Upload folder using huggingface_hub
Browse files- .ipynb_checkpoints/config-checkpoint.json +25 -0
- 1_Pooling/config.json +10 -0
- added_tokens.json +4 -0
- config.json +25 -0
- config_sentence_transformers.json +10 -0
- model.safetensors +3 -0
- modules.json +14 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- sentence_bert_config.json +4 -0
- special_tokens_map.json +37 -0
- tokenizer.json +0 -0
- tokenizer_config.json +102 -0
- trainer_state.json +3105 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
.ipynb_checkpoints/config-checkpoint.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "aubmindlab/bert-base-arabertv02",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.45.2",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 64002
|
| 25 |
+
}
|
1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 768,
|
| 3 |
+
"pooling_mode_cls_token": false,
|
| 4 |
+
"pooling_mode_mean_tokens": true,
|
| 5 |
+
"pooling_mode_max_tokens": false,
|
| 6 |
+
"pooling_mode_mean_sqrt_len_tokens": false,
|
| 7 |
+
"pooling_mode_weightedmean_tokens": false,
|
| 8 |
+
"pooling_mode_lasttoken": false,
|
| 9 |
+
"include_prompt": true
|
| 10 |
+
}
|
added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<passage>": 64001,
|
| 3 |
+
"<query>": 64000
|
| 4 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "aubmindlab/bert-base-arabertv02",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertModel"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 3072,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.45.2",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 64002
|
| 25 |
+
}
|
config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.2.1",
|
| 4 |
+
"transformers": "4.45.2",
|
| 5 |
+
"pytorch": "2.1.0+cu118"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7287986689fab8c867a0b99682022c848ae2a015c75194560b119618f96214cf
|
| 3 |
+
size 540801896
|
modules.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"idx": 0,
|
| 4 |
+
"name": "0",
|
| 5 |
+
"path": "",
|
| 6 |
+
"type": "sentence_transformers.models.Transformer"
|
| 7 |
+
},
|
| 8 |
+
{
|
| 9 |
+
"idx": 1,
|
| 10 |
+
"name": "1",
|
| 11 |
+
"path": "1_Pooling",
|
| 12 |
+
"type": "sentence_transformers.models.Pooling"
|
| 13 |
+
}
|
| 14 |
+
]
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1b31f4ac606cfee5c554f2ed0675f4fbe1b69210f580a41bf81b5c90e67b879
|
| 3 |
+
size 1076999098
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e3a30a127e91bea625ce6a21f7f0c95ca900b9ea5e124097be2534876d0d0acb
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bc5ddfdac7e5a8be097378232363eb8a98fba4780c49a2b78b57dec384e48bbb
|
| 3 |
+
size 1064
|
sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 256,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,102 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"5": {
|
| 44 |
+
"content": "[رابط]",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": true,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": true,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"6": {
|
| 52 |
+
"content": "[بريد]",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": true,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": true,
|
| 57 |
+
"special": true
|
| 58 |
+
},
|
| 59 |
+
"7": {
|
| 60 |
+
"content": "[مستخدم]",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": true,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": true,
|
| 65 |
+
"special": true
|
| 66 |
+
},
|
| 67 |
+
"64000": {
|
| 68 |
+
"content": "<query>",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": false,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": true
|
| 74 |
+
},
|
| 75 |
+
"64001": {
|
| 76 |
+
"content": "<passage>",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": false,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": true
|
| 82 |
+
}
|
| 83 |
+
},
|
| 84 |
+
"clean_up_tokenization_spaces": false,
|
| 85 |
+
"cls_token": "[CLS]",
|
| 86 |
+
"do_basic_tokenize": true,
|
| 87 |
+
"do_lower_case": false,
|
| 88 |
+
"mask_token": "[MASK]",
|
| 89 |
+
"max_len": 512,
|
| 90 |
+
"model_max_length": 512,
|
| 91 |
+
"never_split": [
|
| 92 |
+
"[بريد]",
|
| 93 |
+
"[مستخدم]",
|
| 94 |
+
"[رابط]"
|
| 95 |
+
],
|
| 96 |
+
"pad_token": "[PAD]",
|
| 97 |
+
"sep_token": "[SEP]",
|
| 98 |
+
"strip_accents": null,
|
| 99 |
+
"tokenize_chinese_chars": true,
|
| 100 |
+
"tokenizer_class": "BertTokenizer",
|
| 101 |
+
"unk_token": "[UNK]"
|
| 102 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8163072349117797,
|
| 3 |
+
"best_model_checkpoint": "output/arabert-2stage/checkpoint-4000",
|
| 4 |
+
"epoch": 0.6798096532970768,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.001699524133242692,
|
| 13 |
+
"grad_norm": 18.25199317932129,
|
| 14 |
+
"learning_rate": 1.3582342954159594e-07,
|
| 15 |
+
"loss": 4.8813,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.003399048266485384,
|
| 20 |
+
"grad_norm": 22.529903411865234,
|
| 21 |
+
"learning_rate": 3.0560271646859084e-07,
|
| 22 |
+
"loss": 4.555,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.005098572399728076,
|
| 27 |
+
"grad_norm": 16.15692901611328,
|
| 28 |
+
"learning_rate": 4.7538200339558575e-07,
|
| 29 |
+
"loss": 4.5007,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006798096532970768,
|
| 34 |
+
"grad_norm": 15.619988441467285,
|
| 35 |
+
"learning_rate": 6.451612903225807e-07,
|
| 36 |
+
"loss": 3.9173,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.00849762066621346,
|
| 41 |
+
"grad_norm": 16.881023406982422,
|
| 42 |
+
"learning_rate": 8.149405772495757e-07,
|
| 43 |
+
"loss": 4.3681,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.010197144799456152,
|
| 48 |
+
"grad_norm": 16.038394927978516,
|
| 49 |
+
"learning_rate": 9.67741935483871e-07,
|
| 50 |
+
"loss": 4.2357,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.011896668932698844,
|
| 55 |
+
"grad_norm": 12.993019104003906,
|
| 56 |
+
"learning_rate": 1.137521222410866e-06,
|
| 57 |
+
"loss": 4.1423,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.013596193065941536,
|
| 62 |
+
"grad_norm": 15.696857452392578,
|
| 63 |
+
"learning_rate": 1.307300509337861e-06,
|
| 64 |
+
"loss": 3.8028,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.015295717199184228,
|
| 69 |
+
"grad_norm": 10.482597351074219,
|
| 70 |
+
"learning_rate": 1.477079796264856e-06,
|
| 71 |
+
"loss": 3.8624,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.01699524133242692,
|
| 76 |
+
"grad_norm": 13.150389671325684,
|
| 77 |
+
"learning_rate": 1.6468590831918508e-06,
|
| 78 |
+
"loss": 3.5161,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.018694765465669613,
|
| 83 |
+
"grad_norm": 11.572787284851074,
|
| 84 |
+
"learning_rate": 1.8166383701188457e-06,
|
| 85 |
+
"loss": 3.4791,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.020394289598912305,
|
| 90 |
+
"grad_norm": 13.110374450683594,
|
| 91 |
+
"learning_rate": 1.9864176570458403e-06,
|
| 92 |
+
"loss": 3.6538,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.022093813732154997,
|
| 97 |
+
"grad_norm": 10.92455768585205,
|
| 98 |
+
"learning_rate": 2.1561969439728354e-06,
|
| 99 |
+
"loss": 2.8184,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.02379333786539769,
|
| 104 |
+
"grad_norm": 13.158881187438965,
|
| 105 |
+
"learning_rate": 2.3259762308998304e-06,
|
| 106 |
+
"loss": 3.321,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.02549286199864038,
|
| 111 |
+
"grad_norm": 7.724843502044678,
|
| 112 |
+
"learning_rate": 2.4957555178268255e-06,
|
| 113 |
+
"loss": 2.7731,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.027192386131883073,
|
| 118 |
+
"grad_norm": 9.516618728637695,
|
| 119 |
+
"learning_rate": 2.6655348047538205e-06,
|
| 120 |
+
"loss": 3.0967,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.028891910265125765,
|
| 125 |
+
"grad_norm": 8.883516311645508,
|
| 126 |
+
"learning_rate": 2.835314091680815e-06,
|
| 127 |
+
"loss": 2.8666,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.030591434398368457,
|
| 132 |
+
"grad_norm": 7.585137844085693,
|
| 133 |
+
"learning_rate": 3.0050933786078102e-06,
|
| 134 |
+
"loss": 2.2543,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.032290958531611146,
|
| 139 |
+
"grad_norm": 11.794560432434082,
|
| 140 |
+
"learning_rate": 3.174872665534805e-06,
|
| 141 |
+
"loss": 3.1811,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.03399048266485384,
|
| 146 |
+
"grad_norm": 8.631832122802734,
|
| 147 |
+
"learning_rate": 3.3446519524618e-06,
|
| 148 |
+
"loss": 2.5646,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.03569000679809653,
|
| 153 |
+
"grad_norm": 7.68147611618042,
|
| 154 |
+
"learning_rate": 3.5144312393887946e-06,
|
| 155 |
+
"loss": 2.8453,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.037389530931339225,
|
| 160 |
+
"grad_norm": 9.345367431640625,
|
| 161 |
+
"learning_rate": 3.6842105263157896e-06,
|
| 162 |
+
"loss": 3.0061,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.039089055064581914,
|
| 167 |
+
"grad_norm": 10.231043815612793,
|
| 168 |
+
"learning_rate": 3.853989813242784e-06,
|
| 169 |
+
"loss": 2.5211,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.04078857919782461,
|
| 174 |
+
"grad_norm": 6.374863147735596,
|
| 175 |
+
"learning_rate": 4.02376910016978e-06,
|
| 176 |
+
"loss": 2.6467,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.0424881033310673,
|
| 181 |
+
"grad_norm": 12.607449531555176,
|
| 182 |
+
"learning_rate": 4.193548387096774e-06,
|
| 183 |
+
"loss": 2.1417,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.04418762746430999,
|
| 188 |
+
"grad_norm": 12.550193786621094,
|
| 189 |
+
"learning_rate": 4.36332767402377e-06,
|
| 190 |
+
"loss": 2.7977,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.04588715159755268,
|
| 195 |
+
"grad_norm": 9.875368118286133,
|
| 196 |
+
"learning_rate": 4.5331069609507645e-06,
|
| 197 |
+
"loss": 2.8345,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.04758667573079538,
|
| 202 |
+
"grad_norm": 14.792099952697754,
|
| 203 |
+
"learning_rate": 4.702886247877759e-06,
|
| 204 |
+
"loss": 2.8812,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.049286199864038066,
|
| 209 |
+
"grad_norm": 11.13675308227539,
|
| 210 |
+
"learning_rate": 4.872665534804754e-06,
|
| 211 |
+
"loss": 2.493,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.05098572399728076,
|
| 216 |
+
"grad_norm": 15.685995101928711,
|
| 217 |
+
"learning_rate": 5.042444821731749e-06,
|
| 218 |
+
"loss": 2.5967,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.05268524813052345,
|
| 223 |
+
"grad_norm": 8.819021224975586,
|
| 224 |
+
"learning_rate": 5.212224108658745e-06,
|
| 225 |
+
"loss": 2.4955,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.054384772263766146,
|
| 230 |
+
"grad_norm": 12.058821678161621,
|
| 231 |
+
"learning_rate": 5.3820033955857386e-06,
|
| 232 |
+
"loss": 2.2387,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.056084296397008834,
|
| 237 |
+
"grad_norm": 9.35466194152832,
|
| 238 |
+
"learning_rate": 5.551782682512734e-06,
|
| 239 |
+
"loss": 2.6786,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.05778382053025153,
|
| 244 |
+
"grad_norm": 9.817248344421387,
|
| 245 |
+
"learning_rate": 5.721561969439729e-06,
|
| 246 |
+
"loss": 2.2292,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.05948334466349422,
|
| 251 |
+
"grad_norm": 7.38469934463501,
|
| 252 |
+
"learning_rate": 5.891341256366724e-06,
|
| 253 |
+
"loss": 1.8386,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.061182868796736914,
|
| 258 |
+
"grad_norm": 7.639986038208008,
|
| 259 |
+
"learning_rate": 6.061120543293718e-06,
|
| 260 |
+
"loss": 2.5317,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.0628823929299796,
|
| 265 |
+
"grad_norm": 14.127429962158203,
|
| 266 |
+
"learning_rate": 6.2308998302207134e-06,
|
| 267 |
+
"loss": 2.3225,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.06458191706322229,
|
| 272 |
+
"grad_norm": 6.480667591094971,
|
| 273 |
+
"learning_rate": 6.400679117147709e-06,
|
| 274 |
+
"loss": 2.5349,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.066281441196465,
|
| 279 |
+
"grad_norm": 12.269805908203125,
|
| 280 |
+
"learning_rate": 6.5704584040747036e-06,
|
| 281 |
+
"loss": 2.5192,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.06798096532970768,
|
| 286 |
+
"grad_norm": 10.89499568939209,
|
| 287 |
+
"learning_rate": 6.740237691001699e-06,
|
| 288 |
+
"loss": 2.6806,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.06968048946295037,
|
| 293 |
+
"grad_norm": 6.393769264221191,
|
| 294 |
+
"learning_rate": 6.910016977928693e-06,
|
| 295 |
+
"loss": 2.593,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.07138001359619306,
|
| 300 |
+
"grad_norm": 10.500598907470703,
|
| 301 |
+
"learning_rate": 7.079796264855688e-06,
|
| 302 |
+
"loss": 1.893,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.07307953772943576,
|
| 307 |
+
"grad_norm": 9.394804954528809,
|
| 308 |
+
"learning_rate": 7.249575551782683e-06,
|
| 309 |
+
"loss": 2.4957,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.07477906186267845,
|
| 314 |
+
"grad_norm": 8.060555458068848,
|
| 315 |
+
"learning_rate": 7.4193548387096784e-06,
|
| 316 |
+
"loss": 2.2887,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.07647858599592114,
|
| 321 |
+
"grad_norm": 7.361180305480957,
|
| 322 |
+
"learning_rate": 7.589134125636672e-06,
|
| 323 |
+
"loss": 1.8902,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.07817811012916383,
|
| 328 |
+
"grad_norm": 9.406976699829102,
|
| 329 |
+
"learning_rate": 7.758913412563669e-06,
|
| 330 |
+
"loss": 2.445,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.07987763426240653,
|
| 335 |
+
"grad_norm": 6.804340362548828,
|
| 336 |
+
"learning_rate": 7.928692699490664e-06,
|
| 337 |
+
"loss": 2.3188,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.08157715839564922,
|
| 342 |
+
"grad_norm": 8.055757522583008,
|
| 343 |
+
"learning_rate": 8.098471986417658e-06,
|
| 344 |
+
"loss": 1.7857,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.08327668252889191,
|
| 349 |
+
"grad_norm": 14.029518127441406,
|
| 350 |
+
"learning_rate": 8.268251273344653e-06,
|
| 351 |
+
"loss": 1.9323,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.0849762066621346,
|
| 356 |
+
"grad_norm": 7.238176345825195,
|
| 357 |
+
"learning_rate": 8.438030560271647e-06,
|
| 358 |
+
"loss": 2.2119,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.0849762066621346,
|
| 363 |
+
"eval_cosine_accuracy@1": 0.644,
|
| 364 |
+
"eval_cosine_accuracy@10": 0.89,
|
| 365 |
+
"eval_cosine_accuracy@3": 0.7905,
|
| 366 |
+
"eval_cosine_accuracy@5": 0.842,
|
| 367 |
+
"eval_cosine_map@100": 0.7319568991974511,
|
| 368 |
+
"eval_cosine_mrr@10": 0.727909523809523,
|
| 369 |
+
"eval_cosine_ndcg@10": 0.7672958038269853,
|
| 370 |
+
"eval_cosine_precision@1": 0.644,
|
| 371 |
+
"eval_cosine_precision@10": 0.089,
|
| 372 |
+
"eval_cosine_precision@3": 0.2635,
|
| 373 |
+
"eval_cosine_precision@5": 0.1684,
|
| 374 |
+
"eval_cosine_recall@1": 0.644,
|
| 375 |
+
"eval_cosine_recall@10": 0.89,
|
| 376 |
+
"eval_cosine_recall@3": 0.7905,
|
| 377 |
+
"eval_cosine_recall@5": 0.842,
|
| 378 |
+
"eval_loss": 2.020264148712158,
|
| 379 |
+
"eval_runtime": 2.7651,
|
| 380 |
+
"eval_samples_per_second": 272.684,
|
| 381 |
+
"eval_sequential_score": 0.7319568991974511,
|
| 382 |
+
"eval_steps_per_second": 2.17,
|
| 383 |
+
"eval_sts-dev_pearson_cosine": 0.7945015650907715,
|
| 384 |
+
"eval_sts-dev_pearson_dot": 0.7892461139806375,
|
| 385 |
+
"eval_sts-dev_pearson_euclidean": 0.7772212795710262,
|
| 386 |
+
"eval_sts-dev_pearson_manhattan": 0.7758198419084321,
|
| 387 |
+
"eval_sts-dev_pearson_max": 0.7945015650907715,
|
| 388 |
+
"eval_sts-dev_spearman_cosine": 0.8029369749021982,
|
| 389 |
+
"eval_sts-dev_spearman_dot": 0.7981715789995407,
|
| 390 |
+
"eval_sts-dev_spearman_euclidean": 0.7974969539844081,
|
| 391 |
+
"eval_sts-dev_spearman_manhattan": 0.793442700340276,
|
| 392 |
+
"eval_sts-dev_spearman_max": 0.8029369749021982,
|
| 393 |
+
"step": 500
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.0866757307953773,
|
| 397 |
+
"grad_norm": 8.096364974975586,
|
| 398 |
+
"learning_rate": 8.607809847198643e-06,
|
| 399 |
+
"loss": 2.4202,
|
| 400 |
+
"step": 510
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 0.08837525492861999,
|
| 404 |
+
"grad_norm": 5.928977012634277,
|
| 405 |
+
"learning_rate": 8.777589134125636e-06,
|
| 406 |
+
"loss": 1.9747,
|
| 407 |
+
"step": 520
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.09007477906186268,
|
| 411 |
+
"grad_norm": 5.913745880126953,
|
| 412 |
+
"learning_rate": 8.947368421052632e-06,
|
| 413 |
+
"loss": 2.727,
|
| 414 |
+
"step": 530
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.09177430319510536,
|
| 418 |
+
"grad_norm": 11.565766334533691,
|
| 419 |
+
"learning_rate": 9.117147707979627e-06,
|
| 420 |
+
"loss": 2.2645,
|
| 421 |
+
"step": 540
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 0.09347382732834807,
|
| 425 |
+
"grad_norm": 5.834039688110352,
|
| 426 |
+
"learning_rate": 9.286926994906623e-06,
|
| 427 |
+
"loss": 2.5484,
|
| 428 |
+
"step": 550
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.09517335146159075,
|
| 432 |
+
"grad_norm": 9.030844688415527,
|
| 433 |
+
"learning_rate": 9.456706281833618e-06,
|
| 434 |
+
"loss": 2.0806,
|
| 435 |
+
"step": 560
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.09687287559483344,
|
| 439 |
+
"grad_norm": 8.642946243286133,
|
| 440 |
+
"learning_rate": 9.626485568760612e-06,
|
| 441 |
+
"loss": 2.4332,
|
| 442 |
+
"step": 570
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 0.09857239972807613,
|
| 446 |
+
"grad_norm": 8.626252174377441,
|
| 447 |
+
"learning_rate": 9.796264855687608e-06,
|
| 448 |
+
"loss": 1.866,
|
| 449 |
+
"step": 580
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.10027192386131883,
|
| 453 |
+
"grad_norm": 7.73045539855957,
|
| 454 |
+
"learning_rate": 9.966044142614601e-06,
|
| 455 |
+
"loss": 2.1285,
|
| 456 |
+
"step": 590
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 0.10197144799456152,
|
| 460 |
+
"grad_norm": 8.957172393798828,
|
| 461 |
+
"learning_rate": 9.999943676843767e-06,
|
| 462 |
+
"loss": 2.3799,
|
| 463 |
+
"step": 600
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 0.10367097212780421,
|
| 467 |
+
"grad_norm": 7.423664093017578,
|
| 468 |
+
"learning_rate": 9.999714866196319e-06,
|
| 469 |
+
"loss": 2.6725,
|
| 470 |
+
"step": 610
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.1053704962610469,
|
| 474 |
+
"grad_norm": 9.436025619506836,
|
| 475 |
+
"learning_rate": 9.999310055908776e-06,
|
| 476 |
+
"loss": 2.4767,
|
| 477 |
+
"step": 620
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.1070700203942896,
|
| 481 |
+
"grad_norm": 8.621382713317871,
|
| 482 |
+
"learning_rate": 9.998729260231243e-06,
|
| 483 |
+
"loss": 1.9994,
|
| 484 |
+
"step": 630
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 0.10876954452753229,
|
| 488 |
+
"grad_norm": 8.81654167175293,
|
| 489 |
+
"learning_rate": 9.997972499608852e-06,
|
| 490 |
+
"loss": 2.6699,
|
| 491 |
+
"step": 640
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.11046906866077498,
|
| 495 |
+
"grad_norm": 6.4313154220581055,
|
| 496 |
+
"learning_rate": 9.997039800681044e-06,
|
| 497 |
+
"loss": 2.2264,
|
| 498 |
+
"step": 650
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 0.11216859279401767,
|
| 502 |
+
"grad_norm": 8.586142539978027,
|
| 503 |
+
"learning_rate": 9.995931196280622e-06,
|
| 504 |
+
"loss": 2.0675,
|
| 505 |
+
"step": 660
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.11386811692726037,
|
| 509 |
+
"grad_norm": 6.31871223449707,
|
| 510 |
+
"learning_rate": 9.994646725432611e-06,
|
| 511 |
+
"loss": 1.7862,
|
| 512 |
+
"step": 670
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.11556764106050306,
|
| 516 |
+
"grad_norm": 9.907938957214355,
|
| 517 |
+
"learning_rate": 9.993186433352867e-06,
|
| 518 |
+
"loss": 2.5078,
|
| 519 |
+
"step": 680
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.11726716519374575,
|
| 523 |
+
"grad_norm": 5.77314567565918,
|
| 524 |
+
"learning_rate": 9.9915503714465e-06,
|
| 525 |
+
"loss": 2.135,
|
| 526 |
+
"step": 690
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 0.11896668932698844,
|
| 530 |
+
"grad_norm": 15.68774700164795,
|
| 531 |
+
"learning_rate": 9.989738597306053e-06,
|
| 532 |
+
"loss": 2.5356,
|
| 533 |
+
"step": 700
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.12066621346023114,
|
| 537 |
+
"grad_norm": 13.062637329101562,
|
| 538 |
+
"learning_rate": 9.987751174709489e-06,
|
| 539 |
+
"loss": 2.4131,
|
| 540 |
+
"step": 710
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 0.12236573759347383,
|
| 544 |
+
"grad_norm": 6.619427680969238,
|
| 545 |
+
"learning_rate": 9.98558817361792e-06,
|
| 546 |
+
"loss": 2.0465,
|
| 547 |
+
"step": 720
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 0.12406526172671652,
|
| 551 |
+
"grad_norm": 10.573486328125,
|
| 552 |
+
"learning_rate": 9.98324967017318e-06,
|
| 553 |
+
"loss": 2.5195,
|
| 554 |
+
"step": 730
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.1257647858599592,
|
| 558 |
+
"grad_norm": 10.98731803894043,
|
| 559 |
+
"learning_rate": 9.98073574669511e-06,
|
| 560 |
+
"loss": 2.3977,
|
| 561 |
+
"step": 740
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.1274643099932019,
|
| 565 |
+
"grad_norm": 7.568258285522461,
|
| 566 |
+
"learning_rate": 9.978046491678686e-06,
|
| 567 |
+
"loss": 1.8389,
|
| 568 |
+
"step": 750
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 0.12916383412644458,
|
| 572 |
+
"grad_norm": 6.886653423309326,
|
| 573 |
+
"learning_rate": 9.975181999790888e-06,
|
| 574 |
+
"loss": 2.1653,
|
| 575 |
+
"step": 760
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.13086335825968728,
|
| 579 |
+
"grad_norm": 8.51340103149414,
|
| 580 |
+
"learning_rate": 9.972142371867375e-06,
|
| 581 |
+
"loss": 2.3171,
|
| 582 |
+
"step": 770
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 0.13256288239293,
|
| 586 |
+
"grad_norm": 13.55966567993164,
|
| 587 |
+
"learning_rate": 9.968927714908934e-06,
|
| 588 |
+
"loss": 2.3126,
|
| 589 |
+
"step": 780
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.13426240652617266,
|
| 593 |
+
"grad_norm": 6.183629035949707,
|
| 594 |
+
"learning_rate": 9.965538142077709e-06,
|
| 595 |
+
"loss": 1.8593,
|
| 596 |
+
"step": 790
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.13596193065941536,
|
| 600 |
+
"grad_norm": 8.71070671081543,
|
| 601 |
+
"learning_rate": 9.961973772693226e-06,
|
| 602 |
+
"loss": 2.3248,
|
| 603 |
+
"step": 800
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.13766145479265807,
|
| 607 |
+
"grad_norm": 6.578171253204346,
|
| 608 |
+
"learning_rate": 9.958234732228184e-06,
|
| 609 |
+
"loss": 1.4786,
|
| 610 |
+
"step": 810
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 0.13936097892590074,
|
| 614 |
+
"grad_norm": 8.017528533935547,
|
| 615 |
+
"learning_rate": 9.954321152304049e-06,
|
| 616 |
+
"loss": 2.5924,
|
| 617 |
+
"step": 820
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.14106050305914344,
|
| 621 |
+
"grad_norm": 9.630891799926758,
|
| 622 |
+
"learning_rate": 9.950233170686404e-06,
|
| 623 |
+
"loss": 2.7052,
|
| 624 |
+
"step": 830
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.14276002719238612,
|
| 628 |
+
"grad_norm": 9.43925952911377,
|
| 629 |
+
"learning_rate": 9.945970931280117e-06,
|
| 630 |
+
"loss": 2.9104,
|
| 631 |
+
"step": 840
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 0.14445955132562882,
|
| 635 |
+
"grad_norm": 7.2036967277526855,
|
| 636 |
+
"learning_rate": 9.941534584124262e-06,
|
| 637 |
+
"loss": 1.905,
|
| 638 |
+
"step": 850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.14615907545887152,
|
| 642 |
+
"grad_norm": 8.345267295837402,
|
| 643 |
+
"learning_rate": 9.936924285386849e-06,
|
| 644 |
+
"loss": 1.4475,
|
| 645 |
+
"step": 860
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.1478585995921142,
|
| 649 |
+
"grad_norm": 6.2061614990234375,
|
| 650 |
+
"learning_rate": 9.932140197359312e-06,
|
| 651 |
+
"loss": 1.9715,
|
| 652 |
+
"step": 870
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 0.1495581237253569,
|
| 656 |
+
"grad_norm": 8.147466659545898,
|
| 657 |
+
"learning_rate": 9.927182488450813e-06,
|
| 658 |
+
"loss": 1.922,
|
| 659 |
+
"step": 880
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 0.1512576478585996,
|
| 663 |
+
"grad_norm": 12.269794464111328,
|
| 664 |
+
"learning_rate": 9.9220513331823e-06,
|
| 665 |
+
"loss": 2.5347,
|
| 666 |
+
"step": 890
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 0.15295717199184228,
|
| 670 |
+
"grad_norm": 4.720547199249268,
|
| 671 |
+
"learning_rate": 9.916746912180369e-06,
|
| 672 |
+
"loss": 1.9091,
|
| 673 |
+
"step": 900
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 0.15465669612508498,
|
| 677 |
+
"grad_norm": 10.795013427734375,
|
| 678 |
+
"learning_rate": 9.911269412170906e-06,
|
| 679 |
+
"loss": 2.127,
|
| 680 |
+
"step": 910
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.15635622025832766,
|
| 684 |
+
"grad_norm": 9.331660270690918,
|
| 685 |
+
"learning_rate": 9.905619025972513e-06,
|
| 686 |
+
"loss": 2.0569,
|
| 687 |
+
"step": 920
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.15805574439157036,
|
| 691 |
+
"grad_norm": 8.150259017944336,
|
| 692 |
+
"learning_rate": 9.89979595248972e-06,
|
| 693 |
+
"loss": 1.5199,
|
| 694 |
+
"step": 930
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 0.15975526852481306,
|
| 698 |
+
"grad_norm": 5.694140434265137,
|
| 699 |
+
"learning_rate": 9.893800396705985e-06,
|
| 700 |
+
"loss": 2.1115,
|
| 701 |
+
"step": 940
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 0.16145479265805573,
|
| 705 |
+
"grad_norm": 6.206644535064697,
|
| 706 |
+
"learning_rate": 9.887632569676475e-06,
|
| 707 |
+
"loss": 2.3499,
|
| 708 |
+
"step": 950
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 0.16315431679129844,
|
| 712 |
+
"grad_norm": 4.6663408279418945,
|
| 713 |
+
"learning_rate": 9.881292688520638e-06,
|
| 714 |
+
"loss": 2.2428,
|
| 715 |
+
"step": 960
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 0.16485384092454114,
|
| 719 |
+
"grad_norm": 7.22851037979126,
|
| 720 |
+
"learning_rate": 9.874780976414565e-06,
|
| 721 |
+
"loss": 2.3808,
|
| 722 |
+
"step": 970
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.16655336505778381,
|
| 726 |
+
"grad_norm": 5.55798864364624,
|
| 727 |
+
"learning_rate": 9.86809766258312e-06,
|
| 728 |
+
"loss": 2.0025,
|
| 729 |
+
"step": 980
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.16825288919102652,
|
| 733 |
+
"grad_norm": 5.861523628234863,
|
| 734 |
+
"learning_rate": 9.86124298229189e-06,
|
| 735 |
+
"loss": 2.0427,
|
| 736 |
+
"step": 990
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 0.1699524133242692,
|
| 740 |
+
"grad_norm": 6.302640438079834,
|
| 741 |
+
"learning_rate": 9.854217176838886e-06,
|
| 742 |
+
"loss": 1.8858,
|
| 743 |
+
"step": 1000
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 0.1699524133242692,
|
| 747 |
+
"eval_cosine_accuracy@1": 0.682,
|
| 748 |
+
"eval_cosine_accuracy@10": 0.906,
|
| 749 |
+
"eval_cosine_accuracy@3": 0.8185,
|
| 750 |
+
"eval_cosine_accuracy@5": 0.858,
|
| 751 |
+
"eval_cosine_map@100": 0.762636433194321,
|
| 752 |
+
"eval_cosine_mrr@10": 0.7592498015873005,
|
| 753 |
+
"eval_cosine_ndcg@10": 0.7948932488118442,
|
| 754 |
+
"eval_cosine_precision@1": 0.682,
|
| 755 |
+
"eval_cosine_precision@10": 0.09060000000000001,
|
| 756 |
+
"eval_cosine_precision@3": 0.2728333333333333,
|
| 757 |
+
"eval_cosine_precision@5": 0.1716,
|
| 758 |
+
"eval_cosine_recall@1": 0.682,
|
| 759 |
+
"eval_cosine_recall@10": 0.906,
|
| 760 |
+
"eval_cosine_recall@3": 0.8185,
|
| 761 |
+
"eval_cosine_recall@5": 0.858,
|
| 762 |
+
"eval_loss": 1.8376002311706543,
|
| 763 |
+
"eval_runtime": 2.7558,
|
| 764 |
+
"eval_samples_per_second": 273.604,
|
| 765 |
+
"eval_sequential_score": 0.762636433194321,
|
| 766 |
+
"eval_steps_per_second": 2.177,
|
| 767 |
+
"eval_sts-dev_pearson_cosine": 0.7985157354895591,
|
| 768 |
+
"eval_sts-dev_pearson_dot": 0.7942162925984169,
|
| 769 |
+
"eval_sts-dev_pearson_euclidean": 0.7855985286403331,
|
| 770 |
+
"eval_sts-dev_pearson_manhattan": 0.7854261466194248,
|
| 771 |
+
"eval_sts-dev_pearson_max": 0.7985157354895591,
|
| 772 |
+
"eval_sts-dev_spearman_cosine": 0.8055793147203388,
|
| 773 |
+
"eval_sts-dev_spearman_dot": 0.8010661505883895,
|
| 774 |
+
"eval_sts-dev_spearman_euclidean": 0.8034068438279502,
|
| 775 |
+
"eval_sts-dev_spearman_manhattan": 0.8018568915859646,
|
| 776 |
+
"eval_sts-dev_spearman_max": 0.8055793147203388,
|
| 777 |
+
"step": 1000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.1716519374575119,
|
| 781 |
+
"grad_norm": 7.927140235900879,
|
| 782 |
+
"learning_rate": 9.847020493546058e-06,
|
| 783 |
+
"loss": 1.9554,
|
| 784 |
+
"step": 1010
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.1733514615907546,
|
| 788 |
+
"grad_norm": 8.299236297607422,
|
| 789 |
+
"learning_rate": 9.839653185750588e-06,
|
| 790 |
+
"loss": 2.2341,
|
| 791 |
+
"step": 1020
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.17505098572399727,
|
| 795 |
+
"grad_norm": 8.708532333374023,
|
| 796 |
+
"learning_rate": 9.83211551279597e-06,
|
| 797 |
+
"loss": 2.4421,
|
| 798 |
+
"step": 1030
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.17675050985723997,
|
| 802 |
+
"grad_norm": 9.657612800598145,
|
| 803 |
+
"learning_rate": 9.824407740022878e-06,
|
| 804 |
+
"loss": 1.8227,
|
| 805 |
+
"step": 1040
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.17845003399048268,
|
| 809 |
+
"grad_norm": 6.149839401245117,
|
| 810 |
+
"learning_rate": 9.816530138759837e-06,
|
| 811 |
+
"loss": 2.1157,
|
| 812 |
+
"step": 1050
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.18014955812372535,
|
| 816 |
+
"grad_norm": 5.039497375488281,
|
| 817 |
+
"learning_rate": 9.808482986313653e-06,
|
| 818 |
+
"loss": 1.7899,
|
| 819 |
+
"step": 1060
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.18184908225696805,
|
| 823 |
+
"grad_norm": 8.380406379699707,
|
| 824 |
+
"learning_rate": 9.80026656595967e-06,
|
| 825 |
+
"loss": 1.8153,
|
| 826 |
+
"step": 1070
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.18354860639021073,
|
| 830 |
+
"grad_norm": 5.239027500152588,
|
| 831 |
+
"learning_rate": 9.791881166931788e-06,
|
| 832 |
+
"loss": 2.0095,
|
| 833 |
+
"step": 1080
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 0.18524813052345343,
|
| 837 |
+
"grad_norm": 6.085160732269287,
|
| 838 |
+
"learning_rate": 9.783327084412277e-06,
|
| 839 |
+
"loss": 1.5087,
|
| 840 |
+
"step": 1090
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 0.18694765465669613,
|
| 844 |
+
"grad_norm": 5.803791522979736,
|
| 845 |
+
"learning_rate": 9.774604619521404e-06,
|
| 846 |
+
"loss": 1.9106,
|
| 847 |
+
"step": 1100
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.1886471787899388,
|
| 851 |
+
"grad_norm": 5.130899906158447,
|
| 852 |
+
"learning_rate": 9.765714079306814e-06,
|
| 853 |
+
"loss": 2.0189,
|
| 854 |
+
"step": 1110
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.1903467029231815,
|
| 858 |
+
"grad_norm": 13.715828895568848,
|
| 859 |
+
"learning_rate": 9.756655776732727e-06,
|
| 860 |
+
"loss": 2.3282,
|
| 861 |
+
"step": 1120
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.1920462270564242,
|
| 865 |
+
"grad_norm": 9.30579662322998,
|
| 866 |
+
"learning_rate": 9.74743003066893e-06,
|
| 867 |
+
"loss": 2.147,
|
| 868 |
+
"step": 1130
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.1937457511896669,
|
| 872 |
+
"grad_norm": 6.4321675300598145,
|
| 873 |
+
"learning_rate": 9.738037165879538e-06,
|
| 874 |
+
"loss": 1.9017,
|
| 875 |
+
"step": 1140
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 0.1954452753229096,
|
| 879 |
+
"grad_norm": 5.368739128112793,
|
| 880 |
+
"learning_rate": 9.72847751301157e-06,
|
| 881 |
+
"loss": 2.102,
|
| 882 |
+
"step": 1150
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 0.19714479945615226,
|
| 886 |
+
"grad_norm": 5.458316802978516,
|
| 887 |
+
"learning_rate": 9.718751408583312e-06,
|
| 888 |
+
"loss": 2.1275,
|
| 889 |
+
"step": 1160
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.19884432358939497,
|
| 893 |
+
"grad_norm": 8.618191719055176,
|
| 894 |
+
"learning_rate": 9.708859194972462e-06,
|
| 895 |
+
"loss": 2.6718,
|
| 896 |
+
"step": 1170
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.20054384772263767,
|
| 900 |
+
"grad_norm": 7.549973487854004,
|
| 901 |
+
"learning_rate": 9.698801220404086e-06,
|
| 902 |
+
"loss": 1.8549,
|
| 903 |
+
"step": 1180
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.20224337185588034,
|
| 907 |
+
"grad_norm": 5.044092655181885,
|
| 908 |
+
"learning_rate": 9.688577838938358e-06,
|
| 909 |
+
"loss": 2.1802,
|
| 910 |
+
"step": 1190
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.20394289598912305,
|
| 914 |
+
"grad_norm": 5.195682048797607,
|
| 915 |
+
"learning_rate": 9.67818941045809e-06,
|
| 916 |
+
"loss": 2.2905,
|
| 917 |
+
"step": 1200
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 0.20564242012236575,
|
| 921 |
+
"grad_norm": 11.789019584655762,
|
| 922 |
+
"learning_rate": 9.667636300656073e-06,
|
| 923 |
+
"loss": 2.283,
|
| 924 |
+
"step": 1210
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 0.20734194425560842,
|
| 928 |
+
"grad_norm": 7.097940921783447,
|
| 929 |
+
"learning_rate": 9.656918881022196e-06,
|
| 930 |
+
"loss": 2.1976,
|
| 931 |
+
"step": 1220
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 0.20904146838885113,
|
| 935 |
+
"grad_norm": 10.327863693237305,
|
| 936 |
+
"learning_rate": 9.646037528830374e-06,
|
| 937 |
+
"loss": 2.5894,
|
| 938 |
+
"step": 1230
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.2107409925220938,
|
| 942 |
+
"grad_norm": 6.603916645050049,
|
| 943 |
+
"learning_rate": 9.634992627125264e-06,
|
| 944 |
+
"loss": 1.7659,
|
| 945 |
+
"step": 1240
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.2124405166553365,
|
| 949 |
+
"grad_norm": 4.55158805847168,
|
| 950 |
+
"learning_rate": 9.623784564708782e-06,
|
| 951 |
+
"loss": 1.9581,
|
| 952 |
+
"step": 1250
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 0.2141400407885792,
|
| 956 |
+
"grad_norm": 6.308363437652588,
|
| 957 |
+
"learning_rate": 9.612413736126421e-06,
|
| 958 |
+
"loss": 2.1419,
|
| 959 |
+
"step": 1260
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.21583956492182188,
|
| 963 |
+
"grad_norm": 5.769819259643555,
|
| 964 |
+
"learning_rate": 9.600880541653352e-06,
|
| 965 |
+
"loss": 1.7988,
|
| 966 |
+
"step": 1270
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 0.21753908905506458,
|
| 970 |
+
"grad_norm": 9.009718894958496,
|
| 971 |
+
"learning_rate": 9.589185387280343e-06,
|
| 972 |
+
"loss": 2.3537,
|
| 973 |
+
"step": 1280
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.21923861318830729,
|
| 977 |
+
"grad_norm": 10.911920547485352,
|
| 978 |
+
"learning_rate": 9.577328684699468e-06,
|
| 979 |
+
"loss": 2.2776,
|
| 980 |
+
"step": 1290
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.22093813732154996,
|
| 984 |
+
"grad_norm": 7.810164928436279,
|
| 985 |
+
"learning_rate": 9.565310851289602e-06,
|
| 986 |
+
"loss": 2.0633,
|
| 987 |
+
"step": 1300
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.22263766145479266,
|
| 991 |
+
"grad_norm": 7.256144046783447,
|
| 992 |
+
"learning_rate": 9.553132310101741e-06,
|
| 993 |
+
"loss": 2.247,
|
| 994 |
+
"step": 1310
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.22433718558803534,
|
| 998 |
+
"grad_norm": 8.476386070251465,
|
| 999 |
+
"learning_rate": 9.540793489844106e-06,
|
| 1000 |
+
"loss": 1.6592,
|
| 1001 |
+
"step": 1320
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 0.22603670972127804,
|
| 1005 |
+
"grad_norm": 7.557313919067383,
|
| 1006 |
+
"learning_rate": 9.528294824867054e-06,
|
| 1007 |
+
"loss": 1.8965,
|
| 1008 |
+
"step": 1330
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 0.22773623385452074,
|
| 1012 |
+
"grad_norm": 5.973213195800781,
|
| 1013 |
+
"learning_rate": 9.515636755147781e-06,
|
| 1014 |
+
"loss": 2.3415,
|
| 1015 |
+
"step": 1340
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.22943575798776342,
|
| 1019 |
+
"grad_norm": 6.3328471183776855,
|
| 1020 |
+
"learning_rate": 9.50281972627484e-06,
|
| 1021 |
+
"loss": 1.7645,
|
| 1022 |
+
"step": 1350
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.23113528212100612,
|
| 1026 |
+
"grad_norm": 12.455780982971191,
|
| 1027 |
+
"learning_rate": 9.489844189432456e-06,
|
| 1028 |
+
"loss": 1.9076,
|
| 1029 |
+
"step": 1360
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.23283480625424882,
|
| 1033 |
+
"grad_norm": 6.662944316864014,
|
| 1034 |
+
"learning_rate": 9.476710601384639e-06,
|
| 1035 |
+
"loss": 2.0724,
|
| 1036 |
+
"step": 1370
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 0.2345343303874915,
|
| 1040 |
+
"grad_norm": 8.458195686340332,
|
| 1041 |
+
"learning_rate": 9.463419424459108e-06,
|
| 1042 |
+
"loss": 2.0766,
|
| 1043 |
+
"step": 1380
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.2362338545207342,
|
| 1047 |
+
"grad_norm": 5.073389530181885,
|
| 1048 |
+
"learning_rate": 9.449971126531015e-06,
|
| 1049 |
+
"loss": 1.6035,
|
| 1050 |
+
"step": 1390
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 0.23793337865397687,
|
| 1054 |
+
"grad_norm": 9.805305480957031,
|
| 1055 |
+
"learning_rate": 9.436366181006476e-06,
|
| 1056 |
+
"loss": 2.1389,
|
| 1057 |
+
"step": 1400
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 0.23963290278721958,
|
| 1061 |
+
"grad_norm": 5.322582244873047,
|
| 1062 |
+
"learning_rate": 9.422605066805906e-06,
|
| 1063 |
+
"loss": 1.8332,
|
| 1064 |
+
"step": 1410
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.24133242692046228,
|
| 1068 |
+
"grad_norm": 6.252927780151367,
|
| 1069 |
+
"learning_rate": 9.408688268347157e-06,
|
| 1070 |
+
"loss": 1.8316,
|
| 1071 |
+
"step": 1420
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.24303195105370495,
|
| 1075 |
+
"grad_norm": 6.532680034637451,
|
| 1076 |
+
"learning_rate": 9.394616275528475e-06,
|
| 1077 |
+
"loss": 1.754,
|
| 1078 |
+
"step": 1430
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 0.24473147518694766,
|
| 1082 |
+
"grad_norm": 10.015703201293945,
|
| 1083 |
+
"learning_rate": 9.380389583711238e-06,
|
| 1084 |
+
"loss": 2.0633,
|
| 1085 |
+
"step": 1440
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.24643099932019036,
|
| 1089 |
+
"grad_norm": 4.93398904800415,
|
| 1090 |
+
"learning_rate": 9.366008693702535e-06,
|
| 1091 |
+
"loss": 1.9549,
|
| 1092 |
+
"step": 1450
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.24813052345343303,
|
| 1096 |
+
"grad_norm": 10.338335037231445,
|
| 1097 |
+
"learning_rate": 9.351474111737531e-06,
|
| 1098 |
+
"loss": 2.1118,
|
| 1099 |
+
"step": 1460
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.24983004758667574,
|
| 1103 |
+
"grad_norm": 6.729907512664795,
|
| 1104 |
+
"learning_rate": 9.336786349461637e-06,
|
| 1105 |
+
"loss": 2.1022,
|
| 1106 |
+
"step": 1470
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.2515295717199184,
|
| 1110 |
+
"grad_norm": 6.920368194580078,
|
| 1111 |
+
"learning_rate": 9.321945923912518e-06,
|
| 1112 |
+
"loss": 1.7638,
|
| 1113 |
+
"step": 1480
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.2532290958531611,
|
| 1117 |
+
"grad_norm": 8.145246505737305,
|
| 1118 |
+
"learning_rate": 9.306953357501874e-06,
|
| 1119 |
+
"loss": 2.3552,
|
| 1120 |
+
"step": 1490
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 0.2549286199864038,
|
| 1124 |
+
"grad_norm": 7.566394805908203,
|
| 1125 |
+
"learning_rate": 9.291809177997061e-06,
|
| 1126 |
+
"loss": 2.3227,
|
| 1127 |
+
"step": 1500
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 0.2549286199864038,
|
| 1131 |
+
"eval_cosine_accuracy@1": 0.6785,
|
| 1132 |
+
"eval_cosine_accuracy@10": 0.909,
|
| 1133 |
+
"eval_cosine_accuracy@3": 0.8225,
|
| 1134 |
+
"eval_cosine_accuracy@5": 0.863,
|
| 1135 |
+
"eval_cosine_map@100": 0.7639697235001636,
|
| 1136 |
+
"eval_cosine_mrr@10": 0.7603734126984125,
|
| 1137 |
+
"eval_cosine_ndcg@10": 0.7966821763190639,
|
| 1138 |
+
"eval_cosine_precision@1": 0.6785,
|
| 1139 |
+
"eval_cosine_precision@10": 0.09090000000000001,
|
| 1140 |
+
"eval_cosine_precision@3": 0.2741666666666666,
|
| 1141 |
+
"eval_cosine_precision@5": 0.17260000000000003,
|
| 1142 |
+
"eval_cosine_recall@1": 0.6785,
|
| 1143 |
+
"eval_cosine_recall@10": 0.909,
|
| 1144 |
+
"eval_cosine_recall@3": 0.8225,
|
| 1145 |
+
"eval_cosine_recall@5": 0.863,
|
| 1146 |
+
"eval_loss": 1.7280288934707642,
|
| 1147 |
+
"eval_runtime": 2.966,
|
| 1148 |
+
"eval_samples_per_second": 254.217,
|
| 1149 |
+
"eval_sequential_score": 0.7639697235001636,
|
| 1150 |
+
"eval_steps_per_second": 2.023,
|
| 1151 |
+
"eval_sts-dev_pearson_cosine": 0.7956596138775444,
|
| 1152 |
+
"eval_sts-dev_pearson_dot": 0.7898276993726723,
|
| 1153 |
+
"eval_sts-dev_pearson_euclidean": 0.7849266959832228,
|
| 1154 |
+
"eval_sts-dev_pearson_manhattan": 0.7850471763202945,
|
| 1155 |
+
"eval_sts-dev_pearson_max": 0.7956596138775444,
|
| 1156 |
+
"eval_sts-dev_spearman_cosine": 0.8035475738008022,
|
| 1157 |
+
"eval_sts-dev_spearman_dot": 0.7954777202652852,
|
| 1158 |
+
"eval_sts-dev_spearman_euclidean": 0.7982728913757071,
|
| 1159 |
+
"eval_sts-dev_spearman_manhattan": 0.8004630496417331,
|
| 1160 |
+
"eval_sts-dev_spearman_max": 0.8035475738008022,
|
| 1161 |
+
"step": 1500
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.2566281441196465,
|
| 1165 |
+
"grad_norm": 8.947423934936523,
|
| 1166 |
+
"learning_rate": 9.276513918502508e-06,
|
| 1167 |
+
"loss": 2.3639,
|
| 1168 |
+
"step": 1510
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 0.25832766825288916,
|
| 1172 |
+
"grad_norm": 7.069028377532959,
|
| 1173 |
+
"learning_rate": 9.261068117440955e-06,
|
| 1174 |
+
"loss": 1.717,
|
| 1175 |
+
"step": 1520
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 0.26002719238613187,
|
| 1179 |
+
"grad_norm": 9.209027290344238,
|
| 1180 |
+
"learning_rate": 9.24547231853449e-06,
|
| 1181 |
+
"loss": 2.0703,
|
| 1182 |
+
"step": 1530
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 0.26172671651937457,
|
| 1186 |
+
"grad_norm": 6.722319602966309,
|
| 1187 |
+
"learning_rate": 9.229727070785423e-06,
|
| 1188 |
+
"loss": 1.9918,
|
| 1189 |
+
"step": 1540
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 0.26342624065261727,
|
| 1193 |
+
"grad_norm": 7.064426422119141,
|
| 1194 |
+
"learning_rate": 9.21383292845695e-06,
|
| 1195 |
+
"loss": 2.0636,
|
| 1196 |
+
"step": 1550
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.26512576478586,
|
| 1200 |
+
"grad_norm": 5.03215217590332,
|
| 1201 |
+
"learning_rate": 9.197790451053641e-06,
|
| 1202 |
+
"loss": 1.7062,
|
| 1203 |
+
"step": 1560
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.2668252889191027,
|
| 1207 |
+
"grad_norm": 8.875913619995117,
|
| 1208 |
+
"learning_rate": 9.18160020330175e-06,
|
| 1209 |
+
"loss": 2.2969,
|
| 1210 |
+
"step": 1570
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 0.2685248130523453,
|
| 1214 |
+
"grad_norm": 8.441971778869629,
|
| 1215 |
+
"learning_rate": 9.165262755129337e-06,
|
| 1216 |
+
"loss": 1.9633,
|
| 1217 |
+
"step": 1580
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 0.270224337185588,
|
| 1221 |
+
"grad_norm": 6.123201847076416,
|
| 1222 |
+
"learning_rate": 9.148778681646196e-06,
|
| 1223 |
+
"loss": 1.8068,
|
| 1224 |
+
"step": 1590
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.27192386131883073,
|
| 1228 |
+
"grad_norm": 4.153637886047363,
|
| 1229 |
+
"learning_rate": 9.132148563123617e-06,
|
| 1230 |
+
"loss": 1.6155,
|
| 1231 |
+
"step": 1600
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.27362338545207343,
|
| 1235 |
+
"grad_norm": 5.999012470245361,
|
| 1236 |
+
"learning_rate": 9.115372984973963e-06,
|
| 1237 |
+
"loss": 1.6978,
|
| 1238 |
+
"step": 1610
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.27532290958531613,
|
| 1242 |
+
"grad_norm": 7.118943214416504,
|
| 1243 |
+
"learning_rate": 9.098452537730049e-06,
|
| 1244 |
+
"loss": 2.1703,
|
| 1245 |
+
"step": 1620
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.2770224337185588,
|
| 1249 |
+
"grad_norm": 7.293047904968262,
|
| 1250 |
+
"learning_rate": 9.081387817024371e-06,
|
| 1251 |
+
"loss": 2.5395,
|
| 1252 |
+
"step": 1630
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.2787219578518015,
|
| 1256 |
+
"grad_norm": 4.569684982299805,
|
| 1257 |
+
"learning_rate": 9.064179423568122e-06,
|
| 1258 |
+
"loss": 1.7744,
|
| 1259 |
+
"step": 1640
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 0.2804214819850442,
|
| 1263 |
+
"grad_norm": 5.9083404541015625,
|
| 1264 |
+
"learning_rate": 9.046827963130063e-06,
|
| 1265 |
+
"loss": 1.7637,
|
| 1266 |
+
"step": 1650
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 0.2821210061182869,
|
| 1270 |
+
"grad_norm": 7.332223892211914,
|
| 1271 |
+
"learning_rate": 9.02933404651518e-06,
|
| 1272 |
+
"loss": 1.9622,
|
| 1273 |
+
"step": 1660
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.2838205302515296,
|
| 1277 |
+
"grad_norm": 6.511028289794922,
|
| 1278 |
+
"learning_rate": 9.011698289543199e-06,
|
| 1279 |
+
"loss": 1.3597,
|
| 1280 |
+
"step": 1670
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.28552005438477224,
|
| 1284 |
+
"grad_norm": 6.4721832275390625,
|
| 1285 |
+
"learning_rate": 8.9939213130269e-06,
|
| 1286 |
+
"loss": 1.6605,
|
| 1287 |
+
"step": 1680
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 0.28721957851801494,
|
| 1291 |
+
"grad_norm": 7.247612953186035,
|
| 1292 |
+
"learning_rate": 8.97600374275026e-06,
|
| 1293 |
+
"loss": 2.166,
|
| 1294 |
+
"step": 1690
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 0.28891910265125764,
|
| 1298 |
+
"grad_norm": 10.479696273803711,
|
| 1299 |
+
"learning_rate": 8.957946209446435e-06,
|
| 1300 |
+
"loss": 1.959,
|
| 1301 |
+
"step": 1700
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 0.29061862678450034,
|
| 1305 |
+
"grad_norm": 6.533411026000977,
|
| 1306 |
+
"learning_rate": 8.939749348775544e-06,
|
| 1307 |
+
"loss": 2.0912,
|
| 1308 |
+
"step": 1710
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 0.29231815091774305,
|
| 1312 |
+
"grad_norm": 8.847211837768555,
|
| 1313 |
+
"learning_rate": 8.921413801302305e-06,
|
| 1314 |
+
"loss": 2.0446,
|
| 1315 |
+
"step": 1720
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.29401767505098575,
|
| 1319 |
+
"grad_norm": 6.280109882354736,
|
| 1320 |
+
"learning_rate": 8.902940212473477e-06,
|
| 1321 |
+
"loss": 2.1172,
|
| 1322 |
+
"step": 1730
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.2957171991842284,
|
| 1326 |
+
"grad_norm": 6.145227909088135,
|
| 1327 |
+
"learning_rate": 8.884329232595139e-06,
|
| 1328 |
+
"loss": 1.8379,
|
| 1329 |
+
"step": 1740
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.2974167233174711,
|
| 1333 |
+
"grad_norm": 6.113027095794678,
|
| 1334 |
+
"learning_rate": 8.865581516809806e-06,
|
| 1335 |
+
"loss": 2.0789,
|
| 1336 |
+
"step": 1750
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 0.2991162474507138,
|
| 1340 |
+
"grad_norm": 8.188186645507812,
|
| 1341 |
+
"learning_rate": 8.846697725073359e-06,
|
| 1342 |
+
"loss": 1.8014,
|
| 1343 |
+
"step": 1760
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.3008157715839565,
|
| 1347 |
+
"grad_norm": 7.898641109466553,
|
| 1348 |
+
"learning_rate": 8.827678522131815e-06,
|
| 1349 |
+
"loss": 1.7158,
|
| 1350 |
+
"step": 1770
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 0.3025152957171992,
|
| 1354 |
+
"grad_norm": 8.399277687072754,
|
| 1355 |
+
"learning_rate": 8.80852457749793e-06,
|
| 1356 |
+
"loss": 1.785,
|
| 1357 |
+
"step": 1780
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.30421481985044185,
|
| 1361 |
+
"grad_norm": 4.910240650177002,
|
| 1362 |
+
"learning_rate": 8.789236565427627e-06,
|
| 1363 |
+
"loss": 1.9891,
|
| 1364 |
+
"step": 1790
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.30591434398368456,
|
| 1368 |
+
"grad_norm": 10.41693115234375,
|
| 1369 |
+
"learning_rate": 8.769815164896262e-06,
|
| 1370 |
+
"loss": 2.134,
|
| 1371 |
+
"step": 1800
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 0.30761386811692726,
|
| 1375 |
+
"grad_norm": 8.666768074035645,
|
| 1376 |
+
"learning_rate": 8.750261059574722e-06,
|
| 1377 |
+
"loss": 1.8226,
|
| 1378 |
+
"step": 1810
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 0.30931339225016996,
|
| 1382 |
+
"grad_norm": 4.843673229217529,
|
| 1383 |
+
"learning_rate": 8.730574937805363e-06,
|
| 1384 |
+
"loss": 1.9502,
|
| 1385 |
+
"step": 1820
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.31101291638341266,
|
| 1389 |
+
"grad_norm": 6.370514392852783,
|
| 1390 |
+
"learning_rate": 8.71075749257777e-06,
|
| 1391 |
+
"loss": 2.0422,
|
| 1392 |
+
"step": 1830
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 0.3127124405166553,
|
| 1396 |
+
"grad_norm": 4.965259552001953,
|
| 1397 |
+
"learning_rate": 8.690809421504375e-06,
|
| 1398 |
+
"loss": 1.7759,
|
| 1399 |
+
"step": 1840
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 0.314411964649898,
|
| 1403 |
+
"grad_norm": 6.794595241546631,
|
| 1404 |
+
"learning_rate": 8.670731426795888e-06,
|
| 1405 |
+
"loss": 2.0413,
|
| 1406 |
+
"step": 1850
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.3161114887831407,
|
| 1410 |
+
"grad_norm": 4.695109844207764,
|
| 1411 |
+
"learning_rate": 8.65052421523658e-06,
|
| 1412 |
+
"loss": 1.9104,
|
| 1413 |
+
"step": 1860
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 0.3178110129163834,
|
| 1417 |
+
"grad_norm": 6.058196544647217,
|
| 1418 |
+
"learning_rate": 8.63018849815941e-06,
|
| 1419 |
+
"loss": 2.3626,
|
| 1420 |
+
"step": 1870
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 0.3195105370496261,
|
| 1424 |
+
"grad_norm": 4.891868591308594,
|
| 1425 |
+
"learning_rate": 8.609724991420981e-06,
|
| 1426 |
+
"loss": 1.8832,
|
| 1427 |
+
"step": 1880
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 0.3212100611828688,
|
| 1431 |
+
"grad_norm": 5.250426292419434,
|
| 1432 |
+
"learning_rate": 8.589134415376333e-06,
|
| 1433 |
+
"loss": 2.0618,
|
| 1434 |
+
"step": 1890
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.32290958531611147,
|
| 1438 |
+
"grad_norm": 4.848729133605957,
|
| 1439 |
+
"learning_rate": 8.568417494853598e-06,
|
| 1440 |
+
"loss": 2.0059,
|
| 1441 |
+
"step": 1900
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.32460910944935417,
|
| 1445 |
+
"grad_norm": 8.6367769241333,
|
| 1446 |
+
"learning_rate": 8.547574959128472e-06,
|
| 1447 |
+
"loss": 1.9333,
|
| 1448 |
+
"step": 1910
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.3263086335825969,
|
| 1452 |
+
"grad_norm": 5.955835342407227,
|
| 1453 |
+
"learning_rate": 8.526607541898555e-06,
|
| 1454 |
+
"loss": 2.3307,
|
| 1455 |
+
"step": 1920
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 0.3280081577158396,
|
| 1459 |
+
"grad_norm": 7.700503349304199,
|
| 1460 |
+
"learning_rate": 8.505515981257515e-06,
|
| 1461 |
+
"loss": 2.3562,
|
| 1462 |
+
"step": 1930
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 0.3297076818490823,
|
| 1466 |
+
"grad_norm": 6.47861909866333,
|
| 1467 |
+
"learning_rate": 8.48430101966911e-06,
|
| 1468 |
+
"loss": 1.7028,
|
| 1469 |
+
"step": 1940
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 0.3314072059823249,
|
| 1473 |
+
"grad_norm": 6.167905807495117,
|
| 1474 |
+
"learning_rate": 8.462963403941046e-06,
|
| 1475 |
+
"loss": 1.6633,
|
| 1476 |
+
"step": 1950
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 0.33310673011556763,
|
| 1480 |
+
"grad_norm": 5.260265827178955,
|
| 1481 |
+
"learning_rate": 8.441503885198699e-06,
|
| 1482 |
+
"loss": 1.6875,
|
| 1483 |
+
"step": 1960
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 0.33480625424881033,
|
| 1487 |
+
"grad_norm": 5.841998100280762,
|
| 1488 |
+
"learning_rate": 8.419923218858658e-06,
|
| 1489 |
+
"loss": 1.9344,
|
| 1490 |
+
"step": 1970
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.33650577838205303,
|
| 1494 |
+
"grad_norm": 9.941591262817383,
|
| 1495 |
+
"learning_rate": 8.398222164602151e-06,
|
| 1496 |
+
"loss": 2.1813,
|
| 1497 |
+
"step": 1980
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 0.33820530251529574,
|
| 1501 |
+
"grad_norm": 8.419736862182617,
|
| 1502 |
+
"learning_rate": 8.376401486348286e-06,
|
| 1503 |
+
"loss": 1.9077,
|
| 1504 |
+
"step": 1990
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 0.3399048266485384,
|
| 1508 |
+
"grad_norm": 6.201753616333008,
|
| 1509 |
+
"learning_rate": 8.354461952227175e-06,
|
| 1510 |
+
"loss": 1.9249,
|
| 1511 |
+
"step": 2000
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 0.3399048266485384,
|
| 1515 |
+
"eval_cosine_accuracy@1": 0.695,
|
| 1516 |
+
"eval_cosine_accuracy@10": 0.9095,
|
| 1517 |
+
"eval_cosine_accuracy@3": 0.831,
|
| 1518 |
+
"eval_cosine_accuracy@5": 0.871,
|
| 1519 |
+
"eval_cosine_map@100": 0.7742690226366353,
|
| 1520 |
+
"eval_cosine_mrr@10": 0.7707448412698402,
|
| 1521 |
+
"eval_cosine_ndcg@10": 0.8045879491595074,
|
| 1522 |
+
"eval_cosine_precision@1": 0.695,
|
| 1523 |
+
"eval_cosine_precision@10": 0.09095000000000002,
|
| 1524 |
+
"eval_cosine_precision@3": 0.277,
|
| 1525 |
+
"eval_cosine_precision@5": 0.17420000000000002,
|
| 1526 |
+
"eval_cosine_recall@1": 0.695,
|
| 1527 |
+
"eval_cosine_recall@10": 0.9095,
|
| 1528 |
+
"eval_cosine_recall@3": 0.831,
|
| 1529 |
+
"eval_cosine_recall@5": 0.871,
|
| 1530 |
+
"eval_loss": 1.7007904052734375,
|
| 1531 |
+
"eval_runtime": 2.7948,
|
| 1532 |
+
"eval_samples_per_second": 269.79,
|
| 1533 |
+
"eval_sequential_score": 0.7742690226366353,
|
| 1534 |
+
"eval_steps_per_second": 2.147,
|
| 1535 |
+
"eval_sts-dev_pearson_cosine": 0.7952822056751716,
|
| 1536 |
+
"eval_sts-dev_pearson_dot": 0.7915034361455775,
|
| 1537 |
+
"eval_sts-dev_pearson_euclidean": 0.7874083556892966,
|
| 1538 |
+
"eval_sts-dev_pearson_manhattan": 0.7873169677574228,
|
| 1539 |
+
"eval_sts-dev_pearson_max": 0.7952822056751716,
|
| 1540 |
+
"eval_sts-dev_spearman_cosine": 0.8004780454585122,
|
| 1541 |
+
"eval_sts-dev_spearman_dot": 0.7972326119671326,
|
| 1542 |
+
"eval_sts-dev_spearman_euclidean": 0.7979533651258709,
|
| 1543 |
+
"eval_sts-dev_spearman_manhattan": 0.7990561344213346,
|
| 1544 |
+
"eval_sts-dev_spearman_max": 0.8004780454585122,
|
| 1545 |
+
"step": 2000
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 0.3416043507817811,
|
| 1549 |
+
"grad_norm": 5.67038106918335,
|
| 1550 |
+
"learning_rate": 8.332404334552882e-06,
|
| 1551 |
+
"loss": 1.7941,
|
| 1552 |
+
"step": 2010
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 0.3433038749150238,
|
| 1556 |
+
"grad_norm": 5.878876686096191,
|
| 1557 |
+
"learning_rate": 8.310229409796235e-06,
|
| 1558 |
+
"loss": 2.0737,
|
| 1559 |
+
"step": 2020
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 0.3450033990482665,
|
| 1563 |
+
"grad_norm": 5.597275733947754,
|
| 1564 |
+
"learning_rate": 8.287937958557513e-06,
|
| 1565 |
+
"loss": 1.559,
|
| 1566 |
+
"step": 2030
|
| 1567 |
+
},
|
| 1568 |
+
{
|
| 1569 |
+
"epoch": 0.3467029231815092,
|
| 1570 |
+
"grad_norm": 6.382761478424072,
|
| 1571 |
+
"learning_rate": 8.265530765538938e-06,
|
| 1572 |
+
"loss": 1.3173,
|
| 1573 |
+
"step": 2040
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 0.3484024473147519,
|
| 1577 |
+
"grad_norm": 7.027074813842773,
|
| 1578 |
+
"learning_rate": 8.243008619517074e-06,
|
| 1579 |
+
"loss": 2.1037,
|
| 1580 |
+
"step": 2050
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.35010197144799454,
|
| 1584 |
+
"grad_norm": 6.432947158813477,
|
| 1585 |
+
"learning_rate": 8.220372313315053e-06,
|
| 1586 |
+
"loss": 2.27,
|
| 1587 |
+
"step": 2060
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 0.35180149558123724,
|
| 1591 |
+
"grad_norm": 6.3885416984558105,
|
| 1592 |
+
"learning_rate": 8.197622643774663e-06,
|
| 1593 |
+
"loss": 1.7745,
|
| 1594 |
+
"step": 2070
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 0.35350101971447995,
|
| 1598 |
+
"grad_norm": 6.885595798492432,
|
| 1599 |
+
"learning_rate": 8.174760411728308e-06,
|
| 1600 |
+
"loss": 1.3386,
|
| 1601 |
+
"step": 2080
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.35520054384772265,
|
| 1605 |
+
"grad_norm": 7.397289752960205,
|
| 1606 |
+
"learning_rate": 8.1517864219708e-06,
|
| 1607 |
+
"loss": 1.9689,
|
| 1608 |
+
"step": 2090
|
| 1609 |
+
},
|
| 1610 |
+
{
|
| 1611 |
+
"epoch": 0.35690006798096535,
|
| 1612 |
+
"grad_norm": 4.5459885597229,
|
| 1613 |
+
"learning_rate": 8.128701483231048e-06,
|
| 1614 |
+
"loss": 1.699,
|
| 1615 |
+
"step": 2100
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 0.358599592114208,
|
| 1619 |
+
"grad_norm": 4.440462589263916,
|
| 1620 |
+
"learning_rate": 8.105506408143571e-06,
|
| 1621 |
+
"loss": 1.8324,
|
| 1622 |
+
"step": 2110
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.3602991162474507,
|
| 1626 |
+
"grad_norm": 5.693051338195801,
|
| 1627 |
+
"learning_rate": 8.082202013219912e-06,
|
| 1628 |
+
"loss": 1.8712,
|
| 1629 |
+
"step": 2120
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 0.3619986403806934,
|
| 1633 |
+
"grad_norm": 7.692087650299072,
|
| 1634 |
+
"learning_rate": 8.058789118819872e-06,
|
| 1635 |
+
"loss": 2.4458,
|
| 1636 |
+
"step": 2130
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 0.3636981645139361,
|
| 1640 |
+
"grad_norm": 5.480040073394775,
|
| 1641 |
+
"learning_rate": 8.03526854912265e-06,
|
| 1642 |
+
"loss": 1.7687,
|
| 1643 |
+
"step": 2140
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 0.3653976886471788,
|
| 1647 |
+
"grad_norm": 7.727152347564697,
|
| 1648 |
+
"learning_rate": 8.011641132097822e-06,
|
| 1649 |
+
"loss": 1.9936,
|
| 1650 |
+
"step": 2150
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 0.36709721278042146,
|
| 1654 |
+
"grad_norm": 5.268275260925293,
|
| 1655 |
+
"learning_rate": 7.987907699476202e-06,
|
| 1656 |
+
"loss": 1.7563,
|
| 1657 |
+
"step": 2160
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 0.36879673691366416,
|
| 1661 |
+
"grad_norm": 12.53956413269043,
|
| 1662 |
+
"learning_rate": 7.96406908672055e-06,
|
| 1663 |
+
"loss": 2.0094,
|
| 1664 |
+
"step": 2170
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.37049626104690686,
|
| 1668 |
+
"grad_norm": 6.193202495574951,
|
| 1669 |
+
"learning_rate": 7.940126132996172e-06,
|
| 1670 |
+
"loss": 2.0172,
|
| 1671 |
+
"step": 2180
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 0.37219578518014956,
|
| 1675 |
+
"grad_norm": 8.85962200164795,
|
| 1676 |
+
"learning_rate": 7.916079681141385e-06,
|
| 1677 |
+
"loss": 1.9722,
|
| 1678 |
+
"step": 2190
|
| 1679 |
+
},
|
| 1680 |
+
{
|
| 1681 |
+
"epoch": 0.37389530931339227,
|
| 1682 |
+
"grad_norm": 8.260863304138184,
|
| 1683 |
+
"learning_rate": 7.891930577637832e-06,
|
| 1684 |
+
"loss": 1.8102,
|
| 1685 |
+
"step": 2200
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 0.37559483344663497,
|
| 1689 |
+
"grad_norm": 7.650066375732422,
|
| 1690 |
+
"learning_rate": 7.867679672580696e-06,
|
| 1691 |
+
"loss": 2.1561,
|
| 1692 |
+
"step": 2210
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.3772943575798776,
|
| 1696 |
+
"grad_norm": 9.542856216430664,
|
| 1697 |
+
"learning_rate": 7.843327819648774e-06,
|
| 1698 |
+
"loss": 2.1181,
|
| 1699 |
+
"step": 2220
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 0.3789938817131203,
|
| 1703 |
+
"grad_norm": 9.622944831848145,
|
| 1704 |
+
"learning_rate": 7.818875876074421e-06,
|
| 1705 |
+
"loss": 1.9387,
|
| 1706 |
+
"step": 2230
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.380693405846363,
|
| 1710 |
+
"grad_norm": 5.241047382354736,
|
| 1711 |
+
"learning_rate": 7.794324702613376e-06,
|
| 1712 |
+
"loss": 1.7689,
|
| 1713 |
+
"step": 2240
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 0.3823929299796057,
|
| 1717 |
+
"grad_norm": 10.481565475463867,
|
| 1718 |
+
"learning_rate": 7.769675163514463e-06,
|
| 1719 |
+
"loss": 1.9585,
|
| 1720 |
+
"step": 2250
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 0.3840924541128484,
|
| 1724 |
+
"grad_norm": 8.667862892150879,
|
| 1725 |
+
"learning_rate": 7.744928126489169e-06,
|
| 1726 |
+
"loss": 1.5681,
|
| 1727 |
+
"step": 2260
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 0.38579197824609107,
|
| 1731 |
+
"grad_norm": 5.036750793457031,
|
| 1732 |
+
"learning_rate": 7.720084462681092e-06,
|
| 1733 |
+
"loss": 2.0402,
|
| 1734 |
+
"step": 2270
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 0.3874915023793338,
|
| 1738 |
+
"grad_norm": 5.431954383850098,
|
| 1739 |
+
"learning_rate": 7.695145046635281e-06,
|
| 1740 |
+
"loss": 1.78,
|
| 1741 |
+
"step": 2280
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 0.3891910265125765,
|
| 1745 |
+
"grad_norm": 7.302848815917969,
|
| 1746 |
+
"learning_rate": 7.670110756267455e-06,
|
| 1747 |
+
"loss": 1.9524,
|
| 1748 |
+
"step": 2290
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.3908905506458192,
|
| 1752 |
+
"grad_norm": 8.023348808288574,
|
| 1753 |
+
"learning_rate": 7.644982472833083e-06,
|
| 1754 |
+
"loss": 1.9678,
|
| 1755 |
+
"step": 2300
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 0.3925900747790619,
|
| 1759 |
+
"grad_norm": 7.136281967163086,
|
| 1760 |
+
"learning_rate": 7.61976108089638e-06,
|
| 1761 |
+
"loss": 1.6553,
|
| 1762 |
+
"step": 2310
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 0.39428959891230453,
|
| 1766 |
+
"grad_norm": 8.916802406311035,
|
| 1767 |
+
"learning_rate": 7.594447468299156e-06,
|
| 1768 |
+
"loss": 1.8378,
|
| 1769 |
+
"step": 2320
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 0.39598912304554723,
|
| 1773 |
+
"grad_norm": 8.570812225341797,
|
| 1774 |
+
"learning_rate": 7.56904252612957e-06,
|
| 1775 |
+
"loss": 1.8066,
|
| 1776 |
+
"step": 2330
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.39768864717878993,
|
| 1780 |
+
"grad_norm": 10.376533508300781,
|
| 1781 |
+
"learning_rate": 7.543547148690757e-06,
|
| 1782 |
+
"loss": 1.8585,
|
| 1783 |
+
"step": 2340
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.39938817131203264,
|
| 1787 |
+
"grad_norm": 5.894679546356201,
|
| 1788 |
+
"learning_rate": 7.517962233469345e-06,
|
| 1789 |
+
"loss": 1.4762,
|
| 1790 |
+
"step": 2350
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.40108769544527534,
|
| 1794 |
+
"grad_norm": 6.78580904006958,
|
| 1795 |
+
"learning_rate": 7.492288681103872e-06,
|
| 1796 |
+
"loss": 1.6828,
|
| 1797 |
+
"step": 2360
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 0.40278721957851804,
|
| 1801 |
+
"grad_norm": 6.131538391113281,
|
| 1802 |
+
"learning_rate": 7.4665273953530695e-06,
|
| 1803 |
+
"loss": 1.6288,
|
| 1804 |
+
"step": 2370
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 0.4044867437117607,
|
| 1808 |
+
"grad_norm": 5.721993923187256,
|
| 1809 |
+
"learning_rate": 7.440679283064059e-06,
|
| 1810 |
+
"loss": 1.7679,
|
| 1811 |
+
"step": 2380
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 0.4061862678450034,
|
| 1815 |
+
"grad_norm": 6.973949909210205,
|
| 1816 |
+
"learning_rate": 7.41474525414042e-06,
|
| 1817 |
+
"loss": 2.128,
|
| 1818 |
+
"step": 2390
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 0.4078857919782461,
|
| 1822 |
+
"grad_norm": 6.6178202629089355,
|
| 1823 |
+
"learning_rate": 7.388726221510163e-06,
|
| 1824 |
+
"loss": 1.3543,
|
| 1825 |
+
"step": 2400
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 0.4095853161114888,
|
| 1829 |
+
"grad_norm": 5.370213985443115,
|
| 1830 |
+
"learning_rate": 7.3626231010935974e-06,
|
| 1831 |
+
"loss": 1.806,
|
| 1832 |
+
"step": 2410
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.4112848402447315,
|
| 1836 |
+
"grad_norm": 8.617044448852539,
|
| 1837 |
+
"learning_rate": 7.336436811771085e-06,
|
| 1838 |
+
"loss": 1.8597,
|
| 1839 |
+
"step": 2420
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.41298436437797414,
|
| 1843 |
+
"grad_norm": 12.804397583007812,
|
| 1844 |
+
"learning_rate": 7.310168275350692e-06,
|
| 1845 |
+
"loss": 2.2319,
|
| 1846 |
+
"step": 2430
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.41468388851121685,
|
| 1850 |
+
"grad_norm": 8.611510276794434,
|
| 1851 |
+
"learning_rate": 7.28381841653574e-06,
|
| 1852 |
+
"loss": 1.9433,
|
| 1853 |
+
"step": 2440
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.41638341264445955,
|
| 1857 |
+
"grad_norm": 5.746379375457764,
|
| 1858 |
+
"learning_rate": 7.257388162892261e-06,
|
| 1859 |
+
"loss": 1.7124,
|
| 1860 |
+
"step": 2450
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.41808293677770225,
|
| 1864 |
+
"grad_norm": 9.15091609954834,
|
| 1865 |
+
"learning_rate": 7.23087844481634e-06,
|
| 1866 |
+
"loss": 1.702,
|
| 1867 |
+
"step": 2460
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.41978246091094495,
|
| 1871 |
+
"grad_norm": 5.277801513671875,
|
| 1872 |
+
"learning_rate": 7.204290195501358e-06,
|
| 1873 |
+
"loss": 1.5501,
|
| 1874 |
+
"step": 2470
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.4214819850441876,
|
| 1878 |
+
"grad_norm": 9.658906936645508,
|
| 1879 |
+
"learning_rate": 7.177624350905155e-06,
|
| 1880 |
+
"loss": 2.0232,
|
| 1881 |
+
"step": 2480
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.4231815091774303,
|
| 1885 |
+
"grad_norm": 5.400667667388916,
|
| 1886 |
+
"learning_rate": 7.153559522573959e-06,
|
| 1887 |
+
"loss": 1.9549,
|
| 1888 |
+
"step": 2490
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.424881033310673,
|
| 1892 |
+
"grad_norm": 6.810690402984619,
|
| 1893 |
+
"learning_rate": 7.126748835263528e-06,
|
| 1894 |
+
"loss": 1.9216,
|
| 1895 |
+
"step": 2500
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.424881033310673,
|
| 1899 |
+
"eval_cosine_accuracy@1": 0.706,
|
| 1900 |
+
"eval_cosine_accuracy@10": 0.912,
|
| 1901 |
+
"eval_cosine_accuracy@3": 0.8375,
|
| 1902 |
+
"eval_cosine_accuracy@5": 0.87,
|
| 1903 |
+
"eval_cosine_map@100": 0.782229618428024,
|
| 1904 |
+
"eval_cosine_mrr@10": 0.7788738095238087,
|
| 1905 |
+
"eval_cosine_ndcg@10": 0.811319534720964,
|
| 1906 |
+
"eval_cosine_precision@1": 0.706,
|
| 1907 |
+
"eval_cosine_precision@10": 0.09120000000000002,
|
| 1908 |
+
"eval_cosine_precision@3": 0.2791666666666666,
|
| 1909 |
+
"eval_cosine_precision@5": 0.174,
|
| 1910 |
+
"eval_cosine_recall@1": 0.706,
|
| 1911 |
+
"eval_cosine_recall@10": 0.912,
|
| 1912 |
+
"eval_cosine_recall@3": 0.8375,
|
| 1913 |
+
"eval_cosine_recall@5": 0.87,
|
| 1914 |
+
"eval_loss": 1.6223056316375732,
|
| 1915 |
+
"eval_runtime": 2.7895,
|
| 1916 |
+
"eval_samples_per_second": 270.299,
|
| 1917 |
+
"eval_sequential_score": 0.782229618428024,
|
| 1918 |
+
"eval_steps_per_second": 2.151,
|
| 1919 |
+
"eval_sts-dev_pearson_cosine": 0.7982554363739512,
|
| 1920 |
+
"eval_sts-dev_pearson_dot": 0.7952587919868024,
|
| 1921 |
+
"eval_sts-dev_pearson_euclidean": 0.7869234232815028,
|
| 1922 |
+
"eval_sts-dev_pearson_manhattan": 0.7872366700437463,
|
| 1923 |
+
"eval_sts-dev_pearson_max": 0.7982554363739512,
|
| 1924 |
+
"eval_sts-dev_spearman_cosine": 0.8022561650534226,
|
| 1925 |
+
"eval_sts-dev_spearman_dot": 0.7985926145853256,
|
| 1926 |
+
"eval_sts-dev_spearman_euclidean": 0.8008740888247353,
|
| 1927 |
+
"eval_sts-dev_spearman_manhattan": 0.801728465873034,
|
| 1928 |
+
"eval_sts-dev_spearman_max": 0.8022561650534226,
|
| 1929 |
+
"step": 2500
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.4265805574439157,
|
| 1933 |
+
"grad_norm": 6.482539653778076,
|
| 1934 |
+
"learning_rate": 7.099863282277834e-06,
|
| 1935 |
+
"loss": 1.5548,
|
| 1936 |
+
"step": 2510
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.4282800815771584,
|
| 1940 |
+
"grad_norm": 7.853833198547363,
|
| 1941 |
+
"learning_rate": 7.072903810040381e-06,
|
| 1942 |
+
"loss": 1.6629,
|
| 1943 |
+
"step": 2520
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 0.4299796057104011,
|
| 1947 |
+
"grad_norm": 6.526813507080078,
|
| 1948 |
+
"learning_rate": 7.0458713675767665e-06,
|
| 1949 |
+
"loss": 1.9463,
|
| 1950 |
+
"step": 2530
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.43167912984364376,
|
| 1954 |
+
"grad_norm": 9.038359642028809,
|
| 1955 |
+
"learning_rate": 7.018766906481288e-06,
|
| 1956 |
+
"loss": 2.16,
|
| 1957 |
+
"step": 2540
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 0.43337865397688646,
|
| 1961 |
+
"grad_norm": 5.049062728881836,
|
| 1962 |
+
"learning_rate": 6.991591380883435e-06,
|
| 1963 |
+
"loss": 1.6674,
|
| 1964 |
+
"step": 2550
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.43507817811012917,
|
| 1968 |
+
"grad_norm": 6.659489154815674,
|
| 1969 |
+
"learning_rate": 6.964345747414307e-06,
|
| 1970 |
+
"loss": 1.7055,
|
| 1971 |
+
"step": 2560
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.43677770224337187,
|
| 1975 |
+
"grad_norm": 8.091437339782715,
|
| 1976 |
+
"learning_rate": 6.937030965172935e-06,
|
| 1977 |
+
"loss": 1.8062,
|
| 1978 |
+
"step": 2570
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 0.43847722637661457,
|
| 1982 |
+
"grad_norm": 7.644534587860107,
|
| 1983 |
+
"learning_rate": 6.909647995692522e-06,
|
| 1984 |
+
"loss": 2.1972,
|
| 1985 |
+
"step": 2580
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 0.4401767505098572,
|
| 1989 |
+
"grad_norm": 7.349899768829346,
|
| 1990 |
+
"learning_rate": 6.88219780290659e-06,
|
| 1991 |
+
"loss": 1.9069,
|
| 1992 |
+
"step": 2590
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 0.4418762746430999,
|
| 1996 |
+
"grad_norm": 8.312362670898438,
|
| 1997 |
+
"learning_rate": 6.854681353115056e-06,
|
| 1998 |
+
"loss": 2.6629,
|
| 1999 |
+
"step": 2600
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 0.4435757987763426,
|
| 2003 |
+
"grad_norm": 8.937472343444824,
|
| 2004 |
+
"learning_rate": 6.82709961495021e-06,
|
| 2005 |
+
"loss": 2.4573,
|
| 2006 |
+
"step": 2610
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.4452753229095853,
|
| 2010 |
+
"grad_norm": 6.268832683563232,
|
| 2011 |
+
"learning_rate": 6.799453559342619e-06,
|
| 2012 |
+
"loss": 2.0894,
|
| 2013 |
+
"step": 2620
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.446974847042828,
|
| 2017 |
+
"grad_norm": 9.547843933105469,
|
| 2018 |
+
"learning_rate": 6.771744159486947e-06,
|
| 2019 |
+
"loss": 1.5755,
|
| 2020 |
+
"step": 2630
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 0.4486743711760707,
|
| 2024 |
+
"grad_norm": 8.299346923828125,
|
| 2025 |
+
"learning_rate": 6.7439723908077e-06,
|
| 2026 |
+
"loss": 2.3673,
|
| 2027 |
+
"step": 2640
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.4503738953093134,
|
| 2031 |
+
"grad_norm": 6.921226978302002,
|
| 2032 |
+
"learning_rate": 6.716139230924891e-06,
|
| 2033 |
+
"loss": 1.8997,
|
| 2034 |
+
"step": 2650
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 0.4520734194425561,
|
| 2038 |
+
"grad_norm": 6.874094486236572,
|
| 2039 |
+
"learning_rate": 6.6882456596196125e-06,
|
| 2040 |
+
"loss": 2.0955,
|
| 2041 |
+
"step": 2660
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.4537729435757988,
|
| 2045 |
+
"grad_norm": 7.336355686187744,
|
| 2046 |
+
"learning_rate": 6.660292658799565e-06,
|
| 2047 |
+
"loss": 2.0553,
|
| 2048 |
+
"step": 2670
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.4554724677090415,
|
| 2052 |
+
"grad_norm": 6.311689376831055,
|
| 2053 |
+
"learning_rate": 6.632281212464479e-06,
|
| 2054 |
+
"loss": 1.9258,
|
| 2055 |
+
"step": 2680
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.4571719918422842,
|
| 2059 |
+
"grad_norm": 5.280473232269287,
|
| 2060 |
+
"learning_rate": 6.60421230667148e-06,
|
| 2061 |
+
"loss": 1.7009,
|
| 2062 |
+
"step": 2690
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 0.45887151597552683,
|
| 2066 |
+
"grad_norm": 5.1102986335754395,
|
| 2067 |
+
"learning_rate": 6.5760869295003796e-06,
|
| 2068 |
+
"loss": 1.9853,
|
| 2069 |
+
"step": 2700
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 0.46057104010876954,
|
| 2073 |
+
"grad_norm": 6.345133304595947,
|
| 2074 |
+
"learning_rate": 6.54790607101889e-06,
|
| 2075 |
+
"loss": 1.8553,
|
| 2076 |
+
"step": 2710
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 0.46227056424201224,
|
| 2080 |
+
"grad_norm": 9.678619384765625,
|
| 2081 |
+
"learning_rate": 6.519670723247773e-06,
|
| 2082 |
+
"loss": 1.7589,
|
| 2083 |
+
"step": 2720
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 0.46397008837525494,
|
| 2087 |
+
"grad_norm": 4.833879470825195,
|
| 2088 |
+
"learning_rate": 6.491381880125916e-06,
|
| 2089 |
+
"loss": 1.827,
|
| 2090 |
+
"step": 2730
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.46566961250849764,
|
| 2094 |
+
"grad_norm": 6.266608238220215,
|
| 2095 |
+
"learning_rate": 6.463040537475354e-06,
|
| 2096 |
+
"loss": 1.7476,
|
| 2097 |
+
"step": 2740
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.4673691366417403,
|
| 2101 |
+
"grad_norm": 8.750253677368164,
|
| 2102 |
+
"learning_rate": 6.434647692966199e-06,
|
| 2103 |
+
"loss": 2.1204,
|
| 2104 |
+
"step": 2750
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 0.469068660774983,
|
| 2108 |
+
"grad_norm": 5.197884559631348,
|
| 2109 |
+
"learning_rate": 6.406204346081531e-06,
|
| 2110 |
+
"loss": 1.5477,
|
| 2111 |
+
"step": 2760
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 0.4707681849082257,
|
| 2115 |
+
"grad_norm": 7.584977626800537,
|
| 2116 |
+
"learning_rate": 6.377711498082213e-06,
|
| 2117 |
+
"loss": 1.9078,
|
| 2118 |
+
"step": 2770
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.4724677090414684,
|
| 2122 |
+
"grad_norm": 5.226884365081787,
|
| 2123 |
+
"learning_rate": 6.349170151971641e-06,
|
| 2124 |
+
"loss": 1.53,
|
| 2125 |
+
"step": 2780
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.4741672331747111,
|
| 2129 |
+
"grad_norm": 9.499238967895508,
|
| 2130 |
+
"learning_rate": 6.320581312460439e-06,
|
| 2131 |
+
"loss": 1.7868,
|
| 2132 |
+
"step": 2790
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.47586675730795375,
|
| 2136 |
+
"grad_norm": 6.44546365737915,
|
| 2137 |
+
"learning_rate": 6.291945985931093e-06,
|
| 2138 |
+
"loss": 1.7781,
|
| 2139 |
+
"step": 2800
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.47756628144119645,
|
| 2143 |
+
"grad_norm": 5.152192115783691,
|
| 2144 |
+
"learning_rate": 6.263265180402517e-06,
|
| 2145 |
+
"loss": 2.0043,
|
| 2146 |
+
"step": 2810
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 0.47926580557443915,
|
| 2150 |
+
"grad_norm": 7.129688262939453,
|
| 2151 |
+
"learning_rate": 6.234539905494576e-06,
|
| 2152 |
+
"loss": 1.9584,
|
| 2153 |
+
"step": 2820
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 0.48096532970768185,
|
| 2157 |
+
"grad_norm": 6.289784908294678,
|
| 2158 |
+
"learning_rate": 6.2057711723925455e-06,
|
| 2159 |
+
"loss": 1.9584,
|
| 2160 |
+
"step": 2830
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 0.48266485384092456,
|
| 2164 |
+
"grad_norm": 7.36097526550293,
|
| 2165 |
+
"learning_rate": 6.176959993811511e-06,
|
| 2166 |
+
"loss": 1.9952,
|
| 2167 |
+
"step": 2840
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 0.48436437797416726,
|
| 2171 |
+
"grad_norm": 4.622279644012451,
|
| 2172 |
+
"learning_rate": 6.148107383960723e-06,
|
| 2173 |
+
"loss": 1.9764,
|
| 2174 |
+
"step": 2850
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.4860639021074099,
|
| 2178 |
+
"grad_norm": 8.651782989501953,
|
| 2179 |
+
"learning_rate": 6.119214358507889e-06,
|
| 2180 |
+
"loss": 1.8804,
|
| 2181 |
+
"step": 2860
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.4877634262406526,
|
| 2185 |
+
"grad_norm": 6.295717716217041,
|
| 2186 |
+
"learning_rate": 6.090281934543429e-06,
|
| 2187 |
+
"loss": 1.8634,
|
| 2188 |
+
"step": 2870
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 0.4894629503738953,
|
| 2192 |
+
"grad_norm": 5.505663871765137,
|
| 2193 |
+
"learning_rate": 6.061311130544664e-06,
|
| 2194 |
+
"loss": 1.8853,
|
| 2195 |
+
"step": 2880
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 0.491162474507138,
|
| 2199 |
+
"grad_norm": 10.88358211517334,
|
| 2200 |
+
"learning_rate": 6.032302966339965e-06,
|
| 2201 |
+
"loss": 2.2094,
|
| 2202 |
+
"step": 2890
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 0.4928619986403807,
|
| 2206 |
+
"grad_norm": 6.576531410217285,
|
| 2207 |
+
"learning_rate": 6.003258463072859e-06,
|
| 2208 |
+
"loss": 1.5955,
|
| 2209 |
+
"step": 2900
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.49456152277362336,
|
| 2213 |
+
"grad_norm": 5.74630880355835,
|
| 2214 |
+
"learning_rate": 5.974178643166075e-06,
|
| 2215 |
+
"loss": 1.8756,
|
| 2216 |
+
"step": 2910
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.49626104690686607,
|
| 2220 |
+
"grad_norm": 8.021421432495117,
|
| 2221 |
+
"learning_rate": 5.945064530285556e-06,
|
| 2222 |
+
"loss": 1.9114,
|
| 2223 |
+
"step": 2920
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.49796057104010877,
|
| 2227 |
+
"grad_norm": 4.470407009124756,
|
| 2228 |
+
"learning_rate": 5.915917149304424e-06,
|
| 2229 |
+
"loss": 2.2359,
|
| 2230 |
+
"step": 2930
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 0.49966009517335147,
|
| 2234 |
+
"grad_norm": 8.264505386352539,
|
| 2235 |
+
"learning_rate": 5.8867375262669076e-06,
|
| 2236 |
+
"loss": 1.7543,
|
| 2237 |
+
"step": 2940
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 0.5013596193065941,
|
| 2241 |
+
"grad_norm": 6.8826069831848145,
|
| 2242 |
+
"learning_rate": 5.857526688352211e-06,
|
| 2243 |
+
"loss": 1.7754,
|
| 2244 |
+
"step": 2950
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 0.5030591434398368,
|
| 2248 |
+
"grad_norm": 7.148299694061279,
|
| 2249 |
+
"learning_rate": 5.828285663838365e-06,
|
| 2250 |
+
"loss": 2.1194,
|
| 2251 |
+
"step": 2960
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 0.5047586675730795,
|
| 2255 |
+
"grad_norm": 6.472208499908447,
|
| 2256 |
+
"learning_rate": 5.7990154820660295e-06,
|
| 2257 |
+
"loss": 1.7127,
|
| 2258 |
+
"step": 2970
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.5064581917063222,
|
| 2262 |
+
"grad_norm": 8.591309547424316,
|
| 2263 |
+
"learning_rate": 5.769717173402253e-06,
|
| 2264 |
+
"loss": 1.6715,
|
| 2265 |
+
"step": 2980
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.5081577158395649,
|
| 2269 |
+
"grad_norm": 5.60316801071167,
|
| 2270 |
+
"learning_rate": 5.740391769204209e-06,
|
| 2271 |
+
"loss": 1.4632,
|
| 2272 |
+
"step": 2990
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 0.5098572399728076,
|
| 2276 |
+
"grad_norm": 5.79714822769165,
|
| 2277 |
+
"learning_rate": 5.71104030178288e-06,
|
| 2278 |
+
"loss": 1.7171,
|
| 2279 |
+
"step": 3000
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 0.5098572399728076,
|
| 2283 |
+
"eval_cosine_accuracy@1": 0.707,
|
| 2284 |
+
"eval_cosine_accuracy@10": 0.915,
|
| 2285 |
+
"eval_cosine_accuracy@3": 0.839,
|
| 2286 |
+
"eval_cosine_accuracy@5": 0.874,
|
| 2287 |
+
"eval_cosine_map@100": 0.7842834572194356,
|
| 2288 |
+
"eval_cosine_mrr@10": 0.7809134920634915,
|
| 2289 |
+
"eval_cosine_ndcg@10": 0.8136383995155022,
|
| 2290 |
+
"eval_cosine_precision@1": 0.707,
|
| 2291 |
+
"eval_cosine_precision@10": 0.0915,
|
| 2292 |
+
"eval_cosine_precision@3": 0.2796666666666666,
|
| 2293 |
+
"eval_cosine_precision@5": 0.1748,
|
| 2294 |
+
"eval_cosine_recall@1": 0.707,
|
| 2295 |
+
"eval_cosine_recall@10": 0.915,
|
| 2296 |
+
"eval_cosine_recall@3": 0.839,
|
| 2297 |
+
"eval_cosine_recall@5": 0.874,
|
| 2298 |
+
"eval_loss": 1.6132309436798096,
|
| 2299 |
+
"eval_runtime": 2.7954,
|
| 2300 |
+
"eval_samples_per_second": 269.727,
|
| 2301 |
+
"eval_sequential_score": 0.7842834572194356,
|
| 2302 |
+
"eval_steps_per_second": 2.146,
|
| 2303 |
+
"eval_sts-dev_pearson_cosine": 0.8056038655154354,
|
| 2304 |
+
"eval_sts-dev_pearson_dot": 0.8019913183334593,
|
| 2305 |
+
"eval_sts-dev_pearson_euclidean": 0.7922504417572196,
|
| 2306 |
+
"eval_sts-dev_pearson_manhattan": 0.7937333279050955,
|
| 2307 |
+
"eval_sts-dev_pearson_max": 0.8056038655154354,
|
| 2308 |
+
"eval_sts-dev_spearman_cosine": 0.8083570013976158,
|
| 2309 |
+
"eval_sts-dev_spearman_dot": 0.8047381961082496,
|
| 2310 |
+
"eval_sts-dev_spearman_euclidean": 0.8071350345841625,
|
| 2311 |
+
"eval_sts-dev_spearman_manhattan": 0.8069596988802814,
|
| 2312 |
+
"eval_sts-dev_spearman_max": 0.8083570013976158,
|
| 2313 |
+
"step": 3000
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.5115567641060503,
|
| 2317 |
+
"grad_norm": 6.9394049644470215,
|
| 2318 |
+
"learning_rate": 5.6816638043667294e-06,
|
| 2319 |
+
"loss": 1.4526,
|
| 2320 |
+
"step": 3010
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.513256288239293,
|
| 2324 |
+
"grad_norm": 10.348281860351562,
|
| 2325 |
+
"learning_rate": 5.652263311065325e-06,
|
| 2326 |
+
"loss": 1.8309,
|
| 2327 |
+
"step": 3020
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.5149558123725357,
|
| 2331 |
+
"grad_norm": 8.404902458190918,
|
| 2332 |
+
"learning_rate": 5.6228398568329344e-06,
|
| 2333 |
+
"loss": 1.9336,
|
| 2334 |
+
"step": 3030
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.5166553365057783,
|
| 2338 |
+
"grad_norm": 6.444024562835693,
|
| 2339 |
+
"learning_rate": 5.593394477432096e-06,
|
| 2340 |
+
"loss": 1.7704,
|
| 2341 |
+
"step": 3040
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.518354860639021,
|
| 2345 |
+
"grad_norm": 9.629279136657715,
|
| 2346 |
+
"learning_rate": 5.563928209397151e-06,
|
| 2347 |
+
"loss": 1.943,
|
| 2348 |
+
"step": 3050
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.5200543847722637,
|
| 2352 |
+
"grad_norm": 6.889398097991943,
|
| 2353 |
+
"learning_rate": 5.534442089997771e-06,
|
| 2354 |
+
"loss": 1.8689,
|
| 2355 |
+
"step": 3060
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.5217539089055064,
|
| 2359 |
+
"grad_norm": 7.757649898529053,
|
| 2360 |
+
"learning_rate": 5.5049371572024235e-06,
|
| 2361 |
+
"loss": 1.8908,
|
| 2362 |
+
"step": 3070
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.5234534330387491,
|
| 2366 |
+
"grad_norm": 6.982544422149658,
|
| 2367 |
+
"learning_rate": 5.475414449641852e-06,
|
| 2368 |
+
"loss": 1.7924,
|
| 2369 |
+
"step": 3080
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.5251529571719918,
|
| 2373 |
+
"grad_norm": 7.109350204467773,
|
| 2374 |
+
"learning_rate": 5.445875006572502e-06,
|
| 2375 |
+
"loss": 1.7891,
|
| 2376 |
+
"step": 3090
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.5268524813052345,
|
| 2380 |
+
"grad_norm": 7.507669448852539,
|
| 2381 |
+
"learning_rate": 5.416319867839943e-06,
|
| 2382 |
+
"loss": 2.0208,
|
| 2383 |
+
"step": 3100
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.5285520054384772,
|
| 2387 |
+
"grad_norm": 4.527754783630371,
|
| 2388 |
+
"learning_rate": 5.386750073842259e-06,
|
| 2389 |
+
"loss": 1.5862,
|
| 2390 |
+
"step": 3110
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.53025152957172,
|
| 2394 |
+
"grad_norm": 8.16139030456543,
|
| 2395 |
+
"learning_rate": 5.357166665493434e-06,
|
| 2396 |
+
"loss": 2.2195,
|
| 2397 |
+
"step": 3120
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.5319510537049627,
|
| 2401 |
+
"grad_norm": 7.385548114776611,
|
| 2402 |
+
"learning_rate": 5.327570684186697e-06,
|
| 2403 |
+
"loss": 1.6103,
|
| 2404 |
+
"step": 3130
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.5336505778382054,
|
| 2408 |
+
"grad_norm": 7.784885406494141,
|
| 2409 |
+
"learning_rate": 5.297963171757874e-06,
|
| 2410 |
+
"loss": 2.5839,
|
| 2411 |
+
"step": 3140
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.535350101971448,
|
| 2415 |
+
"grad_norm": 7.986569404602051,
|
| 2416 |
+
"learning_rate": 5.268345170448704e-06,
|
| 2417 |
+
"loss": 2.4304,
|
| 2418 |
+
"step": 3150
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.5370496261046906,
|
| 2422 |
+
"grad_norm": 7.29852294921875,
|
| 2423 |
+
"learning_rate": 5.23871772287016e-06,
|
| 2424 |
+
"loss": 1.9009,
|
| 2425 |
+
"step": 3160
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.5387491502379333,
|
| 2429 |
+
"grad_norm": 8.884086608886719,
|
| 2430 |
+
"learning_rate": 5.209081871965741e-06,
|
| 2431 |
+
"loss": 2.2161,
|
| 2432 |
+
"step": 3170
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.540448674371176,
|
| 2436 |
+
"grad_norm": 6.930916786193848,
|
| 2437 |
+
"learning_rate": 5.179438660974756e-06,
|
| 2438 |
+
"loss": 1.9054,
|
| 2439 |
+
"step": 3180
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.5421481985044188,
|
| 2443 |
+
"grad_norm": 6.724717617034912,
|
| 2444 |
+
"learning_rate": 5.149789133395606e-06,
|
| 2445 |
+
"loss": 1.3603,
|
| 2446 |
+
"step": 3190
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.5438477226376615,
|
| 2450 |
+
"grad_norm": 11.375778198242188,
|
| 2451 |
+
"learning_rate": 5.120134332949049e-06,
|
| 2452 |
+
"loss": 2.1651,
|
| 2453 |
+
"step": 3200
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.5455472467709042,
|
| 2457 |
+
"grad_norm": 9.330193519592285,
|
| 2458 |
+
"learning_rate": 5.0904753035414545e-06,
|
| 2459 |
+
"loss": 2.0636,
|
| 2460 |
+
"step": 3210
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.5472467709041469,
|
| 2464 |
+
"grad_norm": 7.2718000411987305,
|
| 2465 |
+
"learning_rate": 5.060813089228064e-06,
|
| 2466 |
+
"loss": 2.0036,
|
| 2467 |
+
"step": 3220
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.5489462950373896,
|
| 2471 |
+
"grad_norm": 5.489546775817871,
|
| 2472 |
+
"learning_rate": 5.0311487341762285e-06,
|
| 2473 |
+
"loss": 1.7724,
|
| 2474 |
+
"step": 3230
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.5506458191706323,
|
| 2478 |
+
"grad_norm": 7.590505123138428,
|
| 2479 |
+
"learning_rate": 5.001483282628664e-06,
|
| 2480 |
+
"loss": 1.2798,
|
| 2481 |
+
"step": 3240
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.552345343303875,
|
| 2485 |
+
"grad_norm": 5.193188190460205,
|
| 2486 |
+
"learning_rate": 4.971817778866681e-06,
|
| 2487 |
+
"loss": 1.8975,
|
| 2488 |
+
"step": 3250
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.5540448674371176,
|
| 2492 |
+
"grad_norm": 4.518863677978516,
|
| 2493 |
+
"learning_rate": 4.9421532671734265e-06,
|
| 2494 |
+
"loss": 1.8153,
|
| 2495 |
+
"step": 3260
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.5557443915703603,
|
| 2499 |
+
"grad_norm": 7.619812488555908,
|
| 2500 |
+
"learning_rate": 4.912490791797129e-06,
|
| 2501 |
+
"loss": 1.6361,
|
| 2502 |
+
"step": 3270
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.557443915703603,
|
| 2506 |
+
"grad_norm": 8.381867408752441,
|
| 2507 |
+
"learning_rate": 4.882831396914332e-06,
|
| 2508 |
+
"loss": 2.2208,
|
| 2509 |
+
"step": 3280
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.5591434398368457,
|
| 2513 |
+
"grad_norm": 4.878629684448242,
|
| 2514 |
+
"learning_rate": 4.853176126593139e-06,
|
| 2515 |
+
"loss": 1.7356,
|
| 2516 |
+
"step": 3290
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.5608429639700884,
|
| 2520 |
+
"grad_norm": 6.435664653778076,
|
| 2521 |
+
"learning_rate": 4.823526024756464e-06,
|
| 2522 |
+
"loss": 1.8563,
|
| 2523 |
+
"step": 3300
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.5625424881033311,
|
| 2527 |
+
"grad_norm": 6.569293022155762,
|
| 2528 |
+
"learning_rate": 4.793882135145278e-06,
|
| 2529 |
+
"loss": 1.8743,
|
| 2530 |
+
"step": 3310
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.5642420122365738,
|
| 2534 |
+
"grad_norm": 6.728095054626465,
|
| 2535 |
+
"learning_rate": 4.764245501281869e-06,
|
| 2536 |
+
"loss": 1.7494,
|
| 2537 |
+
"step": 3320
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.5659415363698165,
|
| 2541 |
+
"grad_norm": 6.320137977600098,
|
| 2542 |
+
"learning_rate": 4.734617166433113e-06,
|
| 2543 |
+
"loss": 1.5363,
|
| 2544 |
+
"step": 3330
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.5676410605030592,
|
| 2548 |
+
"grad_norm": 5.589328765869141,
|
| 2549 |
+
"learning_rate": 4.7049981735737385e-06,
|
| 2550 |
+
"loss": 2.0732,
|
| 2551 |
+
"step": 3340
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.5693405846363019,
|
| 2555 |
+
"grad_norm": 6.616926193237305,
|
| 2556 |
+
"learning_rate": 4.67538956534962e-06,
|
| 2557 |
+
"loss": 1.7303,
|
| 2558 |
+
"step": 3350
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.5710401087695445,
|
| 2562 |
+
"grad_norm": 4.177309036254883,
|
| 2563 |
+
"learning_rate": 4.645792384041078e-06,
|
| 2564 |
+
"loss": 1.532,
|
| 2565 |
+
"step": 3360
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.5727396329027872,
|
| 2569 |
+
"grad_norm": 6.281280517578125,
|
| 2570 |
+
"learning_rate": 4.616207671526175e-06,
|
| 2571 |
+
"loss": 2.2504,
|
| 2572 |
+
"step": 3370
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.5744391570360299,
|
| 2576 |
+
"grad_norm": 6.781736373901367,
|
| 2577 |
+
"learning_rate": 4.586636469244054e-06,
|
| 2578 |
+
"loss": 1.6377,
|
| 2579 |
+
"step": 3380
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.5761386811692726,
|
| 2583 |
+
"grad_norm": 8.028448104858398,
|
| 2584 |
+
"learning_rate": 4.557079818158269e-06,
|
| 2585 |
+
"loss": 2.1332,
|
| 2586 |
+
"step": 3390
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.5778382053025153,
|
| 2590 |
+
"grad_norm": 5.485851764678955,
|
| 2591 |
+
"learning_rate": 4.527538758720144e-06,
|
| 2592 |
+
"loss": 1.8658,
|
| 2593 |
+
"step": 3400
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.579537729435758,
|
| 2597 |
+
"grad_norm": 5.404677867889404,
|
| 2598 |
+
"learning_rate": 4.498014330832148e-06,
|
| 2599 |
+
"loss": 1.8165,
|
| 2600 |
+
"step": 3410
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.5812372535690007,
|
| 2604 |
+
"grad_norm": 5.938612461090088,
|
| 2605 |
+
"learning_rate": 4.468507573811285e-06,
|
| 2606 |
+
"loss": 1.7847,
|
| 2607 |
+
"step": 3420
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.5829367777022434,
|
| 2611 |
+
"grad_norm": 6.639806270599365,
|
| 2612 |
+
"learning_rate": 4.439019526352511e-06,
|
| 2613 |
+
"loss": 1.747,
|
| 2614 |
+
"step": 3430
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 0.5846363018354861,
|
| 2618 |
+
"grad_norm": 10.548473358154297,
|
| 2619 |
+
"learning_rate": 4.409551226492173e-06,
|
| 2620 |
+
"loss": 1.8863,
|
| 2621 |
+
"step": 3440
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.5863358259687288,
|
| 2625 |
+
"grad_norm": 6.586738109588623,
|
| 2626 |
+
"learning_rate": 4.380103711571459e-06,
|
| 2627 |
+
"loss": 2.0465,
|
| 2628 |
+
"step": 3450
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 0.5880353501019715,
|
| 2632 |
+
"grad_norm": 6.210246562957764,
|
| 2633 |
+
"learning_rate": 4.350678018199887e-06,
|
| 2634 |
+
"loss": 1.6285,
|
| 2635 |
+
"step": 3460
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 0.5897348742352141,
|
| 2639 |
+
"grad_norm": 8.666604995727539,
|
| 2640 |
+
"learning_rate": 4.321275182218821e-06,
|
| 2641 |
+
"loss": 2.278,
|
| 2642 |
+
"step": 3470
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 0.5914343983684568,
|
| 2646 |
+
"grad_norm": 5.970696926116943,
|
| 2647 |
+
"learning_rate": 4.291896238664991e-06,
|
| 2648 |
+
"loss": 1.7391,
|
| 2649 |
+
"step": 3480
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 0.5931339225016995,
|
| 2653 |
+
"grad_norm": 6.197153091430664,
|
| 2654 |
+
"learning_rate": 4.262542221734077e-06,
|
| 2655 |
+
"loss": 1.8116,
|
| 2656 |
+
"step": 3490
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 0.5948334466349422,
|
| 2660 |
+
"grad_norm": 5.297378063201904,
|
| 2661 |
+
"learning_rate": 4.233214164744286e-06,
|
| 2662 |
+
"loss": 1.5104,
|
| 2663 |
+
"step": 3500
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.5948334466349422,
|
| 2667 |
+
"eval_cosine_accuracy@1": 0.7,
|
| 2668 |
+
"eval_cosine_accuracy@10": 0.9105,
|
| 2669 |
+
"eval_cosine_accuracy@3": 0.8315,
|
| 2670 |
+
"eval_cosine_accuracy@5": 0.871,
|
| 2671 |
+
"eval_cosine_map@100": 0.7789008444240532,
|
| 2672 |
+
"eval_cosine_mrr@10": 0.775367857142857,
|
| 2673 |
+
"eval_cosine_ndcg@10": 0.8083837978692977,
|
| 2674 |
+
"eval_cosine_precision@1": 0.7,
|
| 2675 |
+
"eval_cosine_precision@10": 0.09105,
|
| 2676 |
+
"eval_cosine_precision@3": 0.2771666666666666,
|
| 2677 |
+
"eval_cosine_precision@5": 0.17420000000000005,
|
| 2678 |
+
"eval_cosine_recall@1": 0.7,
|
| 2679 |
+
"eval_cosine_recall@10": 0.9105,
|
| 2680 |
+
"eval_cosine_recall@3": 0.8315,
|
| 2681 |
+
"eval_cosine_recall@5": 0.871,
|
| 2682 |
+
"eval_loss": 1.5698647499084473,
|
| 2683 |
+
"eval_runtime": 2.7632,
|
| 2684 |
+
"eval_samples_per_second": 272.875,
|
| 2685 |
+
"eval_sequential_score": 0.7789008444240532,
|
| 2686 |
+
"eval_steps_per_second": 2.171,
|
| 2687 |
+
"eval_sts-dev_pearson_cosine": 0.8017937499678631,
|
| 2688 |
+
"eval_sts-dev_pearson_dot": 0.7968971369444564,
|
| 2689 |
+
"eval_sts-dev_pearson_euclidean": 0.7889399312152248,
|
| 2690 |
+
"eval_sts-dev_pearson_manhattan": 0.7898347176162472,
|
| 2691 |
+
"eval_sts-dev_pearson_max": 0.8017937499678631,
|
| 2692 |
+
"eval_sts-dev_spearman_cosine": 0.8044304044501694,
|
| 2693 |
+
"eval_sts-dev_spearman_dot": 0.8013449184419915,
|
| 2694 |
+
"eval_sts-dev_spearman_euclidean": 0.8042950575910329,
|
| 2695 |
+
"eval_sts-dev_spearman_manhattan": 0.8042523771894302,
|
| 2696 |
+
"eval_sts-dev_spearman_max": 0.8044304044501694,
|
| 2697 |
+
"step": 3500
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.5965329707681849,
|
| 2701 |
+
"grad_norm": 5.6484575271606445,
|
| 2702 |
+
"learning_rate": 4.2039131000999865e-06,
|
| 2703 |
+
"loss": 1.8306,
|
| 2704 |
+
"step": 3510
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.5982324949014276,
|
| 2708 |
+
"grad_norm": 7.109194278717041,
|
| 2709 |
+
"learning_rate": 4.174640059255369e-06,
|
| 2710 |
+
"loss": 1.6575,
|
| 2711 |
+
"step": 3520
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.5999320190346703,
|
| 2715 |
+
"grad_norm": 8.67902660369873,
|
| 2716 |
+
"learning_rate": 4.145396072678126e-06,
|
| 2717 |
+
"loss": 1.56,
|
| 2718 |
+
"step": 3530
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.601631543167913,
|
| 2722 |
+
"grad_norm": 5.615638732910156,
|
| 2723 |
+
"learning_rate": 4.116182169813185e-06,
|
| 2724 |
+
"loss": 1.8674,
|
| 2725 |
+
"step": 3540
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.6033310673011557,
|
| 2729 |
+
"grad_norm": 6.5344085693359375,
|
| 2730 |
+
"learning_rate": 4.086999379046472e-06,
|
| 2731 |
+
"loss": 1.4909,
|
| 2732 |
+
"step": 3550
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.6050305914343984,
|
| 2736 |
+
"grad_norm": 7.410928726196289,
|
| 2737 |
+
"learning_rate": 4.057848727668701e-06,
|
| 2738 |
+
"loss": 1.6653,
|
| 2739 |
+
"step": 3560
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.6067301155676411,
|
| 2743 |
+
"grad_norm": 5.162845134735107,
|
| 2744 |
+
"learning_rate": 4.028731241839224e-06,
|
| 2745 |
+
"loss": 2.2836,
|
| 2746 |
+
"step": 3570
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.6084296397008837,
|
| 2750 |
+
"grad_norm": 9.384202003479004,
|
| 2751 |
+
"learning_rate": 3.999647946549898e-06,
|
| 2752 |
+
"loss": 2.1572,
|
| 2753 |
+
"step": 3580
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.6101291638341264,
|
| 2757 |
+
"grad_norm": 9.560714721679688,
|
| 2758 |
+
"learning_rate": 3.970599865589006e-06,
|
| 2759 |
+
"loss": 1.6969,
|
| 2760 |
+
"step": 3590
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.6118286879673691,
|
| 2764 |
+
"grad_norm": 8.857442855834961,
|
| 2765 |
+
"learning_rate": 3.94158802150522e-06,
|
| 2766 |
+
"loss": 2.0679,
|
| 2767 |
+
"step": 3600
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.6135282121006118,
|
| 2771 |
+
"grad_norm": 4.5523200035095215,
|
| 2772 |
+
"learning_rate": 3.912613435571603e-06,
|
| 2773 |
+
"loss": 1.8555,
|
| 2774 |
+
"step": 3610
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.6152277362338545,
|
| 2778 |
+
"grad_norm": 4.791132926940918,
|
| 2779 |
+
"learning_rate": 3.8836771277496575e-06,
|
| 2780 |
+
"loss": 1.8506,
|
| 2781 |
+
"step": 3620
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.6169272603670972,
|
| 2785 |
+
"grad_norm": 5.0394673347473145,
|
| 2786 |
+
"learning_rate": 3.854780116653429e-06,
|
| 2787 |
+
"loss": 1.4441,
|
| 2788 |
+
"step": 3630
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.6186267845003399,
|
| 2792 |
+
"grad_norm": 8.9247465133667,
|
| 2793 |
+
"learning_rate": 3.8259234195136344e-06,
|
| 2794 |
+
"loss": 1.6772,
|
| 2795 |
+
"step": 3640
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.6203263086335826,
|
| 2799 |
+
"grad_norm": 7.963625907897949,
|
| 2800 |
+
"learning_rate": 3.7971080521418636e-06,
|
| 2801 |
+
"loss": 1.6228,
|
| 2802 |
+
"step": 3650
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.6220258327668253,
|
| 2806 |
+
"grad_norm": 5.572691917419434,
|
| 2807 |
+
"learning_rate": 3.7683350288948204e-06,
|
| 2808 |
+
"loss": 1.7809,
|
| 2809 |
+
"step": 3660
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.623725356900068,
|
| 2813 |
+
"grad_norm": 4.029000282287598,
|
| 2814 |
+
"learning_rate": 3.7396053626386115e-06,
|
| 2815 |
+
"loss": 1.5777,
|
| 2816 |
+
"step": 3670
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.6254248810333106,
|
| 2820 |
+
"grad_norm": 5.965753555297852,
|
| 2821 |
+
"learning_rate": 3.7109200647130983e-06,
|
| 2822 |
+
"loss": 2.0558,
|
| 2823 |
+
"step": 3680
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.6271244051665533,
|
| 2827 |
+
"grad_norm": 5.478359222412109,
|
| 2828 |
+
"learning_rate": 3.6822801448962835e-06,
|
| 2829 |
+
"loss": 1.9044,
|
| 2830 |
+
"step": 3690
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.628823929299796,
|
| 2834 |
+
"grad_norm": 6.156702518463135,
|
| 2835 |
+
"learning_rate": 3.6536866113687754e-06,
|
| 2836 |
+
"loss": 1.9202,
|
| 2837 |
+
"step": 3700
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.6305234534330387,
|
| 2841 |
+
"grad_norm": 5.119917392730713,
|
| 2842 |
+
"learning_rate": 3.6251404706783005e-06,
|
| 2843 |
+
"loss": 1.4821,
|
| 2844 |
+
"step": 3710
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.6322229775662814,
|
| 2848 |
+
"grad_norm": 5.908693790435791,
|
| 2849 |
+
"learning_rate": 3.5966427277042593e-06,
|
| 2850 |
+
"loss": 1.5492,
|
| 2851 |
+
"step": 3720
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.6339225016995241,
|
| 2855 |
+
"grad_norm": 6.6029181480407715,
|
| 2856 |
+
"learning_rate": 3.568194385622359e-06,
|
| 2857 |
+
"loss": 1.7337,
|
| 2858 |
+
"step": 3730
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.6356220258327668,
|
| 2862 |
+
"grad_norm": 6.767298221588135,
|
| 2863 |
+
"learning_rate": 3.5397964458693045e-06,
|
| 2864 |
+
"loss": 2.0108,
|
| 2865 |
+
"step": 3740
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.6373215499660095,
|
| 2869 |
+
"grad_norm": 7.079708576202393,
|
| 2870 |
+
"learning_rate": 3.5114499081075386e-06,
|
| 2871 |
+
"loss": 1.7365,
|
| 2872 |
+
"step": 3750
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.6390210740992522,
|
| 2876 |
+
"grad_norm": 5.657531261444092,
|
| 2877 |
+
"learning_rate": 3.4831557701900507e-06,
|
| 2878 |
+
"loss": 1.5291,
|
| 2879 |
+
"step": 3760
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.6407205982324949,
|
| 2883 |
+
"grad_norm": 9.403407096862793,
|
| 2884 |
+
"learning_rate": 3.4549150281252635e-06,
|
| 2885 |
+
"loss": 1.8147,
|
| 2886 |
+
"step": 3770
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.6424201223657376,
|
| 2890 |
+
"grad_norm": 4.905513763427734,
|
| 2891 |
+
"learning_rate": 3.4267286760419527e-06,
|
| 2892 |
+
"loss": 1.6704,
|
| 2893 |
+
"step": 3780
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.6441196464989802,
|
| 2897 |
+
"grad_norm": 5.188987731933594,
|
| 2898 |
+
"learning_rate": 3.398597706154268e-06,
|
| 2899 |
+
"loss": 1.727,
|
| 2900 |
+
"step": 3790
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.6458191706322229,
|
| 2904 |
+
"grad_norm": 9.689862251281738,
|
| 2905 |
+
"learning_rate": 3.3705231087267955e-06,
|
| 2906 |
+
"loss": 1.741,
|
| 2907 |
+
"step": 3800
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.6475186947654656,
|
| 2911 |
+
"grad_norm": 5.145129203796387,
|
| 2912 |
+
"learning_rate": 3.3425058720396986e-06,
|
| 2913 |
+
"loss": 1.6929,
|
| 2914 |
+
"step": 3810
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.6492182188987083,
|
| 2918 |
+
"grad_norm": 5.635613918304443,
|
| 2919 |
+
"learning_rate": 3.314546982353941e-06,
|
| 2920 |
+
"loss": 1.7855,
|
| 2921 |
+
"step": 3820
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.650917743031951,
|
| 2925 |
+
"grad_norm": 5.333996295928955,
|
| 2926 |
+
"learning_rate": 3.2866474238765522e-06,
|
| 2927 |
+
"loss": 1.4379,
|
| 2928 |
+
"step": 3830
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.6526172671651937,
|
| 2932 |
+
"grad_norm": 7.929337501525879,
|
| 2933 |
+
"learning_rate": 3.258808178725984e-06,
|
| 2934 |
+
"loss": 1.8255,
|
| 2935 |
+
"step": 3840
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.6543167912984365,
|
| 2939 |
+
"grad_norm": 7.606388092041016,
|
| 2940 |
+
"learning_rate": 3.2310302268975545e-06,
|
| 2941 |
+
"loss": 1.8072,
|
| 2942 |
+
"step": 3850
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.6560163154316792,
|
| 2946 |
+
"grad_norm": 5.5722126960754395,
|
| 2947 |
+
"learning_rate": 3.2033145462289284e-06,
|
| 2948 |
+
"loss": 1.4806,
|
| 2949 |
+
"step": 3860
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.6577158395649219,
|
| 2953 |
+
"grad_norm": 4.687691688537598,
|
| 2954 |
+
"learning_rate": 3.175662112365707e-06,
|
| 2955 |
+
"loss": 1.7177,
|
| 2956 |
+
"step": 3870
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.6594153636981646,
|
| 2960 |
+
"grad_norm": 5.950764179229736,
|
| 2961 |
+
"learning_rate": 3.1480738987270854e-06,
|
| 2962 |
+
"loss": 1.3493,
|
| 2963 |
+
"step": 3880
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.6611148878314073,
|
| 2967 |
+
"grad_norm": 6.248993396759033,
|
| 2968 |
+
"learning_rate": 3.120550876471576e-06,
|
| 2969 |
+
"loss": 1.7455,
|
| 2970 |
+
"step": 3890
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.6628144119646499,
|
| 2974 |
+
"grad_norm": 6.485055446624756,
|
| 2975 |
+
"learning_rate": 3.0930940144628385e-06,
|
| 2976 |
+
"loss": 1.5822,
|
| 2977 |
+
"step": 3900
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.6645139360978926,
|
| 2981 |
+
"grad_norm": 6.7886881828308105,
|
| 2982 |
+
"learning_rate": 3.0657042792355528e-06,
|
| 2983 |
+
"loss": 2.4897,
|
| 2984 |
+
"step": 3910
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.6662134602311353,
|
| 2988 |
+
"grad_norm": 4.51090669631958,
|
| 2989 |
+
"learning_rate": 3.0383826349614115e-06,
|
| 2990 |
+
"loss": 1.8648,
|
| 2991 |
+
"step": 3920
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.667912984364378,
|
| 2995 |
+
"grad_norm": 7.129496097564697,
|
| 2996 |
+
"learning_rate": 3.0111300434151746e-06,
|
| 2997 |
+
"loss": 1.8095,
|
| 2998 |
+
"step": 3930
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.6696125084976207,
|
| 3002 |
+
"grad_norm": 8.079377174377441,
|
| 3003 |
+
"learning_rate": 2.9839474639408108e-06,
|
| 3004 |
+
"loss": 2.0793,
|
| 3005 |
+
"step": 3940
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.6713120326308634,
|
| 3009 |
+
"grad_norm": 6.9486284255981445,
|
| 3010 |
+
"learning_rate": 2.956835853417726e-06,
|
| 3011 |
+
"loss": 1.9273,
|
| 3012 |
+
"step": 3950
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.6730115567641061,
|
| 3016 |
+
"grad_norm": 6.231419563293457,
|
| 3017 |
+
"learning_rate": 2.9297961662270893e-06,
|
| 3018 |
+
"loss": 1.6705,
|
| 3019 |
+
"step": 3960
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.6747110808973488,
|
| 3023 |
+
"grad_norm": 9.692667007446289,
|
| 3024 |
+
"learning_rate": 2.902829354218219e-06,
|
| 3025 |
+
"loss": 1.8497,
|
| 3026 |
+
"step": 3970
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.6764106050305915,
|
| 3030 |
+
"grad_norm": 6.941633701324463,
|
| 3031 |
+
"learning_rate": 2.875936366675096e-06,
|
| 3032 |
+
"loss": 2.1609,
|
| 3033 |
+
"step": 3980
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.6781101291638342,
|
| 3037 |
+
"grad_norm": 4.512592315673828,
|
| 3038 |
+
"learning_rate": 2.8491181502829335e-06,
|
| 3039 |
+
"loss": 1.719,
|
| 3040 |
+
"step": 3990
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.6798096532970768,
|
| 3044 |
+
"grad_norm": 13.518653869628906,
|
| 3045 |
+
"learning_rate": 2.822375649094846e-06,
|
| 3046 |
+
"loss": 1.7953,
|
| 3047 |
+
"step": 4000
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.6798096532970768,
|
| 3051 |
+
"eval_cosine_accuracy@1": 0.7145,
|
| 3052 |
+
"eval_cosine_accuracy@10": 0.914,
|
| 3053 |
+
"eval_cosine_accuracy@3": 0.839,
|
| 3054 |
+
"eval_cosine_accuracy@5": 0.8745,
|
| 3055 |
+
"eval_cosine_map@100": 0.7881479093038395,
|
| 3056 |
+
"eval_cosine_mrr@10": 0.7848144841269835,
|
| 3057 |
+
"eval_cosine_ndcg@10": 0.8163072349117797,
|
| 3058 |
+
"eval_cosine_precision@1": 0.7145,
|
| 3059 |
+
"eval_cosine_precision@10": 0.09140000000000001,
|
| 3060 |
+
"eval_cosine_precision@3": 0.2796666666666666,
|
| 3061 |
+
"eval_cosine_precision@5": 0.1749,
|
| 3062 |
+
"eval_cosine_recall@1": 0.7145,
|
| 3063 |
+
"eval_cosine_recall@10": 0.914,
|
| 3064 |
+
"eval_cosine_recall@3": 0.839,
|
| 3065 |
+
"eval_cosine_recall@5": 0.8745,
|
| 3066 |
+
"eval_loss": 1.5573265552520752,
|
| 3067 |
+
"eval_runtime": 2.7757,
|
| 3068 |
+
"eval_samples_per_second": 271.645,
|
| 3069 |
+
"eval_sequential_score": 0.7881479093038395,
|
| 3070 |
+
"eval_steps_per_second": 2.162,
|
| 3071 |
+
"eval_sts-dev_pearson_cosine": 0.8053811920422538,
|
| 3072 |
+
"eval_sts-dev_pearson_dot": 0.8004660633078914,
|
| 3073 |
+
"eval_sts-dev_pearson_euclidean": 0.7928635065778744,
|
| 3074 |
+
"eval_sts-dev_pearson_manhattan": 0.7941416034985971,
|
| 3075 |
+
"eval_sts-dev_pearson_max": 0.8053811920422538,
|
| 3076 |
+
"eval_sts-dev_spearman_cosine": 0.8086109320148909,
|
| 3077 |
+
"eval_sts-dev_spearman_dot": 0.804234875942229,
|
| 3078 |
+
"eval_sts-dev_spearman_euclidean": 0.8084185226972233,
|
| 3079 |
+
"eval_sts-dev_spearman_manhattan": 0.8088695507249705,
|
| 3080 |
+
"eval_sts-dev_spearman_max": 0.8088695507249705,
|
| 3081 |
+
"step": 4000
|
| 3082 |
+
}
|
| 3083 |
+
],
|
| 3084 |
+
"logging_steps": 10,
|
| 3085 |
+
"max_steps": 5884,
|
| 3086 |
+
"num_input_tokens_seen": 0,
|
| 3087 |
+
"num_train_epochs": 1,
|
| 3088 |
+
"save_steps": 500,
|
| 3089 |
+
"stateful_callbacks": {
|
| 3090 |
+
"TrainerControl": {
|
| 3091 |
+
"args": {
|
| 3092 |
+
"should_epoch_stop": false,
|
| 3093 |
+
"should_evaluate": false,
|
| 3094 |
+
"should_log": false,
|
| 3095 |
+
"should_save": true,
|
| 3096 |
+
"should_training_stop": false
|
| 3097 |
+
},
|
| 3098 |
+
"attributes": {}
|
| 3099 |
+
}
|
| 3100 |
+
},
|
| 3101 |
+
"total_flos": 0.0,
|
| 3102 |
+
"train_batch_size": 128,
|
| 3103 |
+
"trial_name": null,
|
| 3104 |
+
"trial_params": null
|
| 3105 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ff73f331abc247df1bb895d7a7aef82db9f32fbce58b78521eec8094f43a2457
|
| 3 |
+
size 5496
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|