Upload 3 files
Browse files- config.json +25 -0
- pytorch_model.bin +3 -0
- trainer_state.json +1936 -0
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.18.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 55
|
| 25 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e227e2d91ccde4e47ec3e9472b7509958a492efe989611d6c28687827bfda623
|
| 3 |
+
size 344412267
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1936 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.515237104206927,
|
| 5 |
+
"global_step": 32000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 5e-05,
|
| 13 |
+
"loss": 3.5407,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 0.0001,
|
| 19 |
+
"loss": 3.2075,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 0.00015,
|
| 25 |
+
"loss": 3.0286,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.12,
|
| 30 |
+
"learning_rate": 0.0002,
|
| 31 |
+
"loss": 2.8212,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.15,
|
| 36 |
+
"learning_rate": 0.00025,
|
| 37 |
+
"loss": 2.3586,
|
| 38 |
+
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.18,
|
| 42 |
+
"learning_rate": 0.0003,
|
| 43 |
+
"loss": 1.676,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.21,
|
| 48 |
+
"learning_rate": 0.00035,
|
| 49 |
+
"loss": 1.3696,
|
| 50 |
+
"step": 700
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.24,
|
| 54 |
+
"learning_rate": 0.0004,
|
| 55 |
+
"loss": 1.2677,
|
| 56 |
+
"step": 800
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.27,
|
| 60 |
+
"learning_rate": 0.00045000000000000004,
|
| 61 |
+
"loss": 1.2271,
|
| 62 |
+
"step": 900
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.3,
|
| 66 |
+
"learning_rate": 0.0005,
|
| 67 |
+
"loss": 1.2006,
|
| 68 |
+
"step": 1000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.33,
|
| 72 |
+
"learning_rate": 0.000498467667790377,
|
| 73 |
+
"loss": 1.1846,
|
| 74 |
+
"step": 1100
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.36,
|
| 78 |
+
"learning_rate": 0.0004969353355807539,
|
| 79 |
+
"loss": 1.1663,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.39,
|
| 84 |
+
"learning_rate": 0.0004954030033711309,
|
| 85 |
+
"loss": 1.1429,
|
| 86 |
+
"step": 1300
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.42,
|
| 90 |
+
"learning_rate": 0.0004938706711615078,
|
| 91 |
+
"loss": 1.1384,
|
| 92 |
+
"step": 1400
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.45,
|
| 96 |
+
"learning_rate": 0.0004923383389518848,
|
| 97 |
+
"loss": 1.1353,
|
| 98 |
+
"step": 1500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.48,
|
| 102 |
+
"learning_rate": 0.0004908060067422617,
|
| 103 |
+
"loss": 1.1384,
|
| 104 |
+
"step": 1600
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.51,
|
| 108 |
+
"learning_rate": 0.0004892736745326388,
|
| 109 |
+
"loss": 1.1461,
|
| 110 |
+
"step": 1700
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.54,
|
| 114 |
+
"learning_rate": 0.00048774134232301567,
|
| 115 |
+
"loss": 1.1333,
|
| 116 |
+
"step": 1800
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.56,
|
| 120 |
+
"learning_rate": 0.0004862090101133926,
|
| 121 |
+
"loss": 1.1205,
|
| 122 |
+
"step": 1900
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.59,
|
| 126 |
+
"learning_rate": 0.00048467667790376954,
|
| 127 |
+
"loss": 1.1141,
|
| 128 |
+
"step": 2000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.62,
|
| 132 |
+
"learning_rate": 0.0004831443456941465,
|
| 133 |
+
"loss": 1.1078,
|
| 134 |
+
"step": 2100
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.65,
|
| 138 |
+
"learning_rate": 0.0004816120134845234,
|
| 139 |
+
"loss": 1.1006,
|
| 140 |
+
"step": 2200
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.68,
|
| 144 |
+
"learning_rate": 0.00048007968127490044,
|
| 145 |
+
"loss": 1.0978,
|
| 146 |
+
"step": 2300
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.71,
|
| 150 |
+
"learning_rate": 0.00047854734906527735,
|
| 151 |
+
"loss": 1.0894,
|
| 152 |
+
"step": 2400
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.74,
|
| 156 |
+
"learning_rate": 0.0004770150168556543,
|
| 157 |
+
"loss": 1.0861,
|
| 158 |
+
"step": 2500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.77,
|
| 162 |
+
"learning_rate": 0.0004754826846460313,
|
| 163 |
+
"loss": 1.083,
|
| 164 |
+
"step": 2600
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.8,
|
| 168 |
+
"learning_rate": 0.00047395035243640824,
|
| 169 |
+
"loss": 1.0758,
|
| 170 |
+
"step": 2700
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.83,
|
| 174 |
+
"learning_rate": 0.00047241802022678515,
|
| 175 |
+
"loss": 1.076,
|
| 176 |
+
"step": 2800
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.86,
|
| 180 |
+
"learning_rate": 0.00047088568801716217,
|
| 181 |
+
"loss": 1.0794,
|
| 182 |
+
"step": 2900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.89,
|
| 186 |
+
"learning_rate": 0.0004693533558075391,
|
| 187 |
+
"loss": 1.0706,
|
| 188 |
+
"step": 3000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.92,
|
| 192 |
+
"learning_rate": 0.00046782102359791604,
|
| 193 |
+
"loss": 1.0725,
|
| 194 |
+
"step": 3100
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.95,
|
| 198 |
+
"learning_rate": 0.000466288691388293,
|
| 199 |
+
"loss": 1.069,
|
| 200 |
+
"step": 3200
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.98,
|
| 204 |
+
"learning_rate": 0.0004647563591786699,
|
| 205 |
+
"loss": 1.0674,
|
| 206 |
+
"step": 3300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.01,
|
| 210 |
+
"learning_rate": 0.00046322402696904693,
|
| 211 |
+
"loss": 1.066,
|
| 212 |
+
"step": 3400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.04,
|
| 216 |
+
"learning_rate": 0.00046169169475942384,
|
| 217 |
+
"loss": 1.0569,
|
| 218 |
+
"step": 3500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.07,
|
| 222 |
+
"learning_rate": 0.0004601593625498008,
|
| 223 |
+
"loss": 1.0579,
|
| 224 |
+
"step": 3600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.1,
|
| 228 |
+
"learning_rate": 0.00045862703034017777,
|
| 229 |
+
"loss": 1.0615,
|
| 230 |
+
"step": 3700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.13,
|
| 234 |
+
"learning_rate": 0.00045709469813055473,
|
| 235 |
+
"loss": 1.055,
|
| 236 |
+
"step": 3800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.16,
|
| 240 |
+
"learning_rate": 0.00045556236592093164,
|
| 241 |
+
"loss": 1.0583,
|
| 242 |
+
"step": 3900
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.19,
|
| 246 |
+
"learning_rate": 0.0004540300337113086,
|
| 247 |
+
"loss": 1.0537,
|
| 248 |
+
"step": 4000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.22,
|
| 252 |
+
"learning_rate": 0.00045249770150168557,
|
| 253 |
+
"loss": 1.0531,
|
| 254 |
+
"step": 4100
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.25,
|
| 258 |
+
"learning_rate": 0.00045096536929206254,
|
| 259 |
+
"loss": 1.0507,
|
| 260 |
+
"step": 4200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.28,
|
| 264 |
+
"learning_rate": 0.0004494330370824395,
|
| 265 |
+
"loss": 1.0449,
|
| 266 |
+
"step": 4300
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.31,
|
| 270 |
+
"learning_rate": 0.0004479007048728164,
|
| 271 |
+
"loss": 1.0463,
|
| 272 |
+
"step": 4400
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.34,
|
| 276 |
+
"learning_rate": 0.00044636837266319343,
|
| 277 |
+
"loss": 1.0495,
|
| 278 |
+
"step": 4500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.37,
|
| 282 |
+
"learning_rate": 0.00044483604045357034,
|
| 283 |
+
"loss": 1.0489,
|
| 284 |
+
"step": 4600
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.4,
|
| 288 |
+
"learning_rate": 0.0004433037082439473,
|
| 289 |
+
"loss": 1.043,
|
| 290 |
+
"step": 4700
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.43,
|
| 294 |
+
"learning_rate": 0.00044177137603432427,
|
| 295 |
+
"loss": 1.0404,
|
| 296 |
+
"step": 4800
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.46,
|
| 300 |
+
"learning_rate": 0.0004402390438247012,
|
| 301 |
+
"loss": 1.0448,
|
| 302 |
+
"step": 4900
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.49,
|
| 306 |
+
"learning_rate": 0.00043870671161507814,
|
| 307 |
+
"loss": 1.0378,
|
| 308 |
+
"step": 5000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.52,
|
| 312 |
+
"learning_rate": 0.0004371743794054551,
|
| 313 |
+
"loss": 1.0359,
|
| 314 |
+
"step": 5100
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.55,
|
| 318 |
+
"learning_rate": 0.00043564204719583207,
|
| 319 |
+
"loss": 1.0419,
|
| 320 |
+
"step": 5200
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.58,
|
| 324 |
+
"learning_rate": 0.000434109714986209,
|
| 325 |
+
"loss": 1.0332,
|
| 326 |
+
"step": 5300
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.61,
|
| 330 |
+
"learning_rate": 0.000432577382776586,
|
| 331 |
+
"loss": 1.0382,
|
| 332 |
+
"step": 5400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.64,
|
| 336 |
+
"learning_rate": 0.0004310450505669629,
|
| 337 |
+
"loss": 1.0312,
|
| 338 |
+
"step": 5500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.67,
|
| 342 |
+
"learning_rate": 0.0004295127183573399,
|
| 343 |
+
"loss": 1.0377,
|
| 344 |
+
"step": 5600
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 1.69,
|
| 348 |
+
"learning_rate": 0.00042798038614771683,
|
| 349 |
+
"loss": 1.0296,
|
| 350 |
+
"step": 5700
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 1.72,
|
| 354 |
+
"learning_rate": 0.00042644805393809374,
|
| 355 |
+
"loss": 1.0316,
|
| 356 |
+
"step": 5800
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.75,
|
| 360 |
+
"learning_rate": 0.00042491572172847076,
|
| 361 |
+
"loss": 1.0322,
|
| 362 |
+
"step": 5900
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.78,
|
| 366 |
+
"learning_rate": 0.0004233833895188477,
|
| 367 |
+
"loss": 1.0325,
|
| 368 |
+
"step": 6000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.81,
|
| 372 |
+
"learning_rate": 0.00042185105730922464,
|
| 373 |
+
"loss": 1.0307,
|
| 374 |
+
"step": 6100
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.84,
|
| 378 |
+
"learning_rate": 0.0004203187250996016,
|
| 379 |
+
"loss": 1.0297,
|
| 380 |
+
"step": 6200
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.87,
|
| 384 |
+
"learning_rate": 0.00041878639288997857,
|
| 385 |
+
"loss": 1.031,
|
| 386 |
+
"step": 6300
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 1.9,
|
| 390 |
+
"learning_rate": 0.0004172540606803555,
|
| 391 |
+
"loss": 1.0304,
|
| 392 |
+
"step": 6400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 1.93,
|
| 396 |
+
"learning_rate": 0.0004157217284707325,
|
| 397 |
+
"loss": 1.0278,
|
| 398 |
+
"step": 6500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.96,
|
| 402 |
+
"learning_rate": 0.0004141893962611094,
|
| 403 |
+
"loss": 1.0211,
|
| 404 |
+
"step": 6600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.99,
|
| 408 |
+
"learning_rate": 0.0004126570640514864,
|
| 409 |
+
"loss": 1.0248,
|
| 410 |
+
"step": 6700
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.02,
|
| 414 |
+
"learning_rate": 0.00041112473184186333,
|
| 415 |
+
"loss": 1.0319,
|
| 416 |
+
"step": 6800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.05,
|
| 420 |
+
"learning_rate": 0.00040959239963224024,
|
| 421 |
+
"loss": 1.0301,
|
| 422 |
+
"step": 6900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.08,
|
| 426 |
+
"learning_rate": 0.00040806006742261726,
|
| 427 |
+
"loss": 1.0295,
|
| 428 |
+
"step": 7000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.11,
|
| 432 |
+
"learning_rate": 0.00040652773521299417,
|
| 433 |
+
"loss": 1.0247,
|
| 434 |
+
"step": 7100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.14,
|
| 438 |
+
"learning_rate": 0.00040499540300337113,
|
| 439 |
+
"loss": 1.0205,
|
| 440 |
+
"step": 7200
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.17,
|
| 444 |
+
"learning_rate": 0.0004034630707937481,
|
| 445 |
+
"loss": 1.0221,
|
| 446 |
+
"step": 7300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.2,
|
| 450 |
+
"learning_rate": 0.00040193073858412506,
|
| 451 |
+
"loss": 1.0251,
|
| 452 |
+
"step": 7400
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.23,
|
| 456 |
+
"learning_rate": 0.00040039840637450197,
|
| 457 |
+
"loss": 1.0164,
|
| 458 |
+
"step": 7500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.26,
|
| 462 |
+
"learning_rate": 0.000398866074164879,
|
| 463 |
+
"loss": 1.019,
|
| 464 |
+
"step": 7600
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.29,
|
| 468 |
+
"learning_rate": 0.0003973337419552559,
|
| 469 |
+
"loss": 1.0167,
|
| 470 |
+
"step": 7700
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.32,
|
| 474 |
+
"learning_rate": 0.0003958014097456328,
|
| 475 |
+
"loss": 1.0202,
|
| 476 |
+
"step": 7800
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.35,
|
| 480 |
+
"learning_rate": 0.00039426907753600983,
|
| 481 |
+
"loss": 1.0183,
|
| 482 |
+
"step": 7900
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.38,
|
| 486 |
+
"learning_rate": 0.00039273674532638674,
|
| 487 |
+
"loss": 1.0234,
|
| 488 |
+
"step": 8000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.41,
|
| 492 |
+
"learning_rate": 0.00039120441311676376,
|
| 493 |
+
"loss": 1.0103,
|
| 494 |
+
"step": 8100
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.44,
|
| 498 |
+
"learning_rate": 0.00038967208090714067,
|
| 499 |
+
"loss": 1.0196,
|
| 500 |
+
"step": 8200
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.47,
|
| 504 |
+
"learning_rate": 0.00038813974869751763,
|
| 505 |
+
"loss": 1.0147,
|
| 506 |
+
"step": 8300
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.5,
|
| 510 |
+
"learning_rate": 0.0003866074164878946,
|
| 511 |
+
"loss": 1.0138,
|
| 512 |
+
"step": 8400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.53,
|
| 516 |
+
"learning_rate": 0.00038507508427827156,
|
| 517 |
+
"loss": 1.0151,
|
| 518 |
+
"step": 8500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.56,
|
| 522 |
+
"learning_rate": 0.00038354275206864847,
|
| 523 |
+
"loss": 1.0118,
|
| 524 |
+
"step": 8600
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.59,
|
| 528 |
+
"learning_rate": 0.0003820104198590255,
|
| 529 |
+
"loss": 1.014,
|
| 530 |
+
"step": 8700
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.62,
|
| 534 |
+
"learning_rate": 0.0003804780876494024,
|
| 535 |
+
"loss": 1.0096,
|
| 536 |
+
"step": 8800
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.65,
|
| 540 |
+
"learning_rate": 0.0003789457554397793,
|
| 541 |
+
"loss": 1.0092,
|
| 542 |
+
"step": 8900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.68,
|
| 546 |
+
"learning_rate": 0.0003774134232301563,
|
| 547 |
+
"loss": 1.0096,
|
| 548 |
+
"step": 9000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.71,
|
| 552 |
+
"learning_rate": 0.00037588109102053323,
|
| 553 |
+
"loss": 1.0148,
|
| 554 |
+
"step": 9100
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.74,
|
| 558 |
+
"learning_rate": 0.00037434875881091025,
|
| 559 |
+
"loss": 1.0102,
|
| 560 |
+
"step": 9200
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.77,
|
| 564 |
+
"learning_rate": 0.00037281642660128716,
|
| 565 |
+
"loss": 1.0095,
|
| 566 |
+
"step": 9300
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.8,
|
| 570 |
+
"learning_rate": 0.0003712840943916641,
|
| 571 |
+
"loss": 1.0099,
|
| 572 |
+
"step": 9400
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.82,
|
| 576 |
+
"learning_rate": 0.0003697517621820411,
|
| 577 |
+
"loss": 1.0083,
|
| 578 |
+
"step": 9500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.85,
|
| 582 |
+
"learning_rate": 0.00036821942997241805,
|
| 583 |
+
"loss": 1.0093,
|
| 584 |
+
"step": 9600
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.88,
|
| 588 |
+
"learning_rate": 0.00036668709776279496,
|
| 589 |
+
"loss": 1.0023,
|
| 590 |
+
"step": 9700
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.91,
|
| 594 |
+
"learning_rate": 0.00036515476555317193,
|
| 595 |
+
"loss": 1.0058,
|
| 596 |
+
"step": 9800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.94,
|
| 600 |
+
"learning_rate": 0.0003636224333435489,
|
| 601 |
+
"loss": 1.0088,
|
| 602 |
+
"step": 9900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.97,
|
| 606 |
+
"learning_rate": 0.0003620901011339258,
|
| 607 |
+
"loss": 1.0046,
|
| 608 |
+
"step": 10000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.0,
|
| 612 |
+
"learning_rate": 0.0003605577689243028,
|
| 613 |
+
"loss": 1.0142,
|
| 614 |
+
"step": 10100
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.03,
|
| 618 |
+
"learning_rate": 0.00035902543671467973,
|
| 619 |
+
"loss": 1.0031,
|
| 620 |
+
"step": 10200
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.06,
|
| 624 |
+
"learning_rate": 0.0003574931045050567,
|
| 625 |
+
"loss": 1.006,
|
| 626 |
+
"step": 10300
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.09,
|
| 630 |
+
"learning_rate": 0.00035596077229543366,
|
| 631 |
+
"loss": 1.0019,
|
| 632 |
+
"step": 10400
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.12,
|
| 636 |
+
"learning_rate": 0.0003544284400858106,
|
| 637 |
+
"loss": 1.0023,
|
| 638 |
+
"step": 10500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.15,
|
| 642 |
+
"learning_rate": 0.0003528961078761876,
|
| 643 |
+
"loss": 0.9993,
|
| 644 |
+
"step": 10600
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.18,
|
| 648 |
+
"learning_rate": 0.0003513637756665645,
|
| 649 |
+
"loss": 0.9987,
|
| 650 |
+
"step": 10700
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.21,
|
| 654 |
+
"learning_rate": 0.00034983144345694146,
|
| 655 |
+
"loss": 0.9987,
|
| 656 |
+
"step": 10800
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.24,
|
| 660 |
+
"learning_rate": 0.0003482991112473184,
|
| 661 |
+
"loss": 1.005,
|
| 662 |
+
"step": 10900
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.27,
|
| 666 |
+
"learning_rate": 0.0003467667790376954,
|
| 667 |
+
"loss": 0.9966,
|
| 668 |
+
"step": 11000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.3,
|
| 672 |
+
"learning_rate": 0.0003452344468280723,
|
| 673 |
+
"loss": 0.9986,
|
| 674 |
+
"step": 11100
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.33,
|
| 678 |
+
"learning_rate": 0.0003437021146184493,
|
| 679 |
+
"loss": 0.9973,
|
| 680 |
+
"step": 11200
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.36,
|
| 684 |
+
"learning_rate": 0.00034216978240882623,
|
| 685 |
+
"loss": 1.0011,
|
| 686 |
+
"step": 11300
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 3.39,
|
| 690 |
+
"learning_rate": 0.0003406374501992032,
|
| 691 |
+
"loss": 0.9944,
|
| 692 |
+
"step": 11400
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 3.42,
|
| 696 |
+
"learning_rate": 0.00033910511798958016,
|
| 697 |
+
"loss": 0.996,
|
| 698 |
+
"step": 11500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 3.45,
|
| 702 |
+
"learning_rate": 0.00033757278577995707,
|
| 703 |
+
"loss": 0.9976,
|
| 704 |
+
"step": 11600
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 3.48,
|
| 708 |
+
"learning_rate": 0.0003360404535703341,
|
| 709 |
+
"loss": 0.9931,
|
| 710 |
+
"step": 11700
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 3.51,
|
| 714 |
+
"learning_rate": 0.000334508121360711,
|
| 715 |
+
"loss": 0.9921,
|
| 716 |
+
"step": 11800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.54,
|
| 720 |
+
"learning_rate": 0.00033297578915108796,
|
| 721 |
+
"loss": 0.9911,
|
| 722 |
+
"step": 11900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.57,
|
| 726 |
+
"learning_rate": 0.0003314434569414649,
|
| 727 |
+
"loss": 0.9916,
|
| 728 |
+
"step": 12000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 3.6,
|
| 732 |
+
"learning_rate": 0.0003299111247318419,
|
| 733 |
+
"loss": 0.9921,
|
| 734 |
+
"step": 12100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 3.63,
|
| 738 |
+
"learning_rate": 0.0003283787925222188,
|
| 739 |
+
"loss": 0.991,
|
| 740 |
+
"step": 12200
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 3.66,
|
| 744 |
+
"learning_rate": 0.0003268464603125958,
|
| 745 |
+
"loss": 0.9971,
|
| 746 |
+
"step": 12300
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 3.69,
|
| 750 |
+
"learning_rate": 0.0003253141281029727,
|
| 751 |
+
"loss": 0.995,
|
| 752 |
+
"step": 12400
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 3.72,
|
| 756 |
+
"learning_rate": 0.0003237817958933497,
|
| 757 |
+
"loss": 0.9891,
|
| 758 |
+
"step": 12500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 3.75,
|
| 762 |
+
"learning_rate": 0.00032224946368372665,
|
| 763 |
+
"loss": 0.9907,
|
| 764 |
+
"step": 12600
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 3.78,
|
| 768 |
+
"learning_rate": 0.00032071713147410356,
|
| 769 |
+
"loss": 0.9912,
|
| 770 |
+
"step": 12700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 3.81,
|
| 774 |
+
"learning_rate": 0.0003191847992644805,
|
| 775 |
+
"loss": 0.9873,
|
| 776 |
+
"step": 12800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 3.84,
|
| 780 |
+
"learning_rate": 0.0003176524670548575,
|
| 781 |
+
"loss": 0.9868,
|
| 782 |
+
"step": 12900
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 3.87,
|
| 786 |
+
"learning_rate": 0.00031612013484523445,
|
| 787 |
+
"loss": 0.9845,
|
| 788 |
+
"step": 13000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 3.9,
|
| 792 |
+
"learning_rate": 0.0003145878026356114,
|
| 793 |
+
"loss": 0.9836,
|
| 794 |
+
"step": 13100
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.92,
|
| 798 |
+
"learning_rate": 0.0003130554704259884,
|
| 799 |
+
"loss": 0.986,
|
| 800 |
+
"step": 13200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.95,
|
| 804 |
+
"learning_rate": 0.0003115231382163653,
|
| 805 |
+
"loss": 0.9902,
|
| 806 |
+
"step": 13300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.98,
|
| 810 |
+
"learning_rate": 0.0003099908060067423,
|
| 811 |
+
"loss": 0.983,
|
| 812 |
+
"step": 13400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.01,
|
| 816 |
+
"learning_rate": 0.0003084584737971192,
|
| 817 |
+
"loss": 0.9872,
|
| 818 |
+
"step": 13500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.04,
|
| 822 |
+
"learning_rate": 0.00030692614158749613,
|
| 823 |
+
"loss": 0.9844,
|
| 824 |
+
"step": 13600
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.07,
|
| 828 |
+
"learning_rate": 0.00030539380937787315,
|
| 829 |
+
"loss": 0.9867,
|
| 830 |
+
"step": 13700
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.1,
|
| 834 |
+
"learning_rate": 0.00030386147716825006,
|
| 835 |
+
"loss": 0.9821,
|
| 836 |
+
"step": 13800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.13,
|
| 840 |
+
"learning_rate": 0.000302329144958627,
|
| 841 |
+
"loss": 0.9809,
|
| 842 |
+
"step": 13900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.16,
|
| 846 |
+
"learning_rate": 0.000300796812749004,
|
| 847 |
+
"loss": 0.984,
|
| 848 |
+
"step": 14000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.19,
|
| 852 |
+
"learning_rate": 0.00029926448053938095,
|
| 853 |
+
"loss": 0.9767,
|
| 854 |
+
"step": 14100
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.22,
|
| 858 |
+
"learning_rate": 0.0002977321483297579,
|
| 859 |
+
"loss": 0.9819,
|
| 860 |
+
"step": 14200
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.25,
|
| 864 |
+
"learning_rate": 0.0002961998161201349,
|
| 865 |
+
"loss": 0.9811,
|
| 866 |
+
"step": 14300
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.28,
|
| 870 |
+
"learning_rate": 0.0002946674839105118,
|
| 871 |
+
"loss": 0.9791,
|
| 872 |
+
"step": 14400
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.31,
|
| 876 |
+
"learning_rate": 0.00029313515170088875,
|
| 877 |
+
"loss": 0.9783,
|
| 878 |
+
"step": 14500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.34,
|
| 882 |
+
"learning_rate": 0.0002916028194912657,
|
| 883 |
+
"loss": 0.9878,
|
| 884 |
+
"step": 14600
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.37,
|
| 888 |
+
"learning_rate": 0.00029007048728164263,
|
| 889 |
+
"loss": 0.975,
|
| 890 |
+
"step": 14700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 4.4,
|
| 894 |
+
"learning_rate": 0.00028853815507201965,
|
| 895 |
+
"loss": 0.9775,
|
| 896 |
+
"step": 14800
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 4.43,
|
| 900 |
+
"learning_rate": 0.00028700582286239656,
|
| 901 |
+
"loss": 0.9775,
|
| 902 |
+
"step": 14900
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 4.46,
|
| 906 |
+
"learning_rate": 0.0002854734906527735,
|
| 907 |
+
"loss": 0.9786,
|
| 908 |
+
"step": 15000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 4.49,
|
| 912 |
+
"learning_rate": 0.0002839411584431505,
|
| 913 |
+
"loss": 0.9753,
|
| 914 |
+
"step": 15100
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 4.52,
|
| 918 |
+
"learning_rate": 0.00028240882623352745,
|
| 919 |
+
"loss": 0.9841,
|
| 920 |
+
"step": 15200
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 4.55,
|
| 924 |
+
"learning_rate": 0.00028087649402390436,
|
| 925 |
+
"loss": 0.9716,
|
| 926 |
+
"step": 15300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 4.58,
|
| 930 |
+
"learning_rate": 0.0002793441618142814,
|
| 931 |
+
"loss": 0.9774,
|
| 932 |
+
"step": 15400
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 4.61,
|
| 936 |
+
"learning_rate": 0.0002778118296046583,
|
| 937 |
+
"loss": 0.9723,
|
| 938 |
+
"step": 15500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 4.64,
|
| 942 |
+
"learning_rate": 0.00027627949739503525,
|
| 943 |
+
"loss": 0.9702,
|
| 944 |
+
"step": 15600
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 4.67,
|
| 948 |
+
"learning_rate": 0.0002747471651854122,
|
| 949 |
+
"loss": 0.9766,
|
| 950 |
+
"step": 15700
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 4.7,
|
| 954 |
+
"learning_rate": 0.0002732148329757891,
|
| 955 |
+
"loss": 0.9843,
|
| 956 |
+
"step": 15800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 4.73,
|
| 960 |
+
"learning_rate": 0.00027168250076616614,
|
| 961 |
+
"loss": 0.9701,
|
| 962 |
+
"step": 15900
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 4.76,
|
| 966 |
+
"learning_rate": 0.00027015016855654305,
|
| 967 |
+
"loss": 0.9715,
|
| 968 |
+
"step": 16000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 4.79,
|
| 972 |
+
"learning_rate": 0.00026861783634692,
|
| 973 |
+
"loss": 0.9695,
|
| 974 |
+
"step": 16100
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 4.82,
|
| 978 |
+
"learning_rate": 0.000267085504137297,
|
| 979 |
+
"loss": 0.9699,
|
| 980 |
+
"step": 16200
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 4.85,
|
| 984 |
+
"learning_rate": 0.00026555317192767394,
|
| 985 |
+
"loss": 0.9665,
|
| 986 |
+
"step": 16300
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 4.88,
|
| 990 |
+
"learning_rate": 0.00026402083971805085,
|
| 991 |
+
"loss": 0.9681,
|
| 992 |
+
"step": 16400
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 4.91,
|
| 996 |
+
"learning_rate": 0.0002624885075084278,
|
| 997 |
+
"loss": 0.9697,
|
| 998 |
+
"step": 16500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 4.94,
|
| 1002 |
+
"learning_rate": 0.0002609561752988048,
|
| 1003 |
+
"loss": 0.9662,
|
| 1004 |
+
"step": 16600
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 4.97,
|
| 1008 |
+
"learning_rate": 0.00025942384308918175,
|
| 1009 |
+
"loss": 0.965,
|
| 1010 |
+
"step": 16700
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.0,
|
| 1014 |
+
"learning_rate": 0.0002578915108795587,
|
| 1015 |
+
"loss": 0.9655,
|
| 1016 |
+
"step": 16800
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.03,
|
| 1020 |
+
"learning_rate": 0.0002563591786699356,
|
| 1021 |
+
"loss": 0.9689,
|
| 1022 |
+
"step": 16900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 5.06,
|
| 1026 |
+
"learning_rate": 0.00025482684646031264,
|
| 1027 |
+
"loss": 0.9641,
|
| 1028 |
+
"step": 17000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 5.08,
|
| 1032 |
+
"learning_rate": 0.00025329451425068955,
|
| 1033 |
+
"loss": 0.9612,
|
| 1034 |
+
"step": 17100
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.11,
|
| 1038 |
+
"learning_rate": 0.0002517621820410665,
|
| 1039 |
+
"loss": 0.9667,
|
| 1040 |
+
"step": 17200
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 5.14,
|
| 1044 |
+
"learning_rate": 0.0002502298498314435,
|
| 1045 |
+
"loss": 0.9623,
|
| 1046 |
+
"step": 17300
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 5.17,
|
| 1050 |
+
"learning_rate": 0.0002486975176218204,
|
| 1051 |
+
"loss": 0.9611,
|
| 1052 |
+
"step": 17400
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.2,
|
| 1056 |
+
"learning_rate": 0.00024716518541219735,
|
| 1057 |
+
"loss": 0.956,
|
| 1058 |
+
"step": 17500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 5.23,
|
| 1062 |
+
"learning_rate": 0.0002456328532025743,
|
| 1063 |
+
"loss": 0.9623,
|
| 1064 |
+
"step": 17600
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 5.26,
|
| 1068 |
+
"learning_rate": 0.00024410052099295128,
|
| 1069 |
+
"loss": 0.9577,
|
| 1070 |
+
"step": 17700
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 5.29,
|
| 1074 |
+
"learning_rate": 0.00024256818878332824,
|
| 1075 |
+
"loss": 0.9584,
|
| 1076 |
+
"step": 17800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.32,
|
| 1080 |
+
"learning_rate": 0.00024103585657370518,
|
| 1081 |
+
"loss": 0.9595,
|
| 1082 |
+
"step": 17900
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 5.35,
|
| 1086 |
+
"learning_rate": 0.00023950352436408212,
|
| 1087 |
+
"loss": 0.954,
|
| 1088 |
+
"step": 18000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 5.38,
|
| 1092 |
+
"learning_rate": 0.00023797119215445908,
|
| 1093 |
+
"loss": 0.958,
|
| 1094 |
+
"step": 18100
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 5.41,
|
| 1098 |
+
"learning_rate": 0.00023643885994483605,
|
| 1099 |
+
"loss": 0.9575,
|
| 1100 |
+
"step": 18200
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 5.44,
|
| 1104 |
+
"learning_rate": 0.000234906527735213,
|
| 1105 |
+
"loss": 0.9499,
|
| 1106 |
+
"step": 18300
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 5.47,
|
| 1110 |
+
"learning_rate": 0.00023337419552558995,
|
| 1111 |
+
"loss": 0.9583,
|
| 1112 |
+
"step": 18400
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 5.5,
|
| 1116 |
+
"learning_rate": 0.0002318418633159669,
|
| 1117 |
+
"loss": 0.9547,
|
| 1118 |
+
"step": 18500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 5.53,
|
| 1122 |
+
"learning_rate": 0.00023030953110634387,
|
| 1123 |
+
"loss": 0.9531,
|
| 1124 |
+
"step": 18600
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 5.56,
|
| 1128 |
+
"learning_rate": 0.0002287771988967208,
|
| 1129 |
+
"loss": 0.9566,
|
| 1130 |
+
"step": 18700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 5.59,
|
| 1134 |
+
"learning_rate": 0.00022724486668709778,
|
| 1135 |
+
"loss": 0.9519,
|
| 1136 |
+
"step": 18800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 5.62,
|
| 1140 |
+
"learning_rate": 0.0002257125344774747,
|
| 1141 |
+
"loss": 0.9473,
|
| 1142 |
+
"step": 18900
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 5.65,
|
| 1146 |
+
"learning_rate": 0.00022418020226785168,
|
| 1147 |
+
"loss": 0.9496,
|
| 1148 |
+
"step": 19000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 5.68,
|
| 1152 |
+
"learning_rate": 0.00022264787005822861,
|
| 1153 |
+
"loss": 0.9469,
|
| 1154 |
+
"step": 19100
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 5.71,
|
| 1158 |
+
"learning_rate": 0.00022111553784860558,
|
| 1159 |
+
"loss": 0.9509,
|
| 1160 |
+
"step": 19200
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 5.74,
|
| 1164 |
+
"learning_rate": 0.00021958320563898254,
|
| 1165 |
+
"loss": 0.9466,
|
| 1166 |
+
"step": 19300
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 5.77,
|
| 1170 |
+
"learning_rate": 0.0002180508734293595,
|
| 1171 |
+
"loss": 0.9499,
|
| 1172 |
+
"step": 19400
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 5.8,
|
| 1176 |
+
"learning_rate": 0.00021651854121973644,
|
| 1177 |
+
"loss": 0.9498,
|
| 1178 |
+
"step": 19500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 5.83,
|
| 1182 |
+
"learning_rate": 0.0002149862090101134,
|
| 1183 |
+
"loss": 0.9483,
|
| 1184 |
+
"step": 19600
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 5.86,
|
| 1188 |
+
"learning_rate": 0.00021345387680049037,
|
| 1189 |
+
"loss": 0.9522,
|
| 1190 |
+
"step": 19700
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 5.89,
|
| 1194 |
+
"learning_rate": 0.00021192154459086728,
|
| 1195 |
+
"loss": 0.9441,
|
| 1196 |
+
"step": 19800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 5.92,
|
| 1200 |
+
"learning_rate": 0.00021038921238124425,
|
| 1201 |
+
"loss": 0.9492,
|
| 1202 |
+
"step": 19900
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 5.95,
|
| 1206 |
+
"learning_rate": 0.0002088568801716212,
|
| 1207 |
+
"loss": 0.9421,
|
| 1208 |
+
"step": 20000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 5.98,
|
| 1212 |
+
"learning_rate": 0.00020732454796199817,
|
| 1213 |
+
"loss": 0.9432,
|
| 1214 |
+
"step": 20100
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 6.01,
|
| 1218 |
+
"learning_rate": 0.0002057922157523751,
|
| 1219 |
+
"loss": 0.9542,
|
| 1220 |
+
"step": 20200
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.04,
|
| 1224 |
+
"learning_rate": 0.00020425988354275207,
|
| 1225 |
+
"loss": 0.9426,
|
| 1226 |
+
"step": 20300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 6.07,
|
| 1230 |
+
"learning_rate": 0.00020272755133312904,
|
| 1231 |
+
"loss": 0.9484,
|
| 1232 |
+
"step": 20400
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.1,
|
| 1236 |
+
"learning_rate": 0.00020119521912350598,
|
| 1237 |
+
"loss": 0.9473,
|
| 1238 |
+
"step": 20500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 6.13,
|
| 1242 |
+
"learning_rate": 0.00019966288691388294,
|
| 1243 |
+
"loss": 0.9413,
|
| 1244 |
+
"step": 20600
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 6.16,
|
| 1248 |
+
"learning_rate": 0.0001981305547042599,
|
| 1249 |
+
"loss": 0.9438,
|
| 1250 |
+
"step": 20700
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 6.18,
|
| 1254 |
+
"learning_rate": 0.00019659822249463684,
|
| 1255 |
+
"loss": 0.9421,
|
| 1256 |
+
"step": 20800
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 6.21,
|
| 1260 |
+
"learning_rate": 0.00019506589028501378,
|
| 1261 |
+
"loss": 0.9406,
|
| 1262 |
+
"step": 20900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.24,
|
| 1266 |
+
"learning_rate": 0.00019353355807539074,
|
| 1267 |
+
"loss": 0.9384,
|
| 1268 |
+
"step": 21000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 6.27,
|
| 1272 |
+
"learning_rate": 0.0001920012258657677,
|
| 1273 |
+
"loss": 0.9397,
|
| 1274 |
+
"step": 21100
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 6.3,
|
| 1278 |
+
"learning_rate": 0.00019046889365614464,
|
| 1279 |
+
"loss": 0.9367,
|
| 1280 |
+
"step": 21200
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 6.33,
|
| 1284 |
+
"learning_rate": 0.0001889365614465216,
|
| 1285 |
+
"loss": 0.9402,
|
| 1286 |
+
"step": 21300
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 6.36,
|
| 1290 |
+
"learning_rate": 0.00018740422923689857,
|
| 1291 |
+
"loss": 0.9319,
|
| 1292 |
+
"step": 21400
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 6.39,
|
| 1296 |
+
"learning_rate": 0.00018587189702727554,
|
| 1297 |
+
"loss": 0.9385,
|
| 1298 |
+
"step": 21500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 6.42,
|
| 1302 |
+
"learning_rate": 0.00018433956481765247,
|
| 1303 |
+
"loss": 0.939,
|
| 1304 |
+
"step": 21600
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 6.45,
|
| 1308 |
+
"learning_rate": 0.0001828072326080294,
|
| 1309 |
+
"loss": 0.9399,
|
| 1310 |
+
"step": 21700
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 6.48,
|
| 1314 |
+
"learning_rate": 0.00018127490039840637,
|
| 1315 |
+
"loss": 0.9407,
|
| 1316 |
+
"step": 21800
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 6.51,
|
| 1320 |
+
"learning_rate": 0.00017974256818878334,
|
| 1321 |
+
"loss": 0.94,
|
| 1322 |
+
"step": 21900
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 6.54,
|
| 1326 |
+
"learning_rate": 0.00017821023597916027,
|
| 1327 |
+
"loss": 0.9407,
|
| 1328 |
+
"step": 22000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 6.57,
|
| 1332 |
+
"learning_rate": 0.00017667790376953724,
|
| 1333 |
+
"loss": 0.9353,
|
| 1334 |
+
"step": 22100
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 6.6,
|
| 1338 |
+
"learning_rate": 0.0001751455715599142,
|
| 1339 |
+
"loss": 0.9405,
|
| 1340 |
+
"step": 22200
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 6.63,
|
| 1344 |
+
"learning_rate": 0.00017361323935029114,
|
| 1345 |
+
"loss": 0.9305,
|
| 1346 |
+
"step": 22300
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 6.66,
|
| 1350 |
+
"learning_rate": 0.0001720809071406681,
|
| 1351 |
+
"loss": 0.938,
|
| 1352 |
+
"step": 22400
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 6.69,
|
| 1356 |
+
"learning_rate": 0.00017054857493104507,
|
| 1357 |
+
"loss": 0.9311,
|
| 1358 |
+
"step": 22500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 6.72,
|
| 1362 |
+
"learning_rate": 0.00016901624272142203,
|
| 1363 |
+
"loss": 0.9343,
|
| 1364 |
+
"step": 22600
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 6.75,
|
| 1368 |
+
"learning_rate": 0.00016748391051179894,
|
| 1369 |
+
"loss": 0.9312,
|
| 1370 |
+
"step": 22700
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 6.78,
|
| 1374 |
+
"learning_rate": 0.0001659515783021759,
|
| 1375 |
+
"loss": 0.9353,
|
| 1376 |
+
"step": 22800
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 6.81,
|
| 1380 |
+
"learning_rate": 0.00016441924609255287,
|
| 1381 |
+
"loss": 0.9341,
|
| 1382 |
+
"step": 22900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 6.84,
|
| 1386 |
+
"learning_rate": 0.0001628869138829298,
|
| 1387 |
+
"loss": 0.9338,
|
| 1388 |
+
"step": 23000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 6.87,
|
| 1392 |
+
"learning_rate": 0.00016135458167330677,
|
| 1393 |
+
"loss": 0.9318,
|
| 1394 |
+
"step": 23100
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 6.9,
|
| 1398 |
+
"learning_rate": 0.00015982224946368373,
|
| 1399 |
+
"loss": 0.9309,
|
| 1400 |
+
"step": 23200
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 6.93,
|
| 1404 |
+
"learning_rate": 0.0001582899172540607,
|
| 1405 |
+
"loss": 0.9291,
|
| 1406 |
+
"step": 23300
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 6.96,
|
| 1410 |
+
"learning_rate": 0.00015675758504443764,
|
| 1411 |
+
"loss": 0.9307,
|
| 1412 |
+
"step": 23400
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 6.99,
|
| 1416 |
+
"learning_rate": 0.0001552252528348146,
|
| 1417 |
+
"loss": 0.9325,
|
| 1418 |
+
"step": 23500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 7.02,
|
| 1422 |
+
"learning_rate": 0.00015369292062519156,
|
| 1423 |
+
"loss": 0.9363,
|
| 1424 |
+
"step": 23600
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 7.05,
|
| 1428 |
+
"learning_rate": 0.00015216058841556847,
|
| 1429 |
+
"loss": 0.9325,
|
| 1430 |
+
"step": 23700
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.08,
|
| 1434 |
+
"learning_rate": 0.00015062825620594544,
|
| 1435 |
+
"loss": 0.9276,
|
| 1436 |
+
"step": 23800
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 7.11,
|
| 1440 |
+
"learning_rate": 0.0001490959239963224,
|
| 1441 |
+
"loss": 0.9328,
|
| 1442 |
+
"step": 23900
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 7.14,
|
| 1446 |
+
"learning_rate": 0.00014756359178669937,
|
| 1447 |
+
"loss": 0.9304,
|
| 1448 |
+
"step": 24000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 7.17,
|
| 1452 |
+
"learning_rate": 0.0001460312595770763,
|
| 1453 |
+
"loss": 0.9274,
|
| 1454 |
+
"step": 24100
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 7.2,
|
| 1458 |
+
"learning_rate": 0.00014449892736745327,
|
| 1459 |
+
"loss": 0.9261,
|
| 1460 |
+
"step": 24200
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 7.23,
|
| 1464 |
+
"learning_rate": 0.00014296659515783023,
|
| 1465 |
+
"loss": 0.9245,
|
| 1466 |
+
"step": 24300
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 7.26,
|
| 1470 |
+
"learning_rate": 0.0001414342629482072,
|
| 1471 |
+
"loss": 0.9233,
|
| 1472 |
+
"step": 24400
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 7.29,
|
| 1476 |
+
"learning_rate": 0.00013990193073858413,
|
| 1477 |
+
"loss": 0.9275,
|
| 1478 |
+
"step": 24500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 7.31,
|
| 1482 |
+
"learning_rate": 0.00013836959852896107,
|
| 1483 |
+
"loss": 0.9265,
|
| 1484 |
+
"step": 24600
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 7.34,
|
| 1488 |
+
"learning_rate": 0.00013683726631933803,
|
| 1489 |
+
"loss": 0.9276,
|
| 1490 |
+
"step": 24700
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 7.37,
|
| 1494 |
+
"learning_rate": 0.00013530493410971497,
|
| 1495 |
+
"loss": 0.9252,
|
| 1496 |
+
"step": 24800
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 7.4,
|
| 1500 |
+
"learning_rate": 0.00013377260190009193,
|
| 1501 |
+
"loss": 0.9224,
|
| 1502 |
+
"step": 24900
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 7.43,
|
| 1506 |
+
"learning_rate": 0.0001322402696904689,
|
| 1507 |
+
"loss": 0.9216,
|
| 1508 |
+
"step": 25000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 7.46,
|
| 1512 |
+
"learning_rate": 0.00013070793748084586,
|
| 1513 |
+
"loss": 0.9233,
|
| 1514 |
+
"step": 25100
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 7.49,
|
| 1518 |
+
"learning_rate": 0.0001291756052712228,
|
| 1519 |
+
"loss": 0.9275,
|
| 1520 |
+
"step": 25200
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 7.52,
|
| 1524 |
+
"learning_rate": 0.00012764327306159976,
|
| 1525 |
+
"loss": 0.9229,
|
| 1526 |
+
"step": 25300
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 7.55,
|
| 1530 |
+
"learning_rate": 0.00012611094085197673,
|
| 1531 |
+
"loss": 0.922,
|
| 1532 |
+
"step": 25400
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 7.58,
|
| 1536 |
+
"learning_rate": 0.00012457860864235367,
|
| 1537 |
+
"loss": 0.9255,
|
| 1538 |
+
"step": 25500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 7.61,
|
| 1542 |
+
"learning_rate": 0.0001230462764327306,
|
| 1543 |
+
"loss": 0.9196,
|
| 1544 |
+
"step": 25600
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 7.64,
|
| 1548 |
+
"learning_rate": 0.00012151394422310758,
|
| 1549 |
+
"loss": 0.9198,
|
| 1550 |
+
"step": 25700
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 7.67,
|
| 1554 |
+
"learning_rate": 0.00011998161201348452,
|
| 1555 |
+
"loss": 0.9226,
|
| 1556 |
+
"step": 25800
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 7.7,
|
| 1560 |
+
"learning_rate": 0.00011844927980386148,
|
| 1561 |
+
"loss": 0.9174,
|
| 1562 |
+
"step": 25900
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 7.73,
|
| 1566 |
+
"learning_rate": 0.00011691694759423843,
|
| 1567 |
+
"loss": 0.9191,
|
| 1568 |
+
"step": 26000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 7.76,
|
| 1572 |
+
"learning_rate": 0.0001153846153846154,
|
| 1573 |
+
"loss": 0.9207,
|
| 1574 |
+
"step": 26100
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 7.79,
|
| 1578 |
+
"learning_rate": 0.00011385228317499235,
|
| 1579 |
+
"loss": 0.9225,
|
| 1580 |
+
"step": 26200
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 7.82,
|
| 1584 |
+
"learning_rate": 0.00011231995096536928,
|
| 1585 |
+
"loss": 0.9198,
|
| 1586 |
+
"step": 26300
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 7.85,
|
| 1590 |
+
"learning_rate": 0.00011078761875574625,
|
| 1591 |
+
"loss": 0.9183,
|
| 1592 |
+
"step": 26400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 7.88,
|
| 1596 |
+
"learning_rate": 0.0001092552865461232,
|
| 1597 |
+
"loss": 0.919,
|
| 1598 |
+
"step": 26500
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 7.91,
|
| 1602 |
+
"learning_rate": 0.00010772295433650016,
|
| 1603 |
+
"loss": 0.9193,
|
| 1604 |
+
"step": 26600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 7.94,
|
| 1608 |
+
"learning_rate": 0.00010619062212687711,
|
| 1609 |
+
"loss": 0.9205,
|
| 1610 |
+
"step": 26700
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 7.97,
|
| 1614 |
+
"learning_rate": 0.00010465828991725406,
|
| 1615 |
+
"loss": 0.92,
|
| 1616 |
+
"step": 26800
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 8.0,
|
| 1620 |
+
"learning_rate": 0.00010312595770763101,
|
| 1621 |
+
"loss": 0.9192,
|
| 1622 |
+
"step": 26900
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 8.03,
|
| 1626 |
+
"learning_rate": 0.00010159362549800798,
|
| 1627 |
+
"loss": 0.9186,
|
| 1628 |
+
"step": 27000
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.06,
|
| 1632 |
+
"learning_rate": 0.00010006129328838493,
|
| 1633 |
+
"loss": 0.9176,
|
| 1634 |
+
"step": 27100
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 8.09,
|
| 1638 |
+
"learning_rate": 9.852896107876188e-05,
|
| 1639 |
+
"loss": 0.9136,
|
| 1640 |
+
"step": 27200
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 8.12,
|
| 1644 |
+
"learning_rate": 9.699662886913883e-05,
|
| 1645 |
+
"loss": 0.9116,
|
| 1646 |
+
"step": 27300
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 8.15,
|
| 1650 |
+
"learning_rate": 9.546429665951578e-05,
|
| 1651 |
+
"loss": 0.9174,
|
| 1652 |
+
"step": 27400
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 8.18,
|
| 1656 |
+
"learning_rate": 9.393196444989274e-05,
|
| 1657 |
+
"loss": 0.9156,
|
| 1658 |
+
"step": 27500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 8.21,
|
| 1662 |
+
"learning_rate": 9.23996322402697e-05,
|
| 1663 |
+
"loss": 0.912,
|
| 1664 |
+
"step": 27600
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 8.24,
|
| 1668 |
+
"learning_rate": 9.086730003064666e-05,
|
| 1669 |
+
"loss": 0.9142,
|
| 1670 |
+
"step": 27700
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 8.27,
|
| 1674 |
+
"learning_rate": 8.93349678210236e-05,
|
| 1675 |
+
"loss": 0.9099,
|
| 1676 |
+
"step": 27800
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 8.3,
|
| 1680 |
+
"learning_rate": 8.780263561140055e-05,
|
| 1681 |
+
"loss": 0.9129,
|
| 1682 |
+
"step": 27900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 8.33,
|
| 1686 |
+
"learning_rate": 8.627030340177751e-05,
|
| 1687 |
+
"loss": 0.9145,
|
| 1688 |
+
"step": 28000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 8.36,
|
| 1692 |
+
"learning_rate": 8.473797119215446e-05,
|
| 1693 |
+
"loss": 0.9117,
|
| 1694 |
+
"step": 28100
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 8.39,
|
| 1698 |
+
"learning_rate": 8.320563898253141e-05,
|
| 1699 |
+
"loss": 0.9112,
|
| 1700 |
+
"step": 28200
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 8.42,
|
| 1704 |
+
"learning_rate": 8.167330677290836e-05,
|
| 1705 |
+
"loss": 0.9128,
|
| 1706 |
+
"step": 28300
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 8.44,
|
| 1710 |
+
"learning_rate": 8.014097456328533e-05,
|
| 1711 |
+
"loss": 0.9122,
|
| 1712 |
+
"step": 28400
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 8.47,
|
| 1716 |
+
"learning_rate": 7.860864235366228e-05,
|
| 1717 |
+
"loss": 0.9113,
|
| 1718 |
+
"step": 28500
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 8.5,
|
| 1722 |
+
"learning_rate": 7.707631014403924e-05,
|
| 1723 |
+
"loss": 0.9115,
|
| 1724 |
+
"step": 28600
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 8.53,
|
| 1728 |
+
"learning_rate": 7.554397793441618e-05,
|
| 1729 |
+
"loss": 0.9098,
|
| 1730 |
+
"step": 28700
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 8.56,
|
| 1734 |
+
"learning_rate": 7.401164572479313e-05,
|
| 1735 |
+
"loss": 0.9101,
|
| 1736 |
+
"step": 28800
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 8.59,
|
| 1740 |
+
"learning_rate": 7.247931351517009e-05,
|
| 1741 |
+
"loss": 0.9063,
|
| 1742 |
+
"step": 28900
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 8.62,
|
| 1746 |
+
"learning_rate": 7.094698130554704e-05,
|
| 1747 |
+
"loss": 0.913,
|
| 1748 |
+
"step": 29000
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 8.65,
|
| 1752 |
+
"learning_rate": 6.9414649095924e-05,
|
| 1753 |
+
"loss": 0.9092,
|
| 1754 |
+
"step": 29100
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 8.68,
|
| 1758 |
+
"learning_rate": 6.788231688630094e-05,
|
| 1759 |
+
"loss": 0.9101,
|
| 1760 |
+
"step": 29200
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 8.71,
|
| 1764 |
+
"learning_rate": 6.634998467667791e-05,
|
| 1765 |
+
"loss": 0.9089,
|
| 1766 |
+
"step": 29300
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 8.74,
|
| 1770 |
+
"learning_rate": 6.481765246705486e-05,
|
| 1771 |
+
"loss": 0.9108,
|
| 1772 |
+
"step": 29400
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 8.77,
|
| 1776 |
+
"learning_rate": 6.328532025743182e-05,
|
| 1777 |
+
"loss": 0.9065,
|
| 1778 |
+
"step": 29500
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 8.8,
|
| 1782 |
+
"learning_rate": 6.175298804780877e-05,
|
| 1783 |
+
"loss": 0.9129,
|
| 1784 |
+
"step": 29600
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 8.83,
|
| 1788 |
+
"learning_rate": 6.022065583818572e-05,
|
| 1789 |
+
"loss": 0.9097,
|
| 1790 |
+
"step": 29700
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 8.86,
|
| 1794 |
+
"learning_rate": 5.8688323628562674e-05,
|
| 1795 |
+
"loss": 0.9115,
|
| 1796 |
+
"step": 29800
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 8.89,
|
| 1800 |
+
"learning_rate": 5.715599141893963e-05,
|
| 1801 |
+
"loss": 0.9088,
|
| 1802 |
+
"step": 29900
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 8.92,
|
| 1806 |
+
"learning_rate": 5.5623659209316575e-05,
|
| 1807 |
+
"loss": 0.9112,
|
| 1808 |
+
"step": 30000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 8.95,
|
| 1812 |
+
"learning_rate": 5.409132699969353e-05,
|
| 1813 |
+
"loss": 0.9086,
|
| 1814 |
+
"step": 30100
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 8.98,
|
| 1818 |
+
"learning_rate": 5.255899479007049e-05,
|
| 1819 |
+
"loss": 0.9106,
|
| 1820 |
+
"step": 30200
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 9.01,
|
| 1824 |
+
"learning_rate": 5.102666258044744e-05,
|
| 1825 |
+
"loss": 0.9104,
|
| 1826 |
+
"step": 30300
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 9.04,
|
| 1830 |
+
"learning_rate": 4.94943303708244e-05,
|
| 1831 |
+
"loss": 0.9037,
|
| 1832 |
+
"step": 30400
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.07,
|
| 1836 |
+
"learning_rate": 4.796199816120135e-05,
|
| 1837 |
+
"loss": 0.9082,
|
| 1838 |
+
"step": 30500
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 9.1,
|
| 1842 |
+
"learning_rate": 4.6429665951578305e-05,
|
| 1843 |
+
"loss": 0.9041,
|
| 1844 |
+
"step": 30600
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 9.13,
|
| 1848 |
+
"learning_rate": 4.489733374195526e-05,
|
| 1849 |
+
"loss": 0.9025,
|
| 1850 |
+
"step": 30700
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 9.16,
|
| 1854 |
+
"learning_rate": 4.3365001532332206e-05,
|
| 1855 |
+
"loss": 0.9006,
|
| 1856 |
+
"step": 30800
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 9.19,
|
| 1860 |
+
"learning_rate": 4.1832669322709164e-05,
|
| 1861 |
+
"loss": 0.9072,
|
| 1862 |
+
"step": 30900
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 9.22,
|
| 1866 |
+
"learning_rate": 4.0300337113086114e-05,
|
| 1867 |
+
"loss": 0.9038,
|
| 1868 |
+
"step": 31000
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 9.25,
|
| 1872 |
+
"learning_rate": 3.876800490346307e-05,
|
| 1873 |
+
"loss": 0.9072,
|
| 1874 |
+
"step": 31100
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 9.28,
|
| 1878 |
+
"learning_rate": 3.723567269384002e-05,
|
| 1879 |
+
"loss": 0.9017,
|
| 1880 |
+
"step": 31200
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 9.31,
|
| 1884 |
+
"learning_rate": 3.570334048421698e-05,
|
| 1885 |
+
"loss": 0.9032,
|
| 1886 |
+
"step": 31300
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 9.34,
|
| 1890 |
+
"learning_rate": 3.4171008274593937e-05,
|
| 1891 |
+
"loss": 0.9026,
|
| 1892 |
+
"step": 31400
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 9.37,
|
| 1896 |
+
"learning_rate": 3.263867606497089e-05,
|
| 1897 |
+
"loss": 0.9008,
|
| 1898 |
+
"step": 31500
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 9.4,
|
| 1902 |
+
"learning_rate": 3.110634385534784e-05,
|
| 1903 |
+
"loss": 0.9065,
|
| 1904 |
+
"step": 31600
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 9.43,
|
| 1908 |
+
"learning_rate": 2.9574011645724795e-05,
|
| 1909 |
+
"loss": 0.9026,
|
| 1910 |
+
"step": 31700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 9.46,
|
| 1914 |
+
"learning_rate": 2.804167943610175e-05,
|
| 1915 |
+
"loss": 0.9011,
|
| 1916 |
+
"step": 31800
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 9.49,
|
| 1920 |
+
"learning_rate": 2.65093472264787e-05,
|
| 1921 |
+
"loss": 0.9023,
|
| 1922 |
+
"step": 31900
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 9.52,
|
| 1926 |
+
"learning_rate": 2.4977015016855653e-05,
|
| 1927 |
+
"loss": 0.9007,
|
| 1928 |
+
"step": 32000
|
| 1929 |
+
}
|
| 1930 |
+
],
|
| 1931 |
+
"max_steps": 33630,
|
| 1932 |
+
"num_train_epochs": 10,
|
| 1933 |
+
"total_flos": 8.418135066885916e+17,
|
| 1934 |
+
"trial_name": null,
|
| 1935 |
+
"trial_params": null
|
| 1936 |
+
}
|