step": 4901
2272
+ },
2273
+ {
2274
+ "epoch": 378.0,
2275
+ "learning_rate": 0.000976,
2276
+ "loss": 1.4033,
2277
+ "step": 4914
2278
+ },
2279
+ {
2280
+ "epoch": 379.0,
2281
+ "learning_rate": 0.000968,
2282
+ "loss": 1.3204,
2283
+ "step": 4927
2284
+ },
2285
+ {
2286
+ "epoch": 380.0,
2287
+ "learning_rate": 0.00096,
2288
+ "loss": 1.3257,
2289
+ "step": 4940
2290
+ },
2291
+ {
2292
+ "epoch": 381.0,
2293
+ "learning_rate": 0.0009519999999999999,
2294
+ "loss": 1.3274,
2295
+ "step": 4953
2296
+ },
2297
+ {
2298
+ "epoch": 382.0,
2299
+ "learning_rate": 0.000944,
2300
+ "loss": 1.3177,
2301
+ "step": 4966
2302
+ },
2303
+ {
2304
+ "epoch": 383.0,
2305
+ "learning_rate": 0.0009360000000000001,
2306
+ "loss": 1.3204,
2307
+ "step": 4979
2308
+ },
2309
+ {
2310
+ "epoch": 384.0,
2311
+ "learning_rate": 0.0009280000000000001,
2312
+ "loss": 1.3349,
2313
+ "step": 4992
2314
+ },
2315
+ {
2316
+ "epoch": 385.0,
2317
+ "learning_rate": 0.00092,
2318
+ "loss": 1.3149,
2319
+ "step": 5005
2320
+ },
2321
+ {
2322
+ "epoch": 386.0,
2323
+ "learning_rate": 0.000912,
2324
+ "loss": 1.2994,
2325
+ "step": 5018
2326
+ },
2327
+ {
2328
+ "epoch": 387.0,
2329
+ "learning_rate": 0.0009040000000000001,
2330
+ "loss": 1.3295,
2331
+ "step": 5031
2332
+ },
2333
+ {
2334
+ "epoch": 388.0,
2335
+ "learning_rate": 0.000896,
2336
+ "loss": 1.2975,
2337
+ "step": 5044
2338
+ },
2339
+ {
2340
+ "epoch": 389.0,
2341
+ "learning_rate": 0.000888,
2342
+ "loss": 1.3118,
2343
+ "step": 5057
2344
+ },
2345
+ {
2346
+ "epoch": 390.0,
2347
+ "learning_rate": 0.00088,
2348
+ "loss": 1.2712,
2349
+ "step": 5070
2350
+ },
2351
+ {
2352
+ "epoch": 391.0,
2353
+ "learning_rate": 0.000872,
2354
+ "loss": 1.3184,
2355
+ "step": 5083
2356
+ },
2357
+ {
2358
+ "epoch": 392.0,
2359
+ "learning_rate": 0.000864,
2360
+ "loss": 1.2687,
2361
+ "step": 5096
2362
+ },
2363
+ {
2364
+ "epoch": 393.0,
2365
+ "learning_rate": 0.000856,
2366
+ "loss": 1.2826,
2367
+ "step": 5109
2368
+ },
2369
+ {
2370
+ "epoch": 394.0,
2371
+ "learning_rate": 0.000848,
2372
+ "loss": 1.2766,
2373
+ "step": 5122
2374
+ },
2375
+ {
2376
+ "epoch": 395.0,
2377
+ "learning_rate": 0.00084,
2378
+ "loss": 1.2935,
2379
+ "step": 5135
2380
+ },
2381
+ {
2382
+ "epoch": 396.0,
2383
+ "learning_rate": 0.000832,
2384
+ "loss": 1.288,
2385
+ "step": 5148
2386
+ },
2387
+ {
2388
+ "epoch": 397.0,
2389
+ "learning_rate": 0.000824,
2390
+ "loss": 1.2617,
2391
+ "step": 5161
2392
+ },
2393
+ {
2394
+ "epoch": 398.0,
2395
+ "learning_rate": 0.000816,
2396
+ "loss": 1.2675,
2397
+ "step": 5174
2398
+ },
2399
+ {
2400
+ "epoch": 399.0,
2401
+ "learning_rate": 0.000808,
2402
+ "loss": 1.2895,
2403
+ "step": 5187
2404
+ },
2405
+ {
2406
+ "epoch": 400.0,
2407
+ "learning_rate": 0.0008,
2408
+ "loss": 1.2721,
2409
+ "step": 5200
2410
+ },
2411
+ {
2412
+ "epoch": 401.0,
2413
+ "learning_rate": 0.0007920000000000001,
2414
+ "loss": 1.2897,
2415
+ "step": 5213
2416
+ },
2417
+ {
2418
+ "epoch": 402.0,
2419
+ "learning_rate": 0.0007840000000000001,
2420
+ "loss": 1.2608,
2421
+ "step": 5226
2422
+ },
2423
+ {
2424
+ "epoch": 403.0,
2425
+ "learning_rate": 0.000776,
2426
+ "loss": 1.271,
2427
+ "step": 5239
2428
+ },
2429
+ {
2430
+ "epoch": 404.0,
2431
+ "learning_rate": 0.000768,
2432
+ "loss": 1.2581,
2433
+ "step": 5252
2434
+ },
2435
+ {
2436
+ "epoch": 405.0,
2437
+ "learning_rate": 0.00076,
2438
+ "loss": 1.2497,
2439
+ "step": 5265
2440
+ },
2441
+ {
2442
+ "epoch": 406.0,
2443
+ "learning_rate": 0.0007520000000000001,
2444
+ "loss": 1.2846,
2445
+ "step": 5278
2446
+ },
2447
+ {
2448
+ "epoch": 407.0,
2449
+ "learning_rate": 0.000744,
2450
+ "loss": 1.2718,
2451
+ "step": 5291
2452
+ },
2453
+ {
2454
+ "epoch": 408.0,
2455
+ "learning_rate": 0.000736,
2456
+ "loss": 1.2733,
2457
+ "step": 5304
2458
+ },
2459
+ {
2460
+ "epoch": 409.0,
2461
+ "learning_rate": 0.000728,
2462
+ "loss": 1.2918,
2463
+ "step": 5317
2464
+ },
2465
+ {
2466
+ "epoch": 410.0,
2467
+ "learning_rate": 0.0007199999999999999,
2468
+ "loss": 1.2659,
2469
+ "step": 5330
2470
+ },
2471
+ {
2472
+ "epoch": 411.0,
2473
+ "learning_rate": 0.000712,
2474
+ "loss": 1.2946,
2475
+ "step": 5343
2476
+ },
2477
+ {
2478
+ "epoch": 412.0,
2479
+ "learning_rate": 0.000704,
2480
+ "loss": 1.2425,
2481
+ "step": 5356
2482
+ },
2483
+ {
2484
+ "epoch": 413.0,
2485
+ "learning_rate": 0.000696,
2486
+ "loss": 1.2293,
2487
+ "step": 5369
2488
+ },
2489
+ {
2490
+ "epoch": 414.0,
2491
+ "learning_rate": 0.0006879999999999999,
2492
+ "loss": 1.2847,
2493
+ "step": 5382
2494
+ },
2495
+ {
2496
+ "epoch": 415.0,
2497
+ "learning_rate": 0.00068,
2498
+ "loss": 1.2318,
2499
+ "step": 5395
2500
+ },
2501
+ {
2502
+ "epoch": 416.0,
2503
+ "learning_rate": 0.0006720000000000001,
2504
+ "loss": 1.237,
2505
+ "step": 5408
2506
+ },
2507
+ {
2508
+ "epoch": 417.0,
2509
+ "learning_rate": 0.0006640000000000001,
2510
+ "loss": 1.1875,
2511
+ "step": 5421
2512
+ },
2513
+ {
2514
+ "epoch": 418.0,
2515
+ "learning_rate": 0.000656,
2516
+ "loss": 1.2204,
2517
+ "step": 5434
2518
+ },
2519
+ {
2520
+ "epoch": 419.0,
2521
+ "learning_rate": 0.000648,
2522
+ "loss": 1.1848,
2523
+ "step": 5447
2524
+ },
2525
+ {
2526
+ "epoch": 420.0,
2527
+ "learning_rate": 0.00064,
2528
+ "loss": 1.2146,
2529
+ "step": 5460
2530
+ },
2531
+ {
2532
+ "epoch": 421.0,
2533
+ "learning_rate": 0.000632,
2534
+ "loss": 1.1621,
2535
+ "step": 5473
2536
+ },
2537
+ {
2538
+ "epoch": 422.0,
2539
+ "learning_rate": 0.000624,
2540
+ "loss": 1.1883,
2541
+ "step": 5486
2542
+ },
2543
+ {
2544
+ "epoch": 423.0,
2545
+ "learning_rate": 0.000616,
2546
+ "loss": 1.183,
2547
+ "step": 5499
2548
+ },
2549
+ {
2550
+ "epoch": 424.0,
2551
+ "learning_rate": 0.000608,
2552
+ "loss": 1.1649,
2553
+ "step": 5512
2554
+ },
2555
+ {
2556
+ "epoch": 425.0,
2557
+ "learning_rate": 0.0006,
2558
+ "loss": 1.1824,
2559
+ "step": 5525
2560
+ },
2561
+ {
2562
+ "epoch": 426.0,
2563
+ "learning_rate": 0.000592,
2564
+ "loss": 1.2073,
2565
+ "step": 5538
2566
+ },
2567
+ {
2568
+ "epoch": 427.0,
2569
+ "learning_rate": 0.000584,
2570
+ "loss": 1.147,
2571
+ "step": 5551
2572
+ },
2573
+ {
2574
+ "epoch": 428.0,
2575
+ "learning_rate": 0.000576,
2576
+ "loss": 1.1798,
2577
+ "step": 5564
2578
+ },
2579
+ {
2580
+ "epoch": 429.0,
2581
+ "learning_rate": 0.0005679999999999999,
2582
+ "loss": 1.14,
2583
+ "step": 5577
2584
+ },
2585
+ {
2586
+ "epoch": 430.0,
2587
+ "learning_rate": 0.0005600000000000001,
2588
+ "loss": 1.1585,
2589
+ "step": 5590
2590
+ },
2591
+ {
2592
+ "epoch": 431.0,
2593
+ "learning_rate": 0.0005520000000000001,
2594
+ "loss": 1.1687,
2595
+ "step": 5603
2596
+ },
2597
+ {
2598
+ "epoch": 432.0,
2599
+ "learning_rate": 0.0005440000000000001,
2600
+ "loss": 1.1285,
2601
+ "step": 5616
2602
+ },
2603
+ {
2604
+ "epoch": 433.0,
2605
+ "learning_rate": 0.000536,
2606
+ "loss": 1.1472,
2607
+ "step": 5629
2608
+ },
2609
+ {
2610
+ "epoch": 434.0,
2611
+ "learning_rate": 0.000528,
2612
+ "loss": 1.1894,
2613
+ "step": 5642
2614
+ },
2615
+ {
2616
+ "epoch": 435.0,
2617
+ "learning_rate": 0.0005200000000000001,
2618
+ "loss": 1.1606,
2619
+ "step": 5655
2620
+ },
2621
+ {
2622
+ "epoch": 436.0,
2623
+ "learning_rate": 0.000512,
2624
+ "loss": 1.1294,
2625
+ "step": 5668
2626
+ },
2627
+ {
2628
+ "epoch": 437.0,
2629
+ "learning_rate": 0.000504,
2630
+ "loss": 1.1597,
2631
+ "step": 5681
2632
+ },
2633
+ {
2634
+ "epoch": 438.0,
2635
+ "learning_rate": 0.000496,
2636
+ "loss": 1.1772,
2637
+ "step": 5694
2638
+ },
2639
+ {
2640
+ "epoch": 439.0,
2641
+ "learning_rate": 0.000488,
2642
+ "loss": 1.2044,
2643
+ "step": 5707
2644
+ },
2645
+ {
2646
+ "epoch": 440.0,
2647
+ "learning_rate": 0.00048,
2648
+ "loss": 1.1543,
2649
+ "step": 5720
2650
+ },
2651
+ {
2652
+ "epoch": 441.0,
2653
+ "learning_rate": 0.000472,
2654
+ "loss": 1.1868,
2655
+ "step": 5733
2656
+ },
2657
+ {
2658
+ "epoch": 442.0,
2659
+ "learning_rate": 0.00046400000000000006,
2660
+ "loss": 1.1821,
2661
+ "step": 5746
2662
+ },
2663
+ {
2664
+ "epoch": 443.0,
2665
+ "learning_rate": 0.000456,
2666
+ "loss": 1.0897,
2667
+ "step": 5759
2668
+ },
2669
+ {
2670
+ "epoch": 444.0,
2671
+ "learning_rate": 0.000448,
2672
+ "loss": 1.0977,
2673
+ "step": 5772
2674
+ },
2675
+ {
2676
+ "epoch": 445.0,
2677
+ "learning_rate": 0.00044,
2678
+ "loss": 1.1695,
2679
+ "step": 5785
2680
+ },
2681
+ {
2682
+ "epoch": 446.0,
2683
+ "learning_rate": 0.000432,
2684
+ "loss": 1.1332,
2685
+ "step": 5798
2686
+ },
2687
+ {
2688
+ "epoch": 447.0,
2689
+ "learning_rate": 0.000424,
2690
+ "loss": 1.1321,
2691
+ "step": 5811
2692
+ },
2693
+ {
2694
+ "epoch": 448.0,
2695
+ "learning_rate": 0.000416,
2696
+ "loss": 1.1315,
2697
+ "step": 5824
2698
+ },
2699
+ {
2700
+ "epoch": 449.0,
2701
+ "learning_rate": 0.000408,
2702
+ "loss": 1.1178,
2703
+ "step": 5837
2704
+ },
2705
+ {
2706
+ "epoch": 450.0,
2707
+ "learning_rate": 0.0004,
2708
+ "loss": 1.1163,
2709
+ "step": 5850
2710
+ },
2711
+ {
2712
+ "epoch": 451.0,
2713
+ "learning_rate": 0.00039200000000000004,
2714
+ "loss": 1.1414,
2715
+ "step": 5863
2716
+ },
2717
+ {
2718
+ "epoch": 452.0,
2719
+ "learning_rate": 0.000384,
2720
+ "loss": 1.1274,
2721
+ "step": 5876
2722
+ },
2723
+ {
2724
+ "epoch": 453.0,
2725
+ "learning_rate": 0.00037600000000000003,
2726
+ "loss": 1.1067,
2727
+ "step": 5889
2728
+ },
2729
+ {
2730
+ "epoch": 454.0,
2731
+ "learning_rate": 0.000368,
2732
+ "loss": 1.0889,
2733
+ "step": 5902
2734
+ },
2735
+ {
2736
+ "epoch": 455.0,
2737
+ "learning_rate": 0.00035999999999999997,
2738
+ "loss": 1.0844,
2739
+ "step": 5915
2740
+ },
2741
+ {
2742
+ "epoch": 456.0,
2743
+ "learning_rate": 0.000352,
2744
+ "loss": 1.1341,
2745
+ "step": 5928
2746
+ },
2747
+ {
2748
+ "epoch": 457.0,
2749
+ "learning_rate": 0.00034399999999999996,
2750
+ "loss": 1.0644,
2751
+ "step": 5941
2752
+ },
2753
+ {
2754
+ "epoch": 458.0,
2755
+ "learning_rate": 0.00033600000000000004,
2756
+ "loss": 1.0991,
2757
+ "step": 5954
2758
+ },
2759
+ {
2760
+ "epoch": 459.0,
2761
+ "learning_rate": 0.000328,
2762
+ "loss": 1.1176,
2763
+ "step": 5967
2764
+ },
2765
+ {
2766
+ "epoch": 460.0,
2767
+ "learning_rate": 0.00032,
2768
+ "loss": 1.0997,
2769
+ "step": 5980
2770
+ },
2771
+ {
2772
+ "epoch": 461.0,
2773
+ "learning_rate": 0.000312,
2774
+ "loss": 1.0997,
2775
+ "step": 5993
2776
+ },
2777
+ {
2778
+ "epoch": 462.0,
2779
+ "learning_rate": 0.000304,
2780
+ "loss": 1.0763,
2781
+ "step": 6006
2782
+ },
2783
+ {
2784
+ "epoch": 463.0,
2785
+ "learning_rate": 0.000296,
2786
+ "loss": 1.1102,
2787
+ "step": 6019
2788
+ },
2789
+ {
2790
+ "epoch": 464.0,
2791
+ "learning_rate": 0.000288,
2792
+ "loss": 1.1236,
2793
+ "step": 6032
2794
+ },
2795
+ {
2796
+ "epoch": 465.0,
2797
+ "learning_rate": 0.00028000000000000003,
2798
+ "loss": 1.0941,
2799
+ "step": 6045
2800
+ },
2801
+ {
2802
+ "epoch": 466.0,
2803
+ "learning_rate": 0.00027200000000000005,
2804
+ "loss": 1.0976,
2805
+ "step": 6058
2806
+ },
2807
+ {
2808
+ "epoch": 467.0,
2809
+ "learning_rate": 0.000264,
2810
+ "loss": 1.0688,
2811
+ "step": 6071
2812
+ },
2813
+ {
2814
+ "epoch": 468.0,
2815
+ "learning_rate": 0.000256,
2816
+ "loss": 1.0591,
2817
+ "step": 6084
2818
+ },
2819
+ {
2820
+ "epoch": 469.0,
2821
+ "learning_rate": 0.000248,
2822
+ "loss": 1.0695,
2823
+ "step": 6097
2824
+ },
2825
+ {
2826
+ "epoch": 470.0,
2827
+ "learning_rate": 0.00024,
2828
+ "loss": 1.071,
2829
+ "step": 6110
2830
+ },
2831
+ {
2832
+ "epoch": 471.0,
2833
+ "learning_rate": 0.00023200000000000003,
2834
+ "loss": 1.0709,
2835
+ "step": 6123
2836
+ },
2837
+ {
2838
+ "epoch": 472.0,
2839
+ "learning_rate": 0.000224,
2840
+ "loss": 1.0767,
2841
+ "step": 6136
2842
+ },
2843
+ {
2844
+ "epoch": 473.0,
2845
+ "learning_rate": 0.000216,
2846
+ "loss": 1.0741,
2847
+ "step": 6149
2848
+ },
2849
+ {
2850
+ "epoch": 474.0,
2851
+ "learning_rate": 0.000208,
2852
+ "loss": 1.0644,
2853
+ "step": 6162
2854
+ },
2855
+ {
2856
+ "epoch": 475.0,
2857
+ "learning_rate": 0.0002,
2858
+ "loss": 1.0699,
2859
+ "step": 6175
2860
+ },
2861
+ {
2862
+ "epoch": 476.0,
2863
+ "learning_rate": 0.000192,
2864
+ "loss": 1.0727,
2865
+ "step": 6188
2866
+ },
2867
+ {
2868
+ "epoch": 477.0,
2869
+ "learning_rate": 0.000184,
2870
+ "loss": 1.06,
2871
+ "step": 6201
2872
+ },
2873
+ {
2874
+ "epoch": 478.0,
2875
+ "learning_rate": 0.000176,
2876
+ "loss": 1.0568,
2877
+ "step": 6214
2878
+ },
2879
+ {
2880
+ "epoch": 479.0,
2881
+ "learning_rate": 0.00016800000000000002,
2882
+ "loss": 1.0616,
2883
+ "step": 6227
2884
+ },
2885
+ {
2886
+ "epoch": 480.0,
2887
+ "learning_rate": 0.00016,
2888
+ "loss": 1.0491,
2889
+ "step": 6240
2890
+ },
2891
+ {
2892
+ "epoch": 481.0,
2893
+ "learning_rate": 0.000152,
2894
+ "loss": 1.0502,
2895
+ "step": 6253
2896
+ },
2897
+ {
2898
+ "epoch": 482.0,
2899
+ "learning_rate": 0.000144,
2900
+ "loss": 1.0742,
2901
+ "step": 6266
2902
+ },
2903
+ {
2904
+ "epoch": 483.0,
2905
+ "learning_rate": 0.00013600000000000003,
2906
+ "loss": 1.0582,
2907
+ "step": 6279
2908
+ },
2909
+ {
2910
+ "epoch": 484.0,
2911
+ "learning_rate": 0.000128,
2912
+ "loss": 1.0803,
2913
+ "step": 6292
2914
+ },
2915
+ {
2916
+ "epoch": 485.0,
2917
+ "learning_rate": 0.00012,
2918
+ "loss": 1.067,
2919
+ "step": 6305
2920
+ },
2921
+ {
2922
+ "epoch": 486.0,
2923
+ "learning_rate": 0.000112,
2924
+ "loss": 1.0397,
2925
+ "step": 6318
2926
+ },
2927
+ {
2928
+ "epoch": 487.0,
2929
+ "learning_rate": 0.000104,
2930
+ "loss": 1.0489,
2931
+ "step": 6331
2932
+ },
2933
+ {
2934
+ "epoch": 488.0,
2935
+ "learning_rate": 9.6e-05,
2936
+ "loss": 1.0378,
2937
+ "step": 6344
2938
+ },
2939
+ {
2940
+ "epoch": 489.0,
2941
+ "learning_rate": 8.8e-05,
2942
+ "loss": 1.0418,
2943
+ "step": 6357
2944
+ },
2945
+ {
2946
+ "epoch": 490.0,
2947
+ "learning_rate": 8e-05,
2948
+ "loss": 1.0344,
2949
+ "step": 6370
2950
+ },
2951
+ {
2952
+ "epoch": 491.0,
2953
+ "learning_rate": 7.2e-05,
2954
+ "loss": 1.0768,
2955
+ "step": 6383
2956
+ },
2957
+ {
2958
+ "epoch": 492.0,
2959
+ "learning_rate": 6.4e-05,
2960
+ "loss": 1.0296,
2961
+ "step": 6396
2962
+ },
2963
+ {
2964
+ "epoch": 493.0,
2965
+ "learning_rate": 5.6e-05,
2966
+ "loss": 1.0336,
2967
+ "step": 6409
2968
+ },
2969
+ {
2970
+ "epoch": 494.0,
2971
+ "learning_rate": 4.8e-05,
2972
+ "loss": 1.0568,
2973
+ "step": 6422
2974
+ },
2975
+ {
2976
+ "epoch": 495.0,
2977
+ "learning_rate": 4e-05,
2978
+ "loss": 1.0647,
2979
+ "step": 6435
2980
+ },
2981
+ {
2982
+ "epoch": 496.0,
2983
+ "learning_rate": 3.2e-05,
2984
+ "loss": 1.0448,
2985
+ "step": 6448
2986
+ },
2987
+ {
2988
+ "epoch": 497.0,
2989
+ "learning_rate": 2.4e-05,
2990
+ "loss": 1.0602,
2991
+ "step": 6461
2992
+ },
2993
+ {
2994
+ "epoch": 498.0,
2995
+ "learning_rate": 1.6e-05,
2996
+ "loss": 1.0615,
2997
+ "step": 6474
2998
+ },
2999
+ {
3000
+ "epoch": 499.0,
3001
+ "learning_rate": 8e-06,
3002
+ "loss": 1.0389,
3003
+ "step": 6487
3004
+ },
3005
+ {
3006
+ "epoch": 500.0,
3007
+ "learning_rate": 0.0,
3008
+ "loss": 1.0629,
3009
+ "step": 6500
3010
+ },
3011
+ {
3012
+ "epoch": 500.0,
3013
+ "step": 6500,
3014
+ "total_flos": 284798065115136.0,
3015
+ "train_loss": 2.748476623535156,
3016
+ "train_runtime": 71445.8185,
3017
+ "train_samples_per_second": 0.7,
3018
+ "train_steps_per_second": 0.091
3019
+ }
3020
+ ],
3021
+ "logging_steps": 500,
3022
+ "max_steps": 6500,
3023
+ "num_train_epochs": 500,
3024
+ "save_steps": 500,
3025
+ "total_flos": 284798065115136.0,
3026
+ "trial_name": null,
3027
+ "trial_params": null
3028
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3be7b15c1d6ac73d5f50e855bc0934537f4c2a3de37d4fc634a7f163bb7534f
3
+ size 4856