push of spaceBERT IEEE paper version
Browse files- config.json +22 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- trainer_state.json +1258 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bert-base-uncased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForMaskedLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 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 |
+
"type_vocab_size": 2,
|
| 21 |
+
"vocab_size": 30522
|
| 22 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c4b54b0193c989e1697e60ef3b031aa2882831cda3118a5ba97980a16517c96
|
| 3 |
+
size 876174701
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32f492df5496f2dc766a061b1f54febafcbb44b93ce5fdc04ae40dd138571f72
|
| 3 |
+
size 438112366
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c5511a75c01c80d003409eeab9acc2edb9a8d26ac9af5c8d38e3d5204e68dcf
|
| 3 |
+
size 326
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "name_or_path": "bert-base-uncased"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1258 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 68.99953929788998,
|
| 5 |
+
"global_step": 46782,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.3331797659633281,
|
| 12 |
+
"learning_rate": 4.976190476190477e-05,
|
| 13 |
+
"loss": 2.685059944085315,
|
| 14 |
+
"step": 226
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.6663595319266562,
|
| 18 |
+
"learning_rate": 4.9523809523809525e-05,
|
| 19 |
+
"loss": 2.5215273156630253,
|
| 20 |
+
"step": 452
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.9995392978899843,
|
| 24 |
+
"learning_rate": 4.928571428571429e-05,
|
| 25 |
+
"loss": 2.449192013360758,
|
| 26 |
+
"step": 678
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 1.3331797659633282,
|
| 30 |
+
"learning_rate": 4.904761904761905e-05,
|
| 31 |
+
"loss": 2.3792454542311945,
|
| 32 |
+
"step": 904
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.6663595319266562,
|
| 36 |
+
"learning_rate": 4.880952380952381e-05,
|
| 37 |
+
"loss": 2.3559764760785398,
|
| 38 |
+
"step": 1130
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 1.9995392978899842,
|
| 42 |
+
"learning_rate": 4.8571428571428576e-05,
|
| 43 |
+
"loss": 2.328374001832135,
|
| 44 |
+
"step": 1356
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.333179765963328,
|
| 48 |
+
"learning_rate": 4.8333333333333334e-05,
|
| 49 |
+
"loss": 2.2863990986241705,
|
| 50 |
+
"step": 1582
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 2.6663595319266564,
|
| 54 |
+
"learning_rate": 4.80952380952381e-05,
|
| 55 |
+
"loss": 2.2656879256256914,
|
| 56 |
+
"step": 1808
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.999539297889984,
|
| 60 |
+
"learning_rate": 4.785714285714286e-05,
|
| 61 |
+
"loss": 2.245354711481955,
|
| 62 |
+
"step": 2034
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 3.333179765963328,
|
| 66 |
+
"learning_rate": 4.761904761904762e-05,
|
| 67 |
+
"loss": 2.2154734113575083,
|
| 68 |
+
"step": 2260
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 3.6663595319266564,
|
| 72 |
+
"learning_rate": 4.738095238095238e-05,
|
| 73 |
+
"loss": 2.19904536694552,
|
| 74 |
+
"step": 2486
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 3.999539297889984,
|
| 78 |
+
"learning_rate": 4.714285714285714e-05,
|
| 79 |
+
"loss": 2.185555112045423,
|
| 80 |
+
"step": 2712
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 4.333179765963328,
|
| 84 |
+
"learning_rate": 4.690476190476191e-05,
|
| 85 |
+
"loss": 2.162922816993916,
|
| 86 |
+
"step": 2938
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.666359531926656,
|
| 90 |
+
"learning_rate": 4.666666666666667e-05,
|
| 91 |
+
"loss": 2.1419683135716263,
|
| 92 |
+
"step": 3164
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 4.999539297889984,
|
| 96 |
+
"learning_rate": 4.642857142857143e-05,
|
| 97 |
+
"loss": 2.1468805633814987,
|
| 98 |
+
"step": 3390
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 5.333179765963328,
|
| 102 |
+
"learning_rate": 4.6190476190476194e-05,
|
| 103 |
+
"loss": 2.114357872346861,
|
| 104 |
+
"step": 3616
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 5.666359531926656,
|
| 108 |
+
"learning_rate": 4.595238095238095e-05,
|
| 109 |
+
"loss": 2.1005791858234235,
|
| 110 |
+
"step": 3842
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 5.999539297889984,
|
| 114 |
+
"learning_rate": 4.5714285714285716e-05,
|
| 115 |
+
"loss": 2.0986059408272264,
|
| 116 |
+
"step": 4068
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 6.333179765963328,
|
| 120 |
+
"learning_rate": 4.547619047619048e-05,
|
| 121 |
+
"loss": 2.063803073579231,
|
| 122 |
+
"step": 4294
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 6.666359531926656,
|
| 126 |
+
"learning_rate": 4.523809523809524e-05,
|
| 127 |
+
"loss": 2.0596065014864493,
|
| 128 |
+
"step": 4520
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 6.999539297889984,
|
| 132 |
+
"learning_rate": 4.5e-05,
|
| 133 |
+
"loss": 2.0634367276081997,
|
| 134 |
+
"step": 4746
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 7.333179765963328,
|
| 138 |
+
"learning_rate": 4.476190476190477e-05,
|
| 139 |
+
"loss": 2.049277921693515,
|
| 140 |
+
"step": 4972
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 7.666359531926656,
|
| 144 |
+
"learning_rate": 4.4523809523809525e-05,
|
| 145 |
+
"loss": 2.0317085738730642,
|
| 146 |
+
"step": 5198
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 7.999539297889984,
|
| 150 |
+
"learning_rate": 4.428571428571428e-05,
|
| 151 |
+
"loss": 2.0282083697023645,
|
| 152 |
+
"step": 5424
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 8.333179765963328,
|
| 156 |
+
"learning_rate": 4.404761904761905e-05,
|
| 157 |
+
"loss": 1.9894960116496128,
|
| 158 |
+
"step": 5650
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 8.666359531926656,
|
| 162 |
+
"learning_rate": 4.380952380952381e-05,
|
| 163 |
+
"loss": 1.996843321133504,
|
| 164 |
+
"step": 5876
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 8.999539297889985,
|
| 168 |
+
"learning_rate": 4.3571428571428576e-05,
|
| 169 |
+
"loss": 2.0045918152395603,
|
| 170 |
+
"step": 6102
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 9.333179765963328,
|
| 174 |
+
"learning_rate": 4.3333333333333334e-05,
|
| 175 |
+
"loss": 1.966206238333103,
|
| 176 |
+
"step": 6328
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 9.666359531926656,
|
| 180 |
+
"learning_rate": 4.30952380952381e-05,
|
| 181 |
+
"loss": 1.9711824940369191,
|
| 182 |
+
"step": 6554
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 9.999539297889985,
|
| 186 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 187 |
+
"loss": 1.9805989223243916,
|
| 188 |
+
"step": 6780
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 10.333179765963328,
|
| 192 |
+
"learning_rate": 4.261904761904762e-05,
|
| 193 |
+
"loss": 1.9500477208500415,
|
| 194 |
+
"step": 7006
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 10.666359531926656,
|
| 198 |
+
"learning_rate": 4.2380952380952385e-05,
|
| 199 |
+
"loss": 1.9578322689090155,
|
| 200 |
+
"step": 7232
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 10.999539297889985,
|
| 204 |
+
"learning_rate": 4.214285714285714e-05,
|
| 205 |
+
"loss": 1.9484224572645878,
|
| 206 |
+
"step": 7458
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 11.333179765963328,
|
| 210 |
+
"learning_rate": 4.190476190476191e-05,
|
| 211 |
+
"loss": 1.9283131287161228,
|
| 212 |
+
"step": 7684
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 11.666359531926656,
|
| 216 |
+
"learning_rate": 4.166666666666667e-05,
|
| 217 |
+
"loss": 1.9156345603740321,
|
| 218 |
+
"step": 7910
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 11.999539297889985,
|
| 222 |
+
"learning_rate": 4.1428571428571437e-05,
|
| 223 |
+
"loss": 1.926576732534223,
|
| 224 |
+
"step": 8136
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 12.333179765963328,
|
| 228 |
+
"learning_rate": 4.119047619047619e-05,
|
| 229 |
+
"loss": 1.907513238687431,
|
| 230 |
+
"step": 8362
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 12.666359531926656,
|
| 234 |
+
"learning_rate": 4.095238095238095e-05,
|
| 235 |
+
"loss": 1.9078028856125553,
|
| 236 |
+
"step": 8588
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 12.999539297889985,
|
| 240 |
+
"learning_rate": 4.0714285714285717e-05,
|
| 241 |
+
"loss": 1.90263299182453,
|
| 242 |
+
"step": 8814
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 13.333179765963328,
|
| 246 |
+
"learning_rate": 4.047619047619048e-05,
|
| 247 |
+
"loss": 1.8844813794161366,
|
| 248 |
+
"step": 9040
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 13.666359531926656,
|
| 252 |
+
"learning_rate": 4.023809523809524e-05,
|
| 253 |
+
"loss": 1.8840125429946764,
|
| 254 |
+
"step": 9266
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 13.999539297889985,
|
| 258 |
+
"learning_rate": 4e-05,
|
| 259 |
+
"loss": 1.8717386701465708,
|
| 260 |
+
"step": 9492
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 14.333179765963328,
|
| 264 |
+
"learning_rate": 3.976190476190476e-05,
|
| 265 |
+
"loss": 1.858954910683421,
|
| 266 |
+
"step": 9718
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 14.666359531926656,
|
| 270 |
+
"learning_rate": 3.9523809523809526e-05,
|
| 271 |
+
"loss": 1.8676287895810288,
|
| 272 |
+
"step": 9944
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 14.999539297889985,
|
| 276 |
+
"learning_rate": 3.928571428571429e-05,
|
| 277 |
+
"loss": 1.8678895393304065,
|
| 278 |
+
"step": 10170
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 15.333179765963328,
|
| 282 |
+
"learning_rate": 3.904761904761905e-05,
|
| 283 |
+
"loss": 1.8486336227011892,
|
| 284 |
+
"step": 10396
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 15.666359531926656,
|
| 288 |
+
"learning_rate": 3.880952380952381e-05,
|
| 289 |
+
"loss": 1.8404835388723728,
|
| 290 |
+
"step": 10622
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 15.999539297889985,
|
| 294 |
+
"learning_rate": 3.857142857142858e-05,
|
| 295 |
+
"loss": 1.8481951688243226,
|
| 296 |
+
"step": 10848
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 16.33317976596333,
|
| 300 |
+
"learning_rate": 3.8333333333333334e-05,
|
| 301 |
+
"loss": 1.8267865476355087,
|
| 302 |
+
"step": 11074
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 16.666359531926656,
|
| 306 |
+
"learning_rate": 3.809523809523809e-05,
|
| 307 |
+
"loss": 1.818214213953609,
|
| 308 |
+
"step": 11300
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 16.999539297889985,
|
| 312 |
+
"learning_rate": 3.785714285714286e-05,
|
| 313 |
+
"loss": 1.8284451645032493,
|
| 314 |
+
"step": 11526
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 17.33317976596333,
|
| 318 |
+
"learning_rate": 3.761904761904762e-05,
|
| 319 |
+
"loss": 1.8039584539632882,
|
| 320 |
+
"step": 11752
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 17.666359531926656,
|
| 324 |
+
"learning_rate": 3.7380952380952386e-05,
|
| 325 |
+
"loss": 1.8007052261217507,
|
| 326 |
+
"step": 11978
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 17.999539297889985,
|
| 330 |
+
"learning_rate": 3.7142857142857143e-05,
|
| 331 |
+
"loss": 1.800786381274198,
|
| 332 |
+
"step": 12204
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 18.33317976596333,
|
| 336 |
+
"learning_rate": 3.690476190476191e-05,
|
| 337 |
+
"loss": 1.7864516266679342,
|
| 338 |
+
"step": 12430
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 18.666359531926656,
|
| 342 |
+
"learning_rate": 3.6666666666666666e-05,
|
| 343 |
+
"loss": 1.790079842626521,
|
| 344 |
+
"step": 12656
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 18.999539297889985,
|
| 348 |
+
"learning_rate": 3.642857142857143e-05,
|
| 349 |
+
"loss": 1.7918755185287611,
|
| 350 |
+
"step": 12882
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 19.33317976596333,
|
| 354 |
+
"learning_rate": 3.619047619047619e-05,
|
| 355 |
+
"loss": 1.7703251121318446,
|
| 356 |
+
"step": 13108
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 19.666359531926656,
|
| 360 |
+
"learning_rate": 3.595238095238095e-05,
|
| 361 |
+
"loss": 1.7819123394721377,
|
| 362 |
+
"step": 13334
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 19.999539297889985,
|
| 366 |
+
"learning_rate": 3.571428571428572e-05,
|
| 367 |
+
"loss": 1.7857482370022124,
|
| 368 |
+
"step": 13560
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 20.33317976596333,
|
| 372 |
+
"learning_rate": 3.547619047619048e-05,
|
| 373 |
+
"loss": 1.7561523977634126,
|
| 374 |
+
"step": 13786
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 20.666359531926656,
|
| 378 |
+
"learning_rate": 3.523809523809524e-05,
|
| 379 |
+
"loss": 1.7562858108925608,
|
| 380 |
+
"step": 14012
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 20.999539297889985,
|
| 384 |
+
"learning_rate": 3.5e-05,
|
| 385 |
+
"loss": 1.767443496569068,
|
| 386 |
+
"step": 14238
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 21.33317976596333,
|
| 390 |
+
"learning_rate": 3.476190476190476e-05,
|
| 391 |
+
"loss": 1.7421298406820382,
|
| 392 |
+
"step": 14464
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 21.666359531926656,
|
| 396 |
+
"learning_rate": 3.4523809523809526e-05,
|
| 397 |
+
"loss": 1.7460298791395878,
|
| 398 |
+
"step": 14690
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 21.999539297889985,
|
| 402 |
+
"learning_rate": 3.428571428571429e-05,
|
| 403 |
+
"loss": 1.7635893695122373,
|
| 404 |
+
"step": 14916
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 22.33317976596333,
|
| 408 |
+
"learning_rate": 3.404761904761905e-05,
|
| 409 |
+
"loss": 1.7318757993985066,
|
| 410 |
+
"step": 15142
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 22.666359531926656,
|
| 414 |
+
"learning_rate": 3.380952380952381e-05,
|
| 415 |
+
"loss": 1.7320329784292035,
|
| 416 |
+
"step": 15368
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 22.999539297889985,
|
| 420 |
+
"learning_rate": 3.357142857142857e-05,
|
| 421 |
+
"loss": 1.7314567903501799,
|
| 422 |
+
"step": 15594
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 23.33317976596333,
|
| 426 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 427 |
+
"loss": 1.7091308863817063,
|
| 428 |
+
"step": 15820
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 23.666359531926656,
|
| 432 |
+
"learning_rate": 3.309523809523809e-05,
|
| 433 |
+
"loss": 1.7127776019341123,
|
| 434 |
+
"step": 16046
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 23.999539297889985,
|
| 438 |
+
"learning_rate": 3.285714285714286e-05,
|
| 439 |
+
"loss": 1.714874267578125,
|
| 440 |
+
"step": 16272
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 24.33317976596333,
|
| 444 |
+
"learning_rate": 3.261904761904762e-05,
|
| 445 |
+
"loss": 1.7092985980278623,
|
| 446 |
+
"step": 16498
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 24.666359531926656,
|
| 450 |
+
"learning_rate": 3.2380952380952386e-05,
|
| 451 |
+
"loss": 1.7022116331927544,
|
| 452 |
+
"step": 16724
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 24.999539297889985,
|
| 456 |
+
"learning_rate": 3.2142857142857144e-05,
|
| 457 |
+
"loss": 1.7083171743207273,
|
| 458 |
+
"step": 16950
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 25.33317976596333,
|
| 462 |
+
"learning_rate": 3.19047619047619e-05,
|
| 463 |
+
"loss": 1.6961614558127074,
|
| 464 |
+
"step": 17176
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 25.666359531926656,
|
| 468 |
+
"learning_rate": 3.1666666666666666e-05,
|
| 469 |
+
"loss": 1.6968883413129148,
|
| 470 |
+
"step": 17402
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 25.999539297889985,
|
| 474 |
+
"learning_rate": 3.142857142857143e-05,
|
| 475 |
+
"loss": 1.6995788844285813,
|
| 476 |
+
"step": 17628
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 26.33317976596333,
|
| 480 |
+
"learning_rate": 3.1190476190476195e-05,
|
| 481 |
+
"loss": 1.6844244762859513,
|
| 482 |
+
"step": 17854
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 26.666359531926656,
|
| 486 |
+
"learning_rate": 3.095238095238095e-05,
|
| 487 |
+
"loss": 1.6839947995886337,
|
| 488 |
+
"step": 18080
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 26.999539297889985,
|
| 492 |
+
"learning_rate": 3.071428571428572e-05,
|
| 493 |
+
"loss": 1.6868854623980227,
|
| 494 |
+
"step": 18306
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 27.33317976596333,
|
| 498 |
+
"learning_rate": 3.0476190476190482e-05,
|
| 499 |
+
"loss": 1.6642917869365321,
|
| 500 |
+
"step": 18532
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 27.666359531926656,
|
| 504 |
+
"learning_rate": 3.0238095238095236e-05,
|
| 505 |
+
"loss": 1.6676389981160122,
|
| 506 |
+
"step": 18758
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 27.999539297889985,
|
| 510 |
+
"learning_rate": 3e-05,
|
| 511 |
+
"loss": 1.6597493939695105,
|
| 512 |
+
"step": 18984
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 28.33317976596333,
|
| 516 |
+
"learning_rate": 2.9761904761904762e-05,
|
| 517 |
+
"loss": 1.652435842868501,
|
| 518 |
+
"step": 19210
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 28.666359531926656,
|
| 522 |
+
"learning_rate": 2.9523809523809526e-05,
|
| 523 |
+
"loss": 1.6600899485360205,
|
| 524 |
+
"step": 19436
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 28.999539297889985,
|
| 528 |
+
"learning_rate": 2.9285714285714288e-05,
|
| 529 |
+
"loss": 1.6490822851130393,
|
| 530 |
+
"step": 19662
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 29.33317976596333,
|
| 534 |
+
"learning_rate": 2.9047619047619052e-05,
|
| 535 |
+
"loss": 1.6511269628474143,
|
| 536 |
+
"step": 19888
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 29.666359531926656,
|
| 540 |
+
"learning_rate": 2.880952380952381e-05,
|
| 541 |
+
"loss": 1.6514149893701604,
|
| 542 |
+
"step": 20114
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 29.999539297889985,
|
| 546 |
+
"learning_rate": 2.857142857142857e-05,
|
| 547 |
+
"loss": 1.6452392308057937,
|
| 548 |
+
"step": 20340
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 30.33317976596333,
|
| 552 |
+
"learning_rate": 2.8333333333333335e-05,
|
| 553 |
+
"loss": 1.6428464366271434,
|
| 554 |
+
"step": 20566
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 30.666359531926656,
|
| 558 |
+
"learning_rate": 2.8095238095238096e-05,
|
| 559 |
+
"loss": 1.632520692538371,
|
| 560 |
+
"step": 20792
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 30.999539297889985,
|
| 564 |
+
"learning_rate": 2.785714285714286e-05,
|
| 565 |
+
"loss": 1.6398790747718472,
|
| 566 |
+
"step": 21018
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 31.33317976596333,
|
| 570 |
+
"learning_rate": 2.7619047619047622e-05,
|
| 571 |
+
"loss": 1.6239518697282909,
|
| 572 |
+
"step": 21244
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 31.666359531926656,
|
| 576 |
+
"learning_rate": 2.7380952380952383e-05,
|
| 577 |
+
"loss": 1.6242760852374862,
|
| 578 |
+
"step": 21470
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 31.999539297889985,
|
| 582 |
+
"learning_rate": 2.714285714285714e-05,
|
| 583 |
+
"loss": 1.627472463962251,
|
| 584 |
+
"step": 21696
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 32.33317976596333,
|
| 588 |
+
"learning_rate": 2.6904761904761905e-05,
|
| 589 |
+
"loss": 1.6053812482715708,
|
| 590 |
+
"step": 21922
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 32.66635953192666,
|
| 594 |
+
"learning_rate": 2.6666666666666667e-05,
|
| 595 |
+
"loss": 1.610074878793902,
|
| 596 |
+
"step": 22148
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 32.99953929788998,
|
| 600 |
+
"learning_rate": 2.642857142857143e-05,
|
| 601 |
+
"loss": 1.620673390616358,
|
| 602 |
+
"step": 22374
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 33.33317976596333,
|
| 606 |
+
"learning_rate": 2.6190476190476192e-05,
|
| 607 |
+
"loss": 1.6081807634471792,
|
| 608 |
+
"step": 22600
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 33.66635953192666,
|
| 612 |
+
"learning_rate": 2.5952380952380957e-05,
|
| 613 |
+
"loss": 1.6156560847189574,
|
| 614 |
+
"step": 22826
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 33.99953929788998,
|
| 618 |
+
"learning_rate": 2.5714285714285714e-05,
|
| 619 |
+
"loss": 1.6041984895689299,
|
| 620 |
+
"step": 23052
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 34.33317976596333,
|
| 624 |
+
"learning_rate": 2.5476190476190476e-05,
|
| 625 |
+
"loss": 1.5947894881256914,
|
| 626 |
+
"step": 23278
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 34.66635953192666,
|
| 630 |
+
"learning_rate": 2.523809523809524e-05,
|
| 631 |
+
"loss": 1.5988714167502074,
|
| 632 |
+
"step": 23504
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 34.99953929788998,
|
| 636 |
+
"learning_rate": 2.5e-05,
|
| 637 |
+
"loss": 1.5926924173810841,
|
| 638 |
+
"step": 23730
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 35.33317976596333,
|
| 642 |
+
"learning_rate": 2.4761904761904762e-05,
|
| 643 |
+
"loss": 1.5834472116115874,
|
| 644 |
+
"step": 23956
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 35.66635953192666,
|
| 648 |
+
"learning_rate": 2.4523809523809523e-05,
|
| 649 |
+
"loss": 1.5848133458500415,
|
| 650 |
+
"step": 24182
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 35.99953929788998,
|
| 654 |
+
"learning_rate": 2.4285714285714288e-05,
|
| 655 |
+
"loss": 1.602300593283324,
|
| 656 |
+
"step": 24408
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 36.33317976596333,
|
| 660 |
+
"learning_rate": 2.404761904761905e-05,
|
| 661 |
+
"loss": 1.5835503772296737,
|
| 662 |
+
"step": 24634
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 36.66635953192666,
|
| 666 |
+
"learning_rate": 2.380952380952381e-05,
|
| 667 |
+
"loss": 1.5831868669628042,
|
| 668 |
+
"step": 24860
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 36.99953929788998,
|
| 672 |
+
"learning_rate": 2.357142857142857e-05,
|
| 673 |
+
"loss": 1.5691194787489628,
|
| 674 |
+
"step": 25086
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 37.33317976596333,
|
| 678 |
+
"learning_rate": 2.3333333333333336e-05,
|
| 679 |
+
"loss": 1.568113208872027,
|
| 680 |
+
"step": 25312
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 37.66635953192666,
|
| 684 |
+
"learning_rate": 2.3095238095238097e-05,
|
| 685 |
+
"loss": 1.5696247742239353,
|
| 686 |
+
"step": 25538
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 37.99953929788998,
|
| 690 |
+
"learning_rate": 2.2857142857142858e-05,
|
| 691 |
+
"loss": 1.5706803313398783,
|
| 692 |
+
"step": 25764
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 38.33317976596333,
|
| 696 |
+
"learning_rate": 2.261904761904762e-05,
|
| 697 |
+
"loss": 1.5539683114110896,
|
| 698 |
+
"step": 25990
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 38.66635953192666,
|
| 702 |
+
"learning_rate": 2.2380952380952384e-05,
|
| 703 |
+
"loss": 1.5682053017405282,
|
| 704 |
+
"step": 26216
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 38.99953929788998,
|
| 708 |
+
"learning_rate": 2.214285714285714e-05,
|
| 709 |
+
"loss": 1.5620073976769913,
|
| 710 |
+
"step": 26442
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 39.33317976596333,
|
| 714 |
+
"learning_rate": 2.1904761904761906e-05,
|
| 715 |
+
"loss": 1.5591868172704646,
|
| 716 |
+
"step": 26668
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 39.66635953192666,
|
| 720 |
+
"learning_rate": 2.1666666666666667e-05,
|
| 721 |
+
"loss": 1.5566102424554065,
|
| 722 |
+
"step": 26894
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 39.99953929788998,
|
| 726 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 727 |
+
"loss": 1.5528145849177268,
|
| 728 |
+
"step": 27120
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 40.33317976596333,
|
| 732 |
+
"learning_rate": 2.1190476190476193e-05,
|
| 733 |
+
"loss": 1.5426600633469303,
|
| 734 |
+
"step": 27346
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 40.66635953192666,
|
| 738 |
+
"learning_rate": 2.0952380952380954e-05,
|
| 739 |
+
"loss": 1.541890642284292,
|
| 740 |
+
"step": 27572
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 40.99953929788998,
|
| 744 |
+
"learning_rate": 2.0714285714285718e-05,
|
| 745 |
+
"loss": 1.5517368823026134,
|
| 746 |
+
"step": 27798
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 41.33317976596333,
|
| 750 |
+
"learning_rate": 2.0476190476190476e-05,
|
| 751 |
+
"loss": 1.540764057530766,
|
| 752 |
+
"step": 28024
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 41.66635953192666,
|
| 756 |
+
"learning_rate": 2.023809523809524e-05,
|
| 757 |
+
"loss": 1.5214318469562362,
|
| 758 |
+
"step": 28250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 41.99953929788998,
|
| 762 |
+
"learning_rate": 2e-05,
|
| 763 |
+
"loss": 1.5345232128042035,
|
| 764 |
+
"step": 28476
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 42.33317976596333,
|
| 768 |
+
"learning_rate": 1.9761904761904763e-05,
|
| 769 |
+
"loss": 1.5242220448181691,
|
| 770 |
+
"step": 28702
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 42.66635953192666,
|
| 774 |
+
"learning_rate": 1.9523809523809524e-05,
|
| 775 |
+
"loss": 1.535507067114906,
|
| 776 |
+
"step": 28928
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 42.99953929788998,
|
| 780 |
+
"learning_rate": 1.928571428571429e-05,
|
| 781 |
+
"loss": 1.5329083468006774,
|
| 782 |
+
"step": 29154
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 43.33317976596333,
|
| 786 |
+
"learning_rate": 1.9047619047619046e-05,
|
| 787 |
+
"loss": 1.5211832502246958,
|
| 788 |
+
"step": 29380
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 43.66635953192666,
|
| 792 |
+
"learning_rate": 1.880952380952381e-05,
|
| 793 |
+
"loss": 1.5096026395274476,
|
| 794 |
+
"step": 29606
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 43.99953929788998,
|
| 798 |
+
"learning_rate": 1.8571428571428572e-05,
|
| 799 |
+
"loss": 1.5239440107767561,
|
| 800 |
+
"step": 29832
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 44.33317976596333,
|
| 804 |
+
"learning_rate": 1.8333333333333333e-05,
|
| 805 |
+
"loss": 1.518264635474281,
|
| 806 |
+
"step": 30058
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 44.66635953192666,
|
| 810 |
+
"learning_rate": 1.8095238095238094e-05,
|
| 811 |
+
"loss": 1.5015280394427544,
|
| 812 |
+
"step": 30284
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 44.99953929788998,
|
| 816 |
+
"learning_rate": 1.785714285714286e-05,
|
| 817 |
+
"loss": 1.5198267033669801,
|
| 818 |
+
"step": 30510
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 45.33317976596333,
|
| 822 |
+
"learning_rate": 1.761904761904762e-05,
|
| 823 |
+
"loss": 1.5029577744745575,
|
| 824 |
+
"step": 30736
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 45.66635953192666,
|
| 828 |
+
"learning_rate": 1.738095238095238e-05,
|
| 829 |
+
"loss": 1.5046313800642976,
|
| 830 |
+
"step": 30962
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 45.99953929788998,
|
| 834 |
+
"learning_rate": 1.7142857142857145e-05,
|
| 835 |
+
"loss": 1.5148508527637583,
|
| 836 |
+
"step": 31188
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 46.33317976596333,
|
| 840 |
+
"learning_rate": 1.6904761904761906e-05,
|
| 841 |
+
"loss": 1.4938382849229122,
|
| 842 |
+
"step": 31414
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 46.66635953192666,
|
| 846 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 847 |
+
"loss": 1.5032175789892146,
|
| 848 |
+
"step": 31640
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 46.99953929788998,
|
| 852 |
+
"learning_rate": 1.642857142857143e-05,
|
| 853 |
+
"loss": 1.4958131503214878,
|
| 854 |
+
"step": 31866
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 47.33317976596333,
|
| 858 |
+
"learning_rate": 1.6190476190476193e-05,
|
| 859 |
+
"loss": 1.5038191533721654,
|
| 860 |
+
"step": 32092
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 47.66635953192666,
|
| 864 |
+
"learning_rate": 1.595238095238095e-05,
|
| 865 |
+
"loss": 1.4897128755012445,
|
| 866 |
+
"step": 32318
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 47.99953929788998,
|
| 870 |
+
"learning_rate": 1.5714285714285715e-05,
|
| 871 |
+
"loss": 1.4913623100888413,
|
| 872 |
+
"step": 32544
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 48.33317976596333,
|
| 876 |
+
"learning_rate": 1.5476190476190476e-05,
|
| 877 |
+
"loss": 1.491287096411781,
|
| 878 |
+
"step": 32770
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 48.66635953192666,
|
| 882 |
+
"learning_rate": 1.5238095238095241e-05,
|
| 883 |
+
"loss": 1.4894442938070382,
|
| 884 |
+
"step": 32996
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 48.99953929788998,
|
| 888 |
+
"learning_rate": 1.5e-05,
|
| 889 |
+
"loss": 1.4941079468853706,
|
| 890 |
+
"step": 33222
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 49.33317976596333,
|
| 894 |
+
"learning_rate": 1.4761904761904763e-05,
|
| 895 |
+
"loss": 1.494901538950152,
|
| 896 |
+
"step": 33448
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 49.66635953192666,
|
| 900 |
+
"learning_rate": 1.4523809523809526e-05,
|
| 901 |
+
"loss": 1.4872477033496958,
|
| 902 |
+
"step": 33674
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 49.99953929788998,
|
| 906 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 907 |
+
"loss": 1.4863664745229535,
|
| 908 |
+
"step": 33900
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 50.33317976596333,
|
| 912 |
+
"learning_rate": 1.4047619047619048e-05,
|
| 913 |
+
"loss": 1.4825258508192753,
|
| 914 |
+
"step": 34126
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 50.66635953192666,
|
| 918 |
+
"learning_rate": 1.3809523809523811e-05,
|
| 919 |
+
"loss": 1.4887811441337113,
|
| 920 |
+
"step": 34352
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 50.99953929788998,
|
| 924 |
+
"learning_rate": 1.357142857142857e-05,
|
| 925 |
+
"loss": 1.475349223719234,
|
| 926 |
+
"step": 34578
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 51.33317976596333,
|
| 930 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 931 |
+
"loss": 1.468778492075152,
|
| 932 |
+
"step": 34804
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 51.66635953192666,
|
| 936 |
+
"learning_rate": 1.3095238095238096e-05,
|
| 937 |
+
"loss": 1.4681135869659154,
|
| 938 |
+
"step": 35030
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 51.99953929788998,
|
| 942 |
+
"learning_rate": 1.2857142857142857e-05,
|
| 943 |
+
"loss": 1.4756152397763413,
|
| 944 |
+
"step": 35256
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 52.33317976596333,
|
| 948 |
+
"learning_rate": 1.261904761904762e-05,
|
| 949 |
+
"loss": 1.461721504684043,
|
| 950 |
+
"step": 35482
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 52.66635953192666,
|
| 954 |
+
"learning_rate": 1.2380952380952381e-05,
|
| 955 |
+
"loss": 1.472177421097207,
|
| 956 |
+
"step": 35708
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 52.99953929788998,
|
| 960 |
+
"learning_rate": 1.2142857142857144e-05,
|
| 961 |
+
"loss": 1.470105331555932,
|
| 962 |
+
"step": 35934
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 53.33317976596333,
|
| 966 |
+
"learning_rate": 1.1904761904761905e-05,
|
| 967 |
+
"loss": 1.463019582022608,
|
| 968 |
+
"step": 36160
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 53.66635953192666,
|
| 972 |
+
"learning_rate": 1.1666666666666668e-05,
|
| 973 |
+
"loss": 1.4712693205976908,
|
| 974 |
+
"step": 36386
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 53.99953929788998,
|
| 978 |
+
"learning_rate": 1.1428571428571429e-05,
|
| 979 |
+
"loss": 1.4639480725853844,
|
| 980 |
+
"step": 36612
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 54.33317976596333,
|
| 984 |
+
"learning_rate": 1.1190476190476192e-05,
|
| 985 |
+
"loss": 1.4621197185685149,
|
| 986 |
+
"step": 36838
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 54.66635953192666,
|
| 990 |
+
"learning_rate": 1.0952380952380953e-05,
|
| 991 |
+
"loss": 1.4438346930309736,
|
| 992 |
+
"step": 37064
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 54.99953929788998,
|
| 996 |
+
"learning_rate": 1.0714285714285714e-05,
|
| 997 |
+
"loss": 1.45255799420112,
|
| 998 |
+
"step": 37290
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 55.33317976596333,
|
| 1002 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 1003 |
+
"loss": 1.4527645955043555,
|
| 1004 |
+
"step": 37516
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 55.66635953192666,
|
| 1008 |
+
"learning_rate": 1.0238095238095238e-05,
|
| 1009 |
+
"loss": 1.4556512073077987,
|
| 1010 |
+
"step": 37742
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 55.99953929788998,
|
| 1014 |
+
"learning_rate": 1e-05,
|
| 1015 |
+
"loss": 1.4523168243138136,
|
| 1016 |
+
"step": 37968
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 56.33317976596333,
|
| 1020 |
+
"learning_rate": 9.761904761904762e-06,
|
| 1021 |
+
"loss": 1.4451588318411228,
|
| 1022 |
+
"step": 38194
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 56.66635953192666,
|
| 1026 |
+
"learning_rate": 9.523809523809523e-06,
|
| 1027 |
+
"loss": 1.4351428546736726,
|
| 1028 |
+
"step": 38420
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 56.99953929788998,
|
| 1032 |
+
"learning_rate": 9.285714285714286e-06,
|
| 1033 |
+
"loss": 1.4480798771951051,
|
| 1034 |
+
"step": 38646
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 57.33317976596333,
|
| 1038 |
+
"learning_rate": 9.047619047619047e-06,
|
| 1039 |
+
"loss": 1.4419262641299087,
|
| 1040 |
+
"step": 38872
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 57.66635953192666,
|
| 1044 |
+
"learning_rate": 8.80952380952381e-06,
|
| 1045 |
+
"loss": 1.4400379552250415,
|
| 1046 |
+
"step": 39098
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 57.99953929788998,
|
| 1050 |
+
"learning_rate": 8.571428571428573e-06,
|
| 1051 |
+
"loss": 1.4458791006982854,
|
| 1052 |
+
"step": 39324
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 58.33317976596333,
|
| 1056 |
+
"learning_rate": 8.333333333333334e-06,
|
| 1057 |
+
"loss": 1.4246890987970133,
|
| 1058 |
+
"step": 39550
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 58.66635953192666,
|
| 1062 |
+
"learning_rate": 8.095238095238097e-06,
|
| 1063 |
+
"loss": 1.4372091377730918,
|
| 1064 |
+
"step": 39776
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 58.99953929788998,
|
| 1068 |
+
"learning_rate": 7.857142857142858e-06,
|
| 1069 |
+
"loss": 1.4388618131654451,
|
| 1070 |
+
"step": 40002
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 59.33317976596333,
|
| 1074 |
+
"learning_rate": 7.6190476190476205e-06,
|
| 1075 |
+
"loss": 1.437955738168902,
|
| 1076 |
+
"step": 40228
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 59.66635953192666,
|
| 1080 |
+
"learning_rate": 7.380952380952382e-06,
|
| 1081 |
+
"loss": 1.4384045896276962,
|
| 1082 |
+
"step": 40454
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 59.99953929788998,
|
| 1086 |
+
"learning_rate": 7.142857142857143e-06,
|
| 1087 |
+
"loss": 1.4317560786694552,
|
| 1088 |
+
"step": 40680
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 60.33317976596333,
|
| 1092 |
+
"learning_rate": 6.9047619047619055e-06,
|
| 1093 |
+
"loss": 1.4312950741928236,
|
| 1094 |
+
"step": 40906
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 60.66635953192666,
|
| 1098 |
+
"learning_rate": 6.666666666666667e-06,
|
| 1099 |
+
"loss": 1.4349378737728153,
|
| 1100 |
+
"step": 41132
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 60.99953929788998,
|
| 1104 |
+
"learning_rate": 6.428571428571429e-06,
|
| 1105 |
+
"loss": 1.4232025146484375,
|
| 1106 |
+
"step": 41358
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 61.33317976596333,
|
| 1110 |
+
"learning_rate": 6.190476190476191e-06,
|
| 1111 |
+
"loss": 1.4273036180344303,
|
| 1112 |
+
"step": 41584
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 61.66635953192666,
|
| 1116 |
+
"learning_rate": 5.9523809523809525e-06,
|
| 1117 |
+
"loss": 1.437505671408324,
|
| 1118 |
+
"step": 41810
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 61.99953929788998,
|
| 1122 |
+
"learning_rate": 5.7142857142857145e-06,
|
| 1123 |
+
"loss": 1.4295697507605087,
|
| 1124 |
+
"step": 42036
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 62.33317976596333,
|
| 1128 |
+
"learning_rate": 5.4761904761904765e-06,
|
| 1129 |
+
"loss": 1.4301999522521434,
|
| 1130 |
+
"step": 42262
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 62.66635953192666,
|
| 1134 |
+
"learning_rate": 5.2380952380952384e-06,
|
| 1135 |
+
"loss": 1.4240087998651825,
|
| 1136 |
+
"step": 42488
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 62.99953929788998,
|
| 1140 |
+
"learning_rate": 5e-06,
|
| 1141 |
+
"loss": 1.418294991012168,
|
| 1142 |
+
"step": 42714
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 63.33317976596333,
|
| 1146 |
+
"learning_rate": 4.7619047619047615e-06,
|
| 1147 |
+
"loss": 1.4275296641662059,
|
| 1148 |
+
"step": 42940
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 63.66635953192666,
|
| 1152 |
+
"learning_rate": 4.5238095238095235e-06,
|
| 1153 |
+
"loss": 1.4242185069396434,
|
| 1154 |
+
"step": 43166
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 63.99953929788998,
|
| 1158 |
+
"learning_rate": 4.285714285714286e-06,
|
| 1159 |
+
"loss": 1.4242925053149198,
|
| 1160 |
+
"step": 43392
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 64.33317976596332,
|
| 1164 |
+
"learning_rate": 4.047619047619048e-06,
|
| 1165 |
+
"loss": 1.4117900206979397,
|
| 1166 |
+
"step": 43618
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 64.66635953192666,
|
| 1170 |
+
"learning_rate": 3.8095238095238102e-06,
|
| 1171 |
+
"loss": 1.4196827306156665,
|
| 1172 |
+
"step": 43844
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 64.99953929788998,
|
| 1176 |
+
"learning_rate": 3.5714285714285714e-06,
|
| 1177 |
+
"loss": 1.4127752253439574,
|
| 1178 |
+
"step": 44070
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 65.33317976596332,
|
| 1182 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 1183 |
+
"loss": 1.415105228930448,
|
| 1184 |
+
"step": 44296
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 65.66635953192666,
|
| 1188 |
+
"learning_rate": 3.0952380952380953e-06,
|
| 1189 |
+
"loss": 1.4221684278640072,
|
| 1190 |
+
"step": 44522
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 65.99953929788998,
|
| 1194 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 1195 |
+
"loss": 1.4126794865701051,
|
| 1196 |
+
"step": 44748
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 66.33317976596332,
|
| 1200 |
+
"learning_rate": 2.6190476190476192e-06,
|
| 1201 |
+
"loss": 1.412820731644082,
|
| 1202 |
+
"step": 44974
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 66.66635953192666,
|
| 1206 |
+
"learning_rate": 2.3809523809523808e-06,
|
| 1207 |
+
"loss": 1.4091276995903623,
|
| 1208 |
+
"step": 45200
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 66.99953929788998,
|
| 1212 |
+
"learning_rate": 2.142857142857143e-06,
|
| 1213 |
+
"loss": 1.413559094994469,
|
| 1214 |
+
"step": 45426
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 67.33317976596332,
|
| 1218 |
+
"learning_rate": 1.9047619047619051e-06,
|
| 1219 |
+
"loss": 1.4078961937828403,
|
| 1220 |
+
"step": 45652
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 67.66635953192666,
|
| 1224 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 1225 |
+
"loss": 1.4104151092799364,
|
| 1226 |
+
"step": 45878
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 67.99953929788998,
|
| 1230 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 1231 |
+
"loss": 1.4099604513792865,
|
| 1232 |
+
"step": 46104
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 68.33317976596332,
|
| 1236 |
+
"learning_rate": 1.1904761904761904e-06,
|
| 1237 |
+
"loss": 1.406501297402171,
|
| 1238 |
+
"step": 46330
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 68.66635953192666,
|
| 1242 |
+
"learning_rate": 9.523809523809526e-07,
|
| 1243 |
+
"loss": 1.4021470061445658,
|
| 1244 |
+
"step": 46556
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 68.99953929788998,
|
| 1248 |
+
"learning_rate": 7.142857142857143e-07,
|
| 1249 |
+
"loss": 1.4063275092470962,
|
| 1250 |
+
"step": 46782
|
| 1251 |
+
}
|
| 1252 |
+
],
|
| 1253 |
+
"max_steps": 47460,
|
| 1254 |
+
"num_train_epochs": 70,
|
| 1255 |
+
"total_flos": 531450425497308624,
|
| 1256 |
+
"trial_name": null,
|
| 1257 |
+
"trial_params": null
|
| 1258 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:540b29f8c5d368e254828bc79018dfdc10a2425ffe352141b6a1cdb83a5f254b
|
| 3 |
+
size 1587
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|