Hugging Face's logo

wza
/
vicuna-7b-lora-fin-1epoch

PEFT
Model card Files Community
wza commited on
Commit
bb47802
•
1 Parent(s): 7e7b820

upload

Browse files
Files changed (18) hide show
  1. README.md +3 -0
  2. adapter_config.json +20 -0
  3. adapter_model.bin +3 -0
  4. latest +1 -0
  5. rng_state_0.pth +3 -0
  6. rng_state_1.pth +3 -0
  7. rng_state_2.pth +3 -0
  8. rng_state_3.pth +3 -0
  9. rng_state_4.pth +3 -0
  10. rng_state_5.pth +3 -0
  11. rng_state_6.pth +3 -0
  12. rng_state_7.pth +3 -0
  13. special_tokens_map.json +24 -0
  14. tokenizer.model +3 -0
  15. tokenizer_config.json +34 -0
  16. trainer_state.json +4762 -0
  17. training_args.bin +3 -0
  18. zero_to_fp32.py +578 -0
README.md ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
adapter_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_model_name_or_path": "vicuna-7B-1.1-HF",
3
+ "bias": "none",
4
+ "fan_in_fan_out": false,
5
+ "inference_mode": true,
6
+ "init_lora_weights": true,
7
+ "layers_pattern": null,
8
+ "layers_to_transform": null,
9
+ "lora_alpha": 8,
10
+ "lora_dropout": 0.05,
11
+ "modules_to_save": null,
12
+ "peft_type": "LORA",
13
+ "r": 8,
14
+ "revision": null,
15
+ "target_modules": [
16
+ "q_proj",
17
+ "v_proj"
18
+ ],
19
+ "task_type": "CAUSAL_LM"
20
+ }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:068a110f89a5e8df9d74b3eb2ca96992930d47c72ae371830a0d92afca18acb3
3
+ size 8434381
latest ADDED
@@ -0,0 +1 @@
 
 
1
+ global_step791
rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:164027b90be87a2968bf9fbce240dc6d15ae030be915ecf048df683dd58ee197
3
+ size 21687
rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21793d347c81dad640f5a564f9df3b46e32e5ee41f3ceb5c91acca15a2b428cb
3
+ size 21687
rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9af1ed3ea469b95b2c417d7e3fbcd0df2636e73bfa24dc772ead349f0ffff6d
3
+ size 21687
rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7343d849b930cea26de1e8962b58d4c6997114c0971aba0b4193488cf82500eb
3
+ size 21687
rng_state_4.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2bcfe14b27b836c5cab9665c1cd9bef4165f95d20dbaa6fc8e34596036ac912a
3
+ size 21687
rng_state_5.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:959784e2399b31f445d25d26e7a34119625fc9bd45376a251fb383cae5e17a71
3
+ size 21687
rng_state_6.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39b652926310151e3251aefbca06bc5415bd557d916325bd72d521bf691f31de
3
+ size 21687
rng_state_7.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:687d501122c2283b410d6a97152909c6156778e78603eeba21eb4054f382a588
3
+ size 21687
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "bos_token": {
5
+ "__type": "AddedToken",
6
+ "content": "",
7
+ "lstrip": false,
8
+ "normalized": true,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "clean_up_tokenization_spaces": false,
13
+ "eos_token": {
14
+ "__type": "AddedToken",
15
+ "content": "",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "model_max_length": 2048,
22
+ "pad_token": null,
23
+ "padding_side": "right",
24
+ "sp_model_kwargs": {},
25
+ "tokenizer_class": "LlamaTokenizer",
26
+ "unk_token": {
27
+ "__type": "AddedToken",
28
+ "content": "",
29
+ "lstrip": false,
30
+ "normalized": true,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
trainer_state.json ADDED
@@ -0,0 +1,4762 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "global_step": 791,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 0,
13
+ "loss": 1.4791,
14
+ "step": 1
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 0,
19
+ "loss": 1.5012,
20
+ "step": 2
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 0,
25
+ "loss": 1.5116,
26
+ "step": 3
27
+ },
28
+ {
29
+ "epoch": 0.01,
30
+ "learning_rate": 0,
31
+ "loss": 1.4817,
32
+ "step": 4
33
+ },
34
+ {
35
+ "epoch": 0.01,
36
+ "learning_rate": 0,
37
+ "loss": 1.4658,
38
+ "step": 5
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 0,
43
+ "loss": 1.4871,
44
+ "step": 6
45
+ },
46
+ {
47
+ "epoch": 0.01,
48
+ "learning_rate": 0,
49
+ "loss": 1.4805,
50
+ "step": 7
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 0,
55
+ "loss": 1.538,
56
+ "step": 8
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 0,
61
+ "loss": 1.4946,
62
+ "step": 9
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 0,
67
+ "loss": 1.4747,
68
+ "step": 10
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 0,
73
+ "loss": 1.4675,
74
+ "step": 11
75
+ },
76
+ {
77
+ "epoch": 0.02,
78
+ "learning_rate": 0,
79
+ "loss": 1.5055,
80
+ "step": 12
81
+ },
82
+ {
83
+ "epoch": 0.02,
84
+ "learning_rate": 0,
85
+ "loss": 1.4786,
86
+ "step": 13
87
+ },
88
+ {
89
+ "epoch": 0.02,
90
+ "learning_rate": 0,
91
+ "loss": 1.5139,
92
+ "step": 14
93
+ },
94
+ {
95
+ "epoch": 0.02,
96
+ "learning_rate": 0,
97
+ "loss": 1.473,
98
+ "step": 15
99
+ },
100
+ {
101
+ "epoch": 0.02,
102
+ "learning_rate": 0,
103
+ "loss": 1.4784,
104
+ "step": 16
105
+ },
106
+ {
107
+ "epoch": 0.02,
108
+ "learning_rate": 0,
109
+ "loss": 1.4888,
110
+ "step": 17
111
+ },
112
+ {
113
+ "epoch": 0.02,
114
+ "learning_rate": 0.0,
115
+ "loss": 1.4879,
116
+ "step": 18
117
+ },
118
+ {
119
+ "epoch": 0.02,
120
+ "learning_rate": 3.581044635020827e-06,
121
+ "loss": 1.5187,
122
+ "step": 19
123
+ },
124
+ {
125
+ "epoch": 0.03,
126
+ "learning_rate": 5.675821459916692e-06,
127
+ "loss": 1.4978,
128
+ "step": 20
129
+ },
130
+ {
131
+ "epoch": 0.03,
132
+ "learning_rate": 7.162089270041654e-06,
133
+ "loss": 1.5043,
134
+ "step": 21
135
+ },
136
+ {
137
+ "epoch": 0.03,
138
+ "learning_rate": 8.31492814710052e-06,
139
+ "loss": 1.5099,
140
+ "step": 22
141
+ },
142
+ {
143
+ "epoch": 0.03,
144
+ "learning_rate": 9.256866094937518e-06,
145
+ "loss": 1.5065,
146
+ "step": 23
147
+ },
148
+ {
149
+ "epoch": 0.03,
150
+ "learning_rate": 1.0053263282233697e-05,
151
+ "loss": 1.4681,
152
+ "step": 24
153
+ },
154
+ {
155
+ "epoch": 0.03,
156
+ "learning_rate": 1.0743133905062482e-05,
157
+ "loss": 1.4753,
158
+ "step": 25
159
+ },
160
+ {
161
+ "epoch": 0.03,
162
+ "learning_rate": 1.1351642919833384e-05,
163
+ "loss": 1.5138,
164
+ "step": 26
165
+ },
166
+ {
167
+ "epoch": 0.03,
168
+ "learning_rate": 1.1895972782121347e-05,
169
+ "loss": 1.4987,
170
+ "step": 27
171
+ },
172
+ {
173
+ "epoch": 0.04,
174
+ "learning_rate": 1.2388379038142511e-05,
175
+ "loss": 1.4604,
176
+ "step": 28
177
+ },
178
+ {
179
+ "epoch": 0.04,
180
+ "learning_rate": 1.2837910729958346e-05,
181
+ "loss": 1.5164,
182
+ "step": 29
183
+ },
184
+ {
185
+ "epoch": 0.04,
186
+ "learning_rate": 1.325143979986714e-05,
187
+ "loss": 1.5106,
188
+ "step": 30
189
+ },
190
+ {
191
+ "epoch": 0.04,
192
+ "learning_rate": 1.3634307917254523e-05,
193
+ "loss": 1.4675,
194
+ "step": 31
195
+ },
196
+ {
197
+ "epoch": 0.04,
198
+ "learning_rate": 1.399074960701721e-05,
199
+ "loss": 1.4595,
200
+ "step": 32
201
+ },
202
+ {
203
+ "epoch": 0.04,
204
+ "learning_rate": 1.4324178540083309e-05,
205
+ "loss": 1.4644,
206
+ "step": 33
207
+ },
208
+ {
209
+ "epoch": 0.04,
210
+ "learning_rate": 1.4637386878506516e-05,
211
+ "loss": 1.5077,
212
+ "step": 34
213
+ },
214
+ {
215
+ "epoch": 0.04,
216
+ "learning_rate": 1.4932687554854209e-05,
217
+ "loss": 1.5314,
218
+ "step": 35
219
+ },
220
+ {
221
+ "epoch": 0.05,
222
+ "learning_rate": 1.5212018031974513e-05,
223
+ "loss": 1.4735,
224
+ "step": 36
225
+ },
226
+ {
227
+ "epoch": 0.05,
228
+ "learning_rate": 1.5477017417142176e-05,
229
+ "loss": 1.4847,
230
+ "step": 37
231
+ },
232
+ {
233
+ "epoch": 0.05,
234
+ "learning_rate": 1.5729084742150388e-05,
235
+ "loss": 1.4679,
236
+ "step": 38
237
+ },
238
+ {
239
+ "epoch": 0.05,
240
+ "learning_rate": 1.596942367316334e-05,
241
+ "loss": 1.5153,
242
+ "step": 39
243
+ },
244
+ {
245
+ "epoch": 0.05,
246
+ "learning_rate": 1.6199077273922286e-05,
247
+ "loss": 1.4827,
248
+ "step": 40
249
+ },
250
+ {
251
+ "epoch": 0.05,
252
+ "learning_rate": 1.6418955364979175e-05,
253
+ "loss": 1.4994,
254
+ "step": 41
255
+ },
256
+ {
257
+ "epoch": 0.05,
258
+ "learning_rate": 1.662985629420104e-05,
259
+ "loss": 1.5087,
260
+ "step": 42
261
+ },
262
+ {
263
+ "epoch": 0.05,
264
+ "learning_rate": 1.6832484434887967e-05,
265
+ "loss": 1.4974,
266
+ "step": 43
267
+ },
268
+ {
269
+ "epoch": 0.06,
270
+ "learning_rate": 1.7027464379750072e-05,
271
+ "loss": 1.4484,
272
+ "step": 44
273
+ },
274
+ {
275
+ "epoch": 0.06,
276
+ "learning_rate": 1.721535255227535e-05,
277
+ "loss": 1.4335,
278
+ "step": 45
279
+ },
280
+ {
281
+ "epoch": 0.06,
282
+ "learning_rate": 1.7396646779634733e-05,
283
+ "loss": 1.4854,
284
+ "step": 46
285
+ },
286
+ {
287
+ "epoch": 0.06,
288
+ "learning_rate": 1.757179424203804e-05,
289
+ "loss": 1.4985,
290
+ "step": 47
291
+ },
292
+ {
293
+ "epoch": 0.06,
294
+ "learning_rate": 1.7741198118150898e-05,
295
+ "loss": 1.4907,
296
+ "step": 48
297
+ },
298
+ {
299
+ "epoch": 0.06,
300
+ "learning_rate": 1.7905223175104135e-05,
301
+ "loss": 1.4369,
302
+ "step": 49
303
+ },
304
+ {
305
+ "epoch": 0.06,
306
+ "learning_rate": 1.80642004980592e-05,
307
+ "loss": 1.5002,
308
+ "step": 50
309
+ },
310
+ {
311
+ "epoch": 0.06,
312
+ "learning_rate": 1.8218431513527344e-05,
313
+ "loss": 1.4749,
314
+ "step": 51
315
+ },
316
+ {
317
+ "epoch": 0.07,
318
+ "learning_rate": 1.8368191429334215e-05,
319
+ "loss": 1.5021,
320
+ "step": 52
321
+ },
322
+ {
323
+ "epoch": 0.07,
324
+ "learning_rate": 1.8513732189875035e-05,
325
+ "loss": 1.4545,
326
+ "step": 53
327
+ },
328
+ {
329
+ "epoch": 0.07,
330
+ "learning_rate": 1.8655285026376745e-05,
331
+ "loss": 1.449,
332
+ "step": 54
333
+ },
334
+ {
335
+ "epoch": 0.07,
336
+ "learning_rate": 1.8793062666995343e-05,
337
+ "loss": 1.4276,
338
+ "step": 55
339
+ },
340
+ {
341
+ "epoch": 0.07,
342
+ "learning_rate": 1.892726125978383e-05,
343
+ "loss": 1.4625,
344
+ "step": 56
345
+ },
346
+ {
347
+ "epoch": 0.07,
348
+ "learning_rate": 1.9058062052163002e-05,
349
+ "loss": 1.4697,
350
+ "step": 57
351
+ },
352
+ {
353
+ "epoch": 0.07,
354
+ "learning_rate": 1.9185632862982668e-05,
355
+ "loss": 1.4584,
356
+ "step": 58
357
+ },
358
+ {
359
+ "epoch": 0.07,
360
+ "learning_rate": 1.9310129377171218e-05,
361
+ "loss": 1.4801,
362
+ "step": 59
363
+ },
364
+ {
365
+ "epoch": 0.08,
366
+ "learning_rate": 1.9431696288028554e-05,
367
+ "loss": 1.4368,
368
+ "step": 60
369
+ },
370
+ {
371
+ "epoch": 0.08,
372
+ "learning_rate": 1.9550468308184164e-05,
373
+ "loss": 1.4424,
374
+ "step": 61
375
+ },
376
+ {
377
+ "epoch": 0.08,
378
+ "learning_rate": 1.96665710669339e-05,
379
+ "loss": 1.418,
380
+ "step": 62
381
+ },
382
+ {
383
+ "epoch": 0.08,
384
+ "learning_rate": 1.9780121908943112e-05,
385
+ "loss": 1.4525,
386
+ "step": 63
387
+ },
388
+ {
389
+ "epoch": 0.08,
390
+ "learning_rate": 1.98912306070467e-05,
391
+ "loss": 1.4115,
392
+ "step": 64
393
+ },
394
+ {
395
+ "epoch": 0.08,
396
+ "learning_rate": 2e-05,
397
+ "loss": 1.4584,
398
+ "step": 65
399
+ },
400
+ {
401
+ "epoch": 0.08,
402
+ "learning_rate": 2e-05,
403
+ "loss": 1.3983,
404
+ "step": 66
405
+ },
406
+ {
407
+ "epoch": 0.08,
408
+ "learning_rate": 1.9986962190352023e-05,
409
+ "loss": 1.4412,
410
+ "step": 67
411
+ },
412
+ {
413
+ "epoch": 0.09,
414
+ "learning_rate": 1.9973924380704045e-05,
415
+ "loss": 1.4321,
416
+ "step": 68
417
+ },
418
+ {
419
+ "epoch": 0.09,
420
+ "learning_rate": 1.9960886571056063e-05,
421
+ "loss": 1.429,
422
+ "step": 69
423
+ },
424
+ {
425
+ "epoch": 0.09,
426
+ "learning_rate": 1.9947848761408084e-05,
427
+ "loss": 1.4509,
428
+ "step": 70
429
+ },
430
+ {
431
+ "epoch": 0.09,
432
+ "learning_rate": 1.9934810951760106e-05,
433
+ "loss": 1.4283,
434
+ "step": 71
435
+ },
436
+ {
437
+ "epoch": 0.09,
438
+ "learning_rate": 1.9921773142112127e-05,
439
+ "loss": 1.4553,
440
+ "step": 72
441
+ },
442
+ {
443
+ "epoch": 0.09,
444
+ "learning_rate": 1.990873533246415e-05,
445
+ "loss": 1.3925,
446
+ "step": 73
447
+ },
448
+ {
449
+ "epoch": 0.09,
450
+ "learning_rate": 1.9895697522816167e-05,
451
+ "loss": 1.4257,
452
+ "step": 74
453
+ },
454
+ {
455
+ "epoch": 0.09,
456
+ "learning_rate": 1.988265971316819e-05,
457
+ "loss": 1.4058,
458
+ "step": 75
459
+ },
460
+ {
461
+ "epoch": 0.1,
462
+ "learning_rate": 1.986962190352021e-05,
463
+ "loss": 1.3916,
464
+ "step": 76
465
+ },
466
+ {
467
+ "epoch": 0.1,
468
+ "learning_rate": 1.9856584093872232e-05,
469
+ "loss": 1.4246,
470
+ "step": 77
471
+ },
472
+ {
473
+ "epoch": 0.1,
474
+ "learning_rate": 1.984354628422425e-05,
475
+ "loss": 1.434,
476
+ "step": 78
477
+ },
478
+ {
479
+ "epoch": 0.1,
480
+ "learning_rate": 1.9830508474576275e-05,
481
+ "loss": 1.3815,
482
+ "step": 79
483
+ },
484
+ {
485
+ "epoch": 0.1,
486
+ "learning_rate": 1.9817470664928293e-05,
487
+ "loss": 1.386,
488
+ "step": 80
489
+ },
490
+ {
491
+ "epoch": 0.1,
492
+ "learning_rate": 1.9804432855280314e-05,
493
+ "loss": 1.3976,
494
+ "step": 81
495
+ },
496
+ {
497
+ "epoch": 0.1,
498
+ "learning_rate": 1.9791395045632336e-05,
499
+ "loss": 1.4288,
500
+ "step": 82
501
+ },
502
+ {
503
+ "epoch": 0.1,
504
+ "learning_rate": 1.9778357235984354e-05,
505
+ "loss": 1.4148,
506
+ "step": 83
507
+ },
508
+ {
509
+ "epoch": 0.11,
510
+ "learning_rate": 1.976531942633638e-05,
511
+ "loss": 1.3914,
512
+ "step": 84
513
+ },
514
+ {
515
+ "epoch": 0.11,
516
+ "learning_rate": 1.9752281616688397e-05,
517
+ "loss": 1.4066,
518
+ "step": 85
519
+ },
520
+ {
521
+ "epoch": 0.11,
522
+ "learning_rate": 1.973924380704042e-05,
523
+ "loss": 1.408,
524
+ "step": 86
525
+ },
526
+ {
527
+ "epoch": 0.11,
528
+ "learning_rate": 1.972620599739244e-05,
529
+ "loss": 1.395,
530
+ "step": 87
531
+ },
532
+ {
533
+ "epoch": 0.11,
534
+ "learning_rate": 1.9713168187744462e-05,
535
+ "loss": 1.421,
536
+ "step": 88
537
+ },
538
+ {
539
+ "epoch": 0.11,
540
+ "learning_rate": 1.970013037809648e-05,
541
+ "loss": 1.3871,
542
+ "step": 89
543
+ },
544
+ {
545
+ "epoch": 0.11,
546
+ "learning_rate": 1.96870925684485e-05,
547
+ "loss": 1.4153,
548
+ "step": 90
549
+ },
550
+ {
551
+ "epoch": 0.12,
552
+ "learning_rate": 1.9674054758800523e-05,
553
+ "loss": 1.3899,
554
+ "step": 91
555
+ },
556
+ {
557
+ "epoch": 0.12,
558
+ "learning_rate": 1.9661016949152545e-05,
559
+ "loss": 1.4215,
560
+ "step": 92
561
+ },
562
+ {
563
+ "epoch": 0.12,
564
+ "learning_rate": 1.9647979139504566e-05,
565
+ "loss": 1.3961,
566
+ "step": 93
567
+ },
568
+ {
569
+ "epoch": 0.12,
570
+ "learning_rate": 1.9634941329856584e-05,
571
+ "loss": 1.4343,
572
+ "step": 94
573
+ },
574
+ {
575
+ "epoch": 0.12,
576
+ "learning_rate": 1.9621903520208606e-05,
577
+ "loss": 1.4023,
578
+ "step": 95
579
+ },
580
+ {
581
+ "epoch": 0.12,
582
+ "learning_rate": 1.9608865710560627e-05,
583
+ "loss": 1.4044,
584
+ "step": 96
585
+ },
586
+ {
587
+ "epoch": 0.12,
588
+ "learning_rate": 1.959582790091265e-05,
589
+ "loss": 1.4089,
590
+ "step": 97
591
+ },
592
+ {
593
+ "epoch": 0.12,
594
+ "learning_rate": 1.958279009126467e-05,
595
+ "loss": 1.4286,
596
+ "step": 98
597
+ },
598
+ {
599
+ "epoch": 0.13,
600
+ "learning_rate": 1.956975228161669e-05,
601
+ "loss": 1.3584,
602
+ "step": 99
603
+ },
604
+ {
605
+ "epoch": 0.13,
606
+ "learning_rate": 1.9556714471968713e-05,
607
+ "loss": 1.4116,
608
+ "step": 100
609
+ },
610
+ {
611
+ "epoch": 0.13,
612
+ "learning_rate": 1.954367666232073e-05,
613
+ "loss": 1.3701,
614
+ "step": 101
615
+ },
616
+ {
617
+ "epoch": 0.13,
618
+ "learning_rate": 1.9530638852672753e-05,
619
+ "loss": 1.4187,
620
+ "step": 102
621
+ },
622
+ {
623
+ "epoch": 0.13,
624
+ "learning_rate": 1.9517601043024775e-05,
625
+ "loss": 1.3879,
626
+ "step": 103
627
+ },
628
+ {
629
+ "epoch": 0.13,
630
+ "learning_rate": 1.9504563233376793e-05,
631
+ "loss": 1.3561,
632
+ "step": 104
633
+ },
634
+ {
635
+ "epoch": 0.13,
636
+ "learning_rate": 1.9491525423728814e-05,
637
+ "loss": 1.3715,
638
+ "step": 105
639
+ },
640
+ {
641
+ "epoch": 0.13,
642
+ "learning_rate": 1.9478487614080836e-05,
643
+ "loss": 1.3772,
644
+ "step": 106
645
+ },
646
+ {
647
+ "epoch": 0.14,
648
+ "learning_rate": 1.9465449804432857e-05,
649
+ "loss": 1.3914,
650
+ "step": 107
651
+ },
652
+ {
653
+ "epoch": 0.14,
654
+ "learning_rate": 1.9452411994784876e-05,
655
+ "loss": 1.4004,
656
+ "step": 108
657
+ },
658
+ {
659
+ "epoch": 0.14,
660
+ "learning_rate": 1.94393741851369e-05,
661
+ "loss": 1.3425,
662
+ "step": 109
663
+ },
664
+ {
665
+ "epoch": 0.14,
666
+ "learning_rate": 1.942633637548892e-05,
667
+ "loss": 1.3588,
668
+ "step": 110
669
+ },
670
+ {
671
+ "epoch": 0.14,
672
+ "learning_rate": 1.941329856584094e-05,
673
+ "loss": 1.4044,
674
+ "step": 111
675
+ },
676
+ {
677
+ "epoch": 0.14,
678
+ "learning_rate": 1.940026075619296e-05,
679
+ "loss": 1.3721,
680
+ "step": 112
681
+ },
682
+ {
683
+ "epoch": 0.14,
684
+ "learning_rate": 1.938722294654498e-05,
685
+ "loss": 1.3787,
686
+ "step": 113
687
+ },
688
+ {
689
+ "epoch": 0.14,
690
+ "learning_rate": 1.9374185136897005e-05,
691
+ "loss": 1.3646,
692
+ "step": 114
693
+ },
694
+ {
695
+ "epoch": 0.15,
696
+ "learning_rate": 1.9361147327249023e-05,
697
+ "loss": 1.3621,
698
+ "step": 115
699
+ },
700
+ {
701
+ "epoch": 0.15,
702
+ "learning_rate": 1.9348109517601044e-05,
703
+ "loss": 1.3546,
704
+ "step": 116
705
+ },
706
+ {
707
+ "epoch": 0.15,
708
+ "learning_rate": 1.9335071707953066e-05,
709
+ "loss": 1.385,
710
+ "step": 117
711
+ },
712
+ {
713
+ "epoch": 0.15,
714
+ "learning_rate": 1.9322033898305087e-05,
715
+ "loss": 1.3973,
716
+ "step": 118
717
+ },
718
+ {
719
+ "epoch": 0.15,
720
+ "learning_rate": 1.930899608865711e-05,
721
+ "loss": 1.376,
722
+ "step": 119
723
+ },
724
+ {
725
+ "epoch": 0.15,
726
+ "learning_rate": 1.9295958279009127e-05,
727
+ "loss": 1.3938,
728
+ "step": 120
729
+ },
730
+ {
731
+ "epoch": 0.15,
732
+ "learning_rate": 1.928292046936115e-05,
733
+ "loss": 1.3755,
734
+ "step": 121
735
+ },
736
+ {
737
+ "epoch": 0.15,
738
+ "learning_rate": 1.926988265971317e-05,
739
+ "loss": 1.3964,
740
+ "step": 122
741
+ },
742
+ {
743
+ "epoch": 0.16,
744
+ "learning_rate": 1.9256844850065192e-05,
745
+ "loss": 1.3369,
746
+ "step": 123
747
+ },
748
+ {
749
+ "epoch": 0.16,
750
+ "learning_rate": 1.924380704041721e-05,
751
+ "loss": 1.3865,
752
+ "step": 124
753
+ },
754
+ {
755
+ "epoch": 0.16,
756
+ "learning_rate": 1.923076923076923e-05,
757
+ "loss": 1.3689,
758
+ "step": 125
759
+ },
760
+ {
761
+ "epoch": 0.16,
762
+ "learning_rate": 1.9217731421121253e-05,
763
+ "loss": 1.3519,
764
+ "step": 126
765
+ },
766
+ {
767
+ "epoch": 0.16,
768
+ "learning_rate": 1.9204693611473274e-05,
769
+ "loss": 1.3663,
770
+ "step": 127
771
+ },
772
+ {
773
+ "epoch": 0.16,
774
+ "learning_rate": 1.9191655801825296e-05,
775
+ "loss": 1.3595,
776
+ "step": 128
777
+ },
778
+ {
779
+ "epoch": 0.16,
780
+ "learning_rate": 1.9178617992177314e-05,
781
+ "loss": 1.3594,
782
+ "step": 129
783
+ },
784
+ {
785
+ "epoch": 0.16,
786
+ "learning_rate": 1.916558018252934e-05,
787
+ "loss": 1.3555,
788
+ "step": 130
789
+ },
790
+ {
791
+ "epoch": 0.17,
792
+ "learning_rate": 1.9152542372881357e-05,
793
+ "loss": 1.3864,
794
+ "step": 131
795
+ },
796
+ {
797
+ "epoch": 0.17,
798
+ "learning_rate": 1.913950456323338e-05,
799
+ "loss": 1.4041,
800
+ "step": 132
801
+ },
802
+ {
803
+ "epoch": 0.17,
804
+ "learning_rate": 1.91264667535854e-05,
805
+ "loss": 1.3717,
806
+ "step": 133
807
+ },
808
+ {
809
+ "epoch": 0.17,
810
+ "learning_rate": 1.911342894393742e-05,
811
+ "loss": 1.3451,
812
+ "step": 134
813
+ },
814
+ {
815
+ "epoch": 0.17,
816
+ "learning_rate": 1.910039113428944e-05,
817
+ "loss": 1.3484,
818
+ "step": 135
819
+ },
820
+ {
821
+ "epoch": 0.17,
822
+ "learning_rate": 1.908735332464146e-05,
823
+ "loss": 1.4012,
824
+ "step": 136
825
+ },
826
+ {
827
+ "epoch": 0.17,
828
+ "learning_rate": 1.9074315514993483e-05,
829
+ "loss": 1.3636,
830
+ "step": 137
831
+ },
832
+ {
833
+ "epoch": 0.17,
834
+ "learning_rate": 1.9061277705345505e-05,
835
+ "loss": 1.3663,
836
+ "step": 138
837
+ },
838
+ {
839
+ "epoch": 0.18,
840
+ "learning_rate": 1.9048239895697526e-05,
841
+ "loss": 1.3757,
842
+ "step": 139
843
+ },
844
+ {
845
+ "epoch": 0.18,
846
+ "learning_rate": 1.9035202086049544e-05,
847
+ "loss": 1.3837,
848
+ "step": 140
849
+ },
850
+ {
851
+ "epoch": 0.18,
852
+ "learning_rate": 1.9022164276401566e-05,
853
+ "loss": 1.3621,
854
+ "step": 141
855
+ },
856
+ {
857
+ "epoch": 0.18,
858
+ "learning_rate": 1.9009126466753587e-05,
859
+ "loss": 1.3636,
860
+ "step": 142
861
+ },
862
+ {
863
+ "epoch": 0.18,
864
+ "learning_rate": 1.8996088657105605e-05,
865
+ "loss": 1.3512,
866
+ "step": 143
867
+ },
868
+ {
869
+ "epoch": 0.18,
870
+ "learning_rate": 1.898305084745763e-05,
871
+ "loss": 1.3347,
872
+ "step": 144
873
+ },
874
+ {
875
+ "epoch": 0.18,
876
+ "learning_rate": 1.897001303780965e-05,
877
+ "loss": 1.3534,
878
+ "step": 145
879
+ },
880
+ {
881
+ "epoch": 0.18,
882
+ "learning_rate": 1.895697522816167e-05,
883
+ "loss": 1.3199,
884
+ "step": 146
885
+ },
886
+ {
887
+ "epoch": 0.19,
888
+ "learning_rate": 1.894393741851369e-05,
889
+ "loss": 1.3707,
890
+ "step": 147
891
+ },
892
+ {
893
+ "epoch": 0.19,
894
+ "learning_rate": 1.8930899608865713e-05,
895
+ "loss": 1.3481,
896
+ "step": 148
897
+ },
898
+ {
899
+ "epoch": 0.19,
900
+ "learning_rate": 1.8917861799217735e-05,
901
+ "loss": 1.3829,
902
+ "step": 149
903
+ },
904
+ {
905
+ "epoch": 0.19,
906
+ "learning_rate": 1.8904823989569753e-05,
907
+ "loss": 1.3748,
908
+ "step": 150
909
+ },
910
+ {
911
+ "epoch": 0.19,
912
+ "learning_rate": 1.8891786179921774e-05,
913
+ "loss": 1.3545,
914
+ "step": 151
915
+ },
916
+ {
917
+ "epoch": 0.19,
918
+ "learning_rate": 1.8878748370273796e-05,
919
+ "loss": 1.4016,
920
+ "step": 152
921
+ },
922
+ {
923
+ "epoch": 0.19,
924
+ "learning_rate": 1.8865710560625817e-05,
925
+ "loss": 1.3523,
926
+ "step": 153
927
+ },
928
+ {
929
+ "epoch": 0.19,
930
+ "learning_rate": 1.8852672750977836e-05,
931
+ "loss": 1.3497,
932
+ "step": 154
933
+ },
934
+ {
935
+ "epoch": 0.2,
936
+ "learning_rate": 1.8839634941329857e-05,
937
+ "loss": 1.344,
938
+ "step": 155
939
+ },
940
+ {
941
+ "epoch": 0.2,
942
+ "learning_rate": 1.882659713168188e-05,
943
+ "loss": 1.347,
944
+ "step": 156
945
+ },
946
+ {
947
+ "epoch": 0.2,
948
+ "learning_rate": 1.88135593220339e-05,
949
+ "loss": 1.306,
950
+ "step": 157
951
+ },
952
+ {
953
+ "epoch": 0.2,
954
+ "learning_rate": 1.880052151238592e-05,
955
+ "loss": 1.3613,
956
+ "step": 158
957
+ },
958
+ {
959
+ "epoch": 0.2,
960
+ "learning_rate": 1.878748370273794e-05,
961
+ "loss": 1.335,
962
+ "step": 159
963
+ },
964
+ {
965
+ "epoch": 0.2,
966
+ "learning_rate": 1.8774445893089965e-05,
967
+ "loss": 1.3098,
968
+ "step": 160
969
+ },
970
+ {
971
+ "epoch": 0.2,
972
+ "learning_rate": 1.8761408083441983e-05,
973
+ "loss": 1.3601,
974
+ "step": 161
975
+ },
976
+ {
977
+ "epoch": 0.2,
978
+ "learning_rate": 1.8748370273794004e-05,
979
+ "loss": 1.3298,
980
+ "step": 162
981
+ },
982
+ {
983
+ "epoch": 0.21,
984
+ "learning_rate": 1.8735332464146026e-05,
985
+ "loss": 1.382,
986
+ "step": 163
987
+ },
988
+ {
989
+ "epoch": 0.21,
990
+ "learning_rate": 1.8722294654498044e-05,
991
+ "loss": 1.3274,
992
+ "step": 164
993
+ },
994
+ {
995
+ "epoch": 0.21,
996
+ "learning_rate": 1.870925684485007e-05,
997
+ "loss": 1.3522,
998
+ "step": 165
999
+ },
1000
+ {
1001
+ "epoch": 0.21,
1002
+ "learning_rate": 1.8696219035202087e-05,
1003
+ "loss": 1.3281,
1004
+ "step": 166
1005
+ },
1006
+ {
1007
+ "epoch": 0.21,
1008
+ "learning_rate": 1.868318122555411e-05,
1009
+ "loss": 1.3514,
1010
+ "step": 167
1011
+ },
1012
+ {
1013
+ "epoch": 0.21,
1014
+ "learning_rate": 1.867014341590613e-05,
1015
+ "loss": 1.3655,
1016
+ "step": 168
1017
+ },
1018
+ {
1019
+ "epoch": 0.21,
1020
+ "learning_rate": 1.8657105606258152e-05,
1021
+ "loss": 1.3663,
1022
+ "step": 169
1023
+ },
1024
+ {
1025
+ "epoch": 0.21,
1026
+ "learning_rate": 1.864406779661017e-05,
1027
+ "loss": 1.3439,
1028
+ "step": 170
1029
+ },
1030
+ {
1031
+ "epoch": 0.22,
1032
+ "learning_rate": 1.863102998696219e-05,
1033
+ "loss": 1.3763,
1034
+ "step": 171
1035
+ },
1036
+ {
1037
+ "epoch": 0.22,
1038
+ "learning_rate": 1.8617992177314213e-05,
1039
+ "loss": 1.3954,
1040
+ "step": 172
1041
+ },
1042
+ {
1043
+ "epoch": 0.22,
1044
+ "learning_rate": 1.860495436766623e-05,
1045
+ "loss": 1.3251,
1046
+ "step": 173
1047
+ },
1048
+ {
1049
+ "epoch": 0.22,
1050
+ "learning_rate": 1.8591916558018256e-05,
1051
+ "loss": 1.3619,
1052
+ "step": 174
1053
+ },
1054
+ {
1055
+ "epoch": 0.22,
1056
+ "learning_rate": 1.8578878748370274e-05,
1057
+ "loss": 1.3516,
1058
+ "step": 175
1059
+ },
1060
+ {
1061
+ "epoch": 0.22,
1062
+ "learning_rate": 1.8565840938722296e-05,
1063
+ "loss": 1.3173,
1064
+ "step": 176
1065
+ },
1066
+ {
1067
+ "epoch": 0.22,
1068
+ "learning_rate": 1.8552803129074317e-05,
1069
+ "loss": 1.3649,
1070
+ "step": 177
1071
+ },
1072
+ {
1073
+ "epoch": 0.23,
1074
+ "learning_rate": 1.853976531942634e-05,
1075
+ "loss": 1.3279,
1076
+ "step": 178
1077
+ },
1078
+ {
1079
+ "epoch": 0.23,
1080
+ "learning_rate": 1.852672750977836e-05,
1081
+ "loss": 1.3372,
1082
+ "step": 179
1083
+ },
1084
+ {
1085
+ "epoch": 0.23,
1086
+ "learning_rate": 1.851368970013038e-05,
1087
+ "loss": 1.3315,
1088
+ "step": 180
1089
+ },
1090
+ {
1091
+ "epoch": 0.23,
1092
+ "learning_rate": 1.85006518904824e-05,
1093
+ "loss": 1.3341,
1094
+ "step": 181
1095
+ },
1096
+ {
1097
+ "epoch": 0.23,
1098
+ "learning_rate": 1.848761408083442e-05,
1099
+ "loss": 1.332,
1100
+ "step": 182
1101
+ },
1102
+ {
1103
+ "epoch": 0.23,
1104
+ "learning_rate": 1.8474576271186443e-05,
1105
+ "loss": 1.3092,
1106
+ "step": 183
1107
+ },
1108
+ {
1109
+ "epoch": 0.23,
1110
+ "learning_rate": 1.8461538461538465e-05,
1111
+ "loss": 1.3214,
1112
+ "step": 184
1113
+ },
1114
+ {
1115
+ "epoch": 0.23,
1116
+ "learning_rate": 1.8448500651890483e-05,
1117
+ "loss": 1.3785,
1118
+ "step": 185
1119
+ },
1120
+ {
1121
+ "epoch": 0.24,
1122
+ "learning_rate": 1.8435462842242504e-05,
1123
+ "loss": 1.3363,
1124
+ "step": 186
1125
+ },
1126
+ {
1127
+ "epoch": 0.24,
1128
+ "learning_rate": 1.8422425032594526e-05,
1129
+ "loss": 1.3572,
1130
+ "step": 187
1131
+ },
1132
+ {
1133
+ "epoch": 0.24,
1134
+ "learning_rate": 1.8409387222946547e-05,
1135
+ "loss": 1.3531,
1136
+ "step": 188
1137
+ },
1138
+ {
1139
+ "epoch": 0.24,
1140
+ "learning_rate": 1.8396349413298566e-05,
1141
+ "loss": 1.3387,
1142
+ "step": 189
1143
+ },
1144
+ {
1145
+ "epoch": 0.24,
1146
+ "learning_rate": 1.838331160365059e-05,
1147
+ "loss": 1.3566,
1148
+ "step": 190
1149
+ },
1150
+ {
1151
+ "epoch": 0.24,
1152
+ "learning_rate": 1.837027379400261e-05,
1153
+ "loss": 1.3018,
1154
+ "step": 191
1155
+ },
1156
+ {
1157
+ "epoch": 0.24,
1158
+ "learning_rate": 1.835723598435463e-05,
1159
+ "loss": 1.276,
1160
+ "step": 192
1161
+ },
1162
+ {
1163
+ "epoch": 0.24,
1164
+ "learning_rate": 1.834419817470665e-05,
1165
+ "loss": 1.3368,
1166
+ "step": 193
1167
+ },
1168
+ {
1169
+ "epoch": 0.25,
1170
+ "learning_rate": 1.833116036505867e-05,
1171
+ "loss": 1.3339,
1172
+ "step": 194
1173
+ },
1174
+ {
1175
+ "epoch": 0.25,
1176
+ "learning_rate": 1.8318122555410695e-05,
1177
+ "loss": 1.2972,
1178
+ "step": 195
1179
+ },
1180
+ {
1181
+ "epoch": 0.25,
1182
+ "learning_rate": 1.8305084745762713e-05,
1183
+ "loss": 1.3182,
1184
+ "step": 196
1185
+ },
1186
+ {
1187
+ "epoch": 0.25,
1188
+ "learning_rate": 1.8292046936114734e-05,
1189
+ "loss": 1.3167,
1190
+ "step": 197
1191
+ },
1192
+ {
1193
+ "epoch": 0.25,
1194
+ "learning_rate": 1.8279009126466756e-05,
1195
+ "loss": 1.3792,
1196
+ "step": 198
1197
+ },
1198
+ {
1199
+ "epoch": 0.25,
1200
+ "learning_rate": 1.8265971316818777e-05,
1201
+ "loss": 1.3479,
1202
+ "step": 199
1203
+ },
1204
+ {
1205
+ "epoch": 0.25,
1206
+ "learning_rate": 1.8252933507170796e-05,
1207
+ "loss": 1.3412,
1208
+ "step": 200
1209
+ },
1210
+ {
1211
+ "epoch": 0.25,
1212
+ "learning_rate": 1.8239895697522817e-05,
1213
+ "loss": 1.3534,
1214
+ "step": 201
1215
+ },
1216
+ {
1217
+ "epoch": 0.26,
1218
+ "learning_rate": 1.822685788787484e-05,
1219
+ "loss": 1.3856,
1220
+ "step": 202
1221
+ },
1222
+ {
1223
+ "epoch": 0.26,
1224
+ "learning_rate": 1.821382007822686e-05,
1225
+ "loss": 1.3234,
1226
+ "step": 203
1227
+ },
1228
+ {
1229
+ "epoch": 0.26,
1230
+ "learning_rate": 1.8200782268578882e-05,
1231
+ "loss": 1.3444,
1232
+ "step": 204
1233
+ },
1234
+ {
1235
+ "epoch": 0.26,
1236
+ "learning_rate": 1.81877444589309e-05,
1237
+ "loss": 1.3439,
1238
+ "step": 205
1239
+ },
1240
+ {
1241
+ "epoch": 0.26,
1242
+ "learning_rate": 1.817470664928292e-05,
1243
+ "loss": 1.3241,
1244
+ "step": 206
1245
+ },
1246
+ {
1247
+ "epoch": 0.26,
1248
+ "learning_rate": 1.8161668839634943e-05,
1249
+ "loss": 1.366,
1250
+ "step": 207
1251
+ },
1252
+ {
1253
+ "epoch": 0.26,
1254
+ "learning_rate": 1.8148631029986964e-05,
1255
+ "loss": 1.2926,
1256
+ "step": 208
1257
+ },
1258
+ {
1259
+ "epoch": 0.26,
1260
+ "learning_rate": 1.8135593220338986e-05,
1261
+ "loss": 1.3304,
1262
+ "step": 209
1263
+ },
1264
+ {
1265
+ "epoch": 0.27,
1266
+ "learning_rate": 1.8122555410691004e-05,
1267
+ "loss": 1.3566,
1268
+ "step": 210
1269
+ },
1270
+ {
1271
+ "epoch": 0.27,
1272
+ "learning_rate": 1.8109517601043026e-05,
1273
+ "loss": 1.2843,
1274
+ "step": 211
1275
+ },
1276
+ {
1277
+ "epoch": 0.27,
1278
+ "learning_rate": 1.8096479791395047e-05,
1279
+ "loss": 1.3363,
1280
+ "step": 212
1281
+ },
1282
+ {
1283
+ "epoch": 0.27,
1284
+ "learning_rate": 1.808344198174707e-05,
1285
+ "loss": 1.3334,
1286
+ "step": 213
1287
+ },
1288
+ {
1289
+ "epoch": 0.27,
1290
+ "learning_rate": 1.807040417209909e-05,
1291
+ "loss": 1.3555,
1292
+ "step": 214
1293
+ },
1294
+ {
1295
+ "epoch": 0.27,
1296
+ "learning_rate": 1.805736636245111e-05,
1297
+ "loss": 1.2939,
1298
+ "step": 215
1299
+ },
1300
+ {
1301
+ "epoch": 0.27,
1302
+ "learning_rate": 1.804432855280313e-05,
1303
+ "loss": 1.3147,
1304
+ "step": 216
1305
+ },
1306
+ {
1307
+ "epoch": 0.27,
1308
+ "learning_rate": 1.803129074315515e-05,
1309
+ "loss": 1.2948,
1310
+ "step": 217
1311
+ },
1312
+ {
1313
+ "epoch": 0.28,
1314
+ "learning_rate": 1.8018252933507173e-05,
1315
+ "loss": 1.3479,
1316
+ "step": 218
1317
+ },
1318
+ {
1319
+ "epoch": 0.28,
1320
+ "learning_rate": 1.800521512385919e-05,
1321
+ "loss": 1.3534,
1322
+ "step": 219
1323
+ },
1324
+ {
1325
+ "epoch": 0.28,
1326
+ "learning_rate": 1.7992177314211213e-05,
1327
+ "loss": 1.3074,
1328
+ "step": 220
1329
+ },
1330
+ {
1331
+ "epoch": 0.28,
1332
+ "learning_rate": 1.7979139504563234e-05,
1333
+ "loss": 1.3268,
1334
+ "step": 221
1335
+ },
1336
+ {
1337
+ "epoch": 0.28,
1338
+ "learning_rate": 1.7966101694915256e-05,
1339
+ "loss": 1.3584,
1340
+ "step": 222
1341
+ },
1342
+ {
1343
+ "epoch": 0.28,
1344
+ "learning_rate": 1.7953063885267277e-05,
1345
+ "loss": 1.33,
1346
+ "step": 223
1347
+ },
1348
+ {
1349
+ "epoch": 0.28,
1350
+ "learning_rate": 1.7940026075619295e-05,
1351
+ "loss": 1.3715,
1352
+ "step": 224
1353
+ },
1354
+ {
1355
+ "epoch": 0.28,
1356
+ "learning_rate": 1.792698826597132e-05,
1357
+ "loss": 1.3215,
1358
+ "step": 225
1359
+ },
1360
+ {
1361
+ "epoch": 0.29,
1362
+ "learning_rate": 1.791395045632334e-05,
1363
+ "loss": 1.3071,
1364
+ "step": 226
1365
+ },
1366
+ {
1367
+ "epoch": 0.29,
1368
+ "learning_rate": 1.790091264667536e-05,
1369
+ "loss": 1.3423,
1370
+ "step": 227
1371
+ },
1372
+ {
1373
+ "epoch": 0.29,
1374
+ "learning_rate": 1.788787483702738e-05,
1375
+ "loss": 1.3315,
1376
+ "step": 228
1377
+ },
1378
+ {
1379
+ "epoch": 0.29,
1380
+ "learning_rate": 1.78748370273794e-05,
1381
+ "loss": 1.3218,
1382
+ "step": 229
1383
+ },
1384
+ {
1385
+ "epoch": 0.29,
1386
+ "learning_rate": 1.7861799217731425e-05,
1387
+ "loss": 1.3346,
1388
+ "step": 230
1389
+ },
1390
+ {
1391
+ "epoch": 0.29,
1392
+ "learning_rate": 1.7848761408083443e-05,
1393
+ "loss": 1.2703,
1394
+ "step": 231
1395
+ },
1396
+ {
1397
+ "epoch": 0.29,
1398
+ "learning_rate": 1.7835723598435464e-05,
1399
+ "loss": 1.3326,
1400
+ "step": 232
1401
+ },
1402
+ {
1403
+ "epoch": 0.29,
1404
+ "learning_rate": 1.7822685788787486e-05,
1405
+ "loss": 1.3458,
1406
+ "step": 233
1407
+ },
1408
+ {
1409
+ "epoch": 0.3,
1410
+ "learning_rate": 1.7809647979139507e-05,
1411
+ "loss": 1.3073,
1412
+ "step": 234
1413
+ },
1414
+ {
1415
+ "epoch": 0.3,
1416
+ "learning_rate": 1.7796610169491526e-05,
1417
+ "loss": 1.3069,
1418
+ "step": 235
1419
+ },
1420
+ {
1421
+ "epoch": 0.3,
1422
+ "learning_rate": 1.7783572359843547e-05,
1423
+ "loss": 1.3293,
1424
+ "step": 236
1425
+ },
1426
+ {
1427
+ "epoch": 0.3,
1428
+ "learning_rate": 1.777053455019557e-05,
1429
+ "loss": 1.3202,
1430
+ "step": 237
1431
+ },
1432
+ {
1433
+ "epoch": 0.3,
1434
+ "learning_rate": 1.7757496740547587e-05,
1435
+ "loss": 1.3181,
1436
+ "step": 238
1437
+ },
1438
+ {
1439
+ "epoch": 0.3,
1440
+ "learning_rate": 1.774445893089961e-05,
1441
+ "loss": 1.3342,
1442
+ "step": 239
1443
+ },
1444
+ {
1445
+ "epoch": 0.3,
1446
+ "learning_rate": 1.773142112125163e-05,
1447
+ "loss": 1.2949,
1448
+ "step": 240
1449
+ },
1450
+ {
1451
+ "epoch": 0.3,
1452
+ "learning_rate": 1.771838331160365e-05,
1453
+ "loss": 1.3351,
1454
+ "step": 241
1455
+ },
1456
+ {
1457
+ "epoch": 0.31,
1458
+ "learning_rate": 1.7705345501955673e-05,
1459
+ "loss": 1.3513,
1460
+ "step": 242
1461
+ },
1462
+ {
1463
+ "epoch": 0.31,
1464
+ "learning_rate": 1.7692307692307694e-05,
1465
+ "loss": 1.2994,
1466
+ "step": 243
1467
+ },
1468
+ {
1469
+ "epoch": 0.31,
1470
+ "learning_rate": 1.7679269882659716e-05,
1471
+ "loss": 1.3175,
1472
+ "step": 244
1473
+ },
1474
+ {
1475
+ "epoch": 0.31,
1476
+ "learning_rate": 1.7666232073011734e-05,
1477
+ "loss": 1.2964,
1478
+ "step": 245
1479
+ },
1480
+ {
1481
+ "epoch": 0.31,
1482
+ "learning_rate": 1.7653194263363756e-05,
1483
+ "loss": 1.3344,
1484
+ "step": 246
1485
+ },
1486
+ {
1487
+ "epoch": 0.31,
1488
+ "learning_rate": 1.7640156453715777e-05,
1489
+ "loss": 1.318,
1490
+ "step": 247
1491
+ },
1492
+ {
1493
+ "epoch": 0.31,
1494
+ "learning_rate": 1.76271186440678e-05,
1495
+ "loss": 1.3169,
1496
+ "step": 248
1497
+ },
1498
+ {
1499
+ "epoch": 0.31,
1500
+ "learning_rate": 1.761408083441982e-05,
1501
+ "loss": 1.2988,
1502
+ "step": 249
1503
+ },
1504
+ {
1505
+ "epoch": 0.32,
1506
+ "learning_rate": 1.760104302477184e-05,
1507
+ "loss": 1.363,
1508
+ "step": 250
1509
+ },
1510
+ {
1511
+ "epoch": 0.32,
1512
+ "learning_rate": 1.758800521512386e-05,
1513
+ "loss": 1.3074,
1514
+ "step": 251
1515
+ },
1516
+ {
1517
+ "epoch": 0.32,
1518
+ "learning_rate": 1.757496740547588e-05,
1519
+ "loss": 1.3013,
1520
+ "step": 252
1521
+ },
1522
+ {
1523
+ "epoch": 0.32,
1524
+ "learning_rate": 1.7561929595827903e-05,
1525
+ "loss": 1.2819,
1526
+ "step": 253
1527
+ },
1528
+ {
1529
+ "epoch": 0.32,
1530
+ "learning_rate": 1.754889178617992e-05,
1531
+ "loss": 1.3164,
1532
+ "step": 254
1533
+ },
1534
+ {
1535
+ "epoch": 0.32,
1536
+ "learning_rate": 1.7535853976531946e-05,
1537
+ "loss": 1.3016,
1538
+ "step": 255
1539
+ },
1540
+ {
1541
+ "epoch": 0.32,
1542
+ "learning_rate": 1.7522816166883964e-05,
1543
+ "loss": 1.3185,
1544
+ "step": 256
1545
+ },
1546
+ {
1547
+ "epoch": 0.32,
1548
+ "learning_rate": 1.7509778357235986e-05,
1549
+ "loss": 1.3228,
1550
+ "step": 257
1551
+ },
1552
+ {
1553
+ "epoch": 0.33,
1554
+ "learning_rate": 1.7496740547588007e-05,
1555
+ "loss": 1.3274,
1556
+ "step": 258
1557
+ },
1558
+ {
1559
+ "epoch": 0.33,
1560
+ "learning_rate": 1.7483702737940025e-05,
1561
+ "loss": 1.2992,
1562
+ "step": 259
1563
+ },
1564
+ {
1565
+ "epoch": 0.33,
1566
+ "learning_rate": 1.747066492829205e-05,
1567
+ "loss": 1.3154,
1568
+ "step": 260
1569
+ },
1570
+ {
1571
+ "epoch": 0.33,
1572
+ "learning_rate": 1.745762711864407e-05,
1573
+ "loss": 1.2939,
1574
+ "step": 261
1575
+ },
1576
+ {
1577
+ "epoch": 0.33,
1578
+ "learning_rate": 1.744458930899609e-05,
1579
+ "loss": 1.3522,
1580
+ "step": 262
1581
+ },
1582
+ {
1583
+ "epoch": 0.33,
1584
+ "learning_rate": 1.743155149934811e-05,
1585
+ "loss": 1.3123,
1586
+ "step": 263
1587
+ },
1588
+ {
1589
+ "epoch": 0.33,
1590
+ "learning_rate": 1.7418513689700133e-05,
1591
+ "loss": 1.349,
1592
+ "step": 264
1593
+ },
1594
+ {
1595
+ "epoch": 0.34,
1596
+ "learning_rate": 1.740547588005215e-05,
1597
+ "loss": 1.3293,
1598
+ "step": 265
1599
+ },
1600
+ {
1601
+ "epoch": 0.34,
1602
+ "learning_rate": 1.7392438070404173e-05,
1603
+ "loss": 1.2897,
1604
+ "step": 266
1605
+ },
1606
+ {
1607
+ "epoch": 0.34,
1608
+ "learning_rate": 1.7379400260756194e-05,
1609
+ "loss": 1.3026,
1610
+ "step": 267
1611
+ },
1612
+ {
1613
+ "epoch": 0.34,
1614
+ "learning_rate": 1.7366362451108216e-05,
1615
+ "loss": 1.3231,
1616
+ "step": 268
1617
+ },
1618
+ {
1619
+ "epoch": 0.34,
1620
+ "learning_rate": 1.7353324641460237e-05,
1621
+ "loss": 1.3545,
1622
+ "step": 269
1623
+ },
1624
+ {
1625
+ "epoch": 0.34,
1626
+ "learning_rate": 1.7340286831812255e-05,
1627
+ "loss": 1.3335,
1628
+ "step": 270
1629
+ },
1630
+ {
1631
+ "epoch": 0.34,
1632
+ "learning_rate": 1.7327249022164277e-05,
1633
+ "loss": 1.2864,
1634
+ "step": 271
1635
+ },
1636
+ {
1637
+ "epoch": 0.34,
1638
+ "learning_rate": 1.73142112125163e-05,
1639
+ "loss": 1.3365,
1640
+ "step": 272
1641
+ },
1642
+ {
1643
+ "epoch": 0.35,
1644
+ "learning_rate": 1.730117340286832e-05,
1645
+ "loss": 1.3276,
1646
+ "step": 273
1647
+ },
1648
+ {
1649
+ "epoch": 0.35,
1650
+ "learning_rate": 1.728813559322034e-05,
1651
+ "loss": 1.2736,
1652
+ "step": 274
1653
+ },
1654
+ {
1655
+ "epoch": 0.35,
1656
+ "learning_rate": 1.727509778357236e-05,
1657
+ "loss": 1.3193,
1658
+ "step": 275
1659
+ },
1660
+ {
1661
+ "epoch": 0.35,
1662
+ "learning_rate": 1.726205997392438e-05,
1663
+ "loss": 1.3533,
1664
+ "step": 276
1665
+ },
1666
+ {
1667
+ "epoch": 0.35,
1668
+ "learning_rate": 1.7249022164276403e-05,
1669
+ "loss": 1.319,
1670
+ "step": 277
1671
+ },
1672
+ {
1673
+ "epoch": 0.35,
1674
+ "learning_rate": 1.7235984354628424e-05,
1675
+ "loss": 1.3008,
1676
+ "step": 278
1677
+ },
1678
+ {
1679
+ "epoch": 0.35,
1680
+ "learning_rate": 1.7222946544980446e-05,
1681
+ "loss": 1.2969,
1682
+ "step": 279
1683
+ },
1684
+ {
1685
+ "epoch": 0.35,
1686
+ "learning_rate": 1.7209908735332464e-05,
1687
+ "loss": 1.3213,
1688
+ "step": 280
1689
+ },
1690
+ {
1691
+ "epoch": 0.36,
1692
+ "learning_rate": 1.7196870925684486e-05,
1693
+ "loss": 1.2974,
1694
+ "step": 281
1695
+ },
1696
+ {
1697
+ "epoch": 0.36,
1698
+ "learning_rate": 1.7183833116036507e-05,
1699
+ "loss": 1.325,
1700
+ "step": 282
1701
+ },
1702
+ {
1703
+ "epoch": 0.36,
1704
+ "learning_rate": 1.717079530638853e-05,
1705
+ "loss": 1.3225,
1706
+ "step": 283
1707
+ },
1708
+ {
1709
+ "epoch": 0.36,
1710
+ "learning_rate": 1.7157757496740547e-05,
1711
+ "loss": 1.3142,
1712
+ "step": 284
1713
+ },
1714
+ {
1715
+ "epoch": 0.36,
1716
+ "learning_rate": 1.714471968709257e-05,
1717
+ "loss": 1.3547,
1718
+ "step": 285
1719
+ },
1720
+ {
1721
+ "epoch": 0.36,
1722
+ "learning_rate": 1.713168187744459e-05,
1723
+ "loss": 1.2861,
1724
+ "step": 286
1725
+ },
1726
+ {
1727
+ "epoch": 0.36,
1728
+ "learning_rate": 1.711864406779661e-05,
1729
+ "loss": 1.3414,
1730
+ "step": 287
1731
+ },
1732
+ {
1733
+ "epoch": 0.36,
1734
+ "learning_rate": 1.7105606258148633e-05,
1735
+ "loss": 1.2943,
1736
+ "step": 288
1737
+ },
1738
+ {
1739
+ "epoch": 0.37,
1740
+ "learning_rate": 1.709256844850065e-05,
1741
+ "loss": 1.3063,
1742
+ "step": 289
1743
+ },
1744
+ {
1745
+ "epoch": 0.37,
1746
+ "learning_rate": 1.7079530638852676e-05,
1747
+ "loss": 1.3047,
1748
+ "step": 290
1749
+ },
1750
+ {
1751
+ "epoch": 0.37,
1752
+ "learning_rate": 1.7066492829204694e-05,
1753
+ "loss": 1.2827,
1754
+ "step": 291
1755
+ },
1756
+ {
1757
+ "epoch": 0.37,
1758
+ "learning_rate": 1.7053455019556716e-05,
1759
+ "loss": 1.3425,
1760
+ "step": 292
1761
+ },
1762
+ {
1763
+ "epoch": 0.37,
1764
+ "learning_rate": 1.7040417209908737e-05,
1765
+ "loss": 1.3091,
1766
+ "step": 293
1767
+ },
1768
+ {
1769
+ "epoch": 0.37,
1770
+ "learning_rate": 1.702737940026076e-05,
1771
+ "loss": 1.3136,
1772
+ "step": 294
1773
+ },
1774
+ {
1775
+ "epoch": 0.37,
1776
+ "learning_rate": 1.7014341590612777e-05,
1777
+ "loss": 1.3516,
1778
+ "step": 295
1779
+ },
1780
+ {
1781
+ "epoch": 0.37,
1782
+ "learning_rate": 1.70013037809648e-05,
1783
+ "loss": 1.3174,
1784
+ "step": 296
1785
+ },
1786
+ {
1787
+ "epoch": 0.38,
1788
+ "learning_rate": 1.698826597131682e-05,
1789
+ "loss": 1.2878,
1790
+ "step": 297
1791
+ },
1792
+ {
1793
+ "epoch": 0.38,
1794
+ "learning_rate": 1.697522816166884e-05,
1795
+ "loss": 1.2913,
1796
+ "step": 298
1797
+ },
1798
+ {
1799
+ "epoch": 0.38,
1800
+ "learning_rate": 1.6962190352020863e-05,
1801
+ "loss": 1.3191,
1802
+ "step": 299
1803
+ },
1804
+ {
1805
+ "epoch": 0.38,
1806
+ "learning_rate": 1.694915254237288e-05,
1807
+ "loss": 1.3154,
1808
+ "step": 300
1809
+ },
1810
+ {
1811
+ "epoch": 0.38,
1812
+ "learning_rate": 1.6936114732724903e-05,
1813
+ "loss": 1.3379,
1814
+ "step": 301
1815
+ },
1816
+ {
1817
+ "epoch": 0.38,
1818
+ "learning_rate": 1.6923076923076924e-05,
1819
+ "loss": 1.3069,
1820
+ "step": 302
1821
+ },
1822
+ {
1823
+ "epoch": 0.38,
1824
+ "learning_rate": 1.6910039113428946e-05,
1825
+ "loss": 1.2919,
1826
+ "step": 303
1827
+ },
1828
+ {
1829
+ "epoch": 0.38,
1830
+ "learning_rate": 1.6897001303780967e-05,
1831
+ "loss": 1.3109,
1832
+ "step": 304
1833
+ },
1834
+ {
1835
+ "epoch": 0.39,
1836
+ "learning_rate": 1.6883963494132985e-05,
1837
+ "loss": 1.3032,
1838
+ "step": 305
1839
+ },
1840
+ {
1841
+ "epoch": 0.39,
1842
+ "learning_rate": 1.687092568448501e-05,
1843
+ "loss": 1.3502,
1844
+ "step": 306
1845
+ },
1846
+ {
1847
+ "epoch": 0.39,
1848
+ "learning_rate": 1.685788787483703e-05,
1849
+ "loss": 1.3052,
1850
+ "step": 307
1851
+ },
1852
+ {
1853
+ "epoch": 0.39,
1854
+ "learning_rate": 1.684485006518905e-05,
1855
+ "loss": 1.2963,
1856
+ "step": 308
1857
+ },
1858
+ {
1859
+ "epoch": 0.39,
1860
+ "learning_rate": 1.683181225554107e-05,
1861
+ "loss": 1.3191,
1862
+ "step": 309
1863
+ },
1864
+ {
1865
+ "epoch": 0.39,
1866
+ "learning_rate": 1.681877444589309e-05,
1867
+ "loss": 1.3306,
1868
+ "step": 310
1869
+ },
1870
+ {
1871
+ "epoch": 0.39,
1872
+ "learning_rate": 1.680573663624511e-05,
1873
+ "loss": 1.2985,
1874
+ "step": 311
1875
+ },
1876
+ {
1877
+ "epoch": 0.39,
1878
+ "learning_rate": 1.6792698826597133e-05,
1879
+ "loss": 1.2842,
1880
+ "step": 312
1881
+ },
1882
+ {
1883
+ "epoch": 0.4,
1884
+ "learning_rate": 1.6779661016949154e-05,
1885
+ "loss": 1.2462,
1886
+ "step": 313
1887
+ },
1888
+ {
1889
+ "epoch": 0.4,
1890
+ "learning_rate": 1.6766623207301176e-05,
1891
+ "loss": 1.3252,
1892
+ "step": 314
1893
+ },
1894
+ {
1895
+ "epoch": 0.4,
1896
+ "learning_rate": 1.6753585397653197e-05,
1897
+ "loss": 1.3273,
1898
+ "step": 315
1899
+ },
1900
+ {
1901
+ "epoch": 0.4,
1902
+ "learning_rate": 1.6740547588005215e-05,
1903
+ "loss": 1.3298,
1904
+ "step": 316
1905
+ },
1906
+ {
1907
+ "epoch": 0.4,
1908
+ "learning_rate": 1.6727509778357237e-05,
1909
+ "loss": 1.3062,
1910
+ "step": 317
1911
+ },
1912
+ {
1913
+ "epoch": 0.4,
1914
+ "learning_rate": 1.671447196870926e-05,
1915
+ "loss": 1.297,
1916
+ "step": 318
1917
+ },
1918
+ {
1919
+ "epoch": 0.4,
1920
+ "learning_rate": 1.6701434159061277e-05,
1921
+ "loss": 1.2899,
1922
+ "step": 319
1923
+ },
1924
+ {
1925
+ "epoch": 0.4,
1926
+ "learning_rate": 1.66883963494133e-05,
1927
+ "loss": 1.3318,
1928
+ "step": 320
1929
+ },
1930
+ {
1931
+ "epoch": 0.41,
1932
+ "learning_rate": 1.667535853976532e-05,
1933
+ "loss": 1.2843,
1934
+ "step": 321
1935
+ },
1936
+ {
1937
+ "epoch": 0.41,
1938
+ "learning_rate": 1.666232073011734e-05,
1939
+ "loss": 1.3131,
1940
+ "step": 322
1941
+ },
1942
+ {
1943
+ "epoch": 0.41,
1944
+ "learning_rate": 1.6649282920469363e-05,
1945
+ "loss": 1.3368,
1946
+ "step": 323
1947
+ },
1948
+ {
1949
+ "epoch": 0.41,
1950
+ "learning_rate": 1.6636245110821384e-05,
1951
+ "loss": 1.278,
1952
+ "step": 324
1953
+ },
1954
+ {
1955
+ "epoch": 0.41,
1956
+ "learning_rate": 1.6623207301173406e-05,
1957
+ "loss": 1.3033,
1958
+ "step": 325
1959
+ },
1960
+ {
1961
+ "epoch": 0.41,
1962
+ "learning_rate": 1.6610169491525424e-05,
1963
+ "loss": 1.3167,
1964
+ "step": 326
1965
+ },
1966
+ {
1967
+ "epoch": 0.41,
1968
+ "learning_rate": 1.6597131681877446e-05,
1969
+ "loss": 1.308,
1970
+ "step": 327
1971
+ },
1972
+ {
1973
+ "epoch": 0.41,
1974
+ "learning_rate": 1.6584093872229467e-05,
1975
+ "loss": 1.2753,
1976
+ "step": 328
1977
+ },
1978
+ {
1979
+ "epoch": 0.42,
1980
+ "learning_rate": 1.657105606258149e-05,
1981
+ "loss": 1.2986,
1982
+ "step": 329
1983
+ },
1984
+ {
1985
+ "epoch": 0.42,
1986
+ "learning_rate": 1.6558018252933507e-05,
1987
+ "loss": 1.3022,
1988
+ "step": 330
1989
+ },
1990
+ {
1991
+ "epoch": 0.42,
1992
+ "learning_rate": 1.654498044328553e-05,
1993
+ "loss": 1.2979,
1994
+ "step": 331
1995
+ },
1996
+ {
1997
+ "epoch": 0.42,
1998
+ "learning_rate": 1.653194263363755e-05,
1999
+ "loss": 1.3029,
2000
+ "step": 332
2001
+ },
2002
+ {
2003
+ "epoch": 0.42,
2004
+ "learning_rate": 1.651890482398957e-05,
2005
+ "loss": 1.2787,
2006
+ "step": 333
2007
+ },
2008
+ {
2009
+ "epoch": 0.42,
2010
+ "learning_rate": 1.6505867014341593e-05,
2011
+ "loss": 1.2913,
2012
+ "step": 334
2013
+ },
2014
+ {
2015
+ "epoch": 0.42,
2016
+ "learning_rate": 1.649282920469361e-05,
2017
+ "loss": 1.2938,
2018
+ "step": 335
2019
+ },
2020
+ {
2021
+ "epoch": 0.42,
2022
+ "learning_rate": 1.6479791395045636e-05,
2023
+ "loss": 1.2875,
2024
+ "step": 336
2025
+ },
2026
+ {
2027
+ "epoch": 0.43,
2028
+ "learning_rate": 1.6466753585397654e-05,
2029
+ "loss": 1.2859,
2030
+ "step": 337
2031
+ },
2032
+ {
2033
+ "epoch": 0.43,
2034
+ "learning_rate": 1.6453715775749676e-05,
2035
+ "loss": 1.3157,
2036
+ "step": 338
2037
+ },
2038
+ {
2039
+ "epoch": 0.43,
2040
+ "learning_rate": 1.6440677966101697e-05,
2041
+ "loss": 1.3082,
2042
+ "step": 339
2043
+ },
2044
+ {
2045
+ "epoch": 0.43,
2046
+ "learning_rate": 1.6427640156453715e-05,
2047
+ "loss": 1.2784,
2048
+ "step": 340
2049
+ },
2050
+ {
2051
+ "epoch": 0.43,
2052
+ "learning_rate": 1.6414602346805737e-05,
2053
+ "loss": 1.3204,
2054
+ "step": 341
2055
+ },
2056
+ {
2057
+ "epoch": 0.43,
2058
+ "learning_rate": 1.640156453715776e-05,
2059
+ "loss": 1.3344,
2060
+ "step": 342
2061
+ },
2062
+ {
2063
+ "epoch": 0.43,
2064
+ "learning_rate": 1.638852672750978e-05,
2065
+ "loss": 1.2981,
2066
+ "step": 343
2067
+ },
2068
+ {
2069
+ "epoch": 0.43,
2070
+ "learning_rate": 1.63754889178618e-05,
2071
+ "loss": 1.2943,
2072
+ "step": 344
2073
+ },
2074
+ {
2075
+ "epoch": 0.44,
2076
+ "learning_rate": 1.6362451108213823e-05,
2077
+ "loss": 1.2911,
2078
+ "step": 345
2079
+ },
2080
+ {
2081
+ "epoch": 0.44,
2082
+ "learning_rate": 1.634941329856584e-05,
2083
+ "loss": 1.3058,
2084
+ "step": 346
2085
+ },
2086
+ {
2087
+ "epoch": 0.44,
2088
+ "learning_rate": 1.6336375488917863e-05,
2089
+ "loss": 1.2999,
2090
+ "step": 347
2091
+ },
2092
+ {
2093
+ "epoch": 0.44,
2094
+ "learning_rate": 1.6323337679269884e-05,
2095
+ "loss": 1.3075,
2096
+ "step": 348
2097
+ },
2098
+ {
2099
+ "epoch": 0.44,
2100
+ "learning_rate": 1.6310299869621902e-05,
2101
+ "loss": 1.307,
2102
+ "step": 349
2103
+ },
2104
+ {
2105
+ "epoch": 0.44,
2106
+ "learning_rate": 1.6297262059973927e-05,
2107
+ "loss": 1.2651,
2108
+ "step": 350
2109
+ },
2110
+ {
2111
+ "epoch": 0.44,
2112
+ "learning_rate": 1.6284224250325945e-05,
2113
+ "loss": 1.307,
2114
+ "step": 351
2115
+ },
2116
+ {
2117
+ "epoch": 0.45,
2118
+ "learning_rate": 1.6271186440677967e-05,
2119
+ "loss": 1.2931,
2120
+ "step": 352
2121
+ },
2122
+ {
2123
+ "epoch": 0.45,
2124
+ "learning_rate": 1.625814863102999e-05,
2125
+ "loss": 1.3058,
2126
+ "step": 353
2127
+ },
2128
+ {
2129
+ "epoch": 0.45,
2130
+ "learning_rate": 1.624511082138201e-05,
2131
+ "loss": 1.3273,
2132
+ "step": 354
2133
+ },
2134
+ {
2135
+ "epoch": 0.45,
2136
+ "learning_rate": 1.623207301173403e-05,
2137
+ "loss": 1.2512,
2138
+ "step": 355
2139
+ },
2140
+ {
2141
+ "epoch": 0.45,
2142
+ "learning_rate": 1.621903520208605e-05,
2143
+ "loss": 1.2958,
2144
+ "step": 356
2145
+ },
2146
+ {
2147
+ "epoch": 0.45,
2148
+ "learning_rate": 1.620599739243807e-05,
2149
+ "loss": 1.291,
2150
+ "step": 357
2151
+ },
2152
+ {
2153
+ "epoch": 0.45,
2154
+ "learning_rate": 1.6192959582790093e-05,
2155
+ "loss": 1.3268,
2156
+ "step": 358
2157
+ },
2158
+ {
2159
+ "epoch": 0.45,
2160
+ "learning_rate": 1.6179921773142114e-05,
2161
+ "loss": 1.334,
2162
+ "step": 359
2163
+ },
2164
+ {
2165
+ "epoch": 0.46,
2166
+ "learning_rate": 1.6166883963494132e-05,
2167
+ "loss": 1.2966,
2168
+ "step": 360
2169
+ },
2170
+ {
2171
+ "epoch": 0.46,
2172
+ "learning_rate": 1.6153846153846154e-05,
2173
+ "loss": 1.3442,
2174
+ "step": 361
2175
+ },
2176
+ {
2177
+ "epoch": 0.46,
2178
+ "learning_rate": 1.6140808344198176e-05,
2179
+ "loss": 1.2639,
2180
+ "step": 362
2181
+ },
2182
+ {
2183
+ "epoch": 0.46,
2184
+ "learning_rate": 1.6127770534550197e-05,
2185
+ "loss": 1.2777,
2186
+ "step": 363
2187
+ },
2188
+ {
2189
+ "epoch": 0.46,
2190
+ "learning_rate": 1.611473272490222e-05,
2191
+ "loss": 1.288,
2192
+ "step": 364
2193
+ },
2194
+ {
2195
+ "epoch": 0.46,
2196
+ "learning_rate": 1.6101694915254237e-05,
2197
+ "loss": 1.3018,
2198
+ "step": 365
2199
+ },
2200
+ {
2201
+ "epoch": 0.46,
2202
+ "learning_rate": 1.608865710560626e-05,
2203
+ "loss": 1.2993,
2204
+ "step": 366
2205
+ },
2206
+ {
2207
+ "epoch": 0.46,
2208
+ "learning_rate": 1.607561929595828e-05,
2209
+ "loss": 1.2826,
2210
+ "step": 367
2211
+ },
2212
+ {
2213
+ "epoch": 0.47,
2214
+ "learning_rate": 1.60625814863103e-05,
2215
+ "loss": 1.3124,
2216
+ "step": 368
2217
+ },
2218
+ {
2219
+ "epoch": 0.47,
2220
+ "learning_rate": 1.6049543676662323e-05,
2221
+ "loss": 1.3529,
2222
+ "step": 369
2223
+ },
2224
+ {
2225
+ "epoch": 0.47,
2226
+ "learning_rate": 1.603650586701434e-05,
2227
+ "loss": 1.2919,
2228
+ "step": 370
2229
+ },
2230
+ {
2231
+ "epoch": 0.47,
2232
+ "learning_rate": 1.6023468057366366e-05,
2233
+ "loss": 1.3044,
2234
+ "step": 371
2235
+ },
2236
+ {
2237
+ "epoch": 0.47,
2238
+ "learning_rate": 1.6010430247718384e-05,
2239
+ "loss": 1.2424,
2240
+ "step": 372
2241
+ },
2242
+ {
2243
+ "epoch": 0.47,
2244
+ "learning_rate": 1.5997392438070406e-05,
2245
+ "loss": 1.2787,
2246
+ "step": 373
2247
+ },
2248
+ {
2249
+ "epoch": 0.47,
2250
+ "learning_rate": 1.5984354628422427e-05,
2251
+ "loss": 1.2701,
2252
+ "step": 374
2253
+ },
2254
+ {
2255
+ "epoch": 0.47,
2256
+ "learning_rate": 1.597131681877445e-05,
2257
+ "loss": 1.2924,
2258
+ "step": 375
2259
+ },
2260
+ {
2261
+ "epoch": 0.48,
2262
+ "learning_rate": 1.5958279009126467e-05,
2263
+ "loss": 1.2633,
2264
+ "step": 376
2265
+ },
2266
+ {
2267
+ "epoch": 0.48,
2268
+ "learning_rate": 1.594524119947849e-05,
2269
+ "loss": 1.2288,
2270
+ "step": 377
2271
+ },
2272
+ {
2273
+ "epoch": 0.48,
2274
+ "learning_rate": 1.593220338983051e-05,
2275
+ "loss": 1.2877,
2276
+ "step": 378
2277
+ },
2278
+ {
2279
+ "epoch": 0.48,
2280
+ "learning_rate": 1.5919165580182528e-05,
2281
+ "loss": 1.2767,
2282
+ "step": 379
2283
+ },
2284
+ {
2285
+ "epoch": 0.48,
2286
+ "learning_rate": 1.5906127770534553e-05,
2287
+ "loss": 1.2788,
2288
+ "step": 380
2289
+ },
2290
+ {
2291
+ "epoch": 0.48,
2292
+ "learning_rate": 1.589308996088657e-05,
2293
+ "loss": 1.2822,
2294
+ "step": 381
2295
+ },
2296
+ {
2297
+ "epoch": 0.48,
2298
+ "learning_rate": 1.5880052151238593e-05,
2299
+ "loss": 1.3312,
2300
+ "step": 382
2301
+ },
2302
+ {
2303
+ "epoch": 0.48,
2304
+ "learning_rate": 1.5867014341590614e-05,
2305
+ "loss": 1.2997,
2306
+ "step": 383
2307
+ },
2308
+ {
2309
+ "epoch": 0.49,
2310
+ "learning_rate": 1.5853976531942636e-05,
2311
+ "loss": 1.2441,
2312
+ "step": 384
2313
+ },
2314
+ {
2315
+ "epoch": 0.49,
2316
+ "learning_rate": 1.5840938722294657e-05,
2317
+ "loss": 1.3047,
2318
+ "step": 385
2319
+ },
2320
+ {
2321
+ "epoch": 0.49,
2322
+ "learning_rate": 1.5827900912646675e-05,
2323
+ "loss": 1.3101,
2324
+ "step": 386
2325
+ },
2326
+ {
2327
+ "epoch": 0.49,
2328
+ "learning_rate": 1.5814863102998697e-05,
2329
+ "loss": 1.2845,
2330
+ "step": 387
2331
+ },
2332
+ {
2333
+ "epoch": 0.49,
2334
+ "learning_rate": 1.580182529335072e-05,
2335
+ "loss": 1.2832,
2336
+ "step": 388
2337
+ },
2338
+ {
2339
+ "epoch": 0.49,
2340
+ "learning_rate": 1.578878748370274e-05,
2341
+ "loss": 1.3157,
2342
+ "step": 389
2343
+ },
2344
+ {
2345
+ "epoch": 0.49,
2346
+ "learning_rate": 1.577574967405476e-05,
2347
+ "loss": 1.248,
2348
+ "step": 390
2349
+ },
2350
+ {
2351
+ "epoch": 0.49,
2352
+ "learning_rate": 1.576271186440678e-05,
2353
+ "loss": 1.2858,
2354
+ "step": 391
2355
+ },
2356
+ {
2357
+ "epoch": 0.5,
2358
+ "learning_rate": 1.57496740547588e-05,
2359
+ "loss": 1.283,
2360
+ "step": 392
2361
+ },
2362
+ {
2363
+ "epoch": 0.5,
2364
+ "learning_rate": 1.5736636245110823e-05,
2365
+ "loss": 1.3123,
2366
+ "step": 393
2367
+ },
2368
+ {
2369
+ "epoch": 0.5,
2370
+ "learning_rate": 1.5723598435462844e-05,
2371
+ "loss": 1.2941,
2372
+ "step": 394
2373
+ },
2374
+ {
2375
+ "epoch": 0.5,
2376
+ "learning_rate": 1.5710560625814862e-05,
2377
+ "loss": 1.3051,
2378
+ "step": 395
2379
+ },
2380
+ {
2381
+ "epoch": 0.5,
2382
+ "learning_rate": 1.5697522816166887e-05,
2383
+ "loss": 1.3005,
2384
+ "step": 396
2385
+ },
2386
+ {
2387
+ "epoch": 0.5,
2388
+ "learning_rate": 1.5684485006518905e-05,
2389
+ "loss": 1.2645,
2390
+ "step": 397
2391
+ },
2392
+ {
2393
+ "epoch": 0.5,
2394
+ "learning_rate": 1.5671447196870927e-05,
2395
+ "loss": 1.2991,
2396
+ "step": 398
2397
+ },
2398
+ {
2399
+ "epoch": 0.5,
2400
+ "learning_rate": 1.565840938722295e-05,
2401
+ "loss": 1.2974,
2402
+ "step": 399
2403
+ },
2404
+ {
2405
+ "epoch": 0.51,
2406
+ "learning_rate": 1.5645371577574967e-05,
2407
+ "loss": 1.2877,
2408
+ "step": 400
2409
+ },
2410
+ {
2411
+ "epoch": 0.51,
2412
+ "learning_rate": 1.563233376792699e-05,
2413
+ "loss": 1.2832,
2414
+ "step": 401
2415
+ },
2416
+ {
2417
+ "epoch": 0.51,
2418
+ "learning_rate": 1.561929595827901e-05,
2419
+ "loss": 1.2827,
2420
+ "step": 402
2421
+ },
2422
+ {
2423
+ "epoch": 0.51,
2424
+ "learning_rate": 1.560625814863103e-05,
2425
+ "loss": 1.3064,
2426
+ "step": 403
2427
+ },
2428
+ {
2429
+ "epoch": 0.51,
2430
+ "learning_rate": 1.5593220338983053e-05,
2431
+ "loss": 1.3051,
2432
+ "step": 404
2433
+ },
2434
+ {
2435
+ "epoch": 0.51,
2436
+ "learning_rate": 1.5580182529335074e-05,
2437
+ "loss": 1.2562,
2438
+ "step": 405
2439
+ },
2440
+ {
2441
+ "epoch": 0.51,
2442
+ "learning_rate": 1.5567144719687092e-05,
2443
+ "loss": 1.3186,
2444
+ "step": 406
2445
+ },
2446
+ {
2447
+ "epoch": 0.51,
2448
+ "learning_rate": 1.5554106910039114e-05,
2449
+ "loss": 1.2731,
2450
+ "step": 407
2451
+ },
2452
+ {
2453
+ "epoch": 0.52,
2454
+ "learning_rate": 1.5541069100391136e-05,
2455
+ "loss": 1.2955,
2456
+ "step": 408
2457
+ },
2458
+ {
2459
+ "epoch": 0.52,
2460
+ "learning_rate": 1.5528031290743157e-05,
2461
+ "loss": 1.2938,
2462
+ "step": 409
2463
+ },
2464
+ {
2465
+ "epoch": 0.52,
2466
+ "learning_rate": 1.551499348109518e-05,
2467
+ "loss": 1.2505,
2468
+ "step": 410
2469
+ },
2470
+ {
2471
+ "epoch": 0.52,
2472
+ "learning_rate": 1.5501955671447197e-05,
2473
+ "loss": 1.3368,
2474
+ "step": 411
2475
+ },
2476
+ {
2477
+ "epoch": 0.52,
2478
+ "learning_rate": 1.5488917861799218e-05,
2479
+ "loss": 1.2964,
2480
+ "step": 412
2481
+ },
2482
+ {
2483
+ "epoch": 0.52,
2484
+ "learning_rate": 1.547588005215124e-05,
2485
+ "loss": 1.2986,
2486
+ "step": 413
2487
+ },
2488
+ {
2489
+ "epoch": 0.52,
2490
+ "learning_rate": 1.546284224250326e-05,
2491
+ "loss": 1.2694,
2492
+ "step": 414
2493
+ },
2494
+ {
2495
+ "epoch": 0.52,
2496
+ "learning_rate": 1.5449804432855283e-05,
2497
+ "loss": 1.3032,
2498
+ "step": 415
2499
+ },
2500
+ {
2501
+ "epoch": 0.53,
2502
+ "learning_rate": 1.54367666232073e-05,
2503
+ "loss": 1.3099,
2504
+ "step": 416
2505
+ },
2506
+ {
2507
+ "epoch": 0.53,
2508
+ "learning_rate": 1.5423728813559326e-05,
2509
+ "loss": 1.2749,
2510
+ "step": 417
2511
+ },
2512
+ {
2513
+ "epoch": 0.53,
2514
+ "learning_rate": 1.5410691003911344e-05,
2515
+ "loss": 1.3003,
2516
+ "step": 418
2517
+ },
2518
+ {
2519
+ "epoch": 0.53,
2520
+ "learning_rate": 1.5397653194263366e-05,
2521
+ "loss": 1.3204,
2522
+ "step": 419
2523
+ },
2524
+ {
2525
+ "epoch": 0.53,
2526
+ "learning_rate": 1.5384615384615387e-05,
2527
+ "loss": 1.3378,
2528
+ "step": 420
2529
+ },
2530
+ {
2531
+ "epoch": 0.53,
2532
+ "learning_rate": 1.5371577574967405e-05,
2533
+ "loss": 1.2841,
2534
+ "step": 421
2535
+ },
2536
+ {
2537
+ "epoch": 0.53,
2538
+ "learning_rate": 1.5358539765319427e-05,
2539
+ "loss": 1.2957,
2540
+ "step": 422
2541
+ },
2542
+ {
2543
+ "epoch": 0.53,
2544
+ "learning_rate": 1.534550195567145e-05,
2545
+ "loss": 1.2842,
2546
+ "step": 423
2547
+ },
2548
+ {
2549
+ "epoch": 0.54,
2550
+ "learning_rate": 1.533246414602347e-05,
2551
+ "loss": 1.2773,
2552
+ "step": 424
2553
+ },
2554
+ {
2555
+ "epoch": 0.54,
2556
+ "learning_rate": 1.5319426336375488e-05,
2557
+ "loss": 1.3038,
2558
+ "step": 425
2559
+ },
2560
+ {
2561
+ "epoch": 0.54,
2562
+ "learning_rate": 1.5306388526727513e-05,
2563
+ "loss": 1.2687,
2564
+ "step": 426
2565
+ },
2566
+ {
2567
+ "epoch": 0.54,
2568
+ "learning_rate": 1.529335071707953e-05,
2569
+ "loss": 1.3098,
2570
+ "step": 427
2571
+ },
2572
+ {
2573
+ "epoch": 0.54,
2574
+ "learning_rate": 1.5280312907431553e-05,
2575
+ "loss": 1.2704,
2576
+ "step": 428
2577
+ },
2578
+ {
2579
+ "epoch": 0.54,
2580
+ "learning_rate": 1.5267275097783574e-05,
2581
+ "loss": 1.2802,
2582
+ "step": 429
2583
+ },
2584
+ {
2585
+ "epoch": 0.54,
2586
+ "learning_rate": 1.5254237288135594e-05,
2587
+ "loss": 1.2667,
2588
+ "step": 430
2589
+ },
2590
+ {
2591
+ "epoch": 0.54,
2592
+ "learning_rate": 1.5241199478487616e-05,
2593
+ "loss": 1.27,
2594
+ "step": 431
2595
+ },
2596
+ {
2597
+ "epoch": 0.55,
2598
+ "learning_rate": 1.5228161668839635e-05,
2599
+ "loss": 1.2521,
2600
+ "step": 432
2601
+ },
2602
+ {
2603
+ "epoch": 0.55,
2604
+ "learning_rate": 1.5215123859191657e-05,
2605
+ "loss": 1.2719,
2606
+ "step": 433
2607
+ },
2608
+ {
2609
+ "epoch": 0.55,
2610
+ "learning_rate": 1.5202086049543678e-05,
2611
+ "loss": 1.3188,
2612
+ "step": 434
2613
+ },
2614
+ {
2615
+ "epoch": 0.55,
2616
+ "learning_rate": 1.5189048239895698e-05,
2617
+ "loss": 1.2958,
2618
+ "step": 435
2619
+ },
2620
+ {
2621
+ "epoch": 0.55,
2622
+ "learning_rate": 1.517601043024772e-05,
2623
+ "loss": 1.271,
2624
+ "step": 436
2625
+ },
2626
+ {
2627
+ "epoch": 0.55,
2628
+ "learning_rate": 1.5162972620599741e-05,
2629
+ "loss": 1.2307,
2630
+ "step": 437
2631
+ },
2632
+ {
2633
+ "epoch": 0.55,
2634
+ "learning_rate": 1.5149934810951761e-05,
2635
+ "loss": 1.2925,
2636
+ "step": 438
2637
+ },
2638
+ {
2639
+ "epoch": 0.55,
2640
+ "learning_rate": 1.5136897001303783e-05,
2641
+ "loss": 1.2579,
2642
+ "step": 439
2643
+ },
2644
+ {
2645
+ "epoch": 0.56,
2646
+ "learning_rate": 1.5123859191655803e-05,
2647
+ "loss": 1.3206,
2648
+ "step": 440
2649
+ },
2650
+ {
2651
+ "epoch": 0.56,
2652
+ "learning_rate": 1.5110821382007822e-05,
2653
+ "loss": 1.3112,
2654
+ "step": 441
2655
+ },
2656
+ {
2657
+ "epoch": 0.56,
2658
+ "learning_rate": 1.5097783572359846e-05,
2659
+ "loss": 1.2642,
2660
+ "step": 442
2661
+ },
2662
+ {
2663
+ "epoch": 0.56,
2664
+ "learning_rate": 1.5084745762711865e-05,
2665
+ "loss": 1.3029,
2666
+ "step": 443
2667
+ },
2668
+ {
2669
+ "epoch": 0.56,
2670
+ "learning_rate": 1.5071707953063885e-05,
2671
+ "loss": 1.2833,
2672
+ "step": 444
2673
+ },
2674
+ {
2675
+ "epoch": 0.56,
2676
+ "learning_rate": 1.5058670143415907e-05,
2677
+ "loss": 1.2645,
2678
+ "step": 445
2679
+ },
2680
+ {
2681
+ "epoch": 0.56,
2682
+ "learning_rate": 1.5045632333767928e-05,
2683
+ "loss": 1.293,
2684
+ "step": 446
2685
+ },
2686
+ {
2687
+ "epoch": 0.57,
2688
+ "learning_rate": 1.503259452411995e-05,
2689
+ "loss": 1.3011,
2690
+ "step": 447
2691
+ },
2692
+ {
2693
+ "epoch": 0.57,
2694
+ "learning_rate": 1.501955671447197e-05,
2695
+ "loss": 1.2661,
2696
+ "step": 448
2697
+ },
2698
+ {
2699
+ "epoch": 0.57,
2700
+ "learning_rate": 1.500651890482399e-05,
2701
+ "loss": 1.2931,
2702
+ "step": 449
2703
+ },
2704
+ {
2705
+ "epoch": 0.57,
2706
+ "learning_rate": 1.4993481095176013e-05,
2707
+ "loss": 1.2683,
2708
+ "step": 450
2709
+ },
2710
+ {
2711
+ "epoch": 0.57,
2712
+ "learning_rate": 1.4980443285528033e-05,
2713
+ "loss": 1.2996,
2714
+ "step": 451
2715
+ },
2716
+ {
2717
+ "epoch": 0.57,
2718
+ "learning_rate": 1.4967405475880053e-05,
2719
+ "loss": 1.3036,
2720
+ "step": 452
2721
+ },
2722
+ {
2723
+ "epoch": 0.57,
2724
+ "learning_rate": 1.4954367666232074e-05,
2725
+ "loss": 1.2673,
2726
+ "step": 453
2727
+ },
2728
+ {
2729
+ "epoch": 0.57,
2730
+ "learning_rate": 1.4941329856584096e-05,
2731
+ "loss": 1.2502,
2732
+ "step": 454
2733
+ },
2734
+ {
2735
+ "epoch": 0.58,
2736
+ "learning_rate": 1.4928292046936117e-05,
2737
+ "loss": 1.2681,
2738
+ "step": 455
2739
+ },
2740
+ {
2741
+ "epoch": 0.58,
2742
+ "learning_rate": 1.4915254237288137e-05,
2743
+ "loss": 1.3107,
2744
+ "step": 456
2745
+ },
2746
+ {
2747
+ "epoch": 0.58,
2748
+ "learning_rate": 1.4902216427640157e-05,
2749
+ "loss": 1.2971,
2750
+ "step": 457
2751
+ },
2752
+ {
2753
+ "epoch": 0.58,
2754
+ "learning_rate": 1.488917861799218e-05,
2755
+ "loss": 1.3087,
2756
+ "step": 458
2757
+ },
2758
+ {
2759
+ "epoch": 0.58,
2760
+ "learning_rate": 1.48761408083442e-05,
2761
+ "loss": 1.281,
2762
+ "step": 459
2763
+ },
2764
+ {
2765
+ "epoch": 0.58,
2766
+ "learning_rate": 1.486310299869622e-05,
2767
+ "loss": 1.2891,
2768
+ "step": 460
2769
+ },
2770
+ {
2771
+ "epoch": 0.58,
2772
+ "learning_rate": 1.4850065189048241e-05,
2773
+ "loss": 1.2732,
2774
+ "step": 461
2775
+ },
2776
+ {
2777
+ "epoch": 0.58,
2778
+ "learning_rate": 1.4837027379400261e-05,
2779
+ "loss": 1.3044,
2780
+ "step": 462
2781
+ },
2782
+ {
2783
+ "epoch": 0.59,
2784
+ "learning_rate": 1.4823989569752283e-05,
2785
+ "loss": 1.2866,
2786
+ "step": 463
2787
+ },
2788
+ {
2789
+ "epoch": 0.59,
2790
+ "learning_rate": 1.4810951760104304e-05,
2791
+ "loss": 1.3029,
2792
+ "step": 464
2793
+ },
2794
+ {
2795
+ "epoch": 0.59,
2796
+ "learning_rate": 1.4797913950456324e-05,
2797
+ "loss": 1.2736,
2798
+ "step": 465
2799
+ },
2800
+ {
2801
+ "epoch": 0.59,
2802
+ "learning_rate": 1.4784876140808346e-05,
2803
+ "loss": 1.2585,
2804
+ "step": 466
2805
+ },
2806
+ {
2807
+ "epoch": 0.59,
2808
+ "learning_rate": 1.4771838331160367e-05,
2809
+ "loss": 1.2804,
2810
+ "step": 467
2811
+ },
2812
+ {
2813
+ "epoch": 0.59,
2814
+ "learning_rate": 1.4758800521512387e-05,
2815
+ "loss": 1.2877,
2816
+ "step": 468
2817
+ },
2818
+ {
2819
+ "epoch": 0.59,
2820
+ "learning_rate": 1.4745762711864408e-05,
2821
+ "loss": 1.2478,
2822
+ "step": 469
2823
+ },
2824
+ {
2825
+ "epoch": 0.59,
2826
+ "learning_rate": 1.4732724902216428e-05,
2827
+ "loss": 1.2736,
2828
+ "step": 470
2829
+ },
2830
+ {
2831
+ "epoch": 0.6,
2832
+ "learning_rate": 1.4719687092568448e-05,
2833
+ "loss": 1.2902,
2834
+ "step": 471
2835
+ },
2836
+ {
2837
+ "epoch": 0.6,
2838
+ "learning_rate": 1.4706649282920471e-05,
2839
+ "loss": 1.2788,
2840
+ "step": 472
2841
+ },
2842
+ {
2843
+ "epoch": 0.6,
2844
+ "learning_rate": 1.4693611473272491e-05,
2845
+ "loss": 1.2657,
2846
+ "step": 473
2847
+ },
2848
+ {
2849
+ "epoch": 0.6,
2850
+ "learning_rate": 1.4680573663624513e-05,
2851
+ "loss": 1.2828,
2852
+ "step": 474
2853
+ },
2854
+ {
2855
+ "epoch": 0.6,
2856
+ "learning_rate": 1.4667535853976533e-05,
2857
+ "loss": 1.2885,
2858
+ "step": 475
2859
+ },
2860
+ {
2861
+ "epoch": 0.6,
2862
+ "learning_rate": 1.4654498044328554e-05,
2863
+ "loss": 1.2744,
2864
+ "step": 476
2865
+ },
2866
+ {
2867
+ "epoch": 0.6,
2868
+ "learning_rate": 1.4641460234680576e-05,
2869
+ "loss": 1.2737,
2870
+ "step": 477
2871
+ },
2872
+ {
2873
+ "epoch": 0.6,
2874
+ "learning_rate": 1.4628422425032595e-05,
2875
+ "loss": 1.2311,
2876
+ "step": 478
2877
+ },
2878
+ {
2879
+ "epoch": 0.61,
2880
+ "learning_rate": 1.4615384615384615e-05,
2881
+ "loss": 1.2834,
2882
+ "step": 479
2883
+ },
2884
+ {
2885
+ "epoch": 0.61,
2886
+ "learning_rate": 1.4602346805736639e-05,
2887
+ "loss": 1.2615,
2888
+ "step": 480
2889
+ },
2890
+ {
2891
+ "epoch": 0.61,
2892
+ "learning_rate": 1.4589308996088658e-05,
2893
+ "loss": 1.2748,
2894
+ "step": 481
2895
+ },
2896
+ {
2897
+ "epoch": 0.61,
2898
+ "learning_rate": 1.4576271186440678e-05,
2899
+ "loss": 1.3031,
2900
+ "step": 482
2901
+ },
2902
+ {
2903
+ "epoch": 0.61,
2904
+ "learning_rate": 1.45632333767927e-05,
2905
+ "loss": 1.3126,
2906
+ "step": 483
2907
+ },
2908
+ {
2909
+ "epoch": 0.61,
2910
+ "learning_rate": 1.455019556714472e-05,
2911
+ "loss": 1.249,
2912
+ "step": 484
2913
+ },
2914
+ {
2915
+ "epoch": 0.61,
2916
+ "learning_rate": 1.4537157757496743e-05,
2917
+ "loss": 1.2816,
2918
+ "step": 485
2919
+ },
2920
+ {
2921
+ "epoch": 0.61,
2922
+ "learning_rate": 1.4524119947848763e-05,
2923
+ "loss": 1.3064,
2924
+ "step": 486
2925
+ },
2926
+ {
2927
+ "epoch": 0.62,
2928
+ "learning_rate": 1.4511082138200782e-05,
2929
+ "loss": 1.299,
2930
+ "step": 487
2931
+ },
2932
+ {
2933
+ "epoch": 0.62,
2934
+ "learning_rate": 1.4498044328552806e-05,
2935
+ "loss": 1.2998,
2936
+ "step": 488
2937
+ },
2938
+ {
2939
+ "epoch": 0.62,
2940
+ "learning_rate": 1.4485006518904826e-05,
2941
+ "loss": 1.2906,
2942
+ "step": 489
2943
+ },
2944
+ {
2945
+ "epoch": 0.62,
2946
+ "learning_rate": 1.4471968709256845e-05,
2947
+ "loss": 1.2656,
2948
+ "step": 490
2949
+ },
2950
+ {
2951
+ "epoch": 0.62,
2952
+ "learning_rate": 1.4458930899608867e-05,
2953
+ "loss": 1.2729,
2954
+ "step": 491
2955
+ },
2956
+ {
2957
+ "epoch": 0.62,
2958
+ "learning_rate": 1.4445893089960887e-05,
2959
+ "loss": 1.2892,
2960
+ "step": 492
2961
+ },
2962
+ {
2963
+ "epoch": 0.62,
2964
+ "learning_rate": 1.443285528031291e-05,
2965
+ "loss": 1.2692,
2966
+ "step": 493
2967
+ },
2968
+ {
2969
+ "epoch": 0.62,
2970
+ "learning_rate": 1.441981747066493e-05,
2971
+ "loss": 1.288,
2972
+ "step": 494
2973
+ },
2974
+ {
2975
+ "epoch": 0.63,
2976
+ "learning_rate": 1.440677966101695e-05,
2977
+ "loss": 1.2915,
2978
+ "step": 495
2979
+ },
2980
+ {
2981
+ "epoch": 0.63,
2982
+ "learning_rate": 1.4393741851368971e-05,
2983
+ "loss": 1.2751,
2984
+ "step": 496
2985
+ },
2986
+ {
2987
+ "epoch": 0.63,
2988
+ "learning_rate": 1.4380704041720993e-05,
2989
+ "loss": 1.285,
2990
+ "step": 497
2991
+ },
2992
+ {
2993
+ "epoch": 0.63,
2994
+ "learning_rate": 1.4367666232073013e-05,
2995
+ "loss": 1.2832,
2996
+ "step": 498
2997
+ },
2998
+ {
2999
+ "epoch": 0.63,
3000
+ "learning_rate": 1.4354628422425034e-05,
3001
+ "loss": 1.2645,
3002
+ "step": 499
3003
+ },
3004
+ {
3005
+ "epoch": 0.63,
3006
+ "learning_rate": 1.4341590612777054e-05,
3007
+ "loss": 1.2577,
3008
+ "step": 500
3009
+ },
3010
+ {
3011
+ "epoch": 0.63,
3012
+ "learning_rate": 1.4328552803129077e-05,
3013
+ "loss": 1.252,
3014
+ "step": 501
3015
+ },
3016
+ {
3017
+ "epoch": 0.63,
3018
+ "learning_rate": 1.4315514993481097e-05,
3019
+ "loss": 1.236,
3020
+ "step": 502
3021
+ },
3022
+ {
3023
+ "epoch": 0.64,
3024
+ "learning_rate": 1.4302477183833117e-05,
3025
+ "loss": 1.2863,
3026
+ "step": 503
3027
+ },
3028
+ {
3029
+ "epoch": 0.64,
3030
+ "learning_rate": 1.4289439374185138e-05,
3031
+ "loss": 1.2549,
3032
+ "step": 504
3033
+ },
3034
+ {
3035
+ "epoch": 0.64,
3036
+ "learning_rate": 1.4276401564537158e-05,
3037
+ "loss": 1.3345,
3038
+ "step": 505
3039
+ },
3040
+ {
3041
+ "epoch": 0.64,
3042
+ "learning_rate": 1.426336375488918e-05,
3043
+ "loss": 1.269,
3044
+ "step": 506
3045
+ },
3046
+ {
3047
+ "epoch": 0.64,
3048
+ "learning_rate": 1.4250325945241201e-05,
3049
+ "loss": 1.2834,
3050
+ "step": 507
3051
+ },
3052
+ {
3053
+ "epoch": 0.64,
3054
+ "learning_rate": 1.4237288135593221e-05,
3055
+ "loss": 1.2853,
3056
+ "step": 508
3057
+ },
3058
+ {
3059
+ "epoch": 0.64,
3060
+ "learning_rate": 1.4224250325945241e-05,
3061
+ "loss": 1.2363,
3062
+ "step": 509
3063
+ },
3064
+ {
3065
+ "epoch": 0.64,
3066
+ "learning_rate": 1.4211212516297264e-05,
3067
+ "loss": 1.2842,
3068
+ "step": 510
3069
+ },
3070
+ {
3071
+ "epoch": 0.65,
3072
+ "learning_rate": 1.4198174706649284e-05,
3073
+ "loss": 1.2705,
3074
+ "step": 511
3075
+ },
3076
+ {
3077
+ "epoch": 0.65,
3078
+ "learning_rate": 1.4185136897001306e-05,
3079
+ "loss": 1.2667,
3080
+ "step": 512
3081
+ },
3082
+ {
3083
+ "epoch": 0.65,
3084
+ "learning_rate": 1.4172099087353325e-05,
3085
+ "loss": 1.2555,
3086
+ "step": 513
3087
+ },
3088
+ {
3089
+ "epoch": 0.65,
3090
+ "learning_rate": 1.4159061277705345e-05,
3091
+ "loss": 1.2753,
3092
+ "step": 514
3093
+ },
3094
+ {
3095
+ "epoch": 0.65,
3096
+ "learning_rate": 1.4146023468057368e-05,
3097
+ "loss": 1.3025,
3098
+ "step": 515
3099
+ },
3100
+ {
3101
+ "epoch": 0.65,
3102
+ "learning_rate": 1.4132985658409388e-05,
3103
+ "loss": 1.2666,
3104
+ "step": 516
3105
+ },
3106
+ {
3107
+ "epoch": 0.65,
3108
+ "learning_rate": 1.4119947848761408e-05,
3109
+ "loss": 1.2577,
3110
+ "step": 517
3111
+ },
3112
+ {
3113
+ "epoch": 0.65,
3114
+ "learning_rate": 1.4106910039113431e-05,
3115
+ "loss": 1.2736,
3116
+ "step": 518
3117
+ },
3118
+ {
3119
+ "epoch": 0.66,
3120
+ "learning_rate": 1.4093872229465451e-05,
3121
+ "loss": 1.2861,
3122
+ "step": 519
3123
+ },
3124
+ {
3125
+ "epoch": 0.66,
3126
+ "learning_rate": 1.4080834419817473e-05,
3127
+ "loss": 1.2975,
3128
+ "step": 520
3129
+ },
3130
+ {
3131
+ "epoch": 0.66,
3132
+ "learning_rate": 1.4067796610169493e-05,
3133
+ "loss": 1.2819,
3134
+ "step": 521
3135
+ },
3136
+ {
3137
+ "epoch": 0.66,
3138
+ "learning_rate": 1.4054758800521512e-05,
3139
+ "loss": 1.282,
3140
+ "step": 522
3141
+ },
3142
+ {
3143
+ "epoch": 0.66,
3144
+ "learning_rate": 1.4041720990873536e-05,
3145
+ "loss": 1.2714,
3146
+ "step": 523
3147
+ },
3148
+ {
3149
+ "epoch": 0.66,
3150
+ "learning_rate": 1.4028683181225555e-05,
3151
+ "loss": 1.2793,
3152
+ "step": 524
3153
+ },
3154
+ {
3155
+ "epoch": 0.66,
3156
+ "learning_rate": 1.4015645371577575e-05,
3157
+ "loss": 1.2814,
3158
+ "step": 525
3159
+ },
3160
+ {
3161
+ "epoch": 0.66,
3162
+ "learning_rate": 1.4002607561929597e-05,
3163
+ "loss": 1.3046,
3164
+ "step": 526
3165
+ },
3166
+ {
3167
+ "epoch": 0.67,
3168
+ "learning_rate": 1.3989569752281618e-05,
3169
+ "loss": 1.244,
3170
+ "step": 527
3171
+ },
3172
+ {
3173
+ "epoch": 0.67,
3174
+ "learning_rate": 1.3976531942633638e-05,
3175
+ "loss": 1.2451,
3176
+ "step": 528
3177
+ },
3178
+ {
3179
+ "epoch": 0.67,
3180
+ "learning_rate": 1.396349413298566e-05,
3181
+ "loss": 1.2856,
3182
+ "step": 529
3183
+ },
3184
+ {
3185
+ "epoch": 0.67,
3186
+ "learning_rate": 1.395045632333768e-05,
3187
+ "loss": 1.2764,
3188
+ "step": 530
3189
+ },
3190
+ {
3191
+ "epoch": 0.67,
3192
+ "learning_rate": 1.3937418513689703e-05,
3193
+ "loss": 1.2517,
3194
+ "step": 531
3195
+ },
3196
+ {
3197
+ "epoch": 0.67,
3198
+ "learning_rate": 1.3924380704041723e-05,
3199
+ "loss": 1.2867,
3200
+ "step": 532
3201
+ },
3202
+ {
3203
+ "epoch": 0.67,
3204
+ "learning_rate": 1.3911342894393742e-05,
3205
+ "loss": 1.245,
3206
+ "step": 533
3207
+ },
3208
+ {
3209
+ "epoch": 0.68,
3210
+ "learning_rate": 1.3898305084745764e-05,
3211
+ "loss": 1.2917,
3212
+ "step": 534
3213
+ },
3214
+ {
3215
+ "epoch": 0.68,
3216
+ "learning_rate": 1.3885267275097784e-05,
3217
+ "loss": 1.2524,
3218
+ "step": 535
3219
+ },
3220
+ {
3221
+ "epoch": 0.68,
3222
+ "learning_rate": 1.3872229465449805e-05,
3223
+ "loss": 1.2679,
3224
+ "step": 536
3225
+ },
3226
+ {
3227
+ "epoch": 0.68,
3228
+ "learning_rate": 1.3859191655801827e-05,
3229
+ "loss": 1.3221,
3230
+ "step": 537
3231
+ },
3232
+ {
3233
+ "epoch": 0.68,
3234
+ "learning_rate": 1.3846153846153847e-05,
3235
+ "loss": 1.2793,
3236
+ "step": 538
3237
+ },
3238
+ {
3239
+ "epoch": 0.68,
3240
+ "learning_rate": 1.383311603650587e-05,
3241
+ "loss": 1.2444,
3242
+ "step": 539
3243
+ },
3244
+ {
3245
+ "epoch": 0.68,
3246
+ "learning_rate": 1.382007822685789e-05,
3247
+ "loss": 1.2708,
3248
+ "step": 540
3249
+ },
3250
+ {
3251
+ "epoch": 0.68,
3252
+ "learning_rate": 1.380704041720991e-05,
3253
+ "loss": 1.2523,
3254
+ "step": 541
3255
+ },
3256
+ {
3257
+ "epoch": 0.69,
3258
+ "learning_rate": 1.3794002607561931e-05,
3259
+ "loss": 1.2609,
3260
+ "step": 542
3261
+ },
3262
+ {
3263
+ "epoch": 0.69,
3264
+ "learning_rate": 1.3780964797913951e-05,
3265
+ "loss": 1.2598,
3266
+ "step": 543
3267
+ },
3268
+ {
3269
+ "epoch": 0.69,
3270
+ "learning_rate": 1.3767926988265971e-05,
3271
+ "loss": 1.2975,
3272
+ "step": 544
3273
+ },
3274
+ {
3275
+ "epoch": 0.69,
3276
+ "learning_rate": 1.3754889178617994e-05,
3277
+ "loss": 1.2717,
3278
+ "step": 545
3279
+ },
3280
+ {
3281
+ "epoch": 0.69,
3282
+ "learning_rate": 1.3741851368970014e-05,
3283
+ "loss": 1.2833,
3284
+ "step": 546
3285
+ },
3286
+ {
3287
+ "epoch": 0.69,
3288
+ "learning_rate": 1.3728813559322034e-05,
3289
+ "loss": 1.2526,
3290
+ "step": 547
3291
+ },
3292
+ {
3293
+ "epoch": 0.69,
3294
+ "learning_rate": 1.3715775749674057e-05,
3295
+ "loss": 1.2871,
3296
+ "step": 548
3297
+ },
3298
+ {
3299
+ "epoch": 0.69,
3300
+ "learning_rate": 1.3702737940026077e-05,
3301
+ "loss": 1.2983,
3302
+ "step": 549
3303
+ },
3304
+ {
3305
+ "epoch": 0.7,
3306
+ "learning_rate": 1.3689700130378098e-05,
3307
+ "loss": 1.2928,
3308
+ "step": 550
3309
+ },
3310
+ {
3311
+ "epoch": 0.7,
3312
+ "learning_rate": 1.3676662320730118e-05,
3313
+ "loss": 1.2762,
3314
+ "step": 551
3315
+ },
3316
+ {
3317
+ "epoch": 0.7,
3318
+ "learning_rate": 1.3663624511082138e-05,
3319
+ "loss": 1.2646,
3320
+ "step": 552
3321
+ },
3322
+ {
3323
+ "epoch": 0.7,
3324
+ "learning_rate": 1.3650586701434161e-05,
3325
+ "loss": 1.2729,
3326
+ "step": 553
3327
+ },
3328
+ {
3329
+ "epoch": 0.7,
3330
+ "learning_rate": 1.3637548891786181e-05,
3331
+ "loss": 1.3102,
3332
+ "step": 554
3333
+ },
3334
+ {
3335
+ "epoch": 0.7,
3336
+ "learning_rate": 1.3624511082138201e-05,
3337
+ "loss": 1.3051,
3338
+ "step": 555
3339
+ },
3340
+ {
3341
+ "epoch": 0.7,
3342
+ "learning_rate": 1.3611473272490223e-05,
3343
+ "loss": 1.2805,
3344
+ "step": 556
3345
+ },
3346
+ {
3347
+ "epoch": 0.7,
3348
+ "learning_rate": 1.3598435462842244e-05,
3349
+ "loss": 1.2861,
3350
+ "step": 557
3351
+ },
3352
+ {
3353
+ "epoch": 0.71,
3354
+ "learning_rate": 1.3585397653194266e-05,
3355
+ "loss": 1.2333,
3356
+ "step": 558
3357
+ },
3358
+ {
3359
+ "epoch": 0.71,
3360
+ "learning_rate": 1.3572359843546285e-05,
3361
+ "loss": 1.2615,
3362
+ "step": 559
3363
+ },
3364
+ {
3365
+ "epoch": 0.71,
3366
+ "learning_rate": 1.3559322033898305e-05,
3367
+ "loss": 1.2748,
3368
+ "step": 560
3369
+ },
3370
+ {
3371
+ "epoch": 0.71,
3372
+ "learning_rate": 1.3546284224250328e-05,
3373
+ "loss": 1.2869,
3374
+ "step": 561
3375
+ },
3376
+ {
3377
+ "epoch": 0.71,
3378
+ "learning_rate": 1.3533246414602348e-05,
3379
+ "loss": 1.2874,
3380
+ "step": 562
3381
+ },
3382
+ {
3383
+ "epoch": 0.71,
3384
+ "learning_rate": 1.3520208604954368e-05,
3385
+ "loss": 1.2952,
3386
+ "step": 563
3387
+ },
3388
+ {
3389
+ "epoch": 0.71,
3390
+ "learning_rate": 1.350717079530639e-05,
3391
+ "loss": 1.2576,
3392
+ "step": 564
3393
+ },
3394
+ {
3395
+ "epoch": 0.71,
3396
+ "learning_rate": 1.349413298565841e-05,
3397
+ "loss": 1.295,
3398
+ "step": 565
3399
+ },
3400
+ {
3401
+ "epoch": 0.72,
3402
+ "learning_rate": 1.3481095176010431e-05,
3403
+ "loss": 1.2789,
3404
+ "step": 566
3405
+ },
3406
+ {
3407
+ "epoch": 0.72,
3408
+ "learning_rate": 1.3468057366362453e-05,
3409
+ "loss": 1.2804,
3410
+ "step": 567
3411
+ },
3412
+ {
3413
+ "epoch": 0.72,
3414
+ "learning_rate": 1.3455019556714472e-05,
3415
+ "loss": 1.2206,
3416
+ "step": 568
3417
+ },
3418
+ {
3419
+ "epoch": 0.72,
3420
+ "learning_rate": 1.3441981747066496e-05,
3421
+ "loss": 1.2588,
3422
+ "step": 569
3423
+ },
3424
+ {
3425
+ "epoch": 0.72,
3426
+ "learning_rate": 1.3428943937418515e-05,
3427
+ "loss": 1.2982,
3428
+ "step": 570
3429
+ },
3430
+ {
3431
+ "epoch": 0.72,
3432
+ "learning_rate": 1.3415906127770535e-05,
3433
+ "loss": 1.3011,
3434
+ "step": 571
3435
+ },
3436
+ {
3437
+ "epoch": 0.72,
3438
+ "learning_rate": 1.3402868318122557e-05,
3439
+ "loss": 1.2406,
3440
+ "step": 572
3441
+ },
3442
+ {
3443
+ "epoch": 0.72,
3444
+ "learning_rate": 1.3389830508474577e-05,
3445
+ "loss": 1.2391,
3446
+ "step": 573
3447
+ },
3448
+ {
3449
+ "epoch": 0.73,
3450
+ "learning_rate": 1.3376792698826597e-05,
3451
+ "loss": 1.2511,
3452
+ "step": 574
3453
+ },
3454
+ {
3455
+ "epoch": 0.73,
3456
+ "learning_rate": 1.336375488917862e-05,
3457
+ "loss": 1.2675,
3458
+ "step": 575
3459
+ },
3460
+ {
3461
+ "epoch": 0.73,
3462
+ "learning_rate": 1.335071707953064e-05,
3463
+ "loss": 1.2623,
3464
+ "step": 576
3465
+ },
3466
+ {
3467
+ "epoch": 0.73,
3468
+ "learning_rate": 1.3337679269882661e-05,
3469
+ "loss": 1.2775,
3470
+ "step": 577
3471
+ },
3472
+ {
3473
+ "epoch": 0.73,
3474
+ "learning_rate": 1.3324641460234683e-05,
3475
+ "loss": 1.2499,
3476
+ "step": 578
3477
+ },
3478
+ {
3479
+ "epoch": 0.73,
3480
+ "learning_rate": 1.3311603650586703e-05,
3481
+ "loss": 1.2494,
3482
+ "step": 579
3483
+ },
3484
+ {
3485
+ "epoch": 0.73,
3486
+ "learning_rate": 1.3298565840938724e-05,
3487
+ "loss": 1.2524,
3488
+ "step": 580
3489
+ },
3490
+ {
3491
+ "epoch": 0.73,
3492
+ "learning_rate": 1.3285528031290744e-05,
3493
+ "loss": 1.2797,
3494
+ "step": 581
3495
+ },
3496
+ {
3497
+ "epoch": 0.74,
3498
+ "learning_rate": 1.3272490221642764e-05,
3499
+ "loss": 1.282,
3500
+ "step": 582
3501
+ },
3502
+ {
3503
+ "epoch": 0.74,
3504
+ "learning_rate": 1.3259452411994787e-05,
3505
+ "loss": 1.2408,
3506
+ "step": 583
3507
+ },
3508
+ {
3509
+ "epoch": 0.74,
3510
+ "learning_rate": 1.3246414602346807e-05,
3511
+ "loss": 1.2726,
3512
+ "step": 584
3513
+ },
3514
+ {
3515
+ "epoch": 0.74,
3516
+ "learning_rate": 1.3233376792698827e-05,
3517
+ "loss": 1.2561,
3518
+ "step": 585
3519
+ },
3520
+ {
3521
+ "epoch": 0.74,
3522
+ "learning_rate": 1.3220338983050848e-05,
3523
+ "loss": 1.2661,
3524
+ "step": 586
3525
+ },
3526
+ {
3527
+ "epoch": 0.74,
3528
+ "learning_rate": 1.320730117340287e-05,
3529
+ "loss": 1.2561,
3530
+ "step": 587
3531
+ },
3532
+ {
3533
+ "epoch": 0.74,
3534
+ "learning_rate": 1.3194263363754891e-05,
3535
+ "loss": 1.243,
3536
+ "step": 588
3537
+ },
3538
+ {
3539
+ "epoch": 0.74,
3540
+ "learning_rate": 1.3181225554106911e-05,
3541
+ "loss": 1.2936,
3542
+ "step": 589
3543
+ },
3544
+ {
3545
+ "epoch": 0.75,
3546
+ "learning_rate": 1.3168187744458931e-05,
3547
+ "loss": 1.2764,
3548
+ "step": 590
3549
+ },
3550
+ {
3551
+ "epoch": 0.75,
3552
+ "learning_rate": 1.3155149934810954e-05,
3553
+ "loss": 1.2767,
3554
+ "step": 591
3555
+ },
3556
+ {
3557
+ "epoch": 0.75,
3558
+ "learning_rate": 1.3142112125162974e-05,
3559
+ "loss": 1.2394,
3560
+ "step": 592
3561
+ },
3562
+ {
3563
+ "epoch": 0.75,
3564
+ "learning_rate": 1.3129074315514994e-05,
3565
+ "loss": 1.2588,
3566
+ "step": 593
3567
+ },
3568
+ {
3569
+ "epoch": 0.75,
3570
+ "learning_rate": 1.3116036505867015e-05,
3571
+ "loss": 1.2523,
3572
+ "step": 594
3573
+ },
3574
+ {
3575
+ "epoch": 0.75,
3576
+ "learning_rate": 1.3102998696219035e-05,
3577
+ "loss": 1.2491,
3578
+ "step": 595
3579
+ },
3580
+ {
3581
+ "epoch": 0.75,
3582
+ "learning_rate": 1.3089960886571058e-05,
3583
+ "loss": 1.2787,
3584
+ "step": 596
3585
+ },
3586
+ {
3587
+ "epoch": 0.75,
3588
+ "learning_rate": 1.3076923076923078e-05,
3589
+ "loss": 1.2155,
3590
+ "step": 597
3591
+ },
3592
+ {
3593
+ "epoch": 0.76,
3594
+ "learning_rate": 1.3063885267275098e-05,
3595
+ "loss": 1.2714,
3596
+ "step": 598
3597
+ },
3598
+ {
3599
+ "epoch": 0.76,
3600
+ "learning_rate": 1.305084745762712e-05,
3601
+ "loss": 1.2273,
3602
+ "step": 599
3603
+ },
3604
+ {
3605
+ "epoch": 0.76,
3606
+ "learning_rate": 1.3037809647979141e-05,
3607
+ "loss": 1.2542,
3608
+ "step": 600
3609
+ },
3610
+ {
3611
+ "epoch": 0.76,
3612
+ "learning_rate": 1.3024771838331161e-05,
3613
+ "loss": 1.2642,
3614
+ "step": 601
3615
+ },
3616
+ {
3617
+ "epoch": 0.76,
3618
+ "learning_rate": 1.3011734028683183e-05,
3619
+ "loss": 1.2722,
3620
+ "step": 602
3621
+ },
3622
+ {
3623
+ "epoch": 0.76,
3624
+ "learning_rate": 1.2998696219035202e-05,
3625
+ "loss": 1.2264,
3626
+ "step": 603
3627
+ },
3628
+ {
3629
+ "epoch": 0.76,
3630
+ "learning_rate": 1.2985658409387226e-05,
3631
+ "loss": 1.2885,
3632
+ "step": 604
3633
+ },
3634
+ {
3635
+ "epoch": 0.76,
3636
+ "learning_rate": 1.2972620599739245e-05,
3637
+ "loss": 1.2732,
3638
+ "step": 605
3639
+ },
3640
+ {
3641
+ "epoch": 0.77,
3642
+ "learning_rate": 1.2959582790091265e-05,
3643
+ "loss": 1.2704,
3644
+ "step": 606
3645
+ },
3646
+ {
3647
+ "epoch": 0.77,
3648
+ "learning_rate": 1.2946544980443287e-05,
3649
+ "loss": 1.2737,
3650
+ "step": 607
3651
+ },
3652
+ {
3653
+ "epoch": 0.77,
3654
+ "learning_rate": 1.2933507170795307e-05,
3655
+ "loss": 1.2621,
3656
+ "step": 608
3657
+ },
3658
+ {
3659
+ "epoch": 0.77,
3660
+ "learning_rate": 1.2920469361147328e-05,
3661
+ "loss": 1.2648,
3662
+ "step": 609
3663
+ },
3664
+ {
3665
+ "epoch": 0.77,
3666
+ "learning_rate": 1.290743155149935e-05,
3667
+ "loss": 1.2455,
3668
+ "step": 610
3669
+ },
3670
+ {
3671
+ "epoch": 0.77,
3672
+ "learning_rate": 1.289439374185137e-05,
3673
+ "loss": 1.2411,
3674
+ "step": 611
3675
+ },
3676
+ {
3677
+ "epoch": 0.77,
3678
+ "learning_rate": 1.288135593220339e-05,
3679
+ "loss": 1.2219,
3680
+ "step": 612
3681
+ },
3682
+ {
3683
+ "epoch": 0.77,
3684
+ "learning_rate": 1.2868318122555413e-05,
3685
+ "loss": 1.266,
3686
+ "step": 613
3687
+ },
3688
+ {
3689
+ "epoch": 0.78,
3690
+ "learning_rate": 1.2855280312907432e-05,
3691
+ "loss": 1.2748,
3692
+ "step": 614
3693
+ },
3694
+ {
3695
+ "epoch": 0.78,
3696
+ "learning_rate": 1.2842242503259454e-05,
3697
+ "loss": 1.257,
3698
+ "step": 615
3699
+ },
3700
+ {
3701
+ "epoch": 0.78,
3702
+ "learning_rate": 1.2829204693611474e-05,
3703
+ "loss": 1.2866,
3704
+ "step": 616
3705
+ },
3706
+ {
3707
+ "epoch": 0.78,
3708
+ "learning_rate": 1.2816166883963494e-05,
3709
+ "loss": 1.2799,
3710
+ "step": 617
3711
+ },
3712
+ {
3713
+ "epoch": 0.78,
3714
+ "learning_rate": 1.2803129074315517e-05,
3715
+ "loss": 1.2476,
3716
+ "step": 618
3717
+ },
3718
+ {
3719
+ "epoch": 0.78,
3720
+ "learning_rate": 1.2790091264667537e-05,
3721
+ "loss": 1.2384,
3722
+ "step": 619
3723
+ },
3724
+ {
3725
+ "epoch": 0.78,
3726
+ "learning_rate": 1.2777053455019557e-05,
3727
+ "loss": 1.2716,
3728
+ "step": 620
3729
+ },
3730
+ {
3731
+ "epoch": 0.79,
3732
+ "learning_rate": 1.276401564537158e-05,
3733
+ "loss": 1.2717,
3734
+ "step": 621
3735
+ },
3736
+ {
3737
+ "epoch": 0.79,
3738
+ "learning_rate": 1.27509778357236e-05,
3739
+ "loss": 1.235,
3740
+ "step": 622
3741
+ },
3742
+ {
3743
+ "epoch": 0.79,
3744
+ "learning_rate": 1.2737940026075621e-05,
3745
+ "loss": 1.2544,
3746
+ "step": 623
3747
+ },
3748
+ {
3749
+ "epoch": 0.79,
3750
+ "learning_rate": 1.2724902216427641e-05,
3751
+ "loss": 1.218,
3752
+ "step": 624
3753
+ },
3754
+ {
3755
+ "epoch": 0.79,
3756
+ "learning_rate": 1.2711864406779661e-05,
3757
+ "loss": 1.276,
3758
+ "step": 625
3759
+ },
3760
+ {
3761
+ "epoch": 0.79,
3762
+ "learning_rate": 1.2698826597131684e-05,
3763
+ "loss": 1.2299,
3764
+ "step": 626
3765
+ },
3766
+ {
3767
+ "epoch": 0.79,
3768
+ "learning_rate": 1.2685788787483704e-05,
3769
+ "loss": 1.2955,
3770
+ "step": 627
3771
+ },
3772
+ {
3773
+ "epoch": 0.79,
3774
+ "learning_rate": 1.2672750977835724e-05,
3775
+ "loss": 1.2668,
3776
+ "step": 628
3777
+ },
3778
+ {
3779
+ "epoch": 0.8,
3780
+ "learning_rate": 1.2659713168187745e-05,
3781
+ "loss": 1.2819,
3782
+ "step": 629
3783
+ },
3784
+ {
3785
+ "epoch": 0.8,
3786
+ "learning_rate": 1.2646675358539767e-05,
3787
+ "loss": 1.2571,
3788
+ "step": 630
3789
+ },
3790
+ {
3791
+ "epoch": 0.8,
3792
+ "learning_rate": 1.2633637548891787e-05,
3793
+ "loss": 1.272,
3794
+ "step": 631
3795
+ },
3796
+ {
3797
+ "epoch": 0.8,
3798
+ "learning_rate": 1.2620599739243808e-05,
3799
+ "loss": 1.3043,
3800
+ "step": 632
3801
+ },
3802
+ {
3803
+ "epoch": 0.8,
3804
+ "learning_rate": 1.2607561929595828e-05,
3805
+ "loss": 1.2443,
3806
+ "step": 633
3807
+ },
3808
+ {
3809
+ "epoch": 0.8,
3810
+ "learning_rate": 1.2594524119947851e-05,
3811
+ "loss": 1.2615,
3812
+ "step": 634
3813
+ },
3814
+ {
3815
+ "epoch": 0.8,
3816
+ "learning_rate": 1.2581486310299871e-05,
3817
+ "loss": 1.2781,
3818
+ "step": 635
3819
+ },
3820
+ {
3821
+ "epoch": 0.8,
3822
+ "learning_rate": 1.2568448500651891e-05,
3823
+ "loss": 1.2577,
3824
+ "step": 636
3825
+ },
3826
+ {
3827
+ "epoch": 0.81,
3828
+ "learning_rate": 1.2555410691003912e-05,
3829
+ "loss": 1.2922,
3830
+ "step": 637
3831
+ },
3832
+ {
3833
+ "epoch": 0.81,
3834
+ "learning_rate": 1.2542372881355932e-05,
3835
+ "loss": 1.2979,
3836
+ "step": 638
3837
+ },
3838
+ {
3839
+ "epoch": 0.81,
3840
+ "learning_rate": 1.2529335071707954e-05,
3841
+ "loss": 1.2463,
3842
+ "step": 639
3843
+ },
3844
+ {
3845
+ "epoch": 0.81,
3846
+ "learning_rate": 1.2516297262059975e-05,
3847
+ "loss": 1.2788,
3848
+ "step": 640
3849
+ },
3850
+ {
3851
+ "epoch": 0.81,
3852
+ "learning_rate": 1.2503259452411995e-05,
3853
+ "loss": 1.2393,
3854
+ "step": 641
3855
+ },
3856
+ {
3857
+ "epoch": 0.81,
3858
+ "learning_rate": 1.2490221642764018e-05,
3859
+ "loss": 1.2751,
3860
+ "step": 642
3861
+ },
3862
+ {
3863
+ "epoch": 0.81,
3864
+ "learning_rate": 1.2477183833116038e-05,
3865
+ "loss": 1.2578,
3866
+ "step": 643
3867
+ },
3868
+ {
3869
+ "epoch": 0.81,
3870
+ "learning_rate": 1.2464146023468058e-05,
3871
+ "loss": 1.2604,
3872
+ "step": 644
3873
+ },
3874
+ {
3875
+ "epoch": 0.82,
3876
+ "learning_rate": 1.245110821382008e-05,
3877
+ "loss": 1.2474,
3878
+ "step": 645
3879
+ },
3880
+ {
3881
+ "epoch": 0.82,
3882
+ "learning_rate": 1.24380704041721e-05,
3883
+ "loss": 1.2448,
3884
+ "step": 646
3885
+ },
3886
+ {
3887
+ "epoch": 0.82,
3888
+ "learning_rate": 1.242503259452412e-05,
3889
+ "loss": 1.282,
3890
+ "step": 647
3891
+ },
3892
+ {
3893
+ "epoch": 0.82,
3894
+ "learning_rate": 1.2411994784876143e-05,
3895
+ "loss": 1.2609,
3896
+ "step": 648
3897
+ },
3898
+ {
3899
+ "epoch": 0.82,
3900
+ "learning_rate": 1.2398956975228162e-05,
3901
+ "loss": 1.2749,
3902
+ "step": 649
3903
+ },
3904
+ {
3905
+ "epoch": 0.82,
3906
+ "learning_rate": 1.2385919165580182e-05,
3907
+ "loss": 1.3005,
3908
+ "step": 650
3909
+ },
3910
+ {
3911
+ "epoch": 0.82,
3912
+ "learning_rate": 1.2372881355932205e-05,
3913
+ "loss": 1.2745,
3914
+ "step": 651
3915
+ },
3916
+ {
3917
+ "epoch": 0.82,
3918
+ "learning_rate": 1.2359843546284225e-05,
3919
+ "loss": 1.2719,
3920
+ "step": 652
3921
+ },
3922
+ {
3923
+ "epoch": 0.83,
3924
+ "learning_rate": 1.2346805736636247e-05,
3925
+ "loss": 1.2699,
3926
+ "step": 653
3927
+ },
3928
+ {
3929
+ "epoch": 0.83,
3930
+ "learning_rate": 1.2333767926988267e-05,
3931
+ "loss": 1.2916,
3932
+ "step": 654
3933
+ },
3934
+ {
3935
+ "epoch": 0.83,
3936
+ "learning_rate": 1.2320730117340287e-05,
3937
+ "loss": 1.2686,
3938
+ "step": 655
3939
+ },
3940
+ {
3941
+ "epoch": 0.83,
3942
+ "learning_rate": 1.230769230769231e-05,
3943
+ "loss": 1.26,
3944
+ "step": 656
3945
+ },
3946
+ {
3947
+ "epoch": 0.83,
3948
+ "learning_rate": 1.229465449804433e-05,
3949
+ "loss": 1.2863,
3950
+ "step": 657
3951
+ },
3952
+ {
3953
+ "epoch": 0.83,
3954
+ "learning_rate": 1.228161668839635e-05,
3955
+ "loss": 1.2948,
3956
+ "step": 658
3957
+ },
3958
+ {
3959
+ "epoch": 0.83,
3960
+ "learning_rate": 1.2268578878748371e-05,
3961
+ "loss": 1.2524,
3962
+ "step": 659
3963
+ },
3964
+ {
3965
+ "epoch": 0.83,
3966
+ "learning_rate": 1.2255541069100392e-05,
3967
+ "loss": 1.287,
3968
+ "step": 660
3969
+ },
3970
+ {
3971
+ "epoch": 0.84,
3972
+ "learning_rate": 1.2242503259452414e-05,
3973
+ "loss": 1.2449,
3974
+ "step": 661
3975
+ },
3976
+ {
3977
+ "epoch": 0.84,
3978
+ "learning_rate": 1.2229465449804434e-05,
3979
+ "loss": 1.2594,
3980
+ "step": 662
3981
+ },
3982
+ {
3983
+ "epoch": 0.84,
3984
+ "learning_rate": 1.2216427640156454e-05,
3985
+ "loss": 1.1996,
3986
+ "step": 663
3987
+ },
3988
+ {
3989
+ "epoch": 0.84,
3990
+ "learning_rate": 1.2203389830508477e-05,
3991
+ "loss": 1.2283,
3992
+ "step": 664
3993
+ },
3994
+ {
3995
+ "epoch": 0.84,
3996
+ "learning_rate": 1.2190352020860497e-05,
3997
+ "loss": 1.2709,
3998
+ "step": 665
3999
+ },
4000
+ {
4001
+ "epoch": 0.84,
4002
+ "learning_rate": 1.2177314211212517e-05,
4003
+ "loss": 1.265,
4004
+ "step": 666
4005
+ },
4006
+ {
4007
+ "epoch": 0.84,
4008
+ "learning_rate": 1.2164276401564538e-05,
4009
+ "loss": 1.2476,
4010
+ "step": 667
4011
+ },
4012
+ {
4013
+ "epoch": 0.84,
4014
+ "learning_rate": 1.2151238591916558e-05,
4015
+ "loss": 1.2646,
4016
+ "step": 668
4017
+ },
4018
+ {
4019
+ "epoch": 0.85,
4020
+ "learning_rate": 1.213820078226858e-05,
4021
+ "loss": 1.2458,
4022
+ "step": 669
4023
+ },
4024
+ {
4025
+ "epoch": 0.85,
4026
+ "learning_rate": 1.2125162972620601e-05,
4027
+ "loss": 1.281,
4028
+ "step": 670
4029
+ },
4030
+ {
4031
+ "epoch": 0.85,
4032
+ "learning_rate": 1.2112125162972621e-05,
4033
+ "loss": 1.2588,
4034
+ "step": 671
4035
+ },
4036
+ {
4037
+ "epoch": 0.85,
4038
+ "learning_rate": 1.2099087353324644e-05,
4039
+ "loss": 1.2638,
4040
+ "step": 672
4041
+ },
4042
+ {
4043
+ "epoch": 0.85,
4044
+ "learning_rate": 1.2086049543676664e-05,
4045
+ "loss": 1.2916,
4046
+ "step": 673
4047
+ },
4048
+ {
4049
+ "epoch": 0.85,
4050
+ "learning_rate": 1.2073011734028684e-05,
4051
+ "loss": 1.2437,
4052
+ "step": 674
4053
+ },
4054
+ {
4055
+ "epoch": 0.85,
4056
+ "learning_rate": 1.2059973924380705e-05,
4057
+ "loss": 1.2164,
4058
+ "step": 675
4059
+ },
4060
+ {
4061
+ "epoch": 0.85,
4062
+ "learning_rate": 1.2046936114732725e-05,
4063
+ "loss": 1.2417,
4064
+ "step": 676
4065
+ },
4066
+ {
4067
+ "epoch": 0.86,
4068
+ "learning_rate": 1.2033898305084745e-05,
4069
+ "loss": 1.2457,
4070
+ "step": 677
4071
+ },
4072
+ {
4073
+ "epoch": 0.86,
4074
+ "learning_rate": 1.2020860495436768e-05,
4075
+ "loss": 1.2517,
4076
+ "step": 678
4077
+ },
4078
+ {
4079
+ "epoch": 0.86,
4080
+ "learning_rate": 1.2007822685788788e-05,
4081
+ "loss": 1.2539,
4082
+ "step": 679
4083
+ },
4084
+ {
4085
+ "epoch": 0.86,
4086
+ "learning_rate": 1.199478487614081e-05,
4087
+ "loss": 1.2645,
4088
+ "step": 680
4089
+ },
4090
+ {
4091
+ "epoch": 0.86,
4092
+ "learning_rate": 1.1981747066492831e-05,
4093
+ "loss": 1.2408,
4094
+ "step": 681
4095
+ },
4096
+ {
4097
+ "epoch": 0.86,
4098
+ "learning_rate": 1.1968709256844851e-05,
4099
+ "loss": 1.2574,
4100
+ "step": 682
4101
+ },
4102
+ {
4103
+ "epoch": 0.86,
4104
+ "learning_rate": 1.1955671447196873e-05,
4105
+ "loss": 1.2322,
4106
+ "step": 683
4107
+ },
4108
+ {
4109
+ "epoch": 0.86,
4110
+ "learning_rate": 1.1942633637548892e-05,
4111
+ "loss": 1.2244,
4112
+ "step": 684
4113
+ },
4114
+ {
4115
+ "epoch": 0.87,
4116
+ "learning_rate": 1.1929595827900912e-05,
4117
+ "loss": 1.272,
4118
+ "step": 685
4119
+ },
4120
+ {
4121
+ "epoch": 0.87,
4122
+ "learning_rate": 1.1916558018252935e-05,
4123
+ "loss": 1.2791,
4124
+ "step": 686
4125
+ },
4126
+ {
4127
+ "epoch": 0.87,
4128
+ "learning_rate": 1.1903520208604955e-05,
4129
+ "loss": 1.2562,
4130
+ "step": 687
4131
+ },
4132
+ {
4133
+ "epoch": 0.87,
4134
+ "learning_rate": 1.1890482398956975e-05,
4135
+ "loss": 1.2441,
4136
+ "step": 688
4137
+ },
4138
+ {
4139
+ "epoch": 0.87,
4140
+ "learning_rate": 1.1877444589308997e-05,
4141
+ "loss": 1.2437,
4142
+ "step": 689
4143
+ },
4144
+ {
4145
+ "epoch": 0.87,
4146
+ "learning_rate": 1.1864406779661018e-05,
4147
+ "loss": 1.2618,
4148
+ "step": 690
4149
+ },
4150
+ {
4151
+ "epoch": 0.87,
4152
+ "learning_rate": 1.185136897001304e-05,
4153
+ "loss": 1.2552,
4154
+ "step": 691
4155
+ },
4156
+ {
4157
+ "epoch": 0.87,
4158
+ "learning_rate": 1.183833116036506e-05,
4159
+ "loss": 1.2729,
4160
+ "step": 692
4161
+ },
4162
+ {
4163
+ "epoch": 0.88,
4164
+ "learning_rate": 1.182529335071708e-05,
4165
+ "loss": 1.2729,
4166
+ "step": 693
4167
+ },
4168
+ {
4169
+ "epoch": 0.88,
4170
+ "learning_rate": 1.1812255541069103e-05,
4171
+ "loss": 1.2659,
4172
+ "step": 694
4173
+ },
4174
+ {
4175
+ "epoch": 0.88,
4176
+ "learning_rate": 1.1799217731421122e-05,
4177
+ "loss": 1.2323,
4178
+ "step": 695
4179
+ },
4180
+ {
4181
+ "epoch": 0.88,
4182
+ "learning_rate": 1.1786179921773142e-05,
4183
+ "loss": 1.2218,
4184
+ "step": 696
4185
+ },
4186
+ {
4187
+ "epoch": 0.88,
4188
+ "learning_rate": 1.1773142112125164e-05,
4189
+ "loss": 1.2419,
4190
+ "step": 697
4191
+ },
4192
+ {
4193
+ "epoch": 0.88,
4194
+ "learning_rate": 1.1760104302477184e-05,
4195
+ "loss": 1.2745,
4196
+ "step": 698
4197
+ },
4198
+ {
4199
+ "epoch": 0.88,
4200
+ "learning_rate": 1.1747066492829207e-05,
4201
+ "loss": 1.2887,
4202
+ "step": 699
4203
+ },
4204
+ {
4205
+ "epoch": 0.88,
4206
+ "learning_rate": 1.1734028683181227e-05,
4207
+ "loss": 1.2111,
4208
+ "step": 700
4209
+ },
4210
+ {
4211
+ "epoch": 0.89,
4212
+ "learning_rate": 1.1720990873533247e-05,
4213
+ "loss": 1.2789,
4214
+ "step": 701
4215
+ },
4216
+ {
4217
+ "epoch": 0.89,
4218
+ "learning_rate": 1.170795306388527e-05,
4219
+ "loss": 1.2781,
4220
+ "step": 702
4221
+ },
4222
+ {
4223
+ "epoch": 0.89,
4224
+ "learning_rate": 1.169491525423729e-05,
4225
+ "loss": 1.2439,
4226
+ "step": 703
4227
+ },
4228
+ {
4229
+ "epoch": 0.89,
4230
+ "learning_rate": 1.168187744458931e-05,
4231
+ "loss": 1.2841,
4232
+ "step": 704
4233
+ },
4234
+ {
4235
+ "epoch": 0.89,
4236
+ "learning_rate": 1.1668839634941331e-05,
4237
+ "loss": 1.2676,
4238
+ "step": 705
4239
+ },
4240
+ {
4241
+ "epoch": 0.89,
4242
+ "learning_rate": 1.165580182529335e-05,
4243
+ "loss": 1.2831,
4244
+ "step": 706
4245
+ },
4246
+ {
4247
+ "epoch": 0.89,
4248
+ "learning_rate": 1.1642764015645374e-05,
4249
+ "loss": 1.2432,
4250
+ "step": 707
4251
+ },
4252
+ {
4253
+ "epoch": 0.9,
4254
+ "learning_rate": 1.1629726205997394e-05,
4255
+ "loss": 1.265,
4256
+ "step": 708
4257
+ },
4258
+ {
4259
+ "epoch": 0.9,
4260
+ "learning_rate": 1.1616688396349414e-05,
4261
+ "loss": 1.2151,
4262
+ "step": 709
4263
+ },
4264
+ {
4265
+ "epoch": 0.9,
4266
+ "learning_rate": 1.1603650586701435e-05,
4267
+ "loss": 1.2303,
4268
+ "step": 710
4269
+ },
4270
+ {
4271
+ "epoch": 0.9,
4272
+ "learning_rate": 1.1590612777053457e-05,
4273
+ "loss": 1.2487,
4274
+ "step": 711
4275
+ },
4276
+ {
4277
+ "epoch": 0.9,
4278
+ "learning_rate": 1.1577574967405477e-05,
4279
+ "loss": 1.2618,
4280
+ "step": 712
4281
+ },
4282
+ {
4283
+ "epoch": 0.9,
4284
+ "learning_rate": 1.1564537157757498e-05,
4285
+ "loss": 1.2507,
4286
+ "step": 713
4287
+ },
4288
+ {
4289
+ "epoch": 0.9,
4290
+ "learning_rate": 1.1551499348109518e-05,
4291
+ "loss": 1.2502,
4292
+ "step": 714
4293
+ },
4294
+ {
4295
+ "epoch": 0.9,
4296
+ "learning_rate": 1.1538461538461538e-05,
4297
+ "loss": 1.2424,
4298
+ "step": 715
4299
+ },
4300
+ {
4301
+ "epoch": 0.91,
4302
+ "learning_rate": 1.1525423728813561e-05,
4303
+ "loss": 1.2504,
4304
+ "step": 716
4305
+ },
4306
+ {
4307
+ "epoch": 0.91,
4308
+ "learning_rate": 1.1512385919165581e-05,
4309
+ "loss": 1.2261,
4310
+ "step": 717
4311
+ },
4312
+ {
4313
+ "epoch": 0.91,
4314
+ "learning_rate": 1.1499348109517602e-05,
4315
+ "loss": 1.2887,
4316
+ "step": 718
4317
+ },
4318
+ {
4319
+ "epoch": 0.91,
4320
+ "learning_rate": 1.1486310299869622e-05,
4321
+ "loss": 1.2443,
4322
+ "step": 719
4323
+ },
4324
+ {
4325
+ "epoch": 0.91,
4326
+ "learning_rate": 1.1473272490221644e-05,
4327
+ "loss": 1.2317,
4328
+ "step": 720
4329
+ },
4330
+ {
4331
+ "epoch": 0.91,
4332
+ "learning_rate": 1.1460234680573665e-05,
4333
+ "loss": 1.2329,
4334
+ "step": 721
4335
+ },
4336
+ {
4337
+ "epoch": 0.91,
4338
+ "learning_rate": 1.1447196870925685e-05,
4339
+ "loss": 1.2781,
4340
+ "step": 722
4341
+ },
4342
+ {
4343
+ "epoch": 0.91,
4344
+ "learning_rate": 1.1434159061277705e-05,
4345
+ "loss": 1.2329,
4346
+ "step": 723
4347
+ },
4348
+ {
4349
+ "epoch": 0.92,
4350
+ "learning_rate": 1.1421121251629728e-05,
4351
+ "loss": 1.2723,
4352
+ "step": 724
4353
+ },
4354
+ {
4355
+ "epoch": 0.92,
4356
+ "learning_rate": 1.1408083441981748e-05,
4357
+ "loss": 1.236,
4358
+ "step": 725
4359
+ },
4360
+ {
4361
+ "epoch": 0.92,
4362
+ "learning_rate": 1.139504563233377e-05,
4363
+ "loss": 1.2114,
4364
+ "step": 726
4365
+ },
4366
+ {
4367
+ "epoch": 0.92,
4368
+ "learning_rate": 1.138200782268579e-05,
4369
+ "loss": 1.2483,
4370
+ "step": 727
4371
+ },
4372
+ {
4373
+ "epoch": 0.92,
4374
+ "learning_rate": 1.136897001303781e-05,
4375
+ "loss": 1.2572,
4376
+ "step": 728
4377
+ },
4378
+ {
4379
+ "epoch": 0.92,
4380
+ "learning_rate": 1.1355932203389833e-05,
4381
+ "loss": 1.2286,
4382
+ "step": 729
4383
+ },
4384
+ {
4385
+ "epoch": 0.92,
4386
+ "learning_rate": 1.1342894393741852e-05,
4387
+ "loss": 1.2546,
4388
+ "step": 730
4389
+ },
4390
+ {
4391
+ "epoch": 0.92,
4392
+ "learning_rate": 1.1329856584093872e-05,
4393
+ "loss": 1.238,
4394
+ "step": 731
4395
+ },
4396
+ {
4397
+ "epoch": 0.93,
4398
+ "learning_rate": 1.1316818774445895e-05,
4399
+ "loss": 1.246,
4400
+ "step": 732
4401
+ },
4402
+ {
4403
+ "epoch": 0.93,
4404
+ "learning_rate": 1.1303780964797915e-05,
4405
+ "loss": 1.2441,
4406
+ "step": 733
4407
+ },
4408
+ {
4409
+ "epoch": 0.93,
4410
+ "learning_rate": 1.1290743155149935e-05,
4411
+ "loss": 1.2487,
4412
+ "step": 734
4413
+ },
4414
+ {
4415
+ "epoch": 0.93,
4416
+ "learning_rate": 1.1277705345501957e-05,
4417
+ "loss": 1.2324,
4418
+ "step": 735
4419
+ },
4420
+ {
4421
+ "epoch": 0.93,
4422
+ "learning_rate": 1.1264667535853976e-05,
4423
+ "loss": 1.2577,
4424
+ "step": 736
4425
+ },
4426
+ {
4427
+ "epoch": 0.93,
4428
+ "learning_rate": 1.1251629726206e-05,
4429
+ "loss": 1.2184,
4430
+ "step": 737
4431
+ },
4432
+ {
4433
+ "epoch": 0.93,
4434
+ "learning_rate": 1.123859191655802e-05,
4435
+ "loss": 1.2179,
4436
+ "step": 738
4437
+ },
4438
+ {
4439
+ "epoch": 0.93,
4440
+ "learning_rate": 1.122555410691004e-05,
4441
+ "loss": 1.2626,
4442
+ "step": 739
4443
+ },
4444
+ {
4445
+ "epoch": 0.94,
4446
+ "learning_rate": 1.1212516297262061e-05,
4447
+ "loss": 1.2292,
4448
+ "step": 740
4449
+ },
4450
+ {
4451
+ "epoch": 0.94,
4452
+ "learning_rate": 1.1199478487614082e-05,
4453
+ "loss": 1.274,
4454
+ "step": 741
4455
+ },
4456
+ {
4457
+ "epoch": 0.94,
4458
+ "learning_rate": 1.1186440677966102e-05,
4459
+ "loss": 1.2303,
4460
+ "step": 742
4461
+ },
4462
+ {
4463
+ "epoch": 0.94,
4464
+ "learning_rate": 1.1173402868318124e-05,
4465
+ "loss": 1.2963,
4466
+ "step": 743
4467
+ },
4468
+ {
4469
+ "epoch": 0.94,
4470
+ "learning_rate": 1.1160365058670144e-05,
4471
+ "loss": 1.231,
4472
+ "step": 744
4473
+ },
4474
+ {
4475
+ "epoch": 0.94,
4476
+ "learning_rate": 1.1147327249022167e-05,
4477
+ "loss": 1.2573,
4478
+ "step": 745
4479
+ },
4480
+ {
4481
+ "epoch": 0.94,
4482
+ "learning_rate": 1.1134289439374187e-05,
4483
+ "loss": 1.2445,
4484
+ "step": 746
4485
+ },
4486
+ {
4487
+ "epoch": 0.94,
4488
+ "learning_rate": 1.1121251629726207e-05,
4489
+ "loss": 1.2458,
4490
+ "step": 747
4491
+ },
4492
+ {
4493
+ "epoch": 0.95,
4494
+ "learning_rate": 1.1108213820078228e-05,
4495
+ "loss": 1.2385,
4496
+ "step": 748
4497
+ },
4498
+ {
4499
+ "epoch": 0.95,
4500
+ "learning_rate": 1.1095176010430248e-05,
4501
+ "loss": 1.2599,
4502
+ "step": 749
4503
+ },
4504
+ {
4505
+ "epoch": 0.95,
4506
+ "learning_rate": 1.108213820078227e-05,
4507
+ "loss": 1.2517,
4508
+ "step": 750
4509
+ },
4510
+ {
4511
+ "epoch": 0.95,
4512
+ "learning_rate": 1.1069100391134291e-05,
4513
+ "loss": 1.255,
4514
+ "step": 751
4515
+ },
4516
+ {
4517
+ "epoch": 0.95,
4518
+ "learning_rate": 1.1056062581486311e-05,
4519
+ "loss": 1.2482,
4520
+ "step": 752
4521
+ },
4522
+ {
4523
+ "epoch": 0.95,
4524
+ "learning_rate": 1.104302477183833e-05,
4525
+ "loss": 1.2484,
4526
+ "step": 753
4527
+ },
4528
+ {
4529
+ "epoch": 0.95,
4530
+ "learning_rate": 1.1029986962190354e-05,
4531
+ "loss": 1.2538,
4532
+ "step": 754
4533
+ },
4534
+ {
4535
+ "epoch": 0.95,
4536
+ "learning_rate": 1.1016949152542374e-05,
4537
+ "loss": 1.244,
4538
+ "step": 755
4539
+ },
4540
+ {
4541
+ "epoch": 0.96,
4542
+ "learning_rate": 1.1003911342894395e-05,
4543
+ "loss": 1.2635,
4544
+ "step": 756
4545
+ },
4546
+ {
4547
+ "epoch": 0.96,
4548
+ "learning_rate": 1.0990873533246415e-05,
4549
+ "loss": 1.2307,
4550
+ "step": 757
4551
+ },
4552
+ {
4553
+ "epoch": 0.96,
4554
+ "learning_rate": 1.0977835723598435e-05,
4555
+ "loss": 1.2157,
4556
+ "step": 758
4557
+ },
4558
+ {
4559
+ "epoch": 0.96,
4560
+ "learning_rate": 1.0964797913950458e-05,
4561
+ "loss": 1.2393,
4562
+ "step": 759
4563
+ },
4564
+ {
4565
+ "epoch": 0.96,
4566
+ "learning_rate": 1.0951760104302478e-05,
4567
+ "loss": 1.2542,
4568
+ "step": 760
4569
+ },
4570
+ {
4571
+ "epoch": 0.96,
4572
+ "learning_rate": 1.0938722294654498e-05,
4573
+ "loss": 1.2732,
4574
+ "step": 761
4575
+ },
4576
+ {
4577
+ "epoch": 0.96,
4578
+ "learning_rate": 1.092568448500652e-05,
4579
+ "loss": 1.2625,
4580
+ "step": 762
4581
+ },
4582
+ {
4583
+ "epoch": 0.96,
4584
+ "learning_rate": 1.0912646675358541e-05,
4585
+ "loss": 1.2423,
4586
+ "step": 763
4587
+ },
4588
+ {
4589
+ "epoch": 0.97,
4590
+ "learning_rate": 1.0899608865710562e-05,
4591
+ "loss": 1.2318,
4592
+ "step": 764
4593
+ },
4594
+ {
4595
+ "epoch": 0.97,
4596
+ "learning_rate": 1.0886571056062582e-05,
4597
+ "loss": 1.2404,
4598
+ "step": 765
4599
+ },
4600
+ {
4601
+ "epoch": 0.97,
4602
+ "learning_rate": 1.0873533246414602e-05,
4603
+ "loss": 1.2513,
4604
+ "step": 766
4605
+ },
4606
+ {
4607
+ "epoch": 0.97,
4608
+ "learning_rate": 1.0860495436766625e-05,
4609
+ "loss": 1.2864,
4610
+ "step": 767
4611
+ },
4612
+ {
4613
+ "epoch": 0.97,
4614
+ "learning_rate": 1.0847457627118645e-05,
4615
+ "loss": 1.2753,
4616
+ "step": 768
4617
+ },
4618
+ {
4619
+ "epoch": 0.97,
4620
+ "learning_rate": 1.0834419817470665e-05,
4621
+ "loss": 1.2426,
4622
+ "step": 769
4623
+ },
4624
+ {
4625
+ "epoch": 0.97,
4626
+ "learning_rate": 1.0821382007822687e-05,
4627
+ "loss": 1.2604,
4628
+ "step": 770
4629
+ },
4630
+ {
4631
+ "epoch": 0.97,
4632
+ "learning_rate": 1.0808344198174706e-05,
4633
+ "loss": 1.2538,
4634
+ "step": 771
4635
+ },
4636
+ {
4637
+ "epoch": 0.98,
4638
+ "learning_rate": 1.0795306388526728e-05,
4639
+ "loss": 1.2528,
4640
+ "step": 772
4641
+ },
4642
+ {
4643
+ "epoch": 0.98,
4644
+ "learning_rate": 1.078226857887875e-05,
4645
+ "loss": 1.2666,
4646
+ "step": 773
4647
+ },
4648
+ {
4649
+ "epoch": 0.98,
4650
+ "learning_rate": 1.076923076923077e-05,
4651
+ "loss": 1.2455,
4652
+ "step": 774
4653
+ },
4654
+ {
4655
+ "epoch": 0.98,
4656
+ "learning_rate": 1.0756192959582793e-05,
4657
+ "loss": 1.2563,
4658
+ "step": 775
4659
+ },
4660
+ {
4661
+ "epoch": 0.98,
4662
+ "learning_rate": 1.0743155149934812e-05,
4663
+ "loss": 1.2577,
4664
+ "step": 776
4665
+ },
4666
+ {
4667
+ "epoch": 0.98,
4668
+ "learning_rate": 1.0730117340286832e-05,
4669
+ "loss": 1.2611,
4670
+ "step": 777
4671
+ },
4672
+ {
4673
+ "epoch": 0.98,
4674
+ "learning_rate": 1.0717079530638854e-05,
4675
+ "loss": 1.2731,
4676
+ "step": 778
4677
+ },
4678
+ {
4679
+ "epoch": 0.98,
4680
+ "learning_rate": 1.0704041720990874e-05,
4681
+ "loss": 1.2245,
4682
+ "step": 779
4683
+ },
4684
+ {
4685
+ "epoch": 0.99,
4686
+ "learning_rate": 1.0691003911342893e-05,
4687
+ "loss": 1.2378,
4688
+ "step": 780
4689
+ },
4690
+ {
4691
+ "epoch": 0.99,
4692
+ "learning_rate": 1.0677966101694917e-05,
4693
+ "loss": 1.2627,
4694
+ "step": 781
4695
+ },
4696
+ {
4697
+ "epoch": 0.99,
4698
+ "learning_rate": 1.0664928292046937e-05,
4699
+ "loss": 1.2847,
4700
+ "step": 782
4701
+ },
4702
+ {
4703
+ "epoch": 0.99,
4704
+ "learning_rate": 1.0651890482398958e-05,
4705
+ "loss": 1.2548,
4706
+ "step": 783
4707
+ },
4708
+ {
4709
+ "epoch": 0.99,
4710
+ "learning_rate": 1.063885267275098e-05,
4711
+ "loss": 1.2675,
4712
+ "step": 784
4713
+ },
4714
+ {
4715
+ "epoch": 0.99,
4716
+ "learning_rate": 1.0625814863103e-05,
4717
+ "loss": 1.2573,
4718
+ "step": 785
4719
+ },
4720
+ {
4721
+ "epoch": 0.99,
4722
+ "learning_rate": 1.0612777053455021e-05,
4723
+ "loss": 1.2499,
4724
+ "step": 786
4725
+ },
4726
+ {
4727
+ "epoch": 0.99,
4728
+ "learning_rate": 1.059973924380704e-05,
4729
+ "loss": 1.2209,
4730
+ "step": 787
4731
+ },
4732
+ {
4733
+ "epoch": 1.0,
4734
+ "learning_rate": 1.058670143415906e-05,
4735
+ "loss": 1.2124,
4736
+ "step": 788
4737
+ },
4738
+ {
4739
+ "epoch": 1.0,
4740
+ "learning_rate": 1.0573663624511084e-05,
4741
+ "loss": 1.2803,
4742
+ "step": 789
4743
+ },
4744
+ {
4745
+ "epoch": 1.0,
4746
+ "learning_rate": 1.0560625814863104e-05,
4747
+ "loss": 1.2343,
4748
+ "step": 790
4749
+ },
4750
+ {
4751
+ "epoch": 1.0,
4752
+ "learning_rate": 1.0547588005215125e-05,
4753
+ "loss": 1.3397,
4754
+ "step": 791
4755
+ }
4756
+ ],
4757
+ "max_steps": 1582,
4758
+ "num_train_epochs": 2,
4759
+ "total_flos": 5423674366623744.0,
4760
+ "trial_name": null,
4761
+ "trial_params": null
4762
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a6ea68292e4928eb9c5b2cab80643242a3348a320d9c9e6ba5d69922500ed91d
3
+ size 5819
zero_to_fp32.py ADDED
@@ -0,0 +1,578 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python
2
+
3
+ # Copyright (c) Microsoft Corporation.
4
+ # SPDX-License-Identifier: Apache-2.0
5
+
6
+ # DeepSpeed Team
7
+
8
+ # This script extracts fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints. It gets
9
+ # copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
10
+ # the future. Once extracted, the weights don't require DeepSpeed and can be used in any
11
+ # application.
12
+ #
13
+ # example: python zero_to_fp32.py . pytorch_model.bin
14
+
15
+ import argparse
16
+ import torch
17
+ import glob
18
+ import math
19
+ import os
20
+ import re
21
+ from collections import OrderedDict
22
+ from dataclasses import dataclass
23
+
24
+ # while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
25
+ # DeepSpeed data structures it has to be available in the current python environment.
26
+ from deepspeed.utils import logger
27
+ from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
28
+ FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
29
+ FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
30
+
31
+
32
+ @dataclass
33
+ class zero_model_state:
34
+ buffers: dict()
35
+ param_shapes: dict()
36
+ shared_params: list
37
+ ds_version: int
38
+ frozen_param_shapes: dict()
39
+ frozen_param_fragments: dict()
40
+
41
+
42
+ debug = 0
43
+
44
+ # load to cpu
45
+ device = torch.device('cpu')
46
+
47
+
48
+ def atoi(text):
49
+ return int(text) if text.isdigit() else text
50
+
51
+
52
+ def natural_keys(text):
53
+ '''
54
+ alist.sort(key=natural_keys) sorts in human order
55
+ http://nedbatchelder.com/blog/200712/human_sorting.html
56
+ (See Toothy's implementation in the comments)
57
+ '''
58
+ return [atoi(c) for c in re.split(r'(\d+)', text)]
59
+
60
+
61
+ def get_model_state_file(checkpoint_dir, zero_stage):
62
+ if not os.path.isdir(checkpoint_dir):
63
+ raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
64
+
65
+ # there should be only one file
66
+ if zero_stage == 2:
67
+ file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
68
+ elif zero_stage == 3:
69
+ file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
70
+
71
+ if not os.path.exists(file):
72
+ raise FileNotFoundError(f"can't find model states file at '{file}'")
73
+
74
+ return file
75
+
76
+
77
+ def get_checkpoint_files(checkpoint_dir, glob_pattern):
78
+ # XXX: need to test that this simple glob rule works for multi-node setup too
79
+ ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
80
+
81
+ if len(ckpt_files) == 0:
82
+ raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
83
+
84
+ return ckpt_files
85
+
86
+
87
+ def get_optim_files(checkpoint_dir):
88
+ return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
89
+
90
+
91
+ def get_model_state_files(checkpoint_dir):
92
+ return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
93
+
94
+
95
+ def parse_model_states(files):
96
+ zero_model_states = []
97
+ for file in files:
98
+ state_dict = torch.load(file, map_location=device)
99
+
100
+ if BUFFER_NAMES not in state_dict:
101
+ raise ValueError(f"{file} is not a model state checkpoint")
102
+ buffer_names = state_dict[BUFFER_NAMES]
103
+ if debug:
104
+ print("Found buffers:", buffer_names)
105
+
106
+ # recover just the buffers while restoring them to fp32 if they were saved in fp16
107
+ buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
108
+ param_shapes = state_dict[PARAM_SHAPES]
109
+
110
+ # collect parameters that are included in param_shapes
111
+ param_names = []
112
+ for s in param_shapes:
113
+ for name in s.keys():
114
+ param_names.append(name)
115
+
116
+ # update with frozen parameters
117
+ frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
118
+ if frozen_param_shapes is not None:
119
+ if debug:
120
+ print(f"Found frozen_param_shapes: {frozen_param_shapes}")
121
+ param_names += list(frozen_param_shapes.keys())
122
+
123
+ # handle shared params
124
+ shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
125
+
126
+ ds_version = state_dict.get(DS_VERSION, None)
127
+
128
+ frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
129
+
130
+ z_model_state = zero_model_state(buffers=buffers,
131
+ param_shapes=param_shapes,
132
+ shared_params=shared_params,
133
+ ds_version=ds_version,
134
+ frozen_param_shapes=frozen_param_shapes,
135
+ frozen_param_fragments=frozen_param_fragments)
136
+ zero_model_states.append(z_model_state)
137
+
138
+ return zero_model_states
139
+
140
+
141
+ def parse_optim_states(files, ds_checkpoint_dir):
142
+
143
+ total_files = len(files)
144
+ state_dicts = []
145
+ for f in files:
146
+ state_dicts.append(torch.load(f, map_location=device))
147
+
148
+ if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
149
+ raise ValueError(f"{files[0]} is not a zero checkpoint")
150
+ zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
151
+ world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
152
+
153
+ # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
154
+ # parameters can be different from data parallelism for non-expert parameters. So we can just
155
+ # use the max of the partition_count to get the dp world_size.
156
+
157
+ if type(world_size) is list:
158
+ world_size = max(world_size)
159
+
160
+ if world_size != total_files:
161
+ raise ValueError(
162
+ f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
163
+ "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
164
+ )
165
+
166
+ # the groups are named differently in each stage
167
+ if zero_stage == 2:
168
+ fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
169
+ elif zero_stage == 3:
170
+ fp32_groups_key = FP32_FLAT_GROUPS
171
+ else:
172
+ raise ValueError(f"unknown zero stage {zero_stage}")
173
+
174
+ if zero_stage == 2:
175
+ fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
176
+ elif zero_stage == 3:
177
+ # if there is more than one param group, there will be multiple flattened tensors - one
178
+ # flattened tensor per group - for simplicity merge them into a single tensor
179
+ #
180
+ # XXX: could make the script more memory efficient for when there are multiple groups - it
181
+ # will require matching the sub-lists of param_shapes for each param group flattened tensor
182
+
183
+ fp32_flat_groups = [
184
+ torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
185
+ ]
186
+
187
+ return zero_stage, world_size, fp32_flat_groups
188
+
189
+
190
+ def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
191
+ """
192
+ Returns fp32 state_dict reconstructed from ds checkpoint
193
+
194
+ Args:
195
+ - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
196
+
197
+ """
198
+ print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
199
+
200
+ optim_files = get_optim_files(ds_checkpoint_dir)
201
+ zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
202
+ print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
203
+
204
+ model_files = get_model_state_files(ds_checkpoint_dir)
205
+
206
+ zero_model_states = parse_model_states(model_files)
207
+ print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
208
+
209
+ if zero_stage == 2:
210
+ return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
211
+ elif zero_stage == 3:
212
+ return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
213
+
214
+
215
+ def _zero2_merge_frozen_params(state_dict, zero_model_states):
216
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
217
+ return
218
+
219
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
220
+ frozen_param_fragments = zero_model_states[0].frozen_param_fragments
221
+
222
+ if debug:
223
+ num_elem = sum(s.numel() for s in frozen_param_shapes.values())
224
+ print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
225
+
226
+ wanted_params = len(frozen_param_shapes)
227
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
228
+ avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
229
+ print(f'Frozen params: Have {avail_numel} numels to process.')
230
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
231
+
232
+ total_params = 0
233
+ total_numel = 0
234
+ for name, shape in frozen_param_shapes.items():
235
+ total_params += 1
236
+ unpartitioned_numel = shape.numel()
237
+ total_numel += unpartitioned_numel
238
+
239
+ state_dict[name] = frozen_param_fragments[name]
240
+
241
+ if debug:
242
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
243
+
244
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
245
+
246
+
247
+ def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
248
+ param_shapes = zero_model_states[0].param_shapes
249
+
250
+ # Reconstruction protocol:
251
+ #
252
+ # XXX: document this
253
+
254
+ if debug:
255
+ for i in range(world_size):
256
+ for j in range(len(fp32_flat_groups[0])):
257
+ print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
258
+
259
+ # XXX: memory usage doubles here (zero2)
260
+ num_param_groups = len(fp32_flat_groups[0])
261
+ merged_single_partition_of_fp32_groups = []
262
+ for i in range(num_param_groups):
263
+ merged_partitions = [sd[i] for sd in fp32_flat_groups]
264
+ full_single_fp32_vector = torch.cat(merged_partitions, 0)
265
+ merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
266
+ avail_numel = sum(
267
+ [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
268
+
269
+ if debug:
270
+ wanted_params = sum([len(shapes) for shapes in param_shapes])
271
+ wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
272
+ # not asserting if there is a mismatch due to possible padding
273
+ print(f"Have {avail_numel} numels to process.")
274
+ print(f"Need {wanted_numel} numels in {wanted_params} params.")
275
+
276
+ # params
277
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
278
+ # out-of-core computing solution
279
+ total_numel = 0
280
+ total_params = 0
281
+ for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
282
+ offset = 0
283
+ avail_numel = full_single_fp32_vector.numel()
284
+ for name, shape in shapes.items():
285
+
286
+ unpartitioned_numel = shape.numel()
287
+ total_numel += unpartitioned_numel
288
+ total_params += 1
289
+
290
+ if debug:
291
+ print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
292
+ state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
293
+ offset += unpartitioned_numel
294
+
295
+ # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
296
+ # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
297
+ # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
298
+ # live optimizer object, so we are checking that the numbers are within the right range
299
+ align_to = 2 * world_size
300
+
301
+ def zero2_align(x):
302
+ return align_to * math.ceil(x / align_to)
303
+
304
+ if debug:
305
+ print(f"original offset={offset}, avail_numel={avail_numel}")
306
+
307
+ offset = zero2_align(offset)
308
+ avail_numel = zero2_align(avail_numel)
309
+
310
+ if debug:
311
+ print(f"aligned offset={offset}, avail_numel={avail_numel}")
312
+
313
+ # Sanity check
314
+ if offset != avail_numel:
315
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
316
+
317
+ print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
318
+
319
+
320
+ def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
321
+ state_dict = OrderedDict()
322
+
323
+ # buffers
324
+ buffers = zero_model_states[0].buffers
325
+ state_dict.update(buffers)
326
+ if debug:
327
+ print(f"added {len(buffers)} buffers")
328
+
329
+ _zero2_merge_frozen_params(state_dict, zero_model_states)
330
+
331
+ _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
332
+
333
+ # recover shared parameters
334
+ for pair in zero_model_states[0].shared_params:
335
+ if pair[1] in state_dict:
336
+ state_dict[pair[0]] = state_dict[pair[1]]
337
+
338
+ return state_dict
339
+
340
+
341
+ def zero3_partitioned_param_info(unpartitioned_numel, world_size):
342
+ remainder = unpartitioned_numel % world_size
343
+ padding_numel = (world_size - remainder) if remainder else 0
344
+ partitioned_numel = math.ceil(unpartitioned_numel / world_size)
345
+ return partitioned_numel, padding_numel
346
+
347
+
348
+ def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
349
+ if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
350
+ return
351
+
352
+ if debug:
353
+ for i in range(world_size):
354
+ num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
355
+ print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
356
+
357
+ frozen_param_shapes = zero_model_states[0].frozen_param_shapes
358
+ wanted_params = len(frozen_param_shapes)
359
+ wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
360
+ avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
361
+ print(f'Frozen params: Have {avail_numel} numels to process.')
362
+ print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
363
+
364
+ total_params = 0
365
+ total_numel = 0
366
+ for name, shape in zero_model_states[0].frozen_param_shapes.items():
367
+ total_params += 1
368
+ unpartitioned_numel = shape.numel()
369
+ total_numel += unpartitioned_numel
370
+
371
+ param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
372
+ state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
373
+
374
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
375
+
376
+ if debug:
377
+ print(
378
+ f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
379
+ )
380
+
381
+ print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
382
+
383
+
384
+ def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
385
+ param_shapes = zero_model_states[0].param_shapes
386
+ avail_numel = fp32_flat_groups[0].numel() * world_size
387
+ # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
388
+ # param, re-consolidating each param, while dealing with padding if any
389
+
390
+ # merge list of dicts, preserving order
391
+ param_shapes = {k: v for d in param_shapes for k, v in d.items()}
392
+
393
+ if debug:
394
+ for i in range(world_size):
395
+ print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
396
+
397
+ wanted_params = len(param_shapes)
398
+ wanted_numel = sum(shape.numel() for shape in param_shapes.values())
399
+ # not asserting if there is a mismatch due to possible padding
400
+ avail_numel = fp32_flat_groups[0].numel() * world_size
401
+ print(f"Trainable params: Have {avail_numel} numels to process.")
402
+ print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
403
+
404
+ # params
405
+ # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
406
+ # out-of-core computing solution
407
+ offset = 0
408
+ total_numel = 0
409
+ total_params = 0
410
+ for name, shape in param_shapes.items():
411
+
412
+ unpartitioned_numel = shape.numel()
413
+ total_numel += unpartitioned_numel
414
+ total_params += 1
415
+
416
+ partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
417
+
418
+ if debug:
419
+ print(
420
+ f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
421
+ )
422
+
423
+ # XXX: memory usage doubles here
424
+ state_dict[name] = torch.cat(
425
+ tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
426
+ 0).narrow(0, 0, unpartitioned_numel).view(shape)
427
+ offset += partitioned_numel
428
+
429
+ offset *= world_size
430
+
431
+ # Sanity check
432
+ if offset != avail_numel:
433
+ raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
434
+
435
+ print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
436
+
437
+
438
+ def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
439
+ state_dict = OrderedDict()
440
+
441
+ # buffers
442
+ buffers = zero_model_states[0].buffers
443
+ state_dict.update(buffers)
444
+ if debug:
445
+ print(f"added {len(buffers)} buffers")
446
+
447
+ _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
448
+
449
+ _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
450
+
451
+ # recover shared parameters
452
+ for pair in zero_model_states[0].shared_params:
453
+ if pair[1] in state_dict:
454
+ state_dict[pair[0]] = state_dict[pair[1]]
455
+
456
+ return state_dict
457
+
458
+
459
+ def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
460
+ """
461
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
462
+ ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
463
+ via a model hub.
464
+
465
+ Args:
466
+ - ``checkpoint_dir``: path to the desired checkpoint folder
467
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
468
+
469
+ Returns:
470
+ - pytorch ``state_dict``
471
+
472
+ Note: this approach may not work if your application doesn't have sufficient free CPU memory and
473
+ you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
474
+ the checkpoint.
475
+
476
+ A typical usage might be ::
477
+
478
+ from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
479
+ # do the training and checkpoint saving
480
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
481
+ model = model.cpu() # move to cpu
482
+ model.load_state_dict(state_dict)
483
+ # submit to model hub or save the model to share with others
484
+
485
+ In this example the ``model`` will no longer be usable in the deepspeed context of the same
486
+ application. i.e. you will need to re-initialize the deepspeed engine, since
487
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
488
+
489
+ If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
490
+
491
+ """
492
+ if tag is None:
493
+ latest_path = os.path.join(checkpoint_dir, 'latest')
494
+ if os.path.isfile(latest_path):
495
+ with open(latest_path, 'r') as fd:
496
+ tag = fd.read().strip()
497
+ else:
498
+ raise ValueError(f"Unable to find 'latest' file at {latest_path}")
499
+
500
+ ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
501
+
502
+ if not os.path.isdir(ds_checkpoint_dir):
503
+ raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
504
+
505
+ return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
506
+
507
+
508
+ def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
509
+ """
510
+ Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
511
+ loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
512
+
513
+ Args:
514
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
515
+ - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
516
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
517
+ """
518
+
519
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
520
+ print(f"Saving fp32 state dict to {output_file}")
521
+ torch.save(state_dict, output_file)
522
+
523
+
524
+ def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
525
+ """
526
+ 1. Put the provided model to cpu
527
+ 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
528
+ 3. Load it into the provided model
529
+
530
+ Args:
531
+ - ``model``: the model object to update
532
+ - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
533
+ - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
534
+
535
+ Returns:
536
+ - ``model`: modified model
537
+
538
+ Make sure you have plenty of CPU memory available before you call this function. If you don't
539
+ have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
540
+ conveniently placed for you in the checkpoint folder.
541
+
542
+ A typical usage might be ::
543
+
544
+ from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
545
+ model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
546
+ # submit to model hub or save the model to share with others
547
+
548
+ Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
549
+ of the same application. i.e. you will need to re-initialize the deepspeed engine, since
550
+ ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
551
+
552
+ """
553
+ logger.info(f"Extracting fp32 weights")
554
+ state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
555
+
556
+ logger.info(f"Overwriting model with fp32 weights")
557
+ model = model.cpu()
558
+ model.load_state_dict(state_dict, strict=False)
559
+
560
+ return model
561
+
562
+
563
+ if __name__ == "__main__":
564
+
565
+ parser = argparse.ArgumentParser()
566
+ parser.add_argument("checkpoint_dir",
567
+ type=str,
568
+ help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
569
+ parser.add_argument(
570
+ "output_file",
571
+ type=str,
572
+ help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
573
+ parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
574
+ args = parser.parse_args()
575
+
576
+ debug = args.debug
577
+
578
+ convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file)