Upload 9 files
Browse files- config.json +24 -0
- eval_results.txt +12 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- train_results.txt +3 -0
- trainer_state.json +808 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "checkpoints/multicse-bert-large-uncased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForCL"
|
| 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": 1024,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 4096,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 16,
|
| 17 |
+
"num_hidden_layers": 24,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"transformers_version": "4.2.1",
|
| 21 |
+
"type_vocab_size": 2,
|
| 22 |
+
"use_cache": true,
|
| 23 |
+
"vocab_size": 30522
|
| 24 |
+
}
|
eval_results.txt
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch = 3.0
|
| 2 |
+
eval_CR = 90.48
|
| 3 |
+
eval_MPQA = 89.26
|
| 4 |
+
eval_MR = 83.67
|
| 5 |
+
eval_MRPC = 76.1
|
| 6 |
+
eval_SST2 = 88.76
|
| 7 |
+
eval_SUBJ = 94.42
|
| 8 |
+
eval_TREC = 78.56
|
| 9 |
+
eval_avg_sts = 0.84970489436635
|
| 10 |
+
eval_avg_transfer = 85.89285714285714
|
| 11 |
+
eval_sickr_spearman = 0.8333206856972879
|
| 12 |
+
eval_stsb_spearman = 0.8660891030354121
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea34bd77f0c8c64115c249126992f28f680859c9536d9acde15d25ab55409e34
|
| 3 |
+
size 1470080962
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "/data/lyhe/data/bert-large-uncased"}
|
train_results.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch = 3.0
|
| 2 |
+
train_runtime = 7555.7932
|
| 3 |
+
train_samples_per_second = 0.358
|
trainer_state.json
ADDED
|
@@ -0,0 +1,808 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.84970489436635,
|
| 3 |
+
"best_model_checkpoint": "checkpoints/multicse-bert-large-uncased",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 2706,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"eval_avg_sts": 0.7624996022459282,
|
| 13 |
+
"eval_sickr_spearman": 0.7306625637849724,
|
| 14 |
+
"eval_stsb_spearman": 0.7943366407068839,
|
| 15 |
+
"step": 25
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.06,
|
| 19 |
+
"eval_avg_sts": 0.8071313365516628,
|
| 20 |
+
"eval_sickr_spearman": 0.7781768986027124,
|
| 21 |
+
"eval_stsb_spearman": 0.8360857745006132,
|
| 22 |
+
"step": 50
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"epoch": 0.08,
|
| 26 |
+
"eval_avg_sts": 0.8149248188031465,
|
| 27 |
+
"eval_sickr_spearman": 0.7874237501893749,
|
| 28 |
+
"eval_stsb_spearman": 0.842425887416918,
|
| 29 |
+
"step": 75
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.11,
|
| 33 |
+
"eval_avg_sts": 0.8218597211542165,
|
| 34 |
+
"eval_sickr_spearman": 0.7970756000233636,
|
| 35 |
+
"eval_stsb_spearman": 0.8466438422850692,
|
| 36 |
+
"step": 100
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 0.14,
|
| 40 |
+
"eval_avg_sts": 0.8222996516785348,
|
| 41 |
+
"eval_sickr_spearman": 0.7976810320569911,
|
| 42 |
+
"eval_stsb_spearman": 0.8469182713000784,
|
| 43 |
+
"step": 125
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 0.17,
|
| 47 |
+
"eval_avg_sts": 0.8296571369885761,
|
| 48 |
+
"eval_sickr_spearman": 0.806694500521766,
|
| 49 |
+
"eval_stsb_spearman": 0.8526197734553864,
|
| 50 |
+
"step": 150
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"eval_avg_sts": 0.8336126250309825,
|
| 55 |
+
"eval_sickr_spearman": 0.8114649014854978,
|
| 56 |
+
"eval_stsb_spearman": 0.8557603485764674,
|
| 57 |
+
"step": 175
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.22,
|
| 61 |
+
"eval_avg_sts": 0.8353162227426572,
|
| 62 |
+
"eval_sickr_spearman": 0.8118504471367415,
|
| 63 |
+
"eval_stsb_spearman": 0.8587819983485729,
|
| 64 |
+
"step": 200
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.25,
|
| 68 |
+
"eval_avg_sts": 0.8300942602644428,
|
| 69 |
+
"eval_sickr_spearman": 0.8061713937960034,
|
| 70 |
+
"eval_stsb_spearman": 0.8540171267328822,
|
| 71 |
+
"step": 225
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.28,
|
| 75 |
+
"eval_avg_sts": 0.8347835912873767,
|
| 76 |
+
"eval_sickr_spearman": 0.8143648272659254,
|
| 77 |
+
"eval_stsb_spearman": 0.855202355308828,
|
| 78 |
+
"step": 250
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.3,
|
| 82 |
+
"eval_avg_sts": 0.8319994856438386,
|
| 83 |
+
"eval_sickr_spearman": 0.8115708100641689,
|
| 84 |
+
"eval_stsb_spearman": 0.8524281612235085,
|
| 85 |
+
"step": 275
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.33,
|
| 89 |
+
"eval_avg_sts": 0.8368355138029437,
|
| 90 |
+
"eval_sickr_spearman": 0.8189529021685982,
|
| 91 |
+
"eval_stsb_spearman": 0.8547181254372894,
|
| 92 |
+
"step": 300
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.36,
|
| 96 |
+
"eval_avg_sts": 0.8374901642309274,
|
| 97 |
+
"eval_sickr_spearman": 0.8199181832142,
|
| 98 |
+
"eval_stsb_spearman": 0.8550621452476548,
|
| 99 |
+
"step": 325
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.39,
|
| 103 |
+
"eval_avg_sts": 0.835477792488754,
|
| 104 |
+
"eval_sickr_spearman": 0.814654118589887,
|
| 105 |
+
"eval_stsb_spearman": 0.8563014663876211,
|
| 106 |
+
"step": 350
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.42,
|
| 110 |
+
"eval_avg_sts": 0.839369517546883,
|
| 111 |
+
"eval_sickr_spearman": 0.8205669393013246,
|
| 112 |
+
"eval_stsb_spearman": 0.8581720957924415,
|
| 113 |
+
"step": 375
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.44,
|
| 117 |
+
"eval_avg_sts": 0.8380452297674821,
|
| 118 |
+
"eval_sickr_spearman": 0.8173899220967006,
|
| 119 |
+
"eval_stsb_spearman": 0.8587005374382635,
|
| 120 |
+
"step": 400
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.47,
|
| 124 |
+
"eval_avg_sts": 0.8416927413750428,
|
| 125 |
+
"eval_sickr_spearman": 0.8239470800339302,
|
| 126 |
+
"eval_stsb_spearman": 0.8594384027161553,
|
| 127 |
+
"step": 425
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.5,
|
| 131 |
+
"eval_avg_sts": 0.8319043646289148,
|
| 132 |
+
"eval_sickr_spearman": 0.8098517769436988,
|
| 133 |
+
"eval_stsb_spearman": 0.8539569523141309,
|
| 134 |
+
"step": 450
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.53,
|
| 138 |
+
"eval_avg_sts": 0.8327259954986015,
|
| 139 |
+
"eval_sickr_spearman": 0.8118681706131722,
|
| 140 |
+
"eval_stsb_spearman": 0.8535838203840307,
|
| 141 |
+
"step": 475
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.55,
|
| 145 |
+
"learning_rate": 8.152254249815226e-06,
|
| 146 |
+
"loss": 0.4849,
|
| 147 |
+
"step": 500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.55,
|
| 151 |
+
"eval_avg_sts": 0.8398049341434459,
|
| 152 |
+
"eval_sickr_spearman": 0.8219116180171858,
|
| 153 |
+
"eval_stsb_spearman": 0.857698250269706,
|
| 154 |
+
"step": 500
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.58,
|
| 158 |
+
"eval_avg_sts": 0.8399287988815654,
|
| 159 |
+
"eval_sickr_spearman": 0.8219933669518336,
|
| 160 |
+
"eval_stsb_spearman": 0.8578642308112973,
|
| 161 |
+
"step": 525
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.61,
|
| 165 |
+
"eval_avg_sts": 0.8372186514713016,
|
| 166 |
+
"eval_sickr_spearman": 0.8191553052300584,
|
| 167 |
+
"eval_stsb_spearman": 0.8552819977125447,
|
| 168 |
+
"step": 550
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 0.64,
|
| 172 |
+
"eval_avg_sts": 0.8397984072854183,
|
| 173 |
+
"eval_sickr_spearman": 0.8222581624140619,
|
| 174 |
+
"eval_stsb_spearman": 0.8573386521567748,
|
| 175 |
+
"step": 575
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.67,
|
| 179 |
+
"eval_avg_sts": 0.8424072290367514,
|
| 180 |
+
"eval_sickr_spearman": 0.8249417080825108,
|
| 181 |
+
"eval_stsb_spearman": 0.859872749990992,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.69,
|
| 186 |
+
"eval_avg_sts": 0.8389197343591988,
|
| 187 |
+
"eval_sickr_spearman": 0.8207355284673723,
|
| 188 |
+
"eval_stsb_spearman": 0.8571039402510253,
|
| 189 |
+
"step": 625
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.72,
|
| 193 |
+
"eval_avg_sts": 0.8418254593701419,
|
| 194 |
+
"eval_sickr_spearman": 0.8221261729473099,
|
| 195 |
+
"eval_stsb_spearman": 0.8615247457929738,
|
| 196 |
+
"step": 650
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 0.75,
|
| 200 |
+
"eval_avg_sts": 0.844244457574689,
|
| 201 |
+
"eval_sickr_spearman": 0.824336131955579,
|
| 202 |
+
"eval_stsb_spearman": 0.8641527831937991,
|
| 203 |
+
"step": 675
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.78,
|
| 207 |
+
"eval_avg_sts": 0.8422699522786281,
|
| 208 |
+
"eval_sickr_spearman": 0.8214674744221877,
|
| 209 |
+
"eval_stsb_spearman": 0.8630724301350685,
|
| 210 |
+
"step": 700
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.8,
|
| 214 |
+
"eval_avg_sts": 0.8437980698256882,
|
| 215 |
+
"eval_sickr_spearman": 0.824677537024601,
|
| 216 |
+
"eval_stsb_spearman": 0.8629186026267753,
|
| 217 |
+
"step": 725
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.83,
|
| 221 |
+
"eval_avg_sts": 0.8439136197370887,
|
| 222 |
+
"eval_sickr_spearman": 0.8249638984513751,
|
| 223 |
+
"eval_stsb_spearman": 0.8628633410228024,
|
| 224 |
+
"step": 750
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.86,
|
| 228 |
+
"eval_avg_sts": 0.8426886128766484,
|
| 229 |
+
"eval_sickr_spearman": 0.8233757020512226,
|
| 230 |
+
"eval_stsb_spearman": 0.8620015237020744,
|
| 231 |
+
"step": 775
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.89,
|
| 235 |
+
"eval_avg_sts": 0.841929144469251,
|
| 236 |
+
"eval_sickr_spearman": 0.8214800585707644,
|
| 237 |
+
"eval_stsb_spearman": 0.8623782303677376,
|
| 238 |
+
"step": 800
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.91,
|
| 242 |
+
"eval_avg_sts": 0.8463814591231265,
|
| 243 |
+
"eval_sickr_spearman": 0.827185240830685,
|
| 244 |
+
"eval_stsb_spearman": 0.8655776774155678,
|
| 245 |
+
"step": 825
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.94,
|
| 249 |
+
"eval_avg_sts": 0.8382240140191579,
|
| 250 |
+
"eval_sickr_spearman": 0.8145595933822568,
|
| 251 |
+
"eval_stsb_spearman": 0.8618884346560589,
|
| 252 |
+
"step": 850
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 0.97,
|
| 256 |
+
"eval_avg_sts": 0.8402986385705378,
|
| 257 |
+
"eval_sickr_spearman": 0.818052799327648,
|
| 258 |
+
"eval_stsb_spearman": 0.8625444778134276,
|
| 259 |
+
"step": 875
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.0,
|
| 263 |
+
"eval_avg_sts": 0.8417166888579826,
|
| 264 |
+
"eval_sickr_spearman": 0.8214943238078914,
|
| 265 |
+
"eval_stsb_spearman": 0.8619390539080737,
|
| 266 |
+
"step": 900
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.03,
|
| 270 |
+
"eval_avg_sts": 0.840021113372968,
|
| 271 |
+
"eval_sickr_spearman": 0.8206462386497989,
|
| 272 |
+
"eval_stsb_spearman": 0.8593959880961373,
|
| 273 |
+
"step": 925
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.05,
|
| 277 |
+
"eval_avg_sts": 0.8407725046557779,
|
| 278 |
+
"eval_sickr_spearman": 0.8213717484370214,
|
| 279 |
+
"eval_stsb_spearman": 0.8601732608745344,
|
| 280 |
+
"step": 950
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.08,
|
| 284 |
+
"eval_avg_sts": 0.8431906318067501,
|
| 285 |
+
"eval_sickr_spearman": 0.8246691796129508,
|
| 286 |
+
"eval_stsb_spearman": 0.8617120840005495,
|
| 287 |
+
"step": 975
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.11,
|
| 291 |
+
"learning_rate": 6.304508499630452e-06,
|
| 292 |
+
"loss": 0.3038,
|
| 293 |
+
"step": 1000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.11,
|
| 297 |
+
"eval_avg_sts": 0.8416510189604846,
|
| 298 |
+
"eval_sickr_spearman": 0.822419114634981,
|
| 299 |
+
"eval_stsb_spearman": 0.8608829232859881,
|
| 300 |
+
"step": 1000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 1.14,
|
| 304 |
+
"eval_avg_sts": 0.8447615088290126,
|
| 305 |
+
"eval_sickr_spearman": 0.8269052195093008,
|
| 306 |
+
"eval_stsb_spearman": 0.8626177981487243,
|
| 307 |
+
"step": 1025
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 1.16,
|
| 311 |
+
"eval_avg_sts": 0.8424856124518805,
|
| 312 |
+
"eval_sickr_spearman": 0.8238848317264664,
|
| 313 |
+
"eval_stsb_spearman": 0.8610863931772945,
|
| 314 |
+
"step": 1050
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.19,
|
| 318 |
+
"eval_avg_sts": 0.8424277323953787,
|
| 319 |
+
"eval_sickr_spearman": 0.8237316125128787,
|
| 320 |
+
"eval_stsb_spearman": 0.8611238522778787,
|
| 321 |
+
"step": 1075
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.22,
|
| 325 |
+
"eval_avg_sts": 0.8442776434640062,
|
| 326 |
+
"eval_sickr_spearman": 0.8274288545771791,
|
| 327 |
+
"eval_stsb_spearman": 0.8611264323508335,
|
| 328 |
+
"step": 1100
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 1.25,
|
| 332 |
+
"eval_avg_sts": 0.8442605597470911,
|
| 333 |
+
"eval_sickr_spearman": 0.8269961423843231,
|
| 334 |
+
"eval_stsb_spearman": 0.8615249771098593,
|
| 335 |
+
"step": 1125
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.27,
|
| 339 |
+
"eval_avg_sts": 0.8460938684955821,
|
| 340 |
+
"eval_sickr_spearman": 0.8296871328734947,
|
| 341 |
+
"eval_stsb_spearman": 0.8625006041176696,
|
| 342 |
+
"step": 1150
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 1.3,
|
| 346 |
+
"eval_avg_sts": 0.8452328292042266,
|
| 347 |
+
"eval_sickr_spearman": 0.8290738237392314,
|
| 348 |
+
"eval_stsb_spearman": 0.8613918346692219,
|
| 349 |
+
"step": 1175
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 1.33,
|
| 353 |
+
"eval_avg_sts": 0.8443542857305267,
|
| 354 |
+
"eval_sickr_spearman": 0.8283039812453815,
|
| 355 |
+
"eval_stsb_spearman": 0.860404590215672,
|
| 356 |
+
"step": 1200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 1.36,
|
| 360 |
+
"eval_avg_sts": 0.8455779584769261,
|
| 361 |
+
"eval_sickr_spearman": 0.8298153759143346,
|
| 362 |
+
"eval_stsb_spearman": 0.8613405410395174,
|
| 363 |
+
"step": 1225
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 1.39,
|
| 367 |
+
"eval_avg_sts": 0.8465431787883896,
|
| 368 |
+
"eval_sickr_spearman": 0.8306829136685097,
|
| 369 |
+
"eval_stsb_spearman": 0.8624034439082695,
|
| 370 |
+
"step": 1250
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 1.41,
|
| 374 |
+
"eval_avg_sts": 0.8452222856659524,
|
| 375 |
+
"eval_sickr_spearman": 0.8282286684783264,
|
| 376 |
+
"eval_stsb_spearman": 0.8622159028535784,
|
| 377 |
+
"step": 1275
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.44,
|
| 381 |
+
"eval_avg_sts": 0.8460353103962084,
|
| 382 |
+
"eval_sickr_spearman": 0.8289619112728808,
|
| 383 |
+
"eval_stsb_spearman": 0.8631087095195359,
|
| 384 |
+
"step": 1300
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.47,
|
| 388 |
+
"eval_avg_sts": 0.8428046052986686,
|
| 389 |
+
"eval_sickr_spearman": 0.8233728201851361,
|
| 390 |
+
"eval_stsb_spearman": 0.862236390412201,
|
| 391 |
+
"step": 1325
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.5,
|
| 395 |
+
"eval_avg_sts": 0.8439709964632203,
|
| 396 |
+
"eval_sickr_spearman": 0.8263711616924103,
|
| 397 |
+
"eval_stsb_spearman": 0.8615708312340303,
|
| 398 |
+
"step": 1350
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.52,
|
| 402 |
+
"eval_avg_sts": 0.8464117078835496,
|
| 403 |
+
"eval_sickr_spearman": 0.8287957716930062,
|
| 404 |
+
"eval_stsb_spearman": 0.8640276440740928,
|
| 405 |
+
"step": 1375
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 1.55,
|
| 409 |
+
"eval_avg_sts": 0.8426142536649475,
|
| 410 |
+
"eval_sickr_spearman": 0.8228226239181626,
|
| 411 |
+
"eval_stsb_spearman": 0.8624058834117323,
|
| 412 |
+
"step": 1400
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 1.58,
|
| 416 |
+
"eval_avg_sts": 0.84970489436635,
|
| 417 |
+
"eval_sickr_spearman": 0.8333206856972879,
|
| 418 |
+
"eval_stsb_spearman": 0.8660891030354121,
|
| 419 |
+
"step": 1425
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.61,
|
| 423 |
+
"eval_avg_sts": 0.8459628076570953,
|
| 424 |
+
"eval_sickr_spearman": 0.8295129240685788,
|
| 425 |
+
"eval_stsb_spearman": 0.8624126912456118,
|
| 426 |
+
"step": 1450
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 1.64,
|
| 430 |
+
"eval_avg_sts": 0.8464511583090627,
|
| 431 |
+
"eval_sickr_spearman": 0.8300802674387656,
|
| 432 |
+
"eval_stsb_spearman": 0.8628220491793598,
|
| 433 |
+
"step": 1475
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 1.66,
|
| 437 |
+
"learning_rate": 4.456762749445676e-06,
|
| 438 |
+
"loss": 0.2711,
|
| 439 |
+
"step": 1500
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 1.66,
|
| 443 |
+
"eval_avg_sts": 0.8432266307908444,
|
| 444 |
+
"eval_sickr_spearman": 0.8264625168473455,
|
| 445 |
+
"eval_stsb_spearman": 0.8599907447343432,
|
| 446 |
+
"step": 1500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.69,
|
| 450 |
+
"eval_avg_sts": 0.8444286205972751,
|
| 451 |
+
"eval_sickr_spearman": 0.8294252673084542,
|
| 452 |
+
"eval_stsb_spearman": 0.859431973886096,
|
| 453 |
+
"step": 1525
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.72,
|
| 457 |
+
"eval_avg_sts": 0.8448525205916575,
|
| 458 |
+
"eval_sickr_spearman": 0.8308334911715182,
|
| 459 |
+
"eval_stsb_spearman": 0.8588715500117969,
|
| 460 |
+
"step": 1550
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.75,
|
| 464 |
+
"eval_avg_sts": 0.8411986614799187,
|
| 465 |
+
"eval_sickr_spearman": 0.8257495912086997,
|
| 466 |
+
"eval_stsb_spearman": 0.8566477317511376,
|
| 467 |
+
"step": 1575
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.77,
|
| 471 |
+
"eval_avg_sts": 0.8429029812617075,
|
| 472 |
+
"eval_sickr_spearman": 0.8264979157691056,
|
| 473 |
+
"eval_stsb_spearman": 0.8593080467543093,
|
| 474 |
+
"step": 1600
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.8,
|
| 478 |
+
"eval_avg_sts": 0.844172778094097,
|
| 479 |
+
"eval_sickr_spearman": 0.8278518644875444,
|
| 480 |
+
"eval_stsb_spearman": 0.8604936917006495,
|
| 481 |
+
"step": 1625
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 1.83,
|
| 485 |
+
"eval_avg_sts": 0.8440312504721765,
|
| 486 |
+
"eval_sickr_spearman": 0.828057053352888,
|
| 487 |
+
"eval_stsb_spearman": 0.8600054475914649,
|
| 488 |
+
"step": 1650
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.86,
|
| 492 |
+
"eval_avg_sts": 0.8466716345393903,
|
| 493 |
+
"eval_sickr_spearman": 0.8323783154870722,
|
| 494 |
+
"eval_stsb_spearman": 0.8609649535917084,
|
| 495 |
+
"step": 1675
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.88,
|
| 499 |
+
"eval_avg_sts": 0.8470752337564638,
|
| 500 |
+
"eval_sickr_spearman": 0.8329202984356999,
|
| 501 |
+
"eval_stsb_spearman": 0.8612301690772276,
|
| 502 |
+
"step": 1700
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 1.91,
|
| 506 |
+
"eval_avg_sts": 0.8491146738016275,
|
| 507 |
+
"eval_sickr_spearman": 0.8357918378351774,
|
| 508 |
+
"eval_stsb_spearman": 0.8624375097680775,
|
| 509 |
+
"step": 1725
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.94,
|
| 513 |
+
"eval_avg_sts": 0.8490639696081586,
|
| 514 |
+
"eval_sickr_spearman": 0.8353252637158076,
|
| 515 |
+
"eval_stsb_spearman": 0.8628026755005095,
|
| 516 |
+
"step": 1750
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 1.97,
|
| 520 |
+
"eval_avg_sts": 0.8490725041726443,
|
| 521 |
+
"eval_sickr_spearman": 0.8349738681776864,
|
| 522 |
+
"eval_stsb_spearman": 0.8631711401676022,
|
| 523 |
+
"step": 1775
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 2.0,
|
| 527 |
+
"eval_avg_sts": 0.849465151594995,
|
| 528 |
+
"eval_sickr_spearman": 0.8350664241101576,
|
| 529 |
+
"eval_stsb_spearman": 0.8638638790798324,
|
| 530 |
+
"step": 1800
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.02,
|
| 534 |
+
"eval_avg_sts": 0.8486443980616268,
|
| 535 |
+
"eval_sickr_spearman": 0.8342315955360619,
|
| 536 |
+
"eval_stsb_spearman": 0.8630572005871918,
|
| 537 |
+
"step": 1825
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 2.05,
|
| 541 |
+
"eval_avg_sts": 0.8476111291438189,
|
| 542 |
+
"eval_sickr_spearman": 0.8335225604166323,
|
| 543 |
+
"eval_stsb_spearman": 0.8616996978710053,
|
| 544 |
+
"step": 1850
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 2.08,
|
| 548 |
+
"eval_avg_sts": 0.8474970961473779,
|
| 549 |
+
"eval_sickr_spearman": 0.8330892238194576,
|
| 550 |
+
"eval_stsb_spearman": 0.8619049684752982,
|
| 551 |
+
"step": 1875
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 2.11,
|
| 555 |
+
"eval_avg_sts": 0.8477943008360005,
|
| 556 |
+
"eval_sickr_spearman": 0.8332226061881515,
|
| 557 |
+
"eval_stsb_spearman": 0.8623659954838494,
|
| 558 |
+
"step": 1900
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 2.13,
|
| 562 |
+
"eval_avg_sts": 0.8464807160435242,
|
| 563 |
+
"eval_sickr_spearman": 0.8316015565146151,
|
| 564 |
+
"eval_stsb_spearman": 0.8613598755724333,
|
| 565 |
+
"step": 1925
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 2.16,
|
| 569 |
+
"eval_avg_sts": 0.8462238944734057,
|
| 570 |
+
"eval_sickr_spearman": 0.831052512994076,
|
| 571 |
+
"eval_stsb_spearman": 0.8613952759527355,
|
| 572 |
+
"step": 1950
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.19,
|
| 576 |
+
"eval_avg_sts": 0.8464057290981915,
|
| 577 |
+
"eval_sickr_spearman": 0.8310086125673615,
|
| 578 |
+
"eval_stsb_spearman": 0.8618028456290214,
|
| 579 |
+
"step": 1975
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 2.22,
|
| 583 |
+
"learning_rate": 2.609016999260902e-06,
|
| 584 |
+
"loss": 0.258,
|
| 585 |
+
"step": 2000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 2.22,
|
| 589 |
+
"eval_avg_sts": 0.8468793796616104,
|
| 590 |
+
"eval_sickr_spearman": 0.8316847944134073,
|
| 591 |
+
"eval_stsb_spearman": 0.8620739649098136,
|
| 592 |
+
"step": 2000
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 2.25,
|
| 596 |
+
"eval_avg_sts": 0.8473069362456496,
|
| 597 |
+
"eval_sickr_spearman": 0.8321533858390374,
|
| 598 |
+
"eval_stsb_spearman": 0.8624604866522616,
|
| 599 |
+
"step": 2025
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 2.27,
|
| 603 |
+
"eval_avg_sts": 0.8458607867715203,
|
| 604 |
+
"eval_sickr_spearman": 0.8302502975378566,
|
| 605 |
+
"eval_stsb_spearman": 0.861471276005184,
|
| 606 |
+
"step": 2050
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 2.3,
|
| 610 |
+
"eval_avg_sts": 0.8465971265934726,
|
| 611 |
+
"eval_sickr_spearman": 0.8308937702038229,
|
| 612 |
+
"eval_stsb_spearman": 0.8623004829831225,
|
| 613 |
+
"step": 2075
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 2.33,
|
| 617 |
+
"eval_avg_sts": 0.8476369809742725,
|
| 618 |
+
"eval_sickr_spearman": 0.8326747154140468,
|
| 619 |
+
"eval_stsb_spearman": 0.8625992465344982,
|
| 620 |
+
"step": 2100
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.36,
|
| 624 |
+
"eval_avg_sts": 0.8487784986875214,
|
| 625 |
+
"eval_sickr_spearman": 0.833981881839685,
|
| 626 |
+
"eval_stsb_spearman": 0.8635751155353578,
|
| 627 |
+
"step": 2125
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 2.38,
|
| 631 |
+
"eval_avg_sts": 0.8482552739490947,
|
| 632 |
+
"eval_sickr_spearman": 0.8330946513339202,
|
| 633 |
+
"eval_stsb_spearman": 0.8634158965642693,
|
| 634 |
+
"step": 2150
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 2.41,
|
| 638 |
+
"eval_avg_sts": 0.8459693374985795,
|
| 639 |
+
"eval_sickr_spearman": 0.8297586992146375,
|
| 640 |
+
"eval_stsb_spearman": 0.8621799757825214,
|
| 641 |
+
"step": 2175
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 2.44,
|
| 645 |
+
"eval_avg_sts": 0.8476635726729035,
|
| 646 |
+
"eval_sickr_spearman": 0.8323805729488398,
|
| 647 |
+
"eval_stsb_spearman": 0.8629465723969671,
|
| 648 |
+
"step": 2200
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.47,
|
| 652 |
+
"eval_avg_sts": 0.8479733460637723,
|
| 653 |
+
"eval_sickr_spearman": 0.8331660735817588,
|
| 654 |
+
"eval_stsb_spearman": 0.8627806185457858,
|
| 655 |
+
"step": 2225
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 2.49,
|
| 659 |
+
"eval_avg_sts": 0.8473017297750429,
|
| 660 |
+
"eval_sickr_spearman": 0.8320290333174142,
|
| 661 |
+
"eval_stsb_spearman": 0.8625744262326718,
|
| 662 |
+
"step": 2250
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.52,
|
| 666 |
+
"eval_avg_sts": 0.8479177356313778,
|
| 667 |
+
"eval_sickr_spearman": 0.8327113151133426,
|
| 668 |
+
"eval_stsb_spearman": 0.863124156149413,
|
| 669 |
+
"step": 2275
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.55,
|
| 673 |
+
"eval_avg_sts": 0.8475060767928835,
|
| 674 |
+
"eval_sickr_spearman": 0.8321914264713766,
|
| 675 |
+
"eval_stsb_spearman": 0.8628207271143905,
|
| 676 |
+
"step": 2300
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 2.58,
|
| 680 |
+
"eval_avg_sts": 0.8471482251359982,
|
| 681 |
+
"eval_sickr_spearman": 0.831474706375717,
|
| 682 |
+
"eval_stsb_spearman": 0.8628217438962794,
|
| 683 |
+
"step": 2325
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 2.61,
|
| 687 |
+
"eval_avg_sts": 0.8471555097744191,
|
| 688 |
+
"eval_sickr_spearman": 0.8313104880399002,
|
| 689 |
+
"eval_stsb_spearman": 0.8630005315089381,
|
| 690 |
+
"step": 2350
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 2.63,
|
| 694 |
+
"eval_avg_sts": 0.8463330579657068,
|
| 695 |
+
"eval_sickr_spearman": 0.8301000562525582,
|
| 696 |
+
"eval_stsb_spearman": 0.8625660596788554,
|
| 697 |
+
"step": 2375
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.66,
|
| 701 |
+
"eval_avg_sts": 0.8462206147101088,
|
| 702 |
+
"eval_sickr_spearman": 0.8299467890078692,
|
| 703 |
+
"eval_stsb_spearman": 0.8624944404123482,
|
| 704 |
+
"step": 2400
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.69,
|
| 708 |
+
"eval_avg_sts": 0.8466291774217871,
|
| 709 |
+
"eval_sickr_spearman": 0.8307020300468819,
|
| 710 |
+
"eval_stsb_spearman": 0.8625563247966924,
|
| 711 |
+
"step": 2425
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.72,
|
| 715 |
+
"eval_avg_sts": 0.846255947089394,
|
| 716 |
+
"eval_sickr_spearman": 0.8300505801238202,
|
| 717 |
+
"eval_stsb_spearman": 0.8624613140549677,
|
| 718 |
+
"step": 2450
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 2.74,
|
| 722 |
+
"eval_avg_sts": 0.84652808904023,
|
| 723 |
+
"eval_sickr_spearman": 0.830373449281944,
|
| 724 |
+
"eval_stsb_spearman": 0.8626827287985159,
|
| 725 |
+
"step": 2475
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 2.77,
|
| 729 |
+
"learning_rate": 7.612712490761272e-07,
|
| 730 |
+
"loss": 0.2459,
|
| 731 |
+
"step": 2500
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 2.77,
|
| 735 |
+
"eval_avg_sts": 0.8462118162645162,
|
| 736 |
+
"eval_sickr_spearman": 0.830052889710946,
|
| 737 |
+
"eval_stsb_spearman": 0.8623707428180863,
|
| 738 |
+
"step": 2500
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 2.8,
|
| 742 |
+
"eval_avg_sts": 0.8459259791683191,
|
| 743 |
+
"eval_sickr_spearman": 0.8294702724505018,
|
| 744 |
+
"eval_stsb_spearman": 0.8623816858861365,
|
| 745 |
+
"step": 2525
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.83,
|
| 749 |
+
"eval_avg_sts": 0.8464449438786035,
|
| 750 |
+
"eval_sickr_spearman": 0.8304476093025643,
|
| 751 |
+
"eval_stsb_spearman": 0.8624422784546426,
|
| 752 |
+
"step": 2550
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.85,
|
| 756 |
+
"eval_avg_sts": 0.84648379412736,
|
| 757 |
+
"eval_sickr_spearman": 0.8305015962605808,
|
| 758 |
+
"eval_stsb_spearman": 0.8624659919941392,
|
| 759 |
+
"step": 2575
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 2.88,
|
| 763 |
+
"eval_avg_sts": 0.8465871172906757,
|
| 764 |
+
"eval_sickr_spearman": 0.8306914151734642,
|
| 765 |
+
"eval_stsb_spearman": 0.8624828194078872,
|
| 766 |
+
"step": 2600
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 2.91,
|
| 770 |
+
"eval_avg_sts": 0.8464858615057194,
|
| 771 |
+
"eval_sickr_spearman": 0.8305917026068788,
|
| 772 |
+
"eval_stsb_spearman": 0.8623800204045601,
|
| 773 |
+
"step": 2625
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 2.94,
|
| 777 |
+
"eval_avg_sts": 0.8466908586012467,
|
| 778 |
+
"eval_sickr_spearman": 0.8309179778789479,
|
| 779 |
+
"eval_stsb_spearman": 0.8624637393235455,
|
| 780 |
+
"step": 2650
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 2.97,
|
| 784 |
+
"eval_avg_sts": 0.8467372932579469,
|
| 785 |
+
"eval_sickr_spearman": 0.830980274217513,
|
| 786 |
+
"eval_stsb_spearman": 0.8624943122983809,
|
| 787 |
+
"step": 2675
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 2.99,
|
| 791 |
+
"eval_avg_sts": 0.8467338856246951,
|
| 792 |
+
"eval_sickr_spearman": 0.8309889198157718,
|
| 793 |
+
"eval_stsb_spearman": 0.8624788514336185,
|
| 794 |
+
"step": 2700
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.0,
|
| 798 |
+
"step": 2706,
|
| 799 |
+
"train_runtime": 7555.7932,
|
| 800 |
+
"train_samples_per_second": 0.358
|
| 801 |
+
}
|
| 802 |
+
],
|
| 803 |
+
"max_steps": 2706,
|
| 804 |
+
"num_train_epochs": 3,
|
| 805 |
+
"total_flos": 292973872557019392,
|
| 806 |
+
"trial_name": null,
|
| 807 |
+
"trial_params": null
|
| 808 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9c01b60dcba94f2808e43bc3bce00802f87c2bb2857f90a2373fa1399a4cb35
|
| 3 |
+
size 2552
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|