Junlei
commited on
Commit
•
59dc35f
1
Parent(s):
cedee7a
init
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 +488 -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_augx_0517_275578",
|
| 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 = 92.07
|
| 3 |
+
eval_MPQA = 88.32
|
| 4 |
+
eval_MR = 84.87
|
| 5 |
+
eval_MRPC = 76.42
|
| 6 |
+
eval_SST2 = 88.99
|
| 7 |
+
eval_SUBJ = 92.84
|
| 8 |
+
eval_TREC = 72.6
|
| 9 |
+
eval_avg_sts = 0.8287429760185343
|
| 10 |
+
eval_avg_transfer = 85.15857142857142
|
| 11 |
+
eval_sickr_spearman = 0.7950705897249312
|
| 12 |
+
eval_stsb_spearman = 0.8624153623121376
|
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 sha256:336df629ac9b7504a9e3dc005f36cae5490292467141114edc1dbbd739c045c1
|
| 3 |
+
size 1311252549
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:835146e3b7b66259613bd987e6c8d98bd383833b9c2220e943b390a124845fef
|
| 3 |
+
size 655636485
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:634cc3bb793f313fce33fb84c1a6bd12ea7325fb0baa13b5526fa7326c10005e
|
| 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 = 2320.0614
|
| 3 |
+
train_samples_per_second = 0.697
|
trainer_state.json
ADDED
|
@@ -0,0 +1,488 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8287429760185343,
|
| 3 |
+
"best_model_checkpoint": "result/my-sup-simcse-roberta-base_filtered_final_augx_0517_275578",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 1617,
|
| 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": 0.7923727821060924,
|
| 13 |
+
"eval_sickr_spearman": 0.7589401062590382,
|
| 14 |
+
"eval_stsb_spearman": 0.8258054579531467,
|
| 15 |
+
"step": 25
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.09,
|
| 19 |
+
"eval_avg_sts": 0.788369556206714,
|
| 20 |
+
"eval_sickr_spearman": 0.7429059316970545,
|
| 21 |
+
"eval_stsb_spearman": 0.8338331807163735,
|
| 22 |
+
"step": 50
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"epoch": 0.14,
|
| 26 |
+
"eval_avg_sts": 0.8104646483652821,
|
| 27 |
+
"eval_sickr_spearman": 0.7750206308339096,
|
| 28 |
+
"eval_stsb_spearman": 0.8459086658966546,
|
| 29 |
+
"step": 75
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.19,
|
| 33 |
+
"eval_avg_sts": 0.8156440464261943,
|
| 34 |
+
"eval_sickr_spearman": 0.7790718621158088,
|
| 35 |
+
"eval_stsb_spearman": 0.8522162307365796,
|
| 36 |
+
"step": 100
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 0.23,
|
| 40 |
+
"eval_avg_sts": 0.8158874725367419,
|
| 41 |
+
"eval_sickr_spearman": 0.7767967729339896,
|
| 42 |
+
"eval_stsb_spearman": 0.8549781721394941,
|
| 43 |
+
"step": 125
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 0.28,
|
| 47 |
+
"eval_avg_sts": 0.8226456308827159,
|
| 48 |
+
"eval_sickr_spearman": 0.7900410609689407,
|
| 49 |
+
"eval_stsb_spearman": 0.8552502007964913,
|
| 50 |
+
"step": 150
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.32,
|
| 54 |
+
"eval_avg_sts": 0.8233733407465815,
|
| 55 |
+
"eval_sickr_spearman": 0.7881731314340124,
|
| 56 |
+
"eval_stsb_spearman": 0.8585735500591508,
|
| 57 |
+
"step": 175
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.37,
|
| 61 |
+
"eval_avg_sts": 0.8246758200303745,
|
| 62 |
+
"eval_sickr_spearman": 0.7910648919271954,
|
| 63 |
+
"eval_stsb_spearman": 0.8582867481335537,
|
| 64 |
+
"step": 200
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.42,
|
| 68 |
+
"eval_avg_sts": 0.8127837390967988,
|
| 69 |
+
"eval_sickr_spearman": 0.7683723098661505,
|
| 70 |
+
"eval_stsb_spearman": 0.8571951683274471,
|
| 71 |
+
"step": 225
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.46,
|
| 75 |
+
"eval_avg_sts": 0.8211001358940954,
|
| 76 |
+
"eval_sickr_spearman": 0.7857309900813916,
|
| 77 |
+
"eval_stsb_spearman": 0.8564692817067991,
|
| 78 |
+
"step": 250
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.51,
|
| 82 |
+
"eval_avg_sts": 0.8231415754884768,
|
| 83 |
+
"eval_sickr_spearman": 0.7870932481803794,
|
| 84 |
+
"eval_stsb_spearman": 0.859189902796574,
|
| 85 |
+
"step": 275
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 0.56,
|
| 89 |
+
"eval_avg_sts": 0.8281698148839484,
|
| 90 |
+
"eval_sickr_spearman": 0.7933201442641206,
|
| 91 |
+
"eval_stsb_spearman": 0.8630194855037762,
|
| 92 |
+
"step": 300
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.6,
|
| 96 |
+
"eval_avg_sts": 0.8261624892170372,
|
| 97 |
+
"eval_sickr_spearman": 0.7934045829404487,
|
| 98 |
+
"eval_stsb_spearman": 0.8589203954936255,
|
| 99 |
+
"step": 325
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.65,
|
| 103 |
+
"eval_avg_sts": 0.8223223884475288,
|
| 104 |
+
"eval_sickr_spearman": 0.7860610117793729,
|
| 105 |
+
"eval_stsb_spearman": 0.8585837651156847,
|
| 106 |
+
"step": 350
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.7,
|
| 110 |
+
"eval_avg_sts": 0.8239117966986287,
|
| 111 |
+
"eval_sickr_spearman": 0.7878053572903008,
|
| 112 |
+
"eval_stsb_spearman": 0.8600182361069565,
|
| 113 |
+
"step": 375
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.74,
|
| 117 |
+
"eval_avg_sts": 0.8189825828946973,
|
| 118 |
+
"eval_sickr_spearman": 0.7824180448596978,
|
| 119 |
+
"eval_stsb_spearman": 0.8555471209296966,
|
| 120 |
+
"step": 400
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.79,
|
| 124 |
+
"eval_avg_sts": 0.8284554395260346,
|
| 125 |
+
"eval_sickr_spearman": 0.7938278330063215,
|
| 126 |
+
"eval_stsb_spearman": 0.8630830460457476,
|
| 127 |
+
"step": 425
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.83,
|
| 131 |
+
"eval_avg_sts": 0.8287429760185343,
|
| 132 |
+
"eval_sickr_spearman": 0.7950705897249312,
|
| 133 |
+
"eval_stsb_spearman": 0.8624153623121376,
|
| 134 |
+
"step": 450
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.88,
|
| 138 |
+
"eval_avg_sts": 0.8175086199936331,
|
| 139 |
+
"eval_sickr_spearman": 0.7783943353989228,
|
| 140 |
+
"eval_stsb_spearman": 0.8566229045883434,
|
| 141 |
+
"step": 475
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.93,
|
| 145 |
+
"learning_rate": 3.453927025355597e-05,
|
| 146 |
+
"loss": 0.4533,
|
| 147 |
+
"step": 500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.93,
|
| 151 |
+
"eval_avg_sts": 0.8230793846551405,
|
| 152 |
+
"eval_sickr_spearman": 0.791305047434386,
|
| 153 |
+
"eval_stsb_spearman": 0.8548537218758949,
|
| 154 |
+
"step": 500
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.97,
|
| 158 |
+
"eval_avg_sts": 0.813507998872834,
|
| 159 |
+
"eval_sickr_spearman": 0.7710531177618173,
|
| 160 |
+
"eval_stsb_spearman": 0.8559628799838506,
|
| 161 |
+
"step": 525
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.02,
|
| 165 |
+
"eval_avg_sts": 0.8152806946649198,
|
| 166 |
+
"eval_sickr_spearman": 0.7748347024402428,
|
| 167 |
+
"eval_stsb_spearman": 0.8557266868895969,
|
| 168 |
+
"step": 550
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.07,
|
| 172 |
+
"eval_avg_sts": 0.8193593042077025,
|
| 173 |
+
"eval_sickr_spearman": 0.7769575330305031,
|
| 174 |
+
"eval_stsb_spearman": 0.8617610753849019,
|
| 175 |
+
"step": 575
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.11,
|
| 179 |
+
"eval_avg_sts": 0.8122703001838674,
|
| 180 |
+
"eval_sickr_spearman": 0.7715038416177126,
|
| 181 |
+
"eval_stsb_spearman": 0.8530367587500223,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.16,
|
| 186 |
+
"eval_avg_sts": 0.8236420464403356,
|
| 187 |
+
"eval_sickr_spearman": 0.7849814167123483,
|
| 188 |
+
"eval_stsb_spearman": 0.8623026761683228,
|
| 189 |
+
"step": 625
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.21,
|
| 193 |
+
"eval_avg_sts": 0.8133106230456628,
|
| 194 |
+
"eval_sickr_spearman": 0.770097202780996,
|
| 195 |
+
"eval_stsb_spearman": 0.8565240433103297,
|
| 196 |
+
"step": 650
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.25,
|
| 200 |
+
"eval_avg_sts": 0.8207134638537716,
|
| 201 |
+
"eval_sickr_spearman": 0.7790919391162099,
|
| 202 |
+
"eval_stsb_spearman": 0.8623349885913333,
|
| 203 |
+
"step": 675
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.3,
|
| 207 |
+
"eval_avg_sts": 0.8130353473025216,
|
| 208 |
+
"eval_sickr_spearman": 0.7694488309426829,
|
| 209 |
+
"eval_stsb_spearman": 0.8566218636623602,
|
| 210 |
+
"step": 700
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.35,
|
| 214 |
+
"eval_avg_sts": 0.8135852616644544,
|
| 215 |
+
"eval_sickr_spearman": 0.7713928417422891,
|
| 216 |
+
"eval_stsb_spearman": 0.8557776815866198,
|
| 217 |
+
"step": 725
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.39,
|
| 221 |
+
"eval_avg_sts": 0.8182033488573868,
|
| 222 |
+
"eval_sickr_spearman": 0.7772465361678562,
|
| 223 |
+
"eval_stsb_spearman": 0.8591601615469172,
|
| 224 |
+
"step": 750
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.44,
|
| 228 |
+
"eval_avg_sts": 0.8248685458177708,
|
| 229 |
+
"eval_sickr_spearman": 0.7874361422135461,
|
| 230 |
+
"eval_stsb_spearman": 0.8623009494219955,
|
| 231 |
+
"step": 775
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.48,
|
| 235 |
+
"eval_avg_sts": 0.8161782720571904,
|
| 236 |
+
"eval_sickr_spearman": 0.7755229400927496,
|
| 237 |
+
"eval_stsb_spearman": 0.8568336040216311,
|
| 238 |
+
"step": 800
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.53,
|
| 242 |
+
"eval_avg_sts": 0.8207483887962784,
|
| 243 |
+
"eval_sickr_spearman": 0.7822897537877566,
|
| 244 |
+
"eval_stsb_spearman": 0.8592070238048003,
|
| 245 |
+
"step": 825
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.58,
|
| 249 |
+
"eval_avg_sts": 0.8239255438283619,
|
| 250 |
+
"eval_sickr_spearman": 0.7867861853488854,
|
| 251 |
+
"eval_stsb_spearman": 0.8610649023078384,
|
| 252 |
+
"step": 850
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.62,
|
| 256 |
+
"eval_avg_sts": 0.8190144607374632,
|
| 257 |
+
"eval_sickr_spearman": 0.7796808964820441,
|
| 258 |
+
"eval_stsb_spearman": 0.8583480249928824,
|
| 259 |
+
"step": 875
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.67,
|
| 263 |
+
"eval_avg_sts": 0.8211177033463906,
|
| 264 |
+
"eval_sickr_spearman": 0.7840137821427762,
|
| 265 |
+
"eval_stsb_spearman": 0.8582216245500051,
|
| 266 |
+
"step": 900
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.72,
|
| 270 |
+
"eval_avg_sts": 0.8201580926083631,
|
| 271 |
+
"eval_sickr_spearman": 0.7816100656713059,
|
| 272 |
+
"eval_stsb_spearman": 0.8587061195454205,
|
| 273 |
+
"step": 925
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.76,
|
| 277 |
+
"eval_avg_sts": 0.8140545936227954,
|
| 278 |
+
"eval_sickr_spearman": 0.7693706363095416,
|
| 279 |
+
"eval_stsb_spearman": 0.8587385509360493,
|
| 280 |
+
"step": 950
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.81,
|
| 284 |
+
"eval_avg_sts": 0.8205478478328316,
|
| 285 |
+
"eval_sickr_spearman": 0.778550484509698,
|
| 286 |
+
"eval_stsb_spearman": 0.8625452111559652,
|
| 287 |
+
"step": 975
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.86,
|
| 291 |
+
"learning_rate": 1.9078540507111937e-05,
|
| 292 |
+
"loss": 0.2352,
|
| 293 |
+
"step": 1000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.86,
|
| 297 |
+
"eval_avg_sts": 0.8194978907732944,
|
| 298 |
+
"eval_sickr_spearman": 0.7797764783739058,
|
| 299 |
+
"eval_stsb_spearman": 0.859219303172683,
|
| 300 |
+
"step": 1000
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 1.9,
|
| 304 |
+
"eval_avg_sts": 0.8179603788237452,
|
| 305 |
+
"eval_sickr_spearman": 0.7792966476705391,
|
| 306 |
+
"eval_stsb_spearman": 0.8566241099769513,
|
| 307 |
+
"step": 1025
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 1.95,
|
| 311 |
+
"eval_avg_sts": 0.814634655693672,
|
| 312 |
+
"eval_sickr_spearman": 0.7726660501792106,
|
| 313 |
+
"eval_stsb_spearman": 0.8566032612081333,
|
| 314 |
+
"step": 1050
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.99,
|
| 318 |
+
"eval_avg_sts": 0.8229387572431492,
|
| 319 |
+
"eval_sickr_spearman": 0.786164518802972,
|
| 320 |
+
"eval_stsb_spearman": 0.8597129956833264,
|
| 321 |
+
"step": 1075
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.04,
|
| 325 |
+
"eval_avg_sts": 0.8170446704890681,
|
| 326 |
+
"eval_sickr_spearman": 0.7740013628302915,
|
| 327 |
+
"eval_stsb_spearman": 0.8600879781478449,
|
| 328 |
+
"step": 1100
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.09,
|
| 332 |
+
"eval_avg_sts": 0.823749807845302,
|
| 333 |
+
"eval_sickr_spearman": 0.7840726202420378,
|
| 334 |
+
"eval_stsb_spearman": 0.8634269954485662,
|
| 335 |
+
"step": 1125
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 2.13,
|
| 339 |
+
"eval_avg_sts": 0.824679097934147,
|
| 340 |
+
"eval_sickr_spearman": 0.786571486325457,
|
| 341 |
+
"eval_stsb_spearman": 0.8627867095428369,
|
| 342 |
+
"step": 1150
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 2.18,
|
| 346 |
+
"eval_avg_sts": 0.8215752053838693,
|
| 347 |
+
"eval_sickr_spearman": 0.7836207916708096,
|
| 348 |
+
"eval_stsb_spearman": 0.859529619096929,
|
| 349 |
+
"step": 1175
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 2.23,
|
| 353 |
+
"eval_avg_sts": 0.8271625177527122,
|
| 354 |
+
"eval_sickr_spearman": 0.790379536140775,
|
| 355 |
+
"eval_stsb_spearman": 0.8639454993646494,
|
| 356 |
+
"step": 1200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.27,
|
| 360 |
+
"eval_avg_sts": 0.8175642564462906,
|
| 361 |
+
"eval_sickr_spearman": 0.7762766441365164,
|
| 362 |
+
"eval_stsb_spearman": 0.8588518687560649,
|
| 363 |
+
"step": 1225
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 2.32,
|
| 367 |
+
"eval_avg_sts": 0.8189636390886708,
|
| 368 |
+
"eval_sickr_spearman": 0.7794043814310648,
|
| 369 |
+
"eval_stsb_spearman": 0.8585228967462767,
|
| 370 |
+
"step": 1250
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 2.37,
|
| 374 |
+
"eval_avg_sts": 0.8235331481562411,
|
| 375 |
+
"eval_sickr_spearman": 0.7853204202262986,
|
| 376 |
+
"eval_stsb_spearman": 0.8617458760861836,
|
| 377 |
+
"step": 1275
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.41,
|
| 381 |
+
"eval_avg_sts": 0.8212932121455792,
|
| 382 |
+
"eval_sickr_spearman": 0.7838025893897528,
|
| 383 |
+
"eval_stsb_spearman": 0.8587838349014056,
|
| 384 |
+
"step": 1300
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.46,
|
| 388 |
+
"eval_avg_sts": 0.8200184185533622,
|
| 389 |
+
"eval_sickr_spearman": 0.7804020834701373,
|
| 390 |
+
"eval_stsb_spearman": 0.8596347536365871,
|
| 391 |
+
"step": 1325
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 2.5,
|
| 395 |
+
"eval_avg_sts": 0.8184879956680027,
|
| 396 |
+
"eval_sickr_spearman": 0.7775680563608829,
|
| 397 |
+
"eval_stsb_spearman": 0.8594079349751224,
|
| 398 |
+
"step": 1350
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.55,
|
| 402 |
+
"eval_avg_sts": 0.8240646293698506,
|
| 403 |
+
"eval_sickr_spearman": 0.7871783112610262,
|
| 404 |
+
"eval_stsb_spearman": 0.8609509474786748,
|
| 405 |
+
"step": 1375
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.6,
|
| 409 |
+
"eval_avg_sts": 0.8232889615255443,
|
| 410 |
+
"eval_sickr_spearman": 0.7868230252036885,
|
| 411 |
+
"eval_stsb_spearman": 0.8597548978474002,
|
| 412 |
+
"step": 1400
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 2.64,
|
| 416 |
+
"eval_avg_sts": 0.8214126106703122,
|
| 417 |
+
"eval_sickr_spearman": 0.7826799584558398,
|
| 418 |
+
"eval_stsb_spearman": 0.8601452628847847,
|
| 419 |
+
"step": 1425
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.69,
|
| 423 |
+
"eval_avg_sts": 0.8256026599230287,
|
| 424 |
+
"eval_sickr_spearman": 0.7879515639630783,
|
| 425 |
+
"eval_stsb_spearman": 0.8632537558829791,
|
| 426 |
+
"step": 1450
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 2.74,
|
| 430 |
+
"eval_avg_sts": 0.8227290144728177,
|
| 431 |
+
"eval_sickr_spearman": 0.7835550370929407,
|
| 432 |
+
"eval_stsb_spearman": 0.8619029918526946,
|
| 433 |
+
"step": 1475
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 2.78,
|
| 437 |
+
"learning_rate": 3.6178107606679037e-06,
|
| 438 |
+
"loss": 0.2129,
|
| 439 |
+
"step": 1500
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 2.78,
|
| 443 |
+
"eval_avg_sts": 0.8228934679819335,
|
| 444 |
+
"eval_sickr_spearman": 0.7842996632585357,
|
| 445 |
+
"eval_stsb_spearman": 0.8614872727053314,
|
| 446 |
+
"step": 1500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.83,
|
| 450 |
+
"eval_avg_sts": 0.8208607916598496,
|
| 451 |
+
"eval_sickr_spearman": 0.7816013240108441,
|
| 452 |
+
"eval_stsb_spearman": 0.860120259308855,
|
| 453 |
+
"step": 1525
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 2.88,
|
| 457 |
+
"eval_avg_sts": 0.8206687609569976,
|
| 458 |
+
"eval_sickr_spearman": 0.7812808124709476,
|
| 459 |
+
"eval_stsb_spearman": 0.8600567094430476,
|
| 460 |
+
"step": 1550
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 2.92,
|
| 464 |
+
"eval_avg_sts": 0.8208944641677367,
|
| 465 |
+
"eval_sickr_spearman": 0.7819390787161569,
|
| 466 |
+
"eval_stsb_spearman": 0.8598498496193164,
|
| 467 |
+
"step": 1575
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 2.97,
|
| 471 |
+
"eval_avg_sts": 0.8211952992867365,
|
| 472 |
+
"eval_sickr_spearman": 0.7822705413471812,
|
| 473 |
+
"eval_stsb_spearman": 0.8601200572262918,
|
| 474 |
+
"step": 1600
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 3.0,
|
| 478 |
+
"step": 1617,
|
| 479 |
+
"train_runtime": 2320.0614,
|
| 480 |
+
"train_samples_per_second": 0.697
|
| 481 |
+
}
|
| 482 |
+
],
|
| 483 |
+
"max_steps": 1617,
|
| 484 |
+
"num_train_epochs": 3,
|
| 485 |
+
"total_flos": 156090475282942080,
|
| 486 |
+
"trial_name": null,
|
| 487 |
+
"trial_params": null
|
| 488 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a464551b7303fca86b12387c4ccbbbcd1619d40f333c9d4b7a742fe1572f134c
|
| 3 |
+
size 2107
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|