Upload folder using huggingface_hub
Browse files- config.json +25 -0
- generation_config.json +6 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1216 -0
- training_args.bin +3 -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 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 3072,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_length": 512,
|
| 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": "4.29.2",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 610
|
| 25 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"max_length": 512,
|
| 4 |
+
"pad_token_id": 0,
|
| 5 |
+
"transformers_version": "4.29.2"
|
| 6 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b94eae11ebac271508993f2f27a4ad5e1f5028638f4e573fe3c5054987b4781
|
| 3 |
+
size 692215499
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a448ff3dd52c8bfb983420e92e43e77892ace1424526d9c10b8580966404381a
|
| 3 |
+
size 346122191
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:136a4f26618b8e44d873ce83286097bab76c8ac85c9f04861291f10d72aa2ac7
|
| 3 |
+
size 14645
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b751cb1c63fcc5090801e168722b88043b19f294b16d123cbf19c81339681021
|
| 3 |
+
size 1465
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 15.120586678763136,
|
| 5 |
+
"global_step": 200000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.08,
|
| 12 |
+
"learning_rate": 4.000000000000001e-06,
|
| 13 |
+
"loss": 5.7461,
|
| 14 |
+
"step": 1000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.15,
|
| 18 |
+
"learning_rate": 8.000000000000001e-06,
|
| 19 |
+
"loss": 5.2435,
|
| 20 |
+
"step": 2000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.23,
|
| 24 |
+
"learning_rate": 1.2e-05,
|
| 25 |
+
"loss": 5.1697,
|
| 26 |
+
"step": 3000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.3,
|
| 30 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 31 |
+
"loss": 5.1422,
|
| 32 |
+
"step": 4000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.38,
|
| 36 |
+
"learning_rate": 2e-05,
|
| 37 |
+
"loss": 5.1265,
|
| 38 |
+
"step": 5000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.45,
|
| 42 |
+
"learning_rate": 2.4e-05,
|
| 43 |
+
"loss": 5.1186,
|
| 44 |
+
"step": 6000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.53,
|
| 48 |
+
"learning_rate": 2.8e-05,
|
| 49 |
+
"loss": 5.1159,
|
| 50 |
+
"step": 7000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.6,
|
| 54 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 55 |
+
"loss": 5.11,
|
| 56 |
+
"step": 8000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.68,
|
| 60 |
+
"learning_rate": 3.6e-05,
|
| 61 |
+
"loss": 5.1084,
|
| 62 |
+
"step": 9000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.76,
|
| 66 |
+
"learning_rate": 4e-05,
|
| 67 |
+
"loss": 5.1073,
|
| 68 |
+
"step": 10000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.83,
|
| 72 |
+
"learning_rate": 3.978947368421053e-05,
|
| 73 |
+
"loss": 5.1033,
|
| 74 |
+
"step": 11000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.91,
|
| 78 |
+
"learning_rate": 3.9578947368421056e-05,
|
| 79 |
+
"loss": 5.103,
|
| 80 |
+
"step": 12000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.98,
|
| 84 |
+
"learning_rate": 3.936842105263158e-05,
|
| 85 |
+
"loss": 5.0998,
|
| 86 |
+
"step": 13000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.06,
|
| 90 |
+
"learning_rate": 3.915789473684211e-05,
|
| 91 |
+
"loss": 5.099,
|
| 92 |
+
"step": 14000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.13,
|
| 96 |
+
"learning_rate": 3.8947368421052636e-05,
|
| 97 |
+
"loss": 5.0958,
|
| 98 |
+
"step": 15000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.21,
|
| 102 |
+
"learning_rate": 3.873684210526316e-05,
|
| 103 |
+
"loss": 5.0956,
|
| 104 |
+
"step": 16000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.29,
|
| 108 |
+
"learning_rate": 3.852631578947369e-05,
|
| 109 |
+
"loss": 5.0925,
|
| 110 |
+
"step": 17000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.36,
|
| 114 |
+
"learning_rate": 3.8315789473684216e-05,
|
| 115 |
+
"loss": 5.0947,
|
| 116 |
+
"step": 18000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.44,
|
| 120 |
+
"learning_rate": 3.8105263157894735e-05,
|
| 121 |
+
"loss": 5.0925,
|
| 122 |
+
"step": 19000
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.51,
|
| 126 |
+
"learning_rate": 3.789473684210526e-05,
|
| 127 |
+
"loss": 5.093,
|
| 128 |
+
"step": 20000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.59,
|
| 132 |
+
"learning_rate": 3.7684210526315795e-05,
|
| 133 |
+
"loss": 5.0922,
|
| 134 |
+
"step": 21000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.66,
|
| 138 |
+
"learning_rate": 3.747368421052632e-05,
|
| 139 |
+
"loss": 5.0894,
|
| 140 |
+
"step": 22000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.74,
|
| 144 |
+
"learning_rate": 3.726315789473685e-05,
|
| 145 |
+
"loss": 5.0699,
|
| 146 |
+
"step": 23000
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.81,
|
| 150 |
+
"learning_rate": 3.705263157894737e-05,
|
| 151 |
+
"loss": 5.037,
|
| 152 |
+
"step": 24000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 1.89,
|
| 156 |
+
"learning_rate": 3.6842105263157895e-05,
|
| 157 |
+
"loss": 5.0146,
|
| 158 |
+
"step": 25000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.97,
|
| 162 |
+
"learning_rate": 3.663157894736842e-05,
|
| 163 |
+
"loss": 4.9928,
|
| 164 |
+
"step": 26000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.04,
|
| 168 |
+
"learning_rate": 3.6421052631578955e-05,
|
| 169 |
+
"loss": 4.9687,
|
| 170 |
+
"step": 27000
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.12,
|
| 174 |
+
"learning_rate": 3.621052631578948e-05,
|
| 175 |
+
"loss": 4.9464,
|
| 176 |
+
"step": 28000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.19,
|
| 180 |
+
"learning_rate": 3.6e-05,
|
| 181 |
+
"loss": 4.922,
|
| 182 |
+
"step": 29000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 2.27,
|
| 186 |
+
"learning_rate": 3.578947368421053e-05,
|
| 187 |
+
"loss": 4.8983,
|
| 188 |
+
"step": 30000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.34,
|
| 192 |
+
"learning_rate": 3.5578947368421054e-05,
|
| 193 |
+
"loss": 4.8721,
|
| 194 |
+
"step": 31000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.42,
|
| 198 |
+
"learning_rate": 3.536842105263158e-05,
|
| 199 |
+
"loss": 4.8441,
|
| 200 |
+
"step": 32000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.49,
|
| 204 |
+
"learning_rate": 3.515789473684211e-05,
|
| 205 |
+
"loss": 4.8125,
|
| 206 |
+
"step": 33000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.57,
|
| 210 |
+
"learning_rate": 3.4947368421052634e-05,
|
| 211 |
+
"loss": 4.7796,
|
| 212 |
+
"step": 34000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 2.65,
|
| 216 |
+
"learning_rate": 3.473684210526316e-05,
|
| 217 |
+
"loss": 4.7264,
|
| 218 |
+
"step": 35000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.72,
|
| 222 |
+
"learning_rate": 3.452631578947369e-05,
|
| 223 |
+
"loss": 4.6306,
|
| 224 |
+
"step": 36000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.8,
|
| 228 |
+
"learning_rate": 3.431578947368421e-05,
|
| 229 |
+
"loss": 4.5571,
|
| 230 |
+
"step": 37000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.87,
|
| 234 |
+
"learning_rate": 3.410526315789474e-05,
|
| 235 |
+
"loss": 4.4757,
|
| 236 |
+
"step": 38000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.95,
|
| 240 |
+
"learning_rate": 3.3894736842105266e-05,
|
| 241 |
+
"loss": 4.3942,
|
| 242 |
+
"step": 39000
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 3.02,
|
| 246 |
+
"learning_rate": 3.368421052631579e-05,
|
| 247 |
+
"loss": 4.3376,
|
| 248 |
+
"step": 40000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.1,
|
| 252 |
+
"learning_rate": 3.347368421052632e-05,
|
| 253 |
+
"loss": 4.2928,
|
| 254 |
+
"step": 41000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 3.18,
|
| 258 |
+
"learning_rate": 3.3263157894736846e-05,
|
| 259 |
+
"loss": 4.2614,
|
| 260 |
+
"step": 42000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 3.25,
|
| 264 |
+
"learning_rate": 3.305263157894737e-05,
|
| 265 |
+
"loss": 4.2338,
|
| 266 |
+
"step": 43000
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 3.33,
|
| 270 |
+
"learning_rate": 3.28421052631579e-05,
|
| 271 |
+
"loss": 4.2183,
|
| 272 |
+
"step": 44000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 3.4,
|
| 276 |
+
"learning_rate": 3.2631578947368426e-05,
|
| 277 |
+
"loss": 4.194,
|
| 278 |
+
"step": 45000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 3.48,
|
| 282 |
+
"learning_rate": 3.2421052631578945e-05,
|
| 283 |
+
"loss": 4.1761,
|
| 284 |
+
"step": 46000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 3.55,
|
| 288 |
+
"learning_rate": 3.221052631578947e-05,
|
| 289 |
+
"loss": 4.1644,
|
| 290 |
+
"step": 47000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 3.63,
|
| 294 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 295 |
+
"loss": 4.1498,
|
| 296 |
+
"step": 48000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 3.7,
|
| 300 |
+
"learning_rate": 3.178947368421053e-05,
|
| 301 |
+
"loss": 4.1397,
|
| 302 |
+
"step": 49000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 3.78,
|
| 306 |
+
"learning_rate": 3.157894736842106e-05,
|
| 307 |
+
"loss": 4.1275,
|
| 308 |
+
"step": 50000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 3.86,
|
| 312 |
+
"learning_rate": 3.136842105263158e-05,
|
| 313 |
+
"loss": 4.1146,
|
| 314 |
+
"step": 51000
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 3.93,
|
| 318 |
+
"learning_rate": 3.1157894736842105e-05,
|
| 319 |
+
"loss": 4.1074,
|
| 320 |
+
"step": 52000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.01,
|
| 324 |
+
"learning_rate": 3.094736842105263e-05,
|
| 325 |
+
"loss": 4.1001,
|
| 326 |
+
"step": 53000
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 4.08,
|
| 330 |
+
"learning_rate": 3.073684210526316e-05,
|
| 331 |
+
"loss": 4.0887,
|
| 332 |
+
"step": 54000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.16,
|
| 336 |
+
"learning_rate": 3.052631578947369e-05,
|
| 337 |
+
"loss": 4.0836,
|
| 338 |
+
"step": 55000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 4.23,
|
| 342 |
+
"learning_rate": 3.0315789473684214e-05,
|
| 343 |
+
"loss": 4.0808,
|
| 344 |
+
"step": 56000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 4.31,
|
| 348 |
+
"learning_rate": 3.010526315789474e-05,
|
| 349 |
+
"loss": 4.0691,
|
| 350 |
+
"step": 57000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 4.38,
|
| 354 |
+
"learning_rate": 2.9894736842105264e-05,
|
| 355 |
+
"loss": 4.0624,
|
| 356 |
+
"step": 58000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 4.46,
|
| 360 |
+
"learning_rate": 2.968421052631579e-05,
|
| 361 |
+
"loss": 4.0501,
|
| 362 |
+
"step": 59000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 4.54,
|
| 366 |
+
"learning_rate": 2.9473684210526317e-05,
|
| 367 |
+
"loss": 4.0523,
|
| 368 |
+
"step": 60000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 4.61,
|
| 372 |
+
"learning_rate": 2.9263157894736847e-05,
|
| 373 |
+
"loss": 4.0434,
|
| 374 |
+
"step": 61000
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 4.69,
|
| 378 |
+
"learning_rate": 2.9052631578947374e-05,
|
| 379 |
+
"loss": 4.0381,
|
| 380 |
+
"step": 62000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 4.76,
|
| 384 |
+
"learning_rate": 2.8842105263157897e-05,
|
| 385 |
+
"loss": 4.0363,
|
| 386 |
+
"step": 63000
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 4.84,
|
| 390 |
+
"learning_rate": 2.8631578947368423e-05,
|
| 391 |
+
"loss": 4.0303,
|
| 392 |
+
"step": 64000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 4.91,
|
| 396 |
+
"learning_rate": 2.842105263157895e-05,
|
| 397 |
+
"loss": 4.0211,
|
| 398 |
+
"step": 65000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 4.99,
|
| 402 |
+
"learning_rate": 2.8210526315789476e-05,
|
| 403 |
+
"loss": 4.0168,
|
| 404 |
+
"step": 66000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 5.07,
|
| 408 |
+
"learning_rate": 2.8e-05,
|
| 409 |
+
"loss": 4.0122,
|
| 410 |
+
"step": 67000
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 5.14,
|
| 414 |
+
"learning_rate": 2.778947368421053e-05,
|
| 415 |
+
"loss": 4.0091,
|
| 416 |
+
"step": 68000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 5.22,
|
| 420 |
+
"learning_rate": 2.7578947368421056e-05,
|
| 421 |
+
"loss": 4.0019,
|
| 422 |
+
"step": 69000
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 5.29,
|
| 426 |
+
"learning_rate": 2.7368421052631583e-05,
|
| 427 |
+
"loss": 3.9973,
|
| 428 |
+
"step": 70000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 5.37,
|
| 432 |
+
"learning_rate": 2.715789473684211e-05,
|
| 433 |
+
"loss": 3.9994,
|
| 434 |
+
"step": 71000
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 5.44,
|
| 438 |
+
"learning_rate": 2.6947368421052632e-05,
|
| 439 |
+
"loss": 3.9907,
|
| 440 |
+
"step": 72000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 5.52,
|
| 444 |
+
"learning_rate": 2.673684210526316e-05,
|
| 445 |
+
"loss": 3.9873,
|
| 446 |
+
"step": 73000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 5.59,
|
| 450 |
+
"learning_rate": 2.6526315789473685e-05,
|
| 451 |
+
"loss": 3.984,
|
| 452 |
+
"step": 74000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 5.67,
|
| 456 |
+
"learning_rate": 2.6315789473684215e-05,
|
| 457 |
+
"loss": 3.9795,
|
| 458 |
+
"step": 75000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 5.75,
|
| 462 |
+
"learning_rate": 2.610526315789474e-05,
|
| 463 |
+
"loss": 3.973,
|
| 464 |
+
"step": 76000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 5.82,
|
| 468 |
+
"learning_rate": 2.5894736842105265e-05,
|
| 469 |
+
"loss": 3.9739,
|
| 470 |
+
"step": 77000
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 5.9,
|
| 474 |
+
"learning_rate": 2.568421052631579e-05,
|
| 475 |
+
"loss": 3.9637,
|
| 476 |
+
"step": 78000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 5.97,
|
| 480 |
+
"learning_rate": 2.5473684210526318e-05,
|
| 481 |
+
"loss": 3.9668,
|
| 482 |
+
"step": 79000
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 6.05,
|
| 486 |
+
"learning_rate": 2.526315789473684e-05,
|
| 487 |
+
"loss": 3.9624,
|
| 488 |
+
"step": 80000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 6.12,
|
| 492 |
+
"learning_rate": 2.5052631578947368e-05,
|
| 493 |
+
"loss": 3.9513,
|
| 494 |
+
"step": 81000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 6.2,
|
| 498 |
+
"learning_rate": 2.4842105263157898e-05,
|
| 499 |
+
"loss": 3.9586,
|
| 500 |
+
"step": 82000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 6.28,
|
| 504 |
+
"learning_rate": 2.4631578947368424e-05,
|
| 505 |
+
"loss": 3.9509,
|
| 506 |
+
"step": 83000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 6.35,
|
| 510 |
+
"learning_rate": 2.442105263157895e-05,
|
| 511 |
+
"loss": 3.9526,
|
| 512 |
+
"step": 84000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 6.43,
|
| 516 |
+
"learning_rate": 2.4210526315789474e-05,
|
| 517 |
+
"loss": 3.9427,
|
| 518 |
+
"step": 85000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 6.5,
|
| 522 |
+
"learning_rate": 2.4e-05,
|
| 523 |
+
"loss": 3.9448,
|
| 524 |
+
"step": 86000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 6.58,
|
| 528 |
+
"learning_rate": 2.3789473684210527e-05,
|
| 529 |
+
"loss": 3.9331,
|
| 530 |
+
"step": 87000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 6.65,
|
| 534 |
+
"learning_rate": 2.3578947368421054e-05,
|
| 535 |
+
"loss": 3.9368,
|
| 536 |
+
"step": 88000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 6.73,
|
| 540 |
+
"learning_rate": 2.3368421052631584e-05,
|
| 541 |
+
"loss": 3.9316,
|
| 542 |
+
"step": 89000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 6.8,
|
| 546 |
+
"learning_rate": 2.3157894736842107e-05,
|
| 547 |
+
"loss": 3.932,
|
| 548 |
+
"step": 90000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 6.88,
|
| 552 |
+
"learning_rate": 2.2947368421052633e-05,
|
| 553 |
+
"loss": 3.9249,
|
| 554 |
+
"step": 91000
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 6.96,
|
| 558 |
+
"learning_rate": 2.273684210526316e-05,
|
| 559 |
+
"loss": 3.9272,
|
| 560 |
+
"step": 92000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 7.03,
|
| 564 |
+
"learning_rate": 2.2526315789473686e-05,
|
| 565 |
+
"loss": 3.9157,
|
| 566 |
+
"step": 93000
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 7.11,
|
| 570 |
+
"learning_rate": 2.231578947368421e-05,
|
| 571 |
+
"loss": 3.9139,
|
| 572 |
+
"step": 94000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 7.18,
|
| 576 |
+
"learning_rate": 2.210526315789474e-05,
|
| 577 |
+
"loss": 3.9091,
|
| 578 |
+
"step": 95000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 7.26,
|
| 582 |
+
"learning_rate": 2.1894736842105266e-05,
|
| 583 |
+
"loss": 3.9164,
|
| 584 |
+
"step": 96000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 7.33,
|
| 588 |
+
"learning_rate": 2.1684210526315793e-05,
|
| 589 |
+
"loss": 3.9119,
|
| 590 |
+
"step": 97000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 7.41,
|
| 594 |
+
"learning_rate": 2.147368421052632e-05,
|
| 595 |
+
"loss": 3.9058,
|
| 596 |
+
"step": 98000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 7.48,
|
| 600 |
+
"learning_rate": 2.1263157894736842e-05,
|
| 601 |
+
"loss": 3.9131,
|
| 602 |
+
"step": 99000
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 7.56,
|
| 606 |
+
"learning_rate": 2.105263157894737e-05,
|
| 607 |
+
"loss": 3.9097,
|
| 608 |
+
"step": 100000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 7.64,
|
| 612 |
+
"learning_rate": 2.0842105263157895e-05,
|
| 613 |
+
"loss": 3.9044,
|
| 614 |
+
"step": 101000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 7.71,
|
| 618 |
+
"learning_rate": 2.0631578947368425e-05,
|
| 619 |
+
"loss": 3.9004,
|
| 620 |
+
"step": 102000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 7.79,
|
| 624 |
+
"learning_rate": 2.0421052631578952e-05,
|
| 625 |
+
"loss": 3.8937,
|
| 626 |
+
"step": 103000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 7.86,
|
| 630 |
+
"learning_rate": 2.0210526315789475e-05,
|
| 631 |
+
"loss": 3.9017,
|
| 632 |
+
"step": 104000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 7.94,
|
| 636 |
+
"learning_rate": 2e-05,
|
| 637 |
+
"loss": 3.899,
|
| 638 |
+
"step": 105000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 8.01,
|
| 642 |
+
"learning_rate": 1.9789473684210528e-05,
|
| 643 |
+
"loss": 3.8979,
|
| 644 |
+
"step": 106000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 8.09,
|
| 648 |
+
"learning_rate": 1.9578947368421055e-05,
|
| 649 |
+
"loss": 3.8889,
|
| 650 |
+
"step": 107000
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 8.17,
|
| 654 |
+
"learning_rate": 1.936842105263158e-05,
|
| 655 |
+
"loss": 3.8857,
|
| 656 |
+
"step": 108000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 8.24,
|
| 660 |
+
"learning_rate": 1.9157894736842108e-05,
|
| 661 |
+
"loss": 3.8892,
|
| 662 |
+
"step": 109000
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 8.32,
|
| 666 |
+
"learning_rate": 1.894736842105263e-05,
|
| 667 |
+
"loss": 3.887,
|
| 668 |
+
"step": 110000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.39,
|
| 672 |
+
"learning_rate": 1.873684210526316e-05,
|
| 673 |
+
"loss": 3.8848,
|
| 674 |
+
"step": 111000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 8.47,
|
| 678 |
+
"learning_rate": 1.8526315789473684e-05,
|
| 679 |
+
"loss": 3.8805,
|
| 680 |
+
"step": 112000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 8.54,
|
| 684 |
+
"learning_rate": 1.831578947368421e-05,
|
| 685 |
+
"loss": 3.8819,
|
| 686 |
+
"step": 113000
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 8.62,
|
| 690 |
+
"learning_rate": 1.810526315789474e-05,
|
| 691 |
+
"loss": 3.8753,
|
| 692 |
+
"step": 114000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 8.69,
|
| 696 |
+
"learning_rate": 1.7894736842105264e-05,
|
| 697 |
+
"loss": 3.881,
|
| 698 |
+
"step": 115000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 8.77,
|
| 702 |
+
"learning_rate": 1.768421052631579e-05,
|
| 703 |
+
"loss": 3.8784,
|
| 704 |
+
"step": 116000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 8.85,
|
| 708 |
+
"learning_rate": 1.7473684210526317e-05,
|
| 709 |
+
"loss": 3.8694,
|
| 710 |
+
"step": 117000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 8.92,
|
| 714 |
+
"learning_rate": 1.7263157894736843e-05,
|
| 715 |
+
"loss": 3.8727,
|
| 716 |
+
"step": 118000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 9.0,
|
| 720 |
+
"learning_rate": 1.705263157894737e-05,
|
| 721 |
+
"loss": 3.8701,
|
| 722 |
+
"step": 119000
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 9.07,
|
| 726 |
+
"learning_rate": 1.6842105263157896e-05,
|
| 727 |
+
"loss": 3.8676,
|
| 728 |
+
"step": 120000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 9.15,
|
| 732 |
+
"learning_rate": 1.6631578947368423e-05,
|
| 733 |
+
"loss": 3.8632,
|
| 734 |
+
"step": 121000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 9.22,
|
| 738 |
+
"learning_rate": 1.642105263157895e-05,
|
| 739 |
+
"loss": 3.8645,
|
| 740 |
+
"step": 122000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 9.3,
|
| 744 |
+
"learning_rate": 1.6210526315789473e-05,
|
| 745 |
+
"loss": 3.8682,
|
| 746 |
+
"step": 123000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 9.37,
|
| 750 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 751 |
+
"loss": 3.8575,
|
| 752 |
+
"step": 124000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 9.45,
|
| 756 |
+
"learning_rate": 1.578947368421053e-05,
|
| 757 |
+
"loss": 3.8633,
|
| 758 |
+
"step": 125000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 9.53,
|
| 762 |
+
"learning_rate": 1.5578947368421052e-05,
|
| 763 |
+
"loss": 3.8594,
|
| 764 |
+
"step": 126000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 9.6,
|
| 768 |
+
"learning_rate": 1.536842105263158e-05,
|
| 769 |
+
"loss": 3.8552,
|
| 770 |
+
"step": 127000
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 9.68,
|
| 774 |
+
"learning_rate": 1.5157894736842107e-05,
|
| 775 |
+
"loss": 3.8563,
|
| 776 |
+
"step": 128000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 9.75,
|
| 780 |
+
"learning_rate": 1.4947368421052632e-05,
|
| 781 |
+
"loss": 3.8596,
|
| 782 |
+
"step": 129000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 9.83,
|
| 786 |
+
"learning_rate": 1.4736842105263159e-05,
|
| 787 |
+
"loss": 3.8565,
|
| 788 |
+
"step": 130000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 9.9,
|
| 792 |
+
"learning_rate": 1.4526315789473687e-05,
|
| 793 |
+
"loss": 3.8512,
|
| 794 |
+
"step": 131000
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 9.98,
|
| 798 |
+
"learning_rate": 1.4315789473684212e-05,
|
| 799 |
+
"loss": 3.8531,
|
| 800 |
+
"step": 132000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 10.06,
|
| 804 |
+
"learning_rate": 1.4105263157894738e-05,
|
| 805 |
+
"loss": 3.8538,
|
| 806 |
+
"step": 133000
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 10.13,
|
| 810 |
+
"learning_rate": 1.3894736842105265e-05,
|
| 811 |
+
"loss": 3.8477,
|
| 812 |
+
"step": 134000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 10.21,
|
| 816 |
+
"learning_rate": 1.3684210526315791e-05,
|
| 817 |
+
"loss": 3.8497,
|
| 818 |
+
"step": 135000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 10.28,
|
| 822 |
+
"learning_rate": 1.3473684210526316e-05,
|
| 823 |
+
"loss": 3.8455,
|
| 824 |
+
"step": 136000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 10.36,
|
| 828 |
+
"learning_rate": 1.3263157894736843e-05,
|
| 829 |
+
"loss": 3.8442,
|
| 830 |
+
"step": 137000
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 10.43,
|
| 834 |
+
"learning_rate": 1.305263157894737e-05,
|
| 835 |
+
"loss": 3.8466,
|
| 836 |
+
"step": 138000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 10.51,
|
| 840 |
+
"learning_rate": 1.2842105263157896e-05,
|
| 841 |
+
"loss": 3.8465,
|
| 842 |
+
"step": 139000
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 10.58,
|
| 846 |
+
"learning_rate": 1.263157894736842e-05,
|
| 847 |
+
"loss": 3.8443,
|
| 848 |
+
"step": 140000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 10.66,
|
| 852 |
+
"learning_rate": 1.2421052631578949e-05,
|
| 853 |
+
"loss": 3.8407,
|
| 854 |
+
"step": 141000
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 10.74,
|
| 858 |
+
"learning_rate": 1.2210526315789475e-05,
|
| 859 |
+
"loss": 3.8414,
|
| 860 |
+
"step": 142000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 10.81,
|
| 864 |
+
"learning_rate": 1.2e-05,
|
| 865 |
+
"loss": 3.835,
|
| 866 |
+
"step": 143000
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 10.89,
|
| 870 |
+
"learning_rate": 1.1789473684210527e-05,
|
| 871 |
+
"loss": 3.8321,
|
| 872 |
+
"step": 144000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 10.96,
|
| 876 |
+
"learning_rate": 1.1578947368421053e-05,
|
| 877 |
+
"loss": 3.8366,
|
| 878 |
+
"step": 145000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 11.04,
|
| 882 |
+
"learning_rate": 1.136842105263158e-05,
|
| 883 |
+
"loss": 3.8353,
|
| 884 |
+
"step": 146000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 11.11,
|
| 888 |
+
"learning_rate": 1.1157894736842105e-05,
|
| 889 |
+
"loss": 3.8347,
|
| 890 |
+
"step": 147000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 11.19,
|
| 894 |
+
"learning_rate": 1.0947368421052633e-05,
|
| 895 |
+
"loss": 3.833,
|
| 896 |
+
"step": 148000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 11.26,
|
| 900 |
+
"learning_rate": 1.073684210526316e-05,
|
| 901 |
+
"loss": 3.8337,
|
| 902 |
+
"step": 149000
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 11.34,
|
| 906 |
+
"learning_rate": 1.0526315789473684e-05,
|
| 907 |
+
"loss": 3.8342,
|
| 908 |
+
"step": 150000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 11.42,
|
| 912 |
+
"learning_rate": 1.0315789473684213e-05,
|
| 913 |
+
"loss": 3.8302,
|
| 914 |
+
"step": 151000
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 11.49,
|
| 918 |
+
"learning_rate": 1.0105263157894738e-05,
|
| 919 |
+
"loss": 3.8326,
|
| 920 |
+
"step": 152000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 11.57,
|
| 924 |
+
"learning_rate": 9.894736842105264e-06,
|
| 925 |
+
"loss": 3.827,
|
| 926 |
+
"step": 153000
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 11.64,
|
| 930 |
+
"learning_rate": 9.68421052631579e-06,
|
| 931 |
+
"loss": 3.8309,
|
| 932 |
+
"step": 154000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 11.72,
|
| 936 |
+
"learning_rate": 9.473684210526315e-06,
|
| 937 |
+
"loss": 3.8293,
|
| 938 |
+
"step": 155000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 11.79,
|
| 942 |
+
"learning_rate": 9.263157894736842e-06,
|
| 943 |
+
"loss": 3.8277,
|
| 944 |
+
"step": 156000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 11.87,
|
| 948 |
+
"learning_rate": 9.05263157894737e-06,
|
| 949 |
+
"loss": 3.8273,
|
| 950 |
+
"step": 157000
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 11.95,
|
| 954 |
+
"learning_rate": 8.842105263157895e-06,
|
| 955 |
+
"loss": 3.8199,
|
| 956 |
+
"step": 158000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 12.02,
|
| 960 |
+
"learning_rate": 8.631578947368422e-06,
|
| 961 |
+
"loss": 3.8253,
|
| 962 |
+
"step": 159000
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 12.1,
|
| 966 |
+
"learning_rate": 8.421052631578948e-06,
|
| 967 |
+
"loss": 3.8267,
|
| 968 |
+
"step": 160000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 12.17,
|
| 972 |
+
"learning_rate": 8.210526315789475e-06,
|
| 973 |
+
"loss": 3.8287,
|
| 974 |
+
"step": 161000
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 12.25,
|
| 978 |
+
"learning_rate": 8.000000000000001e-06,
|
| 979 |
+
"loss": 3.8196,
|
| 980 |
+
"step": 162000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 12.32,
|
| 984 |
+
"learning_rate": 7.789473684210526e-06,
|
| 985 |
+
"loss": 3.8208,
|
| 986 |
+
"step": 163000
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 12.4,
|
| 990 |
+
"learning_rate": 7.578947368421054e-06,
|
| 991 |
+
"loss": 3.8287,
|
| 992 |
+
"step": 164000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 12.47,
|
| 996 |
+
"learning_rate": 7.368421052631579e-06,
|
| 997 |
+
"loss": 3.8223,
|
| 998 |
+
"step": 165000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 12.55,
|
| 1002 |
+
"learning_rate": 7.157894736842106e-06,
|
| 1003 |
+
"loss": 3.8103,
|
| 1004 |
+
"step": 166000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 12.63,
|
| 1008 |
+
"learning_rate": 6.947368421052632e-06,
|
| 1009 |
+
"loss": 3.8183,
|
| 1010 |
+
"step": 167000
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 12.7,
|
| 1014 |
+
"learning_rate": 6.736842105263158e-06,
|
| 1015 |
+
"loss": 3.8174,
|
| 1016 |
+
"step": 168000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 12.78,
|
| 1020 |
+
"learning_rate": 6.526315789473685e-06,
|
| 1021 |
+
"loss": 3.8154,
|
| 1022 |
+
"step": 169000
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 12.85,
|
| 1026 |
+
"learning_rate": 6.31578947368421e-06,
|
| 1027 |
+
"loss": 3.8204,
|
| 1028 |
+
"step": 170000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 12.93,
|
| 1032 |
+
"learning_rate": 6.105263157894738e-06,
|
| 1033 |
+
"loss": 3.8104,
|
| 1034 |
+
"step": 171000
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 13.0,
|
| 1038 |
+
"learning_rate": 5.8947368421052634e-06,
|
| 1039 |
+
"loss": 3.8173,
|
| 1040 |
+
"step": 172000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 13.08,
|
| 1044 |
+
"learning_rate": 5.68421052631579e-06,
|
| 1045 |
+
"loss": 3.8139,
|
| 1046 |
+
"step": 173000
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 13.15,
|
| 1050 |
+
"learning_rate": 5.4736842105263165e-06,
|
| 1051 |
+
"loss": 3.809,
|
| 1052 |
+
"step": 174000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 13.23,
|
| 1056 |
+
"learning_rate": 5.263157894736842e-06,
|
| 1057 |
+
"loss": 3.8151,
|
| 1058 |
+
"step": 175000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 13.31,
|
| 1062 |
+
"learning_rate": 5.052631578947369e-06,
|
| 1063 |
+
"loss": 3.8129,
|
| 1064 |
+
"step": 176000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 13.38,
|
| 1068 |
+
"learning_rate": 4.842105263157895e-06,
|
| 1069 |
+
"loss": 3.815,
|
| 1070 |
+
"step": 177000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 13.46,
|
| 1074 |
+
"learning_rate": 4.631578947368421e-06,
|
| 1075 |
+
"loss": 3.8175,
|
| 1076 |
+
"step": 178000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 13.53,
|
| 1080 |
+
"learning_rate": 4.4210526315789476e-06,
|
| 1081 |
+
"loss": 3.8125,
|
| 1082 |
+
"step": 179000
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 13.61,
|
| 1086 |
+
"learning_rate": 4.210526315789474e-06,
|
| 1087 |
+
"loss": 3.8101,
|
| 1088 |
+
"step": 180000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 13.68,
|
| 1092 |
+
"learning_rate": 4.000000000000001e-06,
|
| 1093 |
+
"loss": 3.8079,
|
| 1094 |
+
"step": 181000
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 13.76,
|
| 1098 |
+
"learning_rate": 3.789473684210527e-06,
|
| 1099 |
+
"loss": 3.8124,
|
| 1100 |
+
"step": 182000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 13.84,
|
| 1104 |
+
"learning_rate": 3.578947368421053e-06,
|
| 1105 |
+
"loss": 3.8134,
|
| 1106 |
+
"step": 183000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 13.91,
|
| 1110 |
+
"learning_rate": 3.368421052631579e-06,
|
| 1111 |
+
"loss": 3.8048,
|
| 1112 |
+
"step": 184000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 13.99,
|
| 1116 |
+
"learning_rate": 3.157894736842105e-06,
|
| 1117 |
+
"loss": 3.8167,
|
| 1118 |
+
"step": 185000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 14.06,
|
| 1122 |
+
"learning_rate": 2.9473684210526317e-06,
|
| 1123 |
+
"loss": 3.812,
|
| 1124 |
+
"step": 186000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 14.14,
|
| 1128 |
+
"learning_rate": 2.7368421052631583e-06,
|
| 1129 |
+
"loss": 3.8089,
|
| 1130 |
+
"step": 187000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 14.21,
|
| 1134 |
+
"learning_rate": 2.5263157894736844e-06,
|
| 1135 |
+
"loss": 3.8119,
|
| 1136 |
+
"step": 188000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 14.29,
|
| 1140 |
+
"learning_rate": 2.3157894736842105e-06,
|
| 1141 |
+
"loss": 3.8058,
|
| 1142 |
+
"step": 189000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 14.36,
|
| 1146 |
+
"learning_rate": 2.105263157894737e-06,
|
| 1147 |
+
"loss": 3.8089,
|
| 1148 |
+
"step": 190000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 14.44,
|
| 1152 |
+
"learning_rate": 1.8947368421052634e-06,
|
| 1153 |
+
"loss": 3.8058,
|
| 1154 |
+
"step": 191000
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 14.52,
|
| 1158 |
+
"learning_rate": 1.6842105263157895e-06,
|
| 1159 |
+
"loss": 3.8099,
|
| 1160 |
+
"step": 192000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 14.59,
|
| 1164 |
+
"learning_rate": 1.4736842105263159e-06,
|
| 1165 |
+
"loss": 3.7993,
|
| 1166 |
+
"step": 193000
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 14.67,
|
| 1170 |
+
"learning_rate": 1.2631578947368422e-06,
|
| 1171 |
+
"loss": 3.8072,
|
| 1172 |
+
"step": 194000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 14.74,
|
| 1176 |
+
"learning_rate": 1.0526315789473685e-06,
|
| 1177 |
+
"loss": 3.806,
|
| 1178 |
+
"step": 195000
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 14.82,
|
| 1182 |
+
"learning_rate": 8.421052631578948e-07,
|
| 1183 |
+
"loss": 3.8084,
|
| 1184 |
+
"step": 196000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 14.89,
|
| 1188 |
+
"learning_rate": 6.315789473684211e-07,
|
| 1189 |
+
"loss": 3.8074,
|
| 1190 |
+
"step": 197000
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 14.97,
|
| 1194 |
+
"learning_rate": 4.210526315789474e-07,
|
| 1195 |
+
"loss": 3.8013,
|
| 1196 |
+
"step": 198000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 15.04,
|
| 1200 |
+
"learning_rate": 2.105263157894737e-07,
|
| 1201 |
+
"loss": 3.8081,
|
| 1202 |
+
"step": 199000
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 15.12,
|
| 1206 |
+
"learning_rate": 0.0,
|
| 1207 |
+
"loss": 3.8009,
|
| 1208 |
+
"step": 200000
|
| 1209 |
+
}
|
| 1210 |
+
],
|
| 1211 |
+
"max_steps": 200000,
|
| 1212 |
+
"num_train_epochs": 16,
|
| 1213 |
+
"total_flos": 5.051724710578883e+18,
|
| 1214 |
+
"trial_name": null,
|
| 1215 |
+
"trial_params": null
|
| 1216 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06ffcd57892162760bf8f1bd73a1a42cf2c64618c52c361b5fe38a722735c150
|
| 3 |
+
size 4945
|