Training in progress, step 11748, checkpoint
Browse files- checkpoint-11748/1_Pooling/config.json +10 -0
- checkpoint-11748/README.md +0 -0
- checkpoint-11748/added_tokens.json +3 -0
- checkpoint-11748/config.json +35 -0
- checkpoint-11748/config_sentence_transformers.json +10 -0
- checkpoint-11748/modules.json +14 -0
- checkpoint-11748/optimizer.pt +3 -0
- checkpoint-11748/pytorch_model.bin +3 -0
- checkpoint-11748/rng_state.pth +3 -0
- checkpoint-11748/scheduler.pt +3 -0
- checkpoint-11748/sentence_bert_config.json +4 -0
- checkpoint-11748/special_tokens_map.json +15 -0
- checkpoint-11748/spm.model +3 -0
- checkpoint-11748/tokenizer.json +0 -0
- checkpoint-11748/tokenizer_config.json +58 -0
- checkpoint-11748/trainer_state.json +2536 -0
- checkpoint-11748/training_args.bin +3 -0
checkpoint-11748/1_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-11748/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-11748/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
checkpoint-11748/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/deberta-v3-small",
|
| 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": "4.41.2",
|
| 33 |
+
"type_vocab_size": 0,
|
| 34 |
+
"vocab_size": 128100
|
| 35 |
+
}
|
checkpoint-11748/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.41.2",
|
| 5 |
+
"pytorch": "2.1.2"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-11748/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-11748/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47c4bbc8e1e46fe8ab5da59f14d367fbd679602e0a1943a1d9dc5109b9bf8155
|
| 3 |
+
size 1130520122
|
checkpoint-11748/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b25d31c3114f97e26da394c5e37d75ee6601df1297fcd4137293447089a0c5a
|
| 3 |
+
size 565251810
|
checkpoint-11748/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4512a4d3613f6b7d86b146939f89e3c62b0366edbc96849275f6672de9354f51
|
| 3 |
+
size 14244
|
checkpoint-11748/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80168c8f4520d6288706a30964c96e9c61e9d14287267f1dc90b6c495f5d6935
|
| 3 |
+
size 1064
|
checkpoint-11748/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-11748/special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "[CLS]",
|
| 3 |
+
"cls_token": "[CLS]",
|
| 4 |
+
"eos_token": "[SEP]",
|
| 5 |
+
"mask_token": "[MASK]",
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"unk_token": {
|
| 9 |
+
"content": "[UNK]",
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"normalized": true,
|
| 12 |
+
"rstrip": false,
|
| 13 |
+
"single_word": false
|
| 14 |
+
}
|
| 15 |
+
}
|
checkpoint-11748/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
checkpoint-11748/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-11748/tokenizer_config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"sp_model_kwargs": {},
|
| 54 |
+
"split_by_punct": false,
|
| 55 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 56 |
+
"unk_token": "[UNK]",
|
| 57 |
+
"vocab_type": "spm"
|
| 58 |
+
}
|
checkpoint-11748/trainer_state.json
ADDED
|
@@ -0,0 +1,2536 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 735,
|
| 6 |
+
"global_step": 11748,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02502553626149132,
|
| 13 |
+
"grad_norm": 65.55949401855469,
|
| 14 |
+
"learning_rate": 4.834865509022812e-07,
|
| 15 |
+
"loss": 16.851,
|
| 16 |
+
"step": 147
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.05005107252298264,
|
| 20 |
+
"grad_norm": 23.207971572875977,
|
| 21 |
+
"learning_rate": 9.805924412665985e-07,
|
| 22 |
+
"loss": 11.2787,
|
| 23 |
+
"step": 294
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.07507660878447395,
|
| 27 |
+
"grad_norm": 176.1532440185547,
|
| 28 |
+
"learning_rate": 1.481103166496425e-06,
|
| 29 |
+
"loss": 8.9166,
|
| 30 |
+
"step": 441
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.10010214504596528,
|
| 34 |
+
"grad_norm": 22.1564998626709,
|
| 35 |
+
"learning_rate": 1.981613891726251e-06,
|
| 36 |
+
"loss": 7.9463,
|
| 37 |
+
"step": 588
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.12512768130745658,
|
| 41 |
+
"grad_norm": 20.11876106262207,
|
| 42 |
+
"learning_rate": 2.4821246169560777e-06,
|
| 43 |
+
"loss": 7.2108,
|
| 44 |
+
"step": 735
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.12512768130745658,
|
| 48 |
+
"eval_nli-pairs_loss": 6.905651569366455,
|
| 49 |
+
"eval_nli-pairs_runtime": 4.0844,
|
| 50 |
+
"eval_nli-pairs_samples_per_second": 36.725,
|
| 51 |
+
"eval_nli-pairs_steps_per_second": 1.224,
|
| 52 |
+
"eval_sts-test_pearson_cosine": 0.3740256550072784,
|
| 53 |
+
"eval_sts-test_pearson_dot": 0.13384893803205677,
|
| 54 |
+
"eval_sts-test_pearson_euclidean": 0.3912387619869807,
|
| 55 |
+
"eval_sts-test_pearson_manhattan": 0.4202605137823524,
|
| 56 |
+
"eval_sts-test_pearson_max": 0.4202605137823524,
|
| 57 |
+
"eval_sts-test_spearman_cosine": 0.37210107338950205,
|
| 58 |
+
"eval_sts-test_spearman_dot": 0.12092409843417483,
|
| 59 |
+
"eval_sts-test_spearman_euclidean": 0.39172287978780546,
|
| 60 |
+
"eval_sts-test_spearman_manhattan": 0.4169664738563951,
|
| 61 |
+
"eval_sts-test_spearman_max": 0.4169664738563951,
|
| 62 |
+
"step": 735
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.12512768130745658,
|
| 66 |
+
"eval_vitaminc-pairs_loss": 5.720878601074219,
|
| 67 |
+
"eval_vitaminc-pairs_runtime": 2.1703,
|
| 68 |
+
"eval_vitaminc-pairs_samples_per_second": 69.115,
|
| 69 |
+
"eval_vitaminc-pairs_steps_per_second": 2.304,
|
| 70 |
+
"step": 735
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.12512768130745658,
|
| 74 |
+
"eval_qnli-contrastive_loss": 8.1649751663208,
|
| 75 |
+
"eval_qnli-contrastive_runtime": 0.4937,
|
| 76 |
+
"eval_qnli-contrastive_samples_per_second": 303.841,
|
| 77 |
+
"eval_qnli-contrastive_steps_per_second": 10.128,
|
| 78 |
+
"step": 735
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.12512768130745658,
|
| 82 |
+
"eval_scitail-pairs-qa_loss": 3.7859296798706055,
|
| 83 |
+
"eval_scitail-pairs-qa_runtime": 1.1509,
|
| 84 |
+
"eval_scitail-pairs-qa_samples_per_second": 130.329,
|
| 85 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.344,
|
| 86 |
+
"step": 735
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.12512768130745658,
|
| 90 |
+
"eval_scitail-pairs-pos_loss": 3.9919917583465576,
|
| 91 |
+
"eval_scitail-pairs-pos_runtime": 2.1442,
|
| 92 |
+
"eval_scitail-pairs-pos_samples_per_second": 69.956,
|
| 93 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.332,
|
| 94 |
+
"step": 735
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.12512768130745658,
|
| 98 |
+
"eval_xsum-pairs_loss": 4.600368976593018,
|
| 99 |
+
"eval_xsum-pairs_runtime": 2.26,
|
| 100 |
+
"eval_xsum-pairs_samples_per_second": 66.371,
|
| 101 |
+
"eval_xsum-pairs_steps_per_second": 2.212,
|
| 102 |
+
"step": 735
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.12512768130745658,
|
| 106 |
+
"eval_compression-pairs_loss": 3.3037569522857666,
|
| 107 |
+
"eval_compression-pairs_runtime": 0.449,
|
| 108 |
+
"eval_compression-pairs_samples_per_second": 334.078,
|
| 109 |
+
"eval_compression-pairs_steps_per_second": 11.136,
|
| 110 |
+
"step": 735
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.12512768130745658,
|
| 114 |
+
"eval_sciq_pairs_loss": 10.214456558227539,
|
| 115 |
+
"eval_sciq_pairs_runtime": 7.1179,
|
| 116 |
+
"eval_sciq_pairs_samples_per_second": 21.074,
|
| 117 |
+
"eval_sciq_pairs_steps_per_second": 0.702,
|
| 118 |
+
"step": 735
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.12512768130745658,
|
| 122 |
+
"eval_qasc_pairs_loss": 10.58031940460205,
|
| 123 |
+
"eval_qasc_pairs_runtime": 2.0175,
|
| 124 |
+
"eval_qasc_pairs_samples_per_second": 74.348,
|
| 125 |
+
"eval_qasc_pairs_steps_per_second": 2.478,
|
| 126 |
+
"step": 735
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.12512768130745658,
|
| 130 |
+
"eval_openbookqa_pairs_loss": 7.862658977508545,
|
| 131 |
+
"eval_openbookqa_pairs_runtime": 0.8571,
|
| 132 |
+
"eval_openbookqa_pairs_samples_per_second": 120.168,
|
| 133 |
+
"eval_openbookqa_pairs_steps_per_second": 4.667,
|
| 134 |
+
"step": 735
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.12512768130745658,
|
| 138 |
+
"eval_msmarco_pairs_loss": 8.754273414611816,
|
| 139 |
+
"eval_msmarco_pairs_runtime": 2.7533,
|
| 140 |
+
"eval_msmarco_pairs_samples_per_second": 54.481,
|
| 141 |
+
"eval_msmarco_pairs_steps_per_second": 1.816,
|
| 142 |
+
"step": 735
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.12512768130745658,
|
| 146 |
+
"eval_nq_pairs_loss": 8.415486335754395,
|
| 147 |
+
"eval_nq_pairs_runtime": 5.0894,
|
| 148 |
+
"eval_nq_pairs_samples_per_second": 29.473,
|
| 149 |
+
"eval_nq_pairs_steps_per_second": 0.982,
|
| 150 |
+
"step": 735
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.12512768130745658,
|
| 154 |
+
"eval_trivia_pairs_loss": 9.051105499267578,
|
| 155 |
+
"eval_trivia_pairs_runtime": 9.5498,
|
| 156 |
+
"eval_trivia_pairs_samples_per_second": 15.707,
|
| 157 |
+
"eval_trivia_pairs_steps_per_second": 0.524,
|
| 158 |
+
"step": 735
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.12512768130745658,
|
| 162 |
+
"eval_quora_pairs_loss": 4.5232110023498535,
|
| 163 |
+
"eval_quora_pairs_runtime": 1.1469,
|
| 164 |
+
"eval_quora_pairs_samples_per_second": 130.785,
|
| 165 |
+
"eval_quora_pairs_steps_per_second": 4.36,
|
| 166 |
+
"step": 735
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 0.12512768130745658,
|
| 170 |
+
"eval_gooaq_pairs_loss": 7.579105854034424,
|
| 171 |
+
"eval_gooaq_pairs_runtime": 2.0491,
|
| 172 |
+
"eval_gooaq_pairs_samples_per_second": 73.203,
|
| 173 |
+
"eval_gooaq_pairs_steps_per_second": 2.44,
|
| 174 |
+
"step": 735
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.1501532175689479,
|
| 178 |
+
"grad_norm": 31.7736759185791,
|
| 179 |
+
"learning_rate": 2.982635342185904e-06,
|
| 180 |
+
"loss": 6.7709,
|
| 181 |
+
"step": 882
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.1751787538304392,
|
| 185 |
+
"grad_norm": 31.57339096069336,
|
| 186 |
+
"learning_rate": 3.4831460674157306e-06,
|
| 187 |
+
"loss": 6.1746,
|
| 188 |
+
"step": 1029
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.20020429009193055,
|
| 192 |
+
"grad_norm": 25.392702102661133,
|
| 193 |
+
"learning_rate": 3.9836567926455565e-06,
|
| 194 |
+
"loss": 5.7706,
|
| 195 |
+
"step": 1176
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.22522982635342187,
|
| 199 |
+
"grad_norm": 32.390472412109375,
|
| 200 |
+
"learning_rate": 4.484167517875383e-06,
|
| 201 |
+
"loss": 5.7283,
|
| 202 |
+
"step": 1323
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.25025536261491316,
|
| 206 |
+
"grad_norm": 18.85039520263672,
|
| 207 |
+
"learning_rate": 4.98467824310521e-06,
|
| 208 |
+
"loss": 5.1856,
|
| 209 |
+
"step": 1470
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 0.25025536261491316,
|
| 213 |
+
"eval_nli-pairs_loss": 4.352054119110107,
|
| 214 |
+
"eval_nli-pairs_runtime": 4.1476,
|
| 215 |
+
"eval_nli-pairs_samples_per_second": 36.165,
|
| 216 |
+
"eval_nli-pairs_steps_per_second": 1.206,
|
| 217 |
+
"eval_sts-test_pearson_cosine": 0.6694155778571752,
|
| 218 |
+
"eval_sts-test_pearson_dot": 0.5201102118957572,
|
| 219 |
+
"eval_sts-test_pearson_euclidean": 0.6613028243200022,
|
| 220 |
+
"eval_sts-test_pearson_manhattan": 0.6670710500315469,
|
| 221 |
+
"eval_sts-test_pearson_max": 0.6694155778571752,
|
| 222 |
+
"eval_sts-test_spearman_cosine": 0.6367853204388882,
|
| 223 |
+
"eval_sts-test_spearman_dot": 0.4940207180607985,
|
| 224 |
+
"eval_sts-test_spearman_euclidean": 0.6391132775161348,
|
| 225 |
+
"eval_sts-test_spearman_manhattan": 0.6446159957787251,
|
| 226 |
+
"eval_sts-test_spearman_max": 0.6446159957787251,
|
| 227 |
+
"step": 1470
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.25025536261491316,
|
| 231 |
+
"eval_vitaminc-pairs_loss": 3.4987735748291016,
|
| 232 |
+
"eval_vitaminc-pairs_runtime": 2.1678,
|
| 233 |
+
"eval_vitaminc-pairs_samples_per_second": 69.194,
|
| 234 |
+
"eval_vitaminc-pairs_steps_per_second": 2.306,
|
| 235 |
+
"step": 1470
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.25025536261491316,
|
| 239 |
+
"eval_qnli-contrastive_loss": 12.915559768676758,
|
| 240 |
+
"eval_qnli-contrastive_runtime": 0.4918,
|
| 241 |
+
"eval_qnli-contrastive_samples_per_second": 304.99,
|
| 242 |
+
"eval_qnli-contrastive_steps_per_second": 10.166,
|
| 243 |
+
"step": 1470
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.25025536261491316,
|
| 247 |
+
"eval_scitail-pairs-qa_loss": 1.3250077962875366,
|
| 248 |
+
"eval_scitail-pairs-qa_runtime": 1.154,
|
| 249 |
+
"eval_scitail-pairs-qa_samples_per_second": 129.984,
|
| 250 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.333,
|
| 251 |
+
"step": 1470
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.25025536261491316,
|
| 255 |
+
"eval_scitail-pairs-pos_loss": 2.457335948944092,
|
| 256 |
+
"eval_scitail-pairs-pos_runtime": 2.1475,
|
| 257 |
+
"eval_scitail-pairs-pos_samples_per_second": 69.85,
|
| 258 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.328,
|
| 259 |
+
"step": 1470
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 0.25025536261491316,
|
| 263 |
+
"eval_xsum-pairs_loss": 3.071201801300049,
|
| 264 |
+
"eval_xsum-pairs_runtime": 2.2634,
|
| 265 |
+
"eval_xsum-pairs_samples_per_second": 66.271,
|
| 266 |
+
"eval_xsum-pairs_steps_per_second": 2.209,
|
| 267 |
+
"step": 1470
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.25025536261491316,
|
| 271 |
+
"eval_compression-pairs_loss": 2.0629916191101074,
|
| 272 |
+
"eval_compression-pairs_runtime": 0.4529,
|
| 273 |
+
"eval_compression-pairs_samples_per_second": 331.23,
|
| 274 |
+
"eval_compression-pairs_steps_per_second": 11.041,
|
| 275 |
+
"step": 1470
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.25025536261491316,
|
| 279 |
+
"eval_sciq_pairs_loss": 9.06814193725586,
|
| 280 |
+
"eval_sciq_pairs_runtime": 7.1445,
|
| 281 |
+
"eval_sciq_pairs_samples_per_second": 20.995,
|
| 282 |
+
"eval_sciq_pairs_steps_per_second": 0.7,
|
| 283 |
+
"step": 1470
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.25025536261491316,
|
| 287 |
+
"eval_qasc_pairs_loss": 9.245658874511719,
|
| 288 |
+
"eval_qasc_pairs_runtime": 2.0471,
|
| 289 |
+
"eval_qasc_pairs_samples_per_second": 73.274,
|
| 290 |
+
"eval_qasc_pairs_steps_per_second": 2.442,
|
| 291 |
+
"step": 1470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.25025536261491316,
|
| 295 |
+
"eval_openbookqa_pairs_loss": 5.652446746826172,
|
| 296 |
+
"eval_openbookqa_pairs_runtime": 0.8946,
|
| 297 |
+
"eval_openbookqa_pairs_samples_per_second": 115.14,
|
| 298 |
+
"eval_openbookqa_pairs_steps_per_second": 4.471,
|
| 299 |
+
"step": 1470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.25025536261491316,
|
| 303 |
+
"eval_msmarco_pairs_loss": 4.844855785369873,
|
| 304 |
+
"eval_msmarco_pairs_runtime": 2.7887,
|
| 305 |
+
"eval_msmarco_pairs_samples_per_second": 53.788,
|
| 306 |
+
"eval_msmarco_pairs_steps_per_second": 1.793,
|
| 307 |
+
"step": 1470
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 0.25025536261491316,
|
| 311 |
+
"eval_nq_pairs_loss": 5.023958206176758,
|
| 312 |
+
"eval_nq_pairs_runtime": 5.0823,
|
| 313 |
+
"eval_nq_pairs_samples_per_second": 29.514,
|
| 314 |
+
"eval_nq_pairs_steps_per_second": 0.984,
|
| 315 |
+
"step": 1470
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.25025536261491316,
|
| 319 |
+
"eval_trivia_pairs_loss": 5.2907304763793945,
|
| 320 |
+
"eval_trivia_pairs_runtime": 9.6673,
|
| 321 |
+
"eval_trivia_pairs_samples_per_second": 15.516,
|
| 322 |
+
"eval_trivia_pairs_steps_per_second": 0.517,
|
| 323 |
+
"step": 1470
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.25025536261491316,
|
| 327 |
+
"eval_quora_pairs_loss": 1.5572240352630615,
|
| 328 |
+
"eval_quora_pairs_runtime": 1.1979,
|
| 329 |
+
"eval_quora_pairs_samples_per_second": 125.218,
|
| 330 |
+
"eval_quora_pairs_steps_per_second": 4.174,
|
| 331 |
+
"step": 1470
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.25025536261491316,
|
| 335 |
+
"eval_gooaq_pairs_loss": 3.970768928527832,
|
| 336 |
+
"eval_gooaq_pairs_runtime": 2.117,
|
| 337 |
+
"eval_gooaq_pairs_samples_per_second": 70.855,
|
| 338 |
+
"eval_gooaq_pairs_steps_per_second": 2.362,
|
| 339 |
+
"step": 1470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.2752808988764045,
|
| 343 |
+
"grad_norm": 40.67585754394531,
|
| 344 |
+
"learning_rate": 5.4851889683350365e-06,
|
| 345 |
+
"loss": 4.185,
|
| 346 |
+
"step": 1617
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.3003064351378958,
|
| 350 |
+
"grad_norm": 45.92570495605469,
|
| 351 |
+
"learning_rate": 5.985699693564862e-06,
|
| 352 |
+
"loss": 4.6367,
|
| 353 |
+
"step": 1764
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.32533197139938713,
|
| 357 |
+
"grad_norm": 13.566838264465332,
|
| 358 |
+
"learning_rate": 6.486210418794688e-06,
|
| 359 |
+
"loss": 4.3615,
|
| 360 |
+
"step": 1911
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.3503575076608784,
|
| 364 |
+
"grad_norm": 9.495999336242676,
|
| 365 |
+
"learning_rate": 6.986721144024515e-06,
|
| 366 |
+
"loss": 4.1791,
|
| 367 |
+
"step": 2058
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.37538304392236976,
|
| 371 |
+
"grad_norm": 32.735416412353516,
|
| 372 |
+
"learning_rate": 7.487231869254341e-06,
|
| 373 |
+
"loss": 4.1051,
|
| 374 |
+
"step": 2205
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.37538304392236976,
|
| 378 |
+
"eval_nli-pairs_loss": 3.2717113494873047,
|
| 379 |
+
"eval_nli-pairs_runtime": 4.0124,
|
| 380 |
+
"eval_nli-pairs_samples_per_second": 37.384,
|
| 381 |
+
"eval_nli-pairs_steps_per_second": 1.246,
|
| 382 |
+
"eval_sts-test_pearson_cosine": 0.6958570089637609,
|
| 383 |
+
"eval_sts-test_pearson_dot": 0.5824298957890577,
|
| 384 |
+
"eval_sts-test_pearson_euclidean": 0.6893962819387462,
|
| 385 |
+
"eval_sts-test_pearson_manhattan": 0.6993681181979946,
|
| 386 |
+
"eval_sts-test_pearson_max": 0.6993681181979946,
|
| 387 |
+
"eval_sts-test_spearman_cosine": 0.6652712160836801,
|
| 388 |
+
"eval_sts-test_spearman_dot": 0.5536505624407877,
|
| 389 |
+
"eval_sts-test_spearman_euclidean": 0.6659844314307678,
|
| 390 |
+
"eval_sts-test_spearman_manhattan": 0.675740852112121,
|
| 391 |
+
"eval_sts-test_spearman_max": 0.675740852112121,
|
| 392 |
+
"step": 2205
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.37538304392236976,
|
| 396 |
+
"eval_vitaminc-pairs_loss": 2.7197911739349365,
|
| 397 |
+
"eval_vitaminc-pairs_runtime": 2.1625,
|
| 398 |
+
"eval_vitaminc-pairs_samples_per_second": 69.365,
|
| 399 |
+
"eval_vitaminc-pairs_steps_per_second": 2.312,
|
| 400 |
+
"step": 2205
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 0.37538304392236976,
|
| 404 |
+
"eval_qnli-contrastive_loss": 9.638714790344238,
|
| 405 |
+
"eval_qnli-contrastive_runtime": 0.4877,
|
| 406 |
+
"eval_qnli-contrastive_samples_per_second": 307.567,
|
| 407 |
+
"eval_qnli-contrastive_steps_per_second": 10.252,
|
| 408 |
+
"step": 2205
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.37538304392236976,
|
| 412 |
+
"eval_scitail-pairs-qa_loss": 0.8106752634048462,
|
| 413 |
+
"eval_scitail-pairs-qa_runtime": 1.1588,
|
| 414 |
+
"eval_scitail-pairs-qa_samples_per_second": 129.449,
|
| 415 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.315,
|
| 416 |
+
"step": 2205
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.37538304392236976,
|
| 420 |
+
"eval_scitail-pairs-pos_loss": 1.8894625902175903,
|
| 421 |
+
"eval_scitail-pairs-pos_runtime": 2.1181,
|
| 422 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.817,
|
| 423 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.361,
|
| 424 |
+
"step": 2205
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.37538304392236976,
|
| 428 |
+
"eval_xsum-pairs_loss": 2.262718439102173,
|
| 429 |
+
"eval_xsum-pairs_runtime": 2.2585,
|
| 430 |
+
"eval_xsum-pairs_samples_per_second": 66.416,
|
| 431 |
+
"eval_xsum-pairs_steps_per_second": 2.214,
|
| 432 |
+
"step": 2205
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 0.37538304392236976,
|
| 436 |
+
"eval_compression-pairs_loss": 1.4910633563995361,
|
| 437 |
+
"eval_compression-pairs_runtime": 0.4462,
|
| 438 |
+
"eval_compression-pairs_samples_per_second": 336.204,
|
| 439 |
+
"eval_compression-pairs_steps_per_second": 11.207,
|
| 440 |
+
"step": 2205
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.37538304392236976,
|
| 444 |
+
"eval_sciq_pairs_loss": 8.59740161895752,
|
| 445 |
+
"eval_sciq_pairs_runtime": 7.1845,
|
| 446 |
+
"eval_sciq_pairs_samples_per_second": 20.878,
|
| 447 |
+
"eval_sciq_pairs_steps_per_second": 0.696,
|
| 448 |
+
"step": 2205
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.37538304392236976,
|
| 452 |
+
"eval_qasc_pairs_loss": 8.103879928588867,
|
| 453 |
+
"eval_qasc_pairs_runtime": 2.0762,
|
| 454 |
+
"eval_qasc_pairs_samples_per_second": 72.246,
|
| 455 |
+
"eval_qasc_pairs_steps_per_second": 2.408,
|
| 456 |
+
"step": 2205
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 0.37538304392236976,
|
| 460 |
+
"eval_openbookqa_pairs_loss": 5.090969562530518,
|
| 461 |
+
"eval_openbookqa_pairs_runtime": 0.89,
|
| 462 |
+
"eval_openbookqa_pairs_samples_per_second": 115.726,
|
| 463 |
+
"eval_openbookqa_pairs_steps_per_second": 4.494,
|
| 464 |
+
"step": 2205
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.37538304392236976,
|
| 468 |
+
"eval_msmarco_pairs_loss": 3.9566943645477295,
|
| 469 |
+
"eval_msmarco_pairs_runtime": 2.8183,
|
| 470 |
+
"eval_msmarco_pairs_samples_per_second": 53.223,
|
| 471 |
+
"eval_msmarco_pairs_steps_per_second": 1.774,
|
| 472 |
+
"step": 2205
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.37538304392236976,
|
| 476 |
+
"eval_nq_pairs_loss": 4.009054183959961,
|
| 477 |
+
"eval_nq_pairs_runtime": 5.0219,
|
| 478 |
+
"eval_nq_pairs_samples_per_second": 29.869,
|
| 479 |
+
"eval_nq_pairs_steps_per_second": 0.996,
|
| 480 |
+
"step": 2205
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 0.37538304392236976,
|
| 484 |
+
"eval_trivia_pairs_loss": 4.286431312561035,
|
| 485 |
+
"eval_trivia_pairs_runtime": 9.4975,
|
| 486 |
+
"eval_trivia_pairs_samples_per_second": 15.794,
|
| 487 |
+
"eval_trivia_pairs_steps_per_second": 0.526,
|
| 488 |
+
"step": 2205
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.37538304392236976,
|
| 492 |
+
"eval_quora_pairs_loss": 1.123273491859436,
|
| 493 |
+
"eval_quora_pairs_runtime": 1.1487,
|
| 494 |
+
"eval_quora_pairs_samples_per_second": 130.586,
|
| 495 |
+
"eval_quora_pairs_steps_per_second": 4.353,
|
| 496 |
+
"step": 2205
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 0.37538304392236976,
|
| 500 |
+
"eval_gooaq_pairs_loss": 3.222414255142212,
|
| 501 |
+
"eval_gooaq_pairs_runtime": 2.0173,
|
| 502 |
+
"eval_gooaq_pairs_samples_per_second": 74.357,
|
| 503 |
+
"eval_gooaq_pairs_steps_per_second": 2.479,
|
| 504 |
+
"step": 2205
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.4004085801838611,
|
| 508 |
+
"grad_norm": 218.56105041503906,
|
| 509 |
+
"learning_rate": 7.987742594484168e-06,
|
| 510 |
+
"loss": 3.7674,
|
| 511 |
+
"step": 2352
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 0.4254341164453524,
|
| 515 |
+
"grad_norm": 27.877609252929688,
|
| 516 |
+
"learning_rate": 8.488253319713993e-06,
|
| 517 |
+
"loss": 3.8729,
|
| 518 |
+
"step": 2499
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.45045965270684374,
|
| 522 |
+
"grad_norm": 33.50013732910156,
|
| 523 |
+
"learning_rate": 8.988764044943822e-06,
|
| 524 |
+
"loss": 3.4527,
|
| 525 |
+
"step": 2646
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.475485188968335,
|
| 529 |
+
"grad_norm": 14.015911102294922,
|
| 530 |
+
"learning_rate": 9.489274770173647e-06,
|
| 531 |
+
"loss": 3.3545,
|
| 532 |
+
"step": 2793
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 0.5005107252298263,
|
| 536 |
+
"grad_norm": 33.59694290161133,
|
| 537 |
+
"learning_rate": 9.989785495403473e-06,
|
| 538 |
+
"loss": 3.3247,
|
| 539 |
+
"step": 2940
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.5005107252298263,
|
| 543 |
+
"eval_nli-pairs_loss": 2.7121565341949463,
|
| 544 |
+
"eval_nli-pairs_runtime": 4.1564,
|
| 545 |
+
"eval_nli-pairs_samples_per_second": 36.089,
|
| 546 |
+
"eval_nli-pairs_steps_per_second": 1.203,
|
| 547 |
+
"eval_sts-test_pearson_cosine": 0.716623047702725,
|
| 548 |
+
"eval_sts-test_pearson_dot": 0.6128451070598809,
|
| 549 |
+
"eval_sts-test_pearson_euclidean": 0.7138791236031807,
|
| 550 |
+
"eval_sts-test_pearson_manhattan": 0.7213151818687454,
|
| 551 |
+
"eval_sts-test_pearson_max": 0.7213151818687454,
|
| 552 |
+
"eval_sts-test_spearman_cosine": 0.6919792400941177,
|
| 553 |
+
"eval_sts-test_spearman_dot": 0.5867158357121192,
|
| 554 |
+
"eval_sts-test_spearman_euclidean": 0.6925037259567834,
|
| 555 |
+
"eval_sts-test_spearman_manhattan": 0.7008895667910079,
|
| 556 |
+
"eval_sts-test_spearman_max": 0.7008895667910079,
|
| 557 |
+
"step": 2940
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.5005107252298263,
|
| 561 |
+
"eval_vitaminc-pairs_loss": 2.225992441177368,
|
| 562 |
+
"eval_vitaminc-pairs_runtime": 2.253,
|
| 563 |
+
"eval_vitaminc-pairs_samples_per_second": 66.577,
|
| 564 |
+
"eval_vitaminc-pairs_steps_per_second": 2.219,
|
| 565 |
+
"step": 2940
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 0.5005107252298263,
|
| 569 |
+
"eval_qnli-contrastive_loss": 4.92629861831665,
|
| 570 |
+
"eval_qnli-contrastive_runtime": 0.5005,
|
| 571 |
+
"eval_qnli-contrastive_samples_per_second": 299.691,
|
| 572 |
+
"eval_qnli-contrastive_steps_per_second": 9.99,
|
| 573 |
+
"step": 2940
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.5005107252298263,
|
| 577 |
+
"eval_scitail-pairs-qa_loss": 0.5898066163063049,
|
| 578 |
+
"eval_scitail-pairs-qa_runtime": 1.2227,
|
| 579 |
+
"eval_scitail-pairs-qa_samples_per_second": 122.682,
|
| 580 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.089,
|
| 581 |
+
"step": 2940
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.5005107252298263,
|
| 585 |
+
"eval_scitail-pairs-pos_loss": 1.4237287044525146,
|
| 586 |
+
"eval_scitail-pairs-pos_runtime": 2.4409,
|
| 587 |
+
"eval_scitail-pairs-pos_samples_per_second": 61.452,
|
| 588 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.048,
|
| 589 |
+
"step": 2940
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 0.5005107252298263,
|
| 593 |
+
"eval_xsum-pairs_loss": 1.8388895988464355,
|
| 594 |
+
"eval_xsum-pairs_runtime": 2.2831,
|
| 595 |
+
"eval_xsum-pairs_samples_per_second": 65.7,
|
| 596 |
+
"eval_xsum-pairs_steps_per_second": 2.19,
|
| 597 |
+
"step": 2940
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.5005107252298263,
|
| 601 |
+
"eval_compression-pairs_loss": 1.1590967178344727,
|
| 602 |
+
"eval_compression-pairs_runtime": 0.5152,
|
| 603 |
+
"eval_compression-pairs_samples_per_second": 291.165,
|
| 604 |
+
"eval_compression-pairs_steps_per_second": 9.706,
|
| 605 |
+
"step": 2940
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.5005107252298263,
|
| 609 |
+
"eval_sciq_pairs_loss": 8.282496452331543,
|
| 610 |
+
"eval_sciq_pairs_runtime": 7.2871,
|
| 611 |
+
"eval_sciq_pairs_samples_per_second": 20.584,
|
| 612 |
+
"eval_sciq_pairs_steps_per_second": 0.686,
|
| 613 |
+
"step": 2940
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.5005107252298263,
|
| 617 |
+
"eval_qasc_pairs_loss": 7.817965507507324,
|
| 618 |
+
"eval_qasc_pairs_runtime": 2.0211,
|
| 619 |
+
"eval_qasc_pairs_samples_per_second": 74.218,
|
| 620 |
+
"eval_qasc_pairs_steps_per_second": 2.474,
|
| 621 |
+
"step": 2940
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.5005107252298263,
|
| 625 |
+
"eval_openbookqa_pairs_loss": 4.619383811950684,
|
| 626 |
+
"eval_openbookqa_pairs_runtime": 0.8531,
|
| 627 |
+
"eval_openbookqa_pairs_samples_per_second": 120.731,
|
| 628 |
+
"eval_openbookqa_pairs_steps_per_second": 4.689,
|
| 629 |
+
"step": 2940
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.5005107252298263,
|
| 633 |
+
"eval_msmarco_pairs_loss": 3.478559970855713,
|
| 634 |
+
"eval_msmarco_pairs_runtime": 2.7512,
|
| 635 |
+
"eval_msmarco_pairs_samples_per_second": 54.522,
|
| 636 |
+
"eval_msmarco_pairs_steps_per_second": 1.817,
|
| 637 |
+
"step": 2940
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 0.5005107252298263,
|
| 641 |
+
"eval_nq_pairs_loss": 3.3449866771698,
|
| 642 |
+
"eval_nq_pairs_runtime": 5.0591,
|
| 643 |
+
"eval_nq_pairs_samples_per_second": 29.649,
|
| 644 |
+
"eval_nq_pairs_steps_per_second": 0.988,
|
| 645 |
+
"step": 2940
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.5005107252298263,
|
| 649 |
+
"eval_trivia_pairs_loss": 3.524484872817993,
|
| 650 |
+
"eval_trivia_pairs_runtime": 9.662,
|
| 651 |
+
"eval_trivia_pairs_samples_per_second": 15.525,
|
| 652 |
+
"eval_trivia_pairs_steps_per_second": 0.517,
|
| 653 |
+
"step": 2940
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.5005107252298263,
|
| 657 |
+
"eval_quora_pairs_loss": 0.9095575213432312,
|
| 658 |
+
"eval_quora_pairs_runtime": 1.2482,
|
| 659 |
+
"eval_quora_pairs_samples_per_second": 120.175,
|
| 660 |
+
"eval_quora_pairs_steps_per_second": 4.006,
|
| 661 |
+
"step": 2940
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.5005107252298263,
|
| 665 |
+
"eval_gooaq_pairs_loss": 2.6586034297943115,
|
| 666 |
+
"eval_gooaq_pairs_runtime": 2.1091,
|
| 667 |
+
"eval_gooaq_pairs_samples_per_second": 71.12,
|
| 668 |
+
"eval_gooaq_pairs_steps_per_second": 2.371,
|
| 669 |
+
"step": 2940
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.5255362614913177,
|
| 673 |
+
"grad_norm": 35.33409118652344,
|
| 674 |
+
"learning_rate": 1.04902962206333e-05,
|
| 675 |
+
"loss": 3.116,
|
| 676 |
+
"step": 3087
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.550561797752809,
|
| 680 |
+
"grad_norm": 22.29003143310547,
|
| 681 |
+
"learning_rate": 1.0990806945863125e-05,
|
| 682 |
+
"loss": 3.2418,
|
| 683 |
+
"step": 3234
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 0.5755873340143003,
|
| 687 |
+
"grad_norm": 31.277965545654297,
|
| 688 |
+
"learning_rate": 1.1491317671092953e-05,
|
| 689 |
+
"loss": 3.0757,
|
| 690 |
+
"step": 3381
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.6006128702757916,
|
| 694 |
+
"grad_norm": 24.612506866455078,
|
| 695 |
+
"learning_rate": 1.1991828396322778e-05,
|
| 696 |
+
"loss": 2.8524,
|
| 697 |
+
"step": 3528
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 0.625638406537283,
|
| 701 |
+
"grad_norm": 25.11741065979004,
|
| 702 |
+
"learning_rate": 1.2492339121552605e-05,
|
| 703 |
+
"loss": 2.6875,
|
| 704 |
+
"step": 3675
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.625638406537283,
|
| 708 |
+
"eval_nli-pairs_loss": 2.479051113128662,
|
| 709 |
+
"eval_nli-pairs_runtime": 3.9943,
|
| 710 |
+
"eval_nli-pairs_samples_per_second": 37.553,
|
| 711 |
+
"eval_nli-pairs_steps_per_second": 1.252,
|
| 712 |
+
"eval_sts-test_pearson_cosine": 0.7278742453545186,
|
| 713 |
+
"eval_sts-test_pearson_dot": 0.6217650825208566,
|
| 714 |
+
"eval_sts-test_pearson_euclidean": 0.7243228472931561,
|
| 715 |
+
"eval_sts-test_pearson_manhattan": 0.7333297580184588,
|
| 716 |
+
"eval_sts-test_pearson_max": 0.7333297580184588,
|
| 717 |
+
"eval_sts-test_spearman_cosine": 0.7013110457844404,
|
| 718 |
+
"eval_sts-test_spearman_dot": 0.5970993074902947,
|
| 719 |
+
"eval_sts-test_spearman_euclidean": 0.701564129266252,
|
| 720 |
+
"eval_sts-test_spearman_manhattan": 0.7116482009924582,
|
| 721 |
+
"eval_sts-test_spearman_max": 0.7116482009924582,
|
| 722 |
+
"step": 3675
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.625638406537283,
|
| 726 |
+
"eval_vitaminc-pairs_loss": 1.974273681640625,
|
| 727 |
+
"eval_vitaminc-pairs_runtime": 2.1754,
|
| 728 |
+
"eval_vitaminc-pairs_samples_per_second": 68.953,
|
| 729 |
+
"eval_vitaminc-pairs_steps_per_second": 2.298,
|
| 730 |
+
"step": 3675
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.625638406537283,
|
| 734 |
+
"eval_qnli-contrastive_loss": 1.7706010341644287,
|
| 735 |
+
"eval_qnli-contrastive_runtime": 0.4866,
|
| 736 |
+
"eval_qnli-contrastive_samples_per_second": 308.244,
|
| 737 |
+
"eval_qnli-contrastive_steps_per_second": 10.275,
|
| 738 |
+
"step": 3675
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.625638406537283,
|
| 742 |
+
"eval_scitail-pairs-qa_loss": 0.4400452673435211,
|
| 743 |
+
"eval_scitail-pairs-qa_runtime": 1.1519,
|
| 744 |
+
"eval_scitail-pairs-qa_samples_per_second": 130.222,
|
| 745 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.341,
|
| 746 |
+
"step": 3675
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.625638406537283,
|
| 750 |
+
"eval_scitail-pairs-pos_loss": 1.1909903287887573,
|
| 751 |
+
"eval_scitail-pairs-pos_runtime": 2.1319,
|
| 752 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.36,
|
| 753 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.345,
|
| 754 |
+
"step": 3675
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.625638406537283,
|
| 758 |
+
"eval_xsum-pairs_loss": 1.4811985492706299,
|
| 759 |
+
"eval_xsum-pairs_runtime": 2.254,
|
| 760 |
+
"eval_xsum-pairs_samples_per_second": 66.548,
|
| 761 |
+
"eval_xsum-pairs_steps_per_second": 2.218,
|
| 762 |
+
"step": 3675
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.625638406537283,
|
| 766 |
+
"eval_compression-pairs_loss": 0.8453781008720398,
|
| 767 |
+
"eval_compression-pairs_runtime": 0.4401,
|
| 768 |
+
"eval_compression-pairs_samples_per_second": 340.826,
|
| 769 |
+
"eval_compression-pairs_steps_per_second": 11.361,
|
| 770 |
+
"step": 3675
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.625638406537283,
|
| 774 |
+
"eval_sciq_pairs_loss": 8.014656066894531,
|
| 775 |
+
"eval_sciq_pairs_runtime": 7.0707,
|
| 776 |
+
"eval_sciq_pairs_samples_per_second": 21.214,
|
| 777 |
+
"eval_sciq_pairs_steps_per_second": 0.707,
|
| 778 |
+
"step": 3675
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.625638406537283,
|
| 782 |
+
"eval_qasc_pairs_loss": 6.9316277503967285,
|
| 783 |
+
"eval_qasc_pairs_runtime": 2.0338,
|
| 784 |
+
"eval_qasc_pairs_samples_per_second": 73.752,
|
| 785 |
+
"eval_qasc_pairs_steps_per_second": 2.458,
|
| 786 |
+
"step": 3675
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.625638406537283,
|
| 790 |
+
"eval_openbookqa_pairs_loss": 4.21690034866333,
|
| 791 |
+
"eval_openbookqa_pairs_runtime": 0.918,
|
| 792 |
+
"eval_openbookqa_pairs_samples_per_second": 112.202,
|
| 793 |
+
"eval_openbookqa_pairs_steps_per_second": 4.357,
|
| 794 |
+
"step": 3675
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.625638406537283,
|
| 798 |
+
"eval_msmarco_pairs_loss": 3.0209598541259766,
|
| 799 |
+
"eval_msmarco_pairs_runtime": 2.7749,
|
| 800 |
+
"eval_msmarco_pairs_samples_per_second": 54.056,
|
| 801 |
+
"eval_msmarco_pairs_steps_per_second": 1.802,
|
| 802 |
+
"step": 3675
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.625638406537283,
|
| 806 |
+
"eval_nq_pairs_loss": 2.956088066101074,
|
| 807 |
+
"eval_nq_pairs_runtime": 5.0024,
|
| 808 |
+
"eval_nq_pairs_samples_per_second": 29.986,
|
| 809 |
+
"eval_nq_pairs_steps_per_second": 1.0,
|
| 810 |
+
"step": 3675
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.625638406537283,
|
| 814 |
+
"eval_trivia_pairs_loss": 3.17364501953125,
|
| 815 |
+
"eval_trivia_pairs_runtime": 9.4856,
|
| 816 |
+
"eval_trivia_pairs_samples_per_second": 15.813,
|
| 817 |
+
"eval_trivia_pairs_steps_per_second": 0.527,
|
| 818 |
+
"step": 3675
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.625638406537283,
|
| 822 |
+
"eval_quora_pairs_loss": 0.763593852519989,
|
| 823 |
+
"eval_quora_pairs_runtime": 1.1441,
|
| 824 |
+
"eval_quora_pairs_samples_per_second": 131.104,
|
| 825 |
+
"eval_quora_pairs_steps_per_second": 4.37,
|
| 826 |
+
"step": 3675
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.625638406537283,
|
| 830 |
+
"eval_gooaq_pairs_loss": 2.3524909019470215,
|
| 831 |
+
"eval_gooaq_pairs_runtime": 2.0161,
|
| 832 |
+
"eval_gooaq_pairs_samples_per_second": 74.4,
|
| 833 |
+
"eval_gooaq_pairs_steps_per_second": 2.48,
|
| 834 |
+
"step": 3675
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.6506639427987743,
|
| 838 |
+
"grad_norm": 31.163997650146484,
|
| 839 |
+
"learning_rate": 1.2992849846782432e-05,
|
| 840 |
+
"loss": 2.7808,
|
| 841 |
+
"step": 3822
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.6756894790602656,
|
| 845 |
+
"grad_norm": 14.883658409118652,
|
| 846 |
+
"learning_rate": 1.3493360572012258e-05,
|
| 847 |
+
"loss": 2.5687,
|
| 848 |
+
"step": 3969
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.7007150153217568,
|
| 852 |
+
"grad_norm": 5.874042987823486,
|
| 853 |
+
"learning_rate": 1.3993871297242083e-05,
|
| 854 |
+
"loss": 2.3034,
|
| 855 |
+
"step": 4116
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.7257405515832482,
|
| 859 |
+
"grad_norm": 31.464054107666016,
|
| 860 |
+
"learning_rate": 1.4494382022471912e-05,
|
| 861 |
+
"loss": 2.4412,
|
| 862 |
+
"step": 4263
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.7507660878447395,
|
| 866 |
+
"grad_norm": 16.43915367126465,
|
| 867 |
+
"learning_rate": 1.4994892747701737e-05,
|
| 868 |
+
"loss": 2.3293,
|
| 869 |
+
"step": 4410
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.7507660878447395,
|
| 873 |
+
"eval_nli-pairs_loss": 2.3226094245910645,
|
| 874 |
+
"eval_nli-pairs_runtime": 4.113,
|
| 875 |
+
"eval_nli-pairs_samples_per_second": 36.47,
|
| 876 |
+
"eval_nli-pairs_steps_per_second": 1.216,
|
| 877 |
+
"eval_sts-test_pearson_cosine": 0.7356971966139032,
|
| 878 |
+
"eval_sts-test_pearson_dot": 0.6150809513049869,
|
| 879 |
+
"eval_sts-test_pearson_euclidean": 0.7330733579988641,
|
| 880 |
+
"eval_sts-test_pearson_manhattan": 0.7423412248131348,
|
| 881 |
+
"eval_sts-test_pearson_max": 0.7423412248131348,
|
| 882 |
+
"eval_sts-test_spearman_cosine": 0.7121899723082045,
|
| 883 |
+
"eval_sts-test_spearman_dot": 0.5926505936679538,
|
| 884 |
+
"eval_sts-test_spearman_euclidean": 0.7130179905407037,
|
| 885 |
+
"eval_sts-test_spearman_manhattan": 0.7227257562995023,
|
| 886 |
+
"eval_sts-test_spearman_max": 0.7227257562995023,
|
| 887 |
+
"step": 4410
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.7507660878447395,
|
| 891 |
+
"eval_vitaminc-pairs_loss": 1.7956713438034058,
|
| 892 |
+
"eval_vitaminc-pairs_runtime": 2.174,
|
| 893 |
+
"eval_vitaminc-pairs_samples_per_second": 68.996,
|
| 894 |
+
"eval_vitaminc-pairs_steps_per_second": 2.3,
|
| 895 |
+
"step": 4410
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.7507660878447395,
|
| 899 |
+
"eval_qnli-contrastive_loss": 1.0078614950180054,
|
| 900 |
+
"eval_qnli-contrastive_runtime": 0.4874,
|
| 901 |
+
"eval_qnli-contrastive_samples_per_second": 307.763,
|
| 902 |
+
"eval_qnli-contrastive_steps_per_second": 10.259,
|
| 903 |
+
"step": 4410
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.7507660878447395,
|
| 907 |
+
"eval_scitail-pairs-qa_loss": 0.36971578001976013,
|
| 908 |
+
"eval_scitail-pairs-qa_runtime": 1.164,
|
| 909 |
+
"eval_scitail-pairs-qa_samples_per_second": 128.863,
|
| 910 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.295,
|
| 911 |
+
"step": 4410
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 0.7507660878447395,
|
| 915 |
+
"eval_scitail-pairs-pos_loss": 1.0497769117355347,
|
| 916 |
+
"eval_scitail-pairs-pos_runtime": 2.1205,
|
| 917 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.74,
|
| 918 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.358,
|
| 919 |
+
"step": 4410
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.7507660878447395,
|
| 923 |
+
"eval_xsum-pairs_loss": 1.1691261529922485,
|
| 924 |
+
"eval_xsum-pairs_runtime": 2.259,
|
| 925 |
+
"eval_xsum-pairs_samples_per_second": 66.401,
|
| 926 |
+
"eval_xsum-pairs_steps_per_second": 2.213,
|
| 927 |
+
"step": 4410
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.7507660878447395,
|
| 931 |
+
"eval_compression-pairs_loss": 0.5027483105659485,
|
| 932 |
+
"eval_compression-pairs_runtime": 0.4403,
|
| 933 |
+
"eval_compression-pairs_samples_per_second": 340.682,
|
| 934 |
+
"eval_compression-pairs_steps_per_second": 11.356,
|
| 935 |
+
"step": 4410
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.7507660878447395,
|
| 939 |
+
"eval_sciq_pairs_loss": 7.823739528656006,
|
| 940 |
+
"eval_sciq_pairs_runtime": 7.0738,
|
| 941 |
+
"eval_sciq_pairs_samples_per_second": 21.205,
|
| 942 |
+
"eval_sciq_pairs_steps_per_second": 0.707,
|
| 943 |
+
"step": 4410
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.7507660878447395,
|
| 947 |
+
"eval_qasc_pairs_loss": 6.404655933380127,
|
| 948 |
+
"eval_qasc_pairs_runtime": 2.0346,
|
| 949 |
+
"eval_qasc_pairs_samples_per_second": 73.723,
|
| 950 |
+
"eval_qasc_pairs_steps_per_second": 2.457,
|
| 951 |
+
"step": 4410
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.7507660878447395,
|
| 955 |
+
"eval_openbookqa_pairs_loss": 3.857389211654663,
|
| 956 |
+
"eval_openbookqa_pairs_runtime": 0.8544,
|
| 957 |
+
"eval_openbookqa_pairs_samples_per_second": 120.547,
|
| 958 |
+
"eval_openbookqa_pairs_steps_per_second": 4.681,
|
| 959 |
+
"step": 4410
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.7507660878447395,
|
| 963 |
+
"eval_msmarco_pairs_loss": 2.7028510570526123,
|
| 964 |
+
"eval_msmarco_pairs_runtime": 2.7448,
|
| 965 |
+
"eval_msmarco_pairs_samples_per_second": 54.649,
|
| 966 |
+
"eval_msmarco_pairs_steps_per_second": 1.822,
|
| 967 |
+
"step": 4410
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 0.7507660878447395,
|
| 971 |
+
"eval_nq_pairs_loss": 2.679351329803467,
|
| 972 |
+
"eval_nq_pairs_runtime": 5.067,
|
| 973 |
+
"eval_nq_pairs_samples_per_second": 29.603,
|
| 974 |
+
"eval_nq_pairs_steps_per_second": 0.987,
|
| 975 |
+
"step": 4410
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.7507660878447395,
|
| 979 |
+
"eval_trivia_pairs_loss": 2.8798065185546875,
|
| 980 |
+
"eval_trivia_pairs_runtime": 9.5449,
|
| 981 |
+
"eval_trivia_pairs_samples_per_second": 15.715,
|
| 982 |
+
"eval_trivia_pairs_steps_per_second": 0.524,
|
| 983 |
+
"step": 4410
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.7507660878447395,
|
| 987 |
+
"eval_quora_pairs_loss": 0.6825175285339355,
|
| 988 |
+
"eval_quora_pairs_runtime": 1.1431,
|
| 989 |
+
"eval_quora_pairs_samples_per_second": 131.221,
|
| 990 |
+
"eval_quora_pairs_steps_per_second": 4.374,
|
| 991 |
+
"step": 4410
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.7507660878447395,
|
| 995 |
+
"eval_gooaq_pairs_loss": 2.0472166538238525,
|
| 996 |
+
"eval_gooaq_pairs_runtime": 2.0218,
|
| 997 |
+
"eval_gooaq_pairs_samples_per_second": 74.191,
|
| 998 |
+
"eval_gooaq_pairs_steps_per_second": 2.473,
|
| 999 |
+
"step": 4410
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.7757916241062308,
|
| 1003 |
+
"grad_norm": 4.2425055503845215,
|
| 1004 |
+
"learning_rate": 1.5495403472931565e-05,
|
| 1005 |
+
"loss": 2.3651,
|
| 1006 |
+
"step": 4557
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.8008171603677222,
|
| 1010 |
+
"grad_norm": 22.42776107788086,
|
| 1011 |
+
"learning_rate": 1.5995914198161388e-05,
|
| 1012 |
+
"loss": 2.6296,
|
| 1013 |
+
"step": 4704
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.8258426966292135,
|
| 1017 |
+
"grad_norm": 21.169517517089844,
|
| 1018 |
+
"learning_rate": 1.6496424923391215e-05,
|
| 1019 |
+
"loss": 2.2108,
|
| 1020 |
+
"step": 4851
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.8508682328907048,
|
| 1024 |
+
"grad_norm": 23.326181411743164,
|
| 1025 |
+
"learning_rate": 1.699693564862104e-05,
|
| 1026 |
+
"loss": 2.1852,
|
| 1027 |
+
"step": 4998
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.8758937691521961,
|
| 1031 |
+
"grad_norm": 24.574176788330078,
|
| 1032 |
+
"learning_rate": 1.7497446373850868e-05,
|
| 1033 |
+
"loss": 2.2944,
|
| 1034 |
+
"step": 5145
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.8758937691521961,
|
| 1038 |
+
"eval_nli-pairs_loss": 2.0634915828704834,
|
| 1039 |
+
"eval_nli-pairs_runtime": 4.0019,
|
| 1040 |
+
"eval_nli-pairs_samples_per_second": 37.482,
|
| 1041 |
+
"eval_nli-pairs_steps_per_second": 1.249,
|
| 1042 |
+
"eval_sts-test_pearson_cosine": 0.7466390532977636,
|
| 1043 |
+
"eval_sts-test_pearson_dot": 0.612259458274589,
|
| 1044 |
+
"eval_sts-test_pearson_euclidean": 0.7432536346376271,
|
| 1045 |
+
"eval_sts-test_pearson_manhattan": 0.7500490179501229,
|
| 1046 |
+
"eval_sts-test_pearson_max": 0.7500490179501229,
|
| 1047 |
+
"eval_sts-test_spearman_cosine": 0.728273260456201,
|
| 1048 |
+
"eval_sts-test_spearman_dot": 0.5960115087190596,
|
| 1049 |
+
"eval_sts-test_spearman_euclidean": 0.7272394395622148,
|
| 1050 |
+
"eval_sts-test_spearman_manhattan": 0.7334149564445704,
|
| 1051 |
+
"eval_sts-test_spearman_max": 0.7334149564445704,
|
| 1052 |
+
"step": 5145
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.8758937691521961,
|
| 1056 |
+
"eval_vitaminc-pairs_loss": 1.638654112815857,
|
| 1057 |
+
"eval_vitaminc-pairs_runtime": 2.1637,
|
| 1058 |
+
"eval_vitaminc-pairs_samples_per_second": 69.327,
|
| 1059 |
+
"eval_vitaminc-pairs_steps_per_second": 2.311,
|
| 1060 |
+
"step": 5145
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.8758937691521961,
|
| 1064 |
+
"eval_qnli-contrastive_loss": 0.9639705419540405,
|
| 1065 |
+
"eval_qnli-contrastive_runtime": 0.4889,
|
| 1066 |
+
"eval_qnli-contrastive_samples_per_second": 306.825,
|
| 1067 |
+
"eval_qnli-contrastive_steps_per_second": 10.228,
|
| 1068 |
+
"step": 5145
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 0.8758937691521961,
|
| 1072 |
+
"eval_scitail-pairs-qa_loss": 0.31595128774642944,
|
| 1073 |
+
"eval_scitail-pairs-qa_runtime": 1.1467,
|
| 1074 |
+
"eval_scitail-pairs-qa_samples_per_second": 130.806,
|
| 1075 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.36,
|
| 1076 |
+
"step": 5145
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.8758937691521961,
|
| 1080 |
+
"eval_scitail-pairs-pos_loss": 0.9187478423118591,
|
| 1081 |
+
"eval_scitail-pairs-pos_runtime": 2.1273,
|
| 1082 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.512,
|
| 1083 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.35,
|
| 1084 |
+
"step": 5145
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 0.8758937691521961,
|
| 1088 |
+
"eval_xsum-pairs_loss": 1.060194492340088,
|
| 1089 |
+
"eval_xsum-pairs_runtime": 2.2836,
|
| 1090 |
+
"eval_xsum-pairs_samples_per_second": 65.686,
|
| 1091 |
+
"eval_xsum-pairs_steps_per_second": 2.19,
|
| 1092 |
+
"step": 5145
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 0.8758937691521961,
|
| 1096 |
+
"eval_compression-pairs_loss": 0.41078585386276245,
|
| 1097 |
+
"eval_compression-pairs_runtime": 0.4434,
|
| 1098 |
+
"eval_compression-pairs_samples_per_second": 338.276,
|
| 1099 |
+
"eval_compression-pairs_steps_per_second": 11.276,
|
| 1100 |
+
"step": 5145
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.8758937691521961,
|
| 1104 |
+
"eval_sciq_pairs_loss": 7.577760696411133,
|
| 1105 |
+
"eval_sciq_pairs_runtime": 7.1025,
|
| 1106 |
+
"eval_sciq_pairs_samples_per_second": 21.119,
|
| 1107 |
+
"eval_sciq_pairs_steps_per_second": 0.704,
|
| 1108 |
+
"step": 5145
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.8758937691521961,
|
| 1112 |
+
"eval_qasc_pairs_loss": 6.353766918182373,
|
| 1113 |
+
"eval_qasc_pairs_runtime": 2.0113,
|
| 1114 |
+
"eval_qasc_pairs_samples_per_second": 74.58,
|
| 1115 |
+
"eval_qasc_pairs_steps_per_second": 2.486,
|
| 1116 |
+
"step": 5145
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.8758937691521961,
|
| 1120 |
+
"eval_openbookqa_pairs_loss": 3.7140932083129883,
|
| 1121 |
+
"eval_openbookqa_pairs_runtime": 0.8529,
|
| 1122 |
+
"eval_openbookqa_pairs_samples_per_second": 120.762,
|
| 1123 |
+
"eval_openbookqa_pairs_steps_per_second": 4.69,
|
| 1124 |
+
"step": 5145
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.8758937691521961,
|
| 1128 |
+
"eval_msmarco_pairs_loss": 2.3862576484680176,
|
| 1129 |
+
"eval_msmarco_pairs_runtime": 2.8953,
|
| 1130 |
+
"eval_msmarco_pairs_samples_per_second": 51.808,
|
| 1131 |
+
"eval_msmarco_pairs_steps_per_second": 1.727,
|
| 1132 |
+
"step": 5145
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 0.8758937691521961,
|
| 1136 |
+
"eval_nq_pairs_loss": 2.3543190956115723,
|
| 1137 |
+
"eval_nq_pairs_runtime": 5.0048,
|
| 1138 |
+
"eval_nq_pairs_samples_per_second": 29.971,
|
| 1139 |
+
"eval_nq_pairs_steps_per_second": 0.999,
|
| 1140 |
+
"step": 5145
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 0.8758937691521961,
|
| 1144 |
+
"eval_trivia_pairs_loss": 2.494807481765747,
|
| 1145 |
+
"eval_trivia_pairs_runtime": 9.5513,
|
| 1146 |
+
"eval_trivia_pairs_samples_per_second": 15.705,
|
| 1147 |
+
"eval_trivia_pairs_steps_per_second": 0.523,
|
| 1148 |
+
"step": 5145
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.8758937691521961,
|
| 1152 |
+
"eval_quora_pairs_loss": 0.6137441992759705,
|
| 1153 |
+
"eval_quora_pairs_runtime": 1.1541,
|
| 1154 |
+
"eval_quora_pairs_samples_per_second": 129.967,
|
| 1155 |
+
"eval_quora_pairs_steps_per_second": 4.332,
|
| 1156 |
+
"step": 5145
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 0.8758937691521961,
|
| 1160 |
+
"eval_gooaq_pairs_loss": 1.8279658555984497,
|
| 1161 |
+
"eval_gooaq_pairs_runtime": 2.0951,
|
| 1162 |
+
"eval_gooaq_pairs_samples_per_second": 71.595,
|
| 1163 |
+
"eval_gooaq_pairs_steps_per_second": 2.387,
|
| 1164 |
+
"step": 5145
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.9009193054136875,
|
| 1168 |
+
"grad_norm": 10.590804100036621,
|
| 1169 |
+
"learning_rate": 1.7997957099080695e-05,
|
| 1170 |
+
"loss": 2.2133,
|
| 1171 |
+
"step": 5292
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.9259448416751788,
|
| 1175 |
+
"grad_norm": 18.527711868286133,
|
| 1176 |
+
"learning_rate": 1.849846782431052e-05,
|
| 1177 |
+
"loss": 2.2255,
|
| 1178 |
+
"step": 5439
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.95097037793667,
|
| 1182 |
+
"grad_norm": 2.617710828781128,
|
| 1183 |
+
"learning_rate": 1.8995573714674838e-05,
|
| 1184 |
+
"loss": 2.3502,
|
| 1185 |
+
"step": 5586
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.9759959141981613,
|
| 1189 |
+
"grad_norm": 19.551551818847656,
|
| 1190 |
+
"learning_rate": 1.9496084439904668e-05,
|
| 1191 |
+
"loss": 1.8964,
|
| 1192 |
+
"step": 5733
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.0010214504596526,
|
| 1196 |
+
"grad_norm": 11.783225059509277,
|
| 1197 |
+
"learning_rate": 1.999319033026898e-05,
|
| 1198 |
+
"loss": 1.913,
|
| 1199 |
+
"step": 5880
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.0010214504596526,
|
| 1203 |
+
"eval_nli-pairs_loss": 1.9677053689956665,
|
| 1204 |
+
"eval_nli-pairs_runtime": 4.3863,
|
| 1205 |
+
"eval_nli-pairs_samples_per_second": 34.198,
|
| 1206 |
+
"eval_nli-pairs_steps_per_second": 1.14,
|
| 1207 |
+
"eval_sts-test_pearson_cosine": 0.7531824359441671,
|
| 1208 |
+
"eval_sts-test_pearson_dot": 0.602579906515822,
|
| 1209 |
+
"eval_sts-test_pearson_euclidean": 0.7486763477944213,
|
| 1210 |
+
"eval_sts-test_pearson_manhattan": 0.7566220287347274,
|
| 1211 |
+
"eval_sts-test_pearson_max": 0.7566220287347274,
|
| 1212 |
+
"eval_sts-test_spearman_cosine": 0.7387792578665129,
|
| 1213 |
+
"eval_sts-test_spearman_dot": 0.5926594656319394,
|
| 1214 |
+
"eval_sts-test_spearman_euclidean": 0.733653805383597,
|
| 1215 |
+
"eval_sts-test_spearman_manhattan": 0.7420657558603486,
|
| 1216 |
+
"eval_sts-test_spearman_max": 0.7420657558603486,
|
| 1217 |
+
"step": 5880
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.0010214504596526,
|
| 1221 |
+
"eval_vitaminc-pairs_loss": 1.4394291639328003,
|
| 1222 |
+
"eval_vitaminc-pairs_runtime": 2.2575,
|
| 1223 |
+
"eval_vitaminc-pairs_samples_per_second": 66.446,
|
| 1224 |
+
"eval_vitaminc-pairs_steps_per_second": 2.215,
|
| 1225 |
+
"step": 5880
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.0010214504596526,
|
| 1229 |
+
"eval_qnli-contrastive_loss": 0.45715218782424927,
|
| 1230 |
+
"eval_qnli-contrastive_runtime": 0.501,
|
| 1231 |
+
"eval_qnli-contrastive_samples_per_second": 299.385,
|
| 1232 |
+
"eval_qnli-contrastive_steps_per_second": 9.979,
|
| 1233 |
+
"step": 5880
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.0010214504596526,
|
| 1237 |
+
"eval_scitail-pairs-qa_loss": 0.26679515838623047,
|
| 1238 |
+
"eval_scitail-pairs-qa_runtime": 1.4342,
|
| 1239 |
+
"eval_scitail-pairs-qa_samples_per_second": 104.587,
|
| 1240 |
+
"eval_scitail-pairs-qa_steps_per_second": 3.486,
|
| 1241 |
+
"step": 5880
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 1.0010214504596526,
|
| 1245 |
+
"eval_scitail-pairs-pos_loss": 0.8628473281860352,
|
| 1246 |
+
"eval_scitail-pairs-pos_runtime": 2.3485,
|
| 1247 |
+
"eval_scitail-pairs-pos_samples_per_second": 63.871,
|
| 1248 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.129,
|
| 1249 |
+
"step": 5880
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.0010214504596526,
|
| 1253 |
+
"eval_xsum-pairs_loss": 0.9014443755149841,
|
| 1254 |
+
"eval_xsum-pairs_runtime": 2.2896,
|
| 1255 |
+
"eval_xsum-pairs_samples_per_second": 65.513,
|
| 1256 |
+
"eval_xsum-pairs_steps_per_second": 2.184,
|
| 1257 |
+
"step": 5880
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.0010214504596526,
|
| 1261 |
+
"eval_compression-pairs_loss": 0.3047434389591217,
|
| 1262 |
+
"eval_compression-pairs_runtime": 0.4852,
|
| 1263 |
+
"eval_compression-pairs_samples_per_second": 309.163,
|
| 1264 |
+
"eval_compression-pairs_steps_per_second": 10.305,
|
| 1265 |
+
"step": 5880
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 1.0010214504596526,
|
| 1269 |
+
"eval_sciq_pairs_loss": 1.091601848602295,
|
| 1270 |
+
"eval_sciq_pairs_runtime": 7.3046,
|
| 1271 |
+
"eval_sciq_pairs_samples_per_second": 20.535,
|
| 1272 |
+
"eval_sciq_pairs_steps_per_second": 0.684,
|
| 1273 |
+
"step": 5880
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.0010214504596526,
|
| 1277 |
+
"eval_qasc_pairs_loss": 5.947833061218262,
|
| 1278 |
+
"eval_qasc_pairs_runtime": 2.1787,
|
| 1279 |
+
"eval_qasc_pairs_samples_per_second": 68.849,
|
| 1280 |
+
"eval_qasc_pairs_steps_per_second": 2.295,
|
| 1281 |
+
"step": 5880
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.0010214504596526,
|
| 1285 |
+
"eval_openbookqa_pairs_loss": 3.4724366664886475,
|
| 1286 |
+
"eval_openbookqa_pairs_runtime": 0.9106,
|
| 1287 |
+
"eval_openbookqa_pairs_samples_per_second": 113.111,
|
| 1288 |
+
"eval_openbookqa_pairs_steps_per_second": 4.393,
|
| 1289 |
+
"step": 5880
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.0010214504596526,
|
| 1293 |
+
"eval_msmarco_pairs_loss": 2.1638240814208984,
|
| 1294 |
+
"eval_msmarco_pairs_runtime": 2.82,
|
| 1295 |
+
"eval_msmarco_pairs_samples_per_second": 53.191,
|
| 1296 |
+
"eval_msmarco_pairs_steps_per_second": 1.773,
|
| 1297 |
+
"step": 5880
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.0010214504596526,
|
| 1301 |
+
"eval_nq_pairs_loss": 2.110903739929199,
|
| 1302 |
+
"eval_nq_pairs_runtime": 5.2303,
|
| 1303 |
+
"eval_nq_pairs_samples_per_second": 28.679,
|
| 1304 |
+
"eval_nq_pairs_steps_per_second": 0.956,
|
| 1305 |
+
"step": 5880
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.0010214504596526,
|
| 1309 |
+
"eval_trivia_pairs_loss": 2.3711097240448,
|
| 1310 |
+
"eval_trivia_pairs_runtime": 9.6247,
|
| 1311 |
+
"eval_trivia_pairs_samples_per_second": 15.585,
|
| 1312 |
+
"eval_trivia_pairs_steps_per_second": 0.519,
|
| 1313 |
+
"step": 5880
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.0010214504596526,
|
| 1317 |
+
"eval_quora_pairs_loss": 0.5216041803359985,
|
| 1318 |
+
"eval_quora_pairs_runtime": 1.3072,
|
| 1319 |
+
"eval_quora_pairs_samples_per_second": 114.749,
|
| 1320 |
+
"eval_quora_pairs_steps_per_second": 3.825,
|
| 1321 |
+
"step": 5880
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.0010214504596526,
|
| 1325 |
+
"eval_gooaq_pairs_loss": 1.7041363716125488,
|
| 1326 |
+
"eval_gooaq_pairs_runtime": 2.0973,
|
| 1327 |
+
"eval_gooaq_pairs_samples_per_second": 71.521,
|
| 1328 |
+
"eval_gooaq_pairs_steps_per_second": 2.384,
|
| 1329 |
+
"step": 5880
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.026046986721144,
|
| 1333 |
+
"grad_norm": 17.308378219604492,
|
| 1334 |
+
"learning_rate": 2.0493701055498808e-05,
|
| 1335 |
+
"loss": 1.7772,
|
| 1336 |
+
"step": 6027
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 1.0510725229826354,
|
| 1340 |
+
"grad_norm": 20.248981475830078,
|
| 1341 |
+
"learning_rate": 2.0994211780728634e-05,
|
| 1342 |
+
"loss": 1.9079,
|
| 1343 |
+
"step": 6174
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.0760980592441267,
|
| 1347 |
+
"grad_norm": 6.012618064880371,
|
| 1348 |
+
"learning_rate": 2.1494722505958464e-05,
|
| 1349 |
+
"loss": 1.8657,
|
| 1350 |
+
"step": 6321
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 1.101123595505618,
|
| 1354 |
+
"grad_norm": 1.1185024976730347,
|
| 1355 |
+
"learning_rate": 2.1995233231188288e-05,
|
| 1356 |
+
"loss": 1.7144,
|
| 1357 |
+
"step": 6468
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.1261491317671093,
|
| 1361 |
+
"grad_norm": 1.2436251640319824,
|
| 1362 |
+
"learning_rate": 2.2495743956418114e-05,
|
| 1363 |
+
"loss": 1.7661,
|
| 1364 |
+
"step": 6615
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.1261491317671093,
|
| 1368 |
+
"eval_nli-pairs_loss": 1.7907973527908325,
|
| 1369 |
+
"eval_nli-pairs_runtime": 4.0147,
|
| 1370 |
+
"eval_nli-pairs_samples_per_second": 37.363,
|
| 1371 |
+
"eval_nli-pairs_steps_per_second": 1.245,
|
| 1372 |
+
"eval_sts-test_pearson_cosine": 0.755444461779583,
|
| 1373 |
+
"eval_sts-test_pearson_dot": 0.5833168145328357,
|
| 1374 |
+
"eval_sts-test_pearson_euclidean": 0.7437155007996056,
|
| 1375 |
+
"eval_sts-test_pearson_manhattan": 0.7524938984567344,
|
| 1376 |
+
"eval_sts-test_pearson_max": 0.755444461779583,
|
| 1377 |
+
"eval_sts-test_spearman_cosine": 0.7446166596886566,
|
| 1378 |
+
"eval_sts-test_spearman_dot": 0.5792340720766105,
|
| 1379 |
+
"eval_sts-test_spearman_euclidean": 0.7317285388028532,
|
| 1380 |
+
"eval_sts-test_spearman_manhattan": 0.7401637904976945,
|
| 1381 |
+
"eval_sts-test_spearman_max": 0.7446166596886566,
|
| 1382 |
+
"step": 6615
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.1261491317671093,
|
| 1386 |
+
"eval_vitaminc-pairs_loss": 1.3403607606887817,
|
| 1387 |
+
"eval_vitaminc-pairs_runtime": 2.168,
|
| 1388 |
+
"eval_vitaminc-pairs_samples_per_second": 69.189,
|
| 1389 |
+
"eval_vitaminc-pairs_steps_per_second": 2.306,
|
| 1390 |
+
"step": 6615
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 1.1261491317671093,
|
| 1394 |
+
"eval_qnli-contrastive_loss": 0.2736852467060089,
|
| 1395 |
+
"eval_qnli-contrastive_runtime": 0.4913,
|
| 1396 |
+
"eval_qnli-contrastive_samples_per_second": 305.336,
|
| 1397 |
+
"eval_qnli-contrastive_steps_per_second": 10.178,
|
| 1398 |
+
"step": 6615
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 1.1261491317671093,
|
| 1402 |
+
"eval_scitail-pairs-qa_loss": 0.22441554069519043,
|
| 1403 |
+
"eval_scitail-pairs-qa_runtime": 1.1614,
|
| 1404 |
+
"eval_scitail-pairs-qa_samples_per_second": 129.152,
|
| 1405 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.305,
|
| 1406 |
+
"step": 6615
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.1261491317671093,
|
| 1410 |
+
"eval_scitail-pairs-pos_loss": 0.7723743915557861,
|
| 1411 |
+
"eval_scitail-pairs-pos_runtime": 2.1567,
|
| 1412 |
+
"eval_scitail-pairs-pos_samples_per_second": 69.55,
|
| 1413 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.318,
|
| 1414 |
+
"step": 6615
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 1.1261491317671093,
|
| 1418 |
+
"eval_xsum-pairs_loss": 0.8370540142059326,
|
| 1419 |
+
"eval_xsum-pairs_runtime": 2.2569,
|
| 1420 |
+
"eval_xsum-pairs_samples_per_second": 66.463,
|
| 1421 |
+
"eval_xsum-pairs_steps_per_second": 2.215,
|
| 1422 |
+
"step": 6615
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 1.1261491317671093,
|
| 1426 |
+
"eval_compression-pairs_loss": 0.265947163105011,
|
| 1427 |
+
"eval_compression-pairs_runtime": 0.4431,
|
| 1428 |
+
"eval_compression-pairs_samples_per_second": 338.529,
|
| 1429 |
+
"eval_compression-pairs_steps_per_second": 11.284,
|
| 1430 |
+
"step": 6615
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.1261491317671093,
|
| 1434 |
+
"eval_sciq_pairs_loss": 0.9383512735366821,
|
| 1435 |
+
"eval_sciq_pairs_runtime": 7.1464,
|
| 1436 |
+
"eval_sciq_pairs_samples_per_second": 20.99,
|
| 1437 |
+
"eval_sciq_pairs_steps_per_second": 0.7,
|
| 1438 |
+
"step": 6615
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.1261491317671093,
|
| 1442 |
+
"eval_qasc_pairs_loss": 5.753899097442627,
|
| 1443 |
+
"eval_qasc_pairs_runtime": 2.0099,
|
| 1444 |
+
"eval_qasc_pairs_samples_per_second": 74.63,
|
| 1445 |
+
"eval_qasc_pairs_steps_per_second": 2.488,
|
| 1446 |
+
"step": 6615
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 1.1261491317671093,
|
| 1450 |
+
"eval_openbookqa_pairs_loss": 3.3517918586730957,
|
| 1451 |
+
"eval_openbookqa_pairs_runtime": 0.8594,
|
| 1452 |
+
"eval_openbookqa_pairs_samples_per_second": 119.858,
|
| 1453 |
+
"eval_openbookqa_pairs_steps_per_second": 4.655,
|
| 1454 |
+
"step": 6615
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 1.1261491317671093,
|
| 1458 |
+
"eval_msmarco_pairs_loss": 2.044360399246216,
|
| 1459 |
+
"eval_msmarco_pairs_runtime": 2.7431,
|
| 1460 |
+
"eval_msmarco_pairs_samples_per_second": 54.682,
|
| 1461 |
+
"eval_msmarco_pairs_steps_per_second": 1.823,
|
| 1462 |
+
"step": 6615
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.1261491317671093,
|
| 1466 |
+
"eval_nq_pairs_loss": 1.9409464597702026,
|
| 1467 |
+
"eval_nq_pairs_runtime": 5.028,
|
| 1468 |
+
"eval_nq_pairs_samples_per_second": 29.833,
|
| 1469 |
+
"eval_nq_pairs_steps_per_second": 0.994,
|
| 1470 |
+
"step": 6615
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 1.1261491317671093,
|
| 1474 |
+
"eval_trivia_pairs_loss": 2.369060754776001,
|
| 1475 |
+
"eval_trivia_pairs_runtime": 9.5137,
|
| 1476 |
+
"eval_trivia_pairs_samples_per_second": 15.767,
|
| 1477 |
+
"eval_trivia_pairs_steps_per_second": 0.526,
|
| 1478 |
+
"step": 6615
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.1261491317671093,
|
| 1482 |
+
"eval_quora_pairs_loss": 0.47849634289741516,
|
| 1483 |
+
"eval_quora_pairs_runtime": 1.1413,
|
| 1484 |
+
"eval_quora_pairs_samples_per_second": 131.424,
|
| 1485 |
+
"eval_quora_pairs_steps_per_second": 4.381,
|
| 1486 |
+
"step": 6615
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.1261491317671093,
|
| 1490 |
+
"eval_gooaq_pairs_loss": 1.5795674324035645,
|
| 1491 |
+
"eval_gooaq_pairs_runtime": 2.0155,
|
| 1492 |
+
"eval_gooaq_pairs_samples_per_second": 74.422,
|
| 1493 |
+
"eval_gooaq_pairs_steps_per_second": 2.481,
|
| 1494 |
+
"step": 6615
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.1511746680286006,
|
| 1498 |
+
"grad_norm": 20.95261001586914,
|
| 1499 |
+
"learning_rate": 2.299625468164794e-05,
|
| 1500 |
+
"loss": 1.8066,
|
| 1501 |
+
"step": 6762
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.1762002042900919,
|
| 1505 |
+
"grad_norm": 20.31597900390625,
|
| 1506 |
+
"learning_rate": 2.3496765406877764e-05,
|
| 1507 |
+
"loss": 1.7438,
|
| 1508 |
+
"step": 6909
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.2012257405515832,
|
| 1512 |
+
"grad_norm": 28.363882064819336,
|
| 1513 |
+
"learning_rate": 2.399727613210759e-05,
|
| 1514 |
+
"loss": 2.0231,
|
| 1515 |
+
"step": 7056
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.2262512768130747,
|
| 1519 |
+
"grad_norm": 14.403656959533691,
|
| 1520 |
+
"learning_rate": 2.449778685733742e-05,
|
| 1521 |
+
"loss": 1.8966,
|
| 1522 |
+
"step": 7203
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.251276813074566,
|
| 1526 |
+
"grad_norm": 17.73562240600586,
|
| 1527 |
+
"learning_rate": 2.4998297582567248e-05,
|
| 1528 |
+
"loss": 1.7958,
|
| 1529 |
+
"step": 7350
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.251276813074566,
|
| 1533 |
+
"eval_nli-pairs_loss": 1.5906368494033813,
|
| 1534 |
+
"eval_nli-pairs_runtime": 4.0261,
|
| 1535 |
+
"eval_nli-pairs_samples_per_second": 37.257,
|
| 1536 |
+
"eval_nli-pairs_steps_per_second": 1.242,
|
| 1537 |
+
"eval_sts-test_pearson_cosine": 0.7626661521495873,
|
| 1538 |
+
"eval_sts-test_pearson_dot": 0.5632604768989181,
|
| 1539 |
+
"eval_sts-test_pearson_euclidean": 0.7370060575260952,
|
| 1540 |
+
"eval_sts-test_pearson_manhattan": 0.7472706980613159,
|
| 1541 |
+
"eval_sts-test_pearson_max": 0.7626661521495873,
|
| 1542 |
+
"eval_sts-test_spearman_cosine": 0.7535266725567149,
|
| 1543 |
+
"eval_sts-test_spearman_dot": 0.5848997224802808,
|
| 1544 |
+
"eval_sts-test_spearman_euclidean": 0.7290608032903477,
|
| 1545 |
+
"eval_sts-test_spearman_manhattan": 0.739032087078249,
|
| 1546 |
+
"eval_sts-test_spearman_max": 0.7535266725567149,
|
| 1547 |
+
"step": 7350
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 1.251276813074566,
|
| 1551 |
+
"eval_vitaminc-pairs_loss": 1.222551941871643,
|
| 1552 |
+
"eval_vitaminc-pairs_runtime": 2.1784,
|
| 1553 |
+
"eval_vitaminc-pairs_samples_per_second": 68.857,
|
| 1554 |
+
"eval_vitaminc-pairs_steps_per_second": 2.295,
|
| 1555 |
+
"step": 7350
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.251276813074566,
|
| 1559 |
+
"eval_qnli-contrastive_loss": 0.3951484262943268,
|
| 1560 |
+
"eval_qnli-contrastive_runtime": 0.4916,
|
| 1561 |
+
"eval_qnli-contrastive_samples_per_second": 305.11,
|
| 1562 |
+
"eval_qnli-contrastive_steps_per_second": 10.17,
|
| 1563 |
+
"step": 7350
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.251276813074566,
|
| 1567 |
+
"eval_scitail-pairs-qa_loss": 0.17783091962337494,
|
| 1568 |
+
"eval_scitail-pairs-qa_runtime": 1.1549,
|
| 1569 |
+
"eval_scitail-pairs-qa_samples_per_second": 129.88,
|
| 1570 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.329,
|
| 1571 |
+
"step": 7350
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.251276813074566,
|
| 1575 |
+
"eval_scitail-pairs-pos_loss": 0.7214661836624146,
|
| 1576 |
+
"eval_scitail-pairs-pos_runtime": 2.132,
|
| 1577 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.357,
|
| 1578 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.345,
|
| 1579 |
+
"step": 7350
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.251276813074566,
|
| 1583 |
+
"eval_xsum-pairs_loss": 0.7919928431510925,
|
| 1584 |
+
"eval_xsum-pairs_runtime": 2.2579,
|
| 1585 |
+
"eval_xsum-pairs_samples_per_second": 66.432,
|
| 1586 |
+
"eval_xsum-pairs_steps_per_second": 2.214,
|
| 1587 |
+
"step": 7350
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 1.251276813074566,
|
| 1591 |
+
"eval_compression-pairs_loss": 0.24975377321243286,
|
| 1592 |
+
"eval_compression-pairs_runtime": 0.447,
|
| 1593 |
+
"eval_compression-pairs_samples_per_second": 335.534,
|
| 1594 |
+
"eval_compression-pairs_steps_per_second": 11.184,
|
| 1595 |
+
"step": 7350
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.251276813074566,
|
| 1599 |
+
"eval_sciq_pairs_loss": 0.8343773484230042,
|
| 1600 |
+
"eval_sciq_pairs_runtime": 7.1288,
|
| 1601 |
+
"eval_sciq_pairs_samples_per_second": 21.042,
|
| 1602 |
+
"eval_sciq_pairs_steps_per_second": 0.701,
|
| 1603 |
+
"step": 7350
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.251276813074566,
|
| 1607 |
+
"eval_qasc_pairs_loss": 5.4840240478515625,
|
| 1608 |
+
"eval_qasc_pairs_runtime": 2.025,
|
| 1609 |
+
"eval_qasc_pairs_samples_per_second": 74.074,
|
| 1610 |
+
"eval_qasc_pairs_steps_per_second": 2.469,
|
| 1611 |
+
"step": 7350
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.251276813074566,
|
| 1615 |
+
"eval_openbookqa_pairs_loss": 3.1631176471710205,
|
| 1616 |
+
"eval_openbookqa_pairs_runtime": 0.8612,
|
| 1617 |
+
"eval_openbookqa_pairs_samples_per_second": 119.598,
|
| 1618 |
+
"eval_openbookqa_pairs_steps_per_second": 4.645,
|
| 1619 |
+
"step": 7350
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.251276813074566,
|
| 1623 |
+
"eval_msmarco_pairs_loss": 1.8952231407165527,
|
| 1624 |
+
"eval_msmarco_pairs_runtime": 2.7585,
|
| 1625 |
+
"eval_msmarco_pairs_samples_per_second": 54.378,
|
| 1626 |
+
"eval_msmarco_pairs_steps_per_second": 1.813,
|
| 1627 |
+
"step": 7350
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.251276813074566,
|
| 1631 |
+
"eval_nq_pairs_loss": 1.6934970617294312,
|
| 1632 |
+
"eval_nq_pairs_runtime": 5.0253,
|
| 1633 |
+
"eval_nq_pairs_samples_per_second": 29.849,
|
| 1634 |
+
"eval_nq_pairs_steps_per_second": 0.995,
|
| 1635 |
+
"step": 7350
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 1.251276813074566,
|
| 1639 |
+
"eval_trivia_pairs_loss": 1.9966663122177124,
|
| 1640 |
+
"eval_trivia_pairs_runtime": 9.5675,
|
| 1641 |
+
"eval_trivia_pairs_samples_per_second": 15.678,
|
| 1642 |
+
"eval_trivia_pairs_steps_per_second": 0.523,
|
| 1643 |
+
"step": 7350
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 1.251276813074566,
|
| 1647 |
+
"eval_quora_pairs_loss": 0.405385285615921,
|
| 1648 |
+
"eval_quora_pairs_runtime": 1.1432,
|
| 1649 |
+
"eval_quora_pairs_samples_per_second": 131.209,
|
| 1650 |
+
"eval_quora_pairs_steps_per_second": 4.374,
|
| 1651 |
+
"step": 7350
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.251276813074566,
|
| 1655 |
+
"eval_gooaq_pairs_loss": 1.3951071500778198,
|
| 1656 |
+
"eval_gooaq_pairs_runtime": 2.038,
|
| 1657 |
+
"eval_gooaq_pairs_samples_per_second": 73.601,
|
| 1658 |
+
"eval_gooaq_pairs_steps_per_second": 2.453,
|
| 1659 |
+
"step": 7350
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 1.2763023493360572,
|
| 1663 |
+
"grad_norm": 21.254159927368164,
|
| 1664 |
+
"learning_rate": 2.549880830779707e-05,
|
| 1665 |
+
"loss": 1.5109,
|
| 1666 |
+
"step": 7497
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 1.3013278855975485,
|
| 1670 |
+
"grad_norm": 20.08012580871582,
|
| 1671 |
+
"learning_rate": 2.5999319033026898e-05,
|
| 1672 |
+
"loss": 1.8119,
|
| 1673 |
+
"step": 7644
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.3263534218590398,
|
| 1677 |
+
"grad_norm": 0.6448306441307068,
|
| 1678 |
+
"learning_rate": 2.6499829758256724e-05,
|
| 1679 |
+
"loss": 1.6833,
|
| 1680 |
+
"step": 7791
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 1.351378958120531,
|
| 1684 |
+
"grad_norm": 16.65821647644043,
|
| 1685 |
+
"learning_rate": 2.7000340483486554e-05,
|
| 1686 |
+
"loss": 1.5917,
|
| 1687 |
+
"step": 7938
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.3764044943820224,
|
| 1691 |
+
"grad_norm": 14.949362754821777,
|
| 1692 |
+
"learning_rate": 2.7500851208716378e-05,
|
| 1693 |
+
"loss": 1.809,
|
| 1694 |
+
"step": 8085
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 1.3764044943820224,
|
| 1698 |
+
"eval_nli-pairs_loss": 1.5967836380004883,
|
| 1699 |
+
"eval_nli-pairs_runtime": 4.0496,
|
| 1700 |
+
"eval_nli-pairs_samples_per_second": 37.041,
|
| 1701 |
+
"eval_nli-pairs_steps_per_second": 1.235,
|
| 1702 |
+
"eval_sts-test_pearson_cosine": 0.7653416933913197,
|
| 1703 |
+
"eval_sts-test_pearson_dot": 0.5401711611334493,
|
| 1704 |
+
"eval_sts-test_pearson_euclidean": 0.7529907774019836,
|
| 1705 |
+
"eval_sts-test_pearson_manhattan": 0.7605105025260754,
|
| 1706 |
+
"eval_sts-test_pearson_max": 0.7653416933913197,
|
| 1707 |
+
"eval_sts-test_spearman_cosine": 0.7593865234485873,
|
| 1708 |
+
"eval_sts-test_spearman_dot": 0.5559615063301898,
|
| 1709 |
+
"eval_sts-test_spearman_euclidean": 0.7436431053840061,
|
| 1710 |
+
"eval_sts-test_spearman_manhattan": 0.7515978828464567,
|
| 1711 |
+
"eval_sts-test_spearman_max": 0.7593865234485873,
|
| 1712 |
+
"step": 8085
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 1.3764044943820224,
|
| 1716 |
+
"eval_vitaminc-pairs_loss": 1.1434590816497803,
|
| 1717 |
+
"eval_vitaminc-pairs_runtime": 2.2066,
|
| 1718 |
+
"eval_vitaminc-pairs_samples_per_second": 67.977,
|
| 1719 |
+
"eval_vitaminc-pairs_steps_per_second": 2.266,
|
| 1720 |
+
"step": 8085
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 1.3764044943820224,
|
| 1724 |
+
"eval_qnli-contrastive_loss": 0.3819103538990021,
|
| 1725 |
+
"eval_qnli-contrastive_runtime": 0.4972,
|
| 1726 |
+
"eval_qnli-contrastive_samples_per_second": 301.706,
|
| 1727 |
+
"eval_qnli-contrastive_steps_per_second": 10.057,
|
| 1728 |
+
"step": 8085
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 1.3764044943820224,
|
| 1732 |
+
"eval_scitail-pairs-qa_loss": 0.15774373710155487,
|
| 1733 |
+
"eval_scitail-pairs-qa_runtime": 1.1704,
|
| 1734 |
+
"eval_scitail-pairs-qa_samples_per_second": 128.161,
|
| 1735 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.272,
|
| 1736 |
+
"step": 8085
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 1.3764044943820224,
|
| 1740 |
+
"eval_scitail-pairs-pos_loss": 0.6571963429450989,
|
| 1741 |
+
"eval_scitail-pairs-pos_runtime": 2.1634,
|
| 1742 |
+
"eval_scitail-pairs-pos_samples_per_second": 69.335,
|
| 1743 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.311,
|
| 1744 |
+
"step": 8085
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 1.3764044943820224,
|
| 1748 |
+
"eval_xsum-pairs_loss": 0.7028753757476807,
|
| 1749 |
+
"eval_xsum-pairs_runtime": 2.2608,
|
| 1750 |
+
"eval_xsum-pairs_samples_per_second": 66.347,
|
| 1751 |
+
"eval_xsum-pairs_steps_per_second": 2.212,
|
| 1752 |
+
"step": 8085
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 1.3764044943820224,
|
| 1756 |
+
"eval_compression-pairs_loss": 0.23010987043380737,
|
| 1757 |
+
"eval_compression-pairs_runtime": 0.4514,
|
| 1758 |
+
"eval_compression-pairs_samples_per_second": 332.284,
|
| 1759 |
+
"eval_compression-pairs_steps_per_second": 11.076,
|
| 1760 |
+
"step": 8085
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 1.3764044943820224,
|
| 1764 |
+
"eval_sciq_pairs_loss": 0.799666702747345,
|
| 1765 |
+
"eval_sciq_pairs_runtime": 7.1816,
|
| 1766 |
+
"eval_sciq_pairs_samples_per_second": 20.887,
|
| 1767 |
+
"eval_sciq_pairs_steps_per_second": 0.696,
|
| 1768 |
+
"step": 8085
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.3764044943820224,
|
| 1772 |
+
"eval_qasc_pairs_loss": 5.433376789093018,
|
| 1773 |
+
"eval_qasc_pairs_runtime": 2.0592,
|
| 1774 |
+
"eval_qasc_pairs_samples_per_second": 72.843,
|
| 1775 |
+
"eval_qasc_pairs_steps_per_second": 2.428,
|
| 1776 |
+
"step": 8085
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 1.3764044943820224,
|
| 1780 |
+
"eval_openbookqa_pairs_loss": 2.9010672569274902,
|
| 1781 |
+
"eval_openbookqa_pairs_runtime": 0.865,
|
| 1782 |
+
"eval_openbookqa_pairs_samples_per_second": 119.074,
|
| 1783 |
+
"eval_openbookqa_pairs_steps_per_second": 4.624,
|
| 1784 |
+
"step": 8085
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 1.3764044943820224,
|
| 1788 |
+
"eval_msmarco_pairs_loss": 1.7567836046218872,
|
| 1789 |
+
"eval_msmarco_pairs_runtime": 2.7812,
|
| 1790 |
+
"eval_msmarco_pairs_samples_per_second": 53.933,
|
| 1791 |
+
"eval_msmarco_pairs_steps_per_second": 1.798,
|
| 1792 |
+
"step": 8085
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 1.3764044943820224,
|
| 1796 |
+
"eval_nq_pairs_loss": 1.5407707691192627,
|
| 1797 |
+
"eval_nq_pairs_runtime": 5.0607,
|
| 1798 |
+
"eval_nq_pairs_samples_per_second": 29.64,
|
| 1799 |
+
"eval_nq_pairs_steps_per_second": 0.988,
|
| 1800 |
+
"step": 8085
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 1.3764044943820224,
|
| 1804 |
+
"eval_trivia_pairs_loss": 1.8419283628463745,
|
| 1805 |
+
"eval_trivia_pairs_runtime": 9.5535,
|
| 1806 |
+
"eval_trivia_pairs_samples_per_second": 15.701,
|
| 1807 |
+
"eval_trivia_pairs_steps_per_second": 0.523,
|
| 1808 |
+
"step": 8085
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 1.3764044943820224,
|
| 1812 |
+
"eval_quora_pairs_loss": 0.3561370372772217,
|
| 1813 |
+
"eval_quora_pairs_runtime": 1.2005,
|
| 1814 |
+
"eval_quora_pairs_samples_per_second": 124.946,
|
| 1815 |
+
"eval_quora_pairs_steps_per_second": 4.165,
|
| 1816 |
+
"step": 8085
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 1.3764044943820224,
|
| 1820 |
+
"eval_gooaq_pairs_loss": 1.1745914220809937,
|
| 1821 |
+
"eval_gooaq_pairs_runtime": 2.0463,
|
| 1822 |
+
"eval_gooaq_pairs_samples_per_second": 73.305,
|
| 1823 |
+
"eval_gooaq_pairs_steps_per_second": 2.443,
|
| 1824 |
+
"step": 8085
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 1.401430030643514,
|
| 1828 |
+
"grad_norm": 14.31106185913086,
|
| 1829 |
+
"learning_rate": 2.8001361933946204e-05,
|
| 1830 |
+
"loss": 1.5561,
|
| 1831 |
+
"step": 8232
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 1.4264555669050052,
|
| 1835 |
+
"grad_norm": 11.82392692565918,
|
| 1836 |
+
"learning_rate": 2.850187265917603e-05,
|
| 1837 |
+
"loss": 1.5325,
|
| 1838 |
+
"step": 8379
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 1.4514811031664965,
|
| 1842 |
+
"grad_norm": 21.716449737548828,
|
| 1843 |
+
"learning_rate": 2.9002383384405858e-05,
|
| 1844 |
+
"loss": 1.5085,
|
| 1845 |
+
"step": 8526
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 1.4765066394279878,
|
| 1849 |
+
"grad_norm": 6.5607147216796875,
|
| 1850 |
+
"learning_rate": 2.950289410963568e-05,
|
| 1851 |
+
"loss": 1.5634,
|
| 1852 |
+
"step": 8673
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 1.501532175689479,
|
| 1856 |
+
"grad_norm": 8.737595558166504,
|
| 1857 |
+
"learning_rate": 2.9999998423842776e-05,
|
| 1858 |
+
"loss": 1.3857,
|
| 1859 |
+
"step": 8820
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.501532175689479,
|
| 1863 |
+
"eval_nli-pairs_loss": 1.454946517944336,
|
| 1864 |
+
"eval_nli-pairs_runtime": 4.3786,
|
| 1865 |
+
"eval_nli-pairs_samples_per_second": 34.257,
|
| 1866 |
+
"eval_nli-pairs_steps_per_second": 1.142,
|
| 1867 |
+
"eval_sts-test_pearson_cosine": 0.758856517299588,
|
| 1868 |
+
"eval_sts-test_pearson_dot": 0.5254244903711445,
|
| 1869 |
+
"eval_sts-test_pearson_euclidean": 0.7467439510002647,
|
| 1870 |
+
"eval_sts-test_pearson_manhattan": 0.7525779346304055,
|
| 1871 |
+
"eval_sts-test_pearson_max": 0.758856517299588,
|
| 1872 |
+
"eval_sts-test_spearman_cosine": 0.7596605816446022,
|
| 1873 |
+
"eval_sts-test_spearman_dot": 0.5600186533991508,
|
| 1874 |
+
"eval_sts-test_spearman_euclidean": 0.7367598380547504,
|
| 1875 |
+
"eval_sts-test_spearman_manhattan": 0.7440123650923844,
|
| 1876 |
+
"eval_sts-test_spearman_max": 0.7596605816446022,
|
| 1877 |
+
"step": 8820
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 1.501532175689479,
|
| 1881 |
+
"eval_vitaminc-pairs_loss": 1.1383781433105469,
|
| 1882 |
+
"eval_vitaminc-pairs_runtime": 2.3314,
|
| 1883 |
+
"eval_vitaminc-pairs_samples_per_second": 64.34,
|
| 1884 |
+
"eval_vitaminc-pairs_steps_per_second": 2.145,
|
| 1885 |
+
"step": 8820
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.501532175689479,
|
| 1889 |
+
"eval_qnli-contrastive_loss": 0.32092100381851196,
|
| 1890 |
+
"eval_qnli-contrastive_runtime": 0.5002,
|
| 1891 |
+
"eval_qnli-contrastive_samples_per_second": 299.881,
|
| 1892 |
+
"eval_qnli-contrastive_steps_per_second": 9.996,
|
| 1893 |
+
"step": 8820
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 1.501532175689479,
|
| 1897 |
+
"eval_scitail-pairs-qa_loss": 0.14513270556926727,
|
| 1898 |
+
"eval_scitail-pairs-qa_runtime": 1.5154,
|
| 1899 |
+
"eval_scitail-pairs-qa_samples_per_second": 98.985,
|
| 1900 |
+
"eval_scitail-pairs-qa_steps_per_second": 3.3,
|
| 1901 |
+
"step": 8820
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 1.501532175689479,
|
| 1905 |
+
"eval_scitail-pairs-pos_loss": 0.6857669353485107,
|
| 1906 |
+
"eval_scitail-pairs-pos_runtime": 2.4178,
|
| 1907 |
+
"eval_scitail-pairs-pos_samples_per_second": 62.041,
|
| 1908 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.068,
|
| 1909 |
+
"step": 8820
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 1.501532175689479,
|
| 1913 |
+
"eval_xsum-pairs_loss": 0.683724045753479,
|
| 1914 |
+
"eval_xsum-pairs_runtime": 2.2766,
|
| 1915 |
+
"eval_xsum-pairs_samples_per_second": 65.887,
|
| 1916 |
+
"eval_xsum-pairs_steps_per_second": 2.196,
|
| 1917 |
+
"step": 8820
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 1.501532175689479,
|
| 1921 |
+
"eval_compression-pairs_loss": 0.20896266400814056,
|
| 1922 |
+
"eval_compression-pairs_runtime": 0.4683,
|
| 1923 |
+
"eval_compression-pairs_samples_per_second": 320.274,
|
| 1924 |
+
"eval_compression-pairs_steps_per_second": 10.676,
|
| 1925 |
+
"step": 8820
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 1.501532175689479,
|
| 1929 |
+
"eval_sciq_pairs_loss": 0.7911179661750793,
|
| 1930 |
+
"eval_sciq_pairs_runtime": 7.3506,
|
| 1931 |
+
"eval_sciq_pairs_samples_per_second": 20.407,
|
| 1932 |
+
"eval_sciq_pairs_steps_per_second": 0.68,
|
| 1933 |
+
"step": 8820
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 1.501532175689479,
|
| 1937 |
+
"eval_qasc_pairs_loss": 5.3092241287231445,
|
| 1938 |
+
"eval_qasc_pairs_runtime": 2.1926,
|
| 1939 |
+
"eval_qasc_pairs_samples_per_second": 68.411,
|
| 1940 |
+
"eval_qasc_pairs_steps_per_second": 2.28,
|
| 1941 |
+
"step": 8820
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 1.501532175689479,
|
| 1945 |
+
"eval_openbookqa_pairs_loss": 2.923464298248291,
|
| 1946 |
+
"eval_openbookqa_pairs_runtime": 0.963,
|
| 1947 |
+
"eval_openbookqa_pairs_samples_per_second": 106.961,
|
| 1948 |
+
"eval_openbookqa_pairs_steps_per_second": 4.154,
|
| 1949 |
+
"step": 8820
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 1.501532175689479,
|
| 1953 |
+
"eval_msmarco_pairs_loss": 1.674107313156128,
|
| 1954 |
+
"eval_msmarco_pairs_runtime": 2.8516,
|
| 1955 |
+
"eval_msmarco_pairs_samples_per_second": 52.602,
|
| 1956 |
+
"eval_msmarco_pairs_steps_per_second": 1.753,
|
| 1957 |
+
"step": 8820
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 1.501532175689479,
|
| 1961 |
+
"eval_nq_pairs_loss": 1.4419037103652954,
|
| 1962 |
+
"eval_nq_pairs_runtime": 5.1485,
|
| 1963 |
+
"eval_nq_pairs_samples_per_second": 29.135,
|
| 1964 |
+
"eval_nq_pairs_steps_per_second": 0.971,
|
| 1965 |
+
"step": 8820
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.501532175689479,
|
| 1969 |
+
"eval_trivia_pairs_loss": 1.7546964883804321,
|
| 1970 |
+
"eval_trivia_pairs_runtime": 9.6901,
|
| 1971 |
+
"eval_trivia_pairs_samples_per_second": 15.48,
|
| 1972 |
+
"eval_trivia_pairs_steps_per_second": 0.516,
|
| 1973 |
+
"step": 8820
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 1.501532175689479,
|
| 1977 |
+
"eval_quora_pairs_loss": 0.31785744428634644,
|
| 1978 |
+
"eval_quora_pairs_runtime": 1.2699,
|
| 1979 |
+
"eval_quora_pairs_samples_per_second": 118.117,
|
| 1980 |
+
"eval_quora_pairs_steps_per_second": 3.937,
|
| 1981 |
+
"step": 8820
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 1.501532175689479,
|
| 1985 |
+
"eval_gooaq_pairs_loss": 1.1328644752502441,
|
| 1986 |
+
"eval_gooaq_pairs_runtime": 2.1292,
|
| 1987 |
+
"eval_gooaq_pairs_samples_per_second": 70.448,
|
| 1988 |
+
"eval_gooaq_pairs_steps_per_second": 2.348,
|
| 1989 |
+
"step": 8820
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 1.5265577119509703,
|
| 1993 |
+
"grad_norm": 15.168123245239258,
|
| 1994 |
+
"learning_rate": 2.9965489092992677e-05,
|
| 1995 |
+
"loss": 1.6167,
|
| 1996 |
+
"step": 8967
|
| 1997 |
+
},
|
| 1998 |
+
{
|
| 1999 |
+
"epoch": 1.5515832482124616,
|
| 2000 |
+
"grad_norm": 29.32268524169922,
|
| 2001 |
+
"learning_rate": 2.9863043834895476e-05,
|
| 2002 |
+
"loss": 1.6664,
|
| 2003 |
+
"step": 9114
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 1.5766087844739531,
|
| 2007 |
+
"grad_norm": 2.5159287452697754,
|
| 2008 |
+
"learning_rate": 2.9693127700413034e-05,
|
| 2009 |
+
"loss": 1.4785,
|
| 2010 |
+
"step": 9261
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 1.6016343207354442,
|
| 2014 |
+
"grad_norm": 17.4219970703125,
|
| 2015 |
+
"learning_rate": 2.9456512024854113e-05,
|
| 2016 |
+
"loss": 1.5881,
|
| 2017 |
+
"step": 9408
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 1.6266598569969357,
|
| 2021 |
+
"grad_norm": 15.60139274597168,
|
| 2022 |
+
"learning_rate": 2.915427092649312e-05,
|
| 2023 |
+
"loss": 1.3379,
|
| 2024 |
+
"step": 9555
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 1.6266598569969357,
|
| 2028 |
+
"eval_nli-pairs_loss": 1.3879741430282593,
|
| 2029 |
+
"eval_nli-pairs_runtime": 4.1363,
|
| 2030 |
+
"eval_nli-pairs_samples_per_second": 36.264,
|
| 2031 |
+
"eval_nli-pairs_steps_per_second": 1.209,
|
| 2032 |
+
"eval_sts-test_pearson_cosine": 0.7733483283639441,
|
| 2033 |
+
"eval_sts-test_pearson_dot": 0.5424296843493538,
|
| 2034 |
+
"eval_sts-test_pearson_euclidean": 0.7555770040784449,
|
| 2035 |
+
"eval_sts-test_pearson_manhattan": 0.7604742759594404,
|
| 2036 |
+
"eval_sts-test_pearson_max": 0.7733483283639441,
|
| 2037 |
+
"eval_sts-test_spearman_cosine": 0.779671933510953,
|
| 2038 |
+
"eval_sts-test_spearman_dot": 0.5784449139725663,
|
| 2039 |
+
"eval_sts-test_spearman_euclidean": 0.7515003599642571,
|
| 2040 |
+
"eval_sts-test_spearman_manhattan": 0.7568440288585417,
|
| 2041 |
+
"eval_sts-test_spearman_max": 0.779671933510953,
|
| 2042 |
+
"step": 9555
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 1.6266598569969357,
|
| 2046 |
+
"eval_vitaminc-pairs_loss": 0.9942379593849182,
|
| 2047 |
+
"eval_vitaminc-pairs_runtime": 2.2185,
|
| 2048 |
+
"eval_vitaminc-pairs_samples_per_second": 67.613,
|
| 2049 |
+
"eval_vitaminc-pairs_steps_per_second": 2.254,
|
| 2050 |
+
"step": 9555
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 1.6266598569969357,
|
| 2054 |
+
"eval_qnli-contrastive_loss": 0.2178214192390442,
|
| 2055 |
+
"eval_qnli-contrastive_runtime": 0.4991,
|
| 2056 |
+
"eval_qnli-contrastive_samples_per_second": 300.549,
|
| 2057 |
+
"eval_qnli-contrastive_steps_per_second": 10.018,
|
| 2058 |
+
"step": 9555
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 1.6266598569969357,
|
| 2062 |
+
"eval_scitail-pairs-qa_loss": 0.13629749417304993,
|
| 2063 |
+
"eval_scitail-pairs-qa_runtime": 1.1751,
|
| 2064 |
+
"eval_scitail-pairs-qa_samples_per_second": 127.653,
|
| 2065 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.255,
|
| 2066 |
+
"step": 9555
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 1.6266598569969357,
|
| 2070 |
+
"eval_scitail-pairs-pos_loss": 0.5964671969413757,
|
| 2071 |
+
"eval_scitail-pairs-pos_runtime": 2.1841,
|
| 2072 |
+
"eval_scitail-pairs-pos_samples_per_second": 68.677,
|
| 2073 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.289,
|
| 2074 |
+
"step": 9555
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.6266598569969357,
|
| 2078 |
+
"eval_xsum-pairs_loss": 0.6746851205825806,
|
| 2079 |
+
"eval_xsum-pairs_runtime": 2.2628,
|
| 2080 |
+
"eval_xsum-pairs_samples_per_second": 66.291,
|
| 2081 |
+
"eval_xsum-pairs_steps_per_second": 2.21,
|
| 2082 |
+
"step": 9555
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.6266598569969357,
|
| 2086 |
+
"eval_compression-pairs_loss": 0.17857055366039276,
|
| 2087 |
+
"eval_compression-pairs_runtime": 0.4506,
|
| 2088 |
+
"eval_compression-pairs_samples_per_second": 332.902,
|
| 2089 |
+
"eval_compression-pairs_steps_per_second": 11.097,
|
| 2090 |
+
"step": 9555
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 1.6266598569969357,
|
| 2094 |
+
"eval_sciq_pairs_loss": 0.7349148988723755,
|
| 2095 |
+
"eval_sciq_pairs_runtime": 7.116,
|
| 2096 |
+
"eval_sciq_pairs_samples_per_second": 21.079,
|
| 2097 |
+
"eval_sciq_pairs_steps_per_second": 0.703,
|
| 2098 |
+
"step": 9555
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 1.6266598569969357,
|
| 2102 |
+
"eval_qasc_pairs_loss": 5.115650177001953,
|
| 2103 |
+
"eval_qasc_pairs_runtime": 2.0271,
|
| 2104 |
+
"eval_qasc_pairs_samples_per_second": 73.997,
|
| 2105 |
+
"eval_qasc_pairs_steps_per_second": 2.467,
|
| 2106 |
+
"step": 9555
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 1.6266598569969357,
|
| 2110 |
+
"eval_openbookqa_pairs_loss": 2.694535255432129,
|
| 2111 |
+
"eval_openbookqa_pairs_runtime": 0.8634,
|
| 2112 |
+
"eval_openbookqa_pairs_samples_per_second": 119.302,
|
| 2113 |
+
"eval_openbookqa_pairs_steps_per_second": 4.633,
|
| 2114 |
+
"step": 9555
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 1.6266598569969357,
|
| 2118 |
+
"eval_msmarco_pairs_loss": 1.5184054374694824,
|
| 2119 |
+
"eval_msmarco_pairs_runtime": 2.7561,
|
| 2120 |
+
"eval_msmarco_pairs_samples_per_second": 54.424,
|
| 2121 |
+
"eval_msmarco_pairs_steps_per_second": 1.814,
|
| 2122 |
+
"step": 9555
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 1.6266598569969357,
|
| 2126 |
+
"eval_nq_pairs_loss": 1.293426752090454,
|
| 2127 |
+
"eval_nq_pairs_runtime": 5.0107,
|
| 2128 |
+
"eval_nq_pairs_samples_per_second": 29.936,
|
| 2129 |
+
"eval_nq_pairs_steps_per_second": 0.998,
|
| 2130 |
+
"step": 9555
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.6266598569969357,
|
| 2134 |
+
"eval_trivia_pairs_loss": 1.5939557552337646,
|
| 2135 |
+
"eval_trivia_pairs_runtime": 9.5368,
|
| 2136 |
+
"eval_trivia_pairs_samples_per_second": 15.728,
|
| 2137 |
+
"eval_trivia_pairs_steps_per_second": 0.524,
|
| 2138 |
+
"step": 9555
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.6266598569969357,
|
| 2142 |
+
"eval_quora_pairs_loss": 0.31308451294898987,
|
| 2143 |
+
"eval_quora_pairs_runtime": 1.1456,
|
| 2144 |
+
"eval_quora_pairs_samples_per_second": 130.932,
|
| 2145 |
+
"eval_quora_pairs_steps_per_second": 4.364,
|
| 2146 |
+
"step": 9555
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 1.6266598569969357,
|
| 2150 |
+
"eval_gooaq_pairs_loss": 1.0807112455368042,
|
| 2151 |
+
"eval_gooaq_pairs_runtime": 2.0197,
|
| 2152 |
+
"eval_gooaq_pairs_samples_per_second": 74.269,
|
| 2153 |
+
"eval_gooaq_pairs_steps_per_second": 2.476,
|
| 2154 |
+
"step": 9555
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 1.651685393258427,
|
| 2158 |
+
"grad_norm": 0.7546759843826294,
|
| 2159 |
+
"learning_rate": 2.878777643060379e-05,
|
| 2160 |
+
"loss": 1.4469,
|
| 2161 |
+
"step": 9702
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 1.6767109295199183,
|
| 2165 |
+
"grad_norm": 0.8483991026878357,
|
| 2166 |
+
"learning_rate": 2.835869224114224e-05,
|
| 2167 |
+
"loss": 1.3878,
|
| 2168 |
+
"step": 9849
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 1.7017364657814096,
|
| 2172 |
+
"grad_norm": 20.814105987548828,
|
| 2173 |
+
"learning_rate": 2.7868966188352908e-05,
|
| 2174 |
+
"loss": 1.2764,
|
| 2175 |
+
"step": 9996
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 1.7267620020429009,
|
| 2179 |
+
"grad_norm": 3.1025094985961914,
|
| 2180 |
+
"learning_rate": 2.73208213865815e-05,
|
| 2181 |
+
"loss": 1.3884,
|
| 2182 |
+
"step": 10143
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 1.7517875383043924,
|
| 2186 |
+
"grad_norm": 14.80810260772705,
|
| 2187 |
+
"learning_rate": 2.671674614243416e-05,
|
| 2188 |
+
"loss": 1.2977,
|
| 2189 |
+
"step": 10290
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 1.7517875383043924,
|
| 2193 |
+
"eval_nli-pairs_loss": 1.3081562519073486,
|
| 2194 |
+
"eval_nli-pairs_runtime": 4.0165,
|
| 2195 |
+
"eval_nli-pairs_samples_per_second": 37.346,
|
| 2196 |
+
"eval_nli-pairs_steps_per_second": 1.245,
|
| 2197 |
+
"eval_sts-test_pearson_cosine": 0.7681143802843627,
|
| 2198 |
+
"eval_sts-test_pearson_dot": 0.5287526695750702,
|
| 2199 |
+
"eval_sts-test_pearson_euclidean": 0.7538805205317111,
|
| 2200 |
+
"eval_sts-test_pearson_manhattan": 0.7596894203751682,
|
| 2201 |
+
"eval_sts-test_pearson_max": 0.7681143802843627,
|
| 2202 |
+
"eval_sts-test_spearman_cosine": 0.770908506196058,
|
| 2203 |
+
"eval_sts-test_spearman_dot": 0.5670572774538138,
|
| 2204 |
+
"eval_sts-test_spearman_euclidean": 0.7452730842318486,
|
| 2205 |
+
"eval_sts-test_spearman_manhattan": 0.7517699916174685,
|
| 2206 |
+
"eval_sts-test_spearman_max": 0.770908506196058,
|
| 2207 |
+
"step": 10290
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 1.7517875383043924,
|
| 2211 |
+
"eval_vitaminc-pairs_loss": 0.9676446318626404,
|
| 2212 |
+
"eval_vitaminc-pairs_runtime": 2.1787,
|
| 2213 |
+
"eval_vitaminc-pairs_samples_per_second": 68.85,
|
| 2214 |
+
"eval_vitaminc-pairs_steps_per_second": 2.295,
|
| 2215 |
+
"step": 10290
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 1.7517875383043924,
|
| 2219 |
+
"eval_qnli-contrastive_loss": 0.244391530752182,
|
| 2220 |
+
"eval_qnli-contrastive_runtime": 0.4884,
|
| 2221 |
+
"eval_qnli-contrastive_samples_per_second": 307.113,
|
| 2222 |
+
"eval_qnli-contrastive_steps_per_second": 10.237,
|
| 2223 |
+
"step": 10290
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 1.7517875383043924,
|
| 2227 |
+
"eval_scitail-pairs-qa_loss": 0.1264333575963974,
|
| 2228 |
+
"eval_scitail-pairs-qa_runtime": 1.1536,
|
| 2229 |
+
"eval_scitail-pairs-qa_samples_per_second": 130.03,
|
| 2230 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.334,
|
| 2231 |
+
"step": 10290
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 1.7517875383043924,
|
| 2235 |
+
"eval_scitail-pairs-pos_loss": 0.5472012162208557,
|
| 2236 |
+
"eval_scitail-pairs-pos_runtime": 2.1213,
|
| 2237 |
+
"eval_scitail-pairs-pos_samples_per_second": 70.711,
|
| 2238 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.357,
|
| 2239 |
+
"step": 10290
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 1.7517875383043924,
|
| 2243 |
+
"eval_xsum-pairs_loss": 0.5869634747505188,
|
| 2244 |
+
"eval_xsum-pairs_runtime": 2.2876,
|
| 2245 |
+
"eval_xsum-pairs_samples_per_second": 65.571,
|
| 2246 |
+
"eval_xsum-pairs_steps_per_second": 2.186,
|
| 2247 |
+
"step": 10290
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 1.7517875383043924,
|
| 2251 |
+
"eval_compression-pairs_loss": 0.16663199663162231,
|
| 2252 |
+
"eval_compression-pairs_runtime": 0.4431,
|
| 2253 |
+
"eval_compression-pairs_samples_per_second": 338.526,
|
| 2254 |
+
"eval_compression-pairs_steps_per_second": 11.284,
|
| 2255 |
+
"step": 10290
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 1.7517875383043924,
|
| 2259 |
+
"eval_sciq_pairs_loss": 0.6884138584136963,
|
| 2260 |
+
"eval_sciq_pairs_runtime": 7.0451,
|
| 2261 |
+
"eval_sciq_pairs_samples_per_second": 21.291,
|
| 2262 |
+
"eval_sciq_pairs_steps_per_second": 0.71,
|
| 2263 |
+
"step": 10290
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 1.7517875383043924,
|
| 2267 |
+
"eval_qasc_pairs_loss": 5.099090099334717,
|
| 2268 |
+
"eval_qasc_pairs_runtime": 2.0309,
|
| 2269 |
+
"eval_qasc_pairs_samples_per_second": 73.86,
|
| 2270 |
+
"eval_qasc_pairs_steps_per_second": 2.462,
|
| 2271 |
+
"step": 10290
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 1.7517875383043924,
|
| 2275 |
+
"eval_openbookqa_pairs_loss": 2.6562159061431885,
|
| 2276 |
+
"eval_openbookqa_pairs_runtime": 0.8531,
|
| 2277 |
+
"eval_openbookqa_pairs_samples_per_second": 120.74,
|
| 2278 |
+
"eval_openbookqa_pairs_steps_per_second": 4.689,
|
| 2279 |
+
"step": 10290
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 1.7517875383043924,
|
| 2283 |
+
"eval_msmarco_pairs_loss": 1.3729219436645508,
|
| 2284 |
+
"eval_msmarco_pairs_runtime": 2.7346,
|
| 2285 |
+
"eval_msmarco_pairs_samples_per_second": 54.853,
|
| 2286 |
+
"eval_msmarco_pairs_steps_per_second": 1.828,
|
| 2287 |
+
"step": 10290
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 1.7517875383043924,
|
| 2291 |
+
"eval_nq_pairs_loss": 1.2174726724624634,
|
| 2292 |
+
"eval_nq_pairs_runtime": 4.9981,
|
| 2293 |
+
"eval_nq_pairs_samples_per_second": 30.012,
|
| 2294 |
+
"eval_nq_pairs_steps_per_second": 1.0,
|
| 2295 |
+
"step": 10290
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 1.7517875383043924,
|
| 2299 |
+
"eval_trivia_pairs_loss": 1.5839861631393433,
|
| 2300 |
+
"eval_trivia_pairs_runtime": 9.4611,
|
| 2301 |
+
"eval_trivia_pairs_samples_per_second": 15.854,
|
| 2302 |
+
"eval_trivia_pairs_steps_per_second": 0.528,
|
| 2303 |
+
"step": 10290
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 1.7517875383043924,
|
| 2307 |
+
"eval_quora_pairs_loss": 0.2804078757762909,
|
| 2308 |
+
"eval_quora_pairs_runtime": 1.1799,
|
| 2309 |
+
"eval_quora_pairs_samples_per_second": 127.13,
|
| 2310 |
+
"eval_quora_pairs_steps_per_second": 4.238,
|
| 2311 |
+
"step": 10290
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 1.7517875383043924,
|
| 2315 |
+
"eval_gooaq_pairs_loss": 0.9541385769844055,
|
| 2316 |
+
"eval_gooaq_pairs_runtime": 2.1014,
|
| 2317 |
+
"eval_gooaq_pairs_samples_per_second": 71.38,
|
| 2318 |
+
"eval_gooaq_pairs_steps_per_second": 2.379,
|
| 2319 |
+
"step": 10290
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 1.7768130745658834,
|
| 2323 |
+
"grad_norm": 51.763004302978516,
|
| 2324 |
+
"learning_rate": 2.6059482659094694e-05,
|
| 2325 |
+
"loss": 1.4422,
|
| 2326 |
+
"step": 10437
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 1.801838610827375,
|
| 2330 |
+
"grad_norm": 3.4887988567352295,
|
| 2331 |
+
"learning_rate": 2.5352014588076858e-05,
|
| 2332 |
+
"loss": 1.4997,
|
| 2333 |
+
"step": 10584
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.8268641470888662,
|
| 2337 |
+
"grad_norm": 6.360722064971924,
|
| 2338 |
+
"learning_rate": 2.4597553484920438e-05,
|
| 2339 |
+
"loss": 1.2797,
|
| 2340 |
+
"step": 10731
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 1.8518896833503575,
|
| 2344 |
+
"grad_norm": 16.216428756713867,
|
| 2345 |
+
"learning_rate": 2.3799524230315696e-05,
|
| 2346 |
+
"loss": 1.2362,
|
| 2347 |
+
"step": 10878
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 1.8769152196118488,
|
| 2351 |
+
"grad_norm": 19.113628387451172,
|
| 2352 |
+
"learning_rate": 2.2961549482836967e-05,
|
| 2353 |
+
"loss": 1.2799,
|
| 2354 |
+
"step": 11025
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 1.8769152196118488,
|
| 2358 |
+
"eval_nli-pairs_loss": 1.2669230699539185,
|
| 2359 |
+
"eval_nli-pairs_runtime": 4.0111,
|
| 2360 |
+
"eval_nli-pairs_samples_per_second": 37.396,
|
| 2361 |
+
"eval_nli-pairs_steps_per_second": 1.247,
|
| 2362 |
+
"eval_sts-test_pearson_cosine": 0.774489523257569,
|
| 2363 |
+
"eval_sts-test_pearson_dot": 0.5150859135257536,
|
| 2364 |
+
"eval_sts-test_pearson_euclidean": 0.7570251269629877,
|
| 2365 |
+
"eval_sts-test_pearson_manhattan": 0.7623769541465137,
|
| 2366 |
+
"eval_sts-test_pearson_max": 0.774489523257569,
|
| 2367 |
+
"eval_sts-test_spearman_cosine": 0.7816800005074528,
|
| 2368 |
+
"eval_sts-test_spearman_dot": 0.565603897190929,
|
| 2369 |
+
"eval_sts-test_spearman_euclidean": 0.7507848233553155,
|
| 2370 |
+
"eval_sts-test_spearman_manhattan": 0.756029656784038,
|
| 2371 |
+
"eval_sts-test_spearman_max": 0.7816800005074528,
|
| 2372 |
+
"step": 11025
|
| 2373 |
+
},
|
| 2374 |
+
{
|
| 2375 |
+
"epoch": 1.8769152196118488,
|
| 2376 |
+
"eval_vitaminc-pairs_loss": 0.875577986240387,
|
| 2377 |
+
"eval_vitaminc-pairs_runtime": 2.2185,
|
| 2378 |
+
"eval_vitaminc-pairs_samples_per_second": 67.614,
|
| 2379 |
+
"eval_vitaminc-pairs_steps_per_second": 2.254,
|
| 2380 |
+
"step": 11025
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 1.8769152196118488,
|
| 2384 |
+
"eval_qnli-contrastive_loss": 0.23095794022083282,
|
| 2385 |
+
"eval_qnli-contrastive_runtime": 0.4906,
|
| 2386 |
+
"eval_qnli-contrastive_samples_per_second": 305.756,
|
| 2387 |
+
"eval_qnli-contrastive_steps_per_second": 10.192,
|
| 2388 |
+
"step": 11025
|
| 2389 |
+
},
|
| 2390 |
+
{
|
| 2391 |
+
"epoch": 1.8769152196118488,
|
| 2392 |
+
"eval_scitail-pairs-qa_loss": 0.11762743443250656,
|
| 2393 |
+
"eval_scitail-pairs-qa_runtime": 1.1505,
|
| 2394 |
+
"eval_scitail-pairs-qa_samples_per_second": 130.379,
|
| 2395 |
+
"eval_scitail-pairs-qa_steps_per_second": 4.346,
|
| 2396 |
+
"step": 11025
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 1.8769152196118488,
|
| 2400 |
+
"eval_scitail-pairs-pos_loss": 0.5055103898048401,
|
| 2401 |
+
"eval_scitail-pairs-pos_runtime": 2.1912,
|
| 2402 |
+
"eval_scitail-pairs-pos_samples_per_second": 68.456,
|
| 2403 |
+
"eval_scitail-pairs-pos_steps_per_second": 2.282,
|
| 2404 |
+
"step": 11025
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 1.8769152196118488,
|
| 2408 |
+
"eval_xsum-pairs_loss": 0.5941822528839111,
|
| 2409 |
+
"eval_xsum-pairs_runtime": 2.26,
|
| 2410 |
+
"eval_xsum-pairs_samples_per_second": 66.371,
|
| 2411 |
+
"eval_xsum-pairs_steps_per_second": 2.212,
|
| 2412 |
+
"step": 11025
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 1.8769152196118488,
|
| 2416 |
+
"eval_compression-pairs_loss": 0.16561630368232727,
|
| 2417 |
+
"eval_compression-pairs_runtime": 0.4447,
|
| 2418 |
+
"eval_compression-pairs_samples_per_second": 337.281,
|
| 2419 |
+
"eval_compression-pairs_steps_per_second": 11.243,
|
| 2420 |
+
"step": 11025
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 1.8769152196118488,
|
| 2424 |
+
"eval_sciq_pairs_loss": 0.6859617233276367,
|
| 2425 |
+
"eval_sciq_pairs_runtime": 7.2855,
|
| 2426 |
+
"eval_sciq_pairs_samples_per_second": 20.589,
|
| 2427 |
+
"eval_sciq_pairs_steps_per_second": 0.686,
|
| 2428 |
+
"step": 11025
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 1.8769152196118488,
|
| 2432 |
+
"eval_qasc_pairs_loss": 4.979205131530762,
|
| 2433 |
+
"eval_qasc_pairs_runtime": 2.0332,
|
| 2434 |
+
"eval_qasc_pairs_samples_per_second": 73.775,
|
| 2435 |
+
"eval_qasc_pairs_steps_per_second": 2.459,
|
| 2436 |
+
"step": 11025
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 1.8769152196118488,
|
| 2440 |
+
"eval_openbookqa_pairs_loss": 2.5103061199188232,
|
| 2441 |
+
"eval_openbookqa_pairs_runtime": 0.8673,
|
| 2442 |
+
"eval_openbookqa_pairs_samples_per_second": 118.755,
|
| 2443 |
+
"eval_openbookqa_pairs_steps_per_second": 4.612,
|
| 2444 |
+
"step": 11025
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 1.8769152196118488,
|
| 2448 |
+
"eval_msmarco_pairs_loss": 1.2753304243087769,
|
| 2449 |
+
"eval_msmarco_pairs_runtime": 2.7942,
|
| 2450 |
+
"eval_msmarco_pairs_samples_per_second": 53.683,
|
| 2451 |
+
"eval_msmarco_pairs_steps_per_second": 1.789,
|
| 2452 |
+
"step": 11025
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 1.8769152196118488,
|
| 2456 |
+
"eval_nq_pairs_loss": 1.057248592376709,
|
| 2457 |
+
"eval_nq_pairs_runtime": 5.0749,
|
| 2458 |
+
"eval_nq_pairs_samples_per_second": 29.557,
|
| 2459 |
+
"eval_nq_pairs_steps_per_second": 0.985,
|
| 2460 |
+
"step": 11025
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 1.8769152196118488,
|
| 2464 |
+
"eval_trivia_pairs_loss": 1.4893617630004883,
|
| 2465 |
+
"eval_trivia_pairs_runtime": 9.5535,
|
| 2466 |
+
"eval_trivia_pairs_samples_per_second": 15.701,
|
| 2467 |
+
"eval_trivia_pairs_steps_per_second": 0.523,
|
| 2468 |
+
"step": 11025
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 1.8769152196118488,
|
| 2472 |
+
"eval_quora_pairs_loss": 0.27783504128456116,
|
| 2473 |
+
"eval_quora_pairs_runtime": 1.1843,
|
| 2474 |
+
"eval_quora_pairs_samples_per_second": 126.653,
|
| 2475 |
+
"eval_quora_pairs_steps_per_second": 4.222,
|
| 2476 |
+
"step": 11025
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 1.8769152196118488,
|
| 2480 |
+
"eval_gooaq_pairs_loss": 0.8971360325813293,
|
| 2481 |
+
"eval_gooaq_pairs_runtime": 2.0278,
|
| 2482 |
+
"eval_gooaq_pairs_samples_per_second": 73.97,
|
| 2483 |
+
"eval_gooaq_pairs_steps_per_second": 2.466,
|
| 2484 |
+
"step": 11025
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 1.90194075587334,
|
| 2488 |
+
"grad_norm": 14.915979385375977,
|
| 2489 |
+
"learning_rate": 2.2087433233862403e-05,
|
| 2490 |
+
"loss": 1.2292,
|
| 2491 |
+
"step": 11172
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 1.9269662921348316,
|
| 2495 |
+
"grad_norm": 13.753366470336914,
|
| 2496 |
+
"learning_rate": 2.118740830659258e-05,
|
| 2497 |
+
"loss": 1.0362,
|
| 2498 |
+
"step": 11319
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 1.9519918283963227,
|
| 2502 |
+
"grad_norm": 8.33267593383789,
|
| 2503 |
+
"learning_rate": 2.0259676306932596e-05,
|
| 2504 |
+
"loss": 1.1851,
|
| 2505 |
+
"step": 11466
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 1.9770173646578142,
|
| 2509 |
+
"grad_norm": 0.6671110987663269,
|
| 2510 |
+
"learning_rate": 1.9301804508269106e-05,
|
| 2511 |
+
"loss": 1.0248,
|
| 2512 |
+
"step": 11613
|
| 2513 |
+
}
|
| 2514 |
+
],
|
| 2515 |
+
"logging_steps": 147,
|
| 2516 |
+
"max_steps": 29370,
|
| 2517 |
+
"num_input_tokens_seen": 0,
|
| 2518 |
+
"num_train_epochs": 5,
|
| 2519 |
+
"save_steps": 2937,
|
| 2520 |
+
"stateful_callbacks": {
|
| 2521 |
+
"TrainerControl": {
|
| 2522 |
+
"args": {
|
| 2523 |
+
"should_epoch_stop": false,
|
| 2524 |
+
"should_evaluate": false,
|
| 2525 |
+
"should_log": false,
|
| 2526 |
+
"should_save": true,
|
| 2527 |
+
"should_training_stop": false
|
| 2528 |
+
},
|
| 2529 |
+
"attributes": {}
|
| 2530 |
+
}
|
| 2531 |
+
},
|
| 2532 |
+
"total_flos": 0.0,
|
| 2533 |
+
"train_batch_size": 32,
|
| 2534 |
+
"trial_name": null,
|
| 2535 |
+
"trial_params": null
|
| 2536 |
+
}
|
checkpoint-11748/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:298cfa7a9e669e98ab55937cac47ecfa89fb6a36f2afe18f46ba782a5f5bf5a4
|
| 3 |
+
size 5624
|