Upload 10 files
Browse files- added_tokens.json +3 -0
- config.json +48 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +15 -0
- spm.model +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +59 -0
- train_results.json +8 -0
- trainer_state.json +2113 -0
- training_args.bin +3 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "models/microsoft/deberta-v3-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DebertaV2ForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"finetuning_task": "ner",
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 1024,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "LABEL_0",
|
| 13 |
+
"1": "LABEL_1",
|
| 14 |
+
"2": "LABEL_2",
|
| 15 |
+
"3": "LABEL_3"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 4096,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"LABEL_0": 0,
|
| 21 |
+
"LABEL_1": 1,
|
| 22 |
+
"LABEL_2": 2,
|
| 23 |
+
"LABEL_3": 3
|
| 24 |
+
},
|
| 25 |
+
"layer_norm_eps": 1e-07,
|
| 26 |
+
"max_position_embeddings": 512,
|
| 27 |
+
"max_relative_positions": -1,
|
| 28 |
+
"model_type": "deberta-v2",
|
| 29 |
+
"norm_rel_ebd": "layer_norm",
|
| 30 |
+
"num_attention_heads": 16,
|
| 31 |
+
"num_hidden_layers": 24,
|
| 32 |
+
"pad_token_id": 0,
|
| 33 |
+
"pooler_dropout": 0,
|
| 34 |
+
"pooler_hidden_act": "gelu",
|
| 35 |
+
"pooler_hidden_size": 1024,
|
| 36 |
+
"pos_att_type": [
|
| 37 |
+
"p2c",
|
| 38 |
+
"c2p"
|
| 39 |
+
],
|
| 40 |
+
"position_biased_input": false,
|
| 41 |
+
"position_buckets": 256,
|
| 42 |
+
"relative_attention": true,
|
| 43 |
+
"share_att_key": true,
|
| 44 |
+
"torch_dtype": "float32",
|
| 45 |
+
"transformers_version": "4.34.1",
|
| 46 |
+
"type_vocab_size": 0,
|
| 47 |
+
"vocab_size": 128100
|
| 48 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54897e828057e3c35e8b36345df614a6b12f61565c6464a940287a0d0898a618
|
| 3 |
+
size 1736201005
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "[PAD]",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "[CLS]",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "[SEP]",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "[UNK]",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"128000": {
|
| 37 |
+
"content": "[MASK]",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "[CLS]",
|
| 46 |
+
"clean_up_tokenization_spaces": true,
|
| 47 |
+
"cls_token": "[CLS]",
|
| 48 |
+
"do_lower_case": false,
|
| 49 |
+
"eos_token": "[SEP]",
|
| 50 |
+
"mask_token": "[MASK]",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"pad_token": "[PAD]",
|
| 53 |
+
"sep_token": "[SEP]",
|
| 54 |
+
"sp_model_kwargs": {},
|
| 55 |
+
"split_by_punct": false,
|
| 56 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 57 |
+
"unk_token": "[UNK]",
|
| 58 |
+
"vocab_type": "spm"
|
| 59 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"train_loss": 0.019633541173813333,
|
| 4 |
+
"train_runtime": 2021.1766,
|
| 5 |
+
"train_samples": 67239,
|
| 6 |
+
"train_samples_per_second": 199.604,
|
| 7 |
+
"train_steps_per_second": 1.561
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9965715706847758,
|
| 3 |
+
"best_model_checkpoint": "output_dev/v48-deberta-v3-large-f1-warmup-greater-patience-subword-len-500-maxlen-512-6epoch/checkpoint-1800",
|
| 4 |
+
"epoch": 6.0,
|
| 5 |
+
"eval_steps": 200,
|
| 6 |
+
"global_step": 3156,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 3.1645569620253167e-06,
|
| 14 |
+
"loss": 1.8539,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 6.329113924050633e-06,
|
| 20 |
+
"loss": 1.0295,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.06,
|
| 25 |
+
"learning_rate": 9.49367088607595e-06,
|
| 26 |
+
"loss": 0.4899,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.08,
|
| 31 |
+
"learning_rate": 1.2658227848101267e-05,
|
| 32 |
+
"loss": 0.2892,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.1,
|
| 37 |
+
"learning_rate": 1.5822784810126583e-05,
|
| 38 |
+
"loss": 0.1988,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.11,
|
| 43 |
+
"learning_rate": 1.89873417721519e-05,
|
| 44 |
+
"loss": 0.1417,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 2.2151898734177217e-05,
|
| 50 |
+
"loss": 0.0904,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.15,
|
| 55 |
+
"learning_rate": 2.5316455696202533e-05,
|
| 56 |
+
"loss": 0.0814,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.17,
|
| 61 |
+
"learning_rate": 2.848101265822785e-05,
|
| 62 |
+
"loss": 0.0335,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.19,
|
| 67 |
+
"learning_rate": 3.1645569620253167e-05,
|
| 68 |
+
"loss": 0.0401,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.21,
|
| 73 |
+
"learning_rate": 3.4810126582278487e-05,
|
| 74 |
+
"loss": 0.0304,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.23,
|
| 79 |
+
"learning_rate": 3.79746835443038e-05,
|
| 80 |
+
"loss": 0.0223,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.25,
|
| 85 |
+
"learning_rate": 4.113924050632912e-05,
|
| 86 |
+
"loss": 0.029,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.27,
|
| 91 |
+
"learning_rate": 4.430379746835443e-05,
|
| 92 |
+
"loss": 0.0312,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.29,
|
| 97 |
+
"learning_rate": 4.7468354430379746e-05,
|
| 98 |
+
"loss": 0.0295,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.3,
|
| 103 |
+
"learning_rate": 4.996664442961975e-05,
|
| 104 |
+
"loss": 0.0315,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.32,
|
| 109 |
+
"learning_rate": 4.979986657771848e-05,
|
| 110 |
+
"loss": 0.027,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.34,
|
| 115 |
+
"learning_rate": 4.963308872581721e-05,
|
| 116 |
+
"loss": 0.021,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.36,
|
| 121 |
+
"learning_rate": 4.946631087391594e-05,
|
| 122 |
+
"loss": 0.0258,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.38,
|
| 127 |
+
"learning_rate": 4.929953302201468e-05,
|
| 128 |
+
"loss": 0.0232,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.38,
|
| 133 |
+
"eval_f1": 0.9829244574920268,
|
| 134 |
+
"eval_f1_0": 0.9946946557701127,
|
| 135 |
+
"eval_f1_1": 0.99571802949886,
|
| 136 |
+
"eval_f1_2": 0.9426792739319484,
|
| 137 |
+
"eval_f1_3": 0.9986058707671863,
|
| 138 |
+
"eval_loss": 0.02559051290154457,
|
| 139 |
+
"eval_runtime": 19.2055,
|
| 140 |
+
"eval_samples_per_second": 389.055,
|
| 141 |
+
"eval_steps_per_second": 0.417,
|
| 142 |
+
"step": 200
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.4,
|
| 146 |
+
"learning_rate": 4.913275517011341e-05,
|
| 147 |
+
"loss": 0.023,
|
| 148 |
+
"step": 210
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 0.42,
|
| 152 |
+
"learning_rate": 4.896597731821214e-05,
|
| 153 |
+
"loss": 0.0145,
|
| 154 |
+
"step": 220
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.44,
|
| 158 |
+
"learning_rate": 4.879919946631087e-05,
|
| 159 |
+
"loss": 0.0319,
|
| 160 |
+
"step": 230
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.46,
|
| 164 |
+
"learning_rate": 4.863242161440961e-05,
|
| 165 |
+
"loss": 0.02,
|
| 166 |
+
"step": 240
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 0.48,
|
| 170 |
+
"learning_rate": 4.846564376250834e-05,
|
| 171 |
+
"loss": 0.0204,
|
| 172 |
+
"step": 250
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.49,
|
| 176 |
+
"learning_rate": 4.829886591060707e-05,
|
| 177 |
+
"loss": 0.0217,
|
| 178 |
+
"step": 260
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.51,
|
| 182 |
+
"learning_rate": 4.813208805870581e-05,
|
| 183 |
+
"loss": 0.0239,
|
| 184 |
+
"step": 270
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.53,
|
| 188 |
+
"learning_rate": 4.796531020680454e-05,
|
| 189 |
+
"loss": 0.0289,
|
| 190 |
+
"step": 280
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.55,
|
| 194 |
+
"learning_rate": 4.779853235490327e-05,
|
| 195 |
+
"loss": 0.0298,
|
| 196 |
+
"step": 290
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 0.57,
|
| 200 |
+
"learning_rate": 4.7631754503002e-05,
|
| 201 |
+
"loss": 0.0198,
|
| 202 |
+
"step": 300
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.59,
|
| 206 |
+
"learning_rate": 4.7464976651100736e-05,
|
| 207 |
+
"loss": 0.0313,
|
| 208 |
+
"step": 310
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.61,
|
| 212 |
+
"learning_rate": 4.729819879919947e-05,
|
| 213 |
+
"loss": 0.0187,
|
| 214 |
+
"step": 320
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.63,
|
| 218 |
+
"learning_rate": 4.71314209472982e-05,
|
| 219 |
+
"loss": 0.0264,
|
| 220 |
+
"step": 330
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.65,
|
| 224 |
+
"learning_rate": 4.6964643095396935e-05,
|
| 225 |
+
"loss": 0.0327,
|
| 226 |
+
"step": 340
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.67,
|
| 230 |
+
"learning_rate": 4.6797865243495666e-05,
|
| 231 |
+
"loss": 0.0181,
|
| 232 |
+
"step": 350
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 0.68,
|
| 236 |
+
"learning_rate": 4.6631087391594396e-05,
|
| 237 |
+
"loss": 0.0162,
|
| 238 |
+
"step": 360
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.7,
|
| 242 |
+
"learning_rate": 4.646430953969313e-05,
|
| 243 |
+
"loss": 0.0178,
|
| 244 |
+
"step": 370
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 0.72,
|
| 248 |
+
"learning_rate": 4.6297531687791864e-05,
|
| 249 |
+
"loss": 0.0068,
|
| 250 |
+
"step": 380
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.74,
|
| 254 |
+
"learning_rate": 4.6130753835890595e-05,
|
| 255 |
+
"loss": 0.0202,
|
| 256 |
+
"step": 390
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 0.76,
|
| 260 |
+
"learning_rate": 4.5963975983989326e-05,
|
| 261 |
+
"loss": 0.0261,
|
| 262 |
+
"step": 400
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.76,
|
| 266 |
+
"eval_f1": 0.9902730835281612,
|
| 267 |
+
"eval_f1_0": 0.9970295053857255,
|
| 268 |
+
"eval_f1_1": 0.9966460102285332,
|
| 269 |
+
"eval_f1_2": 0.968532366748815,
|
| 270 |
+
"eval_f1_3": 0.9988844517495713,
|
| 271 |
+
"eval_loss": 0.01530930120497942,
|
| 272 |
+
"eval_runtime": 19.2644,
|
| 273 |
+
"eval_samples_per_second": 387.867,
|
| 274 |
+
"eval_steps_per_second": 0.415,
|
| 275 |
+
"step": 400
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.78,
|
| 279 |
+
"learning_rate": 4.5797198132088056e-05,
|
| 280 |
+
"loss": 0.0196,
|
| 281 |
+
"step": 410
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.8,
|
| 285 |
+
"learning_rate": 4.5630420280186794e-05,
|
| 286 |
+
"loss": 0.0197,
|
| 287 |
+
"step": 420
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.82,
|
| 291 |
+
"learning_rate": 4.5463642428285524e-05,
|
| 292 |
+
"loss": 0.0213,
|
| 293 |
+
"step": 430
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 0.84,
|
| 297 |
+
"learning_rate": 4.5296864576384255e-05,
|
| 298 |
+
"loss": 0.0192,
|
| 299 |
+
"step": 440
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.86,
|
| 303 |
+
"learning_rate": 4.513008672448299e-05,
|
| 304 |
+
"loss": 0.0078,
|
| 305 |
+
"step": 450
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.87,
|
| 309 |
+
"learning_rate": 4.496330887258172e-05,
|
| 310 |
+
"loss": 0.0219,
|
| 311 |
+
"step": 460
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.89,
|
| 315 |
+
"learning_rate": 4.4796531020680454e-05,
|
| 316 |
+
"loss": 0.0076,
|
| 317 |
+
"step": 470
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.91,
|
| 321 |
+
"learning_rate": 4.4629753168779184e-05,
|
| 322 |
+
"loss": 0.0145,
|
| 323 |
+
"step": 480
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.93,
|
| 327 |
+
"learning_rate": 4.446297531687792e-05,
|
| 328 |
+
"loss": 0.0121,
|
| 329 |
+
"step": 490
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.95,
|
| 333 |
+
"learning_rate": 4.429619746497665e-05,
|
| 334 |
+
"loss": 0.0146,
|
| 335 |
+
"step": 500
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.97,
|
| 339 |
+
"learning_rate": 4.412941961307538e-05,
|
| 340 |
+
"loss": 0.0126,
|
| 341 |
+
"step": 510
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.99,
|
| 345 |
+
"learning_rate": 4.396264176117412e-05,
|
| 346 |
+
"loss": 0.0142,
|
| 347 |
+
"step": 520
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.01,
|
| 351 |
+
"learning_rate": 4.379586390927285e-05,
|
| 352 |
+
"loss": 0.0112,
|
| 353 |
+
"step": 530
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.03,
|
| 357 |
+
"learning_rate": 4.362908605737158e-05,
|
| 358 |
+
"loss": 0.0066,
|
| 359 |
+
"step": 540
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.05,
|
| 363 |
+
"learning_rate": 4.346230820547031e-05,
|
| 364 |
+
"loss": 0.0116,
|
| 365 |
+
"step": 550
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.06,
|
| 369 |
+
"learning_rate": 4.329553035356905e-05,
|
| 370 |
+
"loss": 0.0121,
|
| 371 |
+
"step": 560
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.08,
|
| 375 |
+
"learning_rate": 4.312875250166778e-05,
|
| 376 |
+
"loss": 0.0083,
|
| 377 |
+
"step": 570
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.1,
|
| 381 |
+
"learning_rate": 4.296197464976651e-05,
|
| 382 |
+
"loss": 0.0085,
|
| 383 |
+
"step": 580
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.12,
|
| 387 |
+
"learning_rate": 4.279519679786524e-05,
|
| 388 |
+
"loss": 0.0146,
|
| 389 |
+
"step": 590
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.14,
|
| 393 |
+
"learning_rate": 4.262841894596398e-05,
|
| 394 |
+
"loss": 0.0105,
|
| 395 |
+
"step": 600
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.14,
|
| 399 |
+
"eval_f1": 0.9896097338429783,
|
| 400 |
+
"eval_f1_0": 0.9968138351713054,
|
| 401 |
+
"eval_f1_1": 0.9974716910591519,
|
| 402 |
+
"eval_f1_2": 0.9651378712455954,
|
| 403 |
+
"eval_f1_3": 0.9990155378958605,
|
| 404 |
+
"eval_loss": 0.015092647634446621,
|
| 405 |
+
"eval_runtime": 19.3157,
|
| 406 |
+
"eval_samples_per_second": 386.835,
|
| 407 |
+
"eval_steps_per_second": 0.414,
|
| 408 |
+
"step": 600
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.16,
|
| 412 |
+
"learning_rate": 4.246164109406271e-05,
|
| 413 |
+
"loss": 0.0126,
|
| 414 |
+
"step": 610
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 1.18,
|
| 418 |
+
"learning_rate": 4.229486324216144e-05,
|
| 419 |
+
"loss": 0.0119,
|
| 420 |
+
"step": 620
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.2,
|
| 424 |
+
"learning_rate": 4.212808539026018e-05,
|
| 425 |
+
"loss": 0.0211,
|
| 426 |
+
"step": 630
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.22,
|
| 430 |
+
"learning_rate": 4.196130753835891e-05,
|
| 431 |
+
"loss": 0.0142,
|
| 432 |
+
"step": 640
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.24,
|
| 436 |
+
"learning_rate": 4.179452968645764e-05,
|
| 437 |
+
"loss": 0.0096,
|
| 438 |
+
"step": 650
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 1.25,
|
| 442 |
+
"learning_rate": 4.162775183455637e-05,
|
| 443 |
+
"loss": 0.0084,
|
| 444 |
+
"step": 660
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.27,
|
| 448 |
+
"learning_rate": 4.146097398265511e-05,
|
| 449 |
+
"loss": 0.0044,
|
| 450 |
+
"step": 670
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.29,
|
| 454 |
+
"learning_rate": 4.129419613075384e-05,
|
| 455 |
+
"loss": 0.0059,
|
| 456 |
+
"step": 680
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 1.31,
|
| 460 |
+
"learning_rate": 4.112741827885257e-05,
|
| 461 |
+
"loss": 0.0042,
|
| 462 |
+
"step": 690
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 1.33,
|
| 466 |
+
"learning_rate": 4.09606404269513e-05,
|
| 467 |
+
"loss": 0.0117,
|
| 468 |
+
"step": 700
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 1.35,
|
| 472 |
+
"learning_rate": 4.0793862575050036e-05,
|
| 473 |
+
"loss": 0.0074,
|
| 474 |
+
"step": 710
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.37,
|
| 478 |
+
"learning_rate": 4.062708472314877e-05,
|
| 479 |
+
"loss": 0.0087,
|
| 480 |
+
"step": 720
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.39,
|
| 484 |
+
"learning_rate": 4.04603068712475e-05,
|
| 485 |
+
"loss": 0.0094,
|
| 486 |
+
"step": 730
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.41,
|
| 490 |
+
"learning_rate": 4.0293529019346235e-05,
|
| 491 |
+
"loss": 0.0119,
|
| 492 |
+
"step": 740
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.43,
|
| 496 |
+
"learning_rate": 4.0126751167444966e-05,
|
| 497 |
+
"loss": 0.0138,
|
| 498 |
+
"step": 750
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 1.44,
|
| 502 |
+
"learning_rate": 3.9959973315543696e-05,
|
| 503 |
+
"loss": 0.0073,
|
| 504 |
+
"step": 760
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 1.46,
|
| 508 |
+
"learning_rate": 3.979319546364243e-05,
|
| 509 |
+
"loss": 0.0096,
|
| 510 |
+
"step": 770
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.48,
|
| 514 |
+
"learning_rate": 3.9626417611741164e-05,
|
| 515 |
+
"loss": 0.0119,
|
| 516 |
+
"step": 780
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 1.5,
|
| 520 |
+
"learning_rate": 3.9459639759839895e-05,
|
| 521 |
+
"loss": 0.0045,
|
| 522 |
+
"step": 790
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.52,
|
| 526 |
+
"learning_rate": 3.9292861907938625e-05,
|
| 527 |
+
"loss": 0.0102,
|
| 528 |
+
"step": 800
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.52,
|
| 532 |
+
"eval_f1": 0.9950750308913552,
|
| 533 |
+
"eval_f1_0": 0.9985038101075736,
|
| 534 |
+
"eval_f1_1": 0.9979478535606442,
|
| 535 |
+
"eval_f1_2": 0.9849011258106944,
|
| 536 |
+
"eval_f1_3": 0.9989473340865084,
|
| 537 |
+
"eval_loss": 0.009680411778390408,
|
| 538 |
+
"eval_runtime": 19.3566,
|
| 539 |
+
"eval_samples_per_second": 386.019,
|
| 540 |
+
"eval_steps_per_second": 0.413,
|
| 541 |
+
"step": 800
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.54,
|
| 545 |
+
"learning_rate": 3.912608405603736e-05,
|
| 546 |
+
"loss": 0.0046,
|
| 547 |
+
"step": 810
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.56,
|
| 551 |
+
"learning_rate": 3.8959306204136094e-05,
|
| 552 |
+
"loss": 0.0095,
|
| 553 |
+
"step": 820
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.58,
|
| 557 |
+
"learning_rate": 3.8792528352234824e-05,
|
| 558 |
+
"loss": 0.0092,
|
| 559 |
+
"step": 830
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.6,
|
| 563 |
+
"learning_rate": 3.8625750500333555e-05,
|
| 564 |
+
"loss": 0.0125,
|
| 565 |
+
"step": 840
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.62,
|
| 569 |
+
"learning_rate": 3.845897264843229e-05,
|
| 570 |
+
"loss": 0.0182,
|
| 571 |
+
"step": 850
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.63,
|
| 575 |
+
"learning_rate": 3.829219479653102e-05,
|
| 576 |
+
"loss": 0.0189,
|
| 577 |
+
"step": 860
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.65,
|
| 581 |
+
"learning_rate": 3.8125416944629753e-05,
|
| 582 |
+
"loss": 0.0151,
|
| 583 |
+
"step": 870
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.67,
|
| 587 |
+
"learning_rate": 3.7958639092728484e-05,
|
| 588 |
+
"loss": 0.0097,
|
| 589 |
+
"step": 880
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.69,
|
| 593 |
+
"learning_rate": 3.779186124082722e-05,
|
| 594 |
+
"loss": 0.0047,
|
| 595 |
+
"step": 890
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.71,
|
| 599 |
+
"learning_rate": 3.762508338892595e-05,
|
| 600 |
+
"loss": 0.0068,
|
| 601 |
+
"step": 900
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.73,
|
| 605 |
+
"learning_rate": 3.745830553702468e-05,
|
| 606 |
+
"loss": 0.0165,
|
| 607 |
+
"step": 910
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 1.75,
|
| 611 |
+
"learning_rate": 3.729152768512342e-05,
|
| 612 |
+
"loss": 0.011,
|
| 613 |
+
"step": 920
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.77,
|
| 617 |
+
"learning_rate": 3.712474983322215e-05,
|
| 618 |
+
"loss": 0.0128,
|
| 619 |
+
"step": 930
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.79,
|
| 623 |
+
"learning_rate": 3.695797198132088e-05,
|
| 624 |
+
"loss": 0.0131,
|
| 625 |
+
"step": 940
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.81,
|
| 629 |
+
"learning_rate": 3.679119412941961e-05,
|
| 630 |
+
"loss": 0.0061,
|
| 631 |
+
"step": 950
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.83,
|
| 635 |
+
"learning_rate": 3.662441627751835e-05,
|
| 636 |
+
"loss": 0.006,
|
| 637 |
+
"step": 960
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.84,
|
| 641 |
+
"learning_rate": 3.645763842561708e-05,
|
| 642 |
+
"loss": 0.0053,
|
| 643 |
+
"step": 970
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.86,
|
| 647 |
+
"learning_rate": 3.629086057371581e-05,
|
| 648 |
+
"loss": 0.0062,
|
| 649 |
+
"step": 980
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.88,
|
| 653 |
+
"learning_rate": 3.612408272181455e-05,
|
| 654 |
+
"loss": 0.0039,
|
| 655 |
+
"step": 990
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.9,
|
| 659 |
+
"learning_rate": 3.595730486991328e-05,
|
| 660 |
+
"loss": 0.0123,
|
| 661 |
+
"step": 1000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.9,
|
| 665 |
+
"eval_f1": 0.9952965639551941,
|
| 666 |
+
"eval_f1_0": 0.9985619538388097,
|
| 667 |
+
"eval_f1_1": 0.9977381384652626,
|
| 668 |
+
"eval_f1_2": 0.9859661604309984,
|
| 669 |
+
"eval_f1_3": 0.9989200030857055,
|
| 670 |
+
"eval_loss": 0.008907313458621502,
|
| 671 |
+
"eval_runtime": 19.3729,
|
| 672 |
+
"eval_samples_per_second": 385.693,
|
| 673 |
+
"eval_steps_per_second": 0.413,
|
| 674 |
+
"step": 1000
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.92,
|
| 678 |
+
"learning_rate": 3.579052701801201e-05,
|
| 679 |
+
"loss": 0.0081,
|
| 680 |
+
"step": 1010
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.94,
|
| 684 |
+
"learning_rate": 3.562374916611074e-05,
|
| 685 |
+
"loss": 0.0037,
|
| 686 |
+
"step": 1020
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.96,
|
| 690 |
+
"learning_rate": 3.545697131420948e-05,
|
| 691 |
+
"loss": 0.007,
|
| 692 |
+
"step": 1030
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.98,
|
| 696 |
+
"learning_rate": 3.529019346230821e-05,
|
| 697 |
+
"loss": 0.0043,
|
| 698 |
+
"step": 1040
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.0,
|
| 702 |
+
"learning_rate": 3.512341561040694e-05,
|
| 703 |
+
"loss": 0.0068,
|
| 704 |
+
"step": 1050
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.02,
|
| 708 |
+
"learning_rate": 3.495663775850567e-05,
|
| 709 |
+
"loss": 0.0084,
|
| 710 |
+
"step": 1060
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.03,
|
| 714 |
+
"learning_rate": 3.478985990660441e-05,
|
| 715 |
+
"loss": 0.0023,
|
| 716 |
+
"step": 1070
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.05,
|
| 720 |
+
"learning_rate": 3.462308205470314e-05,
|
| 721 |
+
"loss": 0.0052,
|
| 722 |
+
"step": 1080
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.07,
|
| 726 |
+
"learning_rate": 3.445630420280187e-05,
|
| 727 |
+
"loss": 0.0037,
|
| 728 |
+
"step": 1090
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.09,
|
| 732 |
+
"learning_rate": 3.4289526350900605e-05,
|
| 733 |
+
"loss": 0.0052,
|
| 734 |
+
"step": 1100
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.11,
|
| 738 |
+
"learning_rate": 3.4122748498999336e-05,
|
| 739 |
+
"loss": 0.0074,
|
| 740 |
+
"step": 1110
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.13,
|
| 744 |
+
"learning_rate": 3.395597064709807e-05,
|
| 745 |
+
"loss": 0.0027,
|
| 746 |
+
"step": 1120
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.15,
|
| 750 |
+
"learning_rate": 3.37891927951968e-05,
|
| 751 |
+
"loss": 0.0029,
|
| 752 |
+
"step": 1130
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.17,
|
| 756 |
+
"learning_rate": 3.3622414943295535e-05,
|
| 757 |
+
"loss": 0.0035,
|
| 758 |
+
"step": 1140
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.19,
|
| 762 |
+
"learning_rate": 3.3455637091394265e-05,
|
| 763 |
+
"loss": 0.0037,
|
| 764 |
+
"step": 1150
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 2.21,
|
| 768 |
+
"learning_rate": 3.3288859239492996e-05,
|
| 769 |
+
"loss": 0.0031,
|
| 770 |
+
"step": 1160
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 2.22,
|
| 774 |
+
"learning_rate": 3.3122081387591733e-05,
|
| 775 |
+
"loss": 0.0056,
|
| 776 |
+
"step": 1170
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.24,
|
| 780 |
+
"learning_rate": 3.2955303535690464e-05,
|
| 781 |
+
"loss": 0.0062,
|
| 782 |
+
"step": 1180
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.26,
|
| 786 |
+
"learning_rate": 3.2788525683789195e-05,
|
| 787 |
+
"loss": 0.0025,
|
| 788 |
+
"step": 1190
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.28,
|
| 792 |
+
"learning_rate": 3.2621747831887925e-05,
|
| 793 |
+
"loss": 0.0071,
|
| 794 |
+
"step": 1200
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 2.28,
|
| 798 |
+
"eval_f1": 0.9963329419211635,
|
| 799 |
+
"eval_f1_0": 0.9988693819051201,
|
| 800 |
+
"eval_f1_1": 0.9980296050889892,
|
| 801 |
+
"eval_f1_2": 0.9892069880179274,
|
| 802 |
+
"eval_f1_3": 0.9992257926726171,
|
| 803 |
+
"eval_loss": 0.007162621710449457,
|
| 804 |
+
"eval_runtime": 19.3722,
|
| 805 |
+
"eval_samples_per_second": 385.707,
|
| 806 |
+
"eval_steps_per_second": 0.413,
|
| 807 |
+
"step": 1200
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 2.3,
|
| 811 |
+
"learning_rate": 3.245496997998666e-05,
|
| 812 |
+
"loss": 0.0083,
|
| 813 |
+
"step": 1210
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 2.32,
|
| 817 |
+
"learning_rate": 3.228819212808539e-05,
|
| 818 |
+
"loss": 0.0042,
|
| 819 |
+
"step": 1220
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 2.34,
|
| 823 |
+
"learning_rate": 3.2121414276184124e-05,
|
| 824 |
+
"loss": 0.0073,
|
| 825 |
+
"step": 1230
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.36,
|
| 829 |
+
"learning_rate": 3.1954636424282855e-05,
|
| 830 |
+
"loss": 0.0037,
|
| 831 |
+
"step": 1240
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 2.38,
|
| 835 |
+
"learning_rate": 3.178785857238159e-05,
|
| 836 |
+
"loss": 0.0029,
|
| 837 |
+
"step": 1250
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.4,
|
| 841 |
+
"learning_rate": 3.162108072048032e-05,
|
| 842 |
+
"loss": 0.0031,
|
| 843 |
+
"step": 1260
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 2.41,
|
| 847 |
+
"learning_rate": 3.145430286857905e-05,
|
| 848 |
+
"loss": 0.0018,
|
| 849 |
+
"step": 1270
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 2.43,
|
| 853 |
+
"learning_rate": 3.128752501667779e-05,
|
| 854 |
+
"loss": 0.0067,
|
| 855 |
+
"step": 1280
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 2.45,
|
| 859 |
+
"learning_rate": 3.112074716477652e-05,
|
| 860 |
+
"loss": 0.0023,
|
| 861 |
+
"step": 1290
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 2.47,
|
| 865 |
+
"learning_rate": 3.095396931287525e-05,
|
| 866 |
+
"loss": 0.0097,
|
| 867 |
+
"step": 1300
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 2.49,
|
| 871 |
+
"learning_rate": 3.078719146097398e-05,
|
| 872 |
+
"loss": 0.0178,
|
| 873 |
+
"step": 1310
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.51,
|
| 877 |
+
"learning_rate": 3.062041360907272e-05,
|
| 878 |
+
"loss": 0.0035,
|
| 879 |
+
"step": 1320
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.53,
|
| 883 |
+
"learning_rate": 3.045363575717145e-05,
|
| 884 |
+
"loss": 0.0055,
|
| 885 |
+
"step": 1330
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 2.55,
|
| 889 |
+
"learning_rate": 3.028685790527018e-05,
|
| 890 |
+
"loss": 0.0068,
|
| 891 |
+
"step": 1340
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 2.57,
|
| 895 |
+
"learning_rate": 3.0120080053368915e-05,
|
| 896 |
+
"loss": 0.0032,
|
| 897 |
+
"step": 1350
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 2.59,
|
| 901 |
+
"learning_rate": 2.995330220146765e-05,
|
| 902 |
+
"loss": 0.0047,
|
| 903 |
+
"step": 1360
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 2.6,
|
| 907 |
+
"learning_rate": 2.978652434956638e-05,
|
| 908 |
+
"loss": 0.0062,
|
| 909 |
+
"step": 1370
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 2.62,
|
| 913 |
+
"learning_rate": 2.9619746497665114e-05,
|
| 914 |
+
"loss": 0.0033,
|
| 915 |
+
"step": 1380
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.64,
|
| 919 |
+
"learning_rate": 2.9452968645763845e-05,
|
| 920 |
+
"loss": 0.0024,
|
| 921 |
+
"step": 1390
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.66,
|
| 925 |
+
"learning_rate": 2.928619079386258e-05,
|
| 926 |
+
"loss": 0.0024,
|
| 927 |
+
"step": 1400
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 2.66,
|
| 931 |
+
"eval_f1": 0.9959387557872723,
|
| 932 |
+
"eval_f1_0": 0.998752586951556,
|
| 933 |
+
"eval_f1_1": 0.9977694492109829,
|
| 934 |
+
"eval_f1_2": 0.9880693801743948,
|
| 935 |
+
"eval_f1_3": 0.9991636068121555,
|
| 936 |
+
"eval_loss": 0.00828399695456028,
|
| 937 |
+
"eval_runtime": 19.3722,
|
| 938 |
+
"eval_samples_per_second": 385.707,
|
| 939 |
+
"eval_steps_per_second": 0.413,
|
| 940 |
+
"step": 1400
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 2.68,
|
| 944 |
+
"learning_rate": 2.911941294196131e-05,
|
| 945 |
+
"loss": 0.0035,
|
| 946 |
+
"step": 1410
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 2.7,
|
| 950 |
+
"learning_rate": 2.8952635090060043e-05,
|
| 951 |
+
"loss": 0.0023,
|
| 952 |
+
"step": 1420
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 2.72,
|
| 956 |
+
"learning_rate": 2.8785857238158774e-05,
|
| 957 |
+
"loss": 0.0047,
|
| 958 |
+
"step": 1430
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 2.74,
|
| 962 |
+
"learning_rate": 2.8619079386257508e-05,
|
| 963 |
+
"loss": 0.0049,
|
| 964 |
+
"step": 1440
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 2.76,
|
| 968 |
+
"learning_rate": 2.8452301534356242e-05,
|
| 969 |
+
"loss": 0.0071,
|
| 970 |
+
"step": 1450
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 2.78,
|
| 974 |
+
"learning_rate": 2.8285523682454973e-05,
|
| 975 |
+
"loss": 0.0025,
|
| 976 |
+
"step": 1460
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 2.79,
|
| 980 |
+
"learning_rate": 2.8118745830553707e-05,
|
| 981 |
+
"loss": 0.0028,
|
| 982 |
+
"step": 1470
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 2.81,
|
| 986 |
+
"learning_rate": 2.7951967978652437e-05,
|
| 987 |
+
"loss": 0.0041,
|
| 988 |
+
"step": 1480
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 2.83,
|
| 992 |
+
"learning_rate": 2.778519012675117e-05,
|
| 993 |
+
"loss": 0.0054,
|
| 994 |
+
"step": 1490
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 2.85,
|
| 998 |
+
"learning_rate": 2.7618412274849902e-05,
|
| 999 |
+
"loss": 0.0077,
|
| 1000 |
+
"step": 1500
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 2.87,
|
| 1004 |
+
"learning_rate": 2.7451634422948636e-05,
|
| 1005 |
+
"loss": 0.0046,
|
| 1006 |
+
"step": 1510
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 2.89,
|
| 1010 |
+
"learning_rate": 2.7284856571047366e-05,
|
| 1011 |
+
"loss": 0.004,
|
| 1012 |
+
"step": 1520
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 2.91,
|
| 1016 |
+
"learning_rate": 2.71180787191461e-05,
|
| 1017 |
+
"loss": 0.0053,
|
| 1018 |
+
"step": 1530
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 2.93,
|
| 1022 |
+
"learning_rate": 2.6951300867244835e-05,
|
| 1023 |
+
"loss": 0.0035,
|
| 1024 |
+
"step": 1540
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 2.95,
|
| 1028 |
+
"learning_rate": 2.6784523015343565e-05,
|
| 1029 |
+
"loss": 0.0042,
|
| 1030 |
+
"step": 1550
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 2.97,
|
| 1034 |
+
"learning_rate": 2.66177451634423e-05,
|
| 1035 |
+
"loss": 0.01,
|
| 1036 |
+
"step": 1560
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 2.98,
|
| 1040 |
+
"learning_rate": 2.645096731154103e-05,
|
| 1041 |
+
"loss": 0.0045,
|
| 1042 |
+
"step": 1570
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 3.0,
|
| 1046 |
+
"learning_rate": 2.6284189459639764e-05,
|
| 1047 |
+
"loss": 0.0028,
|
| 1048 |
+
"step": 1580
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 3.02,
|
| 1052 |
+
"learning_rate": 2.6117411607738494e-05,
|
| 1053 |
+
"loss": 0.0028,
|
| 1054 |
+
"step": 1590
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 3.04,
|
| 1058 |
+
"learning_rate": 2.595063375583723e-05,
|
| 1059 |
+
"loss": 0.002,
|
| 1060 |
+
"step": 1600
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 3.04,
|
| 1064 |
+
"eval_f1": 0.9960098503562397,
|
| 1065 |
+
"eval_f1_0": 0.9987963282060577,
|
| 1066 |
+
"eval_f1_1": 0.9984863058871551,
|
| 1067 |
+
"eval_f1_2": 0.9876790221409639,
|
| 1068 |
+
"eval_f1_3": 0.9990777451907822,
|
| 1069 |
+
"eval_loss": 0.009044110774993896,
|
| 1070 |
+
"eval_runtime": 19.4406,
|
| 1071 |
+
"eval_samples_per_second": 384.35,
|
| 1072 |
+
"eval_steps_per_second": 0.412,
|
| 1073 |
+
"step": 1600
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 3.06,
|
| 1077 |
+
"learning_rate": 2.578385590393596e-05,
|
| 1078 |
+
"loss": 0.0038,
|
| 1079 |
+
"step": 1610
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.08,
|
| 1083 |
+
"learning_rate": 2.5617078052034693e-05,
|
| 1084 |
+
"loss": 0.0027,
|
| 1085 |
+
"step": 1620
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 3.1,
|
| 1089 |
+
"learning_rate": 2.5450300200133427e-05,
|
| 1090 |
+
"loss": 0.0021,
|
| 1091 |
+
"step": 1630
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 3.12,
|
| 1095 |
+
"learning_rate": 2.5283522348232158e-05,
|
| 1096 |
+
"loss": 0.0025,
|
| 1097 |
+
"step": 1640
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 3.14,
|
| 1101 |
+
"learning_rate": 2.5116744496330892e-05,
|
| 1102 |
+
"loss": 0.0038,
|
| 1103 |
+
"step": 1650
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 3.16,
|
| 1107 |
+
"learning_rate": 2.494996664442962e-05,
|
| 1108 |
+
"loss": 0.0024,
|
| 1109 |
+
"step": 1660
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 3.17,
|
| 1113 |
+
"learning_rate": 2.4783188792528353e-05,
|
| 1114 |
+
"loss": 0.002,
|
| 1115 |
+
"step": 1670
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 3.19,
|
| 1119 |
+
"learning_rate": 2.4616410940627087e-05,
|
| 1120 |
+
"loss": 0.0034,
|
| 1121 |
+
"step": 1680
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 3.21,
|
| 1125 |
+
"learning_rate": 2.4449633088725818e-05,
|
| 1126 |
+
"loss": 0.002,
|
| 1127 |
+
"step": 1690
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 3.23,
|
| 1131 |
+
"learning_rate": 2.4282855236824552e-05,
|
| 1132 |
+
"loss": 0.0018,
|
| 1133 |
+
"step": 1700
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 3.25,
|
| 1137 |
+
"learning_rate": 2.4116077384923282e-05,
|
| 1138 |
+
"loss": 0.0026,
|
| 1139 |
+
"step": 1710
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 3.27,
|
| 1143 |
+
"learning_rate": 2.3949299533022016e-05,
|
| 1144 |
+
"loss": 0.0019,
|
| 1145 |
+
"step": 1720
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 3.29,
|
| 1149 |
+
"learning_rate": 2.3782521681120747e-05,
|
| 1150 |
+
"loss": 0.0076,
|
| 1151 |
+
"step": 1730
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 3.31,
|
| 1155 |
+
"learning_rate": 2.361574382921948e-05,
|
| 1156 |
+
"loss": 0.003,
|
| 1157 |
+
"step": 1740
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 3.33,
|
| 1161 |
+
"learning_rate": 2.344896597731821e-05,
|
| 1162 |
+
"loss": 0.0018,
|
| 1163 |
+
"step": 1750
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 3.35,
|
| 1167 |
+
"learning_rate": 2.3282188125416946e-05,
|
| 1168 |
+
"loss": 0.0025,
|
| 1169 |
+
"step": 1760
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 3.37,
|
| 1173 |
+
"learning_rate": 2.311541027351568e-05,
|
| 1174 |
+
"loss": 0.0025,
|
| 1175 |
+
"step": 1770
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 3.38,
|
| 1179 |
+
"learning_rate": 2.294863242161441e-05,
|
| 1180 |
+
"loss": 0.0025,
|
| 1181 |
+
"step": 1780
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 3.4,
|
| 1185 |
+
"learning_rate": 2.2781854569713144e-05,
|
| 1186 |
+
"loss": 0.0018,
|
| 1187 |
+
"step": 1790
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 3.42,
|
| 1191 |
+
"learning_rate": 2.2615076717811875e-05,
|
| 1192 |
+
"loss": 0.0017,
|
| 1193 |
+
"step": 1800
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 3.42,
|
| 1197 |
+
"eval_f1": 0.9965715706847758,
|
| 1198 |
+
"eval_f1_0": 0.9989613839913065,
|
| 1199 |
+
"eval_f1_1": 0.9985897813972129,
|
| 1200 |
+
"eval_f1_2": 0.9894973463811291,
|
| 1201 |
+
"eval_f1_3": 0.9992377709694545,
|
| 1202 |
+
"eval_loss": 0.0082913339138031,
|
| 1203 |
+
"eval_runtime": 19.4143,
|
| 1204 |
+
"eval_samples_per_second": 384.87,
|
| 1205 |
+
"eval_steps_per_second": 0.412,
|
| 1206 |
+
"step": 1800
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 3.44,
|
| 1210 |
+
"learning_rate": 2.244829886591061e-05,
|
| 1211 |
+
"loss": 0.0019,
|
| 1212 |
+
"step": 1810
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 3.46,
|
| 1216 |
+
"learning_rate": 2.228152101400934e-05,
|
| 1217 |
+
"loss": 0.0012,
|
| 1218 |
+
"step": 1820
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 3.48,
|
| 1222 |
+
"learning_rate": 2.2114743162108074e-05,
|
| 1223 |
+
"loss": 0.0014,
|
| 1224 |
+
"step": 1830
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 3.5,
|
| 1228 |
+
"learning_rate": 2.1947965310206804e-05,
|
| 1229 |
+
"loss": 0.0018,
|
| 1230 |
+
"step": 1840
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 3.52,
|
| 1234 |
+
"learning_rate": 2.1781187458305538e-05,
|
| 1235 |
+
"loss": 0.0017,
|
| 1236 |
+
"step": 1850
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 3.54,
|
| 1240 |
+
"learning_rate": 2.161440960640427e-05,
|
| 1241 |
+
"loss": 0.0017,
|
| 1242 |
+
"step": 1860
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 3.56,
|
| 1246 |
+
"learning_rate": 2.1447631754503003e-05,
|
| 1247 |
+
"loss": 0.0047,
|
| 1248 |
+
"step": 1870
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 3.57,
|
| 1252 |
+
"learning_rate": 2.1280853902601737e-05,
|
| 1253 |
+
"loss": 0.0024,
|
| 1254 |
+
"step": 1880
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 3.59,
|
| 1258 |
+
"learning_rate": 2.1114076050700468e-05,
|
| 1259 |
+
"loss": 0.0021,
|
| 1260 |
+
"step": 1890
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 3.61,
|
| 1264 |
+
"learning_rate": 2.09472981987992e-05,
|
| 1265 |
+
"loss": 0.0019,
|
| 1266 |
+
"step": 1900
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 3.63,
|
| 1270 |
+
"learning_rate": 2.0780520346897932e-05,
|
| 1271 |
+
"loss": 0.0016,
|
| 1272 |
+
"step": 1910
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 3.65,
|
| 1276 |
+
"learning_rate": 2.0613742494996666e-05,
|
| 1277 |
+
"loss": 0.0017,
|
| 1278 |
+
"step": 1920
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 3.67,
|
| 1282 |
+
"learning_rate": 2.0446964643095397e-05,
|
| 1283 |
+
"loss": 0.0014,
|
| 1284 |
+
"step": 1930
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 3.69,
|
| 1288 |
+
"learning_rate": 2.028018679119413e-05,
|
| 1289 |
+
"loss": 0.0015,
|
| 1290 |
+
"step": 1940
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 3.71,
|
| 1294 |
+
"learning_rate": 2.011340893929286e-05,
|
| 1295 |
+
"loss": 0.0022,
|
| 1296 |
+
"step": 1950
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 3.73,
|
| 1300 |
+
"learning_rate": 1.9946631087391596e-05,
|
| 1301 |
+
"loss": 0.0026,
|
| 1302 |
+
"step": 1960
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 3.75,
|
| 1306 |
+
"learning_rate": 1.977985323549033e-05,
|
| 1307 |
+
"loss": 0.0017,
|
| 1308 |
+
"step": 1970
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 3.76,
|
| 1312 |
+
"learning_rate": 1.961307538358906e-05,
|
| 1313 |
+
"loss": 0.0017,
|
| 1314 |
+
"step": 1980
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 3.78,
|
| 1318 |
+
"learning_rate": 1.9446297531687794e-05,
|
| 1319 |
+
"loss": 0.0038,
|
| 1320 |
+
"step": 1990
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 3.8,
|
| 1324 |
+
"learning_rate": 1.9279519679786525e-05,
|
| 1325 |
+
"loss": 0.0019,
|
| 1326 |
+
"step": 2000
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 3.8,
|
| 1330 |
+
"eval_f1": 0.9954966488952333,
|
| 1331 |
+
"eval_f1_0": 0.9986378180905452,
|
| 1332 |
+
"eval_f1_1": 0.998616706855238,
|
| 1333 |
+
"eval_f1_2": 0.9854914081132,
|
| 1334 |
+
"eval_f1_3": 0.9992406625219497,
|
| 1335 |
+
"eval_loss": 0.010157858952879906,
|
| 1336 |
+
"eval_runtime": 19.3129,
|
| 1337 |
+
"eval_samples_per_second": 386.893,
|
| 1338 |
+
"eval_steps_per_second": 0.414,
|
| 1339 |
+
"step": 2000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 3.82,
|
| 1343 |
+
"learning_rate": 1.911274182788526e-05,
|
| 1344 |
+
"loss": 0.0038,
|
| 1345 |
+
"step": 2010
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 3.84,
|
| 1349 |
+
"learning_rate": 1.894596397598399e-05,
|
| 1350 |
+
"loss": 0.0018,
|
| 1351 |
+
"step": 2020
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 3.86,
|
| 1355 |
+
"learning_rate": 1.8779186124082724e-05,
|
| 1356 |
+
"loss": 0.0014,
|
| 1357 |
+
"step": 2030
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 3.88,
|
| 1361 |
+
"learning_rate": 1.8612408272181454e-05,
|
| 1362 |
+
"loss": 0.0027,
|
| 1363 |
+
"step": 2040
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 3.9,
|
| 1367 |
+
"learning_rate": 1.8445630420280188e-05,
|
| 1368 |
+
"loss": 0.0022,
|
| 1369 |
+
"step": 2050
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 3.92,
|
| 1373 |
+
"learning_rate": 1.8278852568378922e-05,
|
| 1374 |
+
"loss": 0.0034,
|
| 1375 |
+
"step": 2060
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 3.94,
|
| 1379 |
+
"learning_rate": 1.8112074716477653e-05,
|
| 1380 |
+
"loss": 0.0026,
|
| 1381 |
+
"step": 2070
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 3.95,
|
| 1385 |
+
"learning_rate": 1.7945296864576387e-05,
|
| 1386 |
+
"loss": 0.0013,
|
| 1387 |
+
"step": 2080
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 3.97,
|
| 1391 |
+
"learning_rate": 1.7778519012675117e-05,
|
| 1392 |
+
"loss": 0.0028,
|
| 1393 |
+
"step": 2090
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 3.99,
|
| 1397 |
+
"learning_rate": 1.761174116077385e-05,
|
| 1398 |
+
"loss": 0.0024,
|
| 1399 |
+
"step": 2100
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 4.01,
|
| 1403 |
+
"learning_rate": 1.7444963308872582e-05,
|
| 1404 |
+
"loss": 0.0024,
|
| 1405 |
+
"step": 2110
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 4.03,
|
| 1409 |
+
"learning_rate": 1.7278185456971316e-05,
|
| 1410 |
+
"loss": 0.0012,
|
| 1411 |
+
"step": 2120
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 4.05,
|
| 1415 |
+
"learning_rate": 1.7111407605070047e-05,
|
| 1416 |
+
"loss": 0.0013,
|
| 1417 |
+
"step": 2130
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 4.07,
|
| 1421 |
+
"learning_rate": 1.694462975316878e-05,
|
| 1422 |
+
"loss": 0.0013,
|
| 1423 |
+
"step": 2140
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 4.09,
|
| 1427 |
+
"learning_rate": 1.6777851901267515e-05,
|
| 1428 |
+
"loss": 0.002,
|
| 1429 |
+
"step": 2150
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 4.11,
|
| 1433 |
+
"learning_rate": 1.6611074049366245e-05,
|
| 1434 |
+
"loss": 0.0017,
|
| 1435 |
+
"step": 2160
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 4.13,
|
| 1439 |
+
"learning_rate": 1.644429619746498e-05,
|
| 1440 |
+
"loss": 0.0013,
|
| 1441 |
+
"step": 2170
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 4.14,
|
| 1445 |
+
"learning_rate": 1.627751834556371e-05,
|
| 1446 |
+
"loss": 0.0015,
|
| 1447 |
+
"step": 2180
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 4.16,
|
| 1451 |
+
"learning_rate": 1.6110740493662444e-05,
|
| 1452 |
+
"loss": 0.0012,
|
| 1453 |
+
"step": 2190
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 4.18,
|
| 1457 |
+
"learning_rate": 1.5943962641761175e-05,
|
| 1458 |
+
"loss": 0.0016,
|
| 1459 |
+
"step": 2200
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 4.18,
|
| 1463 |
+
"eval_f1": 0.9964539401625592,
|
| 1464 |
+
"eval_f1_0": 0.9989316835915937,
|
| 1465 |
+
"eval_f1_1": 0.9986976730119667,
|
| 1466 |
+
"eval_f1_2": 0.9889220122285377,
|
| 1467 |
+
"eval_f1_3": 0.9992643918181386,
|
| 1468 |
+
"eval_loss": 0.008549452759325504,
|
| 1469 |
+
"eval_runtime": 19.2872,
|
| 1470 |
+
"eval_samples_per_second": 387.407,
|
| 1471 |
+
"eval_steps_per_second": 0.415,
|
| 1472 |
+
"step": 2200
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 4.2,
|
| 1476 |
+
"learning_rate": 1.577718478985991e-05,
|
| 1477 |
+
"loss": 0.0013,
|
| 1478 |
+
"step": 2210
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 4.22,
|
| 1482 |
+
"learning_rate": 1.561040693795864e-05,
|
| 1483 |
+
"loss": 0.0018,
|
| 1484 |
+
"step": 2220
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 4.24,
|
| 1488 |
+
"learning_rate": 1.5443629086057373e-05,
|
| 1489 |
+
"loss": 0.0012,
|
| 1490 |
+
"step": 2230
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 4.26,
|
| 1494 |
+
"learning_rate": 1.5276851234156107e-05,
|
| 1495 |
+
"loss": 0.0014,
|
| 1496 |
+
"step": 2240
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 4.28,
|
| 1500 |
+
"learning_rate": 1.5110073382254838e-05,
|
| 1501 |
+
"loss": 0.0011,
|
| 1502 |
+
"step": 2250
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 4.3,
|
| 1506 |
+
"learning_rate": 1.494329553035357e-05,
|
| 1507 |
+
"loss": 0.0044,
|
| 1508 |
+
"step": 2260
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 4.32,
|
| 1512 |
+
"learning_rate": 1.4776517678452303e-05,
|
| 1513 |
+
"loss": 0.0028,
|
| 1514 |
+
"step": 2270
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 4.33,
|
| 1518 |
+
"learning_rate": 1.4609739826551037e-05,
|
| 1519 |
+
"loss": 0.0014,
|
| 1520 |
+
"step": 2280
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 4.35,
|
| 1524 |
+
"learning_rate": 1.4442961974649769e-05,
|
| 1525 |
+
"loss": 0.001,
|
| 1526 |
+
"step": 2290
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 4.37,
|
| 1530 |
+
"learning_rate": 1.4276184122748501e-05,
|
| 1531 |
+
"loss": 0.0018,
|
| 1532 |
+
"step": 2300
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 4.39,
|
| 1536 |
+
"learning_rate": 1.4109406270847234e-05,
|
| 1537 |
+
"loss": 0.0012,
|
| 1538 |
+
"step": 2310
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 4.41,
|
| 1542 |
+
"learning_rate": 1.3942628418945963e-05,
|
| 1543 |
+
"loss": 0.0011,
|
| 1544 |
+
"step": 2320
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 4.43,
|
| 1548 |
+
"learning_rate": 1.3775850567044695e-05,
|
| 1549 |
+
"loss": 0.0011,
|
| 1550 |
+
"step": 2330
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 4.45,
|
| 1554 |
+
"learning_rate": 1.3609072715143429e-05,
|
| 1555 |
+
"loss": 0.0013,
|
| 1556 |
+
"step": 2340
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 4.47,
|
| 1560 |
+
"learning_rate": 1.3442294863242161e-05,
|
| 1561 |
+
"loss": 0.0022,
|
| 1562 |
+
"step": 2350
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 4.49,
|
| 1566 |
+
"learning_rate": 1.3275517011340894e-05,
|
| 1567 |
+
"loss": 0.0012,
|
| 1568 |
+
"step": 2360
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 4.51,
|
| 1572 |
+
"learning_rate": 1.3108739159439626e-05,
|
| 1573 |
+
"loss": 0.0013,
|
| 1574 |
+
"step": 2370
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 4.52,
|
| 1578 |
+
"learning_rate": 1.2941961307538358e-05,
|
| 1579 |
+
"loss": 0.0014,
|
| 1580 |
+
"step": 2380
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 4.54,
|
| 1584 |
+
"learning_rate": 1.277518345563709e-05,
|
| 1585 |
+
"loss": 0.0011,
|
| 1586 |
+
"step": 2390
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 4.56,
|
| 1590 |
+
"learning_rate": 1.2608405603735823e-05,
|
| 1591 |
+
"loss": 0.0013,
|
| 1592 |
+
"step": 2400
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 4.56,
|
| 1596 |
+
"eval_f1": 0.9965431300136034,
|
| 1597 |
+
"eval_f1_0": 0.9989591777046161,
|
| 1598 |
+
"eval_f1_1": 0.9986785664998649,
|
| 1599 |
+
"eval_f1_2": 0.9892763163558415,
|
| 1600 |
+
"eval_f1_3": 0.9992584594940914,
|
| 1601 |
+
"eval_loss": 0.008574241772294044,
|
| 1602 |
+
"eval_runtime": 19.3355,
|
| 1603 |
+
"eval_samples_per_second": 386.439,
|
| 1604 |
+
"eval_steps_per_second": 0.414,
|
| 1605 |
+
"step": 2400
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 4.58,
|
| 1609 |
+
"learning_rate": 1.2441627751834557e-05,
|
| 1610 |
+
"loss": 0.0018,
|
| 1611 |
+
"step": 2410
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 4.6,
|
| 1615 |
+
"learning_rate": 1.227484989993329e-05,
|
| 1616 |
+
"loss": 0.0011,
|
| 1617 |
+
"step": 2420
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 4.62,
|
| 1621 |
+
"learning_rate": 1.2108072048032022e-05,
|
| 1622 |
+
"loss": 0.0012,
|
| 1623 |
+
"step": 2430
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 4.64,
|
| 1627 |
+
"learning_rate": 1.1941294196130754e-05,
|
| 1628 |
+
"loss": 0.0018,
|
| 1629 |
+
"step": 2440
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 4.66,
|
| 1633 |
+
"learning_rate": 1.1774516344229488e-05,
|
| 1634 |
+
"loss": 0.0012,
|
| 1635 |
+
"step": 2450
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 4.68,
|
| 1639 |
+
"learning_rate": 1.160773849232822e-05,
|
| 1640 |
+
"loss": 0.0019,
|
| 1641 |
+
"step": 2460
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 4.7,
|
| 1645 |
+
"learning_rate": 1.1440960640426953e-05,
|
| 1646 |
+
"loss": 0.0013,
|
| 1647 |
+
"step": 2470
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 4.71,
|
| 1651 |
+
"learning_rate": 1.1274182788525685e-05,
|
| 1652 |
+
"loss": 0.0014,
|
| 1653 |
+
"step": 2480
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 4.73,
|
| 1657 |
+
"learning_rate": 1.1107404936624417e-05,
|
| 1658 |
+
"loss": 0.0012,
|
| 1659 |
+
"step": 2490
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 4.75,
|
| 1663 |
+
"learning_rate": 1.094062708472315e-05,
|
| 1664 |
+
"loss": 0.0015,
|
| 1665 |
+
"step": 2500
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 4.77,
|
| 1669 |
+
"learning_rate": 1.0773849232821882e-05,
|
| 1670 |
+
"loss": 0.0012,
|
| 1671 |
+
"step": 2510
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 4.79,
|
| 1675 |
+
"learning_rate": 1.0607071380920614e-05,
|
| 1676 |
+
"loss": 0.0011,
|
| 1677 |
+
"step": 2520
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 4.81,
|
| 1681 |
+
"learning_rate": 1.0440293529019347e-05,
|
| 1682 |
+
"loss": 0.0016,
|
| 1683 |
+
"step": 2530
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 4.83,
|
| 1687 |
+
"learning_rate": 1.027351567711808e-05,
|
| 1688 |
+
"loss": 0.0025,
|
| 1689 |
+
"step": 2540
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 4.85,
|
| 1693 |
+
"learning_rate": 1.0106737825216811e-05,
|
| 1694 |
+
"loss": 0.0019,
|
| 1695 |
+
"step": 2550
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 4.87,
|
| 1699 |
+
"learning_rate": 9.939959973315544e-06,
|
| 1700 |
+
"loss": 0.0012,
|
| 1701 |
+
"step": 2560
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 4.89,
|
| 1705 |
+
"learning_rate": 9.773182121414276e-06,
|
| 1706 |
+
"loss": 0.0011,
|
| 1707 |
+
"step": 2570
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 4.9,
|
| 1711 |
+
"learning_rate": 9.606404269513008e-06,
|
| 1712 |
+
"loss": 0.0018,
|
| 1713 |
+
"step": 2580
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 4.92,
|
| 1717 |
+
"learning_rate": 9.43962641761174e-06,
|
| 1718 |
+
"loss": 0.001,
|
| 1719 |
+
"step": 2590
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 4.94,
|
| 1723 |
+
"learning_rate": 9.272848565710473e-06,
|
| 1724 |
+
"loss": 0.0016,
|
| 1725 |
+
"step": 2600
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 4.94,
|
| 1729 |
+
"eval_f1": 0.9963671849973992,
|
| 1730 |
+
"eval_f1_0": 0.9988809930374426,
|
| 1731 |
+
"eval_f1_1": 0.9980475327397732,
|
| 1732 |
+
"eval_f1_2": 0.9892787776027725,
|
| 1733 |
+
"eval_f1_3": 0.9992614366096084,
|
| 1734 |
+
"eval_loss": 0.009876401163637638,
|
| 1735 |
+
"eval_runtime": 19.3714,
|
| 1736 |
+
"eval_samples_per_second": 385.724,
|
| 1737 |
+
"eval_steps_per_second": 0.413,
|
| 1738 |
+
"step": 2600
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 4.96,
|
| 1742 |
+
"learning_rate": 9.106070713809207e-06,
|
| 1743 |
+
"loss": 0.001,
|
| 1744 |
+
"step": 2610
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 4.98,
|
| 1748 |
+
"learning_rate": 8.93929286190794e-06,
|
| 1749 |
+
"loss": 0.0023,
|
| 1750 |
+
"step": 2620
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 5.0,
|
| 1754 |
+
"learning_rate": 8.772515010006672e-06,
|
| 1755 |
+
"loss": 0.0024,
|
| 1756 |
+
"step": 2630
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 5.02,
|
| 1760 |
+
"learning_rate": 8.605737158105404e-06,
|
| 1761 |
+
"loss": 0.0026,
|
| 1762 |
+
"step": 2640
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 5.04,
|
| 1766 |
+
"learning_rate": 8.438959306204136e-06,
|
| 1767 |
+
"loss": 0.0011,
|
| 1768 |
+
"step": 2650
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 5.06,
|
| 1772 |
+
"learning_rate": 8.272181454302869e-06,
|
| 1773 |
+
"loss": 0.0012,
|
| 1774 |
+
"step": 2660
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 5.08,
|
| 1778 |
+
"learning_rate": 8.105403602401601e-06,
|
| 1779 |
+
"loss": 0.001,
|
| 1780 |
+
"step": 2670
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 5.1,
|
| 1784 |
+
"learning_rate": 7.938625750500333e-06,
|
| 1785 |
+
"loss": 0.001,
|
| 1786 |
+
"step": 2680
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 5.11,
|
| 1790 |
+
"learning_rate": 7.771847898599065e-06,
|
| 1791 |
+
"loss": 0.001,
|
| 1792 |
+
"step": 2690
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 5.13,
|
| 1796 |
+
"learning_rate": 7.605070046697799e-06,
|
| 1797 |
+
"loss": 0.002,
|
| 1798 |
+
"step": 2700
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 5.15,
|
| 1802 |
+
"learning_rate": 7.438292194796531e-06,
|
| 1803 |
+
"loss": 0.0013,
|
| 1804 |
+
"step": 2710
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 5.17,
|
| 1808 |
+
"learning_rate": 7.271514342895263e-06,
|
| 1809 |
+
"loss": 0.0012,
|
| 1810 |
+
"step": 2720
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 5.19,
|
| 1814 |
+
"learning_rate": 7.1047364909939965e-06,
|
| 1815 |
+
"loss": 0.0011,
|
| 1816 |
+
"step": 2730
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 5.21,
|
| 1820 |
+
"learning_rate": 6.937958639092729e-06,
|
| 1821 |
+
"loss": 0.0009,
|
| 1822 |
+
"step": 2740
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 5.23,
|
| 1826 |
+
"learning_rate": 6.771180787191461e-06,
|
| 1827 |
+
"loss": 0.0011,
|
| 1828 |
+
"step": 2750
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 5.25,
|
| 1832 |
+
"learning_rate": 6.6044029352901935e-06,
|
| 1833 |
+
"loss": 0.0012,
|
| 1834 |
+
"step": 2760
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 5.27,
|
| 1838 |
+
"learning_rate": 6.437625083388927e-06,
|
| 1839 |
+
"loss": 0.0011,
|
| 1840 |
+
"step": 2770
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 5.29,
|
| 1844 |
+
"learning_rate": 6.270847231487659e-06,
|
| 1845 |
+
"loss": 0.001,
|
| 1846 |
+
"step": 2780
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 5.3,
|
| 1850 |
+
"learning_rate": 6.104069379586391e-06,
|
| 1851 |
+
"loss": 0.0008,
|
| 1852 |
+
"step": 2790
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 5.32,
|
| 1856 |
+
"learning_rate": 5.937291527685124e-06,
|
| 1857 |
+
"loss": 0.0011,
|
| 1858 |
+
"step": 2800
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 5.32,
|
| 1862 |
+
"eval_f1": 0.9965125366318218,
|
| 1863 |
+
"eval_f1_0": 0.9989578910626513,
|
| 1864 |
+
"eval_f1_1": 0.9988072938423473,
|
| 1865 |
+
"eval_f1_2": 0.9889968193931058,
|
| 1866 |
+
"eval_f1_3": 0.9992881422291826,
|
| 1867 |
+
"eval_loss": 0.009526900947093964,
|
| 1868 |
+
"eval_runtime": 19.4195,
|
| 1869 |
+
"eval_samples_per_second": 384.768,
|
| 1870 |
+
"eval_steps_per_second": 0.412,
|
| 1871 |
+
"step": 2800
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 5.34,
|
| 1875 |
+
"learning_rate": 5.770513675783856e-06,
|
| 1876 |
+
"loss": 0.0018,
|
| 1877 |
+
"step": 2810
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 5.36,
|
| 1881 |
+
"learning_rate": 5.603735823882589e-06,
|
| 1882 |
+
"loss": 0.001,
|
| 1883 |
+
"step": 2820
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 5.38,
|
| 1887 |
+
"learning_rate": 5.4369579719813214e-06,
|
| 1888 |
+
"loss": 0.0012,
|
| 1889 |
+
"step": 2830
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 5.4,
|
| 1893 |
+
"learning_rate": 5.270180120080054e-06,
|
| 1894 |
+
"loss": 0.0011,
|
| 1895 |
+
"step": 2840
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 5.42,
|
| 1899 |
+
"learning_rate": 5.103402268178786e-06,
|
| 1900 |
+
"loss": 0.0008,
|
| 1901 |
+
"step": 2850
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 5.44,
|
| 1905 |
+
"learning_rate": 4.936624416277518e-06,
|
| 1906 |
+
"loss": 0.0013,
|
| 1907 |
+
"step": 2860
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 5.46,
|
| 1911 |
+
"learning_rate": 4.769846564376252e-06,
|
| 1912 |
+
"loss": 0.0011,
|
| 1913 |
+
"step": 2870
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 5.48,
|
| 1917 |
+
"learning_rate": 4.603068712474984e-06,
|
| 1918 |
+
"loss": 0.0013,
|
| 1919 |
+
"step": 2880
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 5.49,
|
| 1923 |
+
"learning_rate": 4.436290860573716e-06,
|
| 1924 |
+
"loss": 0.0009,
|
| 1925 |
+
"step": 2890
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 5.51,
|
| 1929 |
+
"learning_rate": 4.2695130086724486e-06,
|
| 1930 |
+
"loss": 0.001,
|
| 1931 |
+
"step": 2900
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 5.53,
|
| 1935 |
+
"learning_rate": 4.102735156771181e-06,
|
| 1936 |
+
"loss": 0.001,
|
| 1937 |
+
"step": 2910
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 5.55,
|
| 1941 |
+
"learning_rate": 3.935957304869913e-06,
|
| 1942 |
+
"loss": 0.0022,
|
| 1943 |
+
"step": 2920
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 5.57,
|
| 1947 |
+
"learning_rate": 3.7691794529686455e-06,
|
| 1948 |
+
"loss": 0.001,
|
| 1949 |
+
"step": 2930
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 5.59,
|
| 1953 |
+
"learning_rate": 3.6024016010673783e-06,
|
| 1954 |
+
"loss": 0.0013,
|
| 1955 |
+
"step": 2940
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 5.61,
|
| 1959 |
+
"learning_rate": 3.4356237491661106e-06,
|
| 1960 |
+
"loss": 0.0009,
|
| 1961 |
+
"step": 2950
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 5.63,
|
| 1965 |
+
"learning_rate": 3.2688458972648434e-06,
|
| 1966 |
+
"loss": 0.0011,
|
| 1967 |
+
"step": 2960
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 5.65,
|
| 1971 |
+
"learning_rate": 3.1020680453635757e-06,
|
| 1972 |
+
"loss": 0.0013,
|
| 1973 |
+
"step": 2970
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 5.67,
|
| 1977 |
+
"learning_rate": 2.9352901934623084e-06,
|
| 1978 |
+
"loss": 0.001,
|
| 1979 |
+
"step": 2980
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 5.68,
|
| 1983 |
+
"learning_rate": 2.7685123415610408e-06,
|
| 1984 |
+
"loss": 0.0018,
|
| 1985 |
+
"step": 2990
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 5.7,
|
| 1989 |
+
"learning_rate": 2.601734489659773e-06,
|
| 1990 |
+
"loss": 0.0024,
|
| 1991 |
+
"step": 3000
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 5.7,
|
| 1995 |
+
"eval_f1": 0.9965564466847002,
|
| 1996 |
+
"eval_f1_0": 0.9989728439908399,
|
| 1997 |
+
"eval_f1_1": 0.998860218996728,
|
| 1998 |
+
"eval_f1_2": 0.9891016301666877,
|
| 1999 |
+
"eval_f1_3": 0.9992910935845453,
|
| 2000 |
+
"eval_loss": 0.009422173723578453,
|
| 2001 |
+
"eval_runtime": 19.3758,
|
| 2002 |
+
"eval_samples_per_second": 385.635,
|
| 2003 |
+
"eval_steps_per_second": 0.413,
|
| 2004 |
+
"step": 3000
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 5.72,
|
| 2008 |
+
"learning_rate": 2.434956637758506e-06,
|
| 2009 |
+
"loss": 0.0013,
|
| 2010 |
+
"step": 3010
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 5.74,
|
| 2014 |
+
"learning_rate": 2.268178785857238e-06,
|
| 2015 |
+
"loss": 0.0014,
|
| 2016 |
+
"step": 3020
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 5.76,
|
| 2020 |
+
"learning_rate": 2.101400933955971e-06,
|
| 2021 |
+
"loss": 0.001,
|
| 2022 |
+
"step": 3030
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 5.78,
|
| 2026 |
+
"learning_rate": 1.9346230820547032e-06,
|
| 2027 |
+
"loss": 0.0013,
|
| 2028 |
+
"step": 3040
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 5.8,
|
| 2032 |
+
"learning_rate": 1.7678452301534358e-06,
|
| 2033 |
+
"loss": 0.0015,
|
| 2034 |
+
"step": 3050
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 5.82,
|
| 2038 |
+
"learning_rate": 1.6010673782521683e-06,
|
| 2039 |
+
"loss": 0.0012,
|
| 2040 |
+
"step": 3060
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 5.84,
|
| 2044 |
+
"learning_rate": 1.4342895263509006e-06,
|
| 2045 |
+
"loss": 0.0014,
|
| 2046 |
+
"step": 3070
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 5.86,
|
| 2050 |
+
"learning_rate": 1.2675116744496332e-06,
|
| 2051 |
+
"loss": 0.0012,
|
| 2052 |
+
"step": 3080
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 5.87,
|
| 2056 |
+
"learning_rate": 1.1007338225483657e-06,
|
| 2057 |
+
"loss": 0.0011,
|
| 2058 |
+
"step": 3090
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 5.89,
|
| 2062 |
+
"learning_rate": 9.339559706470982e-07,
|
| 2063 |
+
"loss": 0.0008,
|
| 2064 |
+
"step": 3100
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 5.91,
|
| 2068 |
+
"learning_rate": 7.671781187458306e-07,
|
| 2069 |
+
"loss": 0.0013,
|
| 2070 |
+
"step": 3110
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 5.93,
|
| 2074 |
+
"learning_rate": 6.00400266844563e-07,
|
| 2075 |
+
"loss": 0.0011,
|
| 2076 |
+
"step": 3120
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 5.95,
|
| 2080 |
+
"learning_rate": 4.3362241494329555e-07,
|
| 2081 |
+
"loss": 0.001,
|
| 2082 |
+
"step": 3130
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 5.97,
|
| 2086 |
+
"learning_rate": 2.6684456304202804e-07,
|
| 2087 |
+
"loss": 0.0012,
|
| 2088 |
+
"step": 3140
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 5.99,
|
| 2092 |
+
"learning_rate": 1.0006671114076052e-07,
|
| 2093 |
+
"loss": 0.0011,
|
| 2094 |
+
"step": 3150
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 6.0,
|
| 2098 |
+
"step": 3156,
|
| 2099 |
+
"total_flos": 3.671049324168479e+17,
|
| 2100 |
+
"train_loss": 0.019633541173813333,
|
| 2101 |
+
"train_runtime": 2021.1766,
|
| 2102 |
+
"train_samples_per_second": 199.604,
|
| 2103 |
+
"train_steps_per_second": 1.561
|
| 2104 |
+
}
|
| 2105 |
+
],
|
| 2106 |
+
"logging_steps": 10,
|
| 2107 |
+
"max_steps": 3156,
|
| 2108 |
+
"num_train_epochs": 6,
|
| 2109 |
+
"save_steps": 200,
|
| 2110 |
+
"total_flos": 3.671049324168479e+17,
|
| 2111 |
+
"trial_name": null,
|
| 2112 |
+
"trial_params": null
|
| 2113 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26666eead089f572a38f14b4140b6c0ebdbe5f560121acacbf9ce47c003b48c0
|
| 3 |
+
size 4219
|