Commit
·
0df5f89
1
Parent(s):
8cb5b79
Upload 13 files
Browse files- config.json +1 -1
- optimizer.pt +3 -0
- pytorch_model.bin +2 -2
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +2488 -0
- training_args.bin +2 -2
config.json
CHANGED
|
@@ -37,7 +37,7 @@
|
|
| 37 |
"relative_attention": true,
|
| 38 |
"share_att_key": true,
|
| 39 |
"torch_dtype": "float32",
|
| 40 |
-
"transformers_version": "4.
|
| 41 |
"type_vocab_size": 0,
|
| 42 |
"vocab_size": 128100
|
| 43 |
}
|
|
|
|
| 37 |
"relative_attention": true,
|
| 38 |
"share_att_key": true,
|
| 39 |
"torch_dtype": "float32",
|
| 40 |
+
"transformers_version": "4.26.0",
|
| 41 |
"type_vocab_size": 0,
|
| 42 |
"vocab_size": 128100
|
| 43 |
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8390bd88cb829219221edf785c25d1acbfe23b283de6c6613c4c58c68c53e6f
|
| 3 |
+
size 566764677
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb69ea5f3b512727cdf80ec979bf3311f77b5f69513d45f5fc3db656b4dc2119
|
| 3 |
+
size 283396921
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2127737992fcd0482aafa028b8ffbf6f4f4f77f4951f1b81b70be0e7d399a68e
|
| 3 |
+
size 14575
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:feae672a5ba222a4804519f9b64db3e81116eb8a9d3dbd9e6158cfa460ae18c9
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:242b4344aab1ac33adf4a5de5911c5e7c2a1f5133a2834f5dfddeaaa03e2cf7a
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2488 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.4389118552207947,
|
| 3 |
+
"best_model_checkpoint": "OutModelPolicy\\checkpoint-102317",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 204634,
|
| 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": 1.995123000087962e-05,
|
| 13 |
+
"loss": 0.5994,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 1.990236226629006e-05,
|
| 19 |
+
"loss": 0.5391,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.01,
|
| 24 |
+
"learning_rate": 1.98534945317005e-05,
|
| 25 |
+
"loss": 0.5162,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.02,
|
| 30 |
+
"learning_rate": 1.980462679711094e-05,
|
| 31 |
+
"loss": 0.5139,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.02,
|
| 36 |
+
"learning_rate": 1.975575906252138e-05,
|
| 37 |
+
"loss": 0.4944,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.03,
|
| 42 |
+
"learning_rate": 1.9706891327931822e-05,
|
| 43 |
+
"loss": 0.5148,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"learning_rate": 1.9658121328811442e-05,
|
| 49 |
+
"loss": 0.4783,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.04,
|
| 54 |
+
"learning_rate": 1.960925359422188e-05,
|
| 55 |
+
"loss": 0.492,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.04,
|
| 60 |
+
"learning_rate": 1.956038585963232e-05,
|
| 61 |
+
"loss": 0.4928,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.05,
|
| 66 |
+
"learning_rate": 1.951151812504276e-05,
|
| 67 |
+
"loss": 0.4738,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.05,
|
| 72 |
+
"learning_rate": 1.9462748125922378e-05,
|
| 73 |
+
"loss": 0.4831,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.06,
|
| 78 |
+
"learning_rate": 1.941388039133282e-05,
|
| 79 |
+
"loss": 0.4748,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.06,
|
| 84 |
+
"learning_rate": 1.936501265674326e-05,
|
| 85 |
+
"loss": 0.4787,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.07,
|
| 90 |
+
"learning_rate": 1.93161449221537e-05,
|
| 91 |
+
"loss": 0.4653,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.07,
|
| 96 |
+
"learning_rate": 1.926737492303332e-05,
|
| 97 |
+
"loss": 0.478,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.08,
|
| 102 |
+
"learning_rate": 1.921850718844376e-05,
|
| 103 |
+
"loss": 0.4547,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.08,
|
| 108 |
+
"learning_rate": 1.9169639453854202e-05,
|
| 109 |
+
"loss": 0.4747,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.09,
|
| 114 |
+
"learning_rate": 1.912077171926464e-05,
|
| 115 |
+
"loss": 0.4731,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.09,
|
| 120 |
+
"learning_rate": 1.907190398467508e-05,
|
| 121 |
+
"loss": 0.4617,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.1,
|
| 126 |
+
"learning_rate": 1.9023133985554697e-05,
|
| 127 |
+
"loss": 0.4649,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1,
|
| 132 |
+
"learning_rate": 1.8974266250965138e-05,
|
| 133 |
+
"loss": 0.4846,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.11,
|
| 138 |
+
"learning_rate": 1.892539851637558e-05,
|
| 139 |
+
"loss": 0.4757,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.11,
|
| 144 |
+
"learning_rate": 1.887653078178602e-05,
|
| 145 |
+
"loss": 0.4607,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.12,
|
| 150 |
+
"learning_rate": 1.882766304719646e-05,
|
| 151 |
+
"loss": 0.4671,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.12,
|
| 156 |
+
"learning_rate": 1.877889304807608e-05,
|
| 157 |
+
"loss": 0.4645,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.13,
|
| 162 |
+
"learning_rate": 1.8730025313486518e-05,
|
| 163 |
+
"loss": 0.4648,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.13,
|
| 168 |
+
"learning_rate": 1.868115757889696e-05,
|
| 169 |
+
"loss": 0.4648,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.14,
|
| 174 |
+
"learning_rate": 1.86322898443074e-05,
|
| 175 |
+
"loss": 0.4645,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.14,
|
| 180 |
+
"learning_rate": 1.858351984518702e-05,
|
| 181 |
+
"loss": 0.4504,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.15,
|
| 186 |
+
"learning_rate": 1.8534652110597457e-05,
|
| 187 |
+
"loss": 0.4601,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.15,
|
| 192 |
+
"learning_rate": 1.8485784376007898e-05,
|
| 193 |
+
"loss": 0.4637,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.16,
|
| 198 |
+
"learning_rate": 1.843691664141834e-05,
|
| 199 |
+
"loss": 0.4547,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.16,
|
| 204 |
+
"learning_rate": 1.838804890682878e-05,
|
| 205 |
+
"loss": 0.471,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.17,
|
| 210 |
+
"learning_rate": 1.8339278907708396e-05,
|
| 211 |
+
"loss": 0.4487,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.17,
|
| 216 |
+
"learning_rate": 1.8290411173118837e-05,
|
| 217 |
+
"loss": 0.4755,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.18,
|
| 222 |
+
"learning_rate": 1.8241543438529278e-05,
|
| 223 |
+
"loss": 0.4639,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.18,
|
| 228 |
+
"learning_rate": 1.8192773439408898e-05,
|
| 229 |
+
"loss": 0.4573,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.19,
|
| 234 |
+
"learning_rate": 1.814390570481934e-05,
|
| 235 |
+
"loss": 0.45,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.19,
|
| 240 |
+
"learning_rate": 1.809503797022978e-05,
|
| 241 |
+
"loss": 0.455,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.2,
|
| 246 |
+
"learning_rate": 1.8046170235640217e-05,
|
| 247 |
+
"loss": 0.4481,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2,
|
| 252 |
+
"learning_rate": 1.7997302501050658e-05,
|
| 253 |
+
"loss": 0.4669,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.21,
|
| 258 |
+
"learning_rate": 1.79484347664611e-05,
|
| 259 |
+
"loss": 0.4569,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.21,
|
| 264 |
+
"learning_rate": 1.789956703187154e-05,
|
| 265 |
+
"loss": 0.4503,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.22,
|
| 270 |
+
"learning_rate": 1.7850699297281977e-05,
|
| 271 |
+
"loss": 0.4544,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.22,
|
| 276 |
+
"learning_rate": 1.7801929298161597e-05,
|
| 277 |
+
"loss": 0.4563,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.22,
|
| 282 |
+
"learning_rate": 1.7753061563572038e-05,
|
| 283 |
+
"loss": 0.4463,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.23,
|
| 288 |
+
"learning_rate": 1.7704193828982475e-05,
|
| 289 |
+
"loss": 0.4467,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.23,
|
| 294 |
+
"learning_rate": 1.7655326094392916e-05,
|
| 295 |
+
"loss": 0.46,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.24,
|
| 300 |
+
"learning_rate": 1.760655609527254e-05,
|
| 301 |
+
"loss": 0.4554,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.24,
|
| 306 |
+
"learning_rate": 1.7557688360682977e-05,
|
| 307 |
+
"loss": 0.4516,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.25,
|
| 312 |
+
"learning_rate": 1.7508918361562597e-05,
|
| 313 |
+
"loss": 0.435,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.25,
|
| 318 |
+
"learning_rate": 1.7460050626973034e-05,
|
| 319 |
+
"loss": 0.442,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.26,
|
| 324 |
+
"learning_rate": 1.7411182892383475e-05,
|
| 325 |
+
"loss": 0.4499,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.26,
|
| 330 |
+
"learning_rate": 1.7362315157793916e-05,
|
| 331 |
+
"loss": 0.4358,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.27,
|
| 336 |
+
"learning_rate": 1.7313545158673536e-05,
|
| 337 |
+
"loss": 0.4511,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.27,
|
| 342 |
+
"learning_rate": 1.7264677424083977e-05,
|
| 343 |
+
"loss": 0.4477,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.28,
|
| 348 |
+
"learning_rate": 1.7215809689494418e-05,
|
| 349 |
+
"loss": 0.4423,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.28,
|
| 354 |
+
"learning_rate": 1.7166941954904855e-05,
|
| 355 |
+
"loss": 0.4582,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.29,
|
| 360 |
+
"learning_rate": 1.7118074220315296e-05,
|
| 361 |
+
"loss": 0.4436,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.29,
|
| 366 |
+
"learning_rate": 1.7069206485725737e-05,
|
| 367 |
+
"loss": 0.4645,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.3,
|
| 372 |
+
"learning_rate": 1.7020338751136177e-05,
|
| 373 |
+
"loss": 0.455,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.3,
|
| 378 |
+
"learning_rate": 1.6971471016546618e-05,
|
| 379 |
+
"loss": 0.455,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.31,
|
| 384 |
+
"learning_rate": 1.6922701017426235e-05,
|
| 385 |
+
"loss": 0.4532,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.31,
|
| 390 |
+
"learning_rate": 1.6873833282836676e-05,
|
| 391 |
+
"loss": 0.4467,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.32,
|
| 396 |
+
"learning_rate": 1.6824965548247117e-05,
|
| 397 |
+
"loss": 0.449,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.32,
|
| 402 |
+
"learning_rate": 1.6776097813657554e-05,
|
| 403 |
+
"loss": 0.4515,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.33,
|
| 408 |
+
"learning_rate": 1.6727327814537174e-05,
|
| 409 |
+
"loss": 0.4522,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.33,
|
| 414 |
+
"learning_rate": 1.6678460079947615e-05,
|
| 415 |
+
"loss": 0.4407,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.34,
|
| 420 |
+
"learning_rate": 1.6629592345358056e-05,
|
| 421 |
+
"loss": 0.452,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.34,
|
| 426 |
+
"learning_rate": 1.6580724610768497e-05,
|
| 427 |
+
"loss": 0.4465,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.35,
|
| 432 |
+
"learning_rate": 1.6531856876178934e-05,
|
| 433 |
+
"loss": 0.4485,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.35,
|
| 438 |
+
"learning_rate": 1.6483086877058554e-05,
|
| 439 |
+
"loss": 0.452,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.36,
|
| 444 |
+
"learning_rate": 1.6434316877938174e-05,
|
| 445 |
+
"loss": 0.4326,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.36,
|
| 450 |
+
"learning_rate": 1.638544914334861e-05,
|
| 451 |
+
"loss": 0.4724,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.37,
|
| 456 |
+
"learning_rate": 1.6336581408759052e-05,
|
| 457 |
+
"loss": 0.4384,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.37,
|
| 462 |
+
"learning_rate": 1.6287713674169493e-05,
|
| 463 |
+
"loss": 0.444,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.38,
|
| 468 |
+
"learning_rate": 1.6238845939579934e-05,
|
| 469 |
+
"loss": 0.4541,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.38,
|
| 474 |
+
"learning_rate": 1.6189978204990375e-05,
|
| 475 |
+
"loss": 0.4377,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.39,
|
| 480 |
+
"learning_rate": 1.6141110470400812e-05,
|
| 481 |
+
"loss": 0.4471,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.39,
|
| 486 |
+
"learning_rate": 1.6092242735811253e-05,
|
| 487 |
+
"loss": 0.4452,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.4,
|
| 492 |
+
"learning_rate": 1.6043375001221694e-05,
|
| 493 |
+
"loss": 0.4375,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.4,
|
| 498 |
+
"learning_rate": 1.5994702737570493e-05,
|
| 499 |
+
"loss": 0.4672,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.41,
|
| 504 |
+
"learning_rate": 1.5945835002980934e-05,
|
| 505 |
+
"loss": 0.4423,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.41,
|
| 510 |
+
"learning_rate": 1.589696726839137e-05,
|
| 511 |
+
"loss": 0.4521,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.42,
|
| 516 |
+
"learning_rate": 1.5848099533801812e-05,
|
| 517 |
+
"loss": 0.4389,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.42,
|
| 522 |
+
"learning_rate": 1.5799231799212253e-05,
|
| 523 |
+
"loss": 0.4319,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.43,
|
| 528 |
+
"learning_rate": 1.5750461800091873e-05,
|
| 529 |
+
"loss": 0.4591,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.43,
|
| 534 |
+
"learning_rate": 1.5701594065502314e-05,
|
| 535 |
+
"loss": 0.4435,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.43,
|
| 540 |
+
"learning_rate": 1.5652726330912755e-05,
|
| 541 |
+
"loss": 0.4574,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.44,
|
| 546 |
+
"learning_rate": 1.5603858596323192e-05,
|
| 547 |
+
"loss": 0.4509,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.44,
|
| 552 |
+
"learning_rate": 1.5554990861733633e-05,
|
| 553 |
+
"loss": 0.4493,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.45,
|
| 558 |
+
"learning_rate": 1.5506123127144074e-05,
|
| 559 |
+
"loss": 0.4549,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.45,
|
| 564 |
+
"learning_rate": 1.5457255392554515e-05,
|
| 565 |
+
"loss": 0.4419,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.46,
|
| 570 |
+
"learning_rate": 1.5408387657964955e-05,
|
| 571 |
+
"loss": 0.4512,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.46,
|
| 576 |
+
"learning_rate": 1.5359617658844572e-05,
|
| 577 |
+
"loss": 0.4445,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.47,
|
| 582 |
+
"learning_rate": 1.5310749924255013e-05,
|
| 583 |
+
"loss": 0.4405,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.47,
|
| 588 |
+
"learning_rate": 1.5261882189665454e-05,
|
| 589 |
+
"loss": 0.4507,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.48,
|
| 594 |
+
"learning_rate": 1.5213014455075891e-05,
|
| 595 |
+
"loss": 0.444,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.48,
|
| 600 |
+
"learning_rate": 1.5164146720486332e-05,
|
| 601 |
+
"loss": 0.4433,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.49,
|
| 606 |
+
"learning_rate": 1.5115376721365954e-05,
|
| 607 |
+
"loss": 0.4503,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.49,
|
| 612 |
+
"learning_rate": 1.5066606722245572e-05,
|
| 613 |
+
"loss": 0.4461,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.5,
|
| 618 |
+
"learning_rate": 1.5017738987656011e-05,
|
| 619 |
+
"loss": 0.4413,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.5,
|
| 624 |
+
"learning_rate": 1.4968871253066452e-05,
|
| 625 |
+
"loss": 0.4426,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.51,
|
| 630 |
+
"learning_rate": 1.4920003518476891e-05,
|
| 631 |
+
"loss": 0.4527,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.51,
|
| 636 |
+
"learning_rate": 1.4871135783887332e-05,
|
| 637 |
+
"loss": 0.4545,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.52,
|
| 642 |
+
"learning_rate": 1.4822268049297773e-05,
|
| 643 |
+
"loss": 0.4474,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.52,
|
| 648 |
+
"learning_rate": 1.4773400314708214e-05,
|
| 649 |
+
"loss": 0.4571,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.53,
|
| 654 |
+
"learning_rate": 1.4724532580118651e-05,
|
| 655 |
+
"loss": 0.4419,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.53,
|
| 660 |
+
"learning_rate": 1.4675762580998271e-05,
|
| 661 |
+
"loss": 0.4578,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.54,
|
| 666 |
+
"learning_rate": 1.4626894846408712e-05,
|
| 667 |
+
"loss": 0.4424,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.54,
|
| 672 |
+
"learning_rate": 1.4578027111819151e-05,
|
| 673 |
+
"loss": 0.4489,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.55,
|
| 678 |
+
"learning_rate": 1.4529159377229592e-05,
|
| 679 |
+
"loss": 0.44,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.55,
|
| 684 |
+
"learning_rate": 1.4480291642640033e-05,
|
| 685 |
+
"loss": 0.4619,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.56,
|
| 690 |
+
"learning_rate": 1.4431423908050473e-05,
|
| 691 |
+
"loss": 0.4314,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.56,
|
| 696 |
+
"learning_rate": 1.438255617346091e-05,
|
| 697 |
+
"loss": 0.4403,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.57,
|
| 702 |
+
"learning_rate": 1.4333688438871352e-05,
|
| 703 |
+
"loss": 0.4494,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.57,
|
| 708 |
+
"learning_rate": 1.4284820704281792e-05,
|
| 709 |
+
"loss": 0.4277,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.58,
|
| 714 |
+
"learning_rate": 1.423614844063059e-05,
|
| 715 |
+
"loss": 0.4338,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.58,
|
| 720 |
+
"learning_rate": 1.4187280706041031e-05,
|
| 721 |
+
"loss": 0.4387,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.59,
|
| 726 |
+
"learning_rate": 1.413841297145147e-05,
|
| 727 |
+
"loss": 0.4401,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.59,
|
| 732 |
+
"learning_rate": 1.408954523686191e-05,
|
| 733 |
+
"loss": 0.4518,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.6,
|
| 738 |
+
"learning_rate": 1.404067750227235e-05,
|
| 739 |
+
"loss": 0.4323,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.6,
|
| 744 |
+
"learning_rate": 1.399180976768279e-05,
|
| 745 |
+
"loss": 0.4371,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.61,
|
| 750 |
+
"learning_rate": 1.3942942033093231e-05,
|
| 751 |
+
"loss": 0.4495,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.61,
|
| 756 |
+
"learning_rate": 1.389407429850367e-05,
|
| 757 |
+
"loss": 0.4479,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.62,
|
| 762 |
+
"learning_rate": 1.3845206563914111e-05,
|
| 763 |
+
"loss": 0.4465,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.62,
|
| 768 |
+
"learning_rate": 1.3796436564793732e-05,
|
| 769 |
+
"loss": 0.4371,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.63,
|
| 774 |
+
"learning_rate": 1.374756883020417e-05,
|
| 775 |
+
"loss": 0.4572,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.63,
|
| 780 |
+
"learning_rate": 1.369870109561461e-05,
|
| 781 |
+
"loss": 0.444,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.64,
|
| 786 |
+
"learning_rate": 1.3649931096494228e-05,
|
| 787 |
+
"loss": 0.4338,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.64,
|
| 792 |
+
"learning_rate": 1.3601063361904669e-05,
|
| 793 |
+
"loss": 0.4437,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.65,
|
| 798 |
+
"learning_rate": 1.355219562731511e-05,
|
| 799 |
+
"loss": 0.4385,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.65,
|
| 804 |
+
"learning_rate": 1.350332789272555e-05,
|
| 805 |
+
"loss": 0.4353,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.65,
|
| 810 |
+
"learning_rate": 1.3454460158135991e-05,
|
| 811 |
+
"loss": 0.4438,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.66,
|
| 816 |
+
"learning_rate": 1.340559242354643e-05,
|
| 817 |
+
"loss": 0.4386,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.66,
|
| 822 |
+
"learning_rate": 1.335672468895687e-05,
|
| 823 |
+
"loss": 0.4412,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.67,
|
| 828 |
+
"learning_rate": 1.330785695436731e-05,
|
| 829 |
+
"loss": 0.4341,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.67,
|
| 834 |
+
"learning_rate": 1.3259086955246929e-05,
|
| 835 |
+
"loss": 0.4322,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.68,
|
| 840 |
+
"learning_rate": 1.321021922065737e-05,
|
| 841 |
+
"loss": 0.4344,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.68,
|
| 846 |
+
"learning_rate": 1.316135148606781e-05,
|
| 847 |
+
"loss": 0.4366,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.69,
|
| 852 |
+
"learning_rate": 1.3112483751478251e-05,
|
| 853 |
+
"loss": 0.4513,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.69,
|
| 858 |
+
"learning_rate": 1.3063616016888689e-05,
|
| 859 |
+
"loss": 0.4469,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.7,
|
| 864 |
+
"learning_rate": 1.301484601776831e-05,
|
| 865 |
+
"loss": 0.4464,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.7,
|
| 870 |
+
"learning_rate": 1.2965978283178748e-05,
|
| 871 |
+
"loss": 0.4438,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.71,
|
| 876 |
+
"learning_rate": 1.2917110548589189e-05,
|
| 877 |
+
"loss": 0.4451,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.71,
|
| 882 |
+
"learning_rate": 1.286824281399963e-05,
|
| 883 |
+
"loss": 0.4447,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.72,
|
| 888 |
+
"learning_rate": 1.281937507941007e-05,
|
| 889 |
+
"loss": 0.4327,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.72,
|
| 894 |
+
"learning_rate": 1.2770605080289689e-05,
|
| 895 |
+
"loss": 0.433,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.73,
|
| 900 |
+
"learning_rate": 1.272173734570013e-05,
|
| 901 |
+
"loss": 0.4266,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.73,
|
| 906 |
+
"learning_rate": 1.2672967346579748e-05,
|
| 907 |
+
"loss": 0.4156,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.74,
|
| 912 |
+
"learning_rate": 1.2624099611990189e-05,
|
| 913 |
+
"loss": 0.4401,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.74,
|
| 918 |
+
"learning_rate": 1.257523187740063e-05,
|
| 919 |
+
"loss": 0.4513,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.75,
|
| 924 |
+
"learning_rate": 1.2526364142811069e-05,
|
| 925 |
+
"loss": 0.4493,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.75,
|
| 930 |
+
"learning_rate": 1.2477496408221508e-05,
|
| 931 |
+
"loss": 0.4425,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.76,
|
| 936 |
+
"learning_rate": 1.2428628673631948e-05,
|
| 937 |
+
"loss": 0.4364,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.76,
|
| 942 |
+
"learning_rate": 1.2379858674511569e-05,
|
| 943 |
+
"loss": 0.4314,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.77,
|
| 948 |
+
"learning_rate": 1.2330990939922008e-05,
|
| 949 |
+
"loss": 0.4427,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.77,
|
| 954 |
+
"learning_rate": 1.2282123205332448e-05,
|
| 955 |
+
"loss": 0.4376,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.78,
|
| 960 |
+
"learning_rate": 1.223325547074289e-05,
|
| 961 |
+
"loss": 0.4405,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.78,
|
| 966 |
+
"learning_rate": 1.2184387736153328e-05,
|
| 967 |
+
"loss": 0.4363,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.79,
|
| 972 |
+
"learning_rate": 1.2135520001563767e-05,
|
| 973 |
+
"loss": 0.4502,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.79,
|
| 978 |
+
"learning_rate": 1.2086652266974208e-05,
|
| 979 |
+
"loss": 0.4607,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.8,
|
| 984 |
+
"learning_rate": 1.2037784532384649e-05,
|
| 985 |
+
"loss": 0.4275,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.8,
|
| 990 |
+
"learning_rate": 1.198891679779509e-05,
|
| 991 |
+
"loss": 0.4183,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.81,
|
| 996 |
+
"learning_rate": 1.1940244534143888e-05,
|
| 997 |
+
"loss": 0.4324,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.81,
|
| 1002 |
+
"learning_rate": 1.1891376799554328e-05,
|
| 1003 |
+
"loss": 0.4302,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.82,
|
| 1008 |
+
"learning_rate": 1.1842509064964768e-05,
|
| 1009 |
+
"loss": 0.431,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.82,
|
| 1014 |
+
"learning_rate": 1.1793641330375207e-05,
|
| 1015 |
+
"loss": 0.4311,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.83,
|
| 1020 |
+
"learning_rate": 1.1744773595785647e-05,
|
| 1021 |
+
"loss": 0.4474,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.83,
|
| 1026 |
+
"learning_rate": 1.1695905861196088e-05,
|
| 1027 |
+
"loss": 0.4474,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.84,
|
| 1032 |
+
"learning_rate": 1.1647135862075707e-05,
|
| 1033 |
+
"loss": 0.4345,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.84,
|
| 1038 |
+
"learning_rate": 1.1598268127486147e-05,
|
| 1039 |
+
"loss": 0.4389,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.85,
|
| 1044 |
+
"learning_rate": 1.1549400392896588e-05,
|
| 1045 |
+
"loss": 0.4381,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.85,
|
| 1050 |
+
"learning_rate": 1.1500532658307026e-05,
|
| 1051 |
+
"loss": 0.4461,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.86,
|
| 1056 |
+
"learning_rate": 1.1451664923717466e-05,
|
| 1057 |
+
"loss": 0.4335,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.86,
|
| 1062 |
+
"learning_rate": 1.1402797189127907e-05,
|
| 1063 |
+
"loss": 0.434,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.86,
|
| 1068 |
+
"learning_rate": 1.1354027190007526e-05,
|
| 1069 |
+
"loss": 0.4365,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.87,
|
| 1074 |
+
"learning_rate": 1.1305159455417966e-05,
|
| 1075 |
+
"loss": 0.4269,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.87,
|
| 1080 |
+
"learning_rate": 1.1256291720828407e-05,
|
| 1081 |
+
"loss": 0.4346,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.88,
|
| 1086 |
+
"learning_rate": 1.1207521721708026e-05,
|
| 1087 |
+
"loss": 0.43,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.88,
|
| 1092 |
+
"learning_rate": 1.1158653987118467e-05,
|
| 1093 |
+
"loss": 0.4527,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.89,
|
| 1098 |
+
"learning_rate": 1.1109786252528907e-05,
|
| 1099 |
+
"loss": 0.4367,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.89,
|
| 1104 |
+
"learning_rate": 1.1060918517939345e-05,
|
| 1105 |
+
"loss": 0.4387,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.9,
|
| 1110 |
+
"learning_rate": 1.1012050783349786e-05,
|
| 1111 |
+
"loss": 0.4349,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.9,
|
| 1116 |
+
"learning_rate": 1.0963183048760226e-05,
|
| 1117 |
+
"loss": 0.4428,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.91,
|
| 1122 |
+
"learning_rate": 1.0914315314170667e-05,
|
| 1123 |
+
"loss": 0.4277,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.91,
|
| 1128 |
+
"learning_rate": 1.0865447579581108e-05,
|
| 1129 |
+
"loss": 0.4343,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.92,
|
| 1134 |
+
"learning_rate": 1.0816579844991545e-05,
|
| 1135 |
+
"loss": 0.435,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.92,
|
| 1140 |
+
"learning_rate": 1.0767712110401986e-05,
|
| 1141 |
+
"loss": 0.4335,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.93,
|
| 1146 |
+
"learning_rate": 1.0718844375812427e-05,
|
| 1147 |
+
"loss": 0.437,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.93,
|
| 1152 |
+
"learning_rate": 1.0669976641222868e-05,
|
| 1153 |
+
"loss": 0.4389,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.94,
|
| 1158 |
+
"learning_rate": 1.0621108906633307e-05,
|
| 1159 |
+
"loss": 0.4492,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.94,
|
| 1164 |
+
"learning_rate": 1.0572241172043748e-05,
|
| 1165 |
+
"loss": 0.4287,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.95,
|
| 1170 |
+
"learning_rate": 1.0523471172923368e-05,
|
| 1171 |
+
"loss": 0.4399,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.95,
|
| 1176 |
+
"learning_rate": 1.0474603438333805e-05,
|
| 1177 |
+
"loss": 0.431,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.96,
|
| 1182 |
+
"learning_rate": 1.0425833439213427e-05,
|
| 1183 |
+
"loss": 0.445,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.96,
|
| 1188 |
+
"learning_rate": 1.0376965704623868e-05,
|
| 1189 |
+
"loss": 0.4412,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.97,
|
| 1194 |
+
"learning_rate": 1.0328097970034305e-05,
|
| 1195 |
+
"loss": 0.4237,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.97,
|
| 1200 |
+
"learning_rate": 1.0279230235444746e-05,
|
| 1201 |
+
"loss": 0.4413,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.98,
|
| 1206 |
+
"learning_rate": 1.0230362500855187e-05,
|
| 1207 |
+
"loss": 0.4468,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.98,
|
| 1212 |
+
"learning_rate": 1.0181592501734805e-05,
|
| 1213 |
+
"loss": 0.4386,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.99,
|
| 1218 |
+
"learning_rate": 1.0132724767145246e-05,
|
| 1219 |
+
"loss": 0.4373,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.99,
|
| 1224 |
+
"learning_rate": 1.0083857032555687e-05,
|
| 1225 |
+
"loss": 0.427,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.0,
|
| 1230 |
+
"learning_rate": 1.0034989297966128e-05,
|
| 1231 |
+
"loss": 0.4268,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 1.0,
|
| 1236 |
+
"eval_accuracy": 0.8036805258955265,
|
| 1237 |
+
"eval_loss": 0.4389118552207947,
|
| 1238 |
+
"eval_runtime": 205.8988,
|
| 1239 |
+
"eval_samples_per_second": 261.541,
|
| 1240 |
+
"eval_steps_per_second": 26.158,
|
| 1241 |
+
"step": 102317
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.0,
|
| 1245 |
+
"learning_rate": 9.986121563376567e-06,
|
| 1246 |
+
"loss": 0.4179,
|
| 1247 |
+
"step": 102500
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 1.01,
|
| 1251 |
+
"learning_rate": 9.937253828787006e-06,
|
| 1252 |
+
"loss": 0.4299,
|
| 1253 |
+
"step": 103000
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 1.01,
|
| 1257 |
+
"learning_rate": 9.888483829666626e-06,
|
| 1258 |
+
"loss": 0.4186,
|
| 1259 |
+
"step": 103500
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 1.02,
|
| 1263 |
+
"learning_rate": 9.839616095077065e-06,
|
| 1264 |
+
"loss": 0.4088,
|
| 1265 |
+
"step": 104000
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 1.02,
|
| 1269 |
+
"learning_rate": 9.790748360487506e-06,
|
| 1270 |
+
"loss": 0.4214,
|
| 1271 |
+
"step": 104500
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.03,
|
| 1275 |
+
"learning_rate": 9.741880625897947e-06,
|
| 1276 |
+
"loss": 0.4291,
|
| 1277 |
+
"step": 105000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.03,
|
| 1281 |
+
"learning_rate": 9.693110626777565e-06,
|
| 1282 |
+
"loss": 0.4168,
|
| 1283 |
+
"step": 105500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.04,
|
| 1287 |
+
"learning_rate": 9.644242892188004e-06,
|
| 1288 |
+
"loss": 0.4225,
|
| 1289 |
+
"step": 106000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.04,
|
| 1293 |
+
"learning_rate": 9.595375157598445e-06,
|
| 1294 |
+
"loss": 0.4296,
|
| 1295 |
+
"step": 106500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 1.05,
|
| 1299 |
+
"learning_rate": 9.546507423008886e-06,
|
| 1300 |
+
"loss": 0.4075,
|
| 1301 |
+
"step": 107000
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 1.05,
|
| 1305 |
+
"learning_rate": 9.497639688419325e-06,
|
| 1306 |
+
"loss": 0.4122,
|
| 1307 |
+
"step": 107500
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 1.06,
|
| 1311 |
+
"learning_rate": 9.448869689298943e-06,
|
| 1312 |
+
"loss": 0.4243,
|
| 1313 |
+
"step": 108000
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.06,
|
| 1317 |
+
"learning_rate": 9.400001954709384e-06,
|
| 1318 |
+
"loss": 0.4082,
|
| 1319 |
+
"step": 108500
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.07,
|
| 1323 |
+
"learning_rate": 9.351134220119825e-06,
|
| 1324 |
+
"loss": 0.4196,
|
| 1325 |
+
"step": 109000
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 1.07,
|
| 1329 |
+
"learning_rate": 9.302266485530264e-06,
|
| 1330 |
+
"loss": 0.3987,
|
| 1331 |
+
"step": 109500
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.08,
|
| 1335 |
+
"learning_rate": 9.253398750940705e-06,
|
| 1336 |
+
"loss": 0.4021,
|
| 1337 |
+
"step": 110000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 1.08,
|
| 1341 |
+
"learning_rate": 9.204531016351146e-06,
|
| 1342 |
+
"loss": 0.4194,
|
| 1343 |
+
"step": 110500
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.08,
|
| 1347 |
+
"learning_rate": 9.155761017230764e-06,
|
| 1348 |
+
"loss": 0.4157,
|
| 1349 |
+
"step": 111000
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 1.09,
|
| 1353 |
+
"learning_rate": 9.106893282641203e-06,
|
| 1354 |
+
"loss": 0.4193,
|
| 1355 |
+
"step": 111500
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 1.09,
|
| 1359 |
+
"learning_rate": 9.058025548051644e-06,
|
| 1360 |
+
"loss": 0.4124,
|
| 1361 |
+
"step": 112000
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.1,
|
| 1365 |
+
"learning_rate": 9.009157813462085e-06,
|
| 1366 |
+
"loss": 0.415,
|
| 1367 |
+
"step": 112500
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 1.1,
|
| 1371 |
+
"learning_rate": 8.960387814341703e-06,
|
| 1372 |
+
"loss": 0.4145,
|
| 1373 |
+
"step": 113000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.11,
|
| 1377 |
+
"learning_rate": 8.911520079752144e-06,
|
| 1378 |
+
"loss": 0.4291,
|
| 1379 |
+
"step": 113500
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 1.11,
|
| 1383 |
+
"learning_rate": 8.862652345162583e-06,
|
| 1384 |
+
"loss": 0.4243,
|
| 1385 |
+
"step": 114000
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 1.12,
|
| 1389 |
+
"learning_rate": 8.813784610573024e-06,
|
| 1390 |
+
"loss": 0.4207,
|
| 1391 |
+
"step": 114500
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 1.12,
|
| 1395 |
+
"learning_rate": 8.764916875983463e-06,
|
| 1396 |
+
"loss": 0.4221,
|
| 1397 |
+
"step": 115000
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 1.13,
|
| 1401 |
+
"learning_rate": 8.716049141393904e-06,
|
| 1402 |
+
"loss": 0.4155,
|
| 1403 |
+
"step": 115500
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.13,
|
| 1407 |
+
"learning_rate": 8.667181406804343e-06,
|
| 1408 |
+
"loss": 0.423,
|
| 1409 |
+
"step": 116000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.14,
|
| 1413 |
+
"learning_rate": 8.618313672214784e-06,
|
| 1414 |
+
"loss": 0.4094,
|
| 1415 |
+
"step": 116500
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 1.14,
|
| 1419 |
+
"learning_rate": 8.569543673094404e-06,
|
| 1420 |
+
"loss": 0.4177,
|
| 1421 |
+
"step": 117000
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 1.15,
|
| 1425 |
+
"learning_rate": 8.520675938504843e-06,
|
| 1426 |
+
"loss": 0.4246,
|
| 1427 |
+
"step": 117500
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.15,
|
| 1431 |
+
"learning_rate": 8.471808203915284e-06,
|
| 1432 |
+
"loss": 0.4039,
|
| 1433 |
+
"step": 118000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 1.16,
|
| 1437 |
+
"learning_rate": 8.422940469325723e-06,
|
| 1438 |
+
"loss": 0.4124,
|
| 1439 |
+
"step": 118500
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 1.16,
|
| 1443 |
+
"learning_rate": 8.374170470205343e-06,
|
| 1444 |
+
"loss": 0.4124,
|
| 1445 |
+
"step": 119000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.17,
|
| 1449 |
+
"learning_rate": 8.325302735615784e-06,
|
| 1450 |
+
"loss": 0.4155,
|
| 1451 |
+
"step": 119500
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.17,
|
| 1455 |
+
"learning_rate": 8.276435001026223e-06,
|
| 1456 |
+
"loss": 0.4112,
|
| 1457 |
+
"step": 120000
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 1.18,
|
| 1461 |
+
"learning_rate": 8.227567266436664e-06,
|
| 1462 |
+
"loss": 0.4106,
|
| 1463 |
+
"step": 120500
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 1.18,
|
| 1467 |
+
"learning_rate": 8.178699531847103e-06,
|
| 1468 |
+
"loss": 0.4155,
|
| 1469 |
+
"step": 121000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.19,
|
| 1473 |
+
"learning_rate": 8.129831797257543e-06,
|
| 1474 |
+
"loss": 0.4275,
|
| 1475 |
+
"step": 121500
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 1.19,
|
| 1479 |
+
"learning_rate": 8.080964062667983e-06,
|
| 1480 |
+
"loss": 0.4236,
|
| 1481 |
+
"step": 122000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 1.2,
|
| 1485 |
+
"learning_rate": 8.032096328078423e-06,
|
| 1486 |
+
"loss": 0.4175,
|
| 1487 |
+
"step": 122500
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.2,
|
| 1491 |
+
"learning_rate": 7.983424064427223e-06,
|
| 1492 |
+
"loss": 0.4287,
|
| 1493 |
+
"step": 123000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.21,
|
| 1497 |
+
"learning_rate": 7.934556329837662e-06,
|
| 1498 |
+
"loss": 0.4299,
|
| 1499 |
+
"step": 123500
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 1.21,
|
| 1503 |
+
"learning_rate": 7.885688595248103e-06,
|
| 1504 |
+
"loss": 0.4274,
|
| 1505 |
+
"step": 124000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 1.22,
|
| 1509 |
+
"learning_rate": 7.836820860658543e-06,
|
| 1510 |
+
"loss": 0.4129,
|
| 1511 |
+
"step": 124500
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 1.22,
|
| 1515 |
+
"learning_rate": 7.787953126068983e-06,
|
| 1516 |
+
"loss": 0.4007,
|
| 1517 |
+
"step": 125000
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 1.23,
|
| 1521 |
+
"learning_rate": 7.73928086241778e-06,
|
| 1522 |
+
"loss": 0.4213,
|
| 1523 |
+
"step": 125500
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 1.23,
|
| 1527 |
+
"learning_rate": 7.690413127828221e-06,
|
| 1528 |
+
"loss": 0.411,
|
| 1529 |
+
"step": 126000
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.24,
|
| 1533 |
+
"learning_rate": 7.64154539323866e-06,
|
| 1534 |
+
"loss": 0.4191,
|
| 1535 |
+
"step": 126500
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.24,
|
| 1539 |
+
"learning_rate": 7.592677658649101e-06,
|
| 1540 |
+
"loss": 0.4097,
|
| 1541 |
+
"step": 127000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.25,
|
| 1545 |
+
"learning_rate": 7.543809924059541e-06,
|
| 1546 |
+
"loss": 0.414,
|
| 1547 |
+
"step": 127500
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 1.25,
|
| 1551 |
+
"learning_rate": 7.494942189469982e-06,
|
| 1552 |
+
"loss": 0.4155,
|
| 1553 |
+
"step": 128000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 1.26,
|
| 1557 |
+
"learning_rate": 7.446074454880421e-06,
|
| 1558 |
+
"loss": 0.4054,
|
| 1559 |
+
"step": 128500
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 1.26,
|
| 1563 |
+
"learning_rate": 7.397206720290862e-06,
|
| 1564 |
+
"loss": 0.4308,
|
| 1565 |
+
"step": 129000
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 1.27,
|
| 1569 |
+
"learning_rate": 7.348338985701301e-06,
|
| 1570 |
+
"loss": 0.3952,
|
| 1571 |
+
"step": 129500
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.27,
|
| 1575 |
+
"learning_rate": 7.299471251111742e-06,
|
| 1576 |
+
"loss": 0.4232,
|
| 1577 |
+
"step": 130000
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.28,
|
| 1581 |
+
"learning_rate": 7.250701251991361e-06,
|
| 1582 |
+
"loss": 0.4145,
|
| 1583 |
+
"step": 130500
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.28,
|
| 1587 |
+
"learning_rate": 7.201833517401801e-06,
|
| 1588 |
+
"loss": 0.4133,
|
| 1589 |
+
"step": 131000
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 1.29,
|
| 1593 |
+
"learning_rate": 7.152965782812241e-06,
|
| 1594 |
+
"loss": 0.4265,
|
| 1595 |
+
"step": 131500
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.29,
|
| 1599 |
+
"learning_rate": 7.104098048222681e-06,
|
| 1600 |
+
"loss": 0.4081,
|
| 1601 |
+
"step": 132000
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.29,
|
| 1605 |
+
"learning_rate": 7.0552303136331215e-06,
|
| 1606 |
+
"loss": 0.416,
|
| 1607 |
+
"step": 132500
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.3,
|
| 1611 |
+
"learning_rate": 7.006362579043561e-06,
|
| 1612 |
+
"loss": 0.4321,
|
| 1613 |
+
"step": 133000
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.3,
|
| 1617 |
+
"learning_rate": 6.957494844454001e-06,
|
| 1618 |
+
"loss": 0.4028,
|
| 1619 |
+
"step": 133500
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.31,
|
| 1623 |
+
"learning_rate": 6.908627109864442e-06,
|
| 1624 |
+
"loss": 0.4157,
|
| 1625 |
+
"step": 134000
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.31,
|
| 1629 |
+
"learning_rate": 6.859857110744061e-06,
|
| 1630 |
+
"loss": 0.4027,
|
| 1631 |
+
"step": 134500
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.32,
|
| 1635 |
+
"learning_rate": 6.810989376154501e-06,
|
| 1636 |
+
"loss": 0.4226,
|
| 1637 |
+
"step": 135000
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 1.32,
|
| 1641 |
+
"learning_rate": 6.762219377034121e-06,
|
| 1642 |
+
"loss": 0.4146,
|
| 1643 |
+
"step": 135500
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 1.33,
|
| 1647 |
+
"learning_rate": 6.71335164244456e-06,
|
| 1648 |
+
"loss": 0.4064,
|
| 1649 |
+
"step": 136000
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 1.33,
|
| 1653 |
+
"learning_rate": 6.664483907855001e-06,
|
| 1654 |
+
"loss": 0.405,
|
| 1655 |
+
"step": 136500
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.34,
|
| 1659 |
+
"learning_rate": 6.61561617326544e-06,
|
| 1660 |
+
"loss": 0.4066,
|
| 1661 |
+
"step": 137000
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.34,
|
| 1665 |
+
"learning_rate": 6.5667484386758805e-06,
|
| 1666 |
+
"loss": 0.4113,
|
| 1667 |
+
"step": 137500
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.35,
|
| 1671 |
+
"learning_rate": 6.5178807040863205e-06,
|
| 1672 |
+
"loss": 0.4167,
|
| 1673 |
+
"step": 138000
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.35,
|
| 1677 |
+
"learning_rate": 6.4690129694967604e-06,
|
| 1678 |
+
"loss": 0.4161,
|
| 1679 |
+
"step": 138500
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 1.36,
|
| 1683 |
+
"learning_rate": 6.4201452349072e-06,
|
| 1684 |
+
"loss": 0.4206,
|
| 1685 |
+
"step": 139000
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 1.36,
|
| 1689 |
+
"learning_rate": 6.37137523578682e-06,
|
| 1690 |
+
"loss": 0.4207,
|
| 1691 |
+
"step": 139500
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.37,
|
| 1695 |
+
"learning_rate": 6.3225075011972604e-06,
|
| 1696 |
+
"loss": 0.4107,
|
| 1697 |
+
"step": 140000
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.37,
|
| 1701 |
+
"learning_rate": 6.2736397666076996e-06,
|
| 1702 |
+
"loss": 0.4209,
|
| 1703 |
+
"step": 140500
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.38,
|
| 1707 |
+
"learning_rate": 6.22477203201814e-06,
|
| 1708 |
+
"loss": 0.4029,
|
| 1709 |
+
"step": 141000
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.38,
|
| 1713 |
+
"learning_rate": 6.17600203289776e-06,
|
| 1714 |
+
"loss": 0.4245,
|
| 1715 |
+
"step": 141500
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.39,
|
| 1719 |
+
"learning_rate": 6.1271342983081996e-06,
|
| 1720 |
+
"loss": 0.4141,
|
| 1721 |
+
"step": 142000
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 1.39,
|
| 1725 |
+
"learning_rate": 6.07826656371864e-06,
|
| 1726 |
+
"loss": 0.4195,
|
| 1727 |
+
"step": 142500
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 1.4,
|
| 1731 |
+
"learning_rate": 6.0293988291290795e-06,
|
| 1732 |
+
"loss": 0.4014,
|
| 1733 |
+
"step": 143000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.4,
|
| 1737 |
+
"learning_rate": 5.98053109453952e-06,
|
| 1738 |
+
"loss": 0.4197,
|
| 1739 |
+
"step": 143500
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.41,
|
| 1743 |
+
"learning_rate": 5.9317610954191396e-06,
|
| 1744 |
+
"loss": 0.4165,
|
| 1745 |
+
"step": 144000
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.41,
|
| 1749 |
+
"learning_rate": 5.882893360829579e-06,
|
| 1750 |
+
"loss": 0.3974,
|
| 1751 |
+
"step": 144500
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.42,
|
| 1755 |
+
"learning_rate": 5.8340256262400194e-06,
|
| 1756 |
+
"loss": 0.4293,
|
| 1757 |
+
"step": 145000
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.42,
|
| 1761 |
+
"learning_rate": 5.785157891650459e-06,
|
| 1762 |
+
"loss": 0.4129,
|
| 1763 |
+
"step": 145500
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 1.43,
|
| 1767 |
+
"learning_rate": 5.736290157060899e-06,
|
| 1768 |
+
"loss": 0.4277,
|
| 1769 |
+
"step": 146000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 1.43,
|
| 1773 |
+
"learning_rate": 5.687422422471339e-06,
|
| 1774 |
+
"loss": 0.4153,
|
| 1775 |
+
"step": 146500
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 1.44,
|
| 1779 |
+
"learning_rate": 5.63855468788178e-06,
|
| 1780 |
+
"loss": 0.4126,
|
| 1781 |
+
"step": 147000
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.44,
|
| 1785 |
+
"learning_rate": 5.589686953292219e-06,
|
| 1786 |
+
"loss": 0.4172,
|
| 1787 |
+
"step": 147500
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.45,
|
| 1791 |
+
"learning_rate": 5.54081921870266e-06,
|
| 1792 |
+
"loss": 0.418,
|
| 1793 |
+
"step": 148000
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.45,
|
| 1797 |
+
"learning_rate": 5.492146955051459e-06,
|
| 1798 |
+
"loss": 0.4194,
|
| 1799 |
+
"step": 148500
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 1.46,
|
| 1803 |
+
"learning_rate": 5.443279220461898e-06,
|
| 1804 |
+
"loss": 0.42,
|
| 1805 |
+
"step": 149000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 1.46,
|
| 1809 |
+
"learning_rate": 5.3944114858723385e-06,
|
| 1810 |
+
"loss": 0.4172,
|
| 1811 |
+
"step": 149500
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 1.47,
|
| 1815 |
+
"learning_rate": 5.345543751282778e-06,
|
| 1816 |
+
"loss": 0.402,
|
| 1817 |
+
"step": 150000
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 1.47,
|
| 1821 |
+
"learning_rate": 5.296676016693218e-06,
|
| 1822 |
+
"loss": 0.4239,
|
| 1823 |
+
"step": 150500
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 1.48,
|
| 1827 |
+
"learning_rate": 5.247808282103659e-06,
|
| 1828 |
+
"loss": 0.4212,
|
| 1829 |
+
"step": 151000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.48,
|
| 1833 |
+
"learning_rate": 5.198940547514098e-06,
|
| 1834 |
+
"loss": 0.4168,
|
| 1835 |
+
"step": 151500
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 1.49,
|
| 1839 |
+
"learning_rate": 5.150072812924539e-06,
|
| 1840 |
+
"loss": 0.4188,
|
| 1841 |
+
"step": 152000
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 1.49,
|
| 1845 |
+
"learning_rate": 5.1013028138041575e-06,
|
| 1846 |
+
"loss": 0.4102,
|
| 1847 |
+
"step": 152500
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 1.5,
|
| 1851 |
+
"learning_rate": 5.052435079214598e-06,
|
| 1852 |
+
"loss": 0.4193,
|
| 1853 |
+
"step": 153000
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 1.5,
|
| 1857 |
+
"learning_rate": 5.003567344625039e-06,
|
| 1858 |
+
"loss": 0.4176,
|
| 1859 |
+
"step": 153500
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.51,
|
| 1863 |
+
"learning_rate": 4.954699610035478e-06,
|
| 1864 |
+
"loss": 0.413,
|
| 1865 |
+
"step": 154000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 1.51,
|
| 1869 |
+
"learning_rate": 4.9059296109150975e-06,
|
| 1870 |
+
"loss": 0.4258,
|
| 1871 |
+
"step": 154500
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.51,
|
| 1875 |
+
"learning_rate": 4.857159611794717e-06,
|
| 1876 |
+
"loss": 0.4175,
|
| 1877 |
+
"step": 155000
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 1.52,
|
| 1881 |
+
"learning_rate": 4.808291877205157e-06,
|
| 1882 |
+
"loss": 0.416,
|
| 1883 |
+
"step": 155500
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 1.52,
|
| 1887 |
+
"learning_rate": 4.759424142615597e-06,
|
| 1888 |
+
"loss": 0.4067,
|
| 1889 |
+
"step": 156000
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.53,
|
| 1893 |
+
"learning_rate": 4.710556408026037e-06,
|
| 1894 |
+
"loss": 0.43,
|
| 1895 |
+
"step": 156500
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 1.53,
|
| 1899 |
+
"learning_rate": 4.6616886734364774e-06,
|
| 1900 |
+
"loss": 0.4209,
|
| 1901 |
+
"step": 157000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 1.54,
|
| 1905 |
+
"learning_rate": 4.612820938846917e-06,
|
| 1906 |
+
"loss": 0.4202,
|
| 1907 |
+
"step": 157500
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 1.54,
|
| 1911 |
+
"learning_rate": 4.563953204257357e-06,
|
| 1912 |
+
"loss": 0.4096,
|
| 1913 |
+
"step": 158000
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 1.55,
|
| 1917 |
+
"learning_rate": 4.515183205136977e-06,
|
| 1918 |
+
"loss": 0.4143,
|
| 1919 |
+
"step": 158500
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 1.55,
|
| 1923 |
+
"learning_rate": 4.4663154705474166e-06,
|
| 1924 |
+
"loss": 0.4178,
|
| 1925 |
+
"step": 159000
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 1.56,
|
| 1929 |
+
"learning_rate": 4.4174477359578565e-06,
|
| 1930 |
+
"loss": 0.4139,
|
| 1931 |
+
"step": 159500
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 1.56,
|
| 1935 |
+
"learning_rate": 4.368580001368297e-06,
|
| 1936 |
+
"loss": 0.4003,
|
| 1937 |
+
"step": 160000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 1.57,
|
| 1941 |
+
"learning_rate": 4.319712266778737e-06,
|
| 1942 |
+
"loss": 0.4128,
|
| 1943 |
+
"step": 160500
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 1.57,
|
| 1947 |
+
"learning_rate": 4.270844532189177e-06,
|
| 1948 |
+
"loss": 0.4135,
|
| 1949 |
+
"step": 161000
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.58,
|
| 1953 |
+
"learning_rate": 4.221976797599617e-06,
|
| 1954 |
+
"loss": 0.4209,
|
| 1955 |
+
"step": 161500
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 1.58,
|
| 1959 |
+
"learning_rate": 4.173109063010057e-06,
|
| 1960 |
+
"loss": 0.4132,
|
| 1961 |
+
"step": 162000
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 1.59,
|
| 1965 |
+
"learning_rate": 4.124241328420497e-06,
|
| 1966 |
+
"loss": 0.4064,
|
| 1967 |
+
"step": 162500
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 1.59,
|
| 1971 |
+
"learning_rate": 4.075471329300117e-06,
|
| 1972 |
+
"loss": 0.408,
|
| 1973 |
+
"step": 163000
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 1.6,
|
| 1977 |
+
"learning_rate": 4.026701330179736e-06,
|
| 1978 |
+
"loss": 0.4103,
|
| 1979 |
+
"step": 163500
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 1.6,
|
| 1983 |
+
"learning_rate": 3.9778335955901756e-06,
|
| 1984 |
+
"loss": 0.4046,
|
| 1985 |
+
"step": 164000
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 1.61,
|
| 1989 |
+
"learning_rate": 3.928965861000616e-06,
|
| 1990 |
+
"loss": 0.4165,
|
| 1991 |
+
"step": 164500
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 1.61,
|
| 1995 |
+
"learning_rate": 3.880098126411056e-06,
|
| 1996 |
+
"loss": 0.4181,
|
| 1997 |
+
"step": 165000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.62,
|
| 2001 |
+
"learning_rate": 3.831230391821496e-06,
|
| 2002 |
+
"loss": 0.4118,
|
| 2003 |
+
"step": 165500
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 1.62,
|
| 2007 |
+
"learning_rate": 3.782362657231936e-06,
|
| 2008 |
+
"loss": 0.4114,
|
| 2009 |
+
"step": 166000
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 1.63,
|
| 2013 |
+
"learning_rate": 3.7334949226423766e-06,
|
| 2014 |
+
"loss": 0.4193,
|
| 2015 |
+
"step": 166500
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 1.63,
|
| 2019 |
+
"learning_rate": 3.6846271880528165e-06,
|
| 2020 |
+
"loss": 0.4277,
|
| 2021 |
+
"step": 167000
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 1.64,
|
| 2025 |
+
"learning_rate": 3.635857188932436e-06,
|
| 2026 |
+
"loss": 0.4072,
|
| 2027 |
+
"step": 167500
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 1.64,
|
| 2031 |
+
"learning_rate": 3.5869894543428758e-06,
|
| 2032 |
+
"loss": 0.4078,
|
| 2033 |
+
"step": 168000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 1.65,
|
| 2037 |
+
"learning_rate": 3.5381217197533157e-06,
|
| 2038 |
+
"loss": 0.4193,
|
| 2039 |
+
"step": 168500
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 1.65,
|
| 2043 |
+
"learning_rate": 3.4892539851637557e-06,
|
| 2044 |
+
"loss": 0.4092,
|
| 2045 |
+
"step": 169000
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 1.66,
|
| 2049 |
+
"learning_rate": 3.4404839860433754e-06,
|
| 2050 |
+
"loss": 0.4255,
|
| 2051 |
+
"step": 169500
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 1.66,
|
| 2055 |
+
"learning_rate": 3.3916162514538153e-06,
|
| 2056 |
+
"loss": 0.4015,
|
| 2057 |
+
"step": 170000
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 1.67,
|
| 2061 |
+
"learning_rate": 3.3427485168642553e-06,
|
| 2062 |
+
"loss": 0.4035,
|
| 2063 |
+
"step": 170500
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 1.67,
|
| 2067 |
+
"learning_rate": 3.2938807822746956e-06,
|
| 2068 |
+
"loss": 0.4145,
|
| 2069 |
+
"step": 171000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 1.68,
|
| 2073 |
+
"learning_rate": 3.2450130476851356e-06,
|
| 2074 |
+
"loss": 0.4107,
|
| 2075 |
+
"step": 171500
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 1.68,
|
| 2079 |
+
"learning_rate": 3.1961453130955755e-06,
|
| 2080 |
+
"loss": 0.408,
|
| 2081 |
+
"step": 172000
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 1.69,
|
| 2085 |
+
"learning_rate": 3.1472775785060163e-06,
|
| 2086 |
+
"loss": 0.4192,
|
| 2087 |
+
"step": 172500
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 1.69,
|
| 2091 |
+
"learning_rate": 3.0984098439164563e-06,
|
| 2092 |
+
"loss": 0.4187,
|
| 2093 |
+
"step": 173000
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 1.7,
|
| 2097 |
+
"learning_rate": 3.0495421093268962e-06,
|
| 2098 |
+
"loss": 0.4203,
|
| 2099 |
+
"step": 173500
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 1.7,
|
| 2103 |
+
"learning_rate": 3.000772110206515e-06,
|
| 2104 |
+
"loss": 0.4067,
|
| 2105 |
+
"step": 174000
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 1.71,
|
| 2109 |
+
"learning_rate": 2.951904375616955e-06,
|
| 2110 |
+
"loss": 0.4074,
|
| 2111 |
+
"step": 174500
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 1.71,
|
| 2115 |
+
"learning_rate": 2.9031343764965747e-06,
|
| 2116 |
+
"loss": 0.4027,
|
| 2117 |
+
"step": 175000
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.72,
|
| 2121 |
+
"learning_rate": 2.8542666419070147e-06,
|
| 2122 |
+
"loss": 0.4033,
|
| 2123 |
+
"step": 175500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 1.72,
|
| 2127 |
+
"learning_rate": 2.8053989073174546e-06,
|
| 2128 |
+
"loss": 0.4206,
|
| 2129 |
+
"step": 176000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 1.73,
|
| 2133 |
+
"learning_rate": 2.756531172727895e-06,
|
| 2134 |
+
"loss": 0.4209,
|
| 2135 |
+
"step": 176500
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 1.73,
|
| 2139 |
+
"learning_rate": 2.707663438138335e-06,
|
| 2140 |
+
"loss": 0.4148,
|
| 2141 |
+
"step": 177000
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 1.73,
|
| 2145 |
+
"learning_rate": 2.658795703548775e-06,
|
| 2146 |
+
"loss": 0.4092,
|
| 2147 |
+
"step": 177500
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.74,
|
| 2151 |
+
"learning_rate": 2.609927968959215e-06,
|
| 2152 |
+
"loss": 0.4227,
|
| 2153 |
+
"step": 178000
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 1.74,
|
| 2157 |
+
"learning_rate": 2.5610602343696556e-06,
|
| 2158 |
+
"loss": 0.4159,
|
| 2159 |
+
"step": 178500
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 1.75,
|
| 2163 |
+
"learning_rate": 2.5121924997800956e-06,
|
| 2164 |
+
"loss": 0.4178,
|
| 2165 |
+
"step": 179000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 1.75,
|
| 2169 |
+
"learning_rate": 2.4634225006597145e-06,
|
| 2170 |
+
"loss": 0.4041,
|
| 2171 |
+
"step": 179500
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 1.76,
|
| 2175 |
+
"learning_rate": 2.4145547660701544e-06,
|
| 2176 |
+
"loss": 0.4185,
|
| 2177 |
+
"step": 180000
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.76,
|
| 2181 |
+
"learning_rate": 2.3657847669497737e-06,
|
| 2182 |
+
"loss": 0.4162,
|
| 2183 |
+
"step": 180500
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 1.77,
|
| 2187 |
+
"learning_rate": 2.316917032360214e-06,
|
| 2188 |
+
"loss": 0.4095,
|
| 2189 |
+
"step": 181000
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 1.77,
|
| 2193 |
+
"learning_rate": 2.268049297770654e-06,
|
| 2194 |
+
"loss": 0.4135,
|
| 2195 |
+
"step": 181500
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 1.78,
|
| 2199 |
+
"learning_rate": 2.2191815631810944e-06,
|
| 2200 |
+
"loss": 0.4168,
|
| 2201 |
+
"step": 182000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 1.78,
|
| 2205 |
+
"learning_rate": 2.1703138285915343e-06,
|
| 2206 |
+
"loss": 0.3989,
|
| 2207 |
+
"step": 182500
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 1.79,
|
| 2211 |
+
"learning_rate": 2.1214460940019743e-06,
|
| 2212 |
+
"loss": 0.4027,
|
| 2213 |
+
"step": 183000
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 1.79,
|
| 2217 |
+
"learning_rate": 2.0725783594124146e-06,
|
| 2218 |
+
"loss": 0.4042,
|
| 2219 |
+
"step": 183500
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 1.8,
|
| 2223 |
+
"learning_rate": 2.0237106248228546e-06,
|
| 2224 |
+
"loss": 0.4086,
|
| 2225 |
+
"step": 184000
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 1.8,
|
| 2229 |
+
"learning_rate": 1.9748428902332945e-06,
|
| 2230 |
+
"loss": 0.4128,
|
| 2231 |
+
"step": 184500
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 1.81,
|
| 2235 |
+
"learning_rate": 1.925975155643735e-06,
|
| 2236 |
+
"loss": 0.4162,
|
| 2237 |
+
"step": 185000
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 1.81,
|
| 2241 |
+
"learning_rate": 1.877205156523354e-06,
|
| 2242 |
+
"loss": 0.4162,
|
| 2243 |
+
"step": 185500
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 1.82,
|
| 2247 |
+
"learning_rate": 1.8283374219337941e-06,
|
| 2248 |
+
"loss": 0.3968,
|
| 2249 |
+
"step": 186000
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 1.82,
|
| 2253 |
+
"learning_rate": 1.7794696873442343e-06,
|
| 2254 |
+
"loss": 0.4089,
|
| 2255 |
+
"step": 186500
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 1.83,
|
| 2259 |
+
"learning_rate": 1.7306019527546745e-06,
|
| 2260 |
+
"loss": 0.4124,
|
| 2261 |
+
"step": 187000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 1.83,
|
| 2265 |
+
"learning_rate": 1.6818319536342935e-06,
|
| 2266 |
+
"loss": 0.4041,
|
| 2267 |
+
"step": 187500
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 1.84,
|
| 2271 |
+
"learning_rate": 1.6329642190447337e-06,
|
| 2272 |
+
"loss": 0.4059,
|
| 2273 |
+
"step": 188000
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 1.84,
|
| 2277 |
+
"learning_rate": 1.584194219924353e-06,
|
| 2278 |
+
"loss": 0.4233,
|
| 2279 |
+
"step": 188500
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 1.85,
|
| 2283 |
+
"learning_rate": 1.535326485334793e-06,
|
| 2284 |
+
"loss": 0.4053,
|
| 2285 |
+
"step": 189000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 1.85,
|
| 2289 |
+
"learning_rate": 1.486458750745233e-06,
|
| 2290 |
+
"loss": 0.3967,
|
| 2291 |
+
"step": 189500
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 1.86,
|
| 2295 |
+
"learning_rate": 1.437591016155673e-06,
|
| 2296 |
+
"loss": 0.4158,
|
| 2297 |
+
"step": 190000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 1.86,
|
| 2301 |
+
"learning_rate": 1.3887232815661134e-06,
|
| 2302 |
+
"loss": 0.4152,
|
| 2303 |
+
"step": 190500
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 1.87,
|
| 2307 |
+
"learning_rate": 1.3398555469765534e-06,
|
| 2308 |
+
"loss": 0.4239,
|
| 2309 |
+
"step": 191000
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 1.87,
|
| 2313 |
+
"learning_rate": 1.2910855478561726e-06,
|
| 2314 |
+
"loss": 0.4212,
|
| 2315 |
+
"step": 191500
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 1.88,
|
| 2319 |
+
"learning_rate": 1.2422178132666126e-06,
|
| 2320 |
+
"loss": 0.3981,
|
| 2321 |
+
"step": 192000
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 1.88,
|
| 2325 |
+
"learning_rate": 1.1933500786770528e-06,
|
| 2326 |
+
"loss": 0.4068,
|
| 2327 |
+
"step": 192500
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 1.89,
|
| 2331 |
+
"learning_rate": 1.144482344087493e-06,
|
| 2332 |
+
"loss": 0.4128,
|
| 2333 |
+
"step": 193000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.89,
|
| 2337 |
+
"learning_rate": 1.0957123449671122e-06,
|
| 2338 |
+
"loss": 0.4032,
|
| 2339 |
+
"step": 193500
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 1.9,
|
| 2343 |
+
"learning_rate": 1.0468446103775521e-06,
|
| 2344 |
+
"loss": 0.4141,
|
| 2345 |
+
"step": 194000
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 1.9,
|
| 2349 |
+
"learning_rate": 9.979768757879923e-07,
|
| 2350 |
+
"loss": 0.4089,
|
| 2351 |
+
"step": 194500
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 1.91,
|
| 2355 |
+
"learning_rate": 9.491091411984324e-07,
|
| 2356 |
+
"loss": 0.398,
|
| 2357 |
+
"step": 195000
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 1.91,
|
| 2361 |
+
"learning_rate": 9.002414066088724e-07,
|
| 2362 |
+
"loss": 0.418,
|
| 2363 |
+
"step": 195500
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 1.92,
|
| 2367 |
+
"learning_rate": 8.513736720193126e-07,
|
| 2368 |
+
"loss": 0.4027,
|
| 2369 |
+
"step": 196000
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 1.92,
|
| 2373 |
+
"learning_rate": 8.025059374297526e-07,
|
| 2374 |
+
"loss": 0.4197,
|
| 2375 |
+
"step": 196500
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 1.93,
|
| 2379 |
+
"learning_rate": 7.536382028401929e-07,
|
| 2380 |
+
"loss": 0.4032,
|
| 2381 |
+
"step": 197000
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 1.93,
|
| 2385 |
+
"learning_rate": 7.047704682506328e-07,
|
| 2386 |
+
"loss": 0.407,
|
| 2387 |
+
"step": 197500
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 1.94,
|
| 2391 |
+
"learning_rate": 6.560004691302521e-07,
|
| 2392 |
+
"loss": 0.4111,
|
| 2393 |
+
"step": 198000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 1.94,
|
| 2397 |
+
"learning_rate": 6.072304700098713e-07,
|
| 2398 |
+
"loss": 0.4043,
|
| 2399 |
+
"step": 198500
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 1.94,
|
| 2403 |
+
"learning_rate": 5.583627354203115e-07,
|
| 2404 |
+
"loss": 0.4243,
|
| 2405 |
+
"step": 199000
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 1.95,
|
| 2409 |
+
"learning_rate": 5.094950008307515e-07,
|
| 2410 |
+
"loss": 0.4061,
|
| 2411 |
+
"step": 199500
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 1.95,
|
| 2415 |
+
"learning_rate": 4.606272662411916e-07,
|
| 2416 |
+
"loss": 0.4093,
|
| 2417 |
+
"step": 200000
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 1.96,
|
| 2421 |
+
"learning_rate": 4.1185726712081085e-07,
|
| 2422 |
+
"loss": 0.3918,
|
| 2423 |
+
"step": 200500
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 1.96,
|
| 2427 |
+
"learning_rate": 3.6298953253125096e-07,
|
| 2428 |
+
"loss": 0.4008,
|
| 2429 |
+
"step": 201000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 1.97,
|
| 2433 |
+
"learning_rate": 3.1412179794169107e-07,
|
| 2434 |
+
"loss": 0.4095,
|
| 2435 |
+
"step": 201500
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 1.97,
|
| 2439 |
+
"learning_rate": 2.652540633521311e-07,
|
| 2440 |
+
"loss": 0.4074,
|
| 2441 |
+
"step": 202000
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 1.98,
|
| 2445 |
+
"learning_rate": 2.1638632876257126e-07,
|
| 2446 |
+
"loss": 0.4141,
|
| 2447 |
+
"step": 202500
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 1.98,
|
| 2451 |
+
"learning_rate": 1.6761632964219046e-07,
|
| 2452 |
+
"loss": 0.4112,
|
| 2453 |
+
"step": 203000
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 1.99,
|
| 2457 |
+
"learning_rate": 1.1874859505263056e-07,
|
| 2458 |
+
"loss": 0.4054,
|
| 2459 |
+
"step": 203500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 1.99,
|
| 2463 |
+
"learning_rate": 6.988086046307066e-08,
|
| 2464 |
+
"loss": 0.4067,
|
| 2465 |
+
"step": 204000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.0,
|
| 2469 |
+
"learning_rate": 2.1013125873510755e-08,
|
| 2470 |
+
"loss": 0.3976,
|
| 2471 |
+
"step": 204500
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.0,
|
| 2475 |
+
"eval_accuracy": 0.8056860596832,
|
| 2476 |
+
"eval_loss": 0.4429013729095459,
|
| 2477 |
+
"eval_runtime": 204.9979,
|
| 2478 |
+
"eval_samples_per_second": 262.69,
|
| 2479 |
+
"eval_steps_per_second": 26.273,
|
| 2480 |
+
"step": 204634
|
| 2481 |
+
}
|
| 2482 |
+
],
|
| 2483 |
+
"max_steps": 204634,
|
| 2484 |
+
"num_train_epochs": 2,
|
| 2485 |
+
"total_flos": 5.50000642476408e+16,
|
| 2486 |
+
"trial_name": null,
|
| 2487 |
+
"trial_params": null
|
| 2488 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aed01026b364c3b69f82fc0a4738d0f6f48e28d685e3374b97ae558b2c7d6f2e
|
| 3 |
+
size 3515
|