Cameron Clarke
commited on
Commit
•
f0eaac5
1
Parent(s):
9c4364d
add model files
Browse files- config.json +24 -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 +1472 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bert-base-cased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 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 |
+
"transformers_version": "4.5.0",
|
| 21 |
+
"type_vocab_size": 2,
|
| 22 |
+
"use_cache": true,
|
| 23 |
+
"vocab_size": 28996
|
| 24 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8c19c1a4796b30c7b75d61f7ef61fd8f6633e64b3ce05e29fe11da599f3c9cf2
|
| 3 |
+
size 866621073
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fcac99adeccde07ba3de340e4f2857fec316e6f0d87696c86c68a62101457b41
|
| 3 |
+
size 433336585
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1778b073cba929e6fa6a930807254da0b5e6e69b8b1e5c75ac6aac915c7679e0
|
| 3 |
+
size 623
|
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": false, "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, "special_tokens_map_file": null, "name_or_path": "bert-base-cased"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1472 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.0,
|
| 3 |
+
"best_model_checkpoint": "/scratch/ccc779/jigsaw_severetoxic_model/checkpoint-59839",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 119678,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 1.2130648541044402e-05,
|
| 13 |
+
"loss": 0.0643,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.02,
|
| 18 |
+
"learning_rate": 1.2079755555170145e-05,
|
| 19 |
+
"loss": 0.0878,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.03,
|
| 24 |
+
"learning_rate": 1.2028862569295889e-05,
|
| 25 |
+
"loss": 0.0959,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 1.1977969583421632e-05,
|
| 31 |
+
"loss": 0.0328,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.04,
|
| 36 |
+
"learning_rate": 1.1927076597547376e-05,
|
| 37 |
+
"loss": 0.0508,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.05,
|
| 42 |
+
"learning_rate": 1.1876183611673118e-05,
|
| 43 |
+
"loss": 0.0522,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.06,
|
| 48 |
+
"learning_rate": 1.182529062579886e-05,
|
| 49 |
+
"loss": 0.0431,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.07,
|
| 54 |
+
"learning_rate": 1.1774397639924603e-05,
|
| 55 |
+
"loss": 0.0535,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.08,
|
| 60 |
+
"learning_rate": 1.1723504654050347e-05,
|
| 61 |
+
"loss": 0.0679,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.08,
|
| 66 |
+
"learning_rate": 1.167261166817609e-05,
|
| 67 |
+
"loss": 0.0714,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.09,
|
| 72 |
+
"learning_rate": 1.1621718682301832e-05,
|
| 73 |
+
"loss": 0.0482,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.1,
|
| 78 |
+
"learning_rate": 1.1570825696427575e-05,
|
| 79 |
+
"loss": 0.0648,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.11,
|
| 84 |
+
"learning_rate": 1.1519932710553319e-05,
|
| 85 |
+
"loss": 0.0569,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.12,
|
| 90 |
+
"learning_rate": 1.1469039724679063e-05,
|
| 91 |
+
"loss": 0.0393,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.13,
|
| 96 |
+
"learning_rate": 1.1418146738804806e-05,
|
| 97 |
+
"loss": 0.058,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.13,
|
| 102 |
+
"learning_rate": 1.1367253752930548e-05,
|
| 103 |
+
"loss": 0.0459,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.14,
|
| 108 |
+
"learning_rate": 1.131636076705629e-05,
|
| 109 |
+
"loss": 0.0565,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.15,
|
| 114 |
+
"learning_rate": 1.1265467781182033e-05,
|
| 115 |
+
"loss": 0.049,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.16,
|
| 120 |
+
"learning_rate": 1.1214574795307776e-05,
|
| 121 |
+
"loss": 0.0377,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.17,
|
| 126 |
+
"learning_rate": 1.116368180943352e-05,
|
| 127 |
+
"loss": 0.0548,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.18,
|
| 132 |
+
"learning_rate": 1.1112788823559262e-05,
|
| 133 |
+
"loss": 0.0944,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.18,
|
| 138 |
+
"learning_rate": 1.1061895837685007e-05,
|
| 139 |
+
"loss": 0.045,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.19,
|
| 144 |
+
"learning_rate": 1.1011002851810749e-05,
|
| 145 |
+
"loss": 0.0561,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.2,
|
| 150 |
+
"learning_rate": 1.0960109865936491e-05,
|
| 151 |
+
"loss": 0.0707,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.21,
|
| 156 |
+
"learning_rate": 1.0909216880062236e-05,
|
| 157 |
+
"loss": 0.0581,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.22,
|
| 162 |
+
"learning_rate": 1.0858323894187978e-05,
|
| 163 |
+
"loss": 0.0794,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.23,
|
| 168 |
+
"learning_rate": 1.080743090831372e-05,
|
| 169 |
+
"loss": 0.0863,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.23,
|
| 174 |
+
"learning_rate": 1.0756537922439463e-05,
|
| 175 |
+
"loss": 0.0733,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.24,
|
| 180 |
+
"learning_rate": 1.0705644936565207e-05,
|
| 181 |
+
"loss": 0.0425,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.25,
|
| 186 |
+
"learning_rate": 1.065475195069095e-05,
|
| 187 |
+
"loss": 0.065,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.26,
|
| 192 |
+
"learning_rate": 1.0603858964816694e-05,
|
| 193 |
+
"loss": 0.0392,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.27,
|
| 198 |
+
"learning_rate": 1.0552965978942436e-05,
|
| 199 |
+
"loss": 0.065,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.28,
|
| 204 |
+
"learning_rate": 1.0502072993068179e-05,
|
| 205 |
+
"loss": 0.0679,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.28,
|
| 210 |
+
"learning_rate": 1.0451180007193921e-05,
|
| 211 |
+
"loss": 0.0537,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.29,
|
| 216 |
+
"learning_rate": 1.0400287021319664e-05,
|
| 217 |
+
"loss": 0.0372,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.3,
|
| 222 |
+
"learning_rate": 1.0349394035445408e-05,
|
| 223 |
+
"loss": 0.0705,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.31,
|
| 228 |
+
"learning_rate": 1.029850104957115e-05,
|
| 229 |
+
"loss": 0.0771,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.32,
|
| 234 |
+
"learning_rate": 1.0247608063696895e-05,
|
| 235 |
+
"loss": 0.0689,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.33,
|
| 240 |
+
"learning_rate": 1.0196715077822637e-05,
|
| 241 |
+
"loss": 0.0722,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.33,
|
| 246 |
+
"learning_rate": 1.014582209194838e-05,
|
| 247 |
+
"loss": 0.0495,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.34,
|
| 252 |
+
"learning_rate": 1.0094929106074122e-05,
|
| 253 |
+
"loss": 0.0653,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.35,
|
| 258 |
+
"learning_rate": 1.0044036120199866e-05,
|
| 259 |
+
"loss": 0.0589,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.36,
|
| 264 |
+
"learning_rate": 9.993143134325609e-06,
|
| 265 |
+
"loss": 0.0655,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.37,
|
| 270 |
+
"learning_rate": 9.942250148451351e-06,
|
| 271 |
+
"loss": 0.0524,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.38,
|
| 276 |
+
"learning_rate": 9.891357162577094e-06,
|
| 277 |
+
"loss": 0.0801,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.38,
|
| 282 |
+
"learning_rate": 9.840464176702838e-06,
|
| 283 |
+
"loss": 0.061,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.39,
|
| 288 |
+
"learning_rate": 9.789571190828582e-06,
|
| 289 |
+
"loss": 0.0738,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.4,
|
| 294 |
+
"learning_rate": 9.738678204954325e-06,
|
| 295 |
+
"loss": 0.0558,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.41,
|
| 300 |
+
"learning_rate": 9.687785219080067e-06,
|
| 301 |
+
"loss": 0.0557,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.42,
|
| 306 |
+
"learning_rate": 9.63689223320581e-06,
|
| 307 |
+
"loss": 0.1104,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.43,
|
| 312 |
+
"learning_rate": 9.585999247331552e-06,
|
| 313 |
+
"loss": 0.0464,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.43,
|
| 318 |
+
"learning_rate": 9.535106261457295e-06,
|
| 319 |
+
"loss": 0.0692,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.44,
|
| 324 |
+
"learning_rate": 9.484213275583039e-06,
|
| 325 |
+
"loss": 0.063,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.45,
|
| 330 |
+
"learning_rate": 9.433320289708781e-06,
|
| 331 |
+
"loss": 0.0671,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.46,
|
| 336 |
+
"learning_rate": 9.382427303834525e-06,
|
| 337 |
+
"loss": 0.0622,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.47,
|
| 342 |
+
"learning_rate": 9.331534317960268e-06,
|
| 343 |
+
"loss": 0.0488,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.48,
|
| 348 |
+
"learning_rate": 9.28064133208601e-06,
|
| 349 |
+
"loss": 0.1113,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.48,
|
| 354 |
+
"learning_rate": 9.229748346211755e-06,
|
| 355 |
+
"loss": 0.0815,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.49,
|
| 360 |
+
"learning_rate": 9.178855360337497e-06,
|
| 361 |
+
"loss": 0.0589,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.5,
|
| 366 |
+
"learning_rate": 9.12796237446324e-06,
|
| 367 |
+
"loss": 0.0609,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.51,
|
| 372 |
+
"learning_rate": 9.077069388588982e-06,
|
| 373 |
+
"loss": 0.0898,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.52,
|
| 378 |
+
"learning_rate": 9.026176402714725e-06,
|
| 379 |
+
"loss": 0.0247,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.53,
|
| 384 |
+
"learning_rate": 8.975283416840469e-06,
|
| 385 |
+
"loss": 0.0581,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.53,
|
| 390 |
+
"learning_rate": 8.924390430966213e-06,
|
| 391 |
+
"loss": 0.0489,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.54,
|
| 396 |
+
"learning_rate": 8.873497445091955e-06,
|
| 397 |
+
"loss": 0.1036,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.55,
|
| 402 |
+
"learning_rate": 8.822604459217698e-06,
|
| 403 |
+
"loss": 0.0634,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.56,
|
| 408 |
+
"learning_rate": 8.77171147334344e-06,
|
| 409 |
+
"loss": 0.0456,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.57,
|
| 414 |
+
"learning_rate": 8.720818487469183e-06,
|
| 415 |
+
"loss": 0.0903,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.58,
|
| 420 |
+
"learning_rate": 8.669925501594925e-06,
|
| 421 |
+
"loss": 0.1017,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.58,
|
| 426 |
+
"learning_rate": 8.61903251572067e-06,
|
| 427 |
+
"loss": 0.0675,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.59,
|
| 432 |
+
"learning_rate": 8.568139529846412e-06,
|
| 433 |
+
"loss": 0.0473,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.6,
|
| 438 |
+
"learning_rate": 8.517246543972156e-06,
|
| 439 |
+
"loss": 0.0504,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.61,
|
| 444 |
+
"learning_rate": 8.466353558097899e-06,
|
| 445 |
+
"loss": 0.0489,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.62,
|
| 450 |
+
"learning_rate": 8.415460572223641e-06,
|
| 451 |
+
"loss": 0.0622,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.63,
|
| 456 |
+
"learning_rate": 8.364567586349385e-06,
|
| 457 |
+
"loss": 0.0454,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.64,
|
| 462 |
+
"learning_rate": 8.313674600475128e-06,
|
| 463 |
+
"loss": 0.0253,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.64,
|
| 468 |
+
"learning_rate": 8.26278161460087e-06,
|
| 469 |
+
"loss": 0.0592,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.65,
|
| 474 |
+
"learning_rate": 8.211888628726613e-06,
|
| 475 |
+
"loss": 0.0734,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.66,
|
| 480 |
+
"learning_rate": 8.160995642852357e-06,
|
| 481 |
+
"loss": 0.0595,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.67,
|
| 486 |
+
"learning_rate": 8.1101026569781e-06,
|
| 487 |
+
"loss": 0.0375,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.68,
|
| 492 |
+
"learning_rate": 8.059209671103844e-06,
|
| 493 |
+
"loss": 0.0542,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.69,
|
| 498 |
+
"learning_rate": 8.008316685229586e-06,
|
| 499 |
+
"loss": 0.0207,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.69,
|
| 504 |
+
"learning_rate": 7.957423699355329e-06,
|
| 505 |
+
"loss": 0.0493,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.7,
|
| 510 |
+
"learning_rate": 7.906530713481071e-06,
|
| 511 |
+
"loss": 0.0396,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.71,
|
| 516 |
+
"learning_rate": 7.855637727606814e-06,
|
| 517 |
+
"loss": 0.0734,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.72,
|
| 522 |
+
"learning_rate": 7.804744741732558e-06,
|
| 523 |
+
"loss": 0.0744,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.73,
|
| 528 |
+
"learning_rate": 7.7538517558583e-06,
|
| 529 |
+
"loss": 0.0579,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.74,
|
| 534 |
+
"learning_rate": 7.702958769984044e-06,
|
| 535 |
+
"loss": 0.0528,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.74,
|
| 540 |
+
"learning_rate": 7.652065784109787e-06,
|
| 541 |
+
"loss": 0.052,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.75,
|
| 546 |
+
"learning_rate": 7.601172798235529e-06,
|
| 547 |
+
"loss": 0.0654,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.76,
|
| 552 |
+
"learning_rate": 7.550279812361272e-06,
|
| 553 |
+
"loss": 0.0469,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.77,
|
| 558 |
+
"learning_rate": 7.499386826487016e-06,
|
| 559 |
+
"loss": 0.0457,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.78,
|
| 564 |
+
"learning_rate": 7.4484938406127585e-06,
|
| 565 |
+
"loss": 0.0665,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.79,
|
| 570 |
+
"learning_rate": 7.397600854738501e-06,
|
| 571 |
+
"loss": 0.0464,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.79,
|
| 576 |
+
"learning_rate": 7.346707868864244e-06,
|
| 577 |
+
"loss": 0.0581,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.8,
|
| 582 |
+
"learning_rate": 7.295814882989987e-06,
|
| 583 |
+
"loss": 0.0382,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.81,
|
| 588 |
+
"learning_rate": 7.244921897115731e-06,
|
| 589 |
+
"loss": 0.0675,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.82,
|
| 594 |
+
"learning_rate": 7.1940289112414734e-06,
|
| 595 |
+
"loss": 0.073,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.83,
|
| 600 |
+
"learning_rate": 7.143135925367217e-06,
|
| 601 |
+
"loss": 0.052,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.84,
|
| 606 |
+
"learning_rate": 7.092242939492959e-06,
|
| 607 |
+
"loss": 0.0752,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.84,
|
| 612 |
+
"learning_rate": 7.041349953618702e-06,
|
| 613 |
+
"loss": 0.0699,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.85,
|
| 618 |
+
"learning_rate": 6.990456967744445e-06,
|
| 619 |
+
"loss": 0.0898,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.86,
|
| 624 |
+
"learning_rate": 6.939563981870188e-06,
|
| 625 |
+
"loss": 0.067,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.87,
|
| 630 |
+
"learning_rate": 6.888670995995932e-06,
|
| 631 |
+
"loss": 0.0898,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.88,
|
| 636 |
+
"learning_rate": 6.837778010121674e-06,
|
| 637 |
+
"loss": 0.0536,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.89,
|
| 642 |
+
"learning_rate": 6.786885024247417e-06,
|
| 643 |
+
"loss": 0.0622,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.89,
|
| 648 |
+
"learning_rate": 6.73599203837316e-06,
|
| 649 |
+
"loss": 0.0901,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.9,
|
| 654 |
+
"learning_rate": 6.685099052498904e-06,
|
| 655 |
+
"loss": 0.0474,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.91,
|
| 660 |
+
"learning_rate": 6.634206066624647e-06,
|
| 661 |
+
"loss": 0.0783,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.92,
|
| 666 |
+
"learning_rate": 6.583313080750389e-06,
|
| 667 |
+
"loss": 0.1103,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.93,
|
| 672 |
+
"learning_rate": 6.5324200948761325e-06,
|
| 673 |
+
"loss": 0.0861,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.94,
|
| 678 |
+
"learning_rate": 6.481527109001875e-06,
|
| 679 |
+
"loss": 0.0615,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.94,
|
| 684 |
+
"learning_rate": 6.4306341231276175e-06,
|
| 685 |
+
"loss": 0.0398,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.95,
|
| 690 |
+
"learning_rate": 6.379741137253362e-06,
|
| 691 |
+
"loss": 0.0847,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.96,
|
| 696 |
+
"learning_rate": 6.328848151379104e-06,
|
| 697 |
+
"loss": 0.0403,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.97,
|
| 702 |
+
"learning_rate": 6.2779551655048475e-06,
|
| 703 |
+
"loss": 0.1037,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.98,
|
| 708 |
+
"learning_rate": 6.22706217963059e-06,
|
| 709 |
+
"loss": 0.1065,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.99,
|
| 714 |
+
"learning_rate": 6.1761691937563324e-06,
|
| 715 |
+
"loss": 0.1039,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.99,
|
| 720 |
+
"learning_rate": 6.125276207882077e-06,
|
| 721 |
+
"loss": 0.0597,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.0,
|
| 726 |
+
"eval_accuracy": 0.989822775925601,
|
| 727 |
+
"eval_f1": 0.0,
|
| 728 |
+
"eval_loss": 0.07756483554840088,
|
| 729 |
+
"eval_precision": 0.0,
|
| 730 |
+
"eval_recall": 0.0,
|
| 731 |
+
"eval_runtime": 233.8941,
|
| 732 |
+
"eval_samples_per_second": 170.56,
|
| 733 |
+
"step": 59839
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.0,
|
| 737 |
+
"learning_rate": 6.074383222007819e-06,
|
| 738 |
+
"loss": 0.0682,
|
| 739 |
+
"step": 60000
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.01,
|
| 743 |
+
"learning_rate": 6.0234902361335624e-06,
|
| 744 |
+
"loss": 0.0781,
|
| 745 |
+
"step": 60500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.02,
|
| 749 |
+
"learning_rate": 5.972597250259305e-06,
|
| 750 |
+
"loss": 0.1027,
|
| 751 |
+
"step": 61000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.03,
|
| 755 |
+
"learning_rate": 5.921704264385048e-06,
|
| 756 |
+
"loss": 0.0871,
|
| 757 |
+
"step": 61500
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.04,
|
| 761 |
+
"learning_rate": 5.870811278510792e-06,
|
| 762 |
+
"loss": 0.0525,
|
| 763 |
+
"step": 62000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.04,
|
| 767 |
+
"learning_rate": 5.819918292636534e-06,
|
| 768 |
+
"loss": 0.0616,
|
| 769 |
+
"step": 62500
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.05,
|
| 773 |
+
"learning_rate": 5.769025306762277e-06,
|
| 774 |
+
"loss": 0.0599,
|
| 775 |
+
"step": 63000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.06,
|
| 779 |
+
"learning_rate": 5.71813232088802e-06,
|
| 780 |
+
"loss": 0.0296,
|
| 781 |
+
"step": 63500
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.07,
|
| 785 |
+
"learning_rate": 5.667239335013763e-06,
|
| 786 |
+
"loss": 0.0515,
|
| 787 |
+
"step": 64000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 1.08,
|
| 791 |
+
"learning_rate": 5.6163463491395066e-06,
|
| 792 |
+
"loss": 0.0436,
|
| 793 |
+
"step": 64500
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.09,
|
| 797 |
+
"learning_rate": 5.565453363265249e-06,
|
| 798 |
+
"loss": 0.0819,
|
| 799 |
+
"step": 65000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.09,
|
| 803 |
+
"learning_rate": 5.5145603773909915e-06,
|
| 804 |
+
"loss": 0.0451,
|
| 805 |
+
"step": 65500
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.1,
|
| 809 |
+
"learning_rate": 5.463667391516736e-06,
|
| 810 |
+
"loss": 0.0485,
|
| 811 |
+
"step": 66000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.11,
|
| 815 |
+
"learning_rate": 5.412774405642478e-06,
|
| 816 |
+
"loss": 0.0714,
|
| 817 |
+
"step": 66500
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 1.12,
|
| 821 |
+
"learning_rate": 5.3618814197682215e-06,
|
| 822 |
+
"loss": 0.0573,
|
| 823 |
+
"step": 67000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.13,
|
| 827 |
+
"learning_rate": 5.310988433893964e-06,
|
| 828 |
+
"loss": 0.0972,
|
| 829 |
+
"step": 67500
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.14,
|
| 833 |
+
"learning_rate": 5.260095448019707e-06,
|
| 834 |
+
"loss": 0.0515,
|
| 835 |
+
"step": 68000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.14,
|
| 839 |
+
"learning_rate": 5.209202462145451e-06,
|
| 840 |
+
"loss": 0.0446,
|
| 841 |
+
"step": 68500
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.15,
|
| 845 |
+
"learning_rate": 5.158309476271193e-06,
|
| 846 |
+
"loss": 0.0427,
|
| 847 |
+
"step": 69000
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.16,
|
| 851 |
+
"learning_rate": 5.107416490396936e-06,
|
| 852 |
+
"loss": 0.0566,
|
| 853 |
+
"step": 69500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.17,
|
| 857 |
+
"learning_rate": 5.056523504522679e-06,
|
| 858 |
+
"loss": 0.0424,
|
| 859 |
+
"step": 70000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 1.18,
|
| 863 |
+
"learning_rate": 5.005630518648422e-06,
|
| 864 |
+
"loss": 0.0746,
|
| 865 |
+
"step": 70500
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.19,
|
| 869 |
+
"learning_rate": 4.954737532774165e-06,
|
| 870 |
+
"loss": 0.0207,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.19,
|
| 875 |
+
"learning_rate": 4.903844546899908e-06,
|
| 876 |
+
"loss": 0.0351,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.2,
|
| 881 |
+
"learning_rate": 4.852951561025651e-06,
|
| 882 |
+
"loss": 0.0521,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.21,
|
| 887 |
+
"learning_rate": 4.802058575151395e-06,
|
| 888 |
+
"loss": 0.0747,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.22,
|
| 893 |
+
"learning_rate": 4.751165589277137e-06,
|
| 894 |
+
"loss": 0.0685,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.23,
|
| 899 |
+
"learning_rate": 4.70027260340288e-06,
|
| 900 |
+
"loss": 0.0461,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.24,
|
| 905 |
+
"learning_rate": 4.649379617528623e-06,
|
| 906 |
+
"loss": 0.0988,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.25,
|
| 911 |
+
"learning_rate": 4.598486631654366e-06,
|
| 912 |
+
"loss": 0.0502,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.25,
|
| 917 |
+
"learning_rate": 4.547593645780109e-06,
|
| 918 |
+
"loss": 0.0651,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.26,
|
| 923 |
+
"learning_rate": 4.496700659905852e-06,
|
| 924 |
+
"loss": 0.0731,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.27,
|
| 929 |
+
"learning_rate": 4.445807674031595e-06,
|
| 930 |
+
"loss": 0.0734,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.28,
|
| 935 |
+
"learning_rate": 4.394914688157338e-06,
|
| 936 |
+
"loss": 0.0377,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.29,
|
| 941 |
+
"learning_rate": 4.344021702283081e-06,
|
| 942 |
+
"loss": 0.0667,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.3,
|
| 947 |
+
"learning_rate": 4.293128716408824e-06,
|
| 948 |
+
"loss": 0.0657,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.3,
|
| 953 |
+
"learning_rate": 4.242235730534567e-06,
|
| 954 |
+
"loss": 0.0747,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.31,
|
| 959 |
+
"learning_rate": 4.1913427446603105e-06,
|
| 960 |
+
"loss": 0.0481,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.32,
|
| 965 |
+
"learning_rate": 4.140449758786053e-06,
|
| 966 |
+
"loss": 0.0489,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.33,
|
| 971 |
+
"learning_rate": 4.089556772911796e-06,
|
| 972 |
+
"loss": 0.029,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.34,
|
| 977 |
+
"learning_rate": 4.038663787037539e-06,
|
| 978 |
+
"loss": 0.0818,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.35,
|
| 983 |
+
"learning_rate": 3.987770801163282e-06,
|
| 984 |
+
"loss": 0.0396,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.35,
|
| 989 |
+
"learning_rate": 3.9368778152890255e-06,
|
| 990 |
+
"loss": 0.0758,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.36,
|
| 995 |
+
"learning_rate": 3.885984829414768e-06,
|
| 996 |
+
"loss": 0.0597,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.37,
|
| 1001 |
+
"learning_rate": 3.8350918435405104e-06,
|
| 1002 |
+
"loss": 0.0827,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.38,
|
| 1007 |
+
"learning_rate": 3.784198857666254e-06,
|
| 1008 |
+
"loss": 0.0465,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.39,
|
| 1013 |
+
"learning_rate": 3.7333058717919967e-06,
|
| 1014 |
+
"loss": 0.1027,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.4,
|
| 1019 |
+
"learning_rate": 3.6824128859177396e-06,
|
| 1020 |
+
"loss": 0.0399,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.4,
|
| 1025 |
+
"learning_rate": 3.631519900043483e-06,
|
| 1026 |
+
"loss": 0.0771,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.41,
|
| 1031 |
+
"learning_rate": 3.580626914169226e-06,
|
| 1032 |
+
"loss": 0.0592,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.42,
|
| 1037 |
+
"learning_rate": 3.529733928294969e-06,
|
| 1038 |
+
"loss": 0.0805,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.43,
|
| 1043 |
+
"learning_rate": 3.478840942420712e-06,
|
| 1044 |
+
"loss": 0.0478,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.44,
|
| 1049 |
+
"learning_rate": 3.4279479565464546e-06,
|
| 1050 |
+
"loss": 0.0633,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.45,
|
| 1055 |
+
"learning_rate": 3.3770549706721983e-06,
|
| 1056 |
+
"loss": 0.088,
|
| 1057 |
+
"step": 86500
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.45,
|
| 1061 |
+
"learning_rate": 3.326161984797941e-06,
|
| 1062 |
+
"loss": 0.0506,
|
| 1063 |
+
"step": 87000
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.46,
|
| 1067 |
+
"learning_rate": 3.2752689989236837e-06,
|
| 1068 |
+
"loss": 0.0905,
|
| 1069 |
+
"step": 87500
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.47,
|
| 1073 |
+
"learning_rate": 3.224376013049427e-06,
|
| 1074 |
+
"loss": 0.0395,
|
| 1075 |
+
"step": 88000
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.48,
|
| 1079 |
+
"learning_rate": 3.17348302717517e-06,
|
| 1080 |
+
"loss": 0.05,
|
| 1081 |
+
"step": 88500
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.49,
|
| 1085 |
+
"learning_rate": 3.1225900413009124e-06,
|
| 1086 |
+
"loss": 0.0812,
|
| 1087 |
+
"step": 89000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.5,
|
| 1091 |
+
"learning_rate": 3.071697055426656e-06,
|
| 1092 |
+
"loss": 0.0501,
|
| 1093 |
+
"step": 89500
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.5,
|
| 1097 |
+
"learning_rate": 3.0208040695523987e-06,
|
| 1098 |
+
"loss": 0.0635,
|
| 1099 |
+
"step": 90000
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.51,
|
| 1103 |
+
"learning_rate": 2.969911083678142e-06,
|
| 1104 |
+
"loss": 0.1061,
|
| 1105 |
+
"step": 90500
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.52,
|
| 1109 |
+
"learning_rate": 2.919018097803885e-06,
|
| 1110 |
+
"loss": 0.0743,
|
| 1111 |
+
"step": 91000
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.53,
|
| 1115 |
+
"learning_rate": 2.868125111929628e-06,
|
| 1116 |
+
"loss": 0.0776,
|
| 1117 |
+
"step": 91500
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.54,
|
| 1121 |
+
"learning_rate": 2.8172321260553707e-06,
|
| 1122 |
+
"loss": 0.0753,
|
| 1123 |
+
"step": 92000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.55,
|
| 1127 |
+
"learning_rate": 2.7663391401811136e-06,
|
| 1128 |
+
"loss": 0.0745,
|
| 1129 |
+
"step": 92500
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.55,
|
| 1133 |
+
"learning_rate": 2.715446154306857e-06,
|
| 1134 |
+
"loss": 0.0534,
|
| 1135 |
+
"step": 93000
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.56,
|
| 1139 |
+
"learning_rate": 2.6645531684326e-06,
|
| 1140 |
+
"loss": 0.0882,
|
| 1141 |
+
"step": 93500
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.57,
|
| 1145 |
+
"learning_rate": 2.6136601825583428e-06,
|
| 1146 |
+
"loss": 0.0864,
|
| 1147 |
+
"step": 94000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.58,
|
| 1151 |
+
"learning_rate": 2.5627671966840857e-06,
|
| 1152 |
+
"loss": 0.0671,
|
| 1153 |
+
"step": 94500
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 1.59,
|
| 1157 |
+
"learning_rate": 2.511874210809829e-06,
|
| 1158 |
+
"loss": 0.0446,
|
| 1159 |
+
"step": 95000
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.6,
|
| 1163 |
+
"learning_rate": 2.4609812249355715e-06,
|
| 1164 |
+
"loss": 0.045,
|
| 1165 |
+
"step": 95500
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.6,
|
| 1169 |
+
"learning_rate": 2.410088239061315e-06,
|
| 1170 |
+
"loss": 0.0581,
|
| 1171 |
+
"step": 96000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.61,
|
| 1175 |
+
"learning_rate": 2.3591952531870577e-06,
|
| 1176 |
+
"loss": 0.047,
|
| 1177 |
+
"step": 96500
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.62,
|
| 1181 |
+
"learning_rate": 2.308302267312801e-06,
|
| 1182 |
+
"loss": 0.0367,
|
| 1183 |
+
"step": 97000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.63,
|
| 1187 |
+
"learning_rate": 2.2574092814385436e-06,
|
| 1188 |
+
"loss": 0.0394,
|
| 1189 |
+
"step": 97500
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.64,
|
| 1193 |
+
"learning_rate": 2.206516295564287e-06,
|
| 1194 |
+
"loss": 0.0833,
|
| 1195 |
+
"step": 98000
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 1.65,
|
| 1199 |
+
"learning_rate": 2.15562330969003e-06,
|
| 1200 |
+
"loss": 0.0566,
|
| 1201 |
+
"step": 98500
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.65,
|
| 1205 |
+
"learning_rate": 2.1047303238157727e-06,
|
| 1206 |
+
"loss": 0.0211,
|
| 1207 |
+
"step": 99000
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.66,
|
| 1211 |
+
"learning_rate": 2.0538373379415156e-06,
|
| 1212 |
+
"loss": 0.0526,
|
| 1213 |
+
"step": 99500
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.67,
|
| 1217 |
+
"learning_rate": 2.002944352067259e-06,
|
| 1218 |
+
"loss": 0.0358,
|
| 1219 |
+
"step": 100000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.68,
|
| 1223 |
+
"learning_rate": 1.952051366193002e-06,
|
| 1224 |
+
"loss": 0.102,
|
| 1225 |
+
"step": 100500
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.69,
|
| 1229 |
+
"learning_rate": 1.9011583803187445e-06,
|
| 1230 |
+
"loss": 0.0934,
|
| 1231 |
+
"step": 101000
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.7,
|
| 1235 |
+
"learning_rate": 1.8502653944444877e-06,
|
| 1236 |
+
"loss": 0.0299,
|
| 1237 |
+
"step": 101500
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.7,
|
| 1241 |
+
"learning_rate": 1.7993724085702308e-06,
|
| 1242 |
+
"loss": 0.0945,
|
| 1243 |
+
"step": 102000
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.71,
|
| 1247 |
+
"learning_rate": 1.748479422695974e-06,
|
| 1248 |
+
"loss": 0.1056,
|
| 1249 |
+
"step": 102500
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.72,
|
| 1253 |
+
"learning_rate": 1.6975864368217166e-06,
|
| 1254 |
+
"loss": 0.0283,
|
| 1255 |
+
"step": 103000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.73,
|
| 1259 |
+
"learning_rate": 1.6466934509474597e-06,
|
| 1260 |
+
"loss": 0.0572,
|
| 1261 |
+
"step": 103500
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.74,
|
| 1265 |
+
"learning_rate": 1.5958004650732028e-06,
|
| 1266 |
+
"loss": 0.0695,
|
| 1267 |
+
"step": 104000
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.75,
|
| 1271 |
+
"learning_rate": 1.5449074791989455e-06,
|
| 1272 |
+
"loss": 0.0397,
|
| 1273 |
+
"step": 104500
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.75,
|
| 1277 |
+
"learning_rate": 1.4940144933246887e-06,
|
| 1278 |
+
"loss": 0.0654,
|
| 1279 |
+
"step": 105000
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 1.76,
|
| 1283 |
+
"learning_rate": 1.4431215074504318e-06,
|
| 1284 |
+
"loss": 0.0487,
|
| 1285 |
+
"step": 105500
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.77,
|
| 1289 |
+
"learning_rate": 1.3922285215761747e-06,
|
| 1290 |
+
"loss": 0.0425,
|
| 1291 |
+
"step": 106000
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.78,
|
| 1295 |
+
"learning_rate": 1.3413355357019178e-06,
|
| 1296 |
+
"loss": 0.0469,
|
| 1297 |
+
"step": 106500
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.79,
|
| 1301 |
+
"learning_rate": 1.2904425498276607e-06,
|
| 1302 |
+
"loss": 0.0713,
|
| 1303 |
+
"step": 107000
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.8,
|
| 1307 |
+
"learning_rate": 1.2395495639534036e-06,
|
| 1308 |
+
"loss": 0.0615,
|
| 1309 |
+
"step": 107500
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.8,
|
| 1313 |
+
"learning_rate": 1.1886565780791467e-06,
|
| 1314 |
+
"loss": 0.0501,
|
| 1315 |
+
"step": 108000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.81,
|
| 1319 |
+
"learning_rate": 1.1377635922048896e-06,
|
| 1320 |
+
"loss": 0.0548,
|
| 1321 |
+
"step": 108500
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.82,
|
| 1325 |
+
"learning_rate": 1.0868706063306326e-06,
|
| 1326 |
+
"loss": 0.0402,
|
| 1327 |
+
"step": 109000
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.83,
|
| 1331 |
+
"learning_rate": 1.0359776204563757e-06,
|
| 1332 |
+
"loss": 0.0453,
|
| 1333 |
+
"step": 109500
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.84,
|
| 1337 |
+
"learning_rate": 9.850846345821186e-07,
|
| 1338 |
+
"loss": 0.0665,
|
| 1339 |
+
"step": 110000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.85,
|
| 1343 |
+
"learning_rate": 9.341916487078616e-07,
|
| 1344 |
+
"loss": 0.1019,
|
| 1345 |
+
"step": 110500
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.85,
|
| 1349 |
+
"learning_rate": 8.832986628336047e-07,
|
| 1350 |
+
"loss": 0.0443,
|
| 1351 |
+
"step": 111000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.86,
|
| 1355 |
+
"learning_rate": 8.324056769593476e-07,
|
| 1356 |
+
"loss": 0.0584,
|
| 1357 |
+
"step": 111500
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.87,
|
| 1361 |
+
"learning_rate": 7.815126910850906e-07,
|
| 1362 |
+
"loss": 0.0551,
|
| 1363 |
+
"step": 112000
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 1.88,
|
| 1367 |
+
"learning_rate": 7.306197052108335e-07,
|
| 1368 |
+
"loss": 0.0371,
|
| 1369 |
+
"step": 112500
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.89,
|
| 1373 |
+
"learning_rate": 6.797267193365766e-07,
|
| 1374 |
+
"loss": 0.0901,
|
| 1375 |
+
"step": 113000
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.9,
|
| 1379 |
+
"learning_rate": 6.288337334623196e-07,
|
| 1380 |
+
"loss": 0.0482,
|
| 1381 |
+
"step": 113500
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.91,
|
| 1385 |
+
"learning_rate": 5.779407475880626e-07,
|
| 1386 |
+
"loss": 0.0821,
|
| 1387 |
+
"step": 114000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.91,
|
| 1391 |
+
"learning_rate": 5.270477617138056e-07,
|
| 1392 |
+
"loss": 0.0482,
|
| 1393 |
+
"step": 114500
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.92,
|
| 1397 |
+
"learning_rate": 4.761547758395485e-07,
|
| 1398 |
+
"loss": 0.0686,
|
| 1399 |
+
"step": 115000
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.93,
|
| 1403 |
+
"learning_rate": 4.252617899652915e-07,
|
| 1404 |
+
"loss": 0.0765,
|
| 1405 |
+
"step": 115500
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.94,
|
| 1409 |
+
"learning_rate": 3.7436880409103453e-07,
|
| 1410 |
+
"loss": 0.073,
|
| 1411 |
+
"step": 116000
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 1.95,
|
| 1415 |
+
"learning_rate": 3.234758182167775e-07,
|
| 1416 |
+
"loss": 0.0592,
|
| 1417 |
+
"step": 116500
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.96,
|
| 1421 |
+
"learning_rate": 2.725828323425205e-07,
|
| 1422 |
+
"loss": 0.076,
|
| 1423 |
+
"step": 117000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.96,
|
| 1427 |
+
"learning_rate": 2.2168984646826352e-07,
|
| 1428 |
+
"loss": 0.0532,
|
| 1429 |
+
"step": 117500
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.97,
|
| 1433 |
+
"learning_rate": 1.707968605940065e-07,
|
| 1434 |
+
"loss": 0.053,
|
| 1435 |
+
"step": 118000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.98,
|
| 1439 |
+
"learning_rate": 1.199038747197495e-07,
|
| 1440 |
+
"loss": 0.0708,
|
| 1441 |
+
"step": 118500
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.99,
|
| 1445 |
+
"learning_rate": 6.90108888454925e-08,
|
| 1446 |
+
"loss": 0.083,
|
| 1447 |
+
"step": 119000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 2.0,
|
| 1451 |
+
"learning_rate": 1.8117902971235495e-08,
|
| 1452 |
+
"loss": 0.0802,
|
| 1453 |
+
"step": 119500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 2.0,
|
| 1457 |
+
"eval_accuracy": 0.989822775925601,
|
| 1458 |
+
"eval_f1": 0.0,
|
| 1459 |
+
"eval_loss": 0.061012040823698044,
|
| 1460 |
+
"eval_precision": 0.0,
|
| 1461 |
+
"eval_recall": 0.0,
|
| 1462 |
+
"eval_runtime": 205.9172,
|
| 1463 |
+
"eval_samples_per_second": 193.733,
|
| 1464 |
+
"step": 119678
|
| 1465 |
+
}
|
| 1466 |
+
],
|
| 1467 |
+
"max_steps": 119678,
|
| 1468 |
+
"num_train_epochs": 2,
|
| 1469 |
+
"total_flos": 2.160546753919596e+16,
|
| 1470 |
+
"trial_name": null,
|
| 1471 |
+
"trial_params": null
|
| 1472 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:573bf0960153c1d466ef817e422efb17d7f303a63fd8dccac4dc6568afb108f2
|
| 3 |
+
size 2351
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|