Spaces:
Runtime error
Runtime error
Commit
·
f11ac64
1
Parent(s):
7c43c1d
Upload 6 files
Browse files
checkpoint-210000/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": "4.24.0",
|
23 |
+
"type_vocab_size": 2,
|
24 |
+
"use_cache": true,
|
25 |
+
"vocab_size": 30522
|
26 |
+
}
|
checkpoint-210000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a915aa6900c62f83709495075403a24964821e4f2a80634cc5c9b286ea448ab
|
3 |
+
size 438128811
|
checkpoint-210000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3381bc7c658553a02db0a01f5ffa10ad1f42183d2a3a561b1b1b9c67f2b62c3a
|
3 |
+
size 14503
|
checkpoint-210000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1be4f0217c476cd7d5006edf088e22d817cfe5efd13d5f4428273bd7c26e65fb
|
3 |
+
size 623
|
checkpoint-210000/trainer_state.json
ADDED
@@ -0,0 +1,2536 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 19.4480459344323,
|
5 |
+
"global_step": 210000,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.05,
|
12 |
+
"learning_rate": 4.988423782181886e-05,
|
13 |
+
"loss": 2.359,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.09,
|
18 |
+
"learning_rate": 4.976847564363771e-05,
|
19 |
+
"loss": 2.6154,
|
20 |
+
"step": 1000
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.14,
|
24 |
+
"learning_rate": 4.965271346545657e-05,
|
25 |
+
"loss": 2.6701,
|
26 |
+
"step": 1500
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.19,
|
30 |
+
"learning_rate": 4.953695128727542e-05,
|
31 |
+
"loss": 2.6487,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.23,
|
36 |
+
"learning_rate": 4.9421189109094276e-05,
|
37 |
+
"loss": 2.6192,
|
38 |
+
"step": 2500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 0.28,
|
42 |
+
"learning_rate": 4.930542693091314e-05,
|
43 |
+
"loss": 2.6102,
|
44 |
+
"step": 3000
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.32,
|
48 |
+
"learning_rate": 4.918966475273199e-05,
|
49 |
+
"loss": 2.5748,
|
50 |
+
"step": 3500
|
51 |
+
},
|
52 |
+
{
|
53 |
+
"epoch": 0.37,
|
54 |
+
"learning_rate": 4.907390257455085e-05,
|
55 |
+
"loss": 2.566,
|
56 |
+
"step": 4000
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.42,
|
60 |
+
"learning_rate": 4.89581403963697e-05,
|
61 |
+
"loss": 2.5363,
|
62 |
+
"step": 4500
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.46,
|
66 |
+
"learning_rate": 4.884237821818856e-05,
|
67 |
+
"loss": 2.5141,
|
68 |
+
"step": 5000
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 0.51,
|
72 |
+
"learning_rate": 4.872661604000741e-05,
|
73 |
+
"loss": 2.5071,
|
74 |
+
"step": 5500
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.56,
|
78 |
+
"learning_rate": 4.8610853861826266e-05,
|
79 |
+
"loss": 2.4713,
|
80 |
+
"step": 6000
|
81 |
+
},
|
82 |
+
{
|
83 |
+
"epoch": 0.6,
|
84 |
+
"learning_rate": 4.849509168364512e-05,
|
85 |
+
"loss": 2.4598,
|
86 |
+
"step": 6500
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.65,
|
90 |
+
"learning_rate": 4.8379329505463976e-05,
|
91 |
+
"loss": 2.4441,
|
92 |
+
"step": 7000
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 0.69,
|
96 |
+
"learning_rate": 4.826356732728283e-05,
|
97 |
+
"loss": 2.428,
|
98 |
+
"step": 7500
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.74,
|
102 |
+
"learning_rate": 4.8147805149101685e-05,
|
103 |
+
"loss": 2.4333,
|
104 |
+
"step": 8000
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.79,
|
108 |
+
"learning_rate": 4.803204297092054e-05,
|
109 |
+
"loss": 2.4121,
|
110 |
+
"step": 8500
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 0.83,
|
114 |
+
"learning_rate": 4.79162807927394e-05,
|
115 |
+
"loss": 2.4034,
|
116 |
+
"step": 9000
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 0.88,
|
120 |
+
"learning_rate": 4.7800518614558257e-05,
|
121 |
+
"loss": 2.3879,
|
122 |
+
"step": 9500
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.93,
|
126 |
+
"learning_rate": 4.768475643637711e-05,
|
127 |
+
"loss": 2.3844,
|
128 |
+
"step": 10000
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.97,
|
132 |
+
"learning_rate": 4.7568994258195966e-05,
|
133 |
+
"loss": 2.36,
|
134 |
+
"step": 10500
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 1.02,
|
138 |
+
"learning_rate": 4.745323208001482e-05,
|
139 |
+
"loss": 2.3304,
|
140 |
+
"step": 11000
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 1.07,
|
144 |
+
"learning_rate": 4.7337469901833676e-05,
|
145 |
+
"loss": 2.3104,
|
146 |
+
"step": 11500
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 1.11,
|
150 |
+
"learning_rate": 4.722170772365253e-05,
|
151 |
+
"loss": 2.2965,
|
152 |
+
"step": 12000
|
153 |
+
},
|
154 |
+
{
|
155 |
+
"epoch": 1.16,
|
156 |
+
"learning_rate": 4.7105945545471385e-05,
|
157 |
+
"loss": 2.2925,
|
158 |
+
"step": 12500
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 1.2,
|
162 |
+
"learning_rate": 4.699018336729024e-05,
|
163 |
+
"loss": 2.2881,
|
164 |
+
"step": 13000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 1.25,
|
168 |
+
"learning_rate": 4.6874421189109095e-05,
|
169 |
+
"loss": 2.2873,
|
170 |
+
"step": 13500
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 1.3,
|
174 |
+
"learning_rate": 4.675865901092795e-05,
|
175 |
+
"loss": 2.2997,
|
176 |
+
"step": 14000
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 1.34,
|
180 |
+
"learning_rate": 4.6642896832746804e-05,
|
181 |
+
"loss": 2.2749,
|
182 |
+
"step": 14500
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 1.39,
|
186 |
+
"learning_rate": 4.6527134654565666e-05,
|
187 |
+
"loss": 2.2578,
|
188 |
+
"step": 15000
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 1.44,
|
192 |
+
"learning_rate": 4.641137247638452e-05,
|
193 |
+
"loss": 2.2516,
|
194 |
+
"step": 15500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 1.48,
|
198 |
+
"learning_rate": 4.6295610298203375e-05,
|
199 |
+
"loss": 2.2524,
|
200 |
+
"step": 16000
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 1.53,
|
204 |
+
"learning_rate": 4.617984812002223e-05,
|
205 |
+
"loss": 2.2477,
|
206 |
+
"step": 16500
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 1.57,
|
210 |
+
"learning_rate": 4.6064085941841085e-05,
|
211 |
+
"loss": 2.2351,
|
212 |
+
"step": 17000
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 1.62,
|
216 |
+
"learning_rate": 4.594832376365994e-05,
|
217 |
+
"loss": 2.2216,
|
218 |
+
"step": 17500
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 1.67,
|
222 |
+
"learning_rate": 4.5832561585478794e-05,
|
223 |
+
"loss": 2.223,
|
224 |
+
"step": 18000
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 1.71,
|
228 |
+
"learning_rate": 4.571679940729765e-05,
|
229 |
+
"loss": 2.2147,
|
230 |
+
"step": 18500
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 1.76,
|
234 |
+
"learning_rate": 4.5601037229116504e-05,
|
235 |
+
"loss": 2.1889,
|
236 |
+
"step": 19000
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 1.81,
|
240 |
+
"learning_rate": 4.548527505093536e-05,
|
241 |
+
"loss": 2.205,
|
242 |
+
"step": 19500
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 1.85,
|
246 |
+
"learning_rate": 4.536951287275421e-05,
|
247 |
+
"loss": 2.188,
|
248 |
+
"step": 20000
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 1.9,
|
252 |
+
"learning_rate": 4.5253750694573075e-05,
|
253 |
+
"loss": 2.201,
|
254 |
+
"step": 20500
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 1.94,
|
258 |
+
"learning_rate": 4.513798851639193e-05,
|
259 |
+
"loss": 2.1782,
|
260 |
+
"step": 21000
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 1.99,
|
264 |
+
"learning_rate": 4.5022226338210784e-05,
|
265 |
+
"loss": 2.1824,
|
266 |
+
"step": 21500
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 2.04,
|
270 |
+
"learning_rate": 4.490646416002964e-05,
|
271 |
+
"loss": 2.1512,
|
272 |
+
"step": 22000
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 2.08,
|
276 |
+
"learning_rate": 4.4790701981848494e-05,
|
277 |
+
"loss": 2.1424,
|
278 |
+
"step": 22500
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 2.13,
|
282 |
+
"learning_rate": 4.467493980366735e-05,
|
283 |
+
"loss": 2.1524,
|
284 |
+
"step": 23000
|
285 |
+
},
|
286 |
+
{
|
287 |
+
"epoch": 2.18,
|
288 |
+
"learning_rate": 4.4559177625486204e-05,
|
289 |
+
"loss": 2.1378,
|
290 |
+
"step": 23500
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 2.22,
|
294 |
+
"learning_rate": 4.444341544730506e-05,
|
295 |
+
"loss": 2.1443,
|
296 |
+
"step": 24000
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 2.27,
|
300 |
+
"learning_rate": 4.432765326912391e-05,
|
301 |
+
"loss": 2.1306,
|
302 |
+
"step": 24500
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.32,
|
306 |
+
"learning_rate": 4.421189109094277e-05,
|
307 |
+
"loss": 2.1208,
|
308 |
+
"step": 25000
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 2.36,
|
312 |
+
"learning_rate": 4.409612891276162e-05,
|
313 |
+
"loss": 2.1123,
|
314 |
+
"step": 25500
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.41,
|
318 |
+
"learning_rate": 4.398036673458048e-05,
|
319 |
+
"loss": 2.1156,
|
320 |
+
"step": 26000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.45,
|
324 |
+
"learning_rate": 4.386460455639934e-05,
|
325 |
+
"loss": 2.1152,
|
326 |
+
"step": 26500
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 2.5,
|
330 |
+
"learning_rate": 4.3748842378218194e-05,
|
331 |
+
"loss": 2.1081,
|
332 |
+
"step": 27000
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 2.55,
|
336 |
+
"learning_rate": 4.363308020003705e-05,
|
337 |
+
"loss": 2.1157,
|
338 |
+
"step": 27500
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.59,
|
342 |
+
"learning_rate": 4.35173180218559e-05,
|
343 |
+
"loss": 2.0913,
|
344 |
+
"step": 28000
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 2.64,
|
348 |
+
"learning_rate": 4.340155584367476e-05,
|
349 |
+
"loss": 2.0889,
|
350 |
+
"step": 28500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 2.69,
|
354 |
+
"learning_rate": 4.328579366549361e-05,
|
355 |
+
"loss": 2.0895,
|
356 |
+
"step": 29000
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 2.73,
|
360 |
+
"learning_rate": 4.317003148731247e-05,
|
361 |
+
"loss": 2.0755,
|
362 |
+
"step": 29500
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 2.78,
|
366 |
+
"learning_rate": 4.305426930913132e-05,
|
367 |
+
"loss": 2.097,
|
368 |
+
"step": 30000
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 2.82,
|
372 |
+
"learning_rate": 4.293850713095018e-05,
|
373 |
+
"loss": 2.0809,
|
374 |
+
"step": 30500
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 2.87,
|
378 |
+
"learning_rate": 4.282274495276903e-05,
|
379 |
+
"loss": 2.0865,
|
380 |
+
"step": 31000
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 2.92,
|
384 |
+
"learning_rate": 4.2706982774587887e-05,
|
385 |
+
"loss": 2.0943,
|
386 |
+
"step": 31500
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 2.96,
|
390 |
+
"learning_rate": 4.259122059640674e-05,
|
391 |
+
"loss": 2.0625,
|
392 |
+
"step": 32000
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 3.01,
|
396 |
+
"learning_rate": 4.24754584182256e-05,
|
397 |
+
"loss": 2.063,
|
398 |
+
"step": 32500
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 3.06,
|
402 |
+
"learning_rate": 4.235969624004446e-05,
|
403 |
+
"loss": 2.0345,
|
404 |
+
"step": 33000
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 3.1,
|
408 |
+
"learning_rate": 4.224393406186331e-05,
|
409 |
+
"loss": 2.0417,
|
410 |
+
"step": 33500
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 3.15,
|
414 |
+
"learning_rate": 4.212817188368217e-05,
|
415 |
+
"loss": 2.0288,
|
416 |
+
"step": 34000
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 3.2,
|
420 |
+
"learning_rate": 4.201240970550102e-05,
|
421 |
+
"loss": 2.0252,
|
422 |
+
"step": 34500
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 3.24,
|
426 |
+
"learning_rate": 4.189664752731988e-05,
|
427 |
+
"loss": 2.0456,
|
428 |
+
"step": 35000
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 3.29,
|
432 |
+
"learning_rate": 4.178088534913873e-05,
|
433 |
+
"loss": 2.0232,
|
434 |
+
"step": 35500
|
435 |
+
},
|
436 |
+
{
|
437 |
+
"epoch": 3.33,
|
438 |
+
"learning_rate": 4.1665123170957586e-05,
|
439 |
+
"loss": 2.0188,
|
440 |
+
"step": 36000
|
441 |
+
},
|
442 |
+
{
|
443 |
+
"epoch": 3.38,
|
444 |
+
"learning_rate": 4.154936099277644e-05,
|
445 |
+
"loss": 2.0228,
|
446 |
+
"step": 36500
|
447 |
+
},
|
448 |
+
{
|
449 |
+
"epoch": 3.43,
|
450 |
+
"learning_rate": 4.1433598814595296e-05,
|
451 |
+
"loss": 2.0222,
|
452 |
+
"step": 37000
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 3.47,
|
456 |
+
"learning_rate": 4.131783663641415e-05,
|
457 |
+
"loss": 2.0224,
|
458 |
+
"step": 37500
|
459 |
+
},
|
460 |
+
{
|
461 |
+
"epoch": 3.52,
|
462 |
+
"learning_rate": 4.120207445823301e-05,
|
463 |
+
"loss": 2.0161,
|
464 |
+
"step": 38000
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 3.57,
|
468 |
+
"learning_rate": 4.108631228005187e-05,
|
469 |
+
"loss": 2.0031,
|
470 |
+
"step": 38500
|
471 |
+
},
|
472 |
+
{
|
473 |
+
"epoch": 3.61,
|
474 |
+
"learning_rate": 4.097055010187072e-05,
|
475 |
+
"loss": 2.0084,
|
476 |
+
"step": 39000
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 3.66,
|
480 |
+
"learning_rate": 4.0854787923689576e-05,
|
481 |
+
"loss": 1.9985,
|
482 |
+
"step": 39500
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 3.7,
|
486 |
+
"learning_rate": 4.073902574550843e-05,
|
487 |
+
"loss": 1.9893,
|
488 |
+
"step": 40000
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 3.75,
|
492 |
+
"learning_rate": 4.0623263567327286e-05,
|
493 |
+
"loss": 1.9974,
|
494 |
+
"step": 40500
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 3.8,
|
498 |
+
"learning_rate": 4.050750138914614e-05,
|
499 |
+
"loss": 1.9872,
|
500 |
+
"step": 41000
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 3.84,
|
504 |
+
"learning_rate": 4.0391739210964995e-05,
|
505 |
+
"loss": 2.0026,
|
506 |
+
"step": 41500
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 3.89,
|
510 |
+
"learning_rate": 4.027597703278385e-05,
|
511 |
+
"loss": 1.989,
|
512 |
+
"step": 42000
|
513 |
+
},
|
514 |
+
{
|
515 |
+
"epoch": 3.94,
|
516 |
+
"learning_rate": 4.0160214854602705e-05,
|
517 |
+
"loss": 1.9884,
|
518 |
+
"step": 42500
|
519 |
+
},
|
520 |
+
{
|
521 |
+
"epoch": 3.98,
|
522 |
+
"learning_rate": 4.004445267642156e-05,
|
523 |
+
"loss": 1.9803,
|
524 |
+
"step": 43000
|
525 |
+
},
|
526 |
+
{
|
527 |
+
"epoch": 4.03,
|
528 |
+
"learning_rate": 3.9928690498240415e-05,
|
529 |
+
"loss": 1.9612,
|
530 |
+
"step": 43500
|
531 |
+
},
|
532 |
+
{
|
533 |
+
"epoch": 4.07,
|
534 |
+
"learning_rate": 3.9812928320059276e-05,
|
535 |
+
"loss": 1.9602,
|
536 |
+
"step": 44000
|
537 |
+
},
|
538 |
+
{
|
539 |
+
"epoch": 4.12,
|
540 |
+
"learning_rate": 3.969716614187813e-05,
|
541 |
+
"loss": 1.9474,
|
542 |
+
"step": 44500
|
543 |
+
},
|
544 |
+
{
|
545 |
+
"epoch": 4.17,
|
546 |
+
"learning_rate": 3.9581403963696986e-05,
|
547 |
+
"loss": 1.9457,
|
548 |
+
"step": 45000
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 4.21,
|
552 |
+
"learning_rate": 3.946564178551584e-05,
|
553 |
+
"loss": 1.9483,
|
554 |
+
"step": 45500
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 4.26,
|
558 |
+
"learning_rate": 3.9349879607334695e-05,
|
559 |
+
"loss": 1.9543,
|
560 |
+
"step": 46000
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 4.31,
|
564 |
+
"learning_rate": 3.923411742915355e-05,
|
565 |
+
"loss": 1.946,
|
566 |
+
"step": 46500
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 4.35,
|
570 |
+
"learning_rate": 3.9118355250972405e-05,
|
571 |
+
"loss": 1.9436,
|
572 |
+
"step": 47000
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 4.4,
|
576 |
+
"learning_rate": 3.900259307279126e-05,
|
577 |
+
"loss": 1.9591,
|
578 |
+
"step": 47500
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 4.45,
|
582 |
+
"learning_rate": 3.8886830894610114e-05,
|
583 |
+
"loss": 1.9477,
|
584 |
+
"step": 48000
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 4.49,
|
588 |
+
"learning_rate": 3.877106871642897e-05,
|
589 |
+
"loss": 1.9369,
|
590 |
+
"step": 48500
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 4.54,
|
594 |
+
"learning_rate": 3.8655306538247824e-05,
|
595 |
+
"loss": 1.9349,
|
596 |
+
"step": 49000
|
597 |
+
},
|
598 |
+
{
|
599 |
+
"epoch": 4.58,
|
600 |
+
"learning_rate": 3.853954436006668e-05,
|
601 |
+
"loss": 1.9339,
|
602 |
+
"step": 49500
|
603 |
+
},
|
604 |
+
{
|
605 |
+
"epoch": 4.63,
|
606 |
+
"learning_rate": 3.842378218188554e-05,
|
607 |
+
"loss": 1.9393,
|
608 |
+
"step": 50000
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 4.68,
|
612 |
+
"learning_rate": 3.8308020003704395e-05,
|
613 |
+
"loss": 1.94,
|
614 |
+
"step": 50500
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 4.72,
|
618 |
+
"learning_rate": 3.819225782552325e-05,
|
619 |
+
"loss": 1.9337,
|
620 |
+
"step": 51000
|
621 |
+
},
|
622 |
+
{
|
623 |
+
"epoch": 4.77,
|
624 |
+
"learning_rate": 3.8076495647342104e-05,
|
625 |
+
"loss": 1.9271,
|
626 |
+
"step": 51500
|
627 |
+
},
|
628 |
+
{
|
629 |
+
"epoch": 4.82,
|
630 |
+
"learning_rate": 3.796073346916096e-05,
|
631 |
+
"loss": 1.9113,
|
632 |
+
"step": 52000
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 4.86,
|
636 |
+
"learning_rate": 3.7844971290979814e-05,
|
637 |
+
"loss": 1.9176,
|
638 |
+
"step": 52500
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 4.91,
|
642 |
+
"learning_rate": 3.772920911279867e-05,
|
643 |
+
"loss": 1.9125,
|
644 |
+
"step": 53000
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 4.95,
|
648 |
+
"learning_rate": 3.761344693461752e-05,
|
649 |
+
"loss": 1.9162,
|
650 |
+
"step": 53500
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 5.0,
|
654 |
+
"learning_rate": 3.749768475643638e-05,
|
655 |
+
"loss": 1.9318,
|
656 |
+
"step": 54000
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 5.05,
|
660 |
+
"learning_rate": 3.738192257825523e-05,
|
661 |
+
"loss": 1.8941,
|
662 |
+
"step": 54500
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 5.09,
|
666 |
+
"learning_rate": 3.726616040007409e-05,
|
667 |
+
"loss": 1.9007,
|
668 |
+
"step": 55000
|
669 |
+
},
|
670 |
+
{
|
671 |
+
"epoch": 5.14,
|
672 |
+
"learning_rate": 3.715039822189294e-05,
|
673 |
+
"loss": 1.8976,
|
674 |
+
"step": 55500
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 5.19,
|
678 |
+
"learning_rate": 3.7034636043711804e-05,
|
679 |
+
"loss": 1.9052,
|
680 |
+
"step": 56000
|
681 |
+
},
|
682 |
+
{
|
683 |
+
"epoch": 5.23,
|
684 |
+
"learning_rate": 3.691887386553066e-05,
|
685 |
+
"loss": 1.8927,
|
686 |
+
"step": 56500
|
687 |
+
},
|
688 |
+
{
|
689 |
+
"epoch": 5.28,
|
690 |
+
"learning_rate": 3.6803111687349514e-05,
|
691 |
+
"loss": 1.8941,
|
692 |
+
"step": 57000
|
693 |
+
},
|
694 |
+
{
|
695 |
+
"epoch": 5.33,
|
696 |
+
"learning_rate": 3.668734950916837e-05,
|
697 |
+
"loss": 1.8815,
|
698 |
+
"step": 57500
|
699 |
+
},
|
700 |
+
{
|
701 |
+
"epoch": 5.37,
|
702 |
+
"learning_rate": 3.657158733098722e-05,
|
703 |
+
"loss": 1.8748,
|
704 |
+
"step": 58000
|
705 |
+
},
|
706 |
+
{
|
707 |
+
"epoch": 5.42,
|
708 |
+
"learning_rate": 3.645582515280608e-05,
|
709 |
+
"loss": 1.8955,
|
710 |
+
"step": 58500
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 5.46,
|
714 |
+
"learning_rate": 3.634006297462493e-05,
|
715 |
+
"loss": 1.8792,
|
716 |
+
"step": 59000
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 5.51,
|
720 |
+
"learning_rate": 3.622430079644379e-05,
|
721 |
+
"loss": 1.8902,
|
722 |
+
"step": 59500
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 5.56,
|
726 |
+
"learning_rate": 3.610853861826264e-05,
|
727 |
+
"loss": 1.8842,
|
728 |
+
"step": 60000
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 5.6,
|
732 |
+
"learning_rate": 3.59927764400815e-05,
|
733 |
+
"loss": 1.8881,
|
734 |
+
"step": 60500
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 5.65,
|
738 |
+
"learning_rate": 3.587701426190035e-05,
|
739 |
+
"loss": 1.8708,
|
740 |
+
"step": 61000
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 5.7,
|
744 |
+
"learning_rate": 3.576125208371921e-05,
|
745 |
+
"loss": 1.8691,
|
746 |
+
"step": 61500
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 5.74,
|
750 |
+
"learning_rate": 3.564548990553807e-05,
|
751 |
+
"loss": 1.8636,
|
752 |
+
"step": 62000
|
753 |
+
},
|
754 |
+
{
|
755 |
+
"epoch": 5.79,
|
756 |
+
"learning_rate": 3.552972772735692e-05,
|
757 |
+
"loss": 1.8713,
|
758 |
+
"step": 62500
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 5.83,
|
762 |
+
"learning_rate": 3.541396554917578e-05,
|
763 |
+
"loss": 1.8538,
|
764 |
+
"step": 63000
|
765 |
+
},
|
766 |
+
{
|
767 |
+
"epoch": 5.88,
|
768 |
+
"learning_rate": 3.529820337099463e-05,
|
769 |
+
"loss": 1.855,
|
770 |
+
"step": 63500
|
771 |
+
},
|
772 |
+
{
|
773 |
+
"epoch": 5.93,
|
774 |
+
"learning_rate": 3.518244119281349e-05,
|
775 |
+
"loss": 1.8694,
|
776 |
+
"step": 64000
|
777 |
+
},
|
778 |
+
{
|
779 |
+
"epoch": 5.97,
|
780 |
+
"learning_rate": 3.506667901463234e-05,
|
781 |
+
"loss": 1.8744,
|
782 |
+
"step": 64500
|
783 |
+
},
|
784 |
+
{
|
785 |
+
"epoch": 6.02,
|
786 |
+
"learning_rate": 3.4950916836451197e-05,
|
787 |
+
"loss": 1.8531,
|
788 |
+
"step": 65000
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 6.07,
|
792 |
+
"learning_rate": 3.483515465827005e-05,
|
793 |
+
"loss": 1.8315,
|
794 |
+
"step": 65500
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 6.11,
|
798 |
+
"learning_rate": 3.4719392480088906e-05,
|
799 |
+
"loss": 1.83,
|
800 |
+
"step": 66000
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 6.16,
|
804 |
+
"learning_rate": 3.460363030190776e-05,
|
805 |
+
"loss": 1.8212,
|
806 |
+
"step": 66500
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 6.2,
|
810 |
+
"learning_rate": 3.4487868123726616e-05,
|
811 |
+
"loss": 1.836,
|
812 |
+
"step": 67000
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 6.25,
|
816 |
+
"learning_rate": 3.437210594554548e-05,
|
817 |
+
"loss": 1.8567,
|
818 |
+
"step": 67500
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 6.3,
|
822 |
+
"learning_rate": 3.425634376736433e-05,
|
823 |
+
"loss": 1.848,
|
824 |
+
"step": 68000
|
825 |
+
},
|
826 |
+
{
|
827 |
+
"epoch": 6.34,
|
828 |
+
"learning_rate": 3.414058158918319e-05,
|
829 |
+
"loss": 1.8366,
|
830 |
+
"step": 68500
|
831 |
+
},
|
832 |
+
{
|
833 |
+
"epoch": 6.39,
|
834 |
+
"learning_rate": 3.402481941100204e-05,
|
835 |
+
"loss": 1.8411,
|
836 |
+
"step": 69000
|
837 |
+
},
|
838 |
+
{
|
839 |
+
"epoch": 6.44,
|
840 |
+
"learning_rate": 3.3909057232820896e-05,
|
841 |
+
"loss": 1.8149,
|
842 |
+
"step": 69500
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 6.48,
|
846 |
+
"learning_rate": 3.379329505463975e-05,
|
847 |
+
"loss": 1.8319,
|
848 |
+
"step": 70000
|
849 |
+
},
|
850 |
+
{
|
851 |
+
"epoch": 6.53,
|
852 |
+
"learning_rate": 3.3677532876458606e-05,
|
853 |
+
"loss": 1.8208,
|
854 |
+
"step": 70500
|
855 |
+
},
|
856 |
+
{
|
857 |
+
"epoch": 6.58,
|
858 |
+
"learning_rate": 3.356177069827746e-05,
|
859 |
+
"loss": 1.8214,
|
860 |
+
"step": 71000
|
861 |
+
},
|
862 |
+
{
|
863 |
+
"epoch": 6.62,
|
864 |
+
"learning_rate": 3.3446008520096315e-05,
|
865 |
+
"loss": 1.8264,
|
866 |
+
"step": 71500
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 6.67,
|
870 |
+
"learning_rate": 3.333024634191517e-05,
|
871 |
+
"loss": 1.8273,
|
872 |
+
"step": 72000
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 6.71,
|
876 |
+
"learning_rate": 3.3214484163734025e-05,
|
877 |
+
"loss": 1.8203,
|
878 |
+
"step": 72500
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 6.76,
|
882 |
+
"learning_rate": 3.309872198555288e-05,
|
883 |
+
"loss": 1.8256,
|
884 |
+
"step": 73000
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 6.81,
|
888 |
+
"learning_rate": 3.298295980737174e-05,
|
889 |
+
"loss": 1.828,
|
890 |
+
"step": 73500
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 6.85,
|
894 |
+
"learning_rate": 3.2867197629190596e-05,
|
895 |
+
"loss": 1.818,
|
896 |
+
"step": 74000
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 6.9,
|
900 |
+
"learning_rate": 3.275143545100945e-05,
|
901 |
+
"loss": 1.807,
|
902 |
+
"step": 74500
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 6.95,
|
906 |
+
"learning_rate": 3.2635673272828305e-05,
|
907 |
+
"loss": 1.8138,
|
908 |
+
"step": 75000
|
909 |
+
},
|
910 |
+
{
|
911 |
+
"epoch": 6.99,
|
912 |
+
"learning_rate": 3.251991109464716e-05,
|
913 |
+
"loss": 1.8193,
|
914 |
+
"step": 75500
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 7.04,
|
918 |
+
"learning_rate": 3.2404148916466015e-05,
|
919 |
+
"loss": 1.8036,
|
920 |
+
"step": 76000
|
921 |
+
},
|
922 |
+
{
|
923 |
+
"epoch": 7.08,
|
924 |
+
"learning_rate": 3.228838673828487e-05,
|
925 |
+
"loss": 1.7956,
|
926 |
+
"step": 76500
|
927 |
+
},
|
928 |
+
{
|
929 |
+
"epoch": 7.13,
|
930 |
+
"learning_rate": 3.2172624560103725e-05,
|
931 |
+
"loss": 1.7853,
|
932 |
+
"step": 77000
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 7.18,
|
936 |
+
"learning_rate": 3.205686238192258e-05,
|
937 |
+
"loss": 1.7973,
|
938 |
+
"step": 77500
|
939 |
+
},
|
940 |
+
{
|
941 |
+
"epoch": 7.22,
|
942 |
+
"learning_rate": 3.1941100203741434e-05,
|
943 |
+
"loss": 1.7886,
|
944 |
+
"step": 78000
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 7.27,
|
948 |
+
"learning_rate": 3.182533802556029e-05,
|
949 |
+
"loss": 1.7775,
|
950 |
+
"step": 78500
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 7.32,
|
954 |
+
"learning_rate": 3.170957584737915e-05,
|
955 |
+
"loss": 1.7973,
|
956 |
+
"step": 79000
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 7.36,
|
960 |
+
"learning_rate": 3.1593813669198005e-05,
|
961 |
+
"loss": 1.7879,
|
962 |
+
"step": 79500
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 7.41,
|
966 |
+
"learning_rate": 3.147805149101686e-05,
|
967 |
+
"loss": 1.7839,
|
968 |
+
"step": 80000
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 7.46,
|
972 |
+
"learning_rate": 3.1362289312835715e-05,
|
973 |
+
"loss": 1.7789,
|
974 |
+
"step": 80500
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 7.5,
|
978 |
+
"learning_rate": 3.124652713465457e-05,
|
979 |
+
"loss": 1.7789,
|
980 |
+
"step": 81000
|
981 |
+
},
|
982 |
+
{
|
983 |
+
"epoch": 7.55,
|
984 |
+
"learning_rate": 3.1130764956473424e-05,
|
985 |
+
"loss": 1.7894,
|
986 |
+
"step": 81500
|
987 |
+
},
|
988 |
+
{
|
989 |
+
"epoch": 7.59,
|
990 |
+
"learning_rate": 3.101500277829228e-05,
|
991 |
+
"loss": 1.7802,
|
992 |
+
"step": 82000
|
993 |
+
},
|
994 |
+
{
|
995 |
+
"epoch": 7.64,
|
996 |
+
"learning_rate": 3.0899240600111134e-05,
|
997 |
+
"loss": 1.7719,
|
998 |
+
"step": 82500
|
999 |
+
},
|
1000 |
+
{
|
1001 |
+
"epoch": 7.69,
|
1002 |
+
"learning_rate": 3.078347842192999e-05,
|
1003 |
+
"loss": 1.7815,
|
1004 |
+
"step": 83000
|
1005 |
+
},
|
1006 |
+
{
|
1007 |
+
"epoch": 7.73,
|
1008 |
+
"learning_rate": 3.066771624374884e-05,
|
1009 |
+
"loss": 1.7722,
|
1010 |
+
"step": 83500
|
1011 |
+
},
|
1012 |
+
{
|
1013 |
+
"epoch": 7.78,
|
1014 |
+
"learning_rate": 3.05519540655677e-05,
|
1015 |
+
"loss": 1.7641,
|
1016 |
+
"step": 84000
|
1017 |
+
},
|
1018 |
+
{
|
1019 |
+
"epoch": 7.83,
|
1020 |
+
"learning_rate": 3.0436191887386556e-05,
|
1021 |
+
"loss": 1.7753,
|
1022 |
+
"step": 84500
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 7.87,
|
1026 |
+
"learning_rate": 3.032042970920541e-05,
|
1027 |
+
"loss": 1.7725,
|
1028 |
+
"step": 85000
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 7.92,
|
1032 |
+
"learning_rate": 3.0204667531024266e-05,
|
1033 |
+
"loss": 1.7749,
|
1034 |
+
"step": 85500
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 7.96,
|
1038 |
+
"learning_rate": 3.008890535284312e-05,
|
1039 |
+
"loss": 1.7701,
|
1040 |
+
"step": 86000
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 8.01,
|
1044 |
+
"learning_rate": 2.997314317466198e-05,
|
1045 |
+
"loss": 1.7713,
|
1046 |
+
"step": 86500
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 8.06,
|
1050 |
+
"learning_rate": 2.9857380996480833e-05,
|
1051 |
+
"loss": 1.7582,
|
1052 |
+
"step": 87000
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 8.1,
|
1056 |
+
"learning_rate": 2.9741618818299688e-05,
|
1057 |
+
"loss": 1.7436,
|
1058 |
+
"step": 87500
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 8.15,
|
1062 |
+
"learning_rate": 2.9625856640118543e-05,
|
1063 |
+
"loss": 1.7592,
|
1064 |
+
"step": 88000
|
1065 |
+
},
|
1066 |
+
{
|
1067 |
+
"epoch": 8.2,
|
1068 |
+
"learning_rate": 2.9510094461937398e-05,
|
1069 |
+
"loss": 1.7471,
|
1070 |
+
"step": 88500
|
1071 |
+
},
|
1072 |
+
{
|
1073 |
+
"epoch": 8.24,
|
1074 |
+
"learning_rate": 2.9394332283756252e-05,
|
1075 |
+
"loss": 1.7724,
|
1076 |
+
"step": 89000
|
1077 |
+
},
|
1078 |
+
{
|
1079 |
+
"epoch": 8.29,
|
1080 |
+
"learning_rate": 2.927857010557511e-05,
|
1081 |
+
"loss": 1.7529,
|
1082 |
+
"step": 89500
|
1083 |
+
},
|
1084 |
+
{
|
1085 |
+
"epoch": 8.33,
|
1086 |
+
"learning_rate": 2.9162807927393965e-05,
|
1087 |
+
"loss": 1.7532,
|
1088 |
+
"step": 90000
|
1089 |
+
},
|
1090 |
+
{
|
1091 |
+
"epoch": 8.38,
|
1092 |
+
"learning_rate": 2.904704574921282e-05,
|
1093 |
+
"loss": 1.7542,
|
1094 |
+
"step": 90500
|
1095 |
+
},
|
1096 |
+
{
|
1097 |
+
"epoch": 8.43,
|
1098 |
+
"learning_rate": 2.8931283571031675e-05,
|
1099 |
+
"loss": 1.753,
|
1100 |
+
"step": 91000
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 8.47,
|
1104 |
+
"learning_rate": 2.881552139285053e-05,
|
1105 |
+
"loss": 1.732,
|
1106 |
+
"step": 91500
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 8.52,
|
1110 |
+
"learning_rate": 2.8699759214669384e-05,
|
1111 |
+
"loss": 1.7326,
|
1112 |
+
"step": 92000
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 8.57,
|
1116 |
+
"learning_rate": 2.8583997036488243e-05,
|
1117 |
+
"loss": 1.7351,
|
1118 |
+
"step": 92500
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 8.61,
|
1122 |
+
"learning_rate": 2.8468234858307097e-05,
|
1123 |
+
"loss": 1.7348,
|
1124 |
+
"step": 93000
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 8.66,
|
1128 |
+
"learning_rate": 2.8352472680125952e-05,
|
1129 |
+
"loss": 1.7403,
|
1130 |
+
"step": 93500
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 8.71,
|
1134 |
+
"learning_rate": 2.8236710501944807e-05,
|
1135 |
+
"loss": 1.7589,
|
1136 |
+
"step": 94000
|
1137 |
+
},
|
1138 |
+
{
|
1139 |
+
"epoch": 8.75,
|
1140 |
+
"learning_rate": 2.812094832376366e-05,
|
1141 |
+
"loss": 1.7339,
|
1142 |
+
"step": 94500
|
1143 |
+
},
|
1144 |
+
{
|
1145 |
+
"epoch": 8.8,
|
1146 |
+
"learning_rate": 2.800518614558252e-05,
|
1147 |
+
"loss": 1.7325,
|
1148 |
+
"step": 95000
|
1149 |
+
},
|
1150 |
+
{
|
1151 |
+
"epoch": 8.84,
|
1152 |
+
"learning_rate": 2.7889423967401375e-05,
|
1153 |
+
"loss": 1.7272,
|
1154 |
+
"step": 95500
|
1155 |
+
},
|
1156 |
+
{
|
1157 |
+
"epoch": 8.89,
|
1158 |
+
"learning_rate": 2.777366178922023e-05,
|
1159 |
+
"loss": 1.7211,
|
1160 |
+
"step": 96000
|
1161 |
+
},
|
1162 |
+
{
|
1163 |
+
"epoch": 8.94,
|
1164 |
+
"learning_rate": 2.7657899611039084e-05,
|
1165 |
+
"loss": 1.7379,
|
1166 |
+
"step": 96500
|
1167 |
+
},
|
1168 |
+
{
|
1169 |
+
"epoch": 8.98,
|
1170 |
+
"learning_rate": 2.754213743285794e-05,
|
1171 |
+
"loss": 1.7414,
|
1172 |
+
"step": 97000
|
1173 |
+
},
|
1174 |
+
{
|
1175 |
+
"epoch": 9.03,
|
1176 |
+
"learning_rate": 2.7426375254676794e-05,
|
1177 |
+
"loss": 1.7175,
|
1178 |
+
"step": 97500
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 9.08,
|
1182 |
+
"learning_rate": 2.7310613076495652e-05,
|
1183 |
+
"loss": 1.702,
|
1184 |
+
"step": 98000
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 9.12,
|
1188 |
+
"learning_rate": 2.7194850898314507e-05,
|
1189 |
+
"loss": 1.7245,
|
1190 |
+
"step": 98500
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 9.17,
|
1194 |
+
"learning_rate": 2.707908872013336e-05,
|
1195 |
+
"loss": 1.7124,
|
1196 |
+
"step": 99000
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 9.21,
|
1200 |
+
"learning_rate": 2.6963326541952216e-05,
|
1201 |
+
"loss": 1.7137,
|
1202 |
+
"step": 99500
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 9.26,
|
1206 |
+
"learning_rate": 2.684756436377107e-05,
|
1207 |
+
"loss": 1.7006,
|
1208 |
+
"step": 100000
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 9.31,
|
1212 |
+
"learning_rate": 2.6731802185589926e-05,
|
1213 |
+
"loss": 1.7092,
|
1214 |
+
"step": 100500
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 9.35,
|
1218 |
+
"learning_rate": 2.6616040007408784e-05,
|
1219 |
+
"loss": 1.7175,
|
1220 |
+
"step": 101000
|
1221 |
+
},
|
1222 |
+
{
|
1223 |
+
"epoch": 9.4,
|
1224 |
+
"learning_rate": 2.650027782922764e-05,
|
1225 |
+
"loss": 1.7095,
|
1226 |
+
"step": 101500
|
1227 |
+
},
|
1228 |
+
{
|
1229 |
+
"epoch": 9.45,
|
1230 |
+
"learning_rate": 2.6384515651046493e-05,
|
1231 |
+
"loss": 1.6988,
|
1232 |
+
"step": 102000
|
1233 |
+
},
|
1234 |
+
{
|
1235 |
+
"epoch": 9.49,
|
1236 |
+
"learning_rate": 2.6268753472865348e-05,
|
1237 |
+
"loss": 1.7162,
|
1238 |
+
"step": 102500
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 9.54,
|
1242 |
+
"learning_rate": 2.6152991294684203e-05,
|
1243 |
+
"loss": 1.6978,
|
1244 |
+
"step": 103000
|
1245 |
+
},
|
1246 |
+
{
|
1247 |
+
"epoch": 9.59,
|
1248 |
+
"learning_rate": 2.6037229116503058e-05,
|
1249 |
+
"loss": 1.7047,
|
1250 |
+
"step": 103500
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 9.63,
|
1254 |
+
"learning_rate": 2.5921466938321916e-05,
|
1255 |
+
"loss": 1.7072,
|
1256 |
+
"step": 104000
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 9.68,
|
1260 |
+
"learning_rate": 2.580570476014077e-05,
|
1261 |
+
"loss": 1.7062,
|
1262 |
+
"step": 104500
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 9.72,
|
1266 |
+
"learning_rate": 2.5689942581959625e-05,
|
1267 |
+
"loss": 1.7013,
|
1268 |
+
"step": 105000
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 9.77,
|
1272 |
+
"learning_rate": 2.557418040377848e-05,
|
1273 |
+
"loss": 1.7025,
|
1274 |
+
"step": 105500
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 9.82,
|
1278 |
+
"learning_rate": 2.5458418225597335e-05,
|
1279 |
+
"loss": 1.7198,
|
1280 |
+
"step": 106000
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 9.86,
|
1284 |
+
"learning_rate": 2.534265604741619e-05,
|
1285 |
+
"loss": 1.6934,
|
1286 |
+
"step": 106500
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 9.91,
|
1290 |
+
"learning_rate": 2.5226893869235048e-05,
|
1291 |
+
"loss": 1.7013,
|
1292 |
+
"step": 107000
|
1293 |
+
},
|
1294 |
+
{
|
1295 |
+
"epoch": 9.96,
|
1296 |
+
"learning_rate": 2.5111131691053903e-05,
|
1297 |
+
"loss": 1.7108,
|
1298 |
+
"step": 107500
|
1299 |
+
},
|
1300 |
+
{
|
1301 |
+
"epoch": 10.0,
|
1302 |
+
"learning_rate": 2.4995369512872754e-05,
|
1303 |
+
"loss": 1.702,
|
1304 |
+
"step": 108000
|
1305 |
+
},
|
1306 |
+
{
|
1307 |
+
"epoch": 10.05,
|
1308 |
+
"learning_rate": 2.487960733469161e-05,
|
1309 |
+
"loss": 1.6911,
|
1310 |
+
"step": 108500
|
1311 |
+
},
|
1312 |
+
{
|
1313 |
+
"epoch": 10.09,
|
1314 |
+
"learning_rate": 2.4763845156510463e-05,
|
1315 |
+
"loss": 1.6897,
|
1316 |
+
"step": 109000
|
1317 |
+
},
|
1318 |
+
{
|
1319 |
+
"epoch": 10.14,
|
1320 |
+
"learning_rate": 2.464808297832932e-05,
|
1321 |
+
"loss": 1.6797,
|
1322 |
+
"step": 109500
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 10.19,
|
1326 |
+
"learning_rate": 2.4532320800148176e-05,
|
1327 |
+
"loss": 1.6827,
|
1328 |
+
"step": 110000
|
1329 |
+
},
|
1330 |
+
{
|
1331 |
+
"epoch": 10.23,
|
1332 |
+
"learning_rate": 2.441655862196703e-05,
|
1333 |
+
"loss": 1.6872,
|
1334 |
+
"step": 110500
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 10.28,
|
1338 |
+
"learning_rate": 2.4300796443785886e-05,
|
1339 |
+
"loss": 1.677,
|
1340 |
+
"step": 111000
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 10.33,
|
1344 |
+
"learning_rate": 2.418503426560474e-05,
|
1345 |
+
"loss": 1.6603,
|
1346 |
+
"step": 111500
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 10.37,
|
1350 |
+
"learning_rate": 2.40692720874236e-05,
|
1351 |
+
"loss": 1.6827,
|
1352 |
+
"step": 112000
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 10.42,
|
1356 |
+
"learning_rate": 2.3953509909242454e-05,
|
1357 |
+
"loss": 1.6676,
|
1358 |
+
"step": 112500
|
1359 |
+
},
|
1360 |
+
{
|
1361 |
+
"epoch": 10.46,
|
1362 |
+
"learning_rate": 2.383774773106131e-05,
|
1363 |
+
"loss": 1.6775,
|
1364 |
+
"step": 113000
|
1365 |
+
},
|
1366 |
+
{
|
1367 |
+
"epoch": 10.51,
|
1368 |
+
"learning_rate": 2.3721985552880163e-05,
|
1369 |
+
"loss": 1.6642,
|
1370 |
+
"step": 113500
|
1371 |
+
},
|
1372 |
+
{
|
1373 |
+
"epoch": 10.56,
|
1374 |
+
"learning_rate": 2.3606223374699018e-05,
|
1375 |
+
"loss": 1.6638,
|
1376 |
+
"step": 114000
|
1377 |
+
},
|
1378 |
+
{
|
1379 |
+
"epoch": 10.6,
|
1380 |
+
"learning_rate": 2.3490461196517873e-05,
|
1381 |
+
"loss": 1.6738,
|
1382 |
+
"step": 114500
|
1383 |
+
},
|
1384 |
+
{
|
1385 |
+
"epoch": 10.65,
|
1386 |
+
"learning_rate": 2.337469901833673e-05,
|
1387 |
+
"loss": 1.6731,
|
1388 |
+
"step": 115000
|
1389 |
+
},
|
1390 |
+
{
|
1391 |
+
"epoch": 10.7,
|
1392 |
+
"learning_rate": 2.3258936840155586e-05,
|
1393 |
+
"loss": 1.6628,
|
1394 |
+
"step": 115500
|
1395 |
+
},
|
1396 |
+
{
|
1397 |
+
"epoch": 10.74,
|
1398 |
+
"learning_rate": 2.314317466197444e-05,
|
1399 |
+
"loss": 1.6644,
|
1400 |
+
"step": 116000
|
1401 |
+
},
|
1402 |
+
{
|
1403 |
+
"epoch": 10.79,
|
1404 |
+
"learning_rate": 2.3027412483793295e-05,
|
1405 |
+
"loss": 1.6702,
|
1406 |
+
"step": 116500
|
1407 |
+
},
|
1408 |
+
{
|
1409 |
+
"epoch": 10.84,
|
1410 |
+
"learning_rate": 2.291165030561215e-05,
|
1411 |
+
"loss": 1.6568,
|
1412 |
+
"step": 117000
|
1413 |
+
},
|
1414 |
+
{
|
1415 |
+
"epoch": 10.88,
|
1416 |
+
"learning_rate": 2.2795888127431005e-05,
|
1417 |
+
"loss": 1.6685,
|
1418 |
+
"step": 117500
|
1419 |
+
},
|
1420 |
+
{
|
1421 |
+
"epoch": 10.93,
|
1422 |
+
"learning_rate": 2.2680125949249863e-05,
|
1423 |
+
"loss": 1.673,
|
1424 |
+
"step": 118000
|
1425 |
+
},
|
1426 |
+
{
|
1427 |
+
"epoch": 10.97,
|
1428 |
+
"learning_rate": 2.2564363771068718e-05,
|
1429 |
+
"loss": 1.6614,
|
1430 |
+
"step": 118500
|
1431 |
+
},
|
1432 |
+
{
|
1433 |
+
"epoch": 11.02,
|
1434 |
+
"learning_rate": 2.2448601592887572e-05,
|
1435 |
+
"loss": 1.6594,
|
1436 |
+
"step": 119000
|
1437 |
+
},
|
1438 |
+
{
|
1439 |
+
"epoch": 11.07,
|
1440 |
+
"learning_rate": 2.2332839414706427e-05,
|
1441 |
+
"loss": 1.6438,
|
1442 |
+
"step": 119500
|
1443 |
+
},
|
1444 |
+
{
|
1445 |
+
"epoch": 11.11,
|
1446 |
+
"learning_rate": 2.2217077236525282e-05,
|
1447 |
+
"loss": 1.654,
|
1448 |
+
"step": 120000
|
1449 |
+
},
|
1450 |
+
{
|
1451 |
+
"epoch": 11.16,
|
1452 |
+
"learning_rate": 2.2101315058344137e-05,
|
1453 |
+
"loss": 1.6449,
|
1454 |
+
"step": 120500
|
1455 |
+
},
|
1456 |
+
{
|
1457 |
+
"epoch": 11.21,
|
1458 |
+
"learning_rate": 2.1985552880162995e-05,
|
1459 |
+
"loss": 1.6551,
|
1460 |
+
"step": 121000
|
1461 |
+
},
|
1462 |
+
{
|
1463 |
+
"epoch": 11.25,
|
1464 |
+
"learning_rate": 2.186979070198185e-05,
|
1465 |
+
"loss": 1.6494,
|
1466 |
+
"step": 121500
|
1467 |
+
},
|
1468 |
+
{
|
1469 |
+
"epoch": 11.3,
|
1470 |
+
"learning_rate": 2.1754028523800704e-05,
|
1471 |
+
"loss": 1.6431,
|
1472 |
+
"step": 122000
|
1473 |
+
},
|
1474 |
+
{
|
1475 |
+
"epoch": 11.34,
|
1476 |
+
"learning_rate": 2.163826634561956e-05,
|
1477 |
+
"loss": 1.6609,
|
1478 |
+
"step": 122500
|
1479 |
+
},
|
1480 |
+
{
|
1481 |
+
"epoch": 11.39,
|
1482 |
+
"learning_rate": 2.1522504167438414e-05,
|
1483 |
+
"loss": 1.6547,
|
1484 |
+
"step": 123000
|
1485 |
+
},
|
1486 |
+
{
|
1487 |
+
"epoch": 11.44,
|
1488 |
+
"learning_rate": 2.140674198925727e-05,
|
1489 |
+
"loss": 1.6366,
|
1490 |
+
"step": 123500
|
1491 |
+
},
|
1492 |
+
{
|
1493 |
+
"epoch": 11.48,
|
1494 |
+
"learning_rate": 2.1290979811076127e-05,
|
1495 |
+
"loss": 1.648,
|
1496 |
+
"step": 124000
|
1497 |
+
},
|
1498 |
+
{
|
1499 |
+
"epoch": 11.53,
|
1500 |
+
"learning_rate": 2.117521763289498e-05,
|
1501 |
+
"loss": 1.6543,
|
1502 |
+
"step": 124500
|
1503 |
+
},
|
1504 |
+
{
|
1505 |
+
"epoch": 11.58,
|
1506 |
+
"learning_rate": 2.1059455454713836e-05,
|
1507 |
+
"loss": 1.6465,
|
1508 |
+
"step": 125000
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 11.62,
|
1512 |
+
"learning_rate": 2.094369327653269e-05,
|
1513 |
+
"loss": 1.6457,
|
1514 |
+
"step": 125500
|
1515 |
+
},
|
1516 |
+
{
|
1517 |
+
"epoch": 11.67,
|
1518 |
+
"learning_rate": 2.0827931098351546e-05,
|
1519 |
+
"loss": 1.6374,
|
1520 |
+
"step": 126000
|
1521 |
+
},
|
1522 |
+
{
|
1523 |
+
"epoch": 11.72,
|
1524 |
+
"learning_rate": 2.07121689201704e-05,
|
1525 |
+
"loss": 1.6409,
|
1526 |
+
"step": 126500
|
1527 |
+
},
|
1528 |
+
{
|
1529 |
+
"epoch": 11.76,
|
1530 |
+
"learning_rate": 2.059640674198926e-05,
|
1531 |
+
"loss": 1.6442,
|
1532 |
+
"step": 127000
|
1533 |
+
},
|
1534 |
+
{
|
1535 |
+
"epoch": 11.81,
|
1536 |
+
"learning_rate": 2.0480644563808114e-05,
|
1537 |
+
"loss": 1.6419,
|
1538 |
+
"step": 127500
|
1539 |
+
},
|
1540 |
+
{
|
1541 |
+
"epoch": 11.85,
|
1542 |
+
"learning_rate": 2.0364882385626968e-05,
|
1543 |
+
"loss": 1.6414,
|
1544 |
+
"step": 128000
|
1545 |
+
},
|
1546 |
+
{
|
1547 |
+
"epoch": 11.9,
|
1548 |
+
"learning_rate": 2.0249120207445823e-05,
|
1549 |
+
"loss": 1.6436,
|
1550 |
+
"step": 128500
|
1551 |
+
},
|
1552 |
+
{
|
1553 |
+
"epoch": 11.95,
|
1554 |
+
"learning_rate": 2.0133358029264678e-05,
|
1555 |
+
"loss": 1.6524,
|
1556 |
+
"step": 129000
|
1557 |
+
},
|
1558 |
+
{
|
1559 |
+
"epoch": 11.99,
|
1560 |
+
"learning_rate": 2.0017595851083536e-05,
|
1561 |
+
"loss": 1.6235,
|
1562 |
+
"step": 129500
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 12.04,
|
1566 |
+
"learning_rate": 1.990183367290239e-05,
|
1567 |
+
"loss": 1.6248,
|
1568 |
+
"step": 130000
|
1569 |
+
},
|
1570 |
+
{
|
1571 |
+
"epoch": 12.09,
|
1572 |
+
"learning_rate": 1.9786071494721245e-05,
|
1573 |
+
"loss": 1.6231,
|
1574 |
+
"step": 130500
|
1575 |
+
},
|
1576 |
+
{
|
1577 |
+
"epoch": 12.13,
|
1578 |
+
"learning_rate": 1.96703093165401e-05,
|
1579 |
+
"loss": 1.62,
|
1580 |
+
"step": 131000
|
1581 |
+
},
|
1582 |
+
{
|
1583 |
+
"epoch": 12.18,
|
1584 |
+
"learning_rate": 1.9554547138358955e-05,
|
1585 |
+
"loss": 1.6245,
|
1586 |
+
"step": 131500
|
1587 |
+
},
|
1588 |
+
{
|
1589 |
+
"epoch": 12.22,
|
1590 |
+
"learning_rate": 1.943878496017781e-05,
|
1591 |
+
"loss": 1.6211,
|
1592 |
+
"step": 132000
|
1593 |
+
},
|
1594 |
+
{
|
1595 |
+
"epoch": 12.27,
|
1596 |
+
"learning_rate": 1.9323022781996668e-05,
|
1597 |
+
"loss": 1.6165,
|
1598 |
+
"step": 132500
|
1599 |
+
},
|
1600 |
+
{
|
1601 |
+
"epoch": 12.32,
|
1602 |
+
"learning_rate": 1.9207260603815523e-05,
|
1603 |
+
"loss": 1.627,
|
1604 |
+
"step": 133000
|
1605 |
+
},
|
1606 |
+
{
|
1607 |
+
"epoch": 12.36,
|
1608 |
+
"learning_rate": 1.9091498425634377e-05,
|
1609 |
+
"loss": 1.6094,
|
1610 |
+
"step": 133500
|
1611 |
+
},
|
1612 |
+
{
|
1613 |
+
"epoch": 12.41,
|
1614 |
+
"learning_rate": 1.8975736247453232e-05,
|
1615 |
+
"loss": 1.6132,
|
1616 |
+
"step": 134000
|
1617 |
+
},
|
1618 |
+
{
|
1619 |
+
"epoch": 12.46,
|
1620 |
+
"learning_rate": 1.8859974069272087e-05,
|
1621 |
+
"loss": 1.6304,
|
1622 |
+
"step": 134500
|
1623 |
+
},
|
1624 |
+
{
|
1625 |
+
"epoch": 12.5,
|
1626 |
+
"learning_rate": 1.8744211891090942e-05,
|
1627 |
+
"loss": 1.614,
|
1628 |
+
"step": 135000
|
1629 |
+
},
|
1630 |
+
{
|
1631 |
+
"epoch": 12.55,
|
1632 |
+
"learning_rate": 1.86284497129098e-05,
|
1633 |
+
"loss": 1.6126,
|
1634 |
+
"step": 135500
|
1635 |
+
},
|
1636 |
+
{
|
1637 |
+
"epoch": 12.59,
|
1638 |
+
"learning_rate": 1.8512687534728655e-05,
|
1639 |
+
"loss": 1.6276,
|
1640 |
+
"step": 136000
|
1641 |
+
},
|
1642 |
+
{
|
1643 |
+
"epoch": 12.64,
|
1644 |
+
"learning_rate": 1.839692535654751e-05,
|
1645 |
+
"loss": 1.617,
|
1646 |
+
"step": 136500
|
1647 |
+
},
|
1648 |
+
{
|
1649 |
+
"epoch": 12.69,
|
1650 |
+
"learning_rate": 1.8281163178366364e-05,
|
1651 |
+
"loss": 1.6264,
|
1652 |
+
"step": 137000
|
1653 |
+
},
|
1654 |
+
{
|
1655 |
+
"epoch": 12.73,
|
1656 |
+
"learning_rate": 1.816540100018522e-05,
|
1657 |
+
"loss": 1.6248,
|
1658 |
+
"step": 137500
|
1659 |
+
},
|
1660 |
+
{
|
1661 |
+
"epoch": 12.78,
|
1662 |
+
"learning_rate": 1.8049638822004074e-05,
|
1663 |
+
"loss": 1.613,
|
1664 |
+
"step": 138000
|
1665 |
+
},
|
1666 |
+
{
|
1667 |
+
"epoch": 12.83,
|
1668 |
+
"learning_rate": 1.7933876643822932e-05,
|
1669 |
+
"loss": 1.6153,
|
1670 |
+
"step": 138500
|
1671 |
+
},
|
1672 |
+
{
|
1673 |
+
"epoch": 12.87,
|
1674 |
+
"learning_rate": 1.7818114465641787e-05,
|
1675 |
+
"loss": 1.617,
|
1676 |
+
"step": 139000
|
1677 |
+
},
|
1678 |
+
{
|
1679 |
+
"epoch": 12.92,
|
1680 |
+
"learning_rate": 1.770235228746064e-05,
|
1681 |
+
"loss": 1.614,
|
1682 |
+
"step": 139500
|
1683 |
+
},
|
1684 |
+
{
|
1685 |
+
"epoch": 12.97,
|
1686 |
+
"learning_rate": 1.7586590109279496e-05,
|
1687 |
+
"loss": 1.6237,
|
1688 |
+
"step": 140000
|
1689 |
+
},
|
1690 |
+
{
|
1691 |
+
"epoch": 13.01,
|
1692 |
+
"learning_rate": 1.747082793109835e-05,
|
1693 |
+
"loss": 1.6131,
|
1694 |
+
"step": 140500
|
1695 |
+
},
|
1696 |
+
{
|
1697 |
+
"epoch": 13.06,
|
1698 |
+
"learning_rate": 1.7355065752917206e-05,
|
1699 |
+
"loss": 1.5968,
|
1700 |
+
"step": 141000
|
1701 |
+
},
|
1702 |
+
{
|
1703 |
+
"epoch": 13.1,
|
1704 |
+
"learning_rate": 1.7239303574736064e-05,
|
1705 |
+
"loss": 1.5895,
|
1706 |
+
"step": 141500
|
1707 |
+
},
|
1708 |
+
{
|
1709 |
+
"epoch": 13.15,
|
1710 |
+
"learning_rate": 1.712354139655492e-05,
|
1711 |
+
"loss": 1.5985,
|
1712 |
+
"step": 142000
|
1713 |
+
},
|
1714 |
+
{
|
1715 |
+
"epoch": 13.2,
|
1716 |
+
"learning_rate": 1.7007779218373773e-05,
|
1717 |
+
"loss": 1.5979,
|
1718 |
+
"step": 142500
|
1719 |
+
},
|
1720 |
+
{
|
1721 |
+
"epoch": 13.24,
|
1722 |
+
"learning_rate": 1.6892017040192628e-05,
|
1723 |
+
"loss": 1.6028,
|
1724 |
+
"step": 143000
|
1725 |
+
},
|
1726 |
+
{
|
1727 |
+
"epoch": 13.29,
|
1728 |
+
"learning_rate": 1.6776254862011483e-05,
|
1729 |
+
"loss": 1.5905,
|
1730 |
+
"step": 143500
|
1731 |
+
},
|
1732 |
+
{
|
1733 |
+
"epoch": 13.34,
|
1734 |
+
"learning_rate": 1.6660492683830338e-05,
|
1735 |
+
"loss": 1.5731,
|
1736 |
+
"step": 144000
|
1737 |
+
},
|
1738 |
+
{
|
1739 |
+
"epoch": 13.38,
|
1740 |
+
"learning_rate": 1.6544730505649196e-05,
|
1741 |
+
"loss": 1.5873,
|
1742 |
+
"step": 144500
|
1743 |
+
},
|
1744 |
+
{
|
1745 |
+
"epoch": 13.43,
|
1746 |
+
"learning_rate": 1.642896832746805e-05,
|
1747 |
+
"loss": 1.5985,
|
1748 |
+
"step": 145000
|
1749 |
+
},
|
1750 |
+
{
|
1751 |
+
"epoch": 13.47,
|
1752 |
+
"learning_rate": 1.6313206149286905e-05,
|
1753 |
+
"loss": 1.587,
|
1754 |
+
"step": 145500
|
1755 |
+
},
|
1756 |
+
{
|
1757 |
+
"epoch": 13.52,
|
1758 |
+
"learning_rate": 1.619744397110576e-05,
|
1759 |
+
"loss": 1.6003,
|
1760 |
+
"step": 146000
|
1761 |
+
},
|
1762 |
+
{
|
1763 |
+
"epoch": 13.57,
|
1764 |
+
"learning_rate": 1.6081681792924615e-05,
|
1765 |
+
"loss": 1.581,
|
1766 |
+
"step": 146500
|
1767 |
+
},
|
1768 |
+
{
|
1769 |
+
"epoch": 13.61,
|
1770 |
+
"learning_rate": 1.5965919614743473e-05,
|
1771 |
+
"loss": 1.5993,
|
1772 |
+
"step": 147000
|
1773 |
+
},
|
1774 |
+
{
|
1775 |
+
"epoch": 13.66,
|
1776 |
+
"learning_rate": 1.5850157436562328e-05,
|
1777 |
+
"loss": 1.5928,
|
1778 |
+
"step": 147500
|
1779 |
+
},
|
1780 |
+
{
|
1781 |
+
"epoch": 13.71,
|
1782 |
+
"learning_rate": 1.5734395258381183e-05,
|
1783 |
+
"loss": 1.5953,
|
1784 |
+
"step": 148000
|
1785 |
+
},
|
1786 |
+
{
|
1787 |
+
"epoch": 13.75,
|
1788 |
+
"learning_rate": 1.5618633080200037e-05,
|
1789 |
+
"loss": 1.5915,
|
1790 |
+
"step": 148500
|
1791 |
+
},
|
1792 |
+
{
|
1793 |
+
"epoch": 13.8,
|
1794 |
+
"learning_rate": 1.5502870902018892e-05,
|
1795 |
+
"loss": 1.5883,
|
1796 |
+
"step": 149000
|
1797 |
+
},
|
1798 |
+
{
|
1799 |
+
"epoch": 13.85,
|
1800 |
+
"learning_rate": 1.5387108723837747e-05,
|
1801 |
+
"loss": 1.5988,
|
1802 |
+
"step": 149500
|
1803 |
+
},
|
1804 |
+
{
|
1805 |
+
"epoch": 13.89,
|
1806 |
+
"learning_rate": 1.5271346545656605e-05,
|
1807 |
+
"loss": 1.5909,
|
1808 |
+
"step": 150000
|
1809 |
+
},
|
1810 |
+
{
|
1811 |
+
"epoch": 13.94,
|
1812 |
+
"learning_rate": 1.5155584367475458e-05,
|
1813 |
+
"loss": 1.5992,
|
1814 |
+
"step": 150500
|
1815 |
+
},
|
1816 |
+
{
|
1817 |
+
"epoch": 13.98,
|
1818 |
+
"learning_rate": 1.5039822189294315e-05,
|
1819 |
+
"loss": 1.5851,
|
1820 |
+
"step": 151000
|
1821 |
+
},
|
1822 |
+
{
|
1823 |
+
"epoch": 14.03,
|
1824 |
+
"learning_rate": 1.492406001111317e-05,
|
1825 |
+
"loss": 1.5705,
|
1826 |
+
"step": 151500
|
1827 |
+
},
|
1828 |
+
{
|
1829 |
+
"epoch": 14.08,
|
1830 |
+
"learning_rate": 1.4808297832932024e-05,
|
1831 |
+
"loss": 1.5777,
|
1832 |
+
"step": 152000
|
1833 |
+
},
|
1834 |
+
{
|
1835 |
+
"epoch": 14.12,
|
1836 |
+
"learning_rate": 1.469253565475088e-05,
|
1837 |
+
"loss": 1.5911,
|
1838 |
+
"step": 152500
|
1839 |
+
},
|
1840 |
+
{
|
1841 |
+
"epoch": 14.17,
|
1842 |
+
"learning_rate": 1.4576773476569735e-05,
|
1843 |
+
"loss": 1.5808,
|
1844 |
+
"step": 153000
|
1845 |
+
},
|
1846 |
+
{
|
1847 |
+
"epoch": 14.22,
|
1848 |
+
"learning_rate": 1.446101129838859e-05,
|
1849 |
+
"loss": 1.5692,
|
1850 |
+
"step": 153500
|
1851 |
+
},
|
1852 |
+
{
|
1853 |
+
"epoch": 14.26,
|
1854 |
+
"learning_rate": 1.4345249120207447e-05,
|
1855 |
+
"loss": 1.5751,
|
1856 |
+
"step": 154000
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 14.31,
|
1860 |
+
"learning_rate": 1.4229486942026301e-05,
|
1861 |
+
"loss": 1.5767,
|
1862 |
+
"step": 154500
|
1863 |
+
},
|
1864 |
+
{
|
1865 |
+
"epoch": 14.35,
|
1866 |
+
"learning_rate": 1.4113724763845156e-05,
|
1867 |
+
"loss": 1.5695,
|
1868 |
+
"step": 155000
|
1869 |
+
},
|
1870 |
+
{
|
1871 |
+
"epoch": 14.4,
|
1872 |
+
"learning_rate": 1.3997962585664013e-05,
|
1873 |
+
"loss": 1.5718,
|
1874 |
+
"step": 155500
|
1875 |
+
},
|
1876 |
+
{
|
1877 |
+
"epoch": 14.45,
|
1878 |
+
"learning_rate": 1.3882200407482867e-05,
|
1879 |
+
"loss": 1.5737,
|
1880 |
+
"step": 156000
|
1881 |
+
},
|
1882 |
+
{
|
1883 |
+
"epoch": 14.49,
|
1884 |
+
"learning_rate": 1.3766438229301724e-05,
|
1885 |
+
"loss": 1.5744,
|
1886 |
+
"step": 156500
|
1887 |
+
},
|
1888 |
+
{
|
1889 |
+
"epoch": 14.54,
|
1890 |
+
"learning_rate": 1.3650676051120579e-05,
|
1891 |
+
"loss": 1.5675,
|
1892 |
+
"step": 157000
|
1893 |
+
},
|
1894 |
+
{
|
1895 |
+
"epoch": 14.59,
|
1896 |
+
"learning_rate": 1.3534913872939433e-05,
|
1897 |
+
"loss": 1.5676,
|
1898 |
+
"step": 157500
|
1899 |
+
},
|
1900 |
+
{
|
1901 |
+
"epoch": 14.63,
|
1902 |
+
"learning_rate": 1.341915169475829e-05,
|
1903 |
+
"loss": 1.5693,
|
1904 |
+
"step": 158000
|
1905 |
+
},
|
1906 |
+
{
|
1907 |
+
"epoch": 14.68,
|
1908 |
+
"learning_rate": 1.3303389516577145e-05,
|
1909 |
+
"loss": 1.5728,
|
1910 |
+
"step": 158500
|
1911 |
+
},
|
1912 |
+
{
|
1913 |
+
"epoch": 14.72,
|
1914 |
+
"learning_rate": 1.3187627338396e-05,
|
1915 |
+
"loss": 1.5673,
|
1916 |
+
"step": 159000
|
1917 |
+
},
|
1918 |
+
{
|
1919 |
+
"epoch": 14.77,
|
1920 |
+
"learning_rate": 1.3071865160214856e-05,
|
1921 |
+
"loss": 1.5688,
|
1922 |
+
"step": 159500
|
1923 |
+
},
|
1924 |
+
{
|
1925 |
+
"epoch": 14.82,
|
1926 |
+
"learning_rate": 1.295610298203371e-05,
|
1927 |
+
"loss": 1.5695,
|
1928 |
+
"step": 160000
|
1929 |
+
},
|
1930 |
+
{
|
1931 |
+
"epoch": 14.86,
|
1932 |
+
"learning_rate": 1.2840340803852565e-05,
|
1933 |
+
"loss": 1.5692,
|
1934 |
+
"step": 160500
|
1935 |
+
},
|
1936 |
+
{
|
1937 |
+
"epoch": 14.91,
|
1938 |
+
"learning_rate": 1.2724578625671422e-05,
|
1939 |
+
"loss": 1.5683,
|
1940 |
+
"step": 161000
|
1941 |
+
},
|
1942 |
+
{
|
1943 |
+
"epoch": 14.96,
|
1944 |
+
"learning_rate": 1.2608816447490277e-05,
|
1945 |
+
"loss": 1.5616,
|
1946 |
+
"step": 161500
|
1947 |
+
},
|
1948 |
+
{
|
1949 |
+
"epoch": 15.0,
|
1950 |
+
"learning_rate": 1.2493054269309131e-05,
|
1951 |
+
"loss": 1.5607,
|
1952 |
+
"step": 162000
|
1953 |
+
},
|
1954 |
+
{
|
1955 |
+
"epoch": 15.05,
|
1956 |
+
"learning_rate": 1.2377292091127988e-05,
|
1957 |
+
"loss": 1.5651,
|
1958 |
+
"step": 162500
|
1959 |
+
},
|
1960 |
+
{
|
1961 |
+
"epoch": 15.1,
|
1962 |
+
"learning_rate": 1.2261529912946843e-05,
|
1963 |
+
"loss": 1.5632,
|
1964 |
+
"step": 163000
|
1965 |
+
},
|
1966 |
+
{
|
1967 |
+
"epoch": 15.14,
|
1968 |
+
"learning_rate": 1.2145767734765697e-05,
|
1969 |
+
"loss": 1.557,
|
1970 |
+
"step": 163500
|
1971 |
+
},
|
1972 |
+
{
|
1973 |
+
"epoch": 15.19,
|
1974 |
+
"learning_rate": 1.2030005556584554e-05,
|
1975 |
+
"loss": 1.5654,
|
1976 |
+
"step": 164000
|
1977 |
+
},
|
1978 |
+
{
|
1979 |
+
"epoch": 15.23,
|
1980 |
+
"learning_rate": 1.1914243378403409e-05,
|
1981 |
+
"loss": 1.5583,
|
1982 |
+
"step": 164500
|
1983 |
+
},
|
1984 |
+
{
|
1985 |
+
"epoch": 15.28,
|
1986 |
+
"learning_rate": 1.1798481200222263e-05,
|
1987 |
+
"loss": 1.5521,
|
1988 |
+
"step": 165000
|
1989 |
+
},
|
1990 |
+
{
|
1991 |
+
"epoch": 15.33,
|
1992 |
+
"learning_rate": 1.168271902204112e-05,
|
1993 |
+
"loss": 1.5549,
|
1994 |
+
"step": 165500
|
1995 |
+
},
|
1996 |
+
{
|
1997 |
+
"epoch": 15.37,
|
1998 |
+
"learning_rate": 1.1566956843859975e-05,
|
1999 |
+
"loss": 1.5567,
|
2000 |
+
"step": 166000
|
2001 |
+
},
|
2002 |
+
{
|
2003 |
+
"epoch": 15.42,
|
2004 |
+
"learning_rate": 1.145119466567883e-05,
|
2005 |
+
"loss": 1.5558,
|
2006 |
+
"step": 166500
|
2007 |
+
},
|
2008 |
+
{
|
2009 |
+
"epoch": 15.47,
|
2010 |
+
"learning_rate": 1.1335432487497686e-05,
|
2011 |
+
"loss": 1.566,
|
2012 |
+
"step": 167000
|
2013 |
+
},
|
2014 |
+
{
|
2015 |
+
"epoch": 15.51,
|
2016 |
+
"learning_rate": 1.121967030931654e-05,
|
2017 |
+
"loss": 1.5617,
|
2018 |
+
"step": 167500
|
2019 |
+
},
|
2020 |
+
{
|
2021 |
+
"epoch": 15.56,
|
2022 |
+
"learning_rate": 1.1103908131135395e-05,
|
2023 |
+
"loss": 1.5511,
|
2024 |
+
"step": 168000
|
2025 |
+
},
|
2026 |
+
{
|
2027 |
+
"epoch": 15.6,
|
2028 |
+
"learning_rate": 1.0988145952954252e-05,
|
2029 |
+
"loss": 1.5536,
|
2030 |
+
"step": 168500
|
2031 |
+
},
|
2032 |
+
{
|
2033 |
+
"epoch": 15.65,
|
2034 |
+
"learning_rate": 1.0872383774773107e-05,
|
2035 |
+
"loss": 1.5402,
|
2036 |
+
"step": 169000
|
2037 |
+
},
|
2038 |
+
{
|
2039 |
+
"epoch": 15.7,
|
2040 |
+
"learning_rate": 1.0756621596591961e-05,
|
2041 |
+
"loss": 1.5579,
|
2042 |
+
"step": 169500
|
2043 |
+
},
|
2044 |
+
{
|
2045 |
+
"epoch": 15.74,
|
2046 |
+
"learning_rate": 1.0640859418410818e-05,
|
2047 |
+
"loss": 1.5408,
|
2048 |
+
"step": 170000
|
2049 |
+
},
|
2050 |
+
{
|
2051 |
+
"epoch": 15.79,
|
2052 |
+
"learning_rate": 1.0525097240229673e-05,
|
2053 |
+
"loss": 1.5642,
|
2054 |
+
"step": 170500
|
2055 |
+
},
|
2056 |
+
{
|
2057 |
+
"epoch": 15.84,
|
2058 |
+
"learning_rate": 1.0409335062048527e-05,
|
2059 |
+
"loss": 1.5528,
|
2060 |
+
"step": 171000
|
2061 |
+
},
|
2062 |
+
{
|
2063 |
+
"epoch": 15.88,
|
2064 |
+
"learning_rate": 1.0293572883867384e-05,
|
2065 |
+
"loss": 1.5447,
|
2066 |
+
"step": 171500
|
2067 |
+
},
|
2068 |
+
{
|
2069 |
+
"epoch": 15.93,
|
2070 |
+
"learning_rate": 1.0177810705686239e-05,
|
2071 |
+
"loss": 1.5373,
|
2072 |
+
"step": 172000
|
2073 |
+
},
|
2074 |
+
{
|
2075 |
+
"epoch": 15.98,
|
2076 |
+
"learning_rate": 1.0062048527505093e-05,
|
2077 |
+
"loss": 1.5409,
|
2078 |
+
"step": 172500
|
2079 |
+
},
|
2080 |
+
{
|
2081 |
+
"epoch": 16.02,
|
2082 |
+
"learning_rate": 9.94628634932395e-06,
|
2083 |
+
"loss": 1.5315,
|
2084 |
+
"step": 173000
|
2085 |
+
},
|
2086 |
+
{
|
2087 |
+
"epoch": 16.07,
|
2088 |
+
"learning_rate": 9.830524171142805e-06,
|
2089 |
+
"loss": 1.5454,
|
2090 |
+
"step": 173500
|
2091 |
+
},
|
2092 |
+
{
|
2093 |
+
"epoch": 16.11,
|
2094 |
+
"learning_rate": 9.71476199296166e-06,
|
2095 |
+
"loss": 1.5313,
|
2096 |
+
"step": 174000
|
2097 |
+
},
|
2098 |
+
{
|
2099 |
+
"epoch": 16.16,
|
2100 |
+
"learning_rate": 9.598999814780516e-06,
|
2101 |
+
"loss": 1.5369,
|
2102 |
+
"step": 174500
|
2103 |
+
},
|
2104 |
+
{
|
2105 |
+
"epoch": 16.21,
|
2106 |
+
"learning_rate": 9.48323763659937e-06,
|
2107 |
+
"loss": 1.5318,
|
2108 |
+
"step": 175000
|
2109 |
+
},
|
2110 |
+
{
|
2111 |
+
"epoch": 16.25,
|
2112 |
+
"learning_rate": 9.367475458418227e-06,
|
2113 |
+
"loss": 1.5368,
|
2114 |
+
"step": 175500
|
2115 |
+
},
|
2116 |
+
{
|
2117 |
+
"epoch": 16.3,
|
2118 |
+
"learning_rate": 9.251713280237082e-06,
|
2119 |
+
"loss": 1.5381,
|
2120 |
+
"step": 176000
|
2121 |
+
},
|
2122 |
+
{
|
2123 |
+
"epoch": 16.35,
|
2124 |
+
"learning_rate": 9.135951102055937e-06,
|
2125 |
+
"loss": 1.5295,
|
2126 |
+
"step": 176500
|
2127 |
+
},
|
2128 |
+
{
|
2129 |
+
"epoch": 16.39,
|
2130 |
+
"learning_rate": 9.020188923874793e-06,
|
2131 |
+
"loss": 1.5444,
|
2132 |
+
"step": 177000
|
2133 |
+
},
|
2134 |
+
{
|
2135 |
+
"epoch": 16.44,
|
2136 |
+
"learning_rate": 8.904426745693648e-06,
|
2137 |
+
"loss": 1.5285,
|
2138 |
+
"step": 177500
|
2139 |
+
},
|
2140 |
+
{
|
2141 |
+
"epoch": 16.48,
|
2142 |
+
"learning_rate": 8.788664567512503e-06,
|
2143 |
+
"loss": 1.5327,
|
2144 |
+
"step": 178000
|
2145 |
+
},
|
2146 |
+
{
|
2147 |
+
"epoch": 16.53,
|
2148 |
+
"learning_rate": 8.672902389331359e-06,
|
2149 |
+
"loss": 1.5321,
|
2150 |
+
"step": 178500
|
2151 |
+
},
|
2152 |
+
{
|
2153 |
+
"epoch": 16.58,
|
2154 |
+
"learning_rate": 8.557140211150214e-06,
|
2155 |
+
"loss": 1.5332,
|
2156 |
+
"step": 179000
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 16.62,
|
2160 |
+
"learning_rate": 8.441378032969068e-06,
|
2161 |
+
"loss": 1.5408,
|
2162 |
+
"step": 179500
|
2163 |
+
},
|
2164 |
+
{
|
2165 |
+
"epoch": 16.67,
|
2166 |
+
"learning_rate": 8.325615854787925e-06,
|
2167 |
+
"loss": 1.5332,
|
2168 |
+
"step": 180000
|
2169 |
+
},
|
2170 |
+
{
|
2171 |
+
"epoch": 16.72,
|
2172 |
+
"learning_rate": 8.20985367660678e-06,
|
2173 |
+
"loss": 1.5294,
|
2174 |
+
"step": 180500
|
2175 |
+
},
|
2176 |
+
{
|
2177 |
+
"epoch": 16.76,
|
2178 |
+
"learning_rate": 8.094091498425634e-06,
|
2179 |
+
"loss": 1.5361,
|
2180 |
+
"step": 181000
|
2181 |
+
},
|
2182 |
+
{
|
2183 |
+
"epoch": 16.81,
|
2184 |
+
"learning_rate": 7.978329320244491e-06,
|
2185 |
+
"loss": 1.5229,
|
2186 |
+
"step": 181500
|
2187 |
+
},
|
2188 |
+
{
|
2189 |
+
"epoch": 16.85,
|
2190 |
+
"learning_rate": 7.862567142063346e-06,
|
2191 |
+
"loss": 1.5337,
|
2192 |
+
"step": 182000
|
2193 |
+
},
|
2194 |
+
{
|
2195 |
+
"epoch": 16.9,
|
2196 |
+
"learning_rate": 7.7468049638822e-06,
|
2197 |
+
"loss": 1.5323,
|
2198 |
+
"step": 182500
|
2199 |
+
},
|
2200 |
+
{
|
2201 |
+
"epoch": 16.95,
|
2202 |
+
"learning_rate": 7.631042785701057e-06,
|
2203 |
+
"loss": 1.5376,
|
2204 |
+
"step": 183000
|
2205 |
+
},
|
2206 |
+
{
|
2207 |
+
"epoch": 16.99,
|
2208 |
+
"learning_rate": 7.515280607519912e-06,
|
2209 |
+
"loss": 1.5251,
|
2210 |
+
"step": 183500
|
2211 |
+
},
|
2212 |
+
{
|
2213 |
+
"epoch": 17.04,
|
2214 |
+
"learning_rate": 7.399518429338767e-06,
|
2215 |
+
"loss": 1.523,
|
2216 |
+
"step": 184000
|
2217 |
+
},
|
2218 |
+
{
|
2219 |
+
"epoch": 17.09,
|
2220 |
+
"learning_rate": 7.283756251157622e-06,
|
2221 |
+
"loss": 1.5172,
|
2222 |
+
"step": 184500
|
2223 |
+
},
|
2224 |
+
{
|
2225 |
+
"epoch": 17.13,
|
2226 |
+
"learning_rate": 7.167994072976478e-06,
|
2227 |
+
"loss": 1.511,
|
2228 |
+
"step": 185000
|
2229 |
+
},
|
2230 |
+
{
|
2231 |
+
"epoch": 17.18,
|
2232 |
+
"learning_rate": 7.052231894795333e-06,
|
2233 |
+
"loss": 1.5108,
|
2234 |
+
"step": 185500
|
2235 |
+
},
|
2236 |
+
{
|
2237 |
+
"epoch": 17.23,
|
2238 |
+
"learning_rate": 6.936469716614188e-06,
|
2239 |
+
"loss": 1.5133,
|
2240 |
+
"step": 186000
|
2241 |
+
},
|
2242 |
+
{
|
2243 |
+
"epoch": 17.27,
|
2244 |
+
"learning_rate": 6.820707538433044e-06,
|
2245 |
+
"loss": 1.5204,
|
2246 |
+
"step": 186500
|
2247 |
+
},
|
2248 |
+
{
|
2249 |
+
"epoch": 17.32,
|
2250 |
+
"learning_rate": 6.704945360251899e-06,
|
2251 |
+
"loss": 1.5299,
|
2252 |
+
"step": 187000
|
2253 |
+
},
|
2254 |
+
{
|
2255 |
+
"epoch": 17.36,
|
2256 |
+
"learning_rate": 6.589183182070754e-06,
|
2257 |
+
"loss": 1.5127,
|
2258 |
+
"step": 187500
|
2259 |
+
},
|
2260 |
+
{
|
2261 |
+
"epoch": 17.41,
|
2262 |
+
"learning_rate": 6.47342100388961e-06,
|
2263 |
+
"loss": 1.5191,
|
2264 |
+
"step": 188000
|
2265 |
+
},
|
2266 |
+
{
|
2267 |
+
"epoch": 17.46,
|
2268 |
+
"learning_rate": 6.357658825708465e-06,
|
2269 |
+
"loss": 1.5263,
|
2270 |
+
"step": 188500
|
2271 |
+
},
|
2272 |
+
{
|
2273 |
+
"epoch": 17.5,
|
2274 |
+
"learning_rate": 6.241896647527321e-06,
|
2275 |
+
"loss": 1.5064,
|
2276 |
+
"step": 189000
|
2277 |
+
},
|
2278 |
+
{
|
2279 |
+
"epoch": 17.55,
|
2280 |
+
"learning_rate": 6.126134469346176e-06,
|
2281 |
+
"loss": 1.5267,
|
2282 |
+
"step": 189500
|
2283 |
+
},
|
2284 |
+
{
|
2285 |
+
"epoch": 17.6,
|
2286 |
+
"learning_rate": 6.0103722911650304e-06,
|
2287 |
+
"loss": 1.519,
|
2288 |
+
"step": 190000
|
2289 |
+
},
|
2290 |
+
{
|
2291 |
+
"epoch": 17.64,
|
2292 |
+
"learning_rate": 5.894610112983886e-06,
|
2293 |
+
"loss": 1.5188,
|
2294 |
+
"step": 190500
|
2295 |
+
},
|
2296 |
+
{
|
2297 |
+
"epoch": 17.69,
|
2298 |
+
"learning_rate": 5.778847934802741e-06,
|
2299 |
+
"loss": 1.5149,
|
2300 |
+
"step": 191000
|
2301 |
+
},
|
2302 |
+
{
|
2303 |
+
"epoch": 17.73,
|
2304 |
+
"learning_rate": 5.6630857566215964e-06,
|
2305 |
+
"loss": 1.512,
|
2306 |
+
"step": 191500
|
2307 |
+
},
|
2308 |
+
{
|
2309 |
+
"epoch": 17.78,
|
2310 |
+
"learning_rate": 5.547323578440452e-06,
|
2311 |
+
"loss": 1.5161,
|
2312 |
+
"step": 192000
|
2313 |
+
},
|
2314 |
+
{
|
2315 |
+
"epoch": 17.83,
|
2316 |
+
"learning_rate": 5.431561400259307e-06,
|
2317 |
+
"loss": 1.5128,
|
2318 |
+
"step": 192500
|
2319 |
+
},
|
2320 |
+
{
|
2321 |
+
"epoch": 17.87,
|
2322 |
+
"learning_rate": 5.3157992220781624e-06,
|
2323 |
+
"loss": 1.5151,
|
2324 |
+
"step": 193000
|
2325 |
+
},
|
2326 |
+
{
|
2327 |
+
"epoch": 17.92,
|
2328 |
+
"learning_rate": 5.200037043897018e-06,
|
2329 |
+
"loss": 1.5024,
|
2330 |
+
"step": 193500
|
2331 |
+
},
|
2332 |
+
{
|
2333 |
+
"epoch": 17.97,
|
2334 |
+
"learning_rate": 5.084274865715874e-06,
|
2335 |
+
"loss": 1.5159,
|
2336 |
+
"step": 194000
|
2337 |
+
},
|
2338 |
+
{
|
2339 |
+
"epoch": 18.01,
|
2340 |
+
"learning_rate": 4.9685126875347284e-06,
|
2341 |
+
"loss": 1.5161,
|
2342 |
+
"step": 194500
|
2343 |
+
},
|
2344 |
+
{
|
2345 |
+
"epoch": 18.06,
|
2346 |
+
"learning_rate": 4.852750509353584e-06,
|
2347 |
+
"loss": 1.5198,
|
2348 |
+
"step": 195000
|
2349 |
+
},
|
2350 |
+
{
|
2351 |
+
"epoch": 18.11,
|
2352 |
+
"learning_rate": 4.73698833117244e-06,
|
2353 |
+
"loss": 1.5117,
|
2354 |
+
"step": 195500
|
2355 |
+
},
|
2356 |
+
{
|
2357 |
+
"epoch": 18.15,
|
2358 |
+
"learning_rate": 4.621226152991294e-06,
|
2359 |
+
"loss": 1.5167,
|
2360 |
+
"step": 196000
|
2361 |
+
},
|
2362 |
+
{
|
2363 |
+
"epoch": 18.2,
|
2364 |
+
"learning_rate": 4.50546397481015e-06,
|
2365 |
+
"loss": 1.5066,
|
2366 |
+
"step": 196500
|
2367 |
+
},
|
2368 |
+
{
|
2369 |
+
"epoch": 18.24,
|
2370 |
+
"learning_rate": 4.389701796629006e-06,
|
2371 |
+
"loss": 1.5011,
|
2372 |
+
"step": 197000
|
2373 |
+
},
|
2374 |
+
{
|
2375 |
+
"epoch": 18.29,
|
2376 |
+
"learning_rate": 4.27393961844786e-06,
|
2377 |
+
"loss": 1.5059,
|
2378 |
+
"step": 197500
|
2379 |
+
},
|
2380 |
+
{
|
2381 |
+
"epoch": 18.34,
|
2382 |
+
"learning_rate": 4.158177440266716e-06,
|
2383 |
+
"loss": 1.4973,
|
2384 |
+
"step": 198000
|
2385 |
+
},
|
2386 |
+
{
|
2387 |
+
"epoch": 18.38,
|
2388 |
+
"learning_rate": 4.042415262085572e-06,
|
2389 |
+
"loss": 1.5023,
|
2390 |
+
"step": 198500
|
2391 |
+
},
|
2392 |
+
{
|
2393 |
+
"epoch": 18.43,
|
2394 |
+
"learning_rate": 3.926653083904426e-06,
|
2395 |
+
"loss": 1.5023,
|
2396 |
+
"step": 199000
|
2397 |
+
},
|
2398 |
+
{
|
2399 |
+
"epoch": 18.48,
|
2400 |
+
"learning_rate": 3.810890905723282e-06,
|
2401 |
+
"loss": 1.5106,
|
2402 |
+
"step": 199500
|
2403 |
+
},
|
2404 |
+
{
|
2405 |
+
"epoch": 18.52,
|
2406 |
+
"learning_rate": 3.6951287275421376e-06,
|
2407 |
+
"loss": 1.5022,
|
2408 |
+
"step": 200000
|
2409 |
+
},
|
2410 |
+
{
|
2411 |
+
"epoch": 18.57,
|
2412 |
+
"learning_rate": 3.579366549360993e-06,
|
2413 |
+
"loss": 1.5033,
|
2414 |
+
"step": 200500
|
2415 |
+
},
|
2416 |
+
{
|
2417 |
+
"epoch": 18.61,
|
2418 |
+
"learning_rate": 3.463604371179848e-06,
|
2419 |
+
"loss": 1.5157,
|
2420 |
+
"step": 201000
|
2421 |
+
},
|
2422 |
+
{
|
2423 |
+
"epoch": 18.66,
|
2424 |
+
"learning_rate": 3.3478421929987036e-06,
|
2425 |
+
"loss": 1.4882,
|
2426 |
+
"step": 201500
|
2427 |
+
},
|
2428 |
+
{
|
2429 |
+
"epoch": 18.71,
|
2430 |
+
"learning_rate": 3.232080014817559e-06,
|
2431 |
+
"loss": 1.5025,
|
2432 |
+
"step": 202000
|
2433 |
+
},
|
2434 |
+
{
|
2435 |
+
"epoch": 18.75,
|
2436 |
+
"learning_rate": 3.116317836636414e-06,
|
2437 |
+
"loss": 1.498,
|
2438 |
+
"step": 202500
|
2439 |
+
},
|
2440 |
+
{
|
2441 |
+
"epoch": 18.8,
|
2442 |
+
"learning_rate": 3.0005556584552696e-06,
|
2443 |
+
"loss": 1.5017,
|
2444 |
+
"step": 203000
|
2445 |
+
},
|
2446 |
+
{
|
2447 |
+
"epoch": 18.85,
|
2448 |
+
"learning_rate": 2.884793480274125e-06,
|
2449 |
+
"loss": 1.5127,
|
2450 |
+
"step": 203500
|
2451 |
+
},
|
2452 |
+
{
|
2453 |
+
"epoch": 18.89,
|
2454 |
+
"learning_rate": 2.76903130209298e-06,
|
2455 |
+
"loss": 1.5089,
|
2456 |
+
"step": 204000
|
2457 |
+
},
|
2458 |
+
{
|
2459 |
+
"epoch": 18.94,
|
2460 |
+
"learning_rate": 2.6532691239118356e-06,
|
2461 |
+
"loss": 1.4965,
|
2462 |
+
"step": 204500
|
2463 |
+
},
|
2464 |
+
{
|
2465 |
+
"epoch": 18.98,
|
2466 |
+
"learning_rate": 2.537506945730691e-06,
|
2467 |
+
"loss": 1.5047,
|
2468 |
+
"step": 205000
|
2469 |
+
},
|
2470 |
+
{
|
2471 |
+
"epoch": 19.03,
|
2472 |
+
"learning_rate": 2.4217447675495464e-06,
|
2473 |
+
"loss": 1.4945,
|
2474 |
+
"step": 205500
|
2475 |
+
},
|
2476 |
+
{
|
2477 |
+
"epoch": 19.08,
|
2478 |
+
"learning_rate": 2.3059825893684016e-06,
|
2479 |
+
"loss": 1.4859,
|
2480 |
+
"step": 206000
|
2481 |
+
},
|
2482 |
+
{
|
2483 |
+
"epoch": 19.12,
|
2484 |
+
"learning_rate": 2.190220411187257e-06,
|
2485 |
+
"loss": 1.4915,
|
2486 |
+
"step": 206500
|
2487 |
+
},
|
2488 |
+
{
|
2489 |
+
"epoch": 19.17,
|
2490 |
+
"learning_rate": 2.0744582330061124e-06,
|
2491 |
+
"loss": 1.4989,
|
2492 |
+
"step": 207000
|
2493 |
+
},
|
2494 |
+
{
|
2495 |
+
"epoch": 19.22,
|
2496 |
+
"learning_rate": 1.9586960548249676e-06,
|
2497 |
+
"loss": 1.4926,
|
2498 |
+
"step": 207500
|
2499 |
+
},
|
2500 |
+
{
|
2501 |
+
"epoch": 19.26,
|
2502 |
+
"learning_rate": 1.842933876643823e-06,
|
2503 |
+
"loss": 1.5003,
|
2504 |
+
"step": 208000
|
2505 |
+
},
|
2506 |
+
{
|
2507 |
+
"epoch": 19.31,
|
2508 |
+
"learning_rate": 1.7271716984626784e-06,
|
2509 |
+
"loss": 1.5043,
|
2510 |
+
"step": 208500
|
2511 |
+
},
|
2512 |
+
{
|
2513 |
+
"epoch": 19.36,
|
2514 |
+
"learning_rate": 1.6114095202815336e-06,
|
2515 |
+
"loss": 1.5005,
|
2516 |
+
"step": 209000
|
2517 |
+
},
|
2518 |
+
{
|
2519 |
+
"epoch": 19.4,
|
2520 |
+
"learning_rate": 1.495647342100389e-06,
|
2521 |
+
"loss": 1.4882,
|
2522 |
+
"step": 209500
|
2523 |
+
},
|
2524 |
+
{
|
2525 |
+
"epoch": 19.45,
|
2526 |
+
"learning_rate": 1.3798851639192444e-06,
|
2527 |
+
"loss": 1.4973,
|
2528 |
+
"step": 210000
|
2529 |
+
}
|
2530 |
+
],
|
2531 |
+
"max_steps": 215960,
|
2532 |
+
"num_train_epochs": 20,
|
2533 |
+
"total_flos": 6.015093901152264e+17,
|
2534 |
+
"trial_name": null,
|
2535 |
+
"trial_params": null
|
2536 |
+
}
|
checkpoint-210000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3cce6e9d6e27de927e89729fbc8ce69542e970ce24b3c760a112fd7d97beed8f
|
3 |
+
size 3375
|