Commit
·
064d2e1
1
Parent(s):
f6b5891
add ckpts
Browse files- args.pt +3 -0
- config.json +26 -0
- model.pt +3 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state_0.pth +3 -0
- rng_state_1.pth +3 -0
- rng_state_2.pth +3 -0
- rng_state_3.pth +3 -0
- rng_state_4.pth +3 -0
- rng_state_5.pth +3 -0
- rng_state_6.pth +3 -0
- rng_state_7.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1456 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
args.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bdbf267d3297b7b6f6bdfccaedf794c3dbcdd7ddf27548fd39042a6a67ccf0bb
|
| 3 |
+
size 3951
|
config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bert-base-uncased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 3072,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 512,
|
| 16 |
+
"model_type": "bert",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 12,
|
| 19 |
+
"pad_token_id": 0,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"torch_dtype": "float32",
|
| 22 |
+
"transformers_version": "4.18.0",
|
| 23 |
+
"type_vocab_size": 2,
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 30522
|
| 26 |
+
}
|
model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0e633e9719893083be3ba06a33b71e715dcbbc4b65862be2c2c59bc81fa1cb9c
|
| 3 |
+
size 283595215
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3cd4b27d00fae2cb99f2b79035d8a44832ac7bcc2103ae8ec4d3e07c38a7ec6
|
| 3 |
+
size 1443372131
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9394aec23ba3c84e5c6fc7f1f46552dd42ba415c6844f632935276cc80c0e21a
|
| 3 |
+
size 438134098
|
rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ea6d3f99a0edf6810b96bc1510ea2503f4b7da6ce58a89fba288be9570044ac
|
| 3 |
+
size 14657
|
rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7aa1ce0169de7c4d491b3cb6b86cd3ab778fb0ca33d0bdfb4e6082143866f9d7
|
| 3 |
+
size 14657
|
rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64ee90c25abb8c0cc375b8dacaa320e3f87a5fe2ca18d552d20d8474ca6615bd
|
| 3 |
+
size 14657
|
rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:758bc3008431dd6f358041d77aecc3b0678949a41b28091a00164f581e28de33
|
| 3 |
+
size 14657
|
rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a86a7c4de97edef786e8a9ea3a447c696c7ddb26207d120109839a48b5a8fbaf
|
| 3 |
+
size 14657
|
rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8332c5bb64161a3b7f19e5e6e504f12836d997880684676a231b9d1583523465
|
| 3 |
+
size 14657
|
rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:528ba575f15c3e9f76f28fd32a3f9c90ff80f490fe585e5938c17ba56f4c14a8
|
| 3 |
+
size 14657
|
rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9346b2e149e02e3ee23be437fd96c758f20cb7519a0c53b73d9e451d59bdea40
|
| 3 |
+
size 14657
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa1c761a901d57bec319b6a4902d581d7b8f93da42f78c1a18ed151c96c37d5f
|
| 3 |
+
size 559
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ede8324f9a526b8cbb7d57ae8a7a8b5969551b69ebe9efd06bb2e9919fa06416
|
| 3 |
+
size 623
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"do_lower_case": true, "do_basic_tokenize": true, "never_split": null, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "bert-base-uncased", "tokenizer_class": "BertTokenizer"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1456 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 27.796988998262883,
|
| 5 |
+
"global_step": 120000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 8.669214732453091e-06,
|
| 13 |
+
"loss": 49.8699,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.23,
|
| 18 |
+
"learning_rate": 1.7355802640722724e-05,
|
| 19 |
+
"loss": 41.2735,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.35,
|
| 24 |
+
"learning_rate": 2.6042390548992352e-05,
|
| 25 |
+
"loss": 36.9489,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.46,
|
| 30 |
+
"learning_rate": 3.472897845726199e-05,
|
| 31 |
+
"loss": 33.9779,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.58,
|
| 36 |
+
"learning_rate": 4.341556636553162e-05,
|
| 37 |
+
"loss": 31.7843,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.69,
|
| 42 |
+
"learning_rate": 5.210215427380125e-05,
|
| 43 |
+
"loss": 30.3453,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.81,
|
| 48 |
+
"learning_rate": 6.0788742182070875e-05,
|
| 49 |
+
"loss": 29.3517,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.93,
|
| 54 |
+
"learning_rate": 6.945795691452397e-05,
|
| 55 |
+
"loss": 28.5657,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.04,
|
| 60 |
+
"learning_rate": 7.81445448227936e-05,
|
| 61 |
+
"loss": 27.9059,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 1.16,
|
| 66 |
+
"learning_rate": 8.683113273106324e-05,
|
| 67 |
+
"loss": 27.256,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.27,
|
| 72 |
+
"learning_rate": 9.550034746351632e-05,
|
| 73 |
+
"loss": 26.6368,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.39,
|
| 78 |
+
"learning_rate": 0.00010418693537178596,
|
| 79 |
+
"loss": 26.0616,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.51,
|
| 84 |
+
"learning_rate": 0.00011287352328005559,
|
| 85 |
+
"loss": 25.5779,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.62,
|
| 90 |
+
"learning_rate": 0.00012154273801250867,
|
| 91 |
+
"loss": 25.154,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.74,
|
| 96 |
+
"learning_rate": 0.0001302293259207783,
|
| 97 |
+
"loss": 24.757,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.85,
|
| 102 |
+
"learning_rate": 0.00013891591382904795,
|
| 103 |
+
"loss": 24.3901,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.97,
|
| 108 |
+
"learning_rate": 0.0001476025017373176,
|
| 109 |
+
"loss": 24.0289,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 2.08,
|
| 114 |
+
"learning_rate": 0.0001562890896455872,
|
| 115 |
+
"loss": 23.7116,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.2,
|
| 120 |
+
"learning_rate": 0.0001649756775538568,
|
| 121 |
+
"loss": 23.3895,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 2.32,
|
| 126 |
+
"learning_rate": 0.00017366226546212648,
|
| 127 |
+
"loss": 23.1314,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.43,
|
| 132 |
+
"learning_rate": 0.00018234885337039609,
|
| 133 |
+
"loss": 22.8528,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.55,
|
| 138 |
+
"learning_rate": 0.00019103544127866572,
|
| 139 |
+
"loss": 22.5796,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.66,
|
| 144 |
+
"learning_rate": 0.00019972202918693536,
|
| 145 |
+
"loss": 22.3478,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.78,
|
| 150 |
+
"learning_rate": 0.000208408617095205,
|
| 151 |
+
"loss": 22.1018,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.9,
|
| 156 |
+
"learning_rate": 0.0002170952050034746,
|
| 157 |
+
"loss": 21.8626,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 3.01,
|
| 162 |
+
"learning_rate": 0.0002257644197359277,
|
| 163 |
+
"loss": 21.6681,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 3.13,
|
| 168 |
+
"learning_rate": 0.00023445100764419736,
|
| 169 |
+
"loss": 21.4255,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.24,
|
| 174 |
+
"learning_rate": 0.00024313759555246697,
|
| 175 |
+
"loss": 21.2724,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.36,
|
| 180 |
+
"learning_rate": 0.0002518241834607366,
|
| 181 |
+
"loss": 21.0849,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.47,
|
| 186 |
+
"learning_rate": 0.00026051077136900625,
|
| 187 |
+
"loss": 20.8968,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.59,
|
| 192 |
+
"learning_rate": 0.00026919735927727584,
|
| 193 |
+
"loss": 20.7233,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.71,
|
| 198 |
+
"learning_rate": 0.0002777970813064628,
|
| 199 |
+
"loss": 20.5867,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.82,
|
| 204 |
+
"learning_rate": 0.00028648366921473243,
|
| 205 |
+
"loss": 20.4368,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.94,
|
| 210 |
+
"learning_rate": 0.00029517025712300207,
|
| 211 |
+
"loss": 20.3374,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 4.05,
|
| 216 |
+
"learning_rate": 0.000299571461663192,
|
| 217 |
+
"loss": 20.2117,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 4.17,
|
| 222 |
+
"learning_rate": 0.00029860628522893984,
|
| 223 |
+
"loss": 20.0622,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.29,
|
| 228 |
+
"learning_rate": 0.00029764110879468763,
|
| 229 |
+
"loss": 19.9614,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.4,
|
| 234 |
+
"learning_rate": 0.0002966759323604355,
|
| 235 |
+
"loss": 19.8524,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.52,
|
| 240 |
+
"learning_rate": 0.0002957107559261833,
|
| 241 |
+
"loss": 19.7578,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 4.63,
|
| 246 |
+
"learning_rate": 0.00029474557949193107,
|
| 247 |
+
"loss": 19.6587,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.75,
|
| 252 |
+
"learning_rate": 0.0002937804030576789,
|
| 253 |
+
"loss": 19.5385,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 4.86,
|
| 258 |
+
"learning_rate": 0.0002928152266234267,
|
| 259 |
+
"loss": 19.4285,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.98,
|
| 264 |
+
"learning_rate": 0.00029185005018917456,
|
| 265 |
+
"loss": 19.3376,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 5.1,
|
| 270 |
+
"learning_rate": 0.0002908848737549224,
|
| 271 |
+
"loss": 19.2065,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 5.21,
|
| 276 |
+
"learning_rate": 0.0002899196973206702,
|
| 277 |
+
"loss": 19.1089,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.33,
|
| 282 |
+
"learning_rate": 0.000288954520886418,
|
| 283 |
+
"loss": 19.0411,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 5.44,
|
| 288 |
+
"learning_rate": 0.0002879893444521658,
|
| 289 |
+
"loss": 18.9864,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 5.56,
|
| 294 |
+
"learning_rate": 0.00028702416801791364,
|
| 295 |
+
"loss": 18.9157,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.68,
|
| 300 |
+
"learning_rate": 0.0002860589915836615,
|
| 301 |
+
"loss": 18.8606,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 5.79,
|
| 306 |
+
"learning_rate": 0.0002850938151494093,
|
| 307 |
+
"loss": 18.8049,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 5.91,
|
| 312 |
+
"learning_rate": 0.0002841286387151571,
|
| 313 |
+
"loss": 18.7415,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 6.02,
|
| 318 |
+
"learning_rate": 0.0002831634622809049,
|
| 319 |
+
"loss": 18.695,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 6.14,
|
| 324 |
+
"learning_rate": 0.0002821982858466527,
|
| 325 |
+
"loss": 18.5979,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 6.25,
|
| 330 |
+
"learning_rate": 0.00028123310941240056,
|
| 331 |
+
"loss": 18.5566,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 6.37,
|
| 336 |
+
"learning_rate": 0.0002802698633310169,
|
| 337 |
+
"loss": 18.5145,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.49,
|
| 342 |
+
"learning_rate": 0.0002793046868967647,
|
| 343 |
+
"loss": 18.465,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 6.6,
|
| 348 |
+
"learning_rate": 0.0002783395104625125,
|
| 349 |
+
"loss": 18.4232,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 6.72,
|
| 354 |
+
"learning_rate": 0.00027737433402826034,
|
| 355 |
+
"loss": 18.368,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.83,
|
| 360 |
+
"learning_rate": 0.00027640915759400813,
|
| 361 |
+
"loss": 18.315,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.95,
|
| 366 |
+
"learning_rate": 0.000275443981159756,
|
| 367 |
+
"loss": 18.2882,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 7.07,
|
| 372 |
+
"learning_rate": 0.00027447880472550383,
|
| 373 |
+
"loss": 18.2239,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 7.18,
|
| 378 |
+
"learning_rate": 0.0002735136282912516,
|
| 379 |
+
"loss": 18.163,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 7.3,
|
| 384 |
+
"learning_rate": 0.0002725503822098679,
|
| 385 |
+
"loss": 18.1283,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 7.41,
|
| 390 |
+
"learning_rate": 0.00027158520577561575,
|
| 391 |
+
"loss": 18.1016,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 7.53,
|
| 396 |
+
"learning_rate": 0.0002706200293413636,
|
| 397 |
+
"loss": 18.0663,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.64,
|
| 402 |
+
"learning_rate": 0.0002696567832599799,
|
| 403 |
+
"loss": 18.0263,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.76,
|
| 408 |
+
"learning_rate": 0.00026869160682572773,
|
| 409 |
+
"loss": 17.9946,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 7.88,
|
| 414 |
+
"learning_rate": 0.00026772643039147553,
|
| 415 |
+
"loss": 17.9719,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.99,
|
| 420 |
+
"learning_rate": 0.0002667612539572234,
|
| 421 |
+
"loss": 17.931,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 8.11,
|
| 426 |
+
"learning_rate": 0.00026579607752297117,
|
| 427 |
+
"loss": 17.8646,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 8.22,
|
| 432 |
+
"learning_rate": 0.000264830901088719,
|
| 433 |
+
"loss": 17.8358,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 8.34,
|
| 438 |
+
"learning_rate": 0.0002638676550073353,
|
| 439 |
+
"loss": 17.8081,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 8.45,
|
| 444 |
+
"learning_rate": 0.00026290247857308315,
|
| 445 |
+
"loss": 17.7928,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 8.57,
|
| 450 |
+
"learning_rate": 0.00026193730213883095,
|
| 451 |
+
"loss": 17.7663,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 8.69,
|
| 456 |
+
"learning_rate": 0.0002609721257045788,
|
| 457 |
+
"loss": 17.7224,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 8.8,
|
| 462 |
+
"learning_rate": 0.0002600069492703266,
|
| 463 |
+
"loss": 17.7134,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 8.92,
|
| 468 |
+
"learning_rate": 0.0002590417728360744,
|
| 469 |
+
"loss": 17.6934,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 9.03,
|
| 474 |
+
"learning_rate": 0.00025807659640182223,
|
| 475 |
+
"loss": 17.662,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 9.15,
|
| 480 |
+
"learning_rate": 0.0002571114199675701,
|
| 481 |
+
"loss": 17.5886,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 9.27,
|
| 486 |
+
"learning_rate": 0.00025615010423905485,
|
| 487 |
+
"loss": 17.5854,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 9.38,
|
| 492 |
+
"learning_rate": 0.0002551849278048027,
|
| 493 |
+
"loss": 17.5689,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 9.5,
|
| 498 |
+
"learning_rate": 0.00025421975137055055,
|
| 499 |
+
"loss": 17.5545,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 9.61,
|
| 504 |
+
"learning_rate": 0.00025325650528916683,
|
| 505 |
+
"loss": 17.5269,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 9.73,
|
| 510 |
+
"learning_rate": 0.00025229132885491463,
|
| 511 |
+
"loss": 17.5051,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 9.84,
|
| 516 |
+
"learning_rate": 0.0002513261524206625,
|
| 517 |
+
"loss": 17.4922,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 9.96,
|
| 522 |
+
"learning_rate": 0.0002503609759864103,
|
| 523 |
+
"loss": 17.4682,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 10.08,
|
| 528 |
+
"learning_rate": 0.0002493957995521581,
|
| 529 |
+
"loss": 17.4151,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 10.19,
|
| 534 |
+
"learning_rate": 0.0002484306231179059,
|
| 535 |
+
"loss": 17.3825,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 10.31,
|
| 540 |
+
"learning_rate": 0.00024746544668365376,
|
| 541 |
+
"loss": 17.3681,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 10.42,
|
| 546 |
+
"learning_rate": 0.00024650027024940155,
|
| 547 |
+
"loss": 17.341,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 10.54,
|
| 552 |
+
"learning_rate": 0.0002455350938151494,
|
| 553 |
+
"loss": 17.3344,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 10.66,
|
| 558 |
+
"learning_rate": 0.0002445699173808972,
|
| 559 |
+
"loss": 17.316,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 10.77,
|
| 564 |
+
"learning_rate": 0.00024360474094664502,
|
| 565 |
+
"loss": 17.3002,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 10.89,
|
| 570 |
+
"learning_rate": 0.00024264149486526136,
|
| 571 |
+
"loss": 17.279,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 11.0,
|
| 576 |
+
"learning_rate": 0.00024167631843100918,
|
| 577 |
+
"loss": 17.2714,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 11.12,
|
| 582 |
+
"learning_rate": 0.00024071114199675697,
|
| 583 |
+
"loss": 17.2032,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 11.23,
|
| 588 |
+
"learning_rate": 0.00023974596556250482,
|
| 589 |
+
"loss": 17.1982,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 11.35,
|
| 594 |
+
"learning_rate": 0.00023878271948112113,
|
| 595 |
+
"loss": 17.1764,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 11.47,
|
| 600 |
+
"learning_rate": 0.00023781754304686895,
|
| 601 |
+
"loss": 17.1697,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 11.58,
|
| 606 |
+
"learning_rate": 0.00023685236661261674,
|
| 607 |
+
"loss": 17.1619,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 11.7,
|
| 612 |
+
"learning_rate": 0.0002358871901783646,
|
| 613 |
+
"loss": 17.1475,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 11.81,
|
| 618 |
+
"learning_rate": 0.0002349239440969809,
|
| 619 |
+
"loss": 17.1354,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 11.93,
|
| 624 |
+
"learning_rate": 0.00023395876766272873,
|
| 625 |
+
"loss": 17.1193,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 12.05,
|
| 630 |
+
"learning_rate": 0.00023299359122847657,
|
| 631 |
+
"loss": 17.0976,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 12.16,
|
| 636 |
+
"learning_rate": 0.00023203034514709289,
|
| 637 |
+
"loss": 17.0559,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 12.28,
|
| 642 |
+
"learning_rate": 0.00023106516871284068,
|
| 643 |
+
"loss": 17.0478,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 12.39,
|
| 648 |
+
"learning_rate": 0.0002300999922785885,
|
| 649 |
+
"loss": 17.0457,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 12.51,
|
| 654 |
+
"learning_rate": 0.00022913481584433635,
|
| 655 |
+
"loss": 17.0296,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 12.62,
|
| 660 |
+
"learning_rate": 0.00022816963941008414,
|
| 661 |
+
"loss": 17.0185,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 12.74,
|
| 666 |
+
"learning_rate": 0.00022720639332870045,
|
| 667 |
+
"loss": 17.012,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 12.86,
|
| 672 |
+
"learning_rate": 0.0002262431472473168,
|
| 673 |
+
"loss": 16.9982,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 12.97,
|
| 678 |
+
"learning_rate": 0.00022527990116593313,
|
| 679 |
+
"loss": 16.981,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 13.09,
|
| 684 |
+
"learning_rate": 0.00022431472473168093,
|
| 685 |
+
"loss": 16.9492,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 13.2,
|
| 690 |
+
"learning_rate": 0.00022335147865029726,
|
| 691 |
+
"loss": 16.9352,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 13.32,
|
| 696 |
+
"learning_rate": 0.00022238630221604506,
|
| 697 |
+
"loss": 16.9178,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 13.44,
|
| 702 |
+
"learning_rate": 0.0002214211257817929,
|
| 703 |
+
"loss": 16.9156,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 13.55,
|
| 708 |
+
"learning_rate": 0.00022045787970040922,
|
| 709 |
+
"loss": 16.9093,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 13.67,
|
| 714 |
+
"learning_rate": 0.00021949270326615704,
|
| 715 |
+
"loss": 16.9013,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 13.78,
|
| 720 |
+
"learning_rate": 0.00021852752683190483,
|
| 721 |
+
"loss": 16.8895,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 13.9,
|
| 726 |
+
"learning_rate": 0.00021756235039765268,
|
| 727 |
+
"loss": 16.8733,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 14.01,
|
| 732 |
+
"learning_rate": 0.0002165971739634005,
|
| 733 |
+
"loss": 16.881,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 14.13,
|
| 738 |
+
"learning_rate": 0.0002156319975291483,
|
| 739 |
+
"loss": 16.8109,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 14.25,
|
| 744 |
+
"learning_rate": 0.00021466682109489614,
|
| 745 |
+
"loss": 16.8187,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 14.36,
|
| 750 |
+
"learning_rate": 0.00021370164466064394,
|
| 751 |
+
"loss": 16.8078,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 14.48,
|
| 756 |
+
"learning_rate": 0.00021273646822639176,
|
| 757 |
+
"loss": 16.8041,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 14.59,
|
| 762 |
+
"learning_rate": 0.0002117712917921396,
|
| 763 |
+
"loss": 16.7933,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 14.71,
|
| 768 |
+
"learning_rate": 0.0002108061153578874,
|
| 769 |
+
"loss": 16.7964,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 14.83,
|
| 774 |
+
"learning_rate": 0.00020984093892363522,
|
| 775 |
+
"loss": 16.778,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 14.94,
|
| 780 |
+
"learning_rate": 0.00020887576248938302,
|
| 781 |
+
"loss": 16.7764,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 15.06,
|
| 786 |
+
"learning_rate": 0.00020791058605513086,
|
| 787 |
+
"loss": 16.7603,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 15.17,
|
| 792 |
+
"learning_rate": 0.00020694540962087868,
|
| 793 |
+
"loss": 16.7251,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 15.29,
|
| 798 |
+
"learning_rate": 0.00020598023318662648,
|
| 799 |
+
"loss": 16.712,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 15.4,
|
| 804 |
+
"learning_rate": 0.00020501505675237433,
|
| 805 |
+
"loss": 16.7072,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 15.52,
|
| 810 |
+
"learning_rate": 0.00020404988031812215,
|
| 811 |
+
"loss": 16.7106,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 15.64,
|
| 816 |
+
"learning_rate": 0.00020308470388386994,
|
| 817 |
+
"loss": 16.7037,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 15.75,
|
| 822 |
+
"learning_rate": 0.0002021195274496178,
|
| 823 |
+
"loss": 16.6918,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 15.87,
|
| 828 |
+
"learning_rate": 0.00020115435101536558,
|
| 829 |
+
"loss": 16.6886,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 15.98,
|
| 834 |
+
"learning_rate": 0.0002001891745811134,
|
| 835 |
+
"loss": 16.6947,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 16.1,
|
| 840 |
+
"learning_rate": 0.00019922399814686125,
|
| 841 |
+
"loss": 16.6516,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 16.21,
|
| 846 |
+
"learning_rate": 0.00019825882171260905,
|
| 847 |
+
"loss": 16.6403,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 16.33,
|
| 852 |
+
"learning_rate": 0.00019729364527835687,
|
| 853 |
+
"loss": 16.6343,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 16.45,
|
| 858 |
+
"learning_rate": 0.00019632846884410466,
|
| 859 |
+
"loss": 16.6257,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 16.56,
|
| 864 |
+
"learning_rate": 0.00019536522276272103,
|
| 865 |
+
"loss": 16.6181,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 16.68,
|
| 870 |
+
"learning_rate": 0.00019440004632846882,
|
| 871 |
+
"loss": 16.6204,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 16.79,
|
| 876 |
+
"learning_rate": 0.00019343486989421664,
|
| 877 |
+
"loss": 16.6101,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 16.91,
|
| 882 |
+
"learning_rate": 0.00019247162381283295,
|
| 883 |
+
"loss": 16.6116,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 17.03,
|
| 888 |
+
"learning_rate": 0.0001915083777314493,
|
| 889 |
+
"loss": 16.6047,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 17.14,
|
| 894 |
+
"learning_rate": 0.0001905432012971971,
|
| 895 |
+
"loss": 16.5427,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 17.26,
|
| 900 |
+
"learning_rate": 0.00018957802486294493,
|
| 901 |
+
"loss": 16.5453,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 17.37,
|
| 906 |
+
"learning_rate": 0.00018861284842869276,
|
| 907 |
+
"loss": 16.5489,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 17.49,
|
| 912 |
+
"learning_rate": 0.00018764960234730907,
|
| 913 |
+
"loss": 16.5505,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 17.6,
|
| 918 |
+
"learning_rate": 0.0001866844259130569,
|
| 919 |
+
"loss": 16.5423,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 17.72,
|
| 924 |
+
"learning_rate": 0.0001857192494788047,
|
| 925 |
+
"loss": 16.543,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 17.84,
|
| 930 |
+
"learning_rate": 0.00018475407304455253,
|
| 931 |
+
"loss": 16.5402,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 17.95,
|
| 936 |
+
"learning_rate": 0.00018378889661030035,
|
| 937 |
+
"loss": 16.5373,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 18.07,
|
| 942 |
+
"learning_rate": 0.00018282372017604814,
|
| 943 |
+
"loss": 16.5045,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 18.18,
|
| 948 |
+
"learning_rate": 0.000181858543741796,
|
| 949 |
+
"loss": 16.4782,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 18.3,
|
| 954 |
+
"learning_rate": 0.00018089336730754381,
|
| 955 |
+
"loss": 16.4817,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 18.42,
|
| 960 |
+
"learning_rate": 0.0001799281908732916,
|
| 961 |
+
"loss": 16.4737,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 18.53,
|
| 966 |
+
"learning_rate": 0.00017896494479190797,
|
| 967 |
+
"loss": 16.4797,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 18.65,
|
| 972 |
+
"learning_rate": 0.00017799976835765577,
|
| 973 |
+
"loss": 16.4755,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 18.76,
|
| 978 |
+
"learning_rate": 0.0001770345919234036,
|
| 979 |
+
"loss": 16.464,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 18.88,
|
| 984 |
+
"learning_rate": 0.00017606941548915138,
|
| 985 |
+
"loss": 16.4599,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 18.99,
|
| 990 |
+
"learning_rate": 0.00017510423905489923,
|
| 991 |
+
"loss": 16.4511,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 19.11,
|
| 996 |
+
"learning_rate": 0.00017413906262064705,
|
| 997 |
+
"loss": 16.42,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 19.23,
|
| 1002 |
+
"learning_rate": 0.00017317581653926336,
|
| 1003 |
+
"loss": 16.4039,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 19.34,
|
| 1008 |
+
"learning_rate": 0.00017221064010501116,
|
| 1009 |
+
"loss": 16.4083,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 19.46,
|
| 1014 |
+
"learning_rate": 0.000171245463670759,
|
| 1015 |
+
"loss": 16.413,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 19.57,
|
| 1020 |
+
"learning_rate": 0.00017028028723650683,
|
| 1021 |
+
"loss": 16.4069,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 19.69,
|
| 1026 |
+
"learning_rate": 0.00016931511080225462,
|
| 1027 |
+
"loss": 16.3972,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 19.81,
|
| 1032 |
+
"learning_rate": 0.00016834993436800247,
|
| 1033 |
+
"loss": 16.4007,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 19.92,
|
| 1038 |
+
"learning_rate": 0.0001673847579337503,
|
| 1039 |
+
"loss": 16.3945,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 20.04,
|
| 1044 |
+
"learning_rate": 0.00016641958149949808,
|
| 1045 |
+
"loss": 16.3928,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 20.15,
|
| 1050 |
+
"learning_rate": 0.00016545440506524593,
|
| 1051 |
+
"loss": 16.3468,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 20.27,
|
| 1056 |
+
"learning_rate": 0.00016448922863099373,
|
| 1057 |
+
"loss": 16.3459,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 20.38,
|
| 1062 |
+
"learning_rate": 0.00016352598254961006,
|
| 1063 |
+
"loss": 16.3429,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 20.5,
|
| 1068 |
+
"learning_rate": 0.00016256080611535786,
|
| 1069 |
+
"loss": 16.3394,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 20.62,
|
| 1074 |
+
"learning_rate": 0.0001615956296811057,
|
| 1075 |
+
"loss": 16.3417,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 20.73,
|
| 1080 |
+
"learning_rate": 0.0001606304532468535,
|
| 1081 |
+
"loss": 16.3369,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 20.85,
|
| 1086 |
+
"learning_rate": 0.00015966527681260132,
|
| 1087 |
+
"loss": 16.3392,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 20.96,
|
| 1092 |
+
"learning_rate": 0.00015870203073121763,
|
| 1093 |
+
"loss": 16.3402,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 21.08,
|
| 1098 |
+
"learning_rate": 0.00015773685429696548,
|
| 1099 |
+
"loss": 16.3009,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 21.2,
|
| 1104 |
+
"learning_rate": 0.00015677167786271327,
|
| 1105 |
+
"loss": 16.2924,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 21.31,
|
| 1110 |
+
"learning_rate": 0.0001558065014284611,
|
| 1111 |
+
"loss": 16.2908,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 21.43,
|
| 1116 |
+
"learning_rate": 0.00015484132499420894,
|
| 1117 |
+
"loss": 16.289,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 21.54,
|
| 1122 |
+
"learning_rate": 0.00015387614855995674,
|
| 1123 |
+
"loss": 16.2874,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 21.66,
|
| 1128 |
+
"learning_rate": 0.00015291097212570456,
|
| 1129 |
+
"loss": 16.2869,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 21.77,
|
| 1134 |
+
"learning_rate": 0.00015194772604432087,
|
| 1135 |
+
"loss": 16.2867,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 21.89,
|
| 1140 |
+
"learning_rate": 0.00015098254961006872,
|
| 1141 |
+
"loss": 16.2779,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 22.01,
|
| 1146 |
+
"learning_rate": 0.00015002123388155355,
|
| 1147 |
+
"loss": 16.2885,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 22.12,
|
| 1152 |
+
"learning_rate": 0.00014905605744730134,
|
| 1153 |
+
"loss": 16.2251,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 22.24,
|
| 1158 |
+
"learning_rate": 0.0001480908810130492,
|
| 1159 |
+
"loss": 16.247,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 22.35,
|
| 1164 |
+
"learning_rate": 0.00014712570457879698,
|
| 1165 |
+
"loss": 16.2315,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 22.47,
|
| 1170 |
+
"learning_rate": 0.0001461605281445448,
|
| 1171 |
+
"loss": 16.2464,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 22.59,
|
| 1176 |
+
"learning_rate": 0.00014519535171029263,
|
| 1177 |
+
"loss": 16.2326,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 22.7,
|
| 1182 |
+
"learning_rate": 0.00014423017527604045,
|
| 1183 |
+
"loss": 16.2327,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 22.82,
|
| 1188 |
+
"learning_rate": 0.00014326499884178827,
|
| 1189 |
+
"loss": 16.2269,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 22.93,
|
| 1194 |
+
"learning_rate": 0.0001422998224075361,
|
| 1195 |
+
"loss": 16.2186,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 23.05,
|
| 1200 |
+
"learning_rate": 0.0001413346459732839,
|
| 1201 |
+
"loss": 16.2302,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 23.16,
|
| 1206 |
+
"learning_rate": 0.0001403694695390317,
|
| 1207 |
+
"loss": 16.1879,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 23.28,
|
| 1212 |
+
"learning_rate": 0.00013940622345764804,
|
| 1213 |
+
"loss": 16.1834,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 23.4,
|
| 1218 |
+
"learning_rate": 0.00013844104702339586,
|
| 1219 |
+
"loss": 16.1779,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 23.51,
|
| 1224 |
+
"learning_rate": 0.00013747587058914368,
|
| 1225 |
+
"loss": 16.1835,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 23.63,
|
| 1230 |
+
"learning_rate": 0.00013651455486062851,
|
| 1231 |
+
"loss": 16.1829,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 23.74,
|
| 1236 |
+
"learning_rate": 0.00013554937842637633,
|
| 1237 |
+
"loss": 16.1814,
|
| 1238 |
+
"step": 102500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 23.86,
|
| 1242 |
+
"learning_rate": 0.00013458420199212413,
|
| 1243 |
+
"loss": 16.1864,
|
| 1244 |
+
"step": 103000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 23.97,
|
| 1248 |
+
"learning_rate": 0.00013361902555787198,
|
| 1249 |
+
"loss": 16.1784,
|
| 1250 |
+
"step": 103500
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 24.09,
|
| 1254 |
+
"learning_rate": 0.0001326538491236198,
|
| 1255 |
+
"loss": 16.149,
|
| 1256 |
+
"step": 104000
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 24.21,
|
| 1260 |
+
"learning_rate": 0.0001316886726893676,
|
| 1261 |
+
"loss": 16.1459,
|
| 1262 |
+
"step": 104500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 24.32,
|
| 1266 |
+
"learning_rate": 0.0001307234962551154,
|
| 1267 |
+
"loss": 16.1374,
|
| 1268 |
+
"step": 105000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 24.44,
|
| 1272 |
+
"learning_rate": 0.00012975831982086326,
|
| 1273 |
+
"loss": 16.143,
|
| 1274 |
+
"step": 105500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 24.55,
|
| 1278 |
+
"learning_rate": 0.00012879314338661105,
|
| 1279 |
+
"loss": 16.1267,
|
| 1280 |
+
"step": 106000
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 24.67,
|
| 1284 |
+
"learning_rate": 0.00012782796695235888,
|
| 1285 |
+
"loss": 16.1363,
|
| 1286 |
+
"step": 106500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 24.79,
|
| 1290 |
+
"learning_rate": 0.0001268666512238437,
|
| 1291 |
+
"loss": 16.1328,
|
| 1292 |
+
"step": 107000
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 24.9,
|
| 1296 |
+
"learning_rate": 0.00012590147478959153,
|
| 1297 |
+
"loss": 16.1314,
|
| 1298 |
+
"step": 107500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 25.02,
|
| 1302 |
+
"learning_rate": 0.00012493629835533935,
|
| 1303 |
+
"loss": 16.1427,
|
| 1304 |
+
"step": 108000
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 25.13,
|
| 1308 |
+
"learning_rate": 0.00012397112192108717,
|
| 1309 |
+
"loss": 16.0975,
|
| 1310 |
+
"step": 108500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 25.25,
|
| 1314 |
+
"learning_rate": 0.000123005945486835,
|
| 1315 |
+
"loss": 16.091,
|
| 1316 |
+
"step": 109000
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 25.36,
|
| 1320 |
+
"learning_rate": 0.00012204076905258281,
|
| 1321 |
+
"loss": 16.0959,
|
| 1322 |
+
"step": 109500
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 25.48,
|
| 1326 |
+
"learning_rate": 0.00012107559261833062,
|
| 1327 |
+
"loss": 16.0919,
|
| 1328 |
+
"step": 110000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 25.6,
|
| 1332 |
+
"learning_rate": 0.00012011041618407844,
|
| 1333 |
+
"loss": 16.077,
|
| 1334 |
+
"step": 110500
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 25.71,
|
| 1338 |
+
"learning_rate": 0.00011914523974982626,
|
| 1339 |
+
"loss": 16.0942,
|
| 1340 |
+
"step": 111000
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 25.83,
|
| 1344 |
+
"learning_rate": 0.00011818006331557408,
|
| 1345 |
+
"loss": 16.0786,
|
| 1346 |
+
"step": 111500
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 25.94,
|
| 1350 |
+
"learning_rate": 0.00011721681723419039,
|
| 1351 |
+
"loss": 16.0852,
|
| 1352 |
+
"step": 112000
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 26.06,
|
| 1356 |
+
"learning_rate": 0.00011625164079993821,
|
| 1357 |
+
"loss": 16.0732,
|
| 1358 |
+
"step": 112500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 26.18,
|
| 1362 |
+
"learning_rate": 0.00011528646436568605,
|
| 1363 |
+
"loss": 16.0445,
|
| 1364 |
+
"step": 113000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 26.29,
|
| 1368 |
+
"learning_rate": 0.00011432128793143386,
|
| 1369 |
+
"loss": 16.043,
|
| 1370 |
+
"step": 113500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 26.41,
|
| 1374 |
+
"learning_rate": 0.00011335611149718168,
|
| 1375 |
+
"loss": 16.0484,
|
| 1376 |
+
"step": 114000
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 26.52,
|
| 1380 |
+
"learning_rate": 0.00011239093506292948,
|
| 1381 |
+
"loss": 16.0494,
|
| 1382 |
+
"step": 114500
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 26.64,
|
| 1386 |
+
"learning_rate": 0.00011142575862867732,
|
| 1387 |
+
"loss": 16.043,
|
| 1388 |
+
"step": 115000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 26.75,
|
| 1392 |
+
"learning_rate": 0.00011046058219442514,
|
| 1393 |
+
"loss": 16.0415,
|
| 1394 |
+
"step": 115500
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 26.87,
|
| 1398 |
+
"learning_rate": 0.00010949540576017295,
|
| 1399 |
+
"loss": 16.032,
|
| 1400 |
+
"step": 116000
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 26.99,
|
| 1404 |
+
"learning_rate": 0.00010853022932592077,
|
| 1405 |
+
"loss": 16.0346,
|
| 1406 |
+
"step": 116500
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 27.1,
|
| 1410 |
+
"learning_rate": 0.00010756505289166857,
|
| 1411 |
+
"loss": 16.0145,
|
| 1412 |
+
"step": 117000
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 27.22,
|
| 1416 |
+
"learning_rate": 0.00010659987645741641,
|
| 1417 |
+
"loss": 16.0032,
|
| 1418 |
+
"step": 117500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 27.33,
|
| 1422 |
+
"learning_rate": 0.00010563470002316423,
|
| 1423 |
+
"loss": 15.9992,
|
| 1424 |
+
"step": 118000
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 27.45,
|
| 1428 |
+
"learning_rate": 0.00010467145394178054,
|
| 1429 |
+
"loss": 16.0033,
|
| 1430 |
+
"step": 118500
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 27.57,
|
| 1434 |
+
"learning_rate": 0.00010370627750752838,
|
| 1435 |
+
"loss": 15.9964,
|
| 1436 |
+
"step": 119000
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 27.68,
|
| 1440 |
+
"learning_rate": 0.00010274303142614469,
|
| 1441 |
+
"loss": 15.9994,
|
| 1442 |
+
"step": 119500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 27.8,
|
| 1446 |
+
"learning_rate": 0.00010177785499189251,
|
| 1447 |
+
"loss": 16.0,
|
| 1448 |
+
"step": 120000
|
| 1449 |
+
}
|
| 1450 |
+
],
|
| 1451 |
+
"max_steps": 172680,
|
| 1452 |
+
"num_train_epochs": 40,
|
| 1453 |
+
"total_flos": 0.0,
|
| 1454 |
+
"trial_name": null,
|
| 1455 |
+
"trial_params": null
|
| 1456 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c25de5e38c7cac9ebb8d0a5f4b4d31951aaca3b8082430750b3a626ebebed5ec
|
| 3 |
+
size 3119
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|