Training in progress, step 256, checkpoint
Browse files- checkpoint-256/1_Pooling/config.json +10 -0
- checkpoint-256/README.md +0 -0
- checkpoint-256/added_tokens.json +3 -0
- checkpoint-256/config.json +35 -0
- checkpoint-256/config_sentence_transformers.json +10 -0
- checkpoint-256/modules.json +14 -0
- checkpoint-256/optimizer.pt +3 -0
- checkpoint-256/pytorch_model.bin +3 -0
- checkpoint-256/rng_state.pth +3 -0
- checkpoint-256/scheduler.pt +3 -0
- checkpoint-256/sentence_bert_config.json +4 -0
- checkpoint-256/special_tokens_map.json +51 -0
- checkpoint-256/spm.model +3 -0
- checkpoint-256/tokenizer.json +0 -0
- checkpoint-256/tokenizer_config.json +65 -0
- checkpoint-256/trainer_state.json +2961 -0
- checkpoint-256/training_args.bin +3 -0
checkpoint-256/1_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 |
+
}
|
checkpoint-256/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-256/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
checkpoint-256/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bobox/DeBERTa-small-ST-v1-test",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DebertaV2Model"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 768,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 3072,
|
| 12 |
+
"layer_norm_eps": 1e-07,
|
| 13 |
+
"max_position_embeddings": 512,
|
| 14 |
+
"max_relative_positions": -1,
|
| 15 |
+
"model_type": "deberta-v2",
|
| 16 |
+
"norm_rel_ebd": "layer_norm",
|
| 17 |
+
"num_attention_heads": 12,
|
| 18 |
+
"num_hidden_layers": 6,
|
| 19 |
+
"pad_token_id": 0,
|
| 20 |
+
"pooler_dropout": 0,
|
| 21 |
+
"pooler_hidden_act": "gelu",
|
| 22 |
+
"pooler_hidden_size": 768,
|
| 23 |
+
"pos_att_type": [
|
| 24 |
+
"p2c",
|
| 25 |
+
"c2p"
|
| 26 |
+
],
|
| 27 |
+
"position_biased_input": false,
|
| 28 |
+
"position_buckets": 256,
|
| 29 |
+
"relative_attention": true,
|
| 30 |
+
"share_att_key": true,
|
| 31 |
+
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.42.3",
|
| 33 |
+
"type_vocab_size": 0,
|
| 34 |
+
"vocab_size": 128100
|
| 35 |
+
}
|
checkpoint-256/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.42.3",
|
| 5 |
+
"pytorch": "2.1.2"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-256/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 |
+
]
|
checkpoint-256/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:834fe5b897193f42cefd16b02016d4e8de603f8524a68f05724c2708f8f2c3ae
|
| 3 |
+
size 1130520122
|
checkpoint-256/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7cf20d250eba8880b4d8c48f13e362fc9ff74b0d1f055b2882da485e7ee31624
|
| 3 |
+
size 565251810
|
checkpoint-256/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0df7f5c66d7ba57e794f57af2e385e920041a340531203e1232197d1c5253a5b
|
| 3 |
+
size 14244
|
checkpoint-256/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85f2d383761c230b38964c22e9f0c34fab7bbae40088d6160f5cabcfe8b60caf
|
| 3 |
+
size 1064
|
checkpoint-256/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-256/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "[CLS]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "[SEP]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "[MASK]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "[PAD]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "[SEP]",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "[UNK]",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
checkpoint-256/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
checkpoint-256/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-256/tokenizer_config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "[CLS]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[SEP]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[UNK]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"128000": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "[CLS]",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "[CLS]",
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"eos_token": "[SEP]",
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"max_length": 512,
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"pad_to_multiple_of": null,
|
| 53 |
+
"pad_token": "[PAD]",
|
| 54 |
+
"pad_token_type_id": 0,
|
| 55 |
+
"padding_side": "right",
|
| 56 |
+
"sep_token": "[SEP]",
|
| 57 |
+
"sp_model_kwargs": {},
|
| 58 |
+
"split_by_punct": false,
|
| 59 |
+
"stride": 0,
|
| 60 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 61 |
+
"truncation_side": "right",
|
| 62 |
+
"truncation_strategy": "longest_first",
|
| 63 |
+
"unk_token": "[UNK]",
|
| 64 |
+
"vocab_type": "spm"
|
| 65 |
+
}
|
checkpoint-256/trainer_state.json
ADDED
|
@@ -0,0 +1,2961 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.8,
|
| 5 |
+
"eval_steps": 32,
|
| 6 |
+
"global_step": 256,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.003125,
|
| 13 |
+
"grad_norm": 3.1757984161376953,
|
| 14 |
+
"learning_rate": 3.125e-07,
|
| 15 |
+
"loss": 0.7374,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.00625,
|
| 20 |
+
"grad_norm": 3.137390375137329,
|
| 21 |
+
"learning_rate": 6.25e-07,
|
| 22 |
+
"loss": 0.5723,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.009375,
|
| 27 |
+
"grad_norm": 2.765856981277466,
|
| 28 |
+
"learning_rate": 9.375000000000001e-07,
|
| 29 |
+
"loss": 0.551,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0125,
|
| 34 |
+
"grad_norm": 3.468062162399292,
|
| 35 |
+
"learning_rate": 1.25e-06,
|
| 36 |
+
"loss": 0.7379,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.015625,
|
| 41 |
+
"grad_norm": 2.6695668697357178,
|
| 42 |
+
"learning_rate": 1.5625e-06,
|
| 43 |
+
"loss": 0.5271,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01875,
|
| 48 |
+
"grad_norm": 2.7720863819122314,
|
| 49 |
+
"learning_rate": 1.8750000000000003e-06,
|
| 50 |
+
"loss": 0.5858,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.021875,
|
| 55 |
+
"grad_norm": 3.0211267471313477,
|
| 56 |
+
"learning_rate": 2.1875000000000002e-06,
|
| 57 |
+
"loss": 0.6562,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.025,
|
| 62 |
+
"grad_norm": 3.641108989715576,
|
| 63 |
+
"learning_rate": 2.5e-06,
|
| 64 |
+
"loss": 0.8228,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.028125,
|
| 69 |
+
"grad_norm": 3.9061200618743896,
|
| 70 |
+
"learning_rate": 2.8125e-06,
|
| 71 |
+
"loss": 0.9988,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.03125,
|
| 76 |
+
"grad_norm": 2.642423391342163,
|
| 77 |
+
"learning_rate": 3.125e-06,
|
| 78 |
+
"loss": 0.5582,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.034375,
|
| 83 |
+
"grad_norm": 3.6546943187713623,
|
| 84 |
+
"learning_rate": 3.4375e-06,
|
| 85 |
+
"loss": 0.8546,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0375,
|
| 90 |
+
"grad_norm": 2.5504300594329834,
|
| 91 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 92 |
+
"loss": 0.4235,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.040625,
|
| 97 |
+
"grad_norm": 2.845123529434204,
|
| 98 |
+
"learning_rate": 4.0625000000000005e-06,
|
| 99 |
+
"loss": 0.6418,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04375,
|
| 104 |
+
"grad_norm": 2.8562164306640625,
|
| 105 |
+
"learning_rate": 4.3750000000000005e-06,
|
| 106 |
+
"loss": 0.6577,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.046875,
|
| 111 |
+
"grad_norm": 3.4033620357513428,
|
| 112 |
+
"learning_rate": 4.6875000000000004e-06,
|
| 113 |
+
"loss": 0.8333,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.05,
|
| 118 |
+
"grad_norm": 2.148242473602295,
|
| 119 |
+
"learning_rate": 5e-06,
|
| 120 |
+
"loss": 0.4082,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.053125,
|
| 125 |
+
"grad_norm": 3.685960292816162,
|
| 126 |
+
"learning_rate": 5.3125e-06,
|
| 127 |
+
"loss": 0.8101,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.05625,
|
| 132 |
+
"grad_norm": 2.7071452140808105,
|
| 133 |
+
"learning_rate": 5.625e-06,
|
| 134 |
+
"loss": 0.5259,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.059375,
|
| 139 |
+
"grad_norm": 3.508561611175537,
|
| 140 |
+
"learning_rate": 5.9375e-06,
|
| 141 |
+
"loss": 0.9015,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.0625,
|
| 146 |
+
"grad_norm": 4.140976428985596,
|
| 147 |
+
"learning_rate": 6.25e-06,
|
| 148 |
+
"loss": 1.3915,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.065625,
|
| 153 |
+
"grad_norm": 1.5563820600509644,
|
| 154 |
+
"learning_rate": 6.5625e-06,
|
| 155 |
+
"loss": 0.26,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.06875,
|
| 160 |
+
"grad_norm": 3.1467344760894775,
|
| 161 |
+
"learning_rate": 6.875e-06,
|
| 162 |
+
"loss": 0.6885,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.071875,
|
| 167 |
+
"grad_norm": 3.539327383041382,
|
| 168 |
+
"learning_rate": 7.1875e-06,
|
| 169 |
+
"loss": 0.9357,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.075,
|
| 174 |
+
"grad_norm": 3.1691510677337646,
|
| 175 |
+
"learning_rate": 7.500000000000001e-06,
|
| 176 |
+
"loss": 0.7168,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.078125,
|
| 181 |
+
"grad_norm": 3.9020121097564697,
|
| 182 |
+
"learning_rate": 7.8125e-06,
|
| 183 |
+
"loss": 0.8678,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.08125,
|
| 188 |
+
"grad_norm": 2.3635435104370117,
|
| 189 |
+
"learning_rate": 8.125000000000001e-06,
|
| 190 |
+
"loss": 0.4922,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.084375,
|
| 195 |
+
"grad_norm": 2.5170037746429443,
|
| 196 |
+
"learning_rate": 8.4375e-06,
|
| 197 |
+
"loss": 0.4937,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.0875,
|
| 202 |
+
"grad_norm": 2.7988407611846924,
|
| 203 |
+
"learning_rate": 8.750000000000001e-06,
|
| 204 |
+
"loss": 0.5891,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.090625,
|
| 209 |
+
"grad_norm": 2.99135160446167,
|
| 210 |
+
"learning_rate": 9.0625e-06,
|
| 211 |
+
"loss": 0.6921,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.09375,
|
| 216 |
+
"grad_norm": 3.098013162612915,
|
| 217 |
+
"learning_rate": 9.375000000000001e-06,
|
| 218 |
+
"loss": 0.8087,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.096875,
|
| 223 |
+
"grad_norm": 3.358091115951538,
|
| 224 |
+
"learning_rate": 9.6875e-06,
|
| 225 |
+
"loss": 0.805,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.1,
|
| 230 |
+
"grad_norm": 3.0206046104431152,
|
| 231 |
+
"learning_rate": 1e-05,
|
| 232 |
+
"loss": 0.6141,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1,
|
| 237 |
+
"eval_VitaminC_cosine_accuracy": 0.5546875,
|
| 238 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.8487042188644409,
|
| 239 |
+
"eval_VitaminC_cosine_ap": 0.5467207830251657,
|
| 240 |
+
"eval_VitaminC_cosine_f1": 0.6657824933687002,
|
| 241 |
+
"eval_VitaminC_cosine_f1_threshold": 0.2510407269001007,
|
| 242 |
+
"eval_VitaminC_cosine_precision": 0.4990059642147117,
|
| 243 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 244 |
+
"eval_VitaminC_dot_accuracy": 0.55078125,
|
| 245 |
+
"eval_VitaminC_dot_accuracy_threshold": 318.7947082519531,
|
| 246 |
+
"eval_VitaminC_dot_ap": 0.5360598625078122,
|
| 247 |
+
"eval_VitaminC_dot_f1": 0.6657824933687002,
|
| 248 |
+
"eval_VitaminC_dot_f1_threshold": 98.82717895507812,
|
| 249 |
+
"eval_VitaminC_dot_precision": 0.4990059642147117,
|
| 250 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 251 |
+
"eval_VitaminC_euclidean_accuracy": 0.552734375,
|
| 252 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 15.370981216430664,
|
| 253 |
+
"eval_VitaminC_euclidean_ap": 0.54465834495355,
|
| 254 |
+
"eval_VitaminC_euclidean_f1": 0.6657824933687002,
|
| 255 |
+
"eval_VitaminC_euclidean_f1_threshold": 24.364877700805664,
|
| 256 |
+
"eval_VitaminC_euclidean_precision": 0.4990059642147117,
|
| 257 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 258 |
+
"eval_VitaminC_manhattan_accuracy": 0.5546875,
|
| 259 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 273.6689758300781,
|
| 260 |
+
"eval_VitaminC_manhattan_ap": 0.5450408710915566,
|
| 261 |
+
"eval_VitaminC_manhattan_f1": 0.6675531914893617,
|
| 262 |
+
"eval_VitaminC_manhattan_f1_threshold": 502.82244873046875,
|
| 263 |
+
"eval_VitaminC_manhattan_precision": 0.500998003992016,
|
| 264 |
+
"eval_VitaminC_manhattan_recall": 1.0,
|
| 265 |
+
"eval_VitaminC_max_accuracy": 0.5546875,
|
| 266 |
+
"eval_VitaminC_max_accuracy_threshold": 318.7947082519531,
|
| 267 |
+
"eval_VitaminC_max_ap": 0.5467207830251657,
|
| 268 |
+
"eval_VitaminC_max_f1": 0.6675531914893617,
|
| 269 |
+
"eval_VitaminC_max_f1_threshold": 502.82244873046875,
|
| 270 |
+
"eval_VitaminC_max_precision": 0.500998003992016,
|
| 271 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 272 |
+
"eval_sequential_score": 0.5467207830251657,
|
| 273 |
+
"eval_sts-test_pearson_cosine": 0.8677868917853514,
|
| 274 |
+
"eval_sts-test_pearson_dot": 0.8601917125112223,
|
| 275 |
+
"eval_sts-test_pearson_euclidean": 0.889472619726378,
|
| 276 |
+
"eval_sts-test_pearson_manhattan": 0.890143281884324,
|
| 277 |
+
"eval_sts-test_pearson_max": 0.890143281884324,
|
| 278 |
+
"eval_sts-test_spearman_cosine": 0.8954519734959775,
|
| 279 |
+
"eval_sts-test_spearman_dot": 0.8621348855070287,
|
| 280 |
+
"eval_sts-test_spearman_euclidean": 0.8880001748147683,
|
| 281 |
+
"eval_sts-test_spearman_manhattan": 0.8870461226731652,
|
| 282 |
+
"eval_sts-test_spearman_max": 0.8954519734959775,
|
| 283 |
+
"eval_vitaminc-pairs_loss": 2.332582473754883,
|
| 284 |
+
"eval_vitaminc-pairs_runtime": 2.2432,
|
| 285 |
+
"eval_vitaminc-pairs_samples_per_second": 48.146,
|
| 286 |
+
"eval_vitaminc-pairs_steps_per_second": 0.892,
|
| 287 |
+
"step": 32
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.1,
|
| 291 |
+
"eval_negation-triplets_loss": 0.8681236505508423,
|
| 292 |
+
"eval_negation-triplets_runtime": 0.2927,
|
| 293 |
+
"eval_negation-triplets_samples_per_second": 218.641,
|
| 294 |
+
"eval_negation-triplets_steps_per_second": 3.416,
|
| 295 |
+
"step": 32
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.1,
|
| 299 |
+
"eval_scitail-pairs-pos_loss": 0.07759770005941391,
|
| 300 |
+
"eval_scitail-pairs-pos_runtime": 0.3708,
|
| 301 |
+
"eval_scitail-pairs-pos_samples_per_second": 145.613,
|
| 302 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.697,
|
| 303 |
+
"step": 32
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.1,
|
| 307 |
+
"eval_xsum-pairs_loss": 0.09131219983100891,
|
| 308 |
+
"eval_xsum-pairs_runtime": 2.8486,
|
| 309 |
+
"eval_xsum-pairs_samples_per_second": 44.934,
|
| 310 |
+
"eval_xsum-pairs_steps_per_second": 0.702,
|
| 311 |
+
"step": 32
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.1,
|
| 315 |
+
"eval_sciq_pairs_loss": 0.01965576782822609,
|
| 316 |
+
"eval_sciq_pairs_runtime": 3.6062,
|
| 317 |
+
"eval_sciq_pairs_samples_per_second": 35.494,
|
| 318 |
+
"eval_sciq_pairs_steps_per_second": 0.555,
|
| 319 |
+
"step": 32
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 0.1,
|
| 323 |
+
"eval_qasc_pairs_loss": 0.10996829718351364,
|
| 324 |
+
"eval_qasc_pairs_runtime": 0.5975,
|
| 325 |
+
"eval_qasc_pairs_samples_per_second": 214.235,
|
| 326 |
+
"eval_qasc_pairs_steps_per_second": 3.347,
|
| 327 |
+
"step": 32
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.1,
|
| 331 |
+
"eval_openbookqa_pairs_loss": 0.6932356953620911,
|
| 332 |
+
"eval_openbookqa_pairs_runtime": 0.5729,
|
| 333 |
+
"eval_openbookqa_pairs_samples_per_second": 223.415,
|
| 334 |
+
"eval_openbookqa_pairs_steps_per_second": 3.491,
|
| 335 |
+
"step": 32
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.1,
|
| 339 |
+
"eval_msmarco_pairs_loss": 0.32686129212379456,
|
| 340 |
+
"eval_msmarco_pairs_runtime": 1.4637,
|
| 341 |
+
"eval_msmarco_pairs_samples_per_second": 87.448,
|
| 342 |
+
"eval_msmarco_pairs_steps_per_second": 1.366,
|
| 343 |
+
"step": 32
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 0.1,
|
| 347 |
+
"eval_nq_pairs_loss": 0.1978442668914795,
|
| 348 |
+
"eval_nq_pairs_runtime": 2.8588,
|
| 349 |
+
"eval_nq_pairs_samples_per_second": 44.774,
|
| 350 |
+
"eval_nq_pairs_steps_per_second": 0.7,
|
| 351 |
+
"step": 32
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.1,
|
| 355 |
+
"eval_trivia_pairs_loss": 0.7432661652565002,
|
| 356 |
+
"eval_trivia_pairs_runtime": 4.3895,
|
| 357 |
+
"eval_trivia_pairs_samples_per_second": 29.16,
|
| 358 |
+
"eval_trivia_pairs_steps_per_second": 0.456,
|
| 359 |
+
"step": 32
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.1,
|
| 363 |
+
"eval_gooaq_pairs_loss": 0.3761173486709595,
|
| 364 |
+
"eval_gooaq_pairs_runtime": 1.0043,
|
| 365 |
+
"eval_gooaq_pairs_samples_per_second": 127.452,
|
| 366 |
+
"eval_gooaq_pairs_steps_per_second": 1.991,
|
| 367 |
+
"step": 32
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.1,
|
| 371 |
+
"eval_paws-pos_loss": 0.02476382441818714,
|
| 372 |
+
"eval_paws-pos_runtime": 0.6858,
|
| 373 |
+
"eval_paws-pos_samples_per_second": 186.635,
|
| 374 |
+
"eval_paws-pos_steps_per_second": 2.916,
|
| 375 |
+
"step": 32
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.103125,
|
| 379 |
+
"grad_norm": 3.51029109954834,
|
| 380 |
+
"learning_rate": 1.0312500000000002e-05,
|
| 381 |
+
"loss": 0.7783,
|
| 382 |
+
"step": 33
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.10625,
|
| 386 |
+
"grad_norm": 3.376455783843994,
|
| 387 |
+
"learning_rate": 1.0625e-05,
|
| 388 |
+
"loss": 0.8746,
|
| 389 |
+
"step": 34
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.109375,
|
| 393 |
+
"grad_norm": 2.7385308742523193,
|
| 394 |
+
"learning_rate": 1.0937500000000002e-05,
|
| 395 |
+
"loss": 0.5085,
|
| 396 |
+
"step": 35
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 0.1125,
|
| 400 |
+
"grad_norm": 2.782606840133667,
|
| 401 |
+
"learning_rate": 1.125e-05,
|
| 402 |
+
"loss": 0.4842,
|
| 403 |
+
"step": 36
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.115625,
|
| 407 |
+
"grad_norm": 3.4377782344818115,
|
| 408 |
+
"learning_rate": 1.1562500000000002e-05,
|
| 409 |
+
"loss": 0.8097,
|
| 410 |
+
"step": 37
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.11875,
|
| 414 |
+
"grad_norm": 2.6202378273010254,
|
| 415 |
+
"learning_rate": 1.1875e-05,
|
| 416 |
+
"loss": 0.5325,
|
| 417 |
+
"step": 38
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.121875,
|
| 421 |
+
"grad_norm": 3.0869128704071045,
|
| 422 |
+
"learning_rate": 1.2187500000000001e-05,
|
| 423 |
+
"loss": 0.7221,
|
| 424 |
+
"step": 39
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.125,
|
| 428 |
+
"grad_norm": 3.131516456604004,
|
| 429 |
+
"learning_rate": 1.25e-05,
|
| 430 |
+
"loss": 0.708,
|
| 431 |
+
"step": 40
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.128125,
|
| 435 |
+
"grad_norm": 2.0318033695220947,
|
| 436 |
+
"learning_rate": 1.2812500000000001e-05,
|
| 437 |
+
"loss": 0.2789,
|
| 438 |
+
"step": 41
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 0.13125,
|
| 442 |
+
"grad_norm": 3.2574217319488525,
|
| 443 |
+
"learning_rate": 1.3125e-05,
|
| 444 |
+
"loss": 0.7986,
|
| 445 |
+
"step": 42
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 0.134375,
|
| 449 |
+
"grad_norm": 3.6287729740142822,
|
| 450 |
+
"learning_rate": 1.3437500000000001e-05,
|
| 451 |
+
"loss": 0.9653,
|
| 452 |
+
"step": 43
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.1375,
|
| 456 |
+
"grad_norm": 3.1281752586364746,
|
| 457 |
+
"learning_rate": 1.375e-05,
|
| 458 |
+
"loss": 0.7857,
|
| 459 |
+
"step": 44
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.140625,
|
| 463 |
+
"grad_norm": 2.201566219329834,
|
| 464 |
+
"learning_rate": 1.4062500000000001e-05,
|
| 465 |
+
"loss": 0.2726,
|
| 466 |
+
"step": 45
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.14375,
|
| 470 |
+
"grad_norm": 1.8727688789367676,
|
| 471 |
+
"learning_rate": 1.4375e-05,
|
| 472 |
+
"loss": 0.2458,
|
| 473 |
+
"step": 46
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.146875,
|
| 477 |
+
"grad_norm": 3.156454086303711,
|
| 478 |
+
"learning_rate": 1.4687500000000001e-05,
|
| 479 |
+
"loss": 0.6988,
|
| 480 |
+
"step": 47
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.15,
|
| 484 |
+
"grad_norm": 3.0224971771240234,
|
| 485 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 486 |
+
"loss": 0.6328,
|
| 487 |
+
"step": 48
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 0.153125,
|
| 491 |
+
"grad_norm": 3.4717319011688232,
|
| 492 |
+
"learning_rate": 1.5312500000000003e-05,
|
| 493 |
+
"loss": 0.795,
|
| 494 |
+
"step": 49
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.15625,
|
| 498 |
+
"grad_norm": 2.8961374759674072,
|
| 499 |
+
"learning_rate": 1.5625e-05,
|
| 500 |
+
"loss": 0.6163,
|
| 501 |
+
"step": 50
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.159375,
|
| 505 |
+
"grad_norm": 3.667778491973877,
|
| 506 |
+
"learning_rate": 1.59375e-05,
|
| 507 |
+
"loss": 0.8269,
|
| 508 |
+
"step": 51
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.1625,
|
| 512 |
+
"grad_norm": 2.350587844848633,
|
| 513 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 514 |
+
"loss": 0.52,
|
| 515 |
+
"step": 52
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.165625,
|
| 519 |
+
"grad_norm": 3.312248468399048,
|
| 520 |
+
"learning_rate": 1.6562500000000003e-05,
|
| 521 |
+
"loss": 0.7523,
|
| 522 |
+
"step": 53
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.16875,
|
| 526 |
+
"grad_norm": 2.8101534843444824,
|
| 527 |
+
"learning_rate": 1.6875e-05,
|
| 528 |
+
"loss": 0.6979,
|
| 529 |
+
"step": 54
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.171875,
|
| 533 |
+
"grad_norm": 3.144334077835083,
|
| 534 |
+
"learning_rate": 1.71875e-05,
|
| 535 |
+
"loss": 0.7845,
|
| 536 |
+
"step": 55
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.175,
|
| 540 |
+
"grad_norm": 3.671412229537964,
|
| 541 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 542 |
+
"loss": 0.9325,
|
| 543 |
+
"step": 56
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.178125,
|
| 547 |
+
"grad_norm": 3.204644203186035,
|
| 548 |
+
"learning_rate": 1.7812500000000003e-05,
|
| 549 |
+
"loss": 0.8546,
|
| 550 |
+
"step": 57
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.18125,
|
| 554 |
+
"grad_norm": 2.9951093196868896,
|
| 555 |
+
"learning_rate": 1.8125e-05,
|
| 556 |
+
"loss": 0.6392,
|
| 557 |
+
"step": 58
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.184375,
|
| 561 |
+
"grad_norm": 3.036386013031006,
|
| 562 |
+
"learning_rate": 1.84375e-05,
|
| 563 |
+
"loss": 0.5827,
|
| 564 |
+
"step": 59
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.1875,
|
| 568 |
+
"grad_norm": 3.0899698734283447,
|
| 569 |
+
"learning_rate": 1.8750000000000002e-05,
|
| 570 |
+
"loss": 0.5961,
|
| 571 |
+
"step": 60
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.190625,
|
| 575 |
+
"grad_norm": 2.3574728965759277,
|
| 576 |
+
"learning_rate": 1.9062500000000003e-05,
|
| 577 |
+
"loss": 0.3625,
|
| 578 |
+
"step": 61
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.19375,
|
| 582 |
+
"grad_norm": 2.4232304096221924,
|
| 583 |
+
"learning_rate": 1.9375e-05,
|
| 584 |
+
"loss": 0.2584,
|
| 585 |
+
"step": 62
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.196875,
|
| 589 |
+
"grad_norm": 1.9016233682632446,
|
| 590 |
+
"learning_rate": 1.96875e-05,
|
| 591 |
+
"loss": 0.4047,
|
| 592 |
+
"step": 63
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.2,
|
| 596 |
+
"grad_norm": 3.193114995956421,
|
| 597 |
+
"learning_rate": 2e-05,
|
| 598 |
+
"loss": 0.9429,
|
| 599 |
+
"step": 64
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.2,
|
| 603 |
+
"eval_VitaminC_cosine_accuracy": 0.560546875,
|
| 604 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.8192525506019592,
|
| 605 |
+
"eval_VitaminC_cosine_ap": 0.5485465805560719,
|
| 606 |
+
"eval_VitaminC_cosine_f1": 0.6675531914893617,
|
| 607 |
+
"eval_VitaminC_cosine_f1_threshold": 0.30620089173316956,
|
| 608 |
+
"eval_VitaminC_cosine_precision": 0.500998003992016,
|
| 609 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 610 |
+
"eval_VitaminC_dot_accuracy": 0.5546875,
|
| 611 |
+
"eval_VitaminC_dot_accuracy_threshold": 308.60137939453125,
|
| 612 |
+
"eval_VitaminC_dot_ap": 0.5375184580780159,
|
| 613 |
+
"eval_VitaminC_dot_f1": 0.6657824933687002,
|
| 614 |
+
"eval_VitaminC_dot_f1_threshold": 97.275634765625,
|
| 615 |
+
"eval_VitaminC_dot_precision": 0.4990059642147117,
|
| 616 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 617 |
+
"eval_VitaminC_euclidean_accuracy": 0.552734375,
|
| 618 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 11.976862907409668,
|
| 619 |
+
"eval_VitaminC_euclidean_ap": 0.5494925067012235,
|
| 620 |
+
"eval_VitaminC_euclidean_f1": 0.6666666666666666,
|
| 621 |
+
"eval_VitaminC_euclidean_f1_threshold": 23.21343994140625,
|
| 622 |
+
"eval_VitaminC_euclidean_precision": 0.5,
|
| 623 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 624 |
+
"eval_VitaminC_manhattan_accuracy": 0.552734375,
|
| 625 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 313.34185791015625,
|
| 626 |
+
"eval_VitaminC_manhattan_ap": 0.5475158315491966,
|
| 627 |
+
"eval_VitaminC_manhattan_f1": 0.6666666666666666,
|
| 628 |
+
"eval_VitaminC_manhattan_f1_threshold": 495.06231689453125,
|
| 629 |
+
"eval_VitaminC_manhattan_precision": 0.5,
|
| 630 |
+
"eval_VitaminC_manhattan_recall": 1.0,
|
| 631 |
+
"eval_VitaminC_max_accuracy": 0.560546875,
|
| 632 |
+
"eval_VitaminC_max_accuracy_threshold": 313.34185791015625,
|
| 633 |
+
"eval_VitaminC_max_ap": 0.5494925067012235,
|
| 634 |
+
"eval_VitaminC_max_f1": 0.6675531914893617,
|
| 635 |
+
"eval_VitaminC_max_f1_threshold": 495.06231689453125,
|
| 636 |
+
"eval_VitaminC_max_precision": 0.500998003992016,
|
| 637 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 638 |
+
"eval_sequential_score": 0.5494925067012235,
|
| 639 |
+
"eval_sts-test_pearson_cosine": 0.8681028367252808,
|
| 640 |
+
"eval_sts-test_pearson_dot": 0.8578643818026934,
|
| 641 |
+
"eval_sts-test_pearson_euclidean": 0.8913506886125709,
|
| 642 |
+
"eval_sts-test_pearson_manhattan": 0.8922209656727235,
|
| 643 |
+
"eval_sts-test_pearson_max": 0.8922209656727235,
|
| 644 |
+
"eval_sts-test_spearman_cosine": 0.8960442588011338,
|
| 645 |
+
"eval_sts-test_spearman_dot": 0.8606696844578128,
|
| 646 |
+
"eval_sts-test_spearman_euclidean": 0.8895474944286376,
|
| 647 |
+
"eval_sts-test_spearman_manhattan": 0.8895341585527426,
|
| 648 |
+
"eval_sts-test_spearman_max": 0.8960442588011338,
|
| 649 |
+
"eval_vitaminc-pairs_loss": 2.260099411010742,
|
| 650 |
+
"eval_vitaminc-pairs_runtime": 1.8392,
|
| 651 |
+
"eval_vitaminc-pairs_samples_per_second": 58.723,
|
| 652 |
+
"eval_vitaminc-pairs_steps_per_second": 1.087,
|
| 653 |
+
"step": 64
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.2,
|
| 657 |
+
"eval_negation-triplets_loss": 0.836820662021637,
|
| 658 |
+
"eval_negation-triplets_runtime": 0.294,
|
| 659 |
+
"eval_negation-triplets_samples_per_second": 217.7,
|
| 660 |
+
"eval_negation-triplets_steps_per_second": 3.402,
|
| 661 |
+
"step": 64
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.2,
|
| 665 |
+
"eval_scitail-pairs-pos_loss": 0.08362159878015518,
|
| 666 |
+
"eval_scitail-pairs-pos_runtime": 0.3686,
|
| 667 |
+
"eval_scitail-pairs-pos_samples_per_second": 146.509,
|
| 668 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.713,
|
| 669 |
+
"step": 64
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.2,
|
| 673 |
+
"eval_xsum-pairs_loss": 0.08567425608634949,
|
| 674 |
+
"eval_xsum-pairs_runtime": 2.8489,
|
| 675 |
+
"eval_xsum-pairs_samples_per_second": 44.93,
|
| 676 |
+
"eval_xsum-pairs_steps_per_second": 0.702,
|
| 677 |
+
"step": 64
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 0.2,
|
| 681 |
+
"eval_sciq_pairs_loss": 0.019713517278432846,
|
| 682 |
+
"eval_sciq_pairs_runtime": 3.616,
|
| 683 |
+
"eval_sciq_pairs_samples_per_second": 35.399,
|
| 684 |
+
"eval_sciq_pairs_steps_per_second": 0.553,
|
| 685 |
+
"step": 64
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.2,
|
| 689 |
+
"eval_qasc_pairs_loss": 0.11403815448284149,
|
| 690 |
+
"eval_qasc_pairs_runtime": 0.6024,
|
| 691 |
+
"eval_qasc_pairs_samples_per_second": 212.48,
|
| 692 |
+
"eval_qasc_pairs_steps_per_second": 3.32,
|
| 693 |
+
"step": 64
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.2,
|
| 697 |
+
"eval_openbookqa_pairs_loss": 0.6793034076690674,
|
| 698 |
+
"eval_openbookqa_pairs_runtime": 0.5864,
|
| 699 |
+
"eval_openbookqa_pairs_samples_per_second": 218.266,
|
| 700 |
+
"eval_openbookqa_pairs_steps_per_second": 3.41,
|
| 701 |
+
"step": 64
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 0.2,
|
| 705 |
+
"eval_msmarco_pairs_loss": 0.34600257873535156,
|
| 706 |
+
"eval_msmarco_pairs_runtime": 1.4668,
|
| 707 |
+
"eval_msmarco_pairs_samples_per_second": 87.263,
|
| 708 |
+
"eval_msmarco_pairs_steps_per_second": 1.363,
|
| 709 |
+
"step": 64
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.2,
|
| 713 |
+
"eval_nq_pairs_loss": 0.22141708433628082,
|
| 714 |
+
"eval_nq_pairs_runtime": 2.8596,
|
| 715 |
+
"eval_nq_pairs_samples_per_second": 44.761,
|
| 716 |
+
"eval_nq_pairs_steps_per_second": 0.699,
|
| 717 |
+
"step": 64
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.2,
|
| 721 |
+
"eval_trivia_pairs_loss": 0.7303681969642639,
|
| 722 |
+
"eval_trivia_pairs_runtime": 4.3864,
|
| 723 |
+
"eval_trivia_pairs_samples_per_second": 29.181,
|
| 724 |
+
"eval_trivia_pairs_steps_per_second": 0.456,
|
| 725 |
+
"step": 64
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.2,
|
| 729 |
+
"eval_gooaq_pairs_loss": 0.38013964891433716,
|
| 730 |
+
"eval_gooaq_pairs_runtime": 1.0052,
|
| 731 |
+
"eval_gooaq_pairs_samples_per_second": 127.34,
|
| 732 |
+
"eval_gooaq_pairs_steps_per_second": 1.99,
|
| 733 |
+
"step": 64
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.2,
|
| 737 |
+
"eval_paws-pos_loss": 0.024541139602661133,
|
| 738 |
+
"eval_paws-pos_runtime": 0.6851,
|
| 739 |
+
"eval_paws-pos_samples_per_second": 186.844,
|
| 740 |
+
"eval_paws-pos_steps_per_second": 2.919,
|
| 741 |
+
"step": 64
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.203125,
|
| 745 |
+
"grad_norm": 3.5084540843963623,
|
| 746 |
+
"learning_rate": 2.0312500000000002e-05,
|
| 747 |
+
"loss": 0.7848,
|
| 748 |
+
"step": 65
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.20625,
|
| 752 |
+
"grad_norm": 3.749316453933716,
|
| 753 |
+
"learning_rate": 2.0625000000000003e-05,
|
| 754 |
+
"loss": 0.7589,
|
| 755 |
+
"step": 66
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 0.209375,
|
| 759 |
+
"grad_norm": 3.4131276607513428,
|
| 760 |
+
"learning_rate": 2.09375e-05,
|
| 761 |
+
"loss": 0.5905,
|
| 762 |
+
"step": 67
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.2125,
|
| 766 |
+
"grad_norm": 2.4543726444244385,
|
| 767 |
+
"learning_rate": 2.125e-05,
|
| 768 |
+
"loss": 0.4211,
|
| 769 |
+
"step": 68
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.215625,
|
| 773 |
+
"grad_norm": 2.6270904541015625,
|
| 774 |
+
"learning_rate": 2.1562500000000002e-05,
|
| 775 |
+
"loss": 0.5325,
|
| 776 |
+
"step": 69
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.21875,
|
| 780 |
+
"grad_norm": 2.2518444061279297,
|
| 781 |
+
"learning_rate": 2.1875000000000003e-05,
|
| 782 |
+
"loss": 0.3541,
|
| 783 |
+
"step": 70
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.221875,
|
| 787 |
+
"grad_norm": 3.88729190826416,
|
| 788 |
+
"learning_rate": 2.21875e-05,
|
| 789 |
+
"loss": 0.9396,
|
| 790 |
+
"step": 71
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.225,
|
| 794 |
+
"grad_norm": 3.2759203910827637,
|
| 795 |
+
"learning_rate": 2.25e-05,
|
| 796 |
+
"loss": 0.6997,
|
| 797 |
+
"step": 72
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.228125,
|
| 801 |
+
"grad_norm": 3.149787425994873,
|
| 802 |
+
"learning_rate": 2.2812500000000002e-05,
|
| 803 |
+
"loss": 0.6415,
|
| 804 |
+
"step": 73
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.23125,
|
| 808 |
+
"grad_norm": 4.01395845413208,
|
| 809 |
+
"learning_rate": 2.3125000000000003e-05,
|
| 810 |
+
"loss": 1.1966,
|
| 811 |
+
"step": 74
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.234375,
|
| 815 |
+
"grad_norm": 3.0432724952697754,
|
| 816 |
+
"learning_rate": 2.34375e-05,
|
| 817 |
+
"loss": 0.7142,
|
| 818 |
+
"step": 75
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.2375,
|
| 822 |
+
"grad_norm": 2.960078716278076,
|
| 823 |
+
"learning_rate": 2.375e-05,
|
| 824 |
+
"loss": 0.6048,
|
| 825 |
+
"step": 76
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.240625,
|
| 829 |
+
"grad_norm": 2.414846658706665,
|
| 830 |
+
"learning_rate": 2.4062500000000002e-05,
|
| 831 |
+
"loss": 0.4639,
|
| 832 |
+
"step": 77
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.24375,
|
| 836 |
+
"grad_norm": 4.241907119750977,
|
| 837 |
+
"learning_rate": 2.4375000000000003e-05,
|
| 838 |
+
"loss": 0.9391,
|
| 839 |
+
"step": 78
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 0.246875,
|
| 843 |
+
"grad_norm": 3.350724220275879,
|
| 844 |
+
"learning_rate": 2.46875e-05,
|
| 845 |
+
"loss": 0.6364,
|
| 846 |
+
"step": 79
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 0.25,
|
| 850 |
+
"grad_norm": 2.519324541091919,
|
| 851 |
+
"learning_rate": 2.5e-05,
|
| 852 |
+
"loss": 0.515,
|
| 853 |
+
"step": 80
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.253125,
|
| 857 |
+
"grad_norm": 3.655949592590332,
|
| 858 |
+
"learning_rate": 2.5312500000000002e-05,
|
| 859 |
+
"loss": 0.6505,
|
| 860 |
+
"step": 81
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.25625,
|
| 864 |
+
"grad_norm": 3.1521031856536865,
|
| 865 |
+
"learning_rate": 2.5625000000000003e-05,
|
| 866 |
+
"loss": 0.6149,
|
| 867 |
+
"step": 82
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.259375,
|
| 871 |
+
"grad_norm": 2.637176036834717,
|
| 872 |
+
"learning_rate": 2.5937500000000004e-05,
|
| 873 |
+
"loss": 0.4471,
|
| 874 |
+
"step": 83
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.2625,
|
| 878 |
+
"grad_norm": 4.223080158233643,
|
| 879 |
+
"learning_rate": 2.625e-05,
|
| 880 |
+
"loss": 1.4199,
|
| 881 |
+
"step": 84
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 0.265625,
|
| 885 |
+
"grad_norm": 3.141789436340332,
|
| 886 |
+
"learning_rate": 2.6562500000000002e-05,
|
| 887 |
+
"loss": 0.8484,
|
| 888 |
+
"step": 85
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.26875,
|
| 892 |
+
"grad_norm": 3.2342255115509033,
|
| 893 |
+
"learning_rate": 2.6875000000000003e-05,
|
| 894 |
+
"loss": 0.6412,
|
| 895 |
+
"step": 86
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.271875,
|
| 899 |
+
"grad_norm": 3.445375442504883,
|
| 900 |
+
"learning_rate": 2.7187500000000004e-05,
|
| 901 |
+
"loss": 0.65,
|
| 902 |
+
"step": 87
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.275,
|
| 906 |
+
"grad_norm": 3.395848035812378,
|
| 907 |
+
"learning_rate": 2.75e-05,
|
| 908 |
+
"loss": 0.7453,
|
| 909 |
+
"step": 88
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.278125,
|
| 913 |
+
"grad_norm": 3.752084493637085,
|
| 914 |
+
"learning_rate": 2.7812500000000002e-05,
|
| 915 |
+
"loss": 0.9506,
|
| 916 |
+
"step": 89
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.28125,
|
| 920 |
+
"grad_norm": 3.2424893379211426,
|
| 921 |
+
"learning_rate": 2.8125000000000003e-05,
|
| 922 |
+
"loss": 0.6083,
|
| 923 |
+
"step": 90
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.284375,
|
| 927 |
+
"grad_norm": 2.8851892948150635,
|
| 928 |
+
"learning_rate": 2.8437500000000003e-05,
|
| 929 |
+
"loss": 0.7102,
|
| 930 |
+
"step": 91
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 0.2875,
|
| 934 |
+
"grad_norm": 2.385157823562622,
|
| 935 |
+
"learning_rate": 2.875e-05,
|
| 936 |
+
"loss": 0.4037,
|
| 937 |
+
"step": 92
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.290625,
|
| 941 |
+
"grad_norm": 3.5539441108703613,
|
| 942 |
+
"learning_rate": 2.90625e-05,
|
| 943 |
+
"loss": 0.769,
|
| 944 |
+
"step": 93
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.29375,
|
| 948 |
+
"grad_norm": 3.686418056488037,
|
| 949 |
+
"learning_rate": 2.9375000000000003e-05,
|
| 950 |
+
"loss": 0.8765,
|
| 951 |
+
"step": 94
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.296875,
|
| 955 |
+
"grad_norm": 3.9195055961608887,
|
| 956 |
+
"learning_rate": 2.9687500000000003e-05,
|
| 957 |
+
"loss": 1.2583,
|
| 958 |
+
"step": 95
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.3,
|
| 962 |
+
"grad_norm": 3.5373759269714355,
|
| 963 |
+
"learning_rate": 3.0000000000000004e-05,
|
| 964 |
+
"loss": 0.8885,
|
| 965 |
+
"step": 96
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 0.3,
|
| 969 |
+
"eval_VitaminC_cosine_accuracy": 0.55859375,
|
| 970 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.8407348990440369,
|
| 971 |
+
"eval_VitaminC_cosine_ap": 0.5524635737287826,
|
| 972 |
+
"eval_VitaminC_cosine_f1": 0.6666666666666666,
|
| 973 |
+
"eval_VitaminC_cosine_f1_threshold": 0.2901695668697357,
|
| 974 |
+
"eval_VitaminC_cosine_precision": 0.5,
|
| 975 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 976 |
+
"eval_VitaminC_dot_accuracy": 0.5546875,
|
| 977 |
+
"eval_VitaminC_dot_accuracy_threshold": 331.7409973144531,
|
| 978 |
+
"eval_VitaminC_dot_ap": 0.5393192469559877,
|
| 979 |
+
"eval_VitaminC_dot_f1": 0.6657824933687002,
|
| 980 |
+
"eval_VitaminC_dot_f1_threshold": 104.93923950195312,
|
| 981 |
+
"eval_VitaminC_dot_precision": 0.4990059642147117,
|
| 982 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 983 |
+
"eval_VitaminC_euclidean_accuracy": 0.5546875,
|
| 984 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 10.979323387145996,
|
| 985 |
+
"eval_VitaminC_euclidean_ap": 0.5510789245842218,
|
| 986 |
+
"eval_VitaminC_euclidean_f1": 0.6666666666666666,
|
| 987 |
+
"eval_VitaminC_euclidean_f1_threshold": 23.105466842651367,
|
| 988 |
+
"eval_VitaminC_euclidean_precision": 0.5,
|
| 989 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 990 |
+
"eval_VitaminC_manhattan_accuracy": 0.55078125,
|
| 991 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 228.8612060546875,
|
| 992 |
+
"eval_VitaminC_manhattan_ap": 0.550140326019901,
|
| 993 |
+
"eval_VitaminC_manhattan_f1": 0.6666666666666667,
|
| 994 |
+
"eval_VitaminC_manhattan_f1_threshold": 479.256103515625,
|
| 995 |
+
"eval_VitaminC_manhattan_precision": 0.501002004008016,
|
| 996 |
+
"eval_VitaminC_manhattan_recall": 0.9960159362549801,
|
| 997 |
+
"eval_VitaminC_max_accuracy": 0.55859375,
|
| 998 |
+
"eval_VitaminC_max_accuracy_threshold": 331.7409973144531,
|
| 999 |
+
"eval_VitaminC_max_ap": 0.5524635737287826,
|
| 1000 |
+
"eval_VitaminC_max_f1": 0.6666666666666667,
|
| 1001 |
+
"eval_VitaminC_max_f1_threshold": 479.256103515625,
|
| 1002 |
+
"eval_VitaminC_max_precision": 0.501002004008016,
|
| 1003 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 1004 |
+
"eval_sequential_score": 0.5524635737287826,
|
| 1005 |
+
"eval_sts-test_pearson_cosine": 0.8707252459918289,
|
| 1006 |
+
"eval_sts-test_pearson_dot": 0.8616721319399807,
|
| 1007 |
+
"eval_sts-test_pearson_euclidean": 0.8926205493906139,
|
| 1008 |
+
"eval_sts-test_pearson_manhattan": 0.8931067612799872,
|
| 1009 |
+
"eval_sts-test_pearson_max": 0.8931067612799872,
|
| 1010 |
+
"eval_sts-test_spearman_cosine": 0.8969095691913977,
|
| 1011 |
+
"eval_sts-test_spearman_dot": 0.8614390033923923,
|
| 1012 |
+
"eval_sts-test_spearman_euclidean": 0.8906887410966409,
|
| 1013 |
+
"eval_sts-test_spearman_manhattan": 0.8902939007173846,
|
| 1014 |
+
"eval_sts-test_spearman_max": 0.8969095691913977,
|
| 1015 |
+
"eval_vitaminc-pairs_loss": 2.259434938430786,
|
| 1016 |
+
"eval_vitaminc-pairs_runtime": 1.8587,
|
| 1017 |
+
"eval_vitaminc-pairs_samples_per_second": 58.104,
|
| 1018 |
+
"eval_vitaminc-pairs_steps_per_second": 1.076,
|
| 1019 |
+
"step": 96
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 0.3,
|
| 1023 |
+
"eval_negation-triplets_loss": 0.8346852660179138,
|
| 1024 |
+
"eval_negation-triplets_runtime": 0.2932,
|
| 1025 |
+
"eval_negation-triplets_samples_per_second": 218.315,
|
| 1026 |
+
"eval_negation-triplets_steps_per_second": 3.411,
|
| 1027 |
+
"step": 96
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.3,
|
| 1031 |
+
"eval_scitail-pairs-pos_loss": 0.07568605989217758,
|
| 1032 |
+
"eval_scitail-pairs-pos_runtime": 0.3763,
|
| 1033 |
+
"eval_scitail-pairs-pos_samples_per_second": 143.494,
|
| 1034 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.657,
|
| 1035 |
+
"step": 96
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.3,
|
| 1039 |
+
"eval_xsum-pairs_loss": 0.08208194375038147,
|
| 1040 |
+
"eval_xsum-pairs_runtime": 2.8486,
|
| 1041 |
+
"eval_xsum-pairs_samples_per_second": 44.934,
|
| 1042 |
+
"eval_xsum-pairs_steps_per_second": 0.702,
|
| 1043 |
+
"step": 96
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 0.3,
|
| 1047 |
+
"eval_sciq_pairs_loss": 0.020024314522743225,
|
| 1048 |
+
"eval_sciq_pairs_runtime": 3.6173,
|
| 1049 |
+
"eval_sciq_pairs_samples_per_second": 35.386,
|
| 1050 |
+
"eval_sciq_pairs_steps_per_second": 0.553,
|
| 1051 |
+
"step": 96
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.3,
|
| 1055 |
+
"eval_qasc_pairs_loss": 0.10592304170131683,
|
| 1056 |
+
"eval_qasc_pairs_runtime": 0.5997,
|
| 1057 |
+
"eval_qasc_pairs_samples_per_second": 213.431,
|
| 1058 |
+
"eval_qasc_pairs_steps_per_second": 3.335,
|
| 1059 |
+
"step": 96
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3,
|
| 1063 |
+
"eval_openbookqa_pairs_loss": 0.6809090971946716,
|
| 1064 |
+
"eval_openbookqa_pairs_runtime": 0.5752,
|
| 1065 |
+
"eval_openbookqa_pairs_samples_per_second": 222.54,
|
| 1066 |
+
"eval_openbookqa_pairs_steps_per_second": 3.477,
|
| 1067 |
+
"step": 96
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.3,
|
| 1071 |
+
"eval_msmarco_pairs_loss": 0.3400232195854187,
|
| 1072 |
+
"eval_msmarco_pairs_runtime": 1.4679,
|
| 1073 |
+
"eval_msmarco_pairs_samples_per_second": 87.202,
|
| 1074 |
+
"eval_msmarco_pairs_steps_per_second": 1.363,
|
| 1075 |
+
"step": 96
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.3,
|
| 1079 |
+
"eval_nq_pairs_loss": 0.2074178159236908,
|
| 1080 |
+
"eval_nq_pairs_runtime": 2.8593,
|
| 1081 |
+
"eval_nq_pairs_samples_per_second": 44.766,
|
| 1082 |
+
"eval_nq_pairs_steps_per_second": 0.699,
|
| 1083 |
+
"step": 96
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.3,
|
| 1087 |
+
"eval_trivia_pairs_loss": 0.7431399822235107,
|
| 1088 |
+
"eval_trivia_pairs_runtime": 4.4162,
|
| 1089 |
+
"eval_trivia_pairs_samples_per_second": 28.984,
|
| 1090 |
+
"eval_trivia_pairs_steps_per_second": 0.453,
|
| 1091 |
+
"step": 96
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.3,
|
| 1095 |
+
"eval_gooaq_pairs_loss": 0.3708875775337219,
|
| 1096 |
+
"eval_gooaq_pairs_runtime": 1.0094,
|
| 1097 |
+
"eval_gooaq_pairs_samples_per_second": 126.81,
|
| 1098 |
+
"eval_gooaq_pairs_steps_per_second": 1.981,
|
| 1099 |
+
"step": 96
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 0.3,
|
| 1103 |
+
"eval_paws-pos_loss": 0.024763749912381172,
|
| 1104 |
+
"eval_paws-pos_runtime": 0.6874,
|
| 1105 |
+
"eval_paws-pos_samples_per_second": 186.212,
|
| 1106 |
+
"eval_paws-pos_steps_per_second": 2.91,
|
| 1107 |
+
"step": 96
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.303125,
|
| 1111 |
+
"grad_norm": 3.2354822158813477,
|
| 1112 |
+
"learning_rate": 3.03125e-05,
|
| 1113 |
+
"loss": 0.6398,
|
| 1114 |
+
"step": 97
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.30625,
|
| 1118 |
+
"grad_norm": 3.6665022373199463,
|
| 1119 |
+
"learning_rate": 3.0625000000000006e-05,
|
| 1120 |
+
"loss": 0.8263,
|
| 1121 |
+
"step": 98
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 0.309375,
|
| 1125 |
+
"grad_norm": 3.026954412460327,
|
| 1126 |
+
"learning_rate": 3.09375e-05,
|
| 1127 |
+
"loss": 0.8716,
|
| 1128 |
+
"step": 99
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 0.3125,
|
| 1132 |
+
"grad_norm": 2.445453643798828,
|
| 1133 |
+
"learning_rate": 3.125e-05,
|
| 1134 |
+
"loss": 0.5523,
|
| 1135 |
+
"step": 100
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 0.315625,
|
| 1139 |
+
"grad_norm": 3.4408035278320312,
|
| 1140 |
+
"learning_rate": 3.15625e-05,
|
| 1141 |
+
"loss": 0.5811,
|
| 1142 |
+
"step": 101
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.31875,
|
| 1146 |
+
"grad_norm": 2.8406240940093994,
|
| 1147 |
+
"learning_rate": 3.1875e-05,
|
| 1148 |
+
"loss": 0.7602,
|
| 1149 |
+
"step": 102
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.321875,
|
| 1153 |
+
"grad_norm": 2.5201492309570312,
|
| 1154 |
+
"learning_rate": 3.21875e-05,
|
| 1155 |
+
"loss": 0.5337,
|
| 1156 |
+
"step": 103
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.325,
|
| 1160 |
+
"grad_norm": 3.323239326477051,
|
| 1161 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 1162 |
+
"loss": 0.8182,
|
| 1163 |
+
"step": 104
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 0.328125,
|
| 1167 |
+
"grad_norm": 3.2463977336883545,
|
| 1168 |
+
"learning_rate": 3.2812500000000005e-05,
|
| 1169 |
+
"loss": 0.6641,
|
| 1170 |
+
"step": 105
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 0.33125,
|
| 1174 |
+
"grad_norm": 3.4495010375976562,
|
| 1175 |
+
"learning_rate": 3.3125000000000006e-05,
|
| 1176 |
+
"loss": 1.0088,
|
| 1177 |
+
"step": 106
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 0.334375,
|
| 1181 |
+
"grad_norm": 2.7572243213653564,
|
| 1182 |
+
"learning_rate": 3.34375e-05,
|
| 1183 |
+
"loss": 0.7556,
|
| 1184 |
+
"step": 107
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.3375,
|
| 1188 |
+
"grad_norm": 3.494549512863159,
|
| 1189 |
+
"learning_rate": 3.375e-05,
|
| 1190 |
+
"loss": 0.713,
|
| 1191 |
+
"step": 108
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.340625,
|
| 1195 |
+
"grad_norm": 3.4666013717651367,
|
| 1196 |
+
"learning_rate": 3.40625e-05,
|
| 1197 |
+
"loss": 0.8385,
|
| 1198 |
+
"step": 109
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 0.34375,
|
| 1202 |
+
"grad_norm": 3.05104660987854,
|
| 1203 |
+
"learning_rate": 3.4375e-05,
|
| 1204 |
+
"loss": 0.5181,
|
| 1205 |
+
"step": 110
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 0.346875,
|
| 1209 |
+
"grad_norm": 3.8259003162384033,
|
| 1210 |
+
"learning_rate": 3.46875e-05,
|
| 1211 |
+
"loss": 1.0939,
|
| 1212 |
+
"step": 111
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 0.35,
|
| 1216 |
+
"grad_norm": 3.287792205810547,
|
| 1217 |
+
"learning_rate": 3.5000000000000004e-05,
|
| 1218 |
+
"loss": 0.5826,
|
| 1219 |
+
"step": 112
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.353125,
|
| 1223 |
+
"grad_norm": 3.9174458980560303,
|
| 1224 |
+
"learning_rate": 3.5312500000000005e-05,
|
| 1225 |
+
"loss": 0.7121,
|
| 1226 |
+
"step": 113
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.35625,
|
| 1230 |
+
"grad_norm": 3.424893379211426,
|
| 1231 |
+
"learning_rate": 3.5625000000000005e-05,
|
| 1232 |
+
"loss": 0.9371,
|
| 1233 |
+
"step": 114
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.359375,
|
| 1237 |
+
"grad_norm": 3.5157482624053955,
|
| 1238 |
+
"learning_rate": 3.5937500000000006e-05,
|
| 1239 |
+
"loss": 0.7739,
|
| 1240 |
+
"step": 115
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 0.3625,
|
| 1244 |
+
"grad_norm": 4.468640327453613,
|
| 1245 |
+
"learning_rate": 3.625e-05,
|
| 1246 |
+
"loss": 0.9612,
|
| 1247 |
+
"step": 116
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 0.365625,
|
| 1251 |
+
"grad_norm": 3.4379608631134033,
|
| 1252 |
+
"learning_rate": 3.65625e-05,
|
| 1253 |
+
"loss": 0.7213,
|
| 1254 |
+
"step": 117
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 0.36875,
|
| 1258 |
+
"grad_norm": 2.9453623294830322,
|
| 1259 |
+
"learning_rate": 3.6875e-05,
|
| 1260 |
+
"loss": 0.621,
|
| 1261 |
+
"step": 118
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 0.371875,
|
| 1265 |
+
"grad_norm": 2.4365315437316895,
|
| 1266 |
+
"learning_rate": 3.71875e-05,
|
| 1267 |
+
"loss": 0.5503,
|
| 1268 |
+
"step": 119
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.375,
|
| 1272 |
+
"grad_norm": 3.446967124938965,
|
| 1273 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 1274 |
+
"loss": 0.8439,
|
| 1275 |
+
"step": 120
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 0.378125,
|
| 1279 |
+
"grad_norm": 3.8797788619995117,
|
| 1280 |
+
"learning_rate": 3.7812500000000004e-05,
|
| 1281 |
+
"loss": 0.7813,
|
| 1282 |
+
"step": 121
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 0.38125,
|
| 1286 |
+
"grad_norm": 3.0103230476379395,
|
| 1287 |
+
"learning_rate": 3.8125000000000005e-05,
|
| 1288 |
+
"loss": 0.5637,
|
| 1289 |
+
"step": 122
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 0.384375,
|
| 1293 |
+
"grad_norm": 3.9547793865203857,
|
| 1294 |
+
"learning_rate": 3.8437500000000006e-05,
|
| 1295 |
+
"loss": 0.9052,
|
| 1296 |
+
"step": 123
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 0.3875,
|
| 1300 |
+
"grad_norm": 2.953261375427246,
|
| 1301 |
+
"learning_rate": 3.875e-05,
|
| 1302 |
+
"loss": 0.64,
|
| 1303 |
+
"step": 124
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 0.390625,
|
| 1307 |
+
"grad_norm": 2.914365768432617,
|
| 1308 |
+
"learning_rate": 3.90625e-05,
|
| 1309 |
+
"loss": 0.6529,
|
| 1310 |
+
"step": 125
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.39375,
|
| 1314 |
+
"grad_norm": 3.346844434738159,
|
| 1315 |
+
"learning_rate": 3.9375e-05,
|
| 1316 |
+
"loss": 0.6894,
|
| 1317 |
+
"step": 126
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 0.396875,
|
| 1321 |
+
"grad_norm": 3.946427583694458,
|
| 1322 |
+
"learning_rate": 3.96875e-05,
|
| 1323 |
+
"loss": 0.8604,
|
| 1324 |
+
"step": 127
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 0.4,
|
| 1328 |
+
"grad_norm": 3.3265583515167236,
|
| 1329 |
+
"learning_rate": 4e-05,
|
| 1330 |
+
"loss": 0.8503,
|
| 1331 |
+
"step": 128
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 0.4,
|
| 1335 |
+
"eval_VitaminC_cosine_accuracy": 0.55859375,
|
| 1336 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.835027813911438,
|
| 1337 |
+
"eval_VitaminC_cosine_ap": 0.5482054260732142,
|
| 1338 |
+
"eval_VitaminC_cosine_f1": 0.6666666666666666,
|
| 1339 |
+
"eval_VitaminC_cosine_f1_threshold": 0.28428012132644653,
|
| 1340 |
+
"eval_VitaminC_cosine_precision": 0.5,
|
| 1341 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 1342 |
+
"eval_VitaminC_dot_accuracy": 0.548828125,
|
| 1343 |
+
"eval_VitaminC_dot_accuracy_threshold": 321.1236572265625,
|
| 1344 |
+
"eval_VitaminC_dot_ap": 0.5350248143918641,
|
| 1345 |
+
"eval_VitaminC_dot_f1": 0.6649006622516557,
|
| 1346 |
+
"eval_VitaminC_dot_f1_threshold": 94.1016616821289,
|
| 1347 |
+
"eval_VitaminC_dot_precision": 0.498015873015873,
|
| 1348 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 1349 |
+
"eval_VitaminC_euclidean_accuracy": 0.55859375,
|
| 1350 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 13.260427474975586,
|
| 1351 |
+
"eval_VitaminC_euclidean_ap": 0.551773706587656,
|
| 1352 |
+
"eval_VitaminC_euclidean_f1": 0.6657824933687002,
|
| 1353 |
+
"eval_VitaminC_euclidean_f1_threshold": 23.911056518554688,
|
| 1354 |
+
"eval_VitaminC_euclidean_precision": 0.4990059642147117,
|
| 1355 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 1356 |
+
"eval_VitaminC_manhattan_accuracy": 0.55859375,
|
| 1357 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 273.4624328613281,
|
| 1358 |
+
"eval_VitaminC_manhattan_ap": 0.5494410762635437,
|
| 1359 |
+
"eval_VitaminC_manhattan_f1": 0.6666666666666667,
|
| 1360 |
+
"eval_VitaminC_manhattan_f1_threshold": 472.7373046875,
|
| 1361 |
+
"eval_VitaminC_manhattan_precision": 0.5020161290322581,
|
| 1362 |
+
"eval_VitaminC_manhattan_recall": 0.9920318725099602,
|
| 1363 |
+
"eval_VitaminC_max_accuracy": 0.55859375,
|
| 1364 |
+
"eval_VitaminC_max_accuracy_threshold": 321.1236572265625,
|
| 1365 |
+
"eval_VitaminC_max_ap": 0.551773706587656,
|
| 1366 |
+
"eval_VitaminC_max_f1": 0.6666666666666667,
|
| 1367 |
+
"eval_VitaminC_max_f1_threshold": 472.7373046875,
|
| 1368 |
+
"eval_VitaminC_max_precision": 0.5020161290322581,
|
| 1369 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 1370 |
+
"eval_sequential_score": 0.551773706587656,
|
| 1371 |
+
"eval_sts-test_pearson_cosine": 0.8672675483925697,
|
| 1372 |
+
"eval_sts-test_pearson_dot": 0.8586110849200466,
|
| 1373 |
+
"eval_sts-test_pearson_euclidean": 0.8915515585715386,
|
| 1374 |
+
"eval_sts-test_pearson_manhattan": 0.8913674606593633,
|
| 1375 |
+
"eval_sts-test_pearson_max": 0.8915515585715386,
|
| 1376 |
+
"eval_sts-test_spearman_cosine": 0.8969123885208655,
|
| 1377 |
+
"eval_sts-test_spearman_dot": 0.8619306407500383,
|
| 1378 |
+
"eval_sts-test_spearman_euclidean": 0.8903670690297594,
|
| 1379 |
+
"eval_sts-test_spearman_manhattan": 0.890351227083227,
|
| 1380 |
+
"eval_sts-test_spearman_max": 0.8969123885208655,
|
| 1381 |
+
"eval_vitaminc-pairs_loss": 2.0338199138641357,
|
| 1382 |
+
"eval_vitaminc-pairs_runtime": 1.8309,
|
| 1383 |
+
"eval_vitaminc-pairs_samples_per_second": 58.988,
|
| 1384 |
+
"eval_vitaminc-pairs_steps_per_second": 1.092,
|
| 1385 |
+
"step": 128
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 0.4,
|
| 1389 |
+
"eval_negation-triplets_loss": 0.7916581630706787,
|
| 1390 |
+
"eval_negation-triplets_runtime": 0.2912,
|
| 1391 |
+
"eval_negation-triplets_samples_per_second": 219.766,
|
| 1392 |
+
"eval_negation-triplets_steps_per_second": 3.434,
|
| 1393 |
+
"step": 128
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 0.4,
|
| 1397 |
+
"eval_scitail-pairs-pos_loss": 0.07755717635154724,
|
| 1398 |
+
"eval_scitail-pairs-pos_runtime": 0.3716,
|
| 1399 |
+
"eval_scitail-pairs-pos_samples_per_second": 145.312,
|
| 1400 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.691,
|
| 1401 |
+
"step": 128
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 0.4,
|
| 1405 |
+
"eval_xsum-pairs_loss": 0.08196285367012024,
|
| 1406 |
+
"eval_xsum-pairs_runtime": 2.852,
|
| 1407 |
+
"eval_xsum-pairs_samples_per_second": 44.881,
|
| 1408 |
+
"eval_xsum-pairs_steps_per_second": 0.701,
|
| 1409 |
+
"step": 128
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4,
|
| 1413 |
+
"eval_sciq_pairs_loss": 0.020960956811904907,
|
| 1414 |
+
"eval_sciq_pairs_runtime": 3.5913,
|
| 1415 |
+
"eval_sciq_pairs_samples_per_second": 35.642,
|
| 1416 |
+
"eval_sciq_pairs_steps_per_second": 0.557,
|
| 1417 |
+
"step": 128
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.4,
|
| 1421 |
+
"eval_qasc_pairs_loss": 0.11308694630861282,
|
| 1422 |
+
"eval_qasc_pairs_runtime": 0.595,
|
| 1423 |
+
"eval_qasc_pairs_samples_per_second": 215.137,
|
| 1424 |
+
"eval_qasc_pairs_steps_per_second": 3.362,
|
| 1425 |
+
"step": 128
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.4,
|
| 1429 |
+
"eval_openbookqa_pairs_loss": 0.7888042330741882,
|
| 1430 |
+
"eval_openbookqa_pairs_runtime": 0.5711,
|
| 1431 |
+
"eval_openbookqa_pairs_samples_per_second": 224.114,
|
| 1432 |
+
"eval_openbookqa_pairs_steps_per_second": 3.502,
|
| 1433 |
+
"step": 128
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 0.4,
|
| 1437 |
+
"eval_msmarco_pairs_loss": 0.3428971469402313,
|
| 1438 |
+
"eval_msmarco_pairs_runtime": 1.465,
|
| 1439 |
+
"eval_msmarco_pairs_samples_per_second": 87.373,
|
| 1440 |
+
"eval_msmarco_pairs_steps_per_second": 1.365,
|
| 1441 |
+
"step": 128
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 0.4,
|
| 1445 |
+
"eval_nq_pairs_loss": 0.20846250653266907,
|
| 1446 |
+
"eval_nq_pairs_runtime": 2.8581,
|
| 1447 |
+
"eval_nq_pairs_samples_per_second": 44.786,
|
| 1448 |
+
"eval_nq_pairs_steps_per_second": 0.7,
|
| 1449 |
+
"step": 128
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.4,
|
| 1453 |
+
"eval_trivia_pairs_loss": 0.7110738754272461,
|
| 1454 |
+
"eval_trivia_pairs_runtime": 4.3917,
|
| 1455 |
+
"eval_trivia_pairs_samples_per_second": 29.146,
|
| 1456 |
+
"eval_trivia_pairs_steps_per_second": 0.455,
|
| 1457 |
+
"step": 128
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 0.4,
|
| 1461 |
+
"eval_gooaq_pairs_loss": 0.3744402229785919,
|
| 1462 |
+
"eval_gooaq_pairs_runtime": 1.0043,
|
| 1463 |
+
"eval_gooaq_pairs_samples_per_second": 127.448,
|
| 1464 |
+
"eval_gooaq_pairs_steps_per_second": 1.991,
|
| 1465 |
+
"step": 128
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.4,
|
| 1469 |
+
"eval_paws-pos_loss": 0.024828137829899788,
|
| 1470 |
+
"eval_paws-pos_runtime": 0.6859,
|
| 1471 |
+
"eval_paws-pos_samples_per_second": 186.611,
|
| 1472 |
+
"eval_paws-pos_steps_per_second": 2.916,
|
| 1473 |
+
"step": 128
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.403125,
|
| 1477 |
+
"grad_norm": 3.7963619232177734,
|
| 1478 |
+
"learning_rate": 3.999971762923902e-05,
|
| 1479 |
+
"loss": 0.8171,
|
| 1480 |
+
"step": 129
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.40625,
|
| 1484 |
+
"grad_norm": 3.987645387649536,
|
| 1485 |
+
"learning_rate": 3.999887052758717e-05,
|
| 1486 |
+
"loss": 1.0401,
|
| 1487 |
+
"step": 130
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.409375,
|
| 1491 |
+
"grad_norm": 2.653578758239746,
|
| 1492 |
+
"learning_rate": 3.999745872693735e-05,
|
| 1493 |
+
"loss": 0.4243,
|
| 1494 |
+
"step": 131
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.4125,
|
| 1498 |
+
"grad_norm": 2.3737175464630127,
|
| 1499 |
+
"learning_rate": 3.9995482280443065e-05,
|
| 1500 |
+
"loss": 0.3778,
|
| 1501 |
+
"step": 132
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.415625,
|
| 1505 |
+
"grad_norm": 3.334118127822876,
|
| 1506 |
+
"learning_rate": 3.99929412625164e-05,
|
| 1507 |
+
"loss": 0.7651,
|
| 1508 |
+
"step": 133
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.41875,
|
| 1512 |
+
"grad_norm": 3.5098752975463867,
|
| 1513 |
+
"learning_rate": 3.998983576882524e-05,
|
| 1514 |
+
"loss": 0.6003,
|
| 1515 |
+
"step": 134
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.421875,
|
| 1519 |
+
"grad_norm": 3.023698091506958,
|
| 1520 |
+
"learning_rate": 3.9986165916289686e-05,
|
| 1521 |
+
"loss": 0.6023,
|
| 1522 |
+
"step": 135
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.425,
|
| 1526 |
+
"grad_norm": 3.293668746948242,
|
| 1527 |
+
"learning_rate": 3.998193184307759e-05,
|
| 1528 |
+
"loss": 0.6079,
|
| 1529 |
+
"step": 136
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.428125,
|
| 1533 |
+
"grad_norm": 3.326125144958496,
|
| 1534 |
+
"learning_rate": 3.997713370859942e-05,
|
| 1535 |
+
"loss": 0.6206,
|
| 1536 |
+
"step": 137
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.43125,
|
| 1540 |
+
"grad_norm": 3.322040557861328,
|
| 1541 |
+
"learning_rate": 3.997177169350224e-05,
|
| 1542 |
+
"loss": 0.4694,
|
| 1543 |
+
"step": 138
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.434375,
|
| 1547 |
+
"grad_norm": 3.1219382286071777,
|
| 1548 |
+
"learning_rate": 3.996584599966288e-05,
|
| 1549 |
+
"loss": 0.7528,
|
| 1550 |
+
"step": 139
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.4375,
|
| 1554 |
+
"grad_norm": 3.7076480388641357,
|
| 1555 |
+
"learning_rate": 3.9959356850180354e-05,
|
| 1556 |
+
"loss": 0.8395,
|
| 1557 |
+
"step": 140
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.440625,
|
| 1561 |
+
"grad_norm": 3.1098551750183105,
|
| 1562 |
+
"learning_rate": 3.995230448936749e-05,
|
| 1563 |
+
"loss": 0.6689,
|
| 1564 |
+
"step": 141
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 0.44375,
|
| 1568 |
+
"grad_norm": 3.31339168548584,
|
| 1569 |
+
"learning_rate": 3.9944689182741674e-05,
|
| 1570 |
+
"loss": 0.6547,
|
| 1571 |
+
"step": 142
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.446875,
|
| 1575 |
+
"grad_norm": 4.2841386795043945,
|
| 1576 |
+
"learning_rate": 3.99365112170149e-05,
|
| 1577 |
+
"loss": 0.9242,
|
| 1578 |
+
"step": 143
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.45,
|
| 1582 |
+
"grad_norm": 4.0628132820129395,
|
| 1583 |
+
"learning_rate": 3.992777090008296e-05,
|
| 1584 |
+
"loss": 0.9496,
|
| 1585 |
+
"step": 144
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 0.453125,
|
| 1589 |
+
"grad_norm": 3.484614849090576,
|
| 1590 |
+
"learning_rate": 3.9918468561013834e-05,
|
| 1591 |
+
"loss": 0.6506,
|
| 1592 |
+
"step": 145
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.45625,
|
| 1596 |
+
"grad_norm": 3.4139559268951416,
|
| 1597 |
+
"learning_rate": 3.990860455003534e-05,
|
| 1598 |
+
"loss": 0.786,
|
| 1599 |
+
"step": 146
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.459375,
|
| 1603 |
+
"grad_norm": 3.4322853088378906,
|
| 1604 |
+
"learning_rate": 3.9898179238521916e-05,
|
| 1605 |
+
"loss": 0.7414,
|
| 1606 |
+
"step": 147
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 0.4625,
|
| 1610 |
+
"grad_norm": 2.660554885864258,
|
| 1611 |
+
"learning_rate": 3.9887193018980654e-05,
|
| 1612 |
+
"loss": 0.3978,
|
| 1613 |
+
"step": 148
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 0.465625,
|
| 1617 |
+
"grad_norm": 2.6429054737091064,
|
| 1618 |
+
"learning_rate": 3.9875646305036494e-05,
|
| 1619 |
+
"loss": 0.5635,
|
| 1620 |
+
"step": 149
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.46875,
|
| 1624 |
+
"grad_norm": 4.292131423950195,
|
| 1625 |
+
"learning_rate": 3.98635395314167e-05,
|
| 1626 |
+
"loss": 0.9466,
|
| 1627 |
+
"step": 150
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 0.471875,
|
| 1631 |
+
"grad_norm": 3.1115028858184814,
|
| 1632 |
+
"learning_rate": 3.9850873153934456e-05,
|
| 1633 |
+
"loss": 0.5251,
|
| 1634 |
+
"step": 151
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.475,
|
| 1638 |
+
"grad_norm": 3.307051181793213,
|
| 1639 |
+
"learning_rate": 3.983764764947172e-05,
|
| 1640 |
+
"loss": 0.6636,
|
| 1641 |
+
"step": 152
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.478125,
|
| 1645 |
+
"grad_norm": 3.807854652404785,
|
| 1646 |
+
"learning_rate": 3.9823863515961245e-05,
|
| 1647 |
+
"loss": 0.7834,
|
| 1648 |
+
"step": 153
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.48125,
|
| 1652 |
+
"grad_norm": 2.9957728385925293,
|
| 1653 |
+
"learning_rate": 3.980952127236788e-05,
|
| 1654 |
+
"loss": 0.6177,
|
| 1655 |
+
"step": 154
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 0.484375,
|
| 1659 |
+
"grad_norm": 3.3072471618652344,
|
| 1660 |
+
"learning_rate": 3.979462145866898e-05,
|
| 1661 |
+
"loss": 0.4558,
|
| 1662 |
+
"step": 155
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.4875,
|
| 1666 |
+
"grad_norm": 3.0199949741363525,
|
| 1667 |
+
"learning_rate": 3.977916463583412e-05,
|
| 1668 |
+
"loss": 0.5228,
|
| 1669 |
+
"step": 156
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 0.490625,
|
| 1673 |
+
"grad_norm": 2.8596651554107666,
|
| 1674 |
+
"learning_rate": 3.9763151385803936e-05,
|
| 1675 |
+
"loss": 0.5543,
|
| 1676 |
+
"step": 157
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.49375,
|
| 1680 |
+
"grad_norm": 3.0589263439178467,
|
| 1681 |
+
"learning_rate": 3.974658231146825e-05,
|
| 1682 |
+
"loss": 0.7127,
|
| 1683 |
+
"step": 158
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.496875,
|
| 1687 |
+
"grad_norm": 2.489602565765381,
|
| 1688 |
+
"learning_rate": 3.9729458036643335e-05,
|
| 1689 |
+
"loss": 0.4227,
|
| 1690 |
+
"step": 159
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 0.5,
|
| 1694 |
+
"grad_norm": 3.3471999168395996,
|
| 1695 |
+
"learning_rate": 3.971177920604846e-05,
|
| 1696 |
+
"loss": 0.5914,
|
| 1697 |
+
"step": 160
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 0.5,
|
| 1701 |
+
"eval_VitaminC_cosine_accuracy": 0.55859375,
|
| 1702 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.8433390855789185,
|
| 1703 |
+
"eval_VitaminC_cosine_ap": 0.5529005025024077,
|
| 1704 |
+
"eval_VitaminC_cosine_f1": 0.6657824933687002,
|
| 1705 |
+
"eval_VitaminC_cosine_f1_threshold": 0.3040446639060974,
|
| 1706 |
+
"eval_VitaminC_cosine_precision": 0.4990059642147117,
|
| 1707 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 1708 |
+
"eval_VitaminC_dot_accuracy": 0.55859375,
|
| 1709 |
+
"eval_VitaminC_dot_accuracy_threshold": 309.7912902832031,
|
| 1710 |
+
"eval_VitaminC_dot_ap": 0.5373200658982779,
|
| 1711 |
+
"eval_VitaminC_dot_f1": 0.6666666666666666,
|
| 1712 |
+
"eval_VitaminC_dot_f1_threshold": 122.78400421142578,
|
| 1713 |
+
"eval_VitaminC_dot_precision": 0.5,
|
| 1714 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 1715 |
+
"eval_VitaminC_euclidean_accuracy": 0.5546875,
|
| 1716 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 11.011507034301758,
|
| 1717 |
+
"eval_VitaminC_euclidean_ap": 0.5542686405562732,
|
| 1718 |
+
"eval_VitaminC_euclidean_f1": 0.6675531914893617,
|
| 1719 |
+
"eval_VitaminC_euclidean_f1_threshold": 22.90133285522461,
|
| 1720 |
+
"eval_VitaminC_euclidean_precision": 0.500998003992016,
|
| 1721 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 1722 |
+
"eval_VitaminC_manhattan_accuracy": 0.55859375,
|
| 1723 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 293.54693603515625,
|
| 1724 |
+
"eval_VitaminC_manhattan_ap": 0.5529507613553954,
|
| 1725 |
+
"eval_VitaminC_manhattan_f1": 0.6666666666666667,
|
| 1726 |
+
"eval_VitaminC_manhattan_f1_threshold": 479.09588623046875,
|
| 1727 |
+
"eval_VitaminC_manhattan_precision": 0.501002004008016,
|
| 1728 |
+
"eval_VitaminC_manhattan_recall": 0.9960159362549801,
|
| 1729 |
+
"eval_VitaminC_max_accuracy": 0.55859375,
|
| 1730 |
+
"eval_VitaminC_max_accuracy_threshold": 309.7912902832031,
|
| 1731 |
+
"eval_VitaminC_max_ap": 0.5542686405562732,
|
| 1732 |
+
"eval_VitaminC_max_f1": 0.6675531914893617,
|
| 1733 |
+
"eval_VitaminC_max_f1_threshold": 479.09588623046875,
|
| 1734 |
+
"eval_VitaminC_max_precision": 0.501002004008016,
|
| 1735 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 1736 |
+
"eval_sequential_score": 0.5542686405562732,
|
| 1737 |
+
"eval_sts-test_pearson_cosine": 0.8717931331186477,
|
| 1738 |
+
"eval_sts-test_pearson_dot": 0.8628985772297639,
|
| 1739 |
+
"eval_sts-test_pearson_euclidean": 0.8935960577585327,
|
| 1740 |
+
"eval_sts-test_pearson_manhattan": 0.8926162242871916,
|
| 1741 |
+
"eval_sts-test_pearson_max": 0.8935960577585327,
|
| 1742 |
+
"eval_sts-test_spearman_cosine": 0.8989036406477372,
|
| 1743 |
+
"eval_sts-test_spearman_dot": 0.8620115510306339,
|
| 1744 |
+
"eval_sts-test_spearman_euclidean": 0.8911198747488857,
|
| 1745 |
+
"eval_sts-test_spearman_manhattan": 0.8899440801070879,
|
| 1746 |
+
"eval_sts-test_spearman_max": 0.8989036406477372,
|
| 1747 |
+
"eval_vitaminc-pairs_loss": 2.0564281940460205,
|
| 1748 |
+
"eval_vitaminc-pairs_runtime": 1.8511,
|
| 1749 |
+
"eval_vitaminc-pairs_samples_per_second": 58.343,
|
| 1750 |
+
"eval_vitaminc-pairs_steps_per_second": 1.08,
|
| 1751 |
+
"step": 160
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 0.5,
|
| 1755 |
+
"eval_negation-triplets_loss": 0.7865684032440186,
|
| 1756 |
+
"eval_negation-triplets_runtime": 0.2987,
|
| 1757 |
+
"eval_negation-triplets_samples_per_second": 214.291,
|
| 1758 |
+
"eval_negation-triplets_steps_per_second": 3.348,
|
| 1759 |
+
"step": 160
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5,
|
| 1763 |
+
"eval_scitail-pairs-pos_loss": 0.09969007223844528,
|
| 1764 |
+
"eval_scitail-pairs-pos_runtime": 0.384,
|
| 1765 |
+
"eval_scitail-pairs-pos_samples_per_second": 140.615,
|
| 1766 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.604,
|
| 1767 |
+
"step": 160
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.5,
|
| 1771 |
+
"eval_xsum-pairs_loss": 0.08461853861808777,
|
| 1772 |
+
"eval_xsum-pairs_runtime": 2.8533,
|
| 1773 |
+
"eval_xsum-pairs_samples_per_second": 44.86,
|
| 1774 |
+
"eval_xsum-pairs_steps_per_second": 0.701,
|
| 1775 |
+
"step": 160
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 0.5,
|
| 1779 |
+
"eval_sciq_pairs_loss": 0.020078735426068306,
|
| 1780 |
+
"eval_sciq_pairs_runtime": 3.6458,
|
| 1781 |
+
"eval_sciq_pairs_samples_per_second": 35.109,
|
| 1782 |
+
"eval_sciq_pairs_steps_per_second": 0.549,
|
| 1783 |
+
"step": 160
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.5,
|
| 1787 |
+
"eval_qasc_pairs_loss": 0.12362705171108246,
|
| 1788 |
+
"eval_qasc_pairs_runtime": 0.6028,
|
| 1789 |
+
"eval_qasc_pairs_samples_per_second": 212.356,
|
| 1790 |
+
"eval_qasc_pairs_steps_per_second": 3.318,
|
| 1791 |
+
"step": 160
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 0.5,
|
| 1795 |
+
"eval_openbookqa_pairs_loss": 0.6668081283569336,
|
| 1796 |
+
"eval_openbookqa_pairs_runtime": 0.5784,
|
| 1797 |
+
"eval_openbookqa_pairs_samples_per_second": 221.308,
|
| 1798 |
+
"eval_openbookqa_pairs_steps_per_second": 3.458,
|
| 1799 |
+
"step": 160
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 0.5,
|
| 1803 |
+
"eval_msmarco_pairs_loss": 0.32913729548454285,
|
| 1804 |
+
"eval_msmarco_pairs_runtime": 1.4669,
|
| 1805 |
+
"eval_msmarco_pairs_samples_per_second": 87.26,
|
| 1806 |
+
"eval_msmarco_pairs_steps_per_second": 1.363,
|
| 1807 |
+
"step": 160
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 0.5,
|
| 1811 |
+
"eval_nq_pairs_loss": 0.2085198312997818,
|
| 1812 |
+
"eval_nq_pairs_runtime": 2.8644,
|
| 1813 |
+
"eval_nq_pairs_samples_per_second": 44.687,
|
| 1814 |
+
"eval_nq_pairs_steps_per_second": 0.698,
|
| 1815 |
+
"step": 160
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5,
|
| 1819 |
+
"eval_trivia_pairs_loss": 0.7138605117797852,
|
| 1820 |
+
"eval_trivia_pairs_runtime": 4.3915,
|
| 1821 |
+
"eval_trivia_pairs_samples_per_second": 29.147,
|
| 1822 |
+
"eval_trivia_pairs_steps_per_second": 0.455,
|
| 1823 |
+
"step": 160
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 0.5,
|
| 1827 |
+
"eval_gooaq_pairs_loss": 0.3919322192668915,
|
| 1828 |
+
"eval_gooaq_pairs_runtime": 1.004,
|
| 1829 |
+
"eval_gooaq_pairs_samples_per_second": 127.484,
|
| 1830 |
+
"eval_gooaq_pairs_steps_per_second": 1.992,
|
| 1831 |
+
"step": 160
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 0.5,
|
| 1835 |
+
"eval_paws-pos_loss": 0.025703923776745796,
|
| 1836 |
+
"eval_paws-pos_runtime": 0.6869,
|
| 1837 |
+
"eval_paws-pos_samples_per_second": 186.332,
|
| 1838 |
+
"eval_paws-pos_steps_per_second": 2.911,
|
| 1839 |
+
"step": 160
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 0.503125,
|
| 1843 |
+
"grad_norm": 2.7484354972839355,
|
| 1844 |
+
"learning_rate": 3.9693546485281616e-05,
|
| 1845 |
+
"loss": 0.3874,
|
| 1846 |
+
"step": 161
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 0.50625,
|
| 1850 |
+
"grad_norm": 3.9011173248291016,
|
| 1851 |
+
"learning_rate": 3.967476056079441e-05,
|
| 1852 |
+
"loss": 0.8134,
|
| 1853 |
+
"step": 162
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 0.509375,
|
| 1857 |
+
"grad_norm": 3.723893642425537,
|
| 1858 |
+
"learning_rate": 3.9655422139866315e-05,
|
| 1859 |
+
"loss": 0.5596,
|
| 1860 |
+
"step": 163
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 0.5125,
|
| 1864 |
+
"grad_norm": 1.8328720331192017,
|
| 1865 |
+
"learning_rate": 3.963553195057793e-05,
|
| 1866 |
+
"loss": 0.2877,
|
| 1867 |
+
"step": 164
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 0.515625,
|
| 1871 |
+
"grad_norm": 2.9615490436553955,
|
| 1872 |
+
"learning_rate": 3.9615090741783634e-05,
|
| 1873 |
+
"loss": 0.5218,
|
| 1874 |
+
"step": 165
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.51875,
|
| 1878 |
+
"grad_norm": 3.041154146194458,
|
| 1879 |
+
"learning_rate": 3.959409928308341e-05,
|
| 1880 |
+
"loss": 0.5282,
|
| 1881 |
+
"step": 166
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 0.521875,
|
| 1885 |
+
"grad_norm": 3.439157247543335,
|
| 1886 |
+
"learning_rate": 3.957255836479377e-05,
|
| 1887 |
+
"loss": 0.7528,
|
| 1888 |
+
"step": 167
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 0.525,
|
| 1892 |
+
"grad_norm": 3.576984405517578,
|
| 1893 |
+
"learning_rate": 3.955046879791816e-05,
|
| 1894 |
+
"loss": 0.7174,
|
| 1895 |
+
"step": 168
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 0.528125,
|
| 1899 |
+
"grad_norm": 3.1042630672454834,
|
| 1900 |
+
"learning_rate": 3.952783141411626e-05,
|
| 1901 |
+
"loss": 0.6902,
|
| 1902 |
+
"step": 169
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 0.53125,
|
| 1906 |
+
"grad_norm": 3.0211422443389893,
|
| 1907 |
+
"learning_rate": 3.9504647065672785e-05,
|
| 1908 |
+
"loss": 0.7486,
|
| 1909 |
+
"step": 170
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 0.534375,
|
| 1913 |
+
"grad_norm": 3.5162508487701416,
|
| 1914 |
+
"learning_rate": 3.9480916625465344e-05,
|
| 1915 |
+
"loss": 0.6333,
|
| 1916 |
+
"step": 171
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 0.5375,
|
| 1920 |
+
"grad_norm": 3.9070920944213867,
|
| 1921 |
+
"learning_rate": 3.9456640986931606e-05,
|
| 1922 |
+
"loss": 1.2932,
|
| 1923 |
+
"step": 172
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 0.540625,
|
| 1927 |
+
"grad_norm": 3.548743724822998,
|
| 1928 |
+
"learning_rate": 3.943182106403563e-05,
|
| 1929 |
+
"loss": 0.6259,
|
| 1930 |
+
"step": 173
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 0.54375,
|
| 1934 |
+
"grad_norm": 3.64949893951416,
|
| 1935 |
+
"learning_rate": 3.940645779123349e-05,
|
| 1936 |
+
"loss": 0.8357,
|
| 1937 |
+
"step": 174
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 0.546875,
|
| 1941 |
+
"grad_norm": 2.4284133911132812,
|
| 1942 |
+
"learning_rate": 3.938055212343807e-05,
|
| 1943 |
+
"loss": 0.3604,
|
| 1944 |
+
"step": 175
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 0.55,
|
| 1948 |
+
"grad_norm": 2.9141008853912354,
|
| 1949 |
+
"learning_rate": 3.9354105035983135e-05,
|
| 1950 |
+
"loss": 0.6598,
|
| 1951 |
+
"step": 176
|
| 1952 |
+
},
|
| 1953 |
+
{
|
| 1954 |
+
"epoch": 0.553125,
|
| 1955 |
+
"grad_norm": 2.0430235862731934,
|
| 1956 |
+
"learning_rate": 3.932711752458657e-05,
|
| 1957 |
+
"loss": 0.3169,
|
| 1958 |
+
"step": 177
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 0.55625,
|
| 1962 |
+
"grad_norm": 3.522728204727173,
|
| 1963 |
+
"learning_rate": 3.929959060531291e-05,
|
| 1964 |
+
"loss": 0.8629,
|
| 1965 |
+
"step": 178
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.559375,
|
| 1969 |
+
"grad_norm": 2.419400453567505,
|
| 1970 |
+
"learning_rate": 3.927152531453513e-05,
|
| 1971 |
+
"loss": 0.3648,
|
| 1972 |
+
"step": 179
|
| 1973 |
+
},
|
| 1974 |
+
{
|
| 1975 |
+
"epoch": 0.5625,
|
| 1976 |
+
"grad_norm": 2.826747417449951,
|
| 1977 |
+
"learning_rate": 3.924292270889555e-05,
|
| 1978 |
+
"loss": 0.5103,
|
| 1979 |
+
"step": 180
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 0.565625,
|
| 1983 |
+
"grad_norm": 3.2149524688720703,
|
| 1984 |
+
"learning_rate": 3.921378386526612e-05,
|
| 1985 |
+
"loss": 0.6255,
|
| 1986 |
+
"step": 181
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 0.56875,
|
| 1990 |
+
"grad_norm": 2.2112457752227783,
|
| 1991 |
+
"learning_rate": 3.918410988070782e-05,
|
| 1992 |
+
"loss": 0.4382,
|
| 1993 |
+
"step": 182
|
| 1994 |
+
},
|
| 1995 |
+
{
|
| 1996 |
+
"epoch": 0.571875,
|
| 1997 |
+
"grad_norm": 2.301940441131592,
|
| 1998 |
+
"learning_rate": 3.915390187242941e-05,
|
| 1999 |
+
"loss": 0.4647,
|
| 2000 |
+
"step": 183
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 0.575,
|
| 2004 |
+
"grad_norm": 2.272001266479492,
|
| 2005 |
+
"learning_rate": 3.912316097774532e-05,
|
| 2006 |
+
"loss": 0.4218,
|
| 2007 |
+
"step": 184
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 0.578125,
|
| 2011 |
+
"grad_norm": 3.77436900138855,
|
| 2012 |
+
"learning_rate": 3.909188835403285e-05,
|
| 2013 |
+
"loss": 0.8244,
|
| 2014 |
+
"step": 185
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 0.58125,
|
| 2018 |
+
"grad_norm": 3.236813545227051,
|
| 2019 |
+
"learning_rate": 3.906008517868863e-05,
|
| 2020 |
+
"loss": 0.6579,
|
| 2021 |
+
"step": 186
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 0.584375,
|
| 2025 |
+
"grad_norm": 3.1845405101776123,
|
| 2026 |
+
"learning_rate": 3.9027752649084215e-05,
|
| 2027 |
+
"loss": 0.8384,
|
| 2028 |
+
"step": 187
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 0.5875,
|
| 2032 |
+
"grad_norm": 2.709747791290283,
|
| 2033 |
+
"learning_rate": 3.899489198252108e-05,
|
| 2034 |
+
"loss": 0.5266,
|
| 2035 |
+
"step": 188
|
| 2036 |
+
},
|
| 2037 |
+
{
|
| 2038 |
+
"epoch": 0.590625,
|
| 2039 |
+
"grad_norm": 2.5210235118865967,
|
| 2040 |
+
"learning_rate": 3.896150441618476e-05,
|
| 2041 |
+
"loss": 0.5079,
|
| 2042 |
+
"step": 189
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 0.59375,
|
| 2046 |
+
"grad_norm": 1.9979658126831055,
|
| 2047 |
+
"learning_rate": 3.892759120709824e-05,
|
| 2048 |
+
"loss": 0.2574,
|
| 2049 |
+
"step": 190
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 0.596875,
|
| 2053 |
+
"grad_norm": 2.4257137775421143,
|
| 2054 |
+
"learning_rate": 3.8893153632074675e-05,
|
| 2055 |
+
"loss": 0.4162,
|
| 2056 |
+
"step": 191
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.6,
|
| 2060 |
+
"grad_norm": 3.482635021209717,
|
| 2061 |
+
"learning_rate": 3.88581929876693e-05,
|
| 2062 |
+
"loss": 0.7872,
|
| 2063 |
+
"step": 192
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 0.6,
|
| 2067 |
+
"eval_VitaminC_cosine_accuracy": 0.564453125,
|
| 2068 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.737064003944397,
|
| 2069 |
+
"eval_VitaminC_cosine_ap": 0.5553950127875514,
|
| 2070 |
+
"eval_VitaminC_cosine_f1": 0.6666666666666666,
|
| 2071 |
+
"eval_VitaminC_cosine_f1_threshold": 0.312030553817749,
|
| 2072 |
+
"eval_VitaminC_cosine_precision": 0.5,
|
| 2073 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 2074 |
+
"eval_VitaminC_dot_accuracy": 0.5625,
|
| 2075 |
+
"eval_VitaminC_dot_accuracy_threshold": 326.57232666015625,
|
| 2076 |
+
"eval_VitaminC_dot_ap": 0.5370581483003721,
|
| 2077 |
+
"eval_VitaminC_dot_f1": 0.6649006622516557,
|
| 2078 |
+
"eval_VitaminC_dot_f1_threshold": 116.00311279296875,
|
| 2079 |
+
"eval_VitaminC_dot_precision": 0.498015873015873,
|
| 2080 |
+
"eval_VitaminC_dot_recall": 1.0,
|
| 2081 |
+
"eval_VitaminC_euclidean_accuracy": 0.55859375,
|
| 2082 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 13.492112159729004,
|
| 2083 |
+
"eval_VitaminC_euclidean_ap": 0.5536857778177137,
|
| 2084 |
+
"eval_VitaminC_euclidean_f1": 0.6657824933687002,
|
| 2085 |
+
"eval_VitaminC_euclidean_f1_threshold": 23.840118408203125,
|
| 2086 |
+
"eval_VitaminC_euclidean_precision": 0.4990059642147117,
|
| 2087 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 2088 |
+
"eval_VitaminC_manhattan_accuracy": 0.5625,
|
| 2089 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 306.6820983886719,
|
| 2090 |
+
"eval_VitaminC_manhattan_ap": 0.5520101545849081,
|
| 2091 |
+
"eval_VitaminC_manhattan_f1": 0.6666666666666666,
|
| 2092 |
+
"eval_VitaminC_manhattan_f1_threshold": 490.146728515625,
|
| 2093 |
+
"eval_VitaminC_manhattan_precision": 0.5,
|
| 2094 |
+
"eval_VitaminC_manhattan_recall": 1.0,
|
| 2095 |
+
"eval_VitaminC_max_accuracy": 0.564453125,
|
| 2096 |
+
"eval_VitaminC_max_accuracy_threshold": 326.57232666015625,
|
| 2097 |
+
"eval_VitaminC_max_ap": 0.5553950127875514,
|
| 2098 |
+
"eval_VitaminC_max_f1": 0.6666666666666666,
|
| 2099 |
+
"eval_VitaminC_max_f1_threshold": 490.146728515625,
|
| 2100 |
+
"eval_VitaminC_max_precision": 0.5,
|
| 2101 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 2102 |
+
"eval_sequential_score": 0.5553950127875514,
|
| 2103 |
+
"eval_sts-test_pearson_cosine": 0.8705183135475563,
|
| 2104 |
+
"eval_sts-test_pearson_dot": 0.8575572680200927,
|
| 2105 |
+
"eval_sts-test_pearson_euclidean": 0.894961141451468,
|
| 2106 |
+
"eval_sts-test_pearson_manhattan": 0.8946364485546632,
|
| 2107 |
+
"eval_sts-test_pearson_max": 0.894961141451468,
|
| 2108 |
+
"eval_sts-test_spearman_cosine": 0.8981581293842179,
|
| 2109 |
+
"eval_sts-test_spearman_dot": 0.8574014998383989,
|
| 2110 |
+
"eval_sts-test_spearman_euclidean": 0.8924189591158167,
|
| 2111 |
+
"eval_sts-test_spearman_manhattan": 0.8920942887144219,
|
| 2112 |
+
"eval_sts-test_spearman_max": 0.8981581293842179,
|
| 2113 |
+
"eval_vitaminc-pairs_loss": 2.066204786300659,
|
| 2114 |
+
"eval_vitaminc-pairs_runtime": 1.8428,
|
| 2115 |
+
"eval_vitaminc-pairs_samples_per_second": 58.608,
|
| 2116 |
+
"eval_vitaminc-pairs_steps_per_second": 1.085,
|
| 2117 |
+
"step": 192
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.6,
|
| 2121 |
+
"eval_negation-triplets_loss": 0.763123095035553,
|
| 2122 |
+
"eval_negation-triplets_runtime": 0.297,
|
| 2123 |
+
"eval_negation-triplets_samples_per_second": 215.511,
|
| 2124 |
+
"eval_negation-triplets_steps_per_second": 3.367,
|
| 2125 |
+
"step": 192
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.6,
|
| 2129 |
+
"eval_scitail-pairs-pos_loss": 0.07364190369844437,
|
| 2130 |
+
"eval_scitail-pairs-pos_runtime": 0.3662,
|
| 2131 |
+
"eval_scitail-pairs-pos_samples_per_second": 147.451,
|
| 2132 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.731,
|
| 2133 |
+
"step": 192
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.6,
|
| 2137 |
+
"eval_xsum-pairs_loss": 0.06735075265169144,
|
| 2138 |
+
"eval_xsum-pairs_runtime": 2.8409,
|
| 2139 |
+
"eval_xsum-pairs_samples_per_second": 45.056,
|
| 2140 |
+
"eval_xsum-pairs_steps_per_second": 0.704,
|
| 2141 |
+
"step": 192
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 0.6,
|
| 2145 |
+
"eval_sciq_pairs_loss": 0.01930728368461132,
|
| 2146 |
+
"eval_sciq_pairs_runtime": 3.6003,
|
| 2147 |
+
"eval_sciq_pairs_samples_per_second": 35.552,
|
| 2148 |
+
"eval_sciq_pairs_steps_per_second": 0.556,
|
| 2149 |
+
"step": 192
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 0.6,
|
| 2153 |
+
"eval_qasc_pairs_loss": 0.11278136074542999,
|
| 2154 |
+
"eval_qasc_pairs_runtime": 0.5997,
|
| 2155 |
+
"eval_qasc_pairs_samples_per_second": 213.437,
|
| 2156 |
+
"eval_qasc_pairs_steps_per_second": 3.335,
|
| 2157 |
+
"step": 192
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 0.6,
|
| 2161 |
+
"eval_openbookqa_pairs_loss": 0.7505559921264648,
|
| 2162 |
+
"eval_openbookqa_pairs_runtime": 0.5774,
|
| 2163 |
+
"eval_openbookqa_pairs_samples_per_second": 221.691,
|
| 2164 |
+
"eval_openbookqa_pairs_steps_per_second": 3.464,
|
| 2165 |
+
"step": 192
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.6,
|
| 2169 |
+
"eval_msmarco_pairs_loss": 0.33166375756263733,
|
| 2170 |
+
"eval_msmarco_pairs_runtime": 1.4619,
|
| 2171 |
+
"eval_msmarco_pairs_samples_per_second": 87.558,
|
| 2172 |
+
"eval_msmarco_pairs_steps_per_second": 1.368,
|
| 2173 |
+
"step": 192
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 0.6,
|
| 2177 |
+
"eval_nq_pairs_loss": 0.21051406860351562,
|
| 2178 |
+
"eval_nq_pairs_runtime": 2.858,
|
| 2179 |
+
"eval_nq_pairs_samples_per_second": 44.786,
|
| 2180 |
+
"eval_nq_pairs_steps_per_second": 0.7,
|
| 2181 |
+
"step": 192
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.6,
|
| 2185 |
+
"eval_trivia_pairs_loss": 0.7072564363479614,
|
| 2186 |
+
"eval_trivia_pairs_runtime": 4.3854,
|
| 2187 |
+
"eval_trivia_pairs_samples_per_second": 29.187,
|
| 2188 |
+
"eval_trivia_pairs_steps_per_second": 0.456,
|
| 2189 |
+
"step": 192
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 0.6,
|
| 2193 |
+
"eval_gooaq_pairs_loss": 0.3748788833618164,
|
| 2194 |
+
"eval_gooaq_pairs_runtime": 1.0024,
|
| 2195 |
+
"eval_gooaq_pairs_samples_per_second": 127.692,
|
| 2196 |
+
"eval_gooaq_pairs_steps_per_second": 1.995,
|
| 2197 |
+
"step": 192
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 0.6,
|
| 2201 |
+
"eval_paws-pos_loss": 0.025185449048876762,
|
| 2202 |
+
"eval_paws-pos_runtime": 0.6844,
|
| 2203 |
+
"eval_paws-pos_samples_per_second": 187.016,
|
| 2204 |
+
"eval_paws-pos_steps_per_second": 2.922,
|
| 2205 |
+
"step": 192
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.603125,
|
| 2209 |
+
"grad_norm": 1.527544617652893,
|
| 2210 |
+
"learning_rate": 3.882271059013064e-05,
|
| 2211 |
+
"loss": 0.2606,
|
| 2212 |
+
"step": 193
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 0.60625,
|
| 2216 |
+
"grad_norm": 3.647446870803833,
|
| 2217 |
+
"learning_rate": 3.878670777535087e-05,
|
| 2218 |
+
"loss": 0.8808,
|
| 2219 |
+
"step": 194
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 0.609375,
|
| 2223 |
+
"grad_norm": 3.806488275527954,
|
| 2224 |
+
"learning_rate": 3.875018589881564e-05,
|
| 2225 |
+
"loss": 0.7685,
|
| 2226 |
+
"step": 195
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 0.6125,
|
| 2230 |
+
"grad_norm": 2.9896490573883057,
|
| 2231 |
+
"learning_rate": 3.871314633555296e-05,
|
| 2232 |
+
"loss": 0.7186,
|
| 2233 |
+
"step": 196
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 0.615625,
|
| 2237 |
+
"grad_norm": 1.31754732131958,
|
| 2238 |
+
"learning_rate": 3.8675590480081455e-05,
|
| 2239 |
+
"loss": 0.1147,
|
| 2240 |
+
"step": 197
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 0.61875,
|
| 2244 |
+
"grad_norm": 2.025834798812866,
|
| 2245 |
+
"learning_rate": 3.863751974635784e-05,
|
| 2246 |
+
"loss": 0.2816,
|
| 2247 |
+
"step": 198
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 0.621875,
|
| 2251 |
+
"grad_norm": 2.5674166679382324,
|
| 2252 |
+
"learning_rate": 3.8598935567723734e-05,
|
| 2253 |
+
"loss": 0.506,
|
| 2254 |
+
"step": 199
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 0.625,
|
| 2258 |
+
"grad_norm": 3.270737648010254,
|
| 2259 |
+
"learning_rate": 3.8559839396851656e-05,
|
| 2260 |
+
"loss": 0.5699,
|
| 2261 |
+
"step": 200
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 0.628125,
|
| 2265 |
+
"grad_norm": 1.6074001789093018,
|
| 2266 |
+
"learning_rate": 3.852023270569033e-05,
|
| 2267 |
+
"loss": 0.2746,
|
| 2268 |
+
"step": 201
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 0.63125,
|
| 2272 |
+
"grad_norm": 3.736549139022827,
|
| 2273 |
+
"learning_rate": 3.8480116985409306e-05,
|
| 2274 |
+
"loss": 0.7131,
|
| 2275 |
+
"step": 202
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 0.634375,
|
| 2279 |
+
"grad_norm": 3.9329938888549805,
|
| 2280 |
+
"learning_rate": 3.843949374634278e-05,
|
| 2281 |
+
"loss": 0.9307,
|
| 2282 |
+
"step": 203
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 0.6375,
|
| 2286 |
+
"grad_norm": 3.110591173171997,
|
| 2287 |
+
"learning_rate": 3.839836451793273e-05,
|
| 2288 |
+
"loss": 0.6033,
|
| 2289 |
+
"step": 204
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 0.640625,
|
| 2293 |
+
"grad_norm": 3.889007091522217,
|
| 2294 |
+
"learning_rate": 3.8356730848671374e-05,
|
| 2295 |
+
"loss": 0.7203,
|
| 2296 |
+
"step": 205
|
| 2297 |
+
},
|
| 2298 |
+
{
|
| 2299 |
+
"epoch": 0.64375,
|
| 2300 |
+
"grad_norm": 3.2738683223724365,
|
| 2301 |
+
"learning_rate": 3.8314594306042813e-05,
|
| 2302 |
+
"loss": 0.7422,
|
| 2303 |
+
"step": 206
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 0.646875,
|
| 2307 |
+
"grad_norm": 3.077531099319458,
|
| 2308 |
+
"learning_rate": 3.827195647646407e-05,
|
| 2309 |
+
"loss": 0.6955,
|
| 2310 |
+
"step": 207
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 0.65,
|
| 2314 |
+
"grad_norm": 3.336914539337158,
|
| 2315 |
+
"learning_rate": 3.822881896522533e-05,
|
| 2316 |
+
"loss": 0.7139,
|
| 2317 |
+
"step": 208
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 0.653125,
|
| 2321 |
+
"grad_norm": 2.866854429244995,
|
| 2322 |
+
"learning_rate": 3.818518339642951e-05,
|
| 2323 |
+
"loss": 0.4741,
|
| 2324 |
+
"step": 209
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 0.65625,
|
| 2328 |
+
"grad_norm": 1.8859411478042603,
|
| 2329 |
+
"learning_rate": 3.81410514129311e-05,
|
| 2330 |
+
"loss": 0.2658,
|
| 2331 |
+
"step": 210
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 0.659375,
|
| 2335 |
+
"grad_norm": 2.938387870788574,
|
| 2336 |
+
"learning_rate": 3.809642467627435e-05,
|
| 2337 |
+
"loss": 0.6033,
|
| 2338 |
+
"step": 211
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 0.6625,
|
| 2342 |
+
"grad_norm": 3.269779920578003,
|
| 2343 |
+
"learning_rate": 3.805130486663068e-05,
|
| 2344 |
+
"loss": 0.7776,
|
| 2345 |
+
"step": 212
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 0.665625,
|
| 2349 |
+
"grad_norm": 2.8948724269866943,
|
| 2350 |
+
"learning_rate": 3.800569368273539e-05,
|
| 2351 |
+
"loss": 0.6791,
|
| 2352 |
+
"step": 213
|
| 2353 |
+
},
|
| 2354 |
+
{
|
| 2355 |
+
"epoch": 0.66875,
|
| 2356 |
+
"grad_norm": 2.962749719619751,
|
| 2357 |
+
"learning_rate": 3.795959284182381e-05,
|
| 2358 |
+
"loss": 0.4367,
|
| 2359 |
+
"step": 214
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 0.671875,
|
| 2363 |
+
"grad_norm": 3.2313294410705566,
|
| 2364 |
+
"learning_rate": 3.791300407956651e-05,
|
| 2365 |
+
"loss": 0.7212,
|
| 2366 |
+
"step": 215
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 0.675,
|
| 2370 |
+
"grad_norm": 3.182274580001831,
|
| 2371 |
+
"learning_rate": 3.7865929150004086e-05,
|
| 2372 |
+
"loss": 0.7797,
|
| 2373 |
+
"step": 216
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 0.678125,
|
| 2377 |
+
"grad_norm": 2.7317817211151123,
|
| 2378 |
+
"learning_rate": 3.781836982548101e-05,
|
| 2379 |
+
"loss": 0.4547,
|
| 2380 |
+
"step": 217
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 0.68125,
|
| 2384 |
+
"grad_norm": 3.443126916885376,
|
| 2385 |
+
"learning_rate": 3.777032789657898e-05,
|
| 2386 |
+
"loss": 0.6771,
|
| 2387 |
+
"step": 218
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 0.684375,
|
| 2391 |
+
"grad_norm": 2.923877000808716,
|
| 2392 |
+
"learning_rate": 3.772180517204946e-05,
|
| 2393 |
+
"loss": 0.5488,
|
| 2394 |
+
"step": 219
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 0.6875,
|
| 2398 |
+
"grad_norm": 2.897601366043091,
|
| 2399 |
+
"learning_rate": 3.767280347874561e-05,
|
| 2400 |
+
"loss": 0.7352,
|
| 2401 |
+
"step": 220
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 0.690625,
|
| 2405 |
+
"grad_norm": 4.06088924407959,
|
| 2406 |
+
"learning_rate": 3.762332466155348e-05,
|
| 2407 |
+
"loss": 0.9567,
|
| 2408 |
+
"step": 221
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 0.69375,
|
| 2412 |
+
"grad_norm": 2.582475423812866,
|
| 2413 |
+
"learning_rate": 3.7573370583322575e-05,
|
| 2414 |
+
"loss": 0.4274,
|
| 2415 |
+
"step": 222
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 0.696875,
|
| 2419 |
+
"grad_norm": 3.7177348136901855,
|
| 2420 |
+
"learning_rate": 3.7522943124795706e-05,
|
| 2421 |
+
"loss": 0.7653,
|
| 2422 |
+
"step": 223
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 0.7,
|
| 2426 |
+
"grad_norm": 2.9613823890686035,
|
| 2427 |
+
"learning_rate": 3.7472044184538186e-05,
|
| 2428 |
+
"loss": 0.5672,
|
| 2429 |
+
"step": 224
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 0.7,
|
| 2433 |
+
"eval_VitaminC_cosine_accuracy": 0.55859375,
|
| 2434 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.8318229913711548,
|
| 2435 |
+
"eval_VitaminC_cosine_ap": 0.5483869647391425,
|
| 2436 |
+
"eval_VitaminC_cosine_f1": 0.6657824933687002,
|
| 2437 |
+
"eval_VitaminC_cosine_f1_threshold": 0.2898828089237213,
|
| 2438 |
+
"eval_VitaminC_cosine_precision": 0.4990059642147117,
|
| 2439 |
+
"eval_VitaminC_cosine_recall": 1.0,
|
| 2440 |
+
"eval_VitaminC_dot_accuracy": 0.5546875,
|
| 2441 |
+
"eval_VitaminC_dot_accuracy_threshold": 280.6613464355469,
|
| 2442 |
+
"eval_VitaminC_dot_ap": 0.5352389087249884,
|
| 2443 |
+
"eval_VitaminC_dot_f1": 0.6666666666666667,
|
| 2444 |
+
"eval_VitaminC_dot_f1_threshold": 127.79656982421875,
|
| 2445 |
+
"eval_VitaminC_dot_precision": 0.501002004008016,
|
| 2446 |
+
"eval_VitaminC_dot_recall": 0.9960159362549801,
|
| 2447 |
+
"eval_VitaminC_euclidean_accuracy": 0.556640625,
|
| 2448 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 15.362771987915039,
|
| 2449 |
+
"eval_VitaminC_euclidean_ap": 0.5487471191186046,
|
| 2450 |
+
"eval_VitaminC_euclidean_f1": 0.6657789613848203,
|
| 2451 |
+
"eval_VitaminC_euclidean_f1_threshold": 23.0285587310791,
|
| 2452 |
+
"eval_VitaminC_euclidean_precision": 0.5,
|
| 2453 |
+
"eval_VitaminC_euclidean_recall": 0.9960159362549801,
|
| 2454 |
+
"eval_VitaminC_manhattan_accuracy": 0.556640625,
|
| 2455 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 304.9786376953125,
|
| 2456 |
+
"eval_VitaminC_manhattan_ap": 0.5448852224007886,
|
| 2457 |
+
"eval_VitaminC_manhattan_f1": 0.6657824933687002,
|
| 2458 |
+
"eval_VitaminC_manhattan_f1_threshold": 503.7974548339844,
|
| 2459 |
+
"eval_VitaminC_manhattan_precision": 0.4990059642147117,
|
| 2460 |
+
"eval_VitaminC_manhattan_recall": 1.0,
|
| 2461 |
+
"eval_VitaminC_max_accuracy": 0.55859375,
|
| 2462 |
+
"eval_VitaminC_max_accuracy_threshold": 304.9786376953125,
|
| 2463 |
+
"eval_VitaminC_max_ap": 0.5487471191186046,
|
| 2464 |
+
"eval_VitaminC_max_f1": 0.6666666666666667,
|
| 2465 |
+
"eval_VitaminC_max_f1_threshold": 503.7974548339844,
|
| 2466 |
+
"eval_VitaminC_max_precision": 0.501002004008016,
|
| 2467 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 2468 |
+
"eval_sequential_score": 0.5487471191186046,
|
| 2469 |
+
"eval_sts-test_pearson_cosine": 0.8727242216490746,
|
| 2470 |
+
"eval_sts-test_pearson_dot": 0.8620679649117718,
|
| 2471 |
+
"eval_sts-test_pearson_euclidean": 0.8961291746213003,
|
| 2472 |
+
"eval_sts-test_pearson_manhattan": 0.8961616445842001,
|
| 2473 |
+
"eval_sts-test_pearson_max": 0.8961616445842001,
|
| 2474 |
+
"eval_sts-test_spearman_cosine": 0.9004602237727143,
|
| 2475 |
+
"eval_sts-test_spearman_dot": 0.8617584826474656,
|
| 2476 |
+
"eval_sts-test_spearman_euclidean": 0.8945701970021624,
|
| 2477 |
+
"eval_sts-test_spearman_manhattan": 0.8942019836234342,
|
| 2478 |
+
"eval_sts-test_spearman_max": 0.9004602237727143,
|
| 2479 |
+
"eval_vitaminc-pairs_loss": 2.07513689994812,
|
| 2480 |
+
"eval_vitaminc-pairs_runtime": 1.8489,
|
| 2481 |
+
"eval_vitaminc-pairs_samples_per_second": 58.414,
|
| 2482 |
+
"eval_vitaminc-pairs_steps_per_second": 1.082,
|
| 2483 |
+
"step": 224
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 0.7,
|
| 2487 |
+
"eval_negation-triplets_loss": 0.7822766900062561,
|
| 2488 |
+
"eval_negation-triplets_runtime": 0.2948,
|
| 2489 |
+
"eval_negation-triplets_samples_per_second": 217.093,
|
| 2490 |
+
"eval_negation-triplets_steps_per_second": 3.392,
|
| 2491 |
+
"step": 224
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 0.7,
|
| 2495 |
+
"eval_scitail-pairs-pos_loss": 0.084584079682827,
|
| 2496 |
+
"eval_scitail-pairs-pos_runtime": 0.365,
|
| 2497 |
+
"eval_scitail-pairs-pos_samples_per_second": 147.944,
|
| 2498 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.74,
|
| 2499 |
+
"step": 224
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 0.7,
|
| 2503 |
+
"eval_xsum-pairs_loss": 0.05927089601755142,
|
| 2504 |
+
"eval_xsum-pairs_runtime": 2.8461,
|
| 2505 |
+
"eval_xsum-pairs_samples_per_second": 44.974,
|
| 2506 |
+
"eval_xsum-pairs_steps_per_second": 0.703,
|
| 2507 |
+
"step": 224
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 0.7,
|
| 2511 |
+
"eval_sciq_pairs_loss": 0.019030971452593803,
|
| 2512 |
+
"eval_sciq_pairs_runtime": 3.6465,
|
| 2513 |
+
"eval_sciq_pairs_samples_per_second": 35.102,
|
| 2514 |
+
"eval_sciq_pairs_steps_per_second": 0.548,
|
| 2515 |
+
"step": 224
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.7,
|
| 2519 |
+
"eval_qasc_pairs_loss": 0.12519867718219757,
|
| 2520 |
+
"eval_qasc_pairs_runtime": 0.6003,
|
| 2521 |
+
"eval_qasc_pairs_samples_per_second": 213.235,
|
| 2522 |
+
"eval_qasc_pairs_steps_per_second": 3.332,
|
| 2523 |
+
"step": 224
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.7,
|
| 2527 |
+
"eval_openbookqa_pairs_loss": 0.7141773700714111,
|
| 2528 |
+
"eval_openbookqa_pairs_runtime": 0.5753,
|
| 2529 |
+
"eval_openbookqa_pairs_samples_per_second": 222.508,
|
| 2530 |
+
"eval_openbookqa_pairs_steps_per_second": 3.477,
|
| 2531 |
+
"step": 224
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.7,
|
| 2535 |
+
"eval_msmarco_pairs_loss": 0.3040487468242645,
|
| 2536 |
+
"eval_msmarco_pairs_runtime": 1.4648,
|
| 2537 |
+
"eval_msmarco_pairs_samples_per_second": 87.383,
|
| 2538 |
+
"eval_msmarco_pairs_steps_per_second": 1.365,
|
| 2539 |
+
"step": 224
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 0.7,
|
| 2543 |
+
"eval_nq_pairs_loss": 0.1808711141347885,
|
| 2544 |
+
"eval_nq_pairs_runtime": 2.8595,
|
| 2545 |
+
"eval_nq_pairs_samples_per_second": 44.764,
|
| 2546 |
+
"eval_nq_pairs_steps_per_second": 0.699,
|
| 2547 |
+
"step": 224
|
| 2548 |
+
},
|
| 2549 |
+
{
|
| 2550 |
+
"epoch": 0.7,
|
| 2551 |
+
"eval_trivia_pairs_loss": 0.7160522937774658,
|
| 2552 |
+
"eval_trivia_pairs_runtime": 4.3875,
|
| 2553 |
+
"eval_trivia_pairs_samples_per_second": 29.174,
|
| 2554 |
+
"eval_trivia_pairs_steps_per_second": 0.456,
|
| 2555 |
+
"step": 224
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 0.7,
|
| 2559 |
+
"eval_gooaq_pairs_loss": 0.3398577868938446,
|
| 2560 |
+
"eval_gooaq_pairs_runtime": 1.0189,
|
| 2561 |
+
"eval_gooaq_pairs_samples_per_second": 125.631,
|
| 2562 |
+
"eval_gooaq_pairs_steps_per_second": 1.963,
|
| 2563 |
+
"step": 224
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 0.7,
|
| 2567 |
+
"eval_paws-pos_loss": 0.0250654686242342,
|
| 2568 |
+
"eval_paws-pos_runtime": 0.6965,
|
| 2569 |
+
"eval_paws-pos_samples_per_second": 183.765,
|
| 2570 |
+
"eval_paws-pos_steps_per_second": 2.871,
|
| 2571 |
+
"step": 224
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.703125,
|
| 2575 |
+
"grad_norm": 2.7675271034240723,
|
| 2576 |
+
"learning_rate": 3.742067567886634e-05,
|
| 2577 |
+
"loss": 0.6116,
|
| 2578 |
+
"step": 225
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.70625,
|
| 2582 |
+
"grad_norm": 3.1136417388916016,
|
| 2583 |
+
"learning_rate": 3.7368839541775386e-05,
|
| 2584 |
+
"loss": 0.6484,
|
| 2585 |
+
"step": 226
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.709375,
|
| 2589 |
+
"grad_norm": 3.1425583362579346,
|
| 2590 |
+
"learning_rate": 3.731653772486657e-05,
|
| 2591 |
+
"loss": 0.669,
|
| 2592 |
+
"step": 227
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7125,
|
| 2596 |
+
"grad_norm": 1.8860105276107788,
|
| 2597 |
+
"learning_rate": 3.726377219727376e-05,
|
| 2598 |
+
"loss": 0.263,
|
| 2599 |
+
"step": 228
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.715625,
|
| 2603 |
+
"grad_norm": 2.6990439891815186,
|
| 2604 |
+
"learning_rate": 3.721054494558923e-05,
|
| 2605 |
+
"loss": 0.6181,
|
| 2606 |
+
"step": 229
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.71875,
|
| 2610 |
+
"grad_norm": 3.836609363555908,
|
| 2611 |
+
"learning_rate": 3.7156857973788926e-05,
|
| 2612 |
+
"loss": 0.8956,
|
| 2613 |
+
"step": 230
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.721875,
|
| 2617 |
+
"grad_norm": 3.0837268829345703,
|
| 2618 |
+
"learning_rate": 3.710271330315699e-05,
|
| 2619 |
+
"loss": 0.5363,
|
| 2620 |
+
"step": 231
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.725,
|
| 2624 |
+
"grad_norm": 3.639112710952759,
|
| 2625 |
+
"learning_rate": 3.704811297220967e-05,
|
| 2626 |
+
"loss": 0.823,
|
| 2627 |
+
"step": 232
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.728125,
|
| 2631 |
+
"grad_norm": 3.301112651824951,
|
| 2632 |
+
"learning_rate": 3.699305903661858e-05,
|
| 2633 |
+
"loss": 0.7795,
|
| 2634 |
+
"step": 233
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.73125,
|
| 2638 |
+
"grad_norm": 2.289018154144287,
|
| 2639 |
+
"learning_rate": 3.693755356913326e-05,
|
| 2640 |
+
"loss": 0.3688,
|
| 2641 |
+
"step": 234
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.734375,
|
| 2645 |
+
"grad_norm": 2.259490966796875,
|
| 2646 |
+
"learning_rate": 3.688159865950319e-05,
|
| 2647 |
+
"loss": 0.3835,
|
| 2648 |
+
"step": 235
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7375,
|
| 2652 |
+
"grad_norm": 2.2043821811676025,
|
| 2653 |
+
"learning_rate": 3.6825196414399096e-05,
|
| 2654 |
+
"loss": 0.3393,
|
| 2655 |
+
"step": 236
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.740625,
|
| 2659 |
+
"grad_norm": 2.6866259574890137,
|
| 2660 |
+
"learning_rate": 3.6768348957333635e-05,
|
| 2661 |
+
"loss": 0.4792,
|
| 2662 |
+
"step": 237
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.74375,
|
| 2666 |
+
"grad_norm": 2.561917304992676,
|
| 2667 |
+
"learning_rate": 3.671105842858142e-05,
|
| 2668 |
+
"loss": 0.3966,
|
| 2669 |
+
"step": 238
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.746875,
|
| 2673 |
+
"grad_norm": 2.1512343883514404,
|
| 2674 |
+
"learning_rate": 3.6653326985098486e-05,
|
| 2675 |
+
"loss": 0.2902,
|
| 2676 |
+
"step": 239
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.75,
|
| 2680 |
+
"grad_norm": 3.7423007488250732,
|
| 2681 |
+
"learning_rate": 3.659515680044106e-05,
|
| 2682 |
+
"loss": 0.6716,
|
| 2683 |
+
"step": 240
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.753125,
|
| 2687 |
+
"grad_norm": 2.6502630710601807,
|
| 2688 |
+
"learning_rate": 3.65365500646837e-05,
|
| 2689 |
+
"loss": 0.6783,
|
| 2690 |
+
"step": 241
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.75625,
|
| 2694 |
+
"grad_norm": 2.8291828632354736,
|
| 2695 |
+
"learning_rate": 3.6477508984336886e-05,
|
| 2696 |
+
"loss": 0.4794,
|
| 2697 |
+
"step": 242
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.759375,
|
| 2701 |
+
"grad_norm": 3.7910561561584473,
|
| 2702 |
+
"learning_rate": 3.641803578226393e-05,
|
| 2703 |
+
"loss": 0.8283,
|
| 2704 |
+
"step": 243
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7625,
|
| 2708 |
+
"grad_norm": 3.3968613147735596,
|
| 2709 |
+
"learning_rate": 3.635813269759727e-05,
|
| 2710 |
+
"loss": 0.6875,
|
| 2711 |
+
"step": 244
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.765625,
|
| 2715 |
+
"grad_norm": 3.5861093997955322,
|
| 2716 |
+
"learning_rate": 3.629780198565419e-05,
|
| 2717 |
+
"loss": 0.8384,
|
| 2718 |
+
"step": 245
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.76875,
|
| 2722 |
+
"grad_norm": 2.709362030029297,
|
| 2723 |
+
"learning_rate": 3.623704591785189e-05,
|
| 2724 |
+
"loss": 0.5796,
|
| 2725 |
+
"step": 246
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.771875,
|
| 2729 |
+
"grad_norm": 2.5690431594848633,
|
| 2730 |
+
"learning_rate": 3.6175866781622e-05,
|
| 2731 |
+
"loss": 0.6206,
|
| 2732 |
+
"step": 247
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.775,
|
| 2736 |
+
"grad_norm": 3.5460782051086426,
|
| 2737 |
+
"learning_rate": 3.611426688032439e-05,
|
| 2738 |
+
"loss": 0.7836,
|
| 2739 |
+
"step": 248
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.778125,
|
| 2743 |
+
"grad_norm": 2.9132962226867676,
|
| 2744 |
+
"learning_rate": 3.605224853316055e-05,
|
| 2745 |
+
"loss": 0.615,
|
| 2746 |
+
"step": 249
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.78125,
|
| 2750 |
+
"grad_norm": 2.707908868789673,
|
| 2751 |
+
"learning_rate": 3.5989814075086195e-05,
|
| 2752 |
+
"loss": 0.433,
|
| 2753 |
+
"step": 250
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.784375,
|
| 2757 |
+
"grad_norm": 3.2124290466308594,
|
| 2758 |
+
"learning_rate": 3.592696585672338e-05,
|
| 2759 |
+
"loss": 0.7394,
|
| 2760 |
+
"step": 251
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.7875,
|
| 2764 |
+
"grad_norm": 1.3290472030639648,
|
| 2765 |
+
"learning_rate": 3.5863706244272006e-05,
|
| 2766 |
+
"loss": 0.1203,
|
| 2767 |
+
"step": 252
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.790625,
|
| 2771 |
+
"grad_norm": 3.5975258350372314,
|
| 2772 |
+
"learning_rate": 3.580003761942073e-05,
|
| 2773 |
+
"loss": 1.0909,
|
| 2774 |
+
"step": 253
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.79375,
|
| 2778 |
+
"grad_norm": 3.1402907371520996,
|
| 2779 |
+
"learning_rate": 3.573596237925728e-05,
|
| 2780 |
+
"loss": 0.7107,
|
| 2781 |
+
"step": 254
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.796875,
|
| 2785 |
+
"grad_norm": 2.1686770915985107,
|
| 2786 |
+
"learning_rate": 3.567148293617825e-05,
|
| 2787 |
+
"loss": 0.3464,
|
| 2788 |
+
"step": 255
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8,
|
| 2792 |
+
"grad_norm": 4.112154960632324,
|
| 2793 |
+
"learning_rate": 3.560660171779821e-05,
|
| 2794 |
+
"loss": 0.9347,
|
| 2795 |
+
"step": 256
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.8,
|
| 2799 |
+
"eval_VitaminC_cosine_accuracy": 0.560546875,
|
| 2800 |
+
"eval_VitaminC_cosine_accuracy_threshold": 0.825050950050354,
|
| 2801 |
+
"eval_VitaminC_cosine_ap": 0.5480281823929228,
|
| 2802 |
+
"eval_VitaminC_cosine_f1": 0.6666666666666667,
|
| 2803 |
+
"eval_VitaminC_cosine_f1_threshold": 0.39935123920440674,
|
| 2804 |
+
"eval_VitaminC_cosine_precision": 0.501002004008016,
|
| 2805 |
+
"eval_VitaminC_cosine_recall": 0.9960159362549801,
|
| 2806 |
+
"eval_VitaminC_dot_accuracy": 0.552734375,
|
| 2807 |
+
"eval_VitaminC_dot_accuracy_threshold": 313.59075927734375,
|
| 2808 |
+
"eval_VitaminC_dot_ap": 0.5329984665726657,
|
| 2809 |
+
"eval_VitaminC_dot_f1": 0.6657789613848203,
|
| 2810 |
+
"eval_VitaminC_dot_f1_threshold": 132.71243286132812,
|
| 2811 |
+
"eval_VitaminC_dot_precision": 0.5,
|
| 2812 |
+
"eval_VitaminC_dot_recall": 0.9960159362549801,
|
| 2813 |
+
"eval_VitaminC_euclidean_accuracy": 0.556640625,
|
| 2814 |
+
"eval_VitaminC_euclidean_accuracy_threshold": 11.227453231811523,
|
| 2815 |
+
"eval_VitaminC_euclidean_ap": 0.5496569156706412,
|
| 2816 |
+
"eval_VitaminC_euclidean_f1": 0.6666666666666666,
|
| 2817 |
+
"eval_VitaminC_euclidean_f1_threshold": 22.6641788482666,
|
| 2818 |
+
"eval_VitaminC_euclidean_precision": 0.5,
|
| 2819 |
+
"eval_VitaminC_euclidean_recall": 1.0,
|
| 2820 |
+
"eval_VitaminC_manhattan_accuracy": 0.5546875,
|
| 2821 |
+
"eval_VitaminC_manhattan_accuracy_threshold": 230.89329528808594,
|
| 2822 |
+
"eval_VitaminC_manhattan_ap": 0.545699310794812,
|
| 2823 |
+
"eval_VitaminC_manhattan_f1": 0.6657824933687002,
|
| 2824 |
+
"eval_VitaminC_manhattan_f1_threshold": 483.625244140625,
|
| 2825 |
+
"eval_VitaminC_manhattan_precision": 0.4990059642147117,
|
| 2826 |
+
"eval_VitaminC_manhattan_recall": 1.0,
|
| 2827 |
+
"eval_VitaminC_max_accuracy": 0.560546875,
|
| 2828 |
+
"eval_VitaminC_max_accuracy_threshold": 313.59075927734375,
|
| 2829 |
+
"eval_VitaminC_max_ap": 0.5496569156706412,
|
| 2830 |
+
"eval_VitaminC_max_f1": 0.6666666666666667,
|
| 2831 |
+
"eval_VitaminC_max_f1_threshold": 483.625244140625,
|
| 2832 |
+
"eval_VitaminC_max_precision": 0.501002004008016,
|
| 2833 |
+
"eval_VitaminC_max_recall": 1.0,
|
| 2834 |
+
"eval_sequential_score": 0.5496569156706412,
|
| 2835 |
+
"eval_sts-test_pearson_cosine": 0.8711436629553765,
|
| 2836 |
+
"eval_sts-test_pearson_dot": 0.859333025320038,
|
| 2837 |
+
"eval_sts-test_pearson_euclidean": 0.8967955144362856,
|
| 2838 |
+
"eval_sts-test_pearson_manhattan": 0.8972988934332646,
|
| 2839 |
+
"eval_sts-test_pearson_max": 0.8972988934332646,
|
| 2840 |
+
"eval_sts-test_spearman_cosine": 0.8987000224084064,
|
| 2841 |
+
"eval_sts-test_spearman_dot": 0.8599958647150425,
|
| 2842 |
+
"eval_sts-test_spearman_euclidean": 0.8935259263175941,
|
| 2843 |
+
"eval_sts-test_spearman_manhattan": 0.8939636384052635,
|
| 2844 |
+
"eval_sts-test_spearman_max": 0.8987000224084064,
|
| 2845 |
+
"eval_vitaminc-pairs_loss": 2.0033843517303467,
|
| 2846 |
+
"eval_vitaminc-pairs_runtime": 1.8717,
|
| 2847 |
+
"eval_vitaminc-pairs_samples_per_second": 57.701,
|
| 2848 |
+
"eval_vitaminc-pairs_steps_per_second": 1.069,
|
| 2849 |
+
"step": 256
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 0.8,
|
| 2853 |
+
"eval_negation-triplets_loss": 0.7500894069671631,
|
| 2854 |
+
"eval_negation-triplets_runtime": 0.2995,
|
| 2855 |
+
"eval_negation-triplets_samples_per_second": 213.659,
|
| 2856 |
+
"eval_negation-triplets_steps_per_second": 3.338,
|
| 2857 |
+
"step": 256
|
| 2858 |
+
},
|
| 2859 |
+
{
|
| 2860 |
+
"epoch": 0.8,
|
| 2861 |
+
"eval_scitail-pairs-pos_loss": 0.07255758345127106,
|
| 2862 |
+
"eval_scitail-pairs-pos_runtime": 0.386,
|
| 2863 |
+
"eval_scitail-pairs-pos_samples_per_second": 139.883,
|
| 2864 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.59,
|
| 2865 |
+
"step": 256
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.8,
|
| 2869 |
+
"eval_xsum-pairs_loss": 0.056476954370737076,
|
| 2870 |
+
"eval_xsum-pairs_runtime": 2.8548,
|
| 2871 |
+
"eval_xsum-pairs_samples_per_second": 44.837,
|
| 2872 |
+
"eval_xsum-pairs_steps_per_second": 0.701,
|
| 2873 |
+
"step": 256
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.8,
|
| 2877 |
+
"eval_sciq_pairs_loss": 0.01967025361955166,
|
| 2878 |
+
"eval_sciq_pairs_runtime": 3.7336,
|
| 2879 |
+
"eval_sciq_pairs_samples_per_second": 34.283,
|
| 2880 |
+
"eval_sciq_pairs_steps_per_second": 0.536,
|
| 2881 |
+
"step": 256
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 0.8,
|
| 2885 |
+
"eval_qasc_pairs_loss": 0.1263607293367386,
|
| 2886 |
+
"eval_qasc_pairs_runtime": 0.6107,
|
| 2887 |
+
"eval_qasc_pairs_samples_per_second": 209.594,
|
| 2888 |
+
"eval_qasc_pairs_steps_per_second": 3.275,
|
| 2889 |
+
"step": 256
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.8,
|
| 2893 |
+
"eval_openbookqa_pairs_loss": 0.7773354649543762,
|
| 2894 |
+
"eval_openbookqa_pairs_runtime": 0.5903,
|
| 2895 |
+
"eval_openbookqa_pairs_samples_per_second": 216.831,
|
| 2896 |
+
"eval_openbookqa_pairs_steps_per_second": 3.388,
|
| 2897 |
+
"step": 256
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 0.8,
|
| 2901 |
+
"eval_msmarco_pairs_loss": 0.2844376862049103,
|
| 2902 |
+
"eval_msmarco_pairs_runtime": 1.4722,
|
| 2903 |
+
"eval_msmarco_pairs_samples_per_second": 86.947,
|
| 2904 |
+
"eval_msmarco_pairs_steps_per_second": 1.359,
|
| 2905 |
+
"step": 256
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 0.8,
|
| 2909 |
+
"eval_nq_pairs_loss": 0.17289823293685913,
|
| 2910 |
+
"eval_nq_pairs_runtime": 2.8665,
|
| 2911 |
+
"eval_nq_pairs_samples_per_second": 44.654,
|
| 2912 |
+
"eval_nq_pairs_steps_per_second": 0.698,
|
| 2913 |
+
"step": 256
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 0.8,
|
| 2917 |
+
"eval_trivia_pairs_loss": 0.6546728610992432,
|
| 2918 |
+
"eval_trivia_pairs_runtime": 4.3994,
|
| 2919 |
+
"eval_trivia_pairs_samples_per_second": 29.095,
|
| 2920 |
+
"eval_trivia_pairs_steps_per_second": 0.455,
|
| 2921 |
+
"step": 256
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.8,
|
| 2925 |
+
"eval_gooaq_pairs_loss": 0.31546029448509216,
|
| 2926 |
+
"eval_gooaq_pairs_runtime": 1.0423,
|
| 2927 |
+
"eval_gooaq_pairs_samples_per_second": 122.802,
|
| 2928 |
+
"eval_gooaq_pairs_steps_per_second": 1.919,
|
| 2929 |
+
"step": 256
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 0.8,
|
| 2933 |
+
"eval_paws-pos_loss": 0.02565235085785389,
|
| 2934 |
+
"eval_paws-pos_runtime": 0.6999,
|
| 2935 |
+
"eval_paws-pos_samples_per_second": 182.88,
|
| 2936 |
+
"eval_paws-pos_steps_per_second": 2.857,
|
| 2937 |
+
"step": 256
|
| 2938 |
+
}
|
| 2939 |
+
],
|
| 2940 |
+
"logging_steps": 1,
|
| 2941 |
+
"max_steps": 640,
|
| 2942 |
+
"num_input_tokens_seen": 0,
|
| 2943 |
+
"num_train_epochs": 2,
|
| 2944 |
+
"save_steps": 64,
|
| 2945 |
+
"stateful_callbacks": {
|
| 2946 |
+
"TrainerControl": {
|
| 2947 |
+
"args": {
|
| 2948 |
+
"should_epoch_stop": false,
|
| 2949 |
+
"should_evaluate": false,
|
| 2950 |
+
"should_log": false,
|
| 2951 |
+
"should_save": true,
|
| 2952 |
+
"should_training_stop": false
|
| 2953 |
+
},
|
| 2954 |
+
"attributes": {}
|
| 2955 |
+
}
|
| 2956 |
+
},
|
| 2957 |
+
"total_flos": 0.0,
|
| 2958 |
+
"train_batch_size": 320,
|
| 2959 |
+
"trial_name": null,
|
| 2960 |
+
"trial_params": null
|
| 2961 |
+
}
|
checkpoint-256/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ffe6f56833ae1db889da47aa19784a25c77bb02e655bcfee55cb3e0df77d35b1
|
| 3 |
+
size 5688
|