Upload 12 files
Browse files- config.json +26 -0
- eval_results.txt +12 -0
- merges.txt +0 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- train_results.txt +3 -0
- trainer_state.json +467 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "result/my-sup-simcse-roberta-base_filtered_final_nli_pos_neg",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"pad_token_id": 1,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"transformers_version": "4.2.1",
|
| 23 |
+
"type_vocab_size": 1,
|
| 24 |
+
"use_cache": true,
|
| 25 |
+
"vocab_size": 50265
|
| 26 |
+
}
|
eval_results.txt
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch = 3.0
|
| 2 |
+
eval_CR = 91.9
|
| 3 |
+
eval_MPQA = 89.35
|
| 4 |
+
eval_MR = 84.88
|
| 5 |
+
eval_MRPC = 76.79
|
| 6 |
+
eval_SST2 = 88.99
|
| 7 |
+
eval_SUBJ = 93.27
|
| 8 |
+
eval_TREC = 79.66
|
| 9 |
+
eval_avg_sts = 0.8509230445807242
|
| 10 |
+
eval_avg_transfer = 86.40571428571427
|
| 11 |
+
eval_sickr_spearman = 0.8376631295472063
|
| 12 |
+
eval_stsb_spearman = 0.8641829596142421
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51512a6ffb4b5a5f49089b97f57133580e6e695e87dea5fcae1959d28592c87e
|
| 3 |
+
size 1311252549
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea7663f93ecfb7b5b58bf622ec51634f5280e32657a094b78f1f02c8433d9622
|
| 3 |
+
size 655636485
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9a1df5831574c6c22d13333cb29f582178639327101c0fa69ab51eea37843e37
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "model_max_length": 512, "name_or_path": "roberta-base"}
|
train_results.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch = 3.0
|
| 2 |
+
train_runtime = 2217.7106
|
| 3 |
+
train_samples_per_second = 0.695
|
trainer_state.json
ADDED
|
@@ -0,0 +1,467 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8509230445807242,
|
| 3 |
+
"best_model_checkpoint": "result/my-sup-simcse-roberta-base_filtered_final_nli_pos_neg",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 1542,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.05,
|
| 12 |
+
"eval_avg_sts": 0.7970172556800076,
|
| 13 |
+
"eval_sickr_spearman": 0.75025416187497,
|
| 14 |
+
"eval_stsb_spearman": 0.8437803494850452,
|
| 15 |
+
"step": 25
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.1,
|
| 19 |
+
"eval_avg_sts": 0.8184611939888191,
|
| 20 |
+
"eval_sickr_spearman": 0.787718901307712,
|
| 21 |
+
"eval_stsb_spearman": 0.8492034866699262,
|
| 22 |
+
"step": 50
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"epoch": 0.15,
|
| 26 |
+
"eval_avg_sts": 0.82628745175756,
|
| 27 |
+
"eval_sickr_spearman": 0.7880150610791795,
|
| 28 |
+
"eval_stsb_spearman": 0.8645598424359405,
|
| 29 |
+
"step": 75
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.19,
|
| 33 |
+
"eval_avg_sts": 0.8214663219382123,
|
| 34 |
+
"eval_sickr_spearman": 0.7863345008709616,
|
| 35 |
+
"eval_stsb_spearman": 0.856598143005463,
|
| 36 |
+
"step": 100
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 0.24,
|
| 40 |
+
"eval_avg_sts": 0.8245359107619952,
|
| 41 |
+
"eval_sickr_spearman": 0.7852220525305533,
|
| 42 |
+
"eval_stsb_spearman": 0.8638497689934371,
|
| 43 |
+
"step": 125
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 0.29,
|
| 47 |
+
"eval_avg_sts": 0.8369552464117445,
|
| 48 |
+
"eval_sickr_spearman": 0.8047760902992264,
|
| 49 |
+
"eval_stsb_spearman": 0.8691344025242626,
|
| 50 |
+
"step": 150
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.34,
|
| 54 |
+
"eval_avg_sts": 0.8347841673419429,
|
| 55 |
+
"eval_sickr_spearman": 0.8065769684165468,
|
| 56 |
+
"eval_stsb_spearman": 0.862991366267339,
|
| 57 |
+
"step": 175
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.39,
|
| 61 |
+
"eval_avg_sts": 0.8365894758739485,
|
| 62 |
+
"eval_sickr_spearman": 0.807168999772873,
|
| 63 |
+
"eval_stsb_spearman": 0.866009951975024,
|
| 64 |
+
"step": 200
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.44,
|
| 68 |
+
"eval_avg_sts": 0.8373694984089681,
|
| 69 |
+
"eval_sickr_spearman": 0.8121874333444314,
|
| 70 |
+
"eval_stsb_spearman": 0.8625515634735048,
|
| 71 |
+
"step": 225
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.49,
|
| 75 |
+
"eval_avg_sts": 0.8353378734135991,
|
| 76 |
+
"eval_sickr_spearman": 0.8028517722512096,
|
| 77 |
+
"eval_stsb_spearman": 0.8678239745759886,
|
| 78 |
+
"step": 250
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.54,
|
| 82 |
+
"eval_avg_sts": 0.8421542038447196,
|
| 83 |
+
"eval_sickr_spearman": 0.8175947747443341,
|
| 84 |
+
"eval_stsb_spearman": 0.8667136329451051,
|
| 85 |
+
"step": 275
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.58,
|
| 89 |
+
"eval_avg_sts": 0.8360416637090009,
|
| 90 |
+
"eval_sickr_spearman": 0.8098268968331538,
|
| 91 |
+
"eval_stsb_spearman": 0.8622564305848479,
|
| 92 |
+
"step": 300
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.63,
|
| 96 |
+
"eval_avg_sts": 0.8432321348956329,
|
| 97 |
+
"eval_sickr_spearman": 0.8172194597176966,
|
| 98 |
+
"eval_stsb_spearman": 0.8692448100735692,
|
| 99 |
+
"step": 325
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.68,
|
| 103 |
+
"eval_avg_sts": 0.8406118882697353,
|
| 104 |
+
"eval_sickr_spearman": 0.8133721684925038,
|
| 105 |
+
"eval_stsb_spearman": 0.8678516080469669,
|
| 106 |
+
"step": 350
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.73,
|
| 110 |
+
"eval_avg_sts": 0.8364168982681424,
|
| 111 |
+
"eval_sickr_spearman": 0.8088663228354931,
|
| 112 |
+
"eval_stsb_spearman": 0.8639674737007915,
|
| 113 |
+
"step": 375
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.78,
|
| 117 |
+
"eval_avg_sts": 0.8411726277287294,
|
| 118 |
+
"eval_sickr_spearman": 0.8214387037924262,
|
| 119 |
+
"eval_stsb_spearman": 0.8609065516650325,
|
| 120 |
+
"step": 400
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.83,
|
| 124 |
+
"eval_avg_sts": 0.8446725188995174,
|
| 125 |
+
"eval_sickr_spearman": 0.824545739682255,
|
| 126 |
+
"eval_stsb_spearman": 0.86479929811678,
|
| 127 |
+
"step": 425
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.88,
|
| 131 |
+
"eval_avg_sts": 0.846832980887532,
|
| 132 |
+
"eval_sickr_spearman": 0.8305859869058076,
|
| 133 |
+
"eval_stsb_spearman": 0.8630799748692564,
|
| 134 |
+
"step": 450
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.92,
|
| 138 |
+
"eval_avg_sts": 0.8417173844571794,
|
| 139 |
+
"eval_sickr_spearman": 0.8217314533556915,
|
| 140 |
+
"eval_stsb_spearman": 0.8617033155586672,
|
| 141 |
+
"step": 475
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.97,
|
| 145 |
+
"learning_rate": 3.378728923476006e-05,
|
| 146 |
+
"loss": 0.5309,
|
| 147 |
+
"step": 500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.97,
|
| 151 |
+
"eval_avg_sts": 0.8394389881737694,
|
| 152 |
+
"eval_sickr_spearman": 0.8176411247572218,
|
| 153 |
+
"eval_stsb_spearman": 0.8612368515903169,
|
| 154 |
+
"step": 500
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.02,
|
| 158 |
+
"eval_avg_sts": 0.8424515638291671,
|
| 159 |
+
"eval_sickr_spearman": 0.8244119250336484,
|
| 160 |
+
"eval_stsb_spearman": 0.8604912026246858,
|
| 161 |
+
"step": 525
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.07,
|
| 165 |
+
"eval_avg_sts": 0.843505143147442,
|
| 166 |
+
"eval_sickr_spearman": 0.8292789645734737,
|
| 167 |
+
"eval_stsb_spearman": 0.8577313217214102,
|
| 168 |
+
"step": 550
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.12,
|
| 172 |
+
"eval_avg_sts": 0.843100574356599,
|
| 173 |
+
"eval_sickr_spearman": 0.8253128924344243,
|
| 174 |
+
"eval_stsb_spearman": 0.8608882562787737,
|
| 175 |
+
"step": 575
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.17,
|
| 179 |
+
"eval_avg_sts": 0.8431532651801806,
|
| 180 |
+
"eval_sickr_spearman": 0.8231387646278283,
|
| 181 |
+
"eval_stsb_spearman": 0.863167765732533,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.22,
|
| 186 |
+
"eval_avg_sts": 0.8411549297525207,
|
| 187 |
+
"eval_sickr_spearman": 0.8223049447068626,
|
| 188 |
+
"eval_stsb_spearman": 0.8600049147981788,
|
| 189 |
+
"step": 625
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.26,
|
| 193 |
+
"eval_avg_sts": 0.8444909588234402,
|
| 194 |
+
"eval_sickr_spearman": 0.8264434965311761,
|
| 195 |
+
"eval_stsb_spearman": 0.8625384211157042,
|
| 196 |
+
"step": 650
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.31,
|
| 200 |
+
"eval_avg_sts": 0.8442403156760934,
|
| 201 |
+
"eval_sickr_spearman": 0.8282549414908131,
|
| 202 |
+
"eval_stsb_spearman": 0.8602256898613736,
|
| 203 |
+
"step": 675
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.36,
|
| 207 |
+
"eval_avg_sts": 0.8433546951723836,
|
| 208 |
+
"eval_sickr_spearman": 0.8271446545499697,
|
| 209 |
+
"eval_stsb_spearman": 0.8595647357947974,
|
| 210 |
+
"step": 700
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.41,
|
| 214 |
+
"eval_avg_sts": 0.8420306309738643,
|
| 215 |
+
"eval_sickr_spearman": 0.8233070175761659,
|
| 216 |
+
"eval_stsb_spearman": 0.8607542443715627,
|
| 217 |
+
"step": 725
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.46,
|
| 221 |
+
"eval_avg_sts": 0.8391026032092552,
|
| 222 |
+
"eval_sickr_spearman": 0.8208057019065734,
|
| 223 |
+
"eval_stsb_spearman": 0.8573995045119369,
|
| 224 |
+
"step": 750
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.51,
|
| 228 |
+
"eval_avg_sts": 0.8413993052906461,
|
| 229 |
+
"eval_sickr_spearman": 0.8210936963907962,
|
| 230 |
+
"eval_stsb_spearman": 0.8617049141904959,
|
| 231 |
+
"step": 775
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.56,
|
| 235 |
+
"eval_avg_sts": 0.8434142809317196,
|
| 236 |
+
"eval_sickr_spearman": 0.8264181361096167,
|
| 237 |
+
"eval_stsb_spearman": 0.8604104257538223,
|
| 238 |
+
"step": 800
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.61,
|
| 242 |
+
"eval_avg_sts": 0.8445746067631623,
|
| 243 |
+
"eval_sickr_spearman": 0.8273452804606768,
|
| 244 |
+
"eval_stsb_spearman": 0.8618039330656478,
|
| 245 |
+
"step": 825
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.65,
|
| 249 |
+
"eval_avg_sts": 0.8429633955631959,
|
| 250 |
+
"eval_sickr_spearman": 0.8270336546745461,
|
| 251 |
+
"eval_stsb_spearman": 0.8588931364518456,
|
| 252 |
+
"step": 850
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.7,
|
| 256 |
+
"eval_avg_sts": 0.8472516730528594,
|
| 257 |
+
"eval_sickr_spearman": 0.8334014259788056,
|
| 258 |
+
"eval_stsb_spearman": 0.8611019201269132,
|
| 259 |
+
"step": 875
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.75,
|
| 263 |
+
"eval_avg_sts": 0.8509230445807242,
|
| 264 |
+
"eval_sickr_spearman": 0.8376631295472063,
|
| 265 |
+
"eval_stsb_spearman": 0.8641829596142421,
|
| 266 |
+
"step": 900
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.8,
|
| 270 |
+
"eval_avg_sts": 0.8453984948023106,
|
| 271 |
+
"eval_sickr_spearman": 0.82835503830621,
|
| 272 |
+
"eval_stsb_spearman": 0.8624419512984111,
|
| 273 |
+
"step": 925
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.85,
|
| 277 |
+
"eval_avg_sts": 0.8441717460262391,
|
| 278 |
+
"eval_sickr_spearman": 0.8250034760789602,
|
| 279 |
+
"eval_stsb_spearman": 0.8633400159735181,
|
| 280 |
+
"step": 950
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.9,
|
| 284 |
+
"eval_avg_sts": 0.8480020725945073,
|
| 285 |
+
"eval_sickr_spearman": 0.8306723468261933,
|
| 286 |
+
"eval_stsb_spearman": 0.8653317983628214,
|
| 287 |
+
"step": 975
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.95,
|
| 291 |
+
"learning_rate": 1.7574578469520103e-05,
|
| 292 |
+
"loss": 0.2714,
|
| 293 |
+
"step": 1000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.95,
|
| 297 |
+
"eval_avg_sts": 0.8467505999509635,
|
| 298 |
+
"eval_sickr_spearman": 0.8330692909123609,
|
| 299 |
+
"eval_stsb_spearman": 0.8604319089895659,
|
| 300 |
+
"step": 1000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 1.99,
|
| 304 |
+
"eval_avg_sts": 0.8494326339830918,
|
| 305 |
+
"eval_sickr_spearman": 0.8343606551056261,
|
| 306 |
+
"eval_stsb_spearman": 0.8645046128605576,
|
| 307 |
+
"step": 1025
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 2.04,
|
| 311 |
+
"eval_avg_sts": 0.8452826795684555,
|
| 312 |
+
"eval_sickr_spearman": 0.8291940455861312,
|
| 313 |
+
"eval_stsb_spearman": 0.8613713135507798,
|
| 314 |
+
"step": 1050
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.09,
|
| 318 |
+
"eval_avg_sts": 0.8457791155470658,
|
| 319 |
+
"eval_sickr_spearman": 0.8285575854609747,
|
| 320 |
+
"eval_stsb_spearman": 0.8630006456331568,
|
| 321 |
+
"step": 1075
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.14,
|
| 325 |
+
"eval_avg_sts": 0.8447554891917526,
|
| 326 |
+
"eval_sickr_spearman": 0.8279863996026727,
|
| 327 |
+
"eval_stsb_spearman": 0.8615245787808324,
|
| 328 |
+
"step": 1100
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.19,
|
| 332 |
+
"eval_avg_sts": 0.847764082068214,
|
| 333 |
+
"eval_sickr_spearman": 0.8317096745239522,
|
| 334 |
+
"eval_stsb_spearman": 0.8638184896124759,
|
| 335 |
+
"step": 1125
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 2.24,
|
| 339 |
+
"eval_avg_sts": 0.8451478500577279,
|
| 340 |
+
"eval_sickr_spearman": 0.828890681149448,
|
| 341 |
+
"eval_stsb_spearman": 0.8614050189660079,
|
| 342 |
+
"step": 1150
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 2.29,
|
| 346 |
+
"eval_avg_sts": 0.8439050801814342,
|
| 347 |
+
"eval_sickr_spearman": 0.8240606735888313,
|
| 348 |
+
"eval_stsb_spearman": 0.8637494867740371,
|
| 349 |
+
"step": 1175
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 2.33,
|
| 353 |
+
"eval_avg_sts": 0.8447525482867402,
|
| 354 |
+
"eval_sickr_spearman": 0.8263380682635195,
|
| 355 |
+
"eval_stsb_spearman": 0.863167028309961,
|
| 356 |
+
"step": 1200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.38,
|
| 360 |
+
"eval_avg_sts": 0.8455355460641325,
|
| 361 |
+
"eval_sickr_spearman": 0.8271100241258328,
|
| 362 |
+
"eval_stsb_spearman": 0.8639610680024322,
|
| 363 |
+
"step": 1225
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 2.43,
|
| 367 |
+
"eval_avg_sts": 0.8467670558910284,
|
| 368 |
+
"eval_sickr_spearman": 0.8293558143357747,
|
| 369 |
+
"eval_stsb_spearman": 0.8641782974462822,
|
| 370 |
+
"step": 1250
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 2.48,
|
| 374 |
+
"eval_avg_sts": 0.8438803647863737,
|
| 375 |
+
"eval_sickr_spearman": 0.8270033470495388,
|
| 376 |
+
"eval_stsb_spearman": 0.8607573825232087,
|
| 377 |
+
"step": 1275
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.53,
|
| 381 |
+
"eval_avg_sts": 0.8454224221400783,
|
| 382 |
+
"eval_sickr_spearman": 0.8282425974977434,
|
| 383 |
+
"eval_stsb_spearman": 0.8626022467824133,
|
| 384 |
+
"step": 1300
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.58,
|
| 388 |
+
"eval_avg_sts": 0.8449894067547106,
|
| 389 |
+
"eval_sickr_spearman": 0.8266484932721141,
|
| 390 |
+
"eval_stsb_spearman": 0.8633303202373069,
|
| 391 |
+
"step": 1325
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 2.63,
|
| 395 |
+
"eval_avg_sts": 0.8460181390370703,
|
| 396 |
+
"eval_sickr_spearman": 0.8278362063484757,
|
| 397 |
+
"eval_stsb_spearman": 0.864200071725665,
|
| 398 |
+
"step": 1350
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.68,
|
| 402 |
+
"eval_avg_sts": 0.8465150060489774,
|
| 403 |
+
"eval_sickr_spearman": 0.8285329455059368,
|
| 404 |
+
"eval_stsb_spearman": 0.8644970665920182,
|
| 405 |
+
"step": 1375
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.72,
|
| 409 |
+
"eval_avg_sts": 0.8471917941480929,
|
| 410 |
+
"eval_sickr_spearman": 0.8293239697155211,
|
| 411 |
+
"eval_stsb_spearman": 0.8650596185806647,
|
| 412 |
+
"step": 1400
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 2.77,
|
| 416 |
+
"eval_avg_sts": 0.8469228159218247,
|
| 417 |
+
"eval_sickr_spearman": 0.8292969282054116,
|
| 418 |
+
"eval_stsb_spearman": 0.8645487036382378,
|
| 419 |
+
"step": 1425
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.82,
|
| 423 |
+
"eval_avg_sts": 0.846386304453965,
|
| 424 |
+
"eval_sickr_spearman": 0.8283193992289432,
|
| 425 |
+
"eval_stsb_spearman": 0.864453209678987,
|
| 426 |
+
"step": 1450
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 2.87,
|
| 430 |
+
"eval_avg_sts": 0.8456684357631843,
|
| 431 |
+
"eval_sickr_spearman": 0.8282710799408963,
|
| 432 |
+
"eval_stsb_spearman": 0.8630657915854721,
|
| 433 |
+
"step": 1475
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 2.92,
|
| 437 |
+
"learning_rate": 1.3618677042801557e-06,
|
| 438 |
+
"loss": 0.2376,
|
| 439 |
+
"step": 1500
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 2.92,
|
| 443 |
+
"eval_avg_sts": 0.846184384136446,
|
| 444 |
+
"eval_sickr_spearman": 0.8285680082099867,
|
| 445 |
+
"eval_stsb_spearman": 0.8638007600629054,
|
| 446 |
+
"step": 1500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.97,
|
| 450 |
+
"eval_avg_sts": 0.8465883297884287,
|
| 451 |
+
"eval_sickr_spearman": 0.8287545610079723,
|
| 452 |
+
"eval_stsb_spearman": 0.864422098568885,
|
| 453 |
+
"step": 1525
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 3.0,
|
| 457 |
+
"step": 1542,
|
| 458 |
+
"train_runtime": 2217.7106,
|
| 459 |
+
"train_samples_per_second": 0.695
|
| 460 |
+
}
|
| 461 |
+
],
|
| 462 |
+
"max_steps": 1542,
|
| 463 |
+
"num_train_epochs": 3,
|
| 464 |
+
"total_flos": 148906707659089920,
|
| 465 |
+
"trial_name": null,
|
| 466 |
+
"trial_params": null
|
| 467 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e08a8d638f45fa7e63fc551188ed21719d4b8b2c111e8cb991746f5ba26e551
|
| 3 |
+
size 2107
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|