Commit
·
00429e4
1
Parent(s):
46b03b7
Upload 11 files
Browse files- README (7).md +9 -0
- adapter_config.json +26 -0
- adapter_model.bin +3 -0
- added_tokens.json +7 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer_config.json +54 -0
- trainer_state.json +729 -0
- training_args.bin +3 -0
README (7).md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.6.0.dev0
|
adapter_config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "facebook/esm2_t6_8M_UR50D",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"fan_in_fan_out": false,
|
| 7 |
+
"inference_mode": true,
|
| 8 |
+
"init_lora_weights": true,
|
| 9 |
+
"layers_pattern": null,
|
| 10 |
+
"layers_to_transform": null,
|
| 11 |
+
"lora_alpha": 1,
|
| 12 |
+
"lora_dropout": 0.5,
|
| 13 |
+
"modules_to_save": [
|
| 14 |
+
"classifier"
|
| 15 |
+
],
|
| 16 |
+
"peft_type": "LORA",
|
| 17 |
+
"r": 2,
|
| 18 |
+
"rank_pattern": {},
|
| 19 |
+
"revision": null,
|
| 20 |
+
"target_modules": [
|
| 21 |
+
"query",
|
| 22 |
+
"key",
|
| 23 |
+
"value"
|
| 24 |
+
],
|
| 25 |
+
"task_type": "TOKEN_CLS"
|
| 26 |
+
}
|
adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ee71daeaa76507530da9f065b5186d37be6cb34d73592fe14366719fda5523d
|
| 3 |
+
size 108323
|
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<cls>": 0,
|
| 3 |
+
"<eos>": 2,
|
| 4 |
+
"<mask>": 32,
|
| 5 |
+
"<pad>": 1,
|
| 6 |
+
"<unk>": 3
|
| 7 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7542477d670743dbca32f1157cf900a273fc84e169e30a30b3700d1efc70153
|
| 3 |
+
size 221317
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8d5c9842203c839cf6fa6f2cdeee01a013e470cf6078f27386ceefc419e88cd
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:163ca3797b8ca4d204606b429ad731470bf3aa8b029aa6c460391f627d7a35f7
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "<cls>",
|
| 3 |
+
"eos_token": "<eos>",
|
| 4 |
+
"mask_token": "<mask>",
|
| 5 |
+
"pad_token": "<pad>",
|
| 6 |
+
"unk_token": "<unk>"
|
| 7 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<cls>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "<eos>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"32": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"additional_special_tokens": [],
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<cls>",
|
| 47 |
+
"eos_token": "<eos>",
|
| 48 |
+
"mask_token": "<mask>",
|
| 49 |
+
"model_max_length": 1024,
|
| 50 |
+
"pad_token": "<pad>",
|
| 51 |
+
"tokenizer_class": "EsmTokenizer",
|
| 52 |
+
"tokenizer_file": null,
|
| 53 |
+
"unk_token": "<unk>"
|
| 54 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,729 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.0837364188034316,
|
| 3 |
+
"best_model_checkpoint": "esm2_t6_8M_lora_ptm_sites_2023-10-09_03-08-23/checkpoint-23309",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 23309,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 0.00037014933433791045,
|
| 14 |
+
"loss": 0.4563,
|
| 15 |
+
"step": 200
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.02,
|
| 19 |
+
"learning_rate": 0.0003701270704479516,
|
| 20 |
+
"loss": 0.2291,
|
| 21 |
+
"step": 400
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.03,
|
| 25 |
+
"learning_rate": 0.00037008979195075377,
|
| 26 |
+
"loss": 0.1662,
|
| 27 |
+
"step": 600
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 0.0003700375763811688,
|
| 32 |
+
"loss": 0.1448,
|
| 33 |
+
"step": 800
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.04,
|
| 37 |
+
"learning_rate": 0.00036997042795487093,
|
| 38 |
+
"loss": 0.1194,
|
| 39 |
+
"step": 1000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.05,
|
| 43 |
+
"learning_rate": 0.0003698887995939965,
|
| 44 |
+
"loss": 0.1104,
|
| 45 |
+
"step": 1200
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.06,
|
| 49 |
+
"learning_rate": 0.00036979187750890046,
|
| 50 |
+
"loss": 0.1126,
|
| 51 |
+
"step": 1400
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.07,
|
| 55 |
+
"learning_rate": 0.00036968004240382616,
|
| 56 |
+
"loss": 0.0939,
|
| 57 |
+
"step": 1600
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.08,
|
| 61 |
+
"learning_rate": 0.0003695533033078884,
|
| 62 |
+
"loss": 0.0985,
|
| 63 |
+
"step": 1800
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.09,
|
| 67 |
+
"learning_rate": 0.00036941167045348997,
|
| 68 |
+
"loss": 0.088,
|
| 69 |
+
"step": 2000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.09,
|
| 73 |
+
"learning_rate": 0.0003692551552754955,
|
| 74 |
+
"loss": 0.0839,
|
| 75 |
+
"step": 2200
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.1,
|
| 79 |
+
"learning_rate": 0.0003690837704103084,
|
| 80 |
+
"loss": 0.0898,
|
| 81 |
+
"step": 2400
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.11,
|
| 85 |
+
"learning_rate": 0.00036889752969485046,
|
| 86 |
+
"loss": 0.0825,
|
| 87 |
+
"step": 2600
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.12,
|
| 91 |
+
"learning_rate": 0.00036869644816544503,
|
| 92 |
+
"loss": 0.0825,
|
| 93 |
+
"step": 2800
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.13,
|
| 97 |
+
"learning_rate": 0.00036848054205660264,
|
| 98 |
+
"loss": 0.0798,
|
| 99 |
+
"step": 3000
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.14,
|
| 103 |
+
"learning_rate": 0.00036824982879971063,
|
| 104 |
+
"loss": 0.0692,
|
| 105 |
+
"step": 3200
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.15,
|
| 109 |
+
"learning_rate": 0.0003680055912846709,
|
| 110 |
+
"loss": 0.0622,
|
| 111 |
+
"step": 3400
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.15,
|
| 115 |
+
"learning_rate": 0.00036774539459843937,
|
| 116 |
+
"loss": 0.0686,
|
| 117 |
+
"step": 3600
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.16,
|
| 121 |
+
"learning_rate": 0.00036747045011699506,
|
| 122 |
+
"loss": 0.0755,
|
| 123 |
+
"step": 3800
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.17,
|
| 127 |
+
"learning_rate": 0.00036718078003824503,
|
| 128 |
+
"loss": 0.0675,
|
| 129 |
+
"step": 4000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.18,
|
| 133 |
+
"learning_rate": 0.0003668764077489815,
|
| 134 |
+
"loss": 0.0695,
|
| 135 |
+
"step": 4200
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.19,
|
| 139 |
+
"learning_rate": 0.00036655735782299405,
|
| 140 |
+
"loss": 0.0738,
|
| 141 |
+
"step": 4400
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.2,
|
| 145 |
+
"learning_rate": 0.00036622365601908513,
|
| 146 |
+
"loss": 0.0601,
|
| 147 |
+
"step": 4600
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.21,
|
| 151 |
+
"learning_rate": 0.0003658753292789907,
|
| 152 |
+
"loss": 0.0707,
|
| 153 |
+
"step": 4800
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.21,
|
| 157 |
+
"learning_rate": 0.000365512405725205,
|
| 158 |
+
"loss": 0.0642,
|
| 159 |
+
"step": 5000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.22,
|
| 163 |
+
"learning_rate": 0.0003651349146587103,
|
| 164 |
+
"loss": 0.0529,
|
| 165 |
+
"step": 5200
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.23,
|
| 169 |
+
"learning_rate": 0.00036474288655661055,
|
| 170 |
+
"loss": 0.0527,
|
| 171 |
+
"step": 5400
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.24,
|
| 175 |
+
"learning_rate": 0.0003643384217654436,
|
| 176 |
+
"loss": 0.0625,
|
| 177 |
+
"step": 5600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.25,
|
| 181 |
+
"learning_rate": 0.00036391748799477383,
|
| 182 |
+
"loss": 0.0546,
|
| 183 |
+
"step": 5800
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.26,
|
| 187 |
+
"learning_rate": 0.00036348211547861136,
|
| 188 |
+
"loss": 0.069,
|
| 189 |
+
"step": 6000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.27,
|
| 193 |
+
"learning_rate": 0.00036303233936717483,
|
| 194 |
+
"loss": 0.0597,
|
| 195 |
+
"step": 6200
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.27,
|
| 199 |
+
"learning_rate": 0.00036256819597357105,
|
| 200 |
+
"loss": 0.0598,
|
| 201 |
+
"step": 6400
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.28,
|
| 205 |
+
"learning_rate": 0.00036208972277086324,
|
| 206 |
+
"loss": 0.0499,
|
| 207 |
+
"step": 6600
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.29,
|
| 211 |
+
"learning_rate": 0.00036159695838904537,
|
| 212 |
+
"loss": 0.0526,
|
| 213 |
+
"step": 6800
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.3,
|
| 217 |
+
"learning_rate": 0.0003610899426119236,
|
| 218 |
+
"loss": 0.0708,
|
| 219 |
+
"step": 7000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.31,
|
| 223 |
+
"learning_rate": 0.00036056871637390423,
|
| 224 |
+
"loss": 0.0506,
|
| 225 |
+
"step": 7200
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.32,
|
| 229 |
+
"learning_rate": 0.0003600333217566887,
|
| 230 |
+
"loss": 0.0654,
|
| 231 |
+
"step": 7400
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.33,
|
| 235 |
+
"learning_rate": 0.00035948380198587616,
|
| 236 |
+
"loss": 0.0529,
|
| 237 |
+
"step": 7600
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.33,
|
| 241 |
+
"learning_rate": 0.0003589230543814262,
|
| 242 |
+
"loss": 0.0698,
|
| 243 |
+
"step": 7800
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.34,
|
| 247 |
+
"learning_rate": 0.00035834548859962624,
|
| 248 |
+
"loss": 0.0499,
|
| 249 |
+
"step": 8000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.35,
|
| 253 |
+
"learning_rate": 0.0003577539339330621,
|
| 254 |
+
"loss": 0.0475,
|
| 255 |
+
"step": 8200
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.36,
|
| 259 |
+
"learning_rate": 0.0003571484381414685,
|
| 260 |
+
"loss": 0.0573,
|
| 261 |
+
"step": 8400
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.37,
|
| 265 |
+
"learning_rate": 0.00035652905011013025,
|
| 266 |
+
"loss": 0.0445,
|
| 267 |
+
"step": 8600
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.38,
|
| 271 |
+
"learning_rate": 0.00035589581984593567,
|
| 272 |
+
"loss": 0.0572,
|
| 273 |
+
"step": 8800
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.39,
|
| 277 |
+
"learning_rate": 0.00035524879847333905,
|
| 278 |
+
"loss": 0.0552,
|
| 279 |
+
"step": 9000
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.39,
|
| 283 |
+
"learning_rate": 0.00035458803823023286,
|
| 284 |
+
"loss": 0.0368,
|
| 285 |
+
"step": 9200
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.4,
|
| 289 |
+
"learning_rate": 0.0003539135924637307,
|
| 290 |
+
"loss": 0.0464,
|
| 291 |
+
"step": 9400
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.41,
|
| 295 |
+
"learning_rate": 0.00035322551562586,
|
| 296 |
+
"loss": 0.0478,
|
| 297 |
+
"step": 9600
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.42,
|
| 301 |
+
"learning_rate": 0.0003525238632691658,
|
| 302 |
+
"loss": 0.064,
|
| 303 |
+
"step": 9800
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.43,
|
| 307 |
+
"learning_rate": 0.0003518086920422257,
|
| 308 |
+
"loss": 0.0501,
|
| 309 |
+
"step": 10000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.44,
|
| 313 |
+
"learning_rate": 0.00035108373623456347,
|
| 314 |
+
"loss": 0.0554,
|
| 315 |
+
"step": 10200
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.45,
|
| 319 |
+
"learning_rate": 0.000350341768437432,
|
| 320 |
+
"loss": 0.0433,
|
| 321 |
+
"step": 10400
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.45,
|
| 325 |
+
"learning_rate": 0.00034958645794357904,
|
| 326 |
+
"loss": 0.0546,
|
| 327 |
+
"step": 10600
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.46,
|
| 331 |
+
"learning_rate": 0.0003488178657337234,
|
| 332 |
+
"loss": 0.0556,
|
| 333 |
+
"step": 10800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.47,
|
| 337 |
+
"learning_rate": 0.0003480360538608961,
|
| 338 |
+
"loss": 0.0597,
|
| 339 |
+
"step": 11000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.48,
|
| 343 |
+
"learning_rate": 0.0003472410854454305,
|
| 344 |
+
"loss": 0.0526,
|
| 345 |
+
"step": 11200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.49,
|
| 349 |
+
"learning_rate": 0.00034643302466986575,
|
| 350 |
+
"loss": 0.0559,
|
| 351 |
+
"step": 11400
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.5,
|
| 355 |
+
"learning_rate": 0.0003456119367737657,
|
| 356 |
+
"loss": 0.0514,
|
| 357 |
+
"step": 11600
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.51,
|
| 361 |
+
"learning_rate": 0.00034477788804845107,
|
| 362 |
+
"loss": 0.056,
|
| 363 |
+
"step": 11800
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.51,
|
| 367 |
+
"learning_rate": 0.00034393094583164754,
|
| 368 |
+
"loss": 0.0491,
|
| 369 |
+
"step": 12000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.52,
|
| 373 |
+
"learning_rate": 0.00034307117850204926,
|
| 374 |
+
"loss": 0.0386,
|
| 375 |
+
"step": 12200
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.53,
|
| 379 |
+
"learning_rate": 0.0003421986554737983,
|
| 380 |
+
"loss": 0.047,
|
| 381 |
+
"step": 12400
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.54,
|
| 385 |
+
"learning_rate": 0.0003413134471908803,
|
| 386 |
+
"loss": 0.0587,
|
| 387 |
+
"step": 12600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.55,
|
| 391 |
+
"learning_rate": 0.00034041562512143693,
|
| 392 |
+
"loss": 0.0442,
|
| 393 |
+
"step": 12800
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.56,
|
| 397 |
+
"learning_rate": 0.00033950526175199636,
|
| 398 |
+
"loss": 0.0517,
|
| 399 |
+
"step": 13000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.57,
|
| 403 |
+
"learning_rate": 0.0003385824305816204,
|
| 404 |
+
"loss": 0.0446,
|
| 405 |
+
"step": 13200
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.57,
|
| 409 |
+
"learning_rate": 0.0003376472061159709,
|
| 410 |
+
"loss": 0.0525,
|
| 411 |
+
"step": 13400
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.58,
|
| 415 |
+
"learning_rate": 0.00033670443208693313,
|
| 416 |
+
"loss": 0.0482,
|
| 417 |
+
"step": 13600
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.59,
|
| 421 |
+
"learning_rate": 0.00033574470955847287,
|
| 422 |
+
"loss": 0.0489,
|
| 423 |
+
"step": 13800
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.6,
|
| 427 |
+
"learning_rate": 0.0003347728228408742,
|
| 428 |
+
"loss": 0.0684,
|
| 429 |
+
"step": 14000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.61,
|
| 433 |
+
"learning_rate": 0.0003337888504003458,
|
| 434 |
+
"loss": 0.0506,
|
| 435 |
+
"step": 14200
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.62,
|
| 439 |
+
"learning_rate": 0.00033279287167884867,
|
| 440 |
+
"loss": 0.0563,
|
| 441 |
+
"step": 14400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.63,
|
| 445 |
+
"learning_rate": 0.00033178496708768257,
|
| 446 |
+
"loss": 0.0356,
|
| 447 |
+
"step": 14600
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.63,
|
| 451 |
+
"learning_rate": 0.0003307652180009936,
|
| 452 |
+
"loss": 0.0573,
|
| 453 |
+
"step": 14800
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.64,
|
| 457 |
+
"learning_rate": 0.0003297337067492049,
|
| 458 |
+
"loss": 0.0498,
|
| 459 |
+
"step": 15000
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.65,
|
| 463 |
+
"learning_rate": 0.00032869051661236907,
|
| 464 |
+
"loss": 0.0493,
|
| 465 |
+
"step": 15200
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.66,
|
| 469 |
+
"learning_rate": 0.00032763573181344475,
|
| 470 |
+
"loss": 0.0479,
|
| 471 |
+
"step": 15400
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.67,
|
| 475 |
+
"learning_rate": 0.0003265694375114969,
|
| 476 |
+
"loss": 0.0497,
|
| 477 |
+
"step": 15600
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.68,
|
| 481 |
+
"learning_rate": 0.0003254917197948212,
|
| 482 |
+
"loss": 0.0543,
|
| 483 |
+
"step": 15800
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.69,
|
| 487 |
+
"learning_rate": 0.0003244081389988359,
|
| 488 |
+
"loss": 0.0537,
|
| 489 |
+
"step": 16000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.7,
|
| 493 |
+
"learning_rate": 0.00032330789242185896,
|
| 494 |
+
"loss": 0.0743,
|
| 495 |
+
"step": 16200
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.7,
|
| 499 |
+
"learning_rate": 0.00032219648575413284,
|
| 500 |
+
"loss": 0.0346,
|
| 501 |
+
"step": 16400
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.71,
|
| 505 |
+
"learning_rate": 0.0003210740087261434,
|
| 506 |
+
"loss": 0.0454,
|
| 507 |
+
"step": 16600
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.72,
|
| 511 |
+
"learning_rate": 0.00031994055196215285,
|
| 512 |
+
"loss": 0.0563,
|
| 513 |
+
"step": 16800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.73,
|
| 517 |
+
"learning_rate": 0.0003187962069728827,
|
| 518 |
+
"loss": 0.0456,
|
| 519 |
+
"step": 17000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.74,
|
| 523 |
+
"learning_rate": 0.000317641066148126,
|
| 524 |
+
"loss": 0.0445,
|
| 525 |
+
"step": 17200
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.75,
|
| 529 |
+
"learning_rate": 0.0003164752227492878,
|
| 530 |
+
"loss": 0.0416,
|
| 531 |
+
"step": 17400
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.76,
|
| 535 |
+
"learning_rate": 0.00031529877090185585,
|
| 536 |
+
"loss": 0.0547,
|
| 537 |
+
"step": 17600
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.76,
|
| 541 |
+
"learning_rate": 0.00031411180558780117,
|
| 542 |
+
"loss": 0.0504,
|
| 543 |
+
"step": 17800
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.77,
|
| 547 |
+
"learning_rate": 0.0003129144226379097,
|
| 548 |
+
"loss": 0.0437,
|
| 549 |
+
"step": 18000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.78,
|
| 553 |
+
"learning_rate": 0.00031171278275606484,
|
| 554 |
+
"loss": 0.0479,
|
| 555 |
+
"step": 18200
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.79,
|
| 559 |
+
"learning_rate": 0.0003104949062567452,
|
| 560 |
+
"loss": 0.0512,
|
| 561 |
+
"step": 18400
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.8,
|
| 565 |
+
"learning_rate": 0.0003092669041354337,
|
| 566 |
+
"loss": 0.0477,
|
| 567 |
+
"step": 18600
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.81,
|
| 571 |
+
"learning_rate": 0.0003080288755360624,
|
| 572 |
+
"loss": 0.0431,
|
| 573 |
+
"step": 18800
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.82,
|
| 577 |
+
"learning_rate": 0.0003067809204120605,
|
| 578 |
+
"loss": 0.05,
|
| 579 |
+
"step": 19000
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.82,
|
| 583 |
+
"learning_rate": 0.00030552313951828493,
|
| 584 |
+
"loss": 0.0557,
|
| 585 |
+
"step": 19200
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.83,
|
| 589 |
+
"learning_rate": 0.0003042556344028853,
|
| 590 |
+
"loss": 0.0388,
|
| 591 |
+
"step": 19400
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.84,
|
| 595 |
+
"learning_rate": 0.00030297850739910594,
|
| 596 |
+
"loss": 0.0515,
|
| 597 |
+
"step": 19600
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.85,
|
| 601 |
+
"learning_rate": 0.00030169186161702326,
|
| 602 |
+
"loss": 0.0438,
|
| 603 |
+
"step": 19800
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.86,
|
| 607 |
+
"learning_rate": 0.0003003958009352217,
|
| 608 |
+
"loss": 0.0512,
|
| 609 |
+
"step": 20000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.87,
|
| 613 |
+
"learning_rate": 0.0002990904299924064,
|
| 614 |
+
"loss": 0.0499,
|
| 615 |
+
"step": 20200
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.88,
|
| 619 |
+
"learning_rate": 0.00029778244977988057,
|
| 620 |
+
"loss": 0.0505,
|
| 621 |
+
"step": 20400
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.88,
|
| 625 |
+
"learning_rate": 0.000296458820457796,
|
| 626 |
+
"loss": 0.0458,
|
| 627 |
+
"step": 20600
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.89,
|
| 631 |
+
"learning_rate": 0.0002951261987306019,
|
| 632 |
+
"loss": 0.0368,
|
| 633 |
+
"step": 20800
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.9,
|
| 637 |
+
"learning_rate": 0.0002937846921887947,
|
| 638 |
+
"loss": 0.0575,
|
| 639 |
+
"step": 21000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.91,
|
| 643 |
+
"learning_rate": 0.00029243440914019487,
|
| 644 |
+
"loss": 0.0444,
|
| 645 |
+
"step": 21200
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.92,
|
| 649 |
+
"learning_rate": 0.0002910754586012025,
|
| 650 |
+
"loss": 0.0467,
|
| 651 |
+
"step": 21400
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.93,
|
| 655 |
+
"learning_rate": 0.00028970795028799613,
|
| 656 |
+
"loss": 0.0383,
|
| 657 |
+
"step": 21600
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.94,
|
| 661 |
+
"learning_rate": 0.0002883319946076741,
|
| 662 |
+
"loss": 0.0486,
|
| 663 |
+
"step": 21800
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.94,
|
| 667 |
+
"learning_rate": 0.0002869477026493413,
|
| 668 |
+
"loss": 0.0576,
|
| 669 |
+
"step": 22000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.95,
|
| 673 |
+
"learning_rate": 0.0002855551861751398,
|
| 674 |
+
"loss": 0.0551,
|
| 675 |
+
"step": 22200
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.96,
|
| 679 |
+
"learning_rate": 0.00028415455761122624,
|
| 680 |
+
"loss": 0.0451,
|
| 681 |
+
"step": 22400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.97,
|
| 685 |
+
"learning_rate": 0.00028274593003869416,
|
| 686 |
+
"loss": 0.0479,
|
| 687 |
+
"step": 22600
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.98,
|
| 691 |
+
"learning_rate": 0.0002813365191744383,
|
| 692 |
+
"loss": 0.0519,
|
| 693 |
+
"step": 22800
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.99,
|
| 697 |
+
"learning_rate": 0.0002799122739710694,
|
| 698 |
+
"loss": 0.049,
|
| 699 |
+
"step": 23000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.0,
|
| 703 |
+
"learning_rate": 0.00027848037226393073,
|
| 704 |
+
"loss": 0.0489,
|
| 705 |
+
"step": 23200
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 1.0,
|
| 709 |
+
"eval_accuracy": 0.9363692882415535,
|
| 710 |
+
"eval_auc": 0.853910045782194,
|
| 711 |
+
"eval_f1": 0.0837364188034316,
|
| 712 |
+
"eval_loss": 0.6555151343345642,
|
| 713 |
+
"eval_mcc": 0.17516389459528245,
|
| 714 |
+
"eval_precision": 0.04427294640937209,
|
| 715 |
+
"eval_recall": 0.7708239966353413,
|
| 716 |
+
"eval_runtime": 1080.6969,
|
| 717 |
+
"eval_samples_per_second": 102.587,
|
| 718 |
+
"eval_steps_per_second": 5.399,
|
| 719 |
+
"step": 23309
|
| 720 |
+
}
|
| 721 |
+
],
|
| 722 |
+
"logging_steps": 200,
|
| 723 |
+
"max_steps": 69927,
|
| 724 |
+
"num_train_epochs": 3,
|
| 725 |
+
"save_steps": 500,
|
| 726 |
+
"total_flos": 1.97635804552818e+16,
|
| 727 |
+
"trial_name": null,
|
| 728 |
+
"trial_params": null
|
| 729 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ceb4d0a926525a8ef9bcb65db95def7efe3bbbdd762776f8aee1b7d91f04375
|
| 3 |
+
size 4091
|