Training in progress, step 258, checkpoint
Browse files- checkpoint-258/1_Pooling/config.json +10 -0
- checkpoint-258/README.md +0 -0
- checkpoint-258/added_tokens.json +3 -0
- checkpoint-258/config.json +35 -0
- checkpoint-258/config_sentence_transformers.json +10 -0
- checkpoint-258/modules.json +14 -0
- checkpoint-258/optimizer.pt +3 -0
- checkpoint-258/pytorch_model.bin +3 -0
- checkpoint-258/rng_state.pth +3 -0
- checkpoint-258/scheduler.pt +3 -0
- checkpoint-258/sentence_bert_config.json +4 -0
- checkpoint-258/special_tokens_map.json +15 -0
- checkpoint-258/spm.model +3 -0
- checkpoint-258/tokenizer.json +0 -0
- checkpoint-258/tokenizer_config.json +58 -0
- checkpoint-258/trainer_state.json +2611 -0
- checkpoint-258/training_args.bin +3 -0
checkpoint-258/1_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-258/README.md
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-258/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"[MASK]": 128000
|
| 3 |
+
}
|
checkpoint-258/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/deberta-v3-base",
|
| 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": 12,
|
| 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": "4.44.0",
|
| 33 |
+
"type_vocab_size": 0,
|
| 34 |
+
"vocab_size": 128100
|
| 35 |
+
}
|
checkpoint-258/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.0.1",
|
| 4 |
+
"transformers": "4.44.0",
|
| 5 |
+
"pytorch": "2.4.0"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": null
|
| 10 |
+
}
|
checkpoint-258/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-258/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:52fc312695b9f87a53ef000d546243f118fde3b0883f79ce888bd6cbd744369b
|
| 3 |
+
size 1470818042
|
checkpoint-258/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0bb547068bd1f2049a4a97699a45e62319e3c48227d1a14f1f92dcbf5daa2276
|
| 3 |
+
size 735393442
|
checkpoint-258/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:508097ffb07b37beb0a6a54ec8f5c973c76081fda11fd2131977b14d2340599b
|
| 3 |
+
size 14244
|
checkpoint-258/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb394469aa0db1d679ad75869541aa383405c98ce30610570ed96c2404a2be6a
|
| 3 |
+
size 1064
|
checkpoint-258/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 512,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-258/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-258/spm.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
|
| 3 |
+
size 2464616
|
checkpoint-258/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-258/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-258/trainer_state.json
ADDED
|
@@ -0,0 +1,2611 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.20015515903801395,
|
| 5 |
+
"eval_steps": 60,
|
| 6 |
+
"global_step": 258,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0007757951900698216,
|
| 13 |
+
"grad_norm": NaN,
|
| 14 |
+
"learning_rate": 0.0,
|
| 15 |
+
"loss": 6.6871,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0015515903801396431,
|
| 20 |
+
"grad_norm": 17.51189613342285,
|
| 21 |
+
"learning_rate": 4.112808460634547e-08,
|
| 22 |
+
"loss": 5.8546,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0023273855702094647,
|
| 27 |
+
"grad_norm": 33.663917541503906,
|
| 28 |
+
"learning_rate": 8.225616921269094e-08,
|
| 29 |
+
"loss": 6.576,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0031031807602792862,
|
| 34 |
+
"grad_norm": 23.573055267333984,
|
| 35 |
+
"learning_rate": 1.2338425381903642e-07,
|
| 36 |
+
"loss": 6.8301,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.003878975950349108,
|
| 41 |
+
"grad_norm": 9.426567077636719,
|
| 42 |
+
"learning_rate": 1.6451233842538187e-07,
|
| 43 |
+
"loss": 5.2308,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.004654771140418929,
|
| 48 |
+
"grad_norm": 19.919723510742188,
|
| 49 |
+
"learning_rate": 2.0564042303172737e-07,
|
| 50 |
+
"loss": 6.3588,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.005430566330488751,
|
| 55 |
+
"grad_norm": 20.105239868164062,
|
| 56 |
+
"learning_rate": 2.4676850763807285e-07,
|
| 57 |
+
"loss": 6.8417,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0062063615205585725,
|
| 62 |
+
"grad_norm": Infinity,
|
| 63 |
+
"learning_rate": 2.4676850763807285e-07,
|
| 64 |
+
"loss": 10.3916,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0069821567106283944,
|
| 69 |
+
"grad_norm": 10.925858497619629,
|
| 70 |
+
"learning_rate": 2.878965922444183e-07,
|
| 71 |
+
"loss": 5.5136,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.007757951900698216,
|
| 76 |
+
"grad_norm": 24.988182067871094,
|
| 77 |
+
"learning_rate": 3.2902467685076374e-07,
|
| 78 |
+
"loss": 7.4726,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.008533747090768037,
|
| 83 |
+
"grad_norm": 10.366790771484375,
|
| 84 |
+
"learning_rate": 3.701527614571093e-07,
|
| 85 |
+
"loss": 5.5414,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.009309542280837859,
|
| 90 |
+
"grad_norm": 10.754115104675293,
|
| 91 |
+
"learning_rate": 4.1128084606345474e-07,
|
| 92 |
+
"loss": 4.8507,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.01008533747090768,
|
| 97 |
+
"grad_norm": 11.047926902770996,
|
| 98 |
+
"learning_rate": 4.524089306698002e-07,
|
| 99 |
+
"loss": 4.9058,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.010861132660977503,
|
| 104 |
+
"grad_norm": 15.155884742736816,
|
| 105 |
+
"learning_rate": 4.935370152761457e-07,
|
| 106 |
+
"loss": 6.2301,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.011636927851047323,
|
| 111 |
+
"grad_norm": 32.3827018737793,
|
| 112 |
+
"learning_rate": 5.346650998824911e-07,
|
| 113 |
+
"loss": 7.9813,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.012412723041117145,
|
| 118 |
+
"grad_norm": 16.102903366088867,
|
| 119 |
+
"learning_rate": 5.757931844888366e-07,
|
| 120 |
+
"loss": 5.5269,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.013188518231186967,
|
| 125 |
+
"grad_norm": 53.7273063659668,
|
| 126 |
+
"learning_rate": 6.169212690951821e-07,
|
| 127 |
+
"loss": 10.0786,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.013964313421256789,
|
| 132 |
+
"grad_norm": 51.25851058959961,
|
| 133 |
+
"learning_rate": 6.580493537015275e-07,
|
| 134 |
+
"loss": 9.6075,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.01474010861132661,
|
| 139 |
+
"grad_norm": 10.667213439941406,
|
| 140 |
+
"learning_rate": 6.99177438307873e-07,
|
| 141 |
+
"loss": 4.671,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.015515903801396431,
|
| 146 |
+
"grad_norm": 24.299957275390625,
|
| 147 |
+
"learning_rate": 7.403055229142186e-07,
|
| 148 |
+
"loss": 6.945,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.016291698991466253,
|
| 153 |
+
"grad_norm": 16.173965454101562,
|
| 154 |
+
"learning_rate": 7.81433607520564e-07,
|
| 155 |
+
"loss": 6.233,
|
| 156 |
+
"step": 21
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.017067494181536073,
|
| 160 |
+
"grad_norm": 10.136642456054688,
|
| 161 |
+
"learning_rate": 8.225616921269095e-07,
|
| 162 |
+
"loss": 5.5828,
|
| 163 |
+
"step": 22
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.017843289371605897,
|
| 167 |
+
"grad_norm": 17.316375732421875,
|
| 168 |
+
"learning_rate": 8.636897767332549e-07,
|
| 169 |
+
"loss": 6.6991,
|
| 170 |
+
"step": 23
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.018619084561675717,
|
| 174 |
+
"grad_norm": 12.166691780090332,
|
| 175 |
+
"learning_rate": 9.048178613396004e-07,
|
| 176 |
+
"loss": 5.9928,
|
| 177 |
+
"step": 24
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.019394879751745538,
|
| 181 |
+
"grad_norm": 11.132185935974121,
|
| 182 |
+
"learning_rate": 9.459459459459459e-07,
|
| 183 |
+
"loss": 4.8443,
|
| 184 |
+
"step": 25
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.02017067494181536,
|
| 188 |
+
"grad_norm": 57.284271240234375,
|
| 189 |
+
"learning_rate": 9.870740305522914e-07,
|
| 190 |
+
"loss": 11.0738,
|
| 191 |
+
"step": 26
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.02094647013188518,
|
| 195 |
+
"grad_norm": 19.147293090820312,
|
| 196 |
+
"learning_rate": 1.0282021151586369e-06,
|
| 197 |
+
"loss": 6.4076,
|
| 198 |
+
"step": 27
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.021722265321955005,
|
| 202 |
+
"grad_norm": 14.959942817687988,
|
| 203 |
+
"learning_rate": 1.0693301997649822e-06,
|
| 204 |
+
"loss": 6.2668,
|
| 205 |
+
"step": 28
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.022498060512024826,
|
| 209 |
+
"grad_norm": 10.889761924743652,
|
| 210 |
+
"learning_rate": 1.1104582843713277e-06,
|
| 211 |
+
"loss": 5.092,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.023273855702094646,
|
| 216 |
+
"grad_norm": 13.842042922973633,
|
| 217 |
+
"learning_rate": 1.1515863689776732e-06,
|
| 218 |
+
"loss": 5.9955,
|
| 219 |
+
"step": 30
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.02404965089216447,
|
| 223 |
+
"grad_norm": 19.306137084960938,
|
| 224 |
+
"learning_rate": 1.1927144535840187e-06,
|
| 225 |
+
"loss": 6.6071,
|
| 226 |
+
"step": 31
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.02482544608223429,
|
| 230 |
+
"grad_norm": 10.03764533996582,
|
| 231 |
+
"learning_rate": 1.2338425381903642e-06,
|
| 232 |
+
"loss": 4.7984,
|
| 233 |
+
"step": 32
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.02560124127230411,
|
| 237 |
+
"grad_norm": 15.16897964477539,
|
| 238 |
+
"learning_rate": 1.2749706227967097e-06,
|
| 239 |
+
"loss": 6.173,
|
| 240 |
+
"step": 33
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.026377036462373934,
|
| 244 |
+
"grad_norm": 15.80406665802002,
|
| 245 |
+
"learning_rate": 1.316098707403055e-06,
|
| 246 |
+
"loss": 5.9604,
|
| 247 |
+
"step": 34
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.027152831652443754,
|
| 251 |
+
"grad_norm": 17.38121795654297,
|
| 252 |
+
"learning_rate": 1.3572267920094007e-06,
|
| 253 |
+
"loss": 6.3435,
|
| 254 |
+
"step": 35
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.027928626842513578,
|
| 258 |
+
"grad_norm": 12.920570373535156,
|
| 259 |
+
"learning_rate": 1.398354876615746e-06,
|
| 260 |
+
"loss": 6.0394,
|
| 261 |
+
"step": 36
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.028704422032583398,
|
| 265 |
+
"grad_norm": 14.097166061401367,
|
| 266 |
+
"learning_rate": 1.4394829612220915e-06,
|
| 267 |
+
"loss": 6.0135,
|
| 268 |
+
"step": 37
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.02948021722265322,
|
| 272 |
+
"grad_norm": 48.04243469238281,
|
| 273 |
+
"learning_rate": 1.4806110458284372e-06,
|
| 274 |
+
"loss": 8.0721,
|
| 275 |
+
"step": 38
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.030256012412723042,
|
| 279 |
+
"grad_norm": 13.951966285705566,
|
| 280 |
+
"learning_rate": 1.5217391304347825e-06,
|
| 281 |
+
"loss": 6.2283,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.031031807602792862,
|
| 286 |
+
"grad_norm": 9.040204048156738,
|
| 287 |
+
"learning_rate": 1.562867215041128e-06,
|
| 288 |
+
"loss": 5.365,
|
| 289 |
+
"step": 40
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.03180760279286268,
|
| 293 |
+
"grad_norm": 16.60372543334961,
|
| 294 |
+
"learning_rate": 1.6039952996474733e-06,
|
| 295 |
+
"loss": 6.0378,
|
| 296 |
+
"step": 41
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.032583397982932506,
|
| 300 |
+
"grad_norm": 18.186500549316406,
|
| 301 |
+
"learning_rate": 1.645123384253819e-06,
|
| 302 |
+
"loss": 6.5136,
|
| 303 |
+
"step": 42
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.03335919317300233,
|
| 307 |
+
"grad_norm": 10.157137870788574,
|
| 308 |
+
"learning_rate": 1.6862514688601645e-06,
|
| 309 |
+
"loss": 5.6955,
|
| 310 |
+
"step": 43
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.03413498836307215,
|
| 314 |
+
"grad_norm": 12.988446235656738,
|
| 315 |
+
"learning_rate": 1.7273795534665098e-06,
|
| 316 |
+
"loss": 6.1769,
|
| 317 |
+
"step": 44
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.03491078355314197,
|
| 321 |
+
"grad_norm": 17.56638526916504,
|
| 322 |
+
"learning_rate": 1.7685076380728553e-06,
|
| 323 |
+
"loss": 6.302,
|
| 324 |
+
"step": 45
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.035686578743211794,
|
| 328 |
+
"grad_norm": 14.749937057495117,
|
| 329 |
+
"learning_rate": 1.8096357226792008e-06,
|
| 330 |
+
"loss": 6.0792,
|
| 331 |
+
"step": 46
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.03646237393328161,
|
| 335 |
+
"grad_norm": 10.104000091552734,
|
| 336 |
+
"learning_rate": 1.8507638072855463e-06,
|
| 337 |
+
"loss": 5.4317,
|
| 338 |
+
"step": 47
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.037238169123351435,
|
| 342 |
+
"grad_norm": 11.410161972045898,
|
| 343 |
+
"learning_rate": 1.8918918918918918e-06,
|
| 344 |
+
"loss": 5.7632,
|
| 345 |
+
"step": 48
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.03801396431342126,
|
| 349 |
+
"grad_norm": 10.582472801208496,
|
| 350 |
+
"learning_rate": 1.933019976498237e-06,
|
| 351 |
+
"loss": 4.6767,
|
| 352 |
+
"step": 49
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.038789759503491075,
|
| 356 |
+
"grad_norm": 14.601061820983887,
|
| 357 |
+
"learning_rate": 1.9741480611045828e-06,
|
| 358 |
+
"loss": 6.1871,
|
| 359 |
+
"step": 50
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.0395655546935609,
|
| 363 |
+
"grad_norm": 10.493443489074707,
|
| 364 |
+
"learning_rate": 2.015276145710928e-06,
|
| 365 |
+
"loss": 5.572,
|
| 366 |
+
"step": 51
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.04034134988363072,
|
| 370 |
+
"grad_norm": 24.4001522064209,
|
| 371 |
+
"learning_rate": 2.0564042303172738e-06,
|
| 372 |
+
"loss": 7.351,
|
| 373 |
+
"step": 52
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.041117145073700546,
|
| 377 |
+
"grad_norm": 13.242673873901367,
|
| 378 |
+
"learning_rate": 2.097532314923619e-06,
|
| 379 |
+
"loss": 5.6907,
|
| 380 |
+
"step": 53
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.04189294026377036,
|
| 384 |
+
"grad_norm": 11.323930740356445,
|
| 385 |
+
"learning_rate": 2.1386603995299644e-06,
|
| 386 |
+
"loss": 5.7986,
|
| 387 |
+
"step": 54
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.04266873545384019,
|
| 391 |
+
"grad_norm": 13.408857345581055,
|
| 392 |
+
"learning_rate": 2.17978848413631e-06,
|
| 393 |
+
"loss": 5.9598,
|
| 394 |
+
"step": 55
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.04344453064391001,
|
| 398 |
+
"grad_norm": 8.789331436157227,
|
| 399 |
+
"learning_rate": 2.2209165687426554e-06,
|
| 400 |
+
"loss": 6.024,
|
| 401 |
+
"step": 56
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.04422032583397983,
|
| 405 |
+
"grad_norm": 10.100234985351562,
|
| 406 |
+
"learning_rate": 2.2620446533490006e-06,
|
| 407 |
+
"loss": 5.328,
|
| 408 |
+
"step": 57
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.04499612102404965,
|
| 412 |
+
"grad_norm": 13.21589469909668,
|
| 413 |
+
"learning_rate": 2.3031727379553464e-06,
|
| 414 |
+
"loss": 5.5545,
|
| 415 |
+
"step": 58
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.045771916214119475,
|
| 419 |
+
"grad_norm": 9.624663352966309,
|
| 420 |
+
"learning_rate": 2.344300822561692e-06,
|
| 421 |
+
"loss": 5.3813,
|
| 422 |
+
"step": 59
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.04654771140418929,
|
| 426 |
+
"grad_norm": 5.808309555053711,
|
| 427 |
+
"learning_rate": 2.3854289071680374e-06,
|
| 428 |
+
"loss": 5.3023,
|
| 429 |
+
"step": 60
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.04654771140418929,
|
| 433 |
+
"eval_Qnli-dev_cosine_accuracy": 0.59765625,
|
| 434 |
+
"eval_Qnli-dev_cosine_accuracy_threshold": 0.9559981226921082,
|
| 435 |
+
"eval_Qnli-dev_cosine_ap": 0.564511259877405,
|
| 436 |
+
"eval_Qnli-dev_cosine_f1": 0.6291834002677376,
|
| 437 |
+
"eval_Qnli-dev_cosine_f1_threshold": 0.7179018259048462,
|
| 438 |
+
"eval_Qnli-dev_cosine_precision": 0.4598825831702544,
|
| 439 |
+
"eval_Qnli-dev_cosine_recall": 0.9957627118644068,
|
| 440 |
+
"eval_Qnli-dev_dot_accuracy": 0.55078125,
|
| 441 |
+
"eval_Qnli-dev_dot_accuracy_threshold": 737.3067016601562,
|
| 442 |
+
"eval_Qnli-dev_dot_ap": 0.4732733398818325,
|
| 443 |
+
"eval_Qnli-dev_dot_f1": 0.629878869448183,
|
| 444 |
+
"eval_Qnli-dev_dot_f1_threshold": 463.4846496582031,
|
| 445 |
+
"eval_Qnli-dev_dot_precision": 0.46153846153846156,
|
| 446 |
+
"eval_Qnli-dev_dot_recall": 0.9915254237288136,
|
| 447 |
+
"eval_Qnli-dev_euclidean_accuracy": 0.6015625,
|
| 448 |
+
"eval_Qnli-dev_euclidean_accuracy_threshold": 8.209783554077148,
|
| 449 |
+
"eval_Qnli-dev_euclidean_ap": 0.5833709480494583,
|
| 450 |
+
"eval_Qnli-dev_euclidean_f1": 0.6315789473684211,
|
| 451 |
+
"eval_Qnli-dev_euclidean_f1_threshold": 16.22138214111328,
|
| 452 |
+
"eval_Qnli-dev_euclidean_precision": 0.4633663366336634,
|
| 453 |
+
"eval_Qnli-dev_euclidean_recall": 0.9915254237288136,
|
| 454 |
+
"eval_Qnli-dev_manhattan_accuracy": 0.634765625,
|
| 455 |
+
"eval_Qnli-dev_manhattan_accuracy_threshold": 146.8093719482422,
|
| 456 |
+
"eval_Qnli-dev_manhattan_ap": 0.6173131758641012,
|
| 457 |
+
"eval_Qnli-dev_manhattan_f1": 0.6291834002677376,
|
| 458 |
+
"eval_Qnli-dev_manhattan_f1_threshold": 266.38592529296875,
|
| 459 |
+
"eval_Qnli-dev_manhattan_precision": 0.4598825831702544,
|
| 460 |
+
"eval_Qnli-dev_manhattan_recall": 0.9957627118644068,
|
| 461 |
+
"eval_Qnli-dev_max_accuracy": 0.634765625,
|
| 462 |
+
"eval_Qnli-dev_max_accuracy_threshold": 737.3067016601562,
|
| 463 |
+
"eval_Qnli-dev_max_ap": 0.6173131758641012,
|
| 464 |
+
"eval_Qnli-dev_max_f1": 0.6315789473684211,
|
| 465 |
+
"eval_Qnli-dev_max_f1_threshold": 463.4846496582031,
|
| 466 |
+
"eval_Qnli-dev_max_precision": 0.4633663366336634,
|
| 467 |
+
"eval_Qnli-dev_max_recall": 0.9957627118644068,
|
| 468 |
+
"eval_allNLI-dev_cosine_accuracy": 0.6640625,
|
| 469 |
+
"eval_allNLI-dev_cosine_accuracy_threshold": 0.9944478273391724,
|
| 470 |
+
"eval_allNLI-dev_cosine_ap": 0.37265957387885357,
|
| 471 |
+
"eval_allNLI-dev_cosine_f1": 0.5104477611940298,
|
| 472 |
+
"eval_allNLI-dev_cosine_f1_threshold": 0.8620990514755249,
|
| 473 |
+
"eval_allNLI-dev_cosine_precision": 0.3440643863179074,
|
| 474 |
+
"eval_allNLI-dev_cosine_recall": 0.9884393063583815,
|
| 475 |
+
"eval_allNLI-dev_dot_accuracy": 0.66015625,
|
| 476 |
+
"eval_allNLI-dev_dot_accuracy_threshold": 1061.320556640625,
|
| 477 |
+
"eval_allNLI-dev_dot_ap": 0.329964611589565,
|
| 478 |
+
"eval_allNLI-dev_dot_f1": 0.5088235294117648,
|
| 479 |
+
"eval_allNLI-dev_dot_f1_threshold": 606.7950439453125,
|
| 480 |
+
"eval_allNLI-dev_dot_precision": 0.34122287968441817,
|
| 481 |
+
"eval_allNLI-dev_dot_recall": 1.0,
|
| 482 |
+
"eval_allNLI-dev_euclidean_accuracy": 0.666015625,
|
| 483 |
+
"eval_allNLI-dev_euclidean_accuracy_threshold": 5.134191513061523,
|
| 484 |
+
"eval_allNLI-dev_euclidean_ap": 0.36839120834109285,
|
| 485 |
+
"eval_allNLI-dev_euclidean_f1": 0.5113464447806354,
|
| 486 |
+
"eval_allNLI-dev_euclidean_f1_threshold": 14.78046989440918,
|
| 487 |
+
"eval_allNLI-dev_euclidean_precision": 0.3463114754098361,
|
| 488 |
+
"eval_allNLI-dev_euclidean_recall": 0.976878612716763,
|
| 489 |
+
"eval_allNLI-dev_manhattan_accuracy": 0.6640625,
|
| 490 |
+
"eval_allNLI-dev_manhattan_accuracy_threshold": 54.993194580078125,
|
| 491 |
+
"eval_allNLI-dev_manhattan_ap": 0.38115647907317707,
|
| 492 |
+
"eval_allNLI-dev_manhattan_f1": 0.5111441307578009,
|
| 493 |
+
"eval_allNLI-dev_manhattan_f1_threshold": 225.10543823242188,
|
| 494 |
+
"eval_allNLI-dev_manhattan_precision": 0.344,
|
| 495 |
+
"eval_allNLI-dev_manhattan_recall": 0.9942196531791907,
|
| 496 |
+
"eval_allNLI-dev_max_accuracy": 0.666015625,
|
| 497 |
+
"eval_allNLI-dev_max_accuracy_threshold": 1061.320556640625,
|
| 498 |
+
"eval_allNLI-dev_max_ap": 0.38115647907317707,
|
| 499 |
+
"eval_allNLI-dev_max_f1": 0.5113464447806354,
|
| 500 |
+
"eval_allNLI-dev_max_f1_threshold": 606.7950439453125,
|
| 501 |
+
"eval_allNLI-dev_max_precision": 0.3463114754098361,
|
| 502 |
+
"eval_allNLI-dev_max_recall": 1.0,
|
| 503 |
+
"eval_sequential_score": 0.6173131758641012,
|
| 504 |
+
"eval_sts-test_pearson_cosine": 0.15527534331116627,
|
| 505 |
+
"eval_sts-test_pearson_dot": 0.1562898413481788,
|
| 506 |
+
"eval_sts-test_pearson_euclidean": 0.1565855982058341,
|
| 507 |
+
"eval_sts-test_pearson_manhattan": 0.14924616505059007,
|
| 508 |
+
"eval_sts-test_pearson_max": 0.1565855982058341,
|
| 509 |
+
"eval_sts-test_spearman_cosine": 0.2096244403759917,
|
| 510 |
+
"eval_sts-test_spearman_dot": 0.155353153646277,
|
| 511 |
+
"eval_sts-test_spearman_euclidean": 0.1821313239351398,
|
| 512 |
+
"eval_sts-test_spearman_manhattan": 0.174985308262377,
|
| 513 |
+
"eval_sts-test_spearman_max": 0.2096244403759917,
|
| 514 |
+
"eval_vitaminc-pairs_loss": 3.995089530944824,
|
| 515 |
+
"eval_vitaminc-pairs_runtime": 5.4908,
|
| 516 |
+
"eval_vitaminc-pairs_samples_per_second": 23.312,
|
| 517 |
+
"eval_vitaminc-pairs_steps_per_second": 0.364,
|
| 518 |
+
"step": 60
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.04654771140418929,
|
| 522 |
+
"eval_negation-triplets_loss": 5.3504109382629395,
|
| 523 |
+
"eval_negation-triplets_runtime": 1.115,
|
| 524 |
+
"eval_negation-triplets_samples_per_second": 114.8,
|
| 525 |
+
"eval_negation-triplets_steps_per_second": 1.794,
|
| 526 |
+
"step": 60
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 0.04654771140418929,
|
| 530 |
+
"eval_scitail-pairs-pos_loss": 2.7031445503234863,
|
| 531 |
+
"eval_scitail-pairs-pos_runtime": 1.238,
|
| 532 |
+
"eval_scitail-pairs-pos_samples_per_second": 103.395,
|
| 533 |
+
"eval_scitail-pairs-pos_steps_per_second": 1.616,
|
| 534 |
+
"step": 60
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.04654771140418929,
|
| 538 |
+
"eval_scitail-pairs-qa_loss": 3.3194987773895264,
|
| 539 |
+
"eval_scitail-pairs-qa_runtime": 0.9091,
|
| 540 |
+
"eval_scitail-pairs-qa_samples_per_second": 140.792,
|
| 541 |
+
"eval_scitail-pairs-qa_steps_per_second": 2.2,
|
| 542 |
+
"step": 60
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.04654771140418929,
|
| 546 |
+
"eval_xsum-pairs_loss": 5.565915584564209,
|
| 547 |
+
"eval_xsum-pairs_runtime": 4.303,
|
| 548 |
+
"eval_xsum-pairs_samples_per_second": 29.747,
|
| 549 |
+
"eval_xsum-pairs_steps_per_second": 0.465,
|
| 550 |
+
"step": 60
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.04654771140418929,
|
| 554 |
+
"eval_sciq_pairs_loss": 0.9480071067810059,
|
| 555 |
+
"eval_sciq_pairs_runtime": 5.5715,
|
| 556 |
+
"eval_sciq_pairs_samples_per_second": 22.974,
|
| 557 |
+
"eval_sciq_pairs_steps_per_second": 0.359,
|
| 558 |
+
"step": 60
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 0.04654771140418929,
|
| 562 |
+
"eval_qasc_pairs_loss": 6.708042144775391,
|
| 563 |
+
"eval_qasc_pairs_runtime": 0.978,
|
| 564 |
+
"eval_qasc_pairs_samples_per_second": 130.881,
|
| 565 |
+
"eval_qasc_pairs_steps_per_second": 2.045,
|
| 566 |
+
"step": 60
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.04654771140418929,
|
| 570 |
+
"eval_openbookqa_pairs_loss": 5.462226390838623,
|
| 571 |
+
"eval_openbookqa_pairs_runtime": 0.8592,
|
| 572 |
+
"eval_openbookqa_pairs_samples_per_second": 148.973,
|
| 573 |
+
"eval_openbookqa_pairs_steps_per_second": 2.328,
|
| 574 |
+
"step": 60
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 0.04654771140418929,
|
| 578 |
+
"eval_msmarco_pairs_loss": 8.80248737335205,
|
| 579 |
+
"eval_msmarco_pairs_runtime": 1.9622,
|
| 580 |
+
"eval_msmarco_pairs_samples_per_second": 65.234,
|
| 581 |
+
"eval_msmarco_pairs_steps_per_second": 1.019,
|
| 582 |
+
"step": 60
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 0.04654771140418929,
|
| 586 |
+
"eval_nq_pairs_loss": 7.891404628753662,
|
| 587 |
+
"eval_nq_pairs_runtime": 4.4106,
|
| 588 |
+
"eval_nq_pairs_samples_per_second": 29.021,
|
| 589 |
+
"eval_nq_pairs_steps_per_second": 0.453,
|
| 590 |
+
"step": 60
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.04654771140418929,
|
| 594 |
+
"eval_trivia_pairs_loss": 6.30972146987915,
|
| 595 |
+
"eval_trivia_pairs_runtime": 4.5239,
|
| 596 |
+
"eval_trivia_pairs_samples_per_second": 28.294,
|
| 597 |
+
"eval_trivia_pairs_steps_per_second": 0.442,
|
| 598 |
+
"step": 60
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.04654771140418929,
|
| 602 |
+
"eval_gooaq_pairs_loss": 6.851067066192627,
|
| 603 |
+
"eval_gooaq_pairs_runtime": 1.2706,
|
| 604 |
+
"eval_gooaq_pairs_samples_per_second": 100.74,
|
| 605 |
+
"eval_gooaq_pairs_steps_per_second": 1.574,
|
| 606 |
+
"step": 60
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.04654771140418929,
|
| 610 |
+
"eval_paws-pos_loss": 2.74210262298584,
|
| 611 |
+
"eval_paws-pos_runtime": 1.0213,
|
| 612 |
+
"eval_paws-pos_samples_per_second": 125.326,
|
| 613 |
+
"eval_paws-pos_steps_per_second": 1.958,
|
| 614 |
+
"step": 60
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.04654771140418929,
|
| 618 |
+
"eval_global_dataset_loss": 5.345330715179443,
|
| 619 |
+
"eval_global_dataset_runtime": 21.1331,
|
| 620 |
+
"eval_global_dataset_samples_per_second": 19.685,
|
| 621 |
+
"eval_global_dataset_steps_per_second": 0.331,
|
| 622 |
+
"step": 60
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.047323506594259115,
|
| 626 |
+
"grad_norm": 15.78637981414795,
|
| 627 |
+
"learning_rate": 2.426556991774383e-06,
|
| 628 |
+
"loss": 3.5705,
|
| 629 |
+
"step": 61
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 0.04809930178432894,
|
| 633 |
+
"grad_norm": 7.229628562927246,
|
| 634 |
+
"learning_rate": 2.4676850763807284e-06,
|
| 635 |
+
"loss": 5.8784,
|
| 636 |
+
"step": 62
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 0.048875096974398756,
|
| 640 |
+
"grad_norm": 35.124271392822266,
|
| 641 |
+
"learning_rate": 2.5088131609870737e-06,
|
| 642 |
+
"loss": 8.5966,
|
| 643 |
+
"step": 63
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 0.04965089216446858,
|
| 647 |
+
"grad_norm": 10.43959903717041,
|
| 648 |
+
"learning_rate": 2.5499412455934194e-06,
|
| 649 |
+
"loss": 5.9225,
|
| 650 |
+
"step": 64
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.0504266873545384,
|
| 654 |
+
"grad_norm": 9.716011047363281,
|
| 655 |
+
"learning_rate": 2.5910693301997647e-06,
|
| 656 |
+
"loss": 4.6667,
|
| 657 |
+
"step": 65
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.05120248254460822,
|
| 661 |
+
"grad_norm": 12.058530807495117,
|
| 662 |
+
"learning_rate": 2.63219741480611e-06,
|
| 663 |
+
"loss": 6.12,
|
| 664 |
+
"step": 66
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.051978277734678044,
|
| 668 |
+
"grad_norm": 9.680259704589844,
|
| 669 |
+
"learning_rate": 2.6733254994124557e-06,
|
| 670 |
+
"loss": 5.2713,
|
| 671 |
+
"step": 67
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 0.05275407292474787,
|
| 675 |
+
"grad_norm": 20.48016929626465,
|
| 676 |
+
"learning_rate": 2.7144535840188014e-06,
|
| 677 |
+
"loss": 6.9483,
|
| 678 |
+
"step": 68
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 0.05352986811481769,
|
| 682 |
+
"grad_norm": 15.96925163269043,
|
| 683 |
+
"learning_rate": 2.7555816686251467e-06,
|
| 684 |
+
"loss": 3.6892,
|
| 685 |
+
"step": 69
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.05430566330488751,
|
| 689 |
+
"grad_norm": 25.622777938842773,
|
| 690 |
+
"learning_rate": 2.796709753231492e-06,
|
| 691 |
+
"loss": 7.5618,
|
| 692 |
+
"step": 70
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.05508145849495733,
|
| 696 |
+
"grad_norm": 9.09017562866211,
|
| 697 |
+
"learning_rate": 2.8378378378378377e-06,
|
| 698 |
+
"loss": 5.8086,
|
| 699 |
+
"step": 71
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.055857253685027156,
|
| 703 |
+
"grad_norm": 23.944202423095703,
|
| 704 |
+
"learning_rate": 2.878965922444183e-06,
|
| 705 |
+
"loss": 6.517,
|
| 706 |
+
"step": 72
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.05663304887509697,
|
| 710 |
+
"grad_norm": 9.879794120788574,
|
| 711 |
+
"learning_rate": 2.9200940070505282e-06,
|
| 712 |
+
"loss": 5.5826,
|
| 713 |
+
"step": 73
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 0.057408844065166796,
|
| 717 |
+
"grad_norm": 16.403295516967773,
|
| 718 |
+
"learning_rate": 2.9612220916568744e-06,
|
| 719 |
+
"loss": 6.5543,
|
| 720 |
+
"step": 74
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 0.05818463925523662,
|
| 724 |
+
"grad_norm": 25.534475326538086,
|
| 725 |
+
"learning_rate": 3.0023501762632197e-06,
|
| 726 |
+
"loss": 6.5721,
|
| 727 |
+
"step": 75
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.05896043444530644,
|
| 731 |
+
"grad_norm": 8.420788764953613,
|
| 732 |
+
"learning_rate": 3.043478260869565e-06,
|
| 733 |
+
"loss": 5.4454,
|
| 734 |
+
"step": 76
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.05973622963537626,
|
| 738 |
+
"grad_norm": 20.24285888671875,
|
| 739 |
+
"learning_rate": 3.0846063454759102e-06,
|
| 740 |
+
"loss": 6.3208,
|
| 741 |
+
"step": 77
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.060512024825446084,
|
| 745 |
+
"grad_norm": 15.204730987548828,
|
| 746 |
+
"learning_rate": 3.125734430082256e-06,
|
| 747 |
+
"loss": 6.5299,
|
| 748 |
+
"step": 78
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.0612878200155159,
|
| 752 |
+
"grad_norm": 20.533952713012695,
|
| 753 |
+
"learning_rate": 3.1668625146886012e-06,
|
| 754 |
+
"loss": 7.1424,
|
| 755 |
+
"step": 79
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 0.062063615205585725,
|
| 759 |
+
"grad_norm": 8.96998405456543,
|
| 760 |
+
"learning_rate": 3.2079905992949465e-06,
|
| 761 |
+
"loss": 5.1799,
|
| 762 |
+
"step": 80
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.06283941039565555,
|
| 766 |
+
"grad_norm": NaN,
|
| 767 |
+
"learning_rate": 3.2079905992949465e-06,
|
| 768 |
+
"loss": 6.3358,
|
| 769 |
+
"step": 81
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.06361520558572537,
|
| 773 |
+
"grad_norm": 28.576021194458008,
|
| 774 |
+
"learning_rate": 3.2491186839012927e-06,
|
| 775 |
+
"loss": 7.9256,
|
| 776 |
+
"step": 82
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.0643910007757952,
|
| 780 |
+
"grad_norm": 8.81160831451416,
|
| 781 |
+
"learning_rate": 3.290246768507638e-06,
|
| 782 |
+
"loss": 5.4606,
|
| 783 |
+
"step": 83
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.06516679596586501,
|
| 787 |
+
"grad_norm": 8.355425834655762,
|
| 788 |
+
"learning_rate": 3.3313748531139832e-06,
|
| 789 |
+
"loss": 5.6861,
|
| 790 |
+
"step": 84
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.06594259115593483,
|
| 794 |
+
"grad_norm": 16.12509536743164,
|
| 795 |
+
"learning_rate": 3.372502937720329e-06,
|
| 796 |
+
"loss": 3.6595,
|
| 797 |
+
"step": 85
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.06671838634600466,
|
| 801 |
+
"grad_norm": 7.375875473022461,
|
| 802 |
+
"learning_rate": 3.4136310223266742e-06,
|
| 803 |
+
"loss": 5.3626,
|
| 804 |
+
"step": 86
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.06749418153607448,
|
| 808 |
+
"grad_norm": 7.526387691497803,
|
| 809 |
+
"learning_rate": 3.4547591069330195e-06,
|
| 810 |
+
"loss": 5.7214,
|
| 811 |
+
"step": 87
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 0.0682699767261443,
|
| 815 |
+
"grad_norm": 16.23571014404297,
|
| 816 |
+
"learning_rate": 3.495887191539365e-06,
|
| 817 |
+
"loss": 3.7724,
|
| 818 |
+
"step": 88
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.06904577191621412,
|
| 822 |
+
"grad_norm": 19.215839385986328,
|
| 823 |
+
"learning_rate": 3.5370152761457105e-06,
|
| 824 |
+
"loss": 6.8782,
|
| 825 |
+
"step": 89
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.06982156710628394,
|
| 829 |
+
"grad_norm": 20.674381256103516,
|
| 830 |
+
"learning_rate": 3.5781433607520563e-06,
|
| 831 |
+
"loss": 7.1858,
|
| 832 |
+
"step": 90
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 0.07059736229635376,
|
| 836 |
+
"grad_norm": 12.960572242736816,
|
| 837 |
+
"learning_rate": 3.6192714453584015e-06,
|
| 838 |
+
"loss": 6.1052,
|
| 839 |
+
"step": 91
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 0.07137315748642359,
|
| 843 |
+
"grad_norm": 5.505962371826172,
|
| 844 |
+
"learning_rate": 3.6603995299647473e-06,
|
| 845 |
+
"loss": 5.6906,
|
| 846 |
+
"step": 92
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 0.0721489526764934,
|
| 850 |
+
"grad_norm": 13.981863021850586,
|
| 851 |
+
"learning_rate": 3.7015276145710925e-06,
|
| 852 |
+
"loss": 6.3279,
|
| 853 |
+
"step": 93
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 0.07292474786656322,
|
| 857 |
+
"grad_norm": 15.760379791259766,
|
| 858 |
+
"learning_rate": 3.742655699177438e-06,
|
| 859 |
+
"loss": 6.6561,
|
| 860 |
+
"step": 94
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.07370054305663305,
|
| 864 |
+
"grad_norm": 4.809438705444336,
|
| 865 |
+
"learning_rate": 3.7837837837837835e-06,
|
| 866 |
+
"loss": 5.634,
|
| 867 |
+
"step": 95
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.07447633824670287,
|
| 871 |
+
"grad_norm": 6.97947883605957,
|
| 872 |
+
"learning_rate": 3.824911868390129e-06,
|
| 873 |
+
"loss": 5.6729,
|
| 874 |
+
"step": 96
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.07525213343677269,
|
| 878 |
+
"grad_norm": 6.912349224090576,
|
| 879 |
+
"learning_rate": 3.866039952996474e-06,
|
| 880 |
+
"loss": 5.1085,
|
| 881 |
+
"step": 97
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 0.07602792862684252,
|
| 885 |
+
"grad_norm": 16.25594711303711,
|
| 886 |
+
"learning_rate": 3.907168037602819e-06,
|
| 887 |
+
"loss": 6.7441,
|
| 888 |
+
"step": 98
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.07680372381691233,
|
| 892 |
+
"grad_norm": 12.987081527709961,
|
| 893 |
+
"learning_rate": 3.9482961222091655e-06,
|
| 894 |
+
"loss": 6.2151,
|
| 895 |
+
"step": 99
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.07757951900698215,
|
| 899 |
+
"grad_norm": 16.0250244140625,
|
| 900 |
+
"learning_rate": 3.989424206815511e-06,
|
| 901 |
+
"loss": 6.6636,
|
| 902 |
+
"step": 100
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.07835531419705198,
|
| 906 |
+
"grad_norm": 7.364856243133545,
|
| 907 |
+
"learning_rate": 4.030552291421856e-06,
|
| 908 |
+
"loss": 5.3276,
|
| 909 |
+
"step": 101
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.0791311093871218,
|
| 913 |
+
"grad_norm": 7.465214252471924,
|
| 914 |
+
"learning_rate": 4.071680376028202e-06,
|
| 915 |
+
"loss": 5.1297,
|
| 916 |
+
"step": 102
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.07990690457719161,
|
| 920 |
+
"grad_norm": 7.24612283706665,
|
| 921 |
+
"learning_rate": 4.1128084606345476e-06,
|
| 922 |
+
"loss": 5.0139,
|
| 923 |
+
"step": 103
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.08068269976726145,
|
| 927 |
+
"grad_norm": 6.434986591339111,
|
| 928 |
+
"learning_rate": 4.153936545240893e-06,
|
| 929 |
+
"loss": 5.029,
|
| 930 |
+
"step": 104
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 0.08145849495733126,
|
| 934 |
+
"grad_norm": 4.641157627105713,
|
| 935 |
+
"learning_rate": 4.195064629847238e-06,
|
| 936 |
+
"loss": 5.605,
|
| 937 |
+
"step": 105
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.08223429014740109,
|
| 941 |
+
"grad_norm": 7.490912437438965,
|
| 942 |
+
"learning_rate": 4.236192714453583e-06,
|
| 943 |
+
"loss": 5.0171,
|
| 944 |
+
"step": 106
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.08301008533747091,
|
| 948 |
+
"grad_norm": 17.2343692779541,
|
| 949 |
+
"learning_rate": 4.277320799059929e-06,
|
| 950 |
+
"loss": 3.6544,
|
| 951 |
+
"step": 107
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.08378588052754073,
|
| 955 |
+
"grad_norm": 7.838309288024902,
|
| 956 |
+
"learning_rate": 4.318448883666274e-06,
|
| 957 |
+
"loss": 5.9355,
|
| 958 |
+
"step": 108
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 0.08456167571761056,
|
| 962 |
+
"grad_norm": 8.586014747619629,
|
| 963 |
+
"learning_rate": 4.35957696827262e-06,
|
| 964 |
+
"loss": 5.7908,
|
| 965 |
+
"step": 109
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 0.08533747090768037,
|
| 969 |
+
"grad_norm": 7.035303592681885,
|
| 970 |
+
"learning_rate": 4.400705052878966e-06,
|
| 971 |
+
"loss": 5.2447,
|
| 972 |
+
"step": 110
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 0.08611326609775019,
|
| 976 |
+
"grad_norm": 15.303041458129883,
|
| 977 |
+
"learning_rate": 4.441833137485311e-06,
|
| 978 |
+
"loss": 6.5699,
|
| 979 |
+
"step": 111
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.08688906128782002,
|
| 983 |
+
"grad_norm": 6.91060733795166,
|
| 984 |
+
"learning_rate": 4.482961222091657e-06,
|
| 985 |
+
"loss": 4.8604,
|
| 986 |
+
"step": 112
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.08766485647788984,
|
| 990 |
+
"grad_norm": 4.178639888763428,
|
| 991 |
+
"learning_rate": 4.524089306698001e-06,
|
| 992 |
+
"loss": 5.6599,
|
| 993 |
+
"step": 113
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.08844065166795965,
|
| 997 |
+
"grad_norm": 7.758280277252197,
|
| 998 |
+
"learning_rate": 4.5652173913043474e-06,
|
| 999 |
+
"loss": 4.9483,
|
| 1000 |
+
"step": 114
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 0.08921644685802949,
|
| 1004 |
+
"grad_norm": 8.945616722106934,
|
| 1005 |
+
"learning_rate": 4.606345475910693e-06,
|
| 1006 |
+
"loss": 5.7634,
|
| 1007 |
+
"step": 115
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 0.0899922420480993,
|
| 1011 |
+
"grad_norm": 6.94469690322876,
|
| 1012 |
+
"learning_rate": 4.647473560517038e-06,
|
| 1013 |
+
"loss": 5.0934,
|
| 1014 |
+
"step": 116
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 0.09076803723816912,
|
| 1018 |
+
"grad_norm": 8.813899040222168,
|
| 1019 |
+
"learning_rate": 4.688601645123384e-06,
|
| 1020 |
+
"loss": 4.5253,
|
| 1021 |
+
"step": 117
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 0.09154383242823895,
|
| 1025 |
+
"grad_norm": 10.021103858947754,
|
| 1026 |
+
"learning_rate": 4.7297297297297294e-06,
|
| 1027 |
+
"loss": 4.6447,
|
| 1028 |
+
"step": 118
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.09231962761830877,
|
| 1032 |
+
"grad_norm": 4.420945644378662,
|
| 1033 |
+
"learning_rate": 4.770857814336075e-06,
|
| 1034 |
+
"loss": 5.5944,
|
| 1035 |
+
"step": 119
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.09309542280837858,
|
| 1039 |
+
"grad_norm": 8.930946350097656,
|
| 1040 |
+
"learning_rate": 4.81198589894242e-06,
|
| 1041 |
+
"loss": 4.4379,
|
| 1042 |
+
"step": 120
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.09309542280837858,
|
| 1046 |
+
"eval_Qnli-dev_cosine_accuracy": 0.611328125,
|
| 1047 |
+
"eval_Qnli-dev_cosine_accuracy_threshold": 0.9671535491943359,
|
| 1048 |
+
"eval_Qnli-dev_cosine_ap": 0.5751874558355262,
|
| 1049 |
+
"eval_Qnli-dev_cosine_f1": 0.6291834002677376,
|
| 1050 |
+
"eval_Qnli-dev_cosine_f1_threshold": 0.7754666805267334,
|
| 1051 |
+
"eval_Qnli-dev_cosine_precision": 0.4598825831702544,
|
| 1052 |
+
"eval_Qnli-dev_cosine_recall": 0.9957627118644068,
|
| 1053 |
+
"eval_Qnli-dev_dot_accuracy": 0.548828125,
|
| 1054 |
+
"eval_Qnli-dev_dot_accuracy_threshold": 812.9189453125,
|
| 1055 |
+
"eval_Qnli-dev_dot_ap": 0.47763633423847596,
|
| 1056 |
+
"eval_Qnli-dev_dot_f1": 0.6308724832214765,
|
| 1057 |
+
"eval_Qnli-dev_dot_f1_threshold": 487.3522033691406,
|
| 1058 |
+
"eval_Qnli-dev_dot_precision": 0.46168958742632615,
|
| 1059 |
+
"eval_Qnli-dev_dot_recall": 0.9957627118644068,
|
| 1060 |
+
"eval_Qnli-dev_euclidean_accuracy": 0.609375,
|
| 1061 |
+
"eval_Qnli-dev_euclidean_accuracy_threshold": 8.314947128295898,
|
| 1062 |
+
"eval_Qnli-dev_euclidean_ap": 0.5924255839030361,
|
| 1063 |
+
"eval_Qnli-dev_euclidean_f1": 0.6300268096514745,
|
| 1064 |
+
"eval_Qnli-dev_euclidean_f1_threshold": 16.918926239013672,
|
| 1065 |
+
"eval_Qnli-dev_euclidean_precision": 0.46078431372549017,
|
| 1066 |
+
"eval_Qnli-dev_euclidean_recall": 0.9957627118644068,
|
| 1067 |
+
"eval_Qnli-dev_manhattan_accuracy": 0.630859375,
|
| 1068 |
+
"eval_Qnli-dev_manhattan_accuracy_threshold": 138.335693359375,
|
| 1069 |
+
"eval_Qnli-dev_manhattan_ap": 0.6151670800573863,
|
| 1070 |
+
"eval_Qnli-dev_manhattan_f1": 0.6325966850828729,
|
| 1071 |
+
"eval_Qnli-dev_manhattan_f1_threshold": 208.749267578125,
|
| 1072 |
+
"eval_Qnli-dev_manhattan_precision": 0.4692622950819672,
|
| 1073 |
+
"eval_Qnli-dev_manhattan_recall": 0.9703389830508474,
|
| 1074 |
+
"eval_Qnli-dev_max_accuracy": 0.630859375,
|
| 1075 |
+
"eval_Qnli-dev_max_accuracy_threshold": 812.9189453125,
|
| 1076 |
+
"eval_Qnli-dev_max_ap": 0.6151670800573863,
|
| 1077 |
+
"eval_Qnli-dev_max_f1": 0.6325966850828729,
|
| 1078 |
+
"eval_Qnli-dev_max_f1_threshold": 487.3522033691406,
|
| 1079 |
+
"eval_Qnli-dev_max_precision": 0.4692622950819672,
|
| 1080 |
+
"eval_Qnli-dev_max_recall": 0.9957627118644068,
|
| 1081 |
+
"eval_allNLI-dev_cosine_accuracy": 0.673828125,
|
| 1082 |
+
"eval_allNLI-dev_cosine_accuracy_threshold": 0.9807866811752319,
|
| 1083 |
+
"eval_allNLI-dev_cosine_ap": 0.3981958701708742,
|
| 1084 |
+
"eval_allNLI-dev_cosine_f1": 0.5135951661631419,
|
| 1085 |
+
"eval_allNLI-dev_cosine_f1_threshold": 0.8885090351104736,
|
| 1086 |
+
"eval_allNLI-dev_cosine_precision": 0.3476482617586912,
|
| 1087 |
+
"eval_allNLI-dev_cosine_recall": 0.9826589595375722,
|
| 1088 |
+
"eval_allNLI-dev_dot_accuracy": 0.66015625,
|
| 1089 |
+
"eval_allNLI-dev_dot_accuracy_threshold": 1020.55224609375,
|
| 1090 |
+
"eval_allNLI-dev_dot_ap": 0.3253263107868747,
|
| 1091 |
+
"eval_allNLI-dev_dot_f1": 0.5089285714285714,
|
| 1092 |
+
"eval_allNLI-dev_dot_f1_threshold": 662.3919067382812,
|
| 1093 |
+
"eval_allNLI-dev_dot_precision": 0.342685370741483,
|
| 1094 |
+
"eval_allNLI-dev_dot_recall": 0.9884393063583815,
|
| 1095 |
+
"eval_allNLI-dev_euclidean_accuracy": 0.66796875,
|
| 1096 |
+
"eval_allNLI-dev_euclidean_accuracy_threshold": 5.130853652954102,
|
| 1097 |
+
"eval_allNLI-dev_euclidean_ap": 0.3883458149835009,
|
| 1098 |
+
"eval_allNLI-dev_euclidean_f1": 0.5128205128205128,
|
| 1099 |
+
"eval_allNLI-dev_euclidean_f1_threshold": 14.313749313354492,
|
| 1100 |
+
"eval_allNLI-dev_euclidean_precision": 0.3469387755102041,
|
| 1101 |
+
"eval_allNLI-dev_euclidean_recall": 0.9826589595375722,
|
| 1102 |
+
"eval_allNLI-dev_manhattan_accuracy": 0.66796875,
|
| 1103 |
+
"eval_allNLI-dev_manhattan_accuracy_threshold": 89.57350158691406,
|
| 1104 |
+
"eval_allNLI-dev_manhattan_ap": 0.40232496525900585,
|
| 1105 |
+
"eval_allNLI-dev_manhattan_f1": 0.5089820359281437,
|
| 1106 |
+
"eval_allNLI-dev_manhattan_f1_threshold": 217.869384765625,
|
| 1107 |
+
"eval_allNLI-dev_manhattan_precision": 0.3434343434343434,
|
| 1108 |
+
"eval_allNLI-dev_manhattan_recall": 0.9826589595375722,
|
| 1109 |
+
"eval_allNLI-dev_max_accuracy": 0.673828125,
|
| 1110 |
+
"eval_allNLI-dev_max_accuracy_threshold": 1020.55224609375,
|
| 1111 |
+
"eval_allNLI-dev_max_ap": 0.40232496525900585,
|
| 1112 |
+
"eval_allNLI-dev_max_f1": 0.5135951661631419,
|
| 1113 |
+
"eval_allNLI-dev_max_f1_threshold": 662.3919067382812,
|
| 1114 |
+
"eval_allNLI-dev_max_precision": 0.3476482617586912,
|
| 1115 |
+
"eval_allNLI-dev_max_recall": 0.9884393063583815,
|
| 1116 |
+
"eval_sequential_score": 0.6151670800573863,
|
| 1117 |
+
"eval_sts-test_pearson_cosine": 0.17530815823467685,
|
| 1118 |
+
"eval_sts-test_pearson_dot": 0.18707707355601674,
|
| 1119 |
+
"eval_sts-test_pearson_euclidean": 0.17782964327059717,
|
| 1120 |
+
"eval_sts-test_pearson_manhattan": 0.16938653812007057,
|
| 1121 |
+
"eval_sts-test_pearson_max": 0.18707707355601674,
|
| 1122 |
+
"eval_sts-test_spearman_cosine": 0.23193070661653478,
|
| 1123 |
+
"eval_sts-test_spearman_dot": 0.1855223507670731,
|
| 1124 |
+
"eval_sts-test_spearman_euclidean": 0.2037636361791284,
|
| 1125 |
+
"eval_sts-test_spearman_manhattan": 0.19557191778143276,
|
| 1126 |
+
"eval_sts-test_spearman_max": 0.23193070661653478,
|
| 1127 |
+
"eval_vitaminc-pairs_loss": 3.9059581756591797,
|
| 1128 |
+
"eval_vitaminc-pairs_runtime": 5.4911,
|
| 1129 |
+
"eval_vitaminc-pairs_samples_per_second": 23.311,
|
| 1130 |
+
"eval_vitaminc-pairs_steps_per_second": 0.364,
|
| 1131 |
+
"step": 120
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.09309542280837858,
|
| 1135 |
+
"eval_negation-triplets_loss": 5.277580738067627,
|
| 1136 |
+
"eval_negation-triplets_runtime": 1.1046,
|
| 1137 |
+
"eval_negation-triplets_samples_per_second": 115.878,
|
| 1138 |
+
"eval_negation-triplets_steps_per_second": 1.811,
|
| 1139 |
+
"step": 120
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 0.09309542280837858,
|
| 1143 |
+
"eval_scitail-pairs-pos_loss": 2.574610471725464,
|
| 1144 |
+
"eval_scitail-pairs-pos_runtime": 1.2106,
|
| 1145 |
+
"eval_scitail-pairs-pos_samples_per_second": 105.735,
|
| 1146 |
+
"eval_scitail-pairs-pos_steps_per_second": 1.652,
|
| 1147 |
+
"step": 120
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 0.09309542280837858,
|
| 1151 |
+
"eval_scitail-pairs-qa_loss": 3.310026168823242,
|
| 1152 |
+
"eval_scitail-pairs-qa_runtime": 0.9039,
|
| 1153 |
+
"eval_scitail-pairs-qa_samples_per_second": 141.61,
|
| 1154 |
+
"eval_scitail-pairs-qa_steps_per_second": 2.213,
|
| 1155 |
+
"step": 120
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.09309542280837858,
|
| 1159 |
+
"eval_xsum-pairs_loss": 5.120739936828613,
|
| 1160 |
+
"eval_xsum-pairs_runtime": 4.3092,
|
| 1161 |
+
"eval_xsum-pairs_samples_per_second": 29.704,
|
| 1162 |
+
"eval_xsum-pairs_steps_per_second": 0.464,
|
| 1163 |
+
"step": 120
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 0.09309542280837858,
|
| 1167 |
+
"eval_sciq_pairs_loss": 0.878578245639801,
|
| 1168 |
+
"eval_sciq_pairs_runtime": 5.5099,
|
| 1169 |
+
"eval_sciq_pairs_samples_per_second": 23.231,
|
| 1170 |
+
"eval_sciq_pairs_steps_per_second": 0.363,
|
| 1171 |
+
"step": 120
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.09309542280837858,
|
| 1175 |
+
"eval_qasc_pairs_loss": 5.170502185821533,
|
| 1176 |
+
"eval_qasc_pairs_runtime": 1.0003,
|
| 1177 |
+
"eval_qasc_pairs_samples_per_second": 127.96,
|
| 1178 |
+
"eval_qasc_pairs_steps_per_second": 1.999,
|
| 1179 |
+
"step": 120
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.09309542280837858,
|
| 1183 |
+
"eval_openbookqa_pairs_loss": 5.240739822387695,
|
| 1184 |
+
"eval_openbookqa_pairs_runtime": 0.8534,
|
| 1185 |
+
"eval_openbookqa_pairs_samples_per_second": 149.988,
|
| 1186 |
+
"eval_openbookqa_pairs_steps_per_second": 2.344,
|
| 1187 |
+
"step": 120
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 0.09309542280837858,
|
| 1191 |
+
"eval_msmarco_pairs_loss": 6.330340385437012,
|
| 1192 |
+
"eval_msmarco_pairs_runtime": 1.955,
|
| 1193 |
+
"eval_msmarco_pairs_samples_per_second": 65.475,
|
| 1194 |
+
"eval_msmarco_pairs_steps_per_second": 1.023,
|
| 1195 |
+
"step": 120
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 0.09309542280837858,
|
| 1199 |
+
"eval_nq_pairs_loss": 5.914206504821777,
|
| 1200 |
+
"eval_nq_pairs_runtime": 4.4057,
|
| 1201 |
+
"eval_nq_pairs_samples_per_second": 29.053,
|
| 1202 |
+
"eval_nq_pairs_steps_per_second": 0.454,
|
| 1203 |
+
"step": 120
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.09309542280837858,
|
| 1207 |
+
"eval_trivia_pairs_loss": 5.414165019989014,
|
| 1208 |
+
"eval_trivia_pairs_runtime": 4.5232,
|
| 1209 |
+
"eval_trivia_pairs_samples_per_second": 28.299,
|
| 1210 |
+
"eval_trivia_pairs_steps_per_second": 0.442,
|
| 1211 |
+
"step": 120
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 0.09309542280837858,
|
| 1215 |
+
"eval_gooaq_pairs_loss": 5.54625129699707,
|
| 1216 |
+
"eval_gooaq_pairs_runtime": 1.2712,
|
| 1217 |
+
"eval_gooaq_pairs_samples_per_second": 100.693,
|
| 1218 |
+
"eval_gooaq_pairs_steps_per_second": 1.573,
|
| 1219 |
+
"step": 120
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 0.09309542280837858,
|
| 1223 |
+
"eval_paws-pos_loss": 2.632805585861206,
|
| 1224 |
+
"eval_paws-pos_runtime": 1.0285,
|
| 1225 |
+
"eval_paws-pos_samples_per_second": 124.448,
|
| 1226 |
+
"eval_paws-pos_steps_per_second": 1.944,
|
| 1227 |
+
"step": 120
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.09309542280837858,
|
| 1231 |
+
"eval_global_dataset_loss": 4.557917594909668,
|
| 1232 |
+
"eval_global_dataset_runtime": 21.1343,
|
| 1233 |
+
"eval_global_dataset_samples_per_second": 19.684,
|
| 1234 |
+
"eval_global_dataset_steps_per_second": 0.331,
|
| 1235 |
+
"step": 120
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.09387121799844841,
|
| 1239 |
+
"grad_norm": 8.41954517364502,
|
| 1240 |
+
"learning_rate": 4.853113983548766e-06,
|
| 1241 |
+
"loss": 4.9112,
|
| 1242 |
+
"step": 121
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.09464701318851823,
|
| 1246 |
+
"grad_norm": 11.133651733398438,
|
| 1247 |
+
"learning_rate": 4.894242068155111e-06,
|
| 1248 |
+
"loss": 4.6273,
|
| 1249 |
+
"step": 122
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.09542280837858805,
|
| 1253 |
+
"grad_norm": 8.026440620422363,
|
| 1254 |
+
"learning_rate": 4.935370152761457e-06,
|
| 1255 |
+
"loss": 5.5262,
|
| 1256 |
+
"step": 123
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.09619860356865788,
|
| 1260 |
+
"grad_norm": 8.465253829956055,
|
| 1261 |
+
"learning_rate": 4.976498237367803e-06,
|
| 1262 |
+
"loss": 4.837,
|
| 1263 |
+
"step": 124
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.0969743987587277,
|
| 1267 |
+
"grad_norm": 16.9398193359375,
|
| 1268 |
+
"learning_rate": 5.017626321974147e-06,
|
| 1269 |
+
"loss": 6.3641,
|
| 1270 |
+
"step": 125
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.09775019394879751,
|
| 1274 |
+
"grad_norm": 8.868680953979492,
|
| 1275 |
+
"learning_rate": 5.0587544065804934e-06,
|
| 1276 |
+
"loss": 4.6542,
|
| 1277 |
+
"step": 126
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.09852598913886734,
|
| 1281 |
+
"grad_norm": 21.058149337768555,
|
| 1282 |
+
"learning_rate": 5.099882491186839e-06,
|
| 1283 |
+
"loss": 3.2895,
|
| 1284 |
+
"step": 127
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.09930178432893716,
|
| 1288 |
+
"grad_norm": 10.110532760620117,
|
| 1289 |
+
"learning_rate": 5.141010575793184e-06,
|
| 1290 |
+
"loss": 5.5488,
|
| 1291 |
+
"step": 128
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.10007757951900698,
|
| 1295 |
+
"grad_norm": 17.80249786376953,
|
| 1296 |
+
"learning_rate": 5.182138660399529e-06,
|
| 1297 |
+
"loss": 6.3668,
|
| 1298 |
+
"step": 129
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.1008533747090768,
|
| 1302 |
+
"grad_norm": 11.01941204071045,
|
| 1303 |
+
"learning_rate": 5.2232667450058754e-06,
|
| 1304 |
+
"loss": 6.0549,
|
| 1305 |
+
"step": 130
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.10162916989914662,
|
| 1309 |
+
"grad_norm": 9.745173454284668,
|
| 1310 |
+
"learning_rate": 5.26439482961222e-06,
|
| 1311 |
+
"loss": 4.7537,
|
| 1312 |
+
"step": 131
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.10240496508921644,
|
| 1316 |
+
"grad_norm": 10.010377883911133,
|
| 1317 |
+
"learning_rate": 5.305522914218566e-06,
|
| 1318 |
+
"loss": 4.245,
|
| 1319 |
+
"step": 132
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.10318076027928627,
|
| 1323 |
+
"grad_norm": 13.334067344665527,
|
| 1324 |
+
"learning_rate": 5.346650998824911e-06,
|
| 1325 |
+
"loss": 4.3526,
|
| 1326 |
+
"step": 133
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.10395655546935609,
|
| 1330 |
+
"grad_norm": 23.037643432617188,
|
| 1331 |
+
"learning_rate": 5.387779083431257e-06,
|
| 1332 |
+
"loss": 2.9629,
|
| 1333 |
+
"step": 134
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.10473235065942592,
|
| 1337 |
+
"grad_norm": 10.07296085357666,
|
| 1338 |
+
"learning_rate": 5.428907168037603e-06,
|
| 1339 |
+
"loss": 4.4471,
|
| 1340 |
+
"step": 135
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.10550814584949574,
|
| 1344 |
+
"grad_norm": 9.801810264587402,
|
| 1345 |
+
"learning_rate": 5.470035252643947e-06,
|
| 1346 |
+
"loss": 4.5066,
|
| 1347 |
+
"step": 136
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.10628394103956555,
|
| 1351 |
+
"grad_norm": 9.673730850219727,
|
| 1352 |
+
"learning_rate": 5.511163337250293e-06,
|
| 1353 |
+
"loss": 4.7698,
|
| 1354 |
+
"step": 137
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.10705973622963538,
|
| 1358 |
+
"grad_norm": 15.595386505126953,
|
| 1359 |
+
"learning_rate": 5.5522914218566394e-06,
|
| 1360 |
+
"loss": 4.706,
|
| 1361 |
+
"step": 138
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.1078355314197052,
|
| 1365 |
+
"grad_norm": 10.181653022766113,
|
| 1366 |
+
"learning_rate": 5.593419506462984e-06,
|
| 1367 |
+
"loss": 4.4883,
|
| 1368 |
+
"step": 139
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.10861132660977502,
|
| 1372 |
+
"grad_norm": 16.41044807434082,
|
| 1373 |
+
"learning_rate": 5.63454759106933e-06,
|
| 1374 |
+
"loss": 6.0553,
|
| 1375 |
+
"step": 140
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.10938712179984485,
|
| 1379 |
+
"grad_norm": 15.553648948669434,
|
| 1380 |
+
"learning_rate": 5.675675675675675e-06,
|
| 1381 |
+
"loss": 5.8958,
|
| 1382 |
+
"step": 141
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.11016291698991466,
|
| 1386 |
+
"grad_norm": 11.734206199645996,
|
| 1387 |
+
"learning_rate": 5.716803760282021e-06,
|
| 1388 |
+
"loss": 4.1842,
|
| 1389 |
+
"step": 142
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.11093871217998448,
|
| 1393 |
+
"grad_norm": 20.94911766052246,
|
| 1394 |
+
"learning_rate": 5.757931844888366e-06,
|
| 1395 |
+
"loss": 6.3887,
|
| 1396 |
+
"step": 143
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.11171450737005431,
|
| 1400 |
+
"grad_norm": 17.768552780151367,
|
| 1401 |
+
"learning_rate": 5.799059929494712e-06,
|
| 1402 |
+
"loss": 4.0725,
|
| 1403 |
+
"step": 144
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.11249030256012413,
|
| 1407 |
+
"grad_norm": 11.716679573059082,
|
| 1408 |
+
"learning_rate": 5.8401880141010565e-06,
|
| 1409 |
+
"loss": 4.6545,
|
| 1410 |
+
"step": 145
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.11326609775019394,
|
| 1414 |
+
"grad_norm": 16.473339080810547,
|
| 1415 |
+
"learning_rate": 5.881316098707403e-06,
|
| 1416 |
+
"loss": 6.1092,
|
| 1417 |
+
"step": 146
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.11404189294026378,
|
| 1421 |
+
"grad_norm": 11.649435997009277,
|
| 1422 |
+
"learning_rate": 5.922444183313749e-06,
|
| 1423 |
+
"loss": 3.7272,
|
| 1424 |
+
"step": 147
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.11481768813033359,
|
| 1428 |
+
"grad_norm": 12.287504196166992,
|
| 1429 |
+
"learning_rate": 5.963572267920093e-06,
|
| 1430 |
+
"loss": 4.1651,
|
| 1431 |
+
"step": 148
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.11559348332040341,
|
| 1435 |
+
"grad_norm": 12.262931823730469,
|
| 1436 |
+
"learning_rate": 6.004700352526439e-06,
|
| 1437 |
+
"loss": 3.8952,
|
| 1438 |
+
"step": 149
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.11636927851047324,
|
| 1442 |
+
"grad_norm": 14.264664649963379,
|
| 1443 |
+
"learning_rate": 6.045828437132785e-06,
|
| 1444 |
+
"loss": 4.6196,
|
| 1445 |
+
"step": 150
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.11714507370054306,
|
| 1449 |
+
"grad_norm": 12.257019996643066,
|
| 1450 |
+
"learning_rate": 6.08695652173913e-06,
|
| 1451 |
+
"loss": 3.7705,
|
| 1452 |
+
"step": 151
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.11792086889061287,
|
| 1456 |
+
"grad_norm": 29.75714874267578,
|
| 1457 |
+
"learning_rate": 6.128084606345475e-06,
|
| 1458 |
+
"loss": 6.073,
|
| 1459 |
+
"step": 152
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.1186966640806827,
|
| 1463 |
+
"grad_norm": 11.830326080322266,
|
| 1464 |
+
"learning_rate": 6.1692126909518205e-06,
|
| 1465 |
+
"loss": 3.7738,
|
| 1466 |
+
"step": 153
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.11947245927075252,
|
| 1470 |
+
"grad_norm": 17.84813117980957,
|
| 1471 |
+
"learning_rate": 6.210340775558167e-06,
|
| 1472 |
+
"loss": 3.6523,
|
| 1473 |
+
"step": 154
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.12024825446082234,
|
| 1477 |
+
"grad_norm": 18.622961044311523,
|
| 1478 |
+
"learning_rate": 6.251468860164512e-06,
|
| 1479 |
+
"loss": 5.4226,
|
| 1480 |
+
"step": 155
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.12102404965089217,
|
| 1484 |
+
"grad_norm": 11.234856605529785,
|
| 1485 |
+
"learning_rate": 6.292596944770857e-06,
|
| 1486 |
+
"loss": 4.5508,
|
| 1487 |
+
"step": 156
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.12179984484096198,
|
| 1491 |
+
"grad_norm": 12.604881286621094,
|
| 1492 |
+
"learning_rate": 6.3337250293772025e-06,
|
| 1493 |
+
"loss": 3.9043,
|
| 1494 |
+
"step": 157
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.1225756400310318,
|
| 1498 |
+
"grad_norm": 13.025086402893066,
|
| 1499 |
+
"learning_rate": 6.374853113983549e-06,
|
| 1500 |
+
"loss": 3.66,
|
| 1501 |
+
"step": 158
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.12335143522110163,
|
| 1505 |
+
"grad_norm": 36.854610443115234,
|
| 1506 |
+
"learning_rate": 6.415981198589893e-06,
|
| 1507 |
+
"loss": 6.0984,
|
| 1508 |
+
"step": 159
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.12412723041117145,
|
| 1512 |
+
"grad_norm": 11.726686477661133,
|
| 1513 |
+
"learning_rate": 6.457109283196239e-06,
|
| 1514 |
+
"loss": 3.9498,
|
| 1515 |
+
"step": 160
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.12490302560124127,
|
| 1519 |
+
"grad_norm": 11.277478218078613,
|
| 1520 |
+
"learning_rate": 6.498237367802585e-06,
|
| 1521 |
+
"loss": 4.4688,
|
| 1522 |
+
"step": 161
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.1256788207913111,
|
| 1526 |
+
"grad_norm": 13.619148254394531,
|
| 1527 |
+
"learning_rate": 6.53936545240893e-06,
|
| 1528 |
+
"loss": 3.6255,
|
| 1529 |
+
"step": 162
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.1264546159813809,
|
| 1533 |
+
"grad_norm": 13.29480266571045,
|
| 1534 |
+
"learning_rate": 6.580493537015276e-06,
|
| 1535 |
+
"loss": 3.658,
|
| 1536 |
+
"step": 163
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.12723041117145073,
|
| 1540 |
+
"grad_norm": 12.314528465270996,
|
| 1541 |
+
"learning_rate": 6.621621621621621e-06,
|
| 1542 |
+
"loss": 3.4856,
|
| 1543 |
+
"step": 164
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.12800620636152055,
|
| 1547 |
+
"grad_norm": 13.117006301879883,
|
| 1548 |
+
"learning_rate": 6.6627497062279665e-06,
|
| 1549 |
+
"loss": 5.3092,
|
| 1550 |
+
"step": 165
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.1287820015515904,
|
| 1554 |
+
"grad_norm": 16.65713882446289,
|
| 1555 |
+
"learning_rate": 6.703877790834312e-06,
|
| 1556 |
+
"loss": 3.7321,
|
| 1557 |
+
"step": 166
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.1295577967416602,
|
| 1561 |
+
"grad_norm": 13.305086135864258,
|
| 1562 |
+
"learning_rate": 6.745005875440658e-06,
|
| 1563 |
+
"loss": 3.2995,
|
| 1564 |
+
"step": 167
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 0.13033359193173003,
|
| 1568 |
+
"grad_norm": 17.29372215270996,
|
| 1569 |
+
"learning_rate": 6.786133960047002e-06,
|
| 1570 |
+
"loss": 5.1161,
|
| 1571 |
+
"step": 168
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.13110938712179984,
|
| 1575 |
+
"grad_norm": 12.293781280517578,
|
| 1576 |
+
"learning_rate": 6.8272620446533485e-06,
|
| 1577 |
+
"loss": 3.614,
|
| 1578 |
+
"step": 169
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.13188518231186966,
|
| 1582 |
+
"grad_norm": 14.023563385009766,
|
| 1583 |
+
"learning_rate": 6.868390129259695e-06,
|
| 1584 |
+
"loss": 4.0901,
|
| 1585 |
+
"step": 170
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 0.13266097750193948,
|
| 1589 |
+
"grad_norm": 12.515665054321289,
|
| 1590 |
+
"learning_rate": 6.909518213866039e-06,
|
| 1591 |
+
"loss": 3.4437,
|
| 1592 |
+
"step": 171
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.13343677269200932,
|
| 1596 |
+
"grad_norm": 19.10329246520996,
|
| 1597 |
+
"learning_rate": 6.950646298472385e-06,
|
| 1598 |
+
"loss": 5.0212,
|
| 1599 |
+
"step": 172
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.13421256788207914,
|
| 1603 |
+
"grad_norm": 22.987899780273438,
|
| 1604 |
+
"learning_rate": 6.99177438307873e-06,
|
| 1605 |
+
"loss": 1.3904,
|
| 1606 |
+
"step": 173
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 0.13498836307214895,
|
| 1610 |
+
"grad_norm": 23.559511184692383,
|
| 1611 |
+
"learning_rate": 7.032902467685076e-06,
|
| 1612 |
+
"loss": 5.6536,
|
| 1613 |
+
"step": 174
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 0.13576415826221877,
|
| 1617 |
+
"grad_norm": 15.472027778625488,
|
| 1618 |
+
"learning_rate": 7.074030552291421e-06,
|
| 1619 |
+
"loss": 5.0981,
|
| 1620 |
+
"step": 175
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.1365399534522886,
|
| 1624 |
+
"grad_norm": 15.673625946044922,
|
| 1625 |
+
"learning_rate": 7.115158636897766e-06,
|
| 1626 |
+
"loss": 4.7528,
|
| 1627 |
+
"step": 176
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 0.1373157486423584,
|
| 1631 |
+
"grad_norm": 31.196725845336914,
|
| 1632 |
+
"learning_rate": 7.1562867215041125e-06,
|
| 1633 |
+
"loss": 5.4556,
|
| 1634 |
+
"step": 177
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.13809154383242825,
|
| 1638 |
+
"grad_norm": 13.222643852233887,
|
| 1639 |
+
"learning_rate": 7.197414806110458e-06,
|
| 1640 |
+
"loss": 2.8553,
|
| 1641 |
+
"step": 178
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.13886733902249807,
|
| 1645 |
+
"grad_norm": 25.982681274414062,
|
| 1646 |
+
"learning_rate": 7.238542890716803e-06,
|
| 1647 |
+
"loss": 5.4703,
|
| 1648 |
+
"step": 179
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.13964313421256788,
|
| 1652 |
+
"grad_norm": 20.191654205322266,
|
| 1653 |
+
"learning_rate": 7.279670975323148e-06,
|
| 1654 |
+
"loss": 4.7665,
|
| 1655 |
+
"step": 180
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 0.13964313421256788,
|
| 1659 |
+
"eval_Qnli-dev_cosine_accuracy": 0.634765625,
|
| 1660 |
+
"eval_Qnli-dev_cosine_accuracy_threshold": 0.9040530323982239,
|
| 1661 |
+
"eval_Qnli-dev_cosine_ap": 0.6065022391968738,
|
| 1662 |
+
"eval_Qnli-dev_cosine_f1": 0.6353591160220995,
|
| 1663 |
+
"eval_Qnli-dev_cosine_f1_threshold": 0.7576093673706055,
|
| 1664 |
+
"eval_Qnli-dev_cosine_precision": 0.4713114754098361,
|
| 1665 |
+
"eval_Qnli-dev_cosine_recall": 0.9745762711864406,
|
| 1666 |
+
"eval_Qnli-dev_dot_accuracy": 0.576171875,
|
| 1667 |
+
"eval_Qnli-dev_dot_accuracy_threshold": 628.8133544921875,
|
| 1668 |
+
"eval_Qnli-dev_dot_ap": 0.526966347597625,
|
| 1669 |
+
"eval_Qnli-dev_dot_f1": 0.6307277628032345,
|
| 1670 |
+
"eval_Qnli-dev_dot_f1_threshold": 381.04156494140625,
|
| 1671 |
+
"eval_Qnli-dev_dot_precision": 0.4624505928853755,
|
| 1672 |
+
"eval_Qnli-dev_dot_recall": 0.9915254237288136,
|
| 1673 |
+
"eval_Qnli-dev_euclidean_accuracy": 0.623046875,
|
| 1674 |
+
"eval_Qnli-dev_euclidean_accuracy_threshold": 11.349706649780273,
|
| 1675 |
+
"eval_Qnli-dev_euclidean_ap": 0.6145366635531581,
|
| 1676 |
+
"eval_Qnli-dev_euclidean_f1": 0.6321525885558583,
|
| 1677 |
+
"eval_Qnli-dev_euclidean_f1_threshold": 18.34234619140625,
|
| 1678 |
+
"eval_Qnli-dev_euclidean_precision": 0.46586345381526106,
|
| 1679 |
+
"eval_Qnli-dev_euclidean_recall": 0.9830508474576272,
|
| 1680 |
+
"eval_Qnli-dev_manhattan_accuracy": 0.619140625,
|
| 1681 |
+
"eval_Qnli-dev_manhattan_accuracy_threshold": 173.58065795898438,
|
| 1682 |
+
"eval_Qnli-dev_manhattan_ap": 0.6265850112923517,
|
| 1683 |
+
"eval_Qnli-dev_manhattan_f1": 0.6324324324324324,
|
| 1684 |
+
"eval_Qnli-dev_manhattan_f1_threshold": 324.4728698730469,
|
| 1685 |
+
"eval_Qnli-dev_manhattan_precision": 0.4642857142857143,
|
| 1686 |
+
"eval_Qnli-dev_manhattan_recall": 0.9915254237288136,
|
| 1687 |
+
"eval_Qnli-dev_max_accuracy": 0.634765625,
|
| 1688 |
+
"eval_Qnli-dev_max_accuracy_threshold": 628.8133544921875,
|
| 1689 |
+
"eval_Qnli-dev_max_ap": 0.6265850112923517,
|
| 1690 |
+
"eval_Qnli-dev_max_f1": 0.6353591160220995,
|
| 1691 |
+
"eval_Qnli-dev_max_f1_threshold": 381.04156494140625,
|
| 1692 |
+
"eval_Qnli-dev_max_precision": 0.4713114754098361,
|
| 1693 |
+
"eval_Qnli-dev_max_recall": 0.9915254237288136,
|
| 1694 |
+
"eval_allNLI-dev_cosine_accuracy": 0.685546875,
|
| 1695 |
+
"eval_allNLI-dev_cosine_accuracy_threshold": 0.9723453521728516,
|
| 1696 |
+
"eval_allNLI-dev_cosine_ap": 0.43881901898113645,
|
| 1697 |
+
"eval_allNLI-dev_cosine_f1": 0.5192982456140351,
|
| 1698 |
+
"eval_allNLI-dev_cosine_f1_threshold": 0.8869714736938477,
|
| 1699 |
+
"eval_allNLI-dev_cosine_precision": 0.37279596977329976,
|
| 1700 |
+
"eval_allNLI-dev_cosine_recall": 0.8554913294797688,
|
| 1701 |
+
"eval_allNLI-dev_dot_accuracy": 0.66796875,
|
| 1702 |
+
"eval_allNLI-dev_dot_accuracy_threshold": 798.5744018554688,
|
| 1703 |
+
"eval_allNLI-dev_dot_ap": 0.3613142873320213,
|
| 1704 |
+
"eval_allNLI-dev_dot_f1": 0.5066273932253312,
|
| 1705 |
+
"eval_allNLI-dev_dot_f1_threshold": 449.14764404296875,
|
| 1706 |
+
"eval_allNLI-dev_dot_precision": 0.33992094861660077,
|
| 1707 |
+
"eval_allNLI-dev_dot_recall": 0.9942196531791907,
|
| 1708 |
+
"eval_allNLI-dev_euclidean_accuracy": 0.681640625,
|
| 1709 |
+
"eval_allNLI-dev_euclidean_accuracy_threshold": 6.218497276306152,
|
| 1710 |
+
"eval_allNLI-dev_euclidean_ap": 0.43303972275330854,
|
| 1711 |
+
"eval_allNLI-dev_euclidean_f1": 0.5192307692307693,
|
| 1712 |
+
"eval_allNLI-dev_euclidean_f1_threshold": 14.741231918334961,
|
| 1713 |
+
"eval_allNLI-dev_euclidean_precision": 0.35920177383592017,
|
| 1714 |
+
"eval_allNLI-dev_euclidean_recall": 0.9364161849710982,
|
| 1715 |
+
"eval_allNLI-dev_manhattan_accuracy": 0.6796875,
|
| 1716 |
+
"eval_allNLI-dev_manhattan_accuracy_threshold": 106.66935729980469,
|
| 1717 |
+
"eval_allNLI-dev_manhattan_ap": 0.4411371390279413,
|
| 1718 |
+
"eval_allNLI-dev_manhattan_f1": 0.514820592823713,
|
| 1719 |
+
"eval_allNLI-dev_manhattan_f1_threshold": 245.81549072265625,
|
| 1720 |
+
"eval_allNLI-dev_manhattan_precision": 0.3525641025641026,
|
| 1721 |
+
"eval_allNLI-dev_manhattan_recall": 0.953757225433526,
|
| 1722 |
+
"eval_allNLI-dev_max_accuracy": 0.685546875,
|
| 1723 |
+
"eval_allNLI-dev_max_accuracy_threshold": 798.5744018554688,
|
| 1724 |
+
"eval_allNLI-dev_max_ap": 0.4411371390279413,
|
| 1725 |
+
"eval_allNLI-dev_max_f1": 0.5192982456140351,
|
| 1726 |
+
"eval_allNLI-dev_max_f1_threshold": 449.14764404296875,
|
| 1727 |
+
"eval_allNLI-dev_max_precision": 0.37279596977329976,
|
| 1728 |
+
"eval_allNLI-dev_max_recall": 0.9942196531791907,
|
| 1729 |
+
"eval_sequential_score": 0.6265850112923517,
|
| 1730 |
+
"eval_sts-test_pearson_cosine": 0.21333982770993898,
|
| 1731 |
+
"eval_sts-test_pearson_dot": 0.21730500216309007,
|
| 1732 |
+
"eval_sts-test_pearson_euclidean": 0.23256570859273834,
|
| 1733 |
+
"eval_sts-test_pearson_manhattan": 0.2446910502330658,
|
| 1734 |
+
"eval_sts-test_pearson_max": 0.2446910502330658,
|
| 1735 |
+
"eval_sts-test_spearman_cosine": 0.29322398745215855,
|
| 1736 |
+
"eval_sts-test_spearman_dot": 0.22164131759237862,
|
| 1737 |
+
"eval_sts-test_spearman_euclidean": 0.2638306587716433,
|
| 1738 |
+
"eval_sts-test_spearman_manhattan": 0.27389212962153303,
|
| 1739 |
+
"eval_sts-test_spearman_max": 0.29322398745215855,
|
| 1740 |
+
"eval_vitaminc-pairs_loss": 3.76989483833313,
|
| 1741 |
+
"eval_vitaminc-pairs_runtime": 5.4587,
|
| 1742 |
+
"eval_vitaminc-pairs_samples_per_second": 23.449,
|
| 1743 |
+
"eval_vitaminc-pairs_steps_per_second": 0.366,
|
| 1744 |
+
"step": 180
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 0.13964313421256788,
|
| 1748 |
+
"eval_negation-triplets_loss": 4.8690900802612305,
|
| 1749 |
+
"eval_negation-triplets_runtime": 1.1072,
|
| 1750 |
+
"eval_negation-triplets_samples_per_second": 115.606,
|
| 1751 |
+
"eval_negation-triplets_steps_per_second": 1.806,
|
| 1752 |
+
"step": 180
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.13964313421256788,
|
| 1756 |
+
"eval_scitail-pairs-pos_loss": 1.604143738746643,
|
| 1757 |
+
"eval_scitail-pairs-pos_runtime": 1.2067,
|
| 1758 |
+
"eval_scitail-pairs-pos_samples_per_second": 106.076,
|
| 1759 |
+
"eval_scitail-pairs-pos_steps_per_second": 1.657,
|
| 1760 |
+
"step": 180
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.13964313421256788,
|
| 1764 |
+
"eval_scitail-pairs-qa_loss": 1.654570460319519,
|
| 1765 |
+
"eval_scitail-pairs-qa_runtime": 0.9037,
|
| 1766 |
+
"eval_scitail-pairs-qa_samples_per_second": 141.636,
|
| 1767 |
+
"eval_scitail-pairs-qa_steps_per_second": 2.213,
|
| 1768 |
+
"step": 180
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.13964313421256788,
|
| 1772 |
+
"eval_xsum-pairs_loss": 4.028111457824707,
|
| 1773 |
+
"eval_xsum-pairs_runtime": 4.3063,
|
| 1774 |
+
"eval_xsum-pairs_samples_per_second": 29.724,
|
| 1775 |
+
"eval_xsum-pairs_steps_per_second": 0.464,
|
| 1776 |
+
"step": 180
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 0.13964313421256788,
|
| 1780 |
+
"eval_sciq_pairs_loss": 0.7544235587120056,
|
| 1781 |
+
"eval_sciq_pairs_runtime": 5.4956,
|
| 1782 |
+
"eval_sciq_pairs_samples_per_second": 23.291,
|
| 1783 |
+
"eval_sciq_pairs_steps_per_second": 0.364,
|
| 1784 |
+
"step": 180
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.13964313421256788,
|
| 1788 |
+
"eval_qasc_pairs_loss": 4.187389850616455,
|
| 1789 |
+
"eval_qasc_pairs_runtime": 0.9748,
|
| 1790 |
+
"eval_qasc_pairs_samples_per_second": 131.305,
|
| 1791 |
+
"eval_qasc_pairs_steps_per_second": 2.052,
|
| 1792 |
+
"step": 180
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 0.13964313421256788,
|
| 1796 |
+
"eval_openbookqa_pairs_loss": 5.340748310089111,
|
| 1797 |
+
"eval_openbookqa_pairs_runtime": 0.8705,
|
| 1798 |
+
"eval_openbookqa_pairs_samples_per_second": 147.041,
|
| 1799 |
+
"eval_openbookqa_pairs_steps_per_second": 2.298,
|
| 1800 |
+
"step": 180
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 0.13964313421256788,
|
| 1804 |
+
"eval_msmarco_pairs_loss": 5.431215286254883,
|
| 1805 |
+
"eval_msmarco_pairs_runtime": 1.9525,
|
| 1806 |
+
"eval_msmarco_pairs_samples_per_second": 65.558,
|
| 1807 |
+
"eval_msmarco_pairs_steps_per_second": 1.024,
|
| 1808 |
+
"step": 180
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.13964313421256788,
|
| 1812 |
+
"eval_nq_pairs_loss": 4.6815714836120605,
|
| 1813 |
+
"eval_nq_pairs_runtime": 4.4065,
|
| 1814 |
+
"eval_nq_pairs_samples_per_second": 29.048,
|
| 1815 |
+
"eval_nq_pairs_steps_per_second": 0.454,
|
| 1816 |
+
"step": 180
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 0.13964313421256788,
|
| 1820 |
+
"eval_trivia_pairs_loss": 5.035283088684082,
|
| 1821 |
+
"eval_trivia_pairs_runtime": 4.5305,
|
| 1822 |
+
"eval_trivia_pairs_samples_per_second": 28.253,
|
| 1823 |
+
"eval_trivia_pairs_steps_per_second": 0.441,
|
| 1824 |
+
"step": 180
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.13964313421256788,
|
| 1828 |
+
"eval_gooaq_pairs_loss": 4.48758602142334,
|
| 1829 |
+
"eval_gooaq_pairs_runtime": 1.2618,
|
| 1830 |
+
"eval_gooaq_pairs_samples_per_second": 101.442,
|
| 1831 |
+
"eval_gooaq_pairs_steps_per_second": 1.585,
|
| 1832 |
+
"step": 180
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 0.13964313421256788,
|
| 1836 |
+
"eval_paws-pos_loss": 0.5738699436187744,
|
| 1837 |
+
"eval_paws-pos_runtime": 1.0233,
|
| 1838 |
+
"eval_paws-pos_samples_per_second": 125.089,
|
| 1839 |
+
"eval_paws-pos_steps_per_second": 1.955,
|
| 1840 |
+
"step": 180
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 0.13964313421256788,
|
| 1844 |
+
"eval_global_dataset_loss": 3.1652355194091797,
|
| 1845 |
+
"eval_global_dataset_runtime": 21.133,
|
| 1846 |
+
"eval_global_dataset_samples_per_second": 19.685,
|
| 1847 |
+
"eval_global_dataset_steps_per_second": 0.331,
|
| 1848 |
+
"step": 180
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 0.1404189294026377,
|
| 1852 |
+
"grad_norm": 15.412625312805176,
|
| 1853 |
+
"learning_rate": 7.3207990599294945e-06,
|
| 1854 |
+
"loss": 4.0213,
|
| 1855 |
+
"step": 181
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 0.14119472459270752,
|
| 1859 |
+
"grad_norm": 17.41399383544922,
|
| 1860 |
+
"learning_rate": 7.361927144535839e-06,
|
| 1861 |
+
"loss": 4.6874,
|
| 1862 |
+
"step": 182
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 0.14197051978277736,
|
| 1866 |
+
"grad_norm": 13.163263320922852,
|
| 1867 |
+
"learning_rate": 7.403055229142185e-06,
|
| 1868 |
+
"loss": 3.1823,
|
| 1869 |
+
"step": 183
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 0.14274631497284718,
|
| 1873 |
+
"grad_norm": 14.50450325012207,
|
| 1874 |
+
"learning_rate": 7.444183313748531e-06,
|
| 1875 |
+
"loss": 3.1686,
|
| 1876 |
+
"step": 184
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 0.143522110162917,
|
| 1880 |
+
"grad_norm": 20.749032974243164,
|
| 1881 |
+
"learning_rate": 7.485311398354876e-06,
|
| 1882 |
+
"loss": 2.8957,
|
| 1883 |
+
"step": 185
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 0.1442979053529868,
|
| 1887 |
+
"grad_norm": 13.969277381896973,
|
| 1888 |
+
"learning_rate": 7.526439482961222e-06,
|
| 1889 |
+
"loss": 4.5781,
|
| 1890 |
+
"step": 186
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 0.14507370054305663,
|
| 1894 |
+
"grad_norm": 15.575841903686523,
|
| 1895 |
+
"learning_rate": 7.567567567567567e-06,
|
| 1896 |
+
"loss": 3.7329,
|
| 1897 |
+
"step": 187
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 0.14584949573312644,
|
| 1901 |
+
"grad_norm": 14.548502922058105,
|
| 1902 |
+
"learning_rate": 7.608695652173912e-06,
|
| 1903 |
+
"loss": 3.4419,
|
| 1904 |
+
"step": 188
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.1466252909231963,
|
| 1908 |
+
"grad_norm": 24.58466339111328,
|
| 1909 |
+
"learning_rate": 7.649823736780258e-06,
|
| 1910 |
+
"loss": 5.6953,
|
| 1911 |
+
"step": 189
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 0.1474010861132661,
|
| 1915 |
+
"grad_norm": 14.317368507385254,
|
| 1916 |
+
"learning_rate": 7.690951821386603e-06,
|
| 1917 |
+
"loss": 3.1869,
|
| 1918 |
+
"step": 190
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 0.14817688130333592,
|
| 1922 |
+
"grad_norm": 16.537460327148438,
|
| 1923 |
+
"learning_rate": 7.732079905992948e-06,
|
| 1924 |
+
"loss": 3.6055,
|
| 1925 |
+
"step": 191
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 0.14895267649340574,
|
| 1929 |
+
"grad_norm": 20.389019012451172,
|
| 1930 |
+
"learning_rate": 7.773207990599295e-06,
|
| 1931 |
+
"loss": 4.6231,
|
| 1932 |
+
"step": 192
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 0.14972847168347556,
|
| 1936 |
+
"grad_norm": 17.62578773498535,
|
| 1937 |
+
"learning_rate": 7.814336075205639e-06,
|
| 1938 |
+
"loss": 4.2417,
|
| 1939 |
+
"step": 193
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 0.15050426687354537,
|
| 1943 |
+
"grad_norm": 28.10770034790039,
|
| 1944 |
+
"learning_rate": 7.855464159811986e-06,
|
| 1945 |
+
"loss": 5.2779,
|
| 1946 |
+
"step": 194
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 0.15128006206361522,
|
| 1950 |
+
"grad_norm": 19.735567092895508,
|
| 1951 |
+
"learning_rate": 7.896592244418331e-06,
|
| 1952 |
+
"loss": 4.3213,
|
| 1953 |
+
"step": 195
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 0.15205585725368503,
|
| 1957 |
+
"grad_norm": 16.370105743408203,
|
| 1958 |
+
"learning_rate": 7.937720329024676e-06,
|
| 1959 |
+
"loss": 3.1158,
|
| 1960 |
+
"step": 196
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 0.15283165244375485,
|
| 1964 |
+
"grad_norm": 21.58885383605957,
|
| 1965 |
+
"learning_rate": 7.978848413631022e-06,
|
| 1966 |
+
"loss": 2.27,
|
| 1967 |
+
"step": 197
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 0.15360744763382467,
|
| 1971 |
+
"grad_norm": 14.355427742004395,
|
| 1972 |
+
"learning_rate": 8.019976498237367e-06,
|
| 1973 |
+
"loss": 3.5041,
|
| 1974 |
+
"step": 198
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 0.15438324282389448,
|
| 1978 |
+
"grad_norm": 13.562695503234863,
|
| 1979 |
+
"learning_rate": 8.061104582843712e-06,
|
| 1980 |
+
"loss": 2.6007,
|
| 1981 |
+
"step": 199
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 0.1551590380139643,
|
| 1985 |
+
"grad_norm": 14.875197410583496,
|
| 1986 |
+
"learning_rate": 8.102232667450058e-06,
|
| 1987 |
+
"loss": 2.4875,
|
| 1988 |
+
"step": 200
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 0.15593483320403415,
|
| 1992 |
+
"grad_norm": 29.655797958374023,
|
| 1993 |
+
"learning_rate": 8.143360752056405e-06,
|
| 1994 |
+
"loss": 5.3046,
|
| 1995 |
+
"step": 201
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 0.15671062839410396,
|
| 1999 |
+
"grad_norm": 21.172327041625977,
|
| 2000 |
+
"learning_rate": 8.184488836662748e-06,
|
| 2001 |
+
"loss": 3.0582,
|
| 2002 |
+
"step": 202
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 0.15748642358417378,
|
| 2006 |
+
"grad_norm": 24.037527084350586,
|
| 2007 |
+
"learning_rate": 8.225616921269095e-06,
|
| 2008 |
+
"loss": 4.9347,
|
| 2009 |
+
"step": 203
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 0.1582622187742436,
|
| 2013 |
+
"grad_norm": 18.551284790039062,
|
| 2014 |
+
"learning_rate": 8.26674500587544e-06,
|
| 2015 |
+
"loss": 2.855,
|
| 2016 |
+
"step": 204
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 0.1590380139643134,
|
| 2020 |
+
"grad_norm": 18.76624870300293,
|
| 2021 |
+
"learning_rate": 8.307873090481786e-06,
|
| 2022 |
+
"loss": 1.7434,
|
| 2023 |
+
"step": 205
|
| 2024 |
+
},
|
| 2025 |
+
{
|
| 2026 |
+
"epoch": 0.15981380915438323,
|
| 2027 |
+
"grad_norm": 12.969701766967773,
|
| 2028 |
+
"learning_rate": 8.349001175088131e-06,
|
| 2029 |
+
"loss": 3.4045,
|
| 2030 |
+
"step": 206
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 0.16058960434445307,
|
| 2034 |
+
"grad_norm": 15.40457820892334,
|
| 2035 |
+
"learning_rate": 8.390129259694476e-06,
|
| 2036 |
+
"loss": 3.4427,
|
| 2037 |
+
"step": 207
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 0.1613653995345229,
|
| 2041 |
+
"grad_norm": 16.32777976989746,
|
| 2042 |
+
"learning_rate": 8.431257344300822e-06,
|
| 2043 |
+
"loss": 3.3483,
|
| 2044 |
+
"step": 208
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 0.1621411947245927,
|
| 2048 |
+
"grad_norm": 17.573177337646484,
|
| 2049 |
+
"learning_rate": 8.472385428907167e-06,
|
| 2050 |
+
"loss": 1.5883,
|
| 2051 |
+
"step": 209
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 0.16291698991466252,
|
| 2055 |
+
"grad_norm": 28.443233489990234,
|
| 2056 |
+
"learning_rate": 8.513513513513514e-06,
|
| 2057 |
+
"loss": 5.3066,
|
| 2058 |
+
"step": 210
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 0.16369278510473234,
|
| 2062 |
+
"grad_norm": 13.738469123840332,
|
| 2063 |
+
"learning_rate": 8.554641598119857e-06,
|
| 2064 |
+
"loss": 0.6047,
|
| 2065 |
+
"step": 211
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 0.16446858029480219,
|
| 2069 |
+
"grad_norm": 15.555298805236816,
|
| 2070 |
+
"learning_rate": 8.595769682726204e-06,
|
| 2071 |
+
"loss": 0.8018,
|
| 2072 |
+
"step": 212
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 0.165244375484872,
|
| 2076 |
+
"grad_norm": 19.836238861083984,
|
| 2077 |
+
"learning_rate": 8.636897767332548e-06,
|
| 2078 |
+
"loss": 2.8775,
|
| 2079 |
+
"step": 213
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 0.16602017067494182,
|
| 2083 |
+
"grad_norm": 10.346473693847656,
|
| 2084 |
+
"learning_rate": 8.678025851938895e-06,
|
| 2085 |
+
"loss": 0.5198,
|
| 2086 |
+
"step": 214
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 0.16679596586501164,
|
| 2090 |
+
"grad_norm": 15.540273666381836,
|
| 2091 |
+
"learning_rate": 8.71915393654524e-06,
|
| 2092 |
+
"loss": 3.1591,
|
| 2093 |
+
"step": 215
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 0.16757176105508145,
|
| 2097 |
+
"grad_norm": 23.148704528808594,
|
| 2098 |
+
"learning_rate": 8.760282021151586e-06,
|
| 2099 |
+
"loss": 2.7381,
|
| 2100 |
+
"step": 216
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 0.16834755624515127,
|
| 2104 |
+
"grad_norm": 36.07377624511719,
|
| 2105 |
+
"learning_rate": 8.801410105757933e-06,
|
| 2106 |
+
"loss": 5.5722,
|
| 2107 |
+
"step": 217
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 0.16912335143522111,
|
| 2111 |
+
"grad_norm": 31.731229782104492,
|
| 2112 |
+
"learning_rate": 8.842538190364276e-06,
|
| 2113 |
+
"loss": 3.998,
|
| 2114 |
+
"step": 218
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 0.16989914662529093,
|
| 2118 |
+
"grad_norm": 19.629518508911133,
|
| 2119 |
+
"learning_rate": 8.883666274970621e-06,
|
| 2120 |
+
"loss": 2.2858,
|
| 2121 |
+
"step": 219
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 0.17067494181536075,
|
| 2125 |
+
"grad_norm": 17.323598861694336,
|
| 2126 |
+
"learning_rate": 8.924794359576967e-06,
|
| 2127 |
+
"loss": 1.556,
|
| 2128 |
+
"step": 220
|
| 2129 |
+
},
|
| 2130 |
+
{
|
| 2131 |
+
"epoch": 0.17145073700543056,
|
| 2132 |
+
"grad_norm": 20.71698760986328,
|
| 2133 |
+
"learning_rate": 8.965922444183314e-06,
|
| 2134 |
+
"loss": 2.5352,
|
| 2135 |
+
"step": 221
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 0.17222653219550038,
|
| 2139 |
+
"grad_norm": 16.524059295654297,
|
| 2140 |
+
"learning_rate": 9.007050528789659e-06,
|
| 2141 |
+
"loss": 3.1682,
|
| 2142 |
+
"step": 222
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 0.1730023273855702,
|
| 2146 |
+
"grad_norm": 18.413827896118164,
|
| 2147 |
+
"learning_rate": 9.048178613396003e-06,
|
| 2148 |
+
"loss": 2.9073,
|
| 2149 |
+
"step": 223
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 0.17377812257564004,
|
| 2153 |
+
"grad_norm": 18.246061325073242,
|
| 2154 |
+
"learning_rate": 9.08930669800235e-06,
|
| 2155 |
+
"loss": 2.547,
|
| 2156 |
+
"step": 224
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 0.17455391776570986,
|
| 2160 |
+
"grad_norm": 20.940176010131836,
|
| 2161 |
+
"learning_rate": 9.130434782608695e-06,
|
| 2162 |
+
"loss": 4.1815,
|
| 2163 |
+
"step": 225
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 0.17532971295577968,
|
| 2167 |
+
"grad_norm": 22.81971549987793,
|
| 2168 |
+
"learning_rate": 9.17156286721504e-06,
|
| 2169 |
+
"loss": 3.7504,
|
| 2170 |
+
"step": 226
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 0.1761055081458495,
|
| 2174 |
+
"grad_norm": 29.75227165222168,
|
| 2175 |
+
"learning_rate": 9.212690951821385e-06,
|
| 2176 |
+
"loss": 5.033,
|
| 2177 |
+
"step": 227
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 0.1768813033359193,
|
| 2181 |
+
"grad_norm": 27.30958366394043,
|
| 2182 |
+
"learning_rate": 9.253819036427732e-06,
|
| 2183 |
+
"loss": 5.2809,
|
| 2184 |
+
"step": 228
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 0.17765709852598913,
|
| 2188 |
+
"grad_norm": 14.53124713897705,
|
| 2189 |
+
"learning_rate": 9.294947121034076e-06,
|
| 2190 |
+
"loss": 2.5598,
|
| 2191 |
+
"step": 229
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 0.17843289371605897,
|
| 2195 |
+
"grad_norm": 10.330536842346191,
|
| 2196 |
+
"learning_rate": 9.336075205640421e-06,
|
| 2197 |
+
"loss": 0.4476,
|
| 2198 |
+
"step": 230
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 0.1792086889061288,
|
| 2202 |
+
"grad_norm": 16.898452758789062,
|
| 2203 |
+
"learning_rate": 9.377203290246768e-06,
|
| 2204 |
+
"loss": 3.4592,
|
| 2205 |
+
"step": 231
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 0.1799844840961986,
|
| 2209 |
+
"grad_norm": 13.675553321838379,
|
| 2210 |
+
"learning_rate": 9.418331374853114e-06,
|
| 2211 |
+
"loss": 2.9202,
|
| 2212 |
+
"step": 232
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 0.18076027928626842,
|
| 2216 |
+
"grad_norm": 13.829231262207031,
|
| 2217 |
+
"learning_rate": 9.459459459459459e-06,
|
| 2218 |
+
"loss": 1.9092,
|
| 2219 |
+
"step": 233
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 0.18153607447633824,
|
| 2223 |
+
"grad_norm": 19.693933486938477,
|
| 2224 |
+
"learning_rate": 9.500587544065804e-06,
|
| 2225 |
+
"loss": 1.9204,
|
| 2226 |
+
"step": 234
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 0.18231186966640806,
|
| 2230 |
+
"grad_norm": 14.612395286560059,
|
| 2231 |
+
"learning_rate": 9.54171562867215e-06,
|
| 2232 |
+
"loss": 2.083,
|
| 2233 |
+
"step": 235
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 0.1830876648564779,
|
| 2237 |
+
"grad_norm": 12.987626075744629,
|
| 2238 |
+
"learning_rate": 9.582843713278495e-06,
|
| 2239 |
+
"loss": 3.0433,
|
| 2240 |
+
"step": 236
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 0.18386346004654772,
|
| 2244 |
+
"grad_norm": 13.259800910949707,
|
| 2245 |
+
"learning_rate": 9.62397179788484e-06,
|
| 2246 |
+
"loss": 1.5429,
|
| 2247 |
+
"step": 237
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 0.18463925523661753,
|
| 2251 |
+
"grad_norm": 8.798643112182617,
|
| 2252 |
+
"learning_rate": 9.665099882491187e-06,
|
| 2253 |
+
"loss": 0.3347,
|
| 2254 |
+
"step": 238
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 0.18541505042668735,
|
| 2258 |
+
"grad_norm": 13.148309707641602,
|
| 2259 |
+
"learning_rate": 9.706227967097532e-06,
|
| 2260 |
+
"loss": 1.8698,
|
| 2261 |
+
"step": 239
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 0.18619084561675717,
|
| 2265 |
+
"grad_norm": 7.301493167877197,
|
| 2266 |
+
"learning_rate": 9.747356051703878e-06,
|
| 2267 |
+
"loss": 0.3031,
|
| 2268 |
+
"step": 240
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 0.18619084561675717,
|
| 2272 |
+
"eval_Qnli-dev_cosine_accuracy": 0.6328125,
|
| 2273 |
+
"eval_Qnli-dev_cosine_accuracy_threshold": 0.8666848540306091,
|
| 2274 |
+
"eval_Qnli-dev_cosine_ap": 0.6314460649736229,
|
| 2275 |
+
"eval_Qnli-dev_cosine_f1": 0.6366366366366366,
|
| 2276 |
+
"eval_Qnli-dev_cosine_f1_threshold": 0.7626283168792725,
|
| 2277 |
+
"eval_Qnli-dev_cosine_precision": 0.4930232558139535,
|
| 2278 |
+
"eval_Qnli-dev_cosine_recall": 0.8983050847457628,
|
| 2279 |
+
"eval_Qnli-dev_dot_accuracy": 0.6015625,
|
| 2280 |
+
"eval_Qnli-dev_dot_accuracy_threshold": 507.15960693359375,
|
| 2281 |
+
"eval_Qnli-dev_dot_ap": 0.5696093250757446,
|
| 2282 |
+
"eval_Qnli-dev_dot_f1": 0.638680659670165,
|
| 2283 |
+
"eval_Qnli-dev_dot_f1_threshold": 389.59454345703125,
|
| 2284 |
+
"eval_Qnli-dev_dot_precision": 0.494199535962877,
|
| 2285 |
+
"eval_Qnli-dev_dot_recall": 0.902542372881356,
|
| 2286 |
+
"eval_Qnli-dev_euclidean_accuracy": 0.626953125,
|
| 2287 |
+
"eval_Qnli-dev_euclidean_accuracy_threshold": 11.648075103759766,
|
| 2288 |
+
"eval_Qnli-dev_euclidean_ap": 0.630384295725278,
|
| 2289 |
+
"eval_Qnli-dev_euclidean_f1": 0.6362098138747886,
|
| 2290 |
+
"eval_Qnli-dev_euclidean_f1_threshold": 14.59055233001709,
|
| 2291 |
+
"eval_Qnli-dev_euclidean_precision": 0.5295774647887324,
|
| 2292 |
+
"eval_Qnli-dev_euclidean_recall": 0.7966101694915254,
|
| 2293 |
+
"eval_Qnli-dev_manhattan_accuracy": 0.6328125,
|
| 2294 |
+
"eval_Qnli-dev_manhattan_accuracy_threshold": 210.99937438964844,
|
| 2295 |
+
"eval_Qnli-dev_manhattan_ap": 0.6305567474066167,
|
| 2296 |
+
"eval_Qnli-dev_manhattan_f1": 0.6392961876832844,
|
| 2297 |
+
"eval_Qnli-dev_manhattan_f1_threshold": 295.44036865234375,
|
| 2298 |
+
"eval_Qnli-dev_manhattan_precision": 0.48878923766816146,
|
| 2299 |
+
"eval_Qnli-dev_manhattan_recall": 0.923728813559322,
|
| 2300 |
+
"eval_Qnli-dev_max_accuracy": 0.6328125,
|
| 2301 |
+
"eval_Qnli-dev_max_accuracy_threshold": 507.15960693359375,
|
| 2302 |
+
"eval_Qnli-dev_max_ap": 0.6314460649736229,
|
| 2303 |
+
"eval_Qnli-dev_max_f1": 0.6392961876832844,
|
| 2304 |
+
"eval_Qnli-dev_max_f1_threshold": 389.59454345703125,
|
| 2305 |
+
"eval_Qnli-dev_max_precision": 0.5295774647887324,
|
| 2306 |
+
"eval_Qnli-dev_max_recall": 0.923728813559322,
|
| 2307 |
+
"eval_allNLI-dev_cosine_accuracy": 0.693359375,
|
| 2308 |
+
"eval_allNLI-dev_cosine_accuracy_threshold": 0.9488164782524109,
|
| 2309 |
+
"eval_allNLI-dev_cosine_ap": 0.5017454777328199,
|
| 2310 |
+
"eval_allNLI-dev_cosine_f1": 0.5440613026819924,
|
| 2311 |
+
"eval_allNLI-dev_cosine_f1_threshold": 0.8723046779632568,
|
| 2312 |
+
"eval_allNLI-dev_cosine_precision": 0.4068767908309456,
|
| 2313 |
+
"eval_allNLI-dev_cosine_recall": 0.8208092485549133,
|
| 2314 |
+
"eval_allNLI-dev_dot_accuracy": 0.673828125,
|
| 2315 |
+
"eval_allNLI-dev_dot_accuracy_threshold": 668.8308715820312,
|
| 2316 |
+
"eval_allNLI-dev_dot_ap": 0.3969704201342592,
|
| 2317 |
+
"eval_allNLI-dev_dot_f1": 0.5069124423963134,
|
| 2318 |
+
"eval_allNLI-dev_dot_f1_threshold": 449.8724365234375,
|
| 2319 |
+
"eval_allNLI-dev_dot_precision": 0.34518828451882844,
|
| 2320 |
+
"eval_allNLI-dev_dot_recall": 0.953757225433526,
|
| 2321 |
+
"eval_allNLI-dev_euclidean_accuracy": 0.6875,
|
| 2322 |
+
"eval_allNLI-dev_euclidean_accuracy_threshold": 7.274426460266113,
|
| 2323 |
+
"eval_allNLI-dev_euclidean_ap": 0.5011125896179964,
|
| 2324 |
+
"eval_allNLI-dev_euclidean_f1": 0.5429141716566867,
|
| 2325 |
+
"eval_allNLI-dev_euclidean_f1_threshold": 12.62590217590332,
|
| 2326 |
+
"eval_allNLI-dev_euclidean_precision": 0.4146341463414634,
|
| 2327 |
+
"eval_allNLI-dev_euclidean_recall": 0.7861271676300579,
|
| 2328 |
+
"eval_allNLI-dev_manhattan_accuracy": 0.689453125,
|
| 2329 |
+
"eval_allNLI-dev_manhattan_accuracy_threshold": 127.58004760742188,
|
| 2330 |
+
"eval_allNLI-dev_manhattan_ap": 0.5053834291774996,
|
| 2331 |
+
"eval_allNLI-dev_manhattan_f1": 0.5362035225048923,
|
| 2332 |
+
"eval_allNLI-dev_manhattan_f1_threshold": 221.26995849609375,
|
| 2333 |
+
"eval_allNLI-dev_manhattan_precision": 0.40532544378698226,
|
| 2334 |
+
"eval_allNLI-dev_manhattan_recall": 0.791907514450867,
|
| 2335 |
+
"eval_allNLI-dev_max_accuracy": 0.693359375,
|
| 2336 |
+
"eval_allNLI-dev_max_accuracy_threshold": 668.8308715820312,
|
| 2337 |
+
"eval_allNLI-dev_max_ap": 0.5053834291774996,
|
| 2338 |
+
"eval_allNLI-dev_max_f1": 0.5440613026819924,
|
| 2339 |
+
"eval_allNLI-dev_max_f1_threshold": 449.8724365234375,
|
| 2340 |
+
"eval_allNLI-dev_max_precision": 0.4146341463414634,
|
| 2341 |
+
"eval_allNLI-dev_max_recall": 0.953757225433526,
|
| 2342 |
+
"eval_sequential_score": 0.6314460649736229,
|
| 2343 |
+
"eval_sts-test_pearson_cosine": 0.4297034177530014,
|
| 2344 |
+
"eval_sts-test_pearson_dot": 0.3153020725919898,
|
| 2345 |
+
"eval_sts-test_pearson_euclidean": 0.4381297122634254,
|
| 2346 |
+
"eval_sts-test_pearson_manhattan": 0.4527271663445909,
|
| 2347 |
+
"eval_sts-test_pearson_max": 0.4527271663445909,
|
| 2348 |
+
"eval_sts-test_spearman_cosine": 0.49463792972220844,
|
| 2349 |
+
"eval_sts-test_spearman_dot": 0.31075978504991664,
|
| 2350 |
+
"eval_sts-test_spearman_euclidean": 0.455142390722399,
|
| 2351 |
+
"eval_sts-test_spearman_manhattan": 0.46803254405357775,
|
| 2352 |
+
"eval_sts-test_spearman_max": 0.49463792972220844,
|
| 2353 |
+
"eval_vitaminc-pairs_loss": 3.9216716289520264,
|
| 2354 |
+
"eval_vitaminc-pairs_runtime": 5.46,
|
| 2355 |
+
"eval_vitaminc-pairs_samples_per_second": 23.443,
|
| 2356 |
+
"eval_vitaminc-pairs_steps_per_second": 0.366,
|
| 2357 |
+
"step": 240
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 0.18619084561675717,
|
| 2361 |
+
"eval_negation-triplets_loss": 3.8406004905700684,
|
| 2362 |
+
"eval_negation-triplets_runtime": 1.1041,
|
| 2363 |
+
"eval_negation-triplets_samples_per_second": 115.932,
|
| 2364 |
+
"eval_negation-triplets_steps_per_second": 1.811,
|
| 2365 |
+
"step": 240
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 0.18619084561675717,
|
| 2369 |
+
"eval_scitail-pairs-pos_loss": 0.597940981388092,
|
| 2370 |
+
"eval_scitail-pairs-pos_runtime": 1.2771,
|
| 2371 |
+
"eval_scitail-pairs-pos_samples_per_second": 100.224,
|
| 2372 |
+
"eval_scitail-pairs-pos_steps_per_second": 1.566,
|
| 2373 |
+
"step": 240
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 0.18619084561675717,
|
| 2377 |
+
"eval_scitail-pairs-qa_loss": 0.5431724786758423,
|
| 2378 |
+
"eval_scitail-pairs-qa_runtime": 0.9108,
|
| 2379 |
+
"eval_scitail-pairs-qa_samples_per_second": 140.54,
|
| 2380 |
+
"eval_scitail-pairs-qa_steps_per_second": 2.196,
|
| 2381 |
+
"step": 240
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 0.18619084561675717,
|
| 2385 |
+
"eval_xsum-pairs_loss": 2.5905959606170654,
|
| 2386 |
+
"eval_xsum-pairs_runtime": 4.3069,
|
| 2387 |
+
"eval_xsum-pairs_samples_per_second": 29.72,
|
| 2388 |
+
"eval_xsum-pairs_steps_per_second": 0.464,
|
| 2389 |
+
"step": 240
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.18619084561675717,
|
| 2393 |
+
"eval_sciq_pairs_loss": 0.44124865531921387,
|
| 2394 |
+
"eval_sciq_pairs_runtime": 5.5223,
|
| 2395 |
+
"eval_sciq_pairs_samples_per_second": 23.179,
|
| 2396 |
+
"eval_sciq_pairs_steps_per_second": 0.362,
|
| 2397 |
+
"step": 240
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.18619084561675717,
|
| 2401 |
+
"eval_qasc_pairs_loss": 2.3103857040405273,
|
| 2402 |
+
"eval_qasc_pairs_runtime": 0.9809,
|
| 2403 |
+
"eval_qasc_pairs_samples_per_second": 130.488,
|
| 2404 |
+
"eval_qasc_pairs_steps_per_second": 2.039,
|
| 2405 |
+
"step": 240
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.18619084561675717,
|
| 2409 |
+
"eval_openbookqa_pairs_loss": 3.97737979888916,
|
| 2410 |
+
"eval_openbookqa_pairs_runtime": 0.8546,
|
| 2411 |
+
"eval_openbookqa_pairs_samples_per_second": 149.783,
|
| 2412 |
+
"eval_openbookqa_pairs_steps_per_second": 2.34,
|
| 2413 |
+
"step": 240
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 0.18619084561675717,
|
| 2417 |
+
"eval_msmarco_pairs_loss": 3.6929101943969727,
|
| 2418 |
+
"eval_msmarco_pairs_runtime": 1.9549,
|
| 2419 |
+
"eval_msmarco_pairs_samples_per_second": 65.476,
|
| 2420 |
+
"eval_msmarco_pairs_steps_per_second": 1.023,
|
| 2421 |
+
"step": 240
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 0.18619084561675717,
|
| 2425 |
+
"eval_nq_pairs_loss": 3.8183703422546387,
|
| 2426 |
+
"eval_nq_pairs_runtime": 4.4066,
|
| 2427 |
+
"eval_nq_pairs_samples_per_second": 29.047,
|
| 2428 |
+
"eval_nq_pairs_steps_per_second": 0.454,
|
| 2429 |
+
"step": 240
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 0.18619084561675717,
|
| 2433 |
+
"eval_trivia_pairs_loss": 3.726828098297119,
|
| 2434 |
+
"eval_trivia_pairs_runtime": 4.5284,
|
| 2435 |
+
"eval_trivia_pairs_samples_per_second": 28.266,
|
| 2436 |
+
"eval_trivia_pairs_steps_per_second": 0.442,
|
| 2437 |
+
"step": 240
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 0.18619084561675717,
|
| 2441 |
+
"eval_gooaq_pairs_loss": 3.068204402923584,
|
| 2442 |
+
"eval_gooaq_pairs_runtime": 1.2681,
|
| 2443 |
+
"eval_gooaq_pairs_samples_per_second": 100.939,
|
| 2444 |
+
"eval_gooaq_pairs_steps_per_second": 1.577,
|
| 2445 |
+
"step": 240
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.18619084561675717,
|
| 2449 |
+
"eval_paws-pos_loss": 0.11907516419887543,
|
| 2450 |
+
"eval_paws-pos_runtime": 1.0209,
|
| 2451 |
+
"eval_paws-pos_samples_per_second": 125.382,
|
| 2452 |
+
"eval_paws-pos_steps_per_second": 1.959,
|
| 2453 |
+
"step": 240
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.18619084561675717,
|
| 2457 |
+
"eval_global_dataset_loss": 1.8651976585388184,
|
| 2458 |
+
"eval_global_dataset_runtime": 21.15,
|
| 2459 |
+
"eval_global_dataset_samples_per_second": 19.669,
|
| 2460 |
+
"eval_global_dataset_steps_per_second": 0.331,
|
| 2461 |
+
"step": 240
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.186966640806827,
|
| 2465 |
+
"grad_norm": 17.317655563354492,
|
| 2466 |
+
"learning_rate": 9.788484136310221e-06,
|
| 2467 |
+
"loss": 3.2076,
|
| 2468 |
+
"step": 241
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.18774243599689683,
|
| 2472 |
+
"grad_norm": 16.139144897460938,
|
| 2473 |
+
"learning_rate": 9.829612220916568e-06,
|
| 2474 |
+
"loss": 2.684,
|
| 2475 |
+
"step": 242
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.18851823118696664,
|
| 2479 |
+
"grad_norm": 17.887540817260742,
|
| 2480 |
+
"learning_rate": 9.870740305522913e-06,
|
| 2481 |
+
"loss": 2.234,
|
| 2482 |
+
"step": 243
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.18929402637703646,
|
| 2486 |
+
"grad_norm": 15.649737358093262,
|
| 2487 |
+
"learning_rate": 9.911868390129259e-06,
|
| 2488 |
+
"loss": 3.0141,
|
| 2489 |
+
"step": 244
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.19006982156710628,
|
| 2493 |
+
"grad_norm": 6.042151927947998,
|
| 2494 |
+
"learning_rate": 9.952996474735606e-06,
|
| 2495 |
+
"loss": 0.271,
|
| 2496 |
+
"step": 245
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.1908456167571761,
|
| 2500 |
+
"grad_norm": 16.89675521850586,
|
| 2501 |
+
"learning_rate": 9.99412455934195e-06,
|
| 2502 |
+
"loss": 1.878,
|
| 2503 |
+
"step": 246
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.19162141194724594,
|
| 2507 |
+
"grad_norm": 21.4814395904541,
|
| 2508 |
+
"learning_rate": 1.0035252643948295e-05,
|
| 2509 |
+
"loss": 2.4108,
|
| 2510 |
+
"step": 247
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.19239720713731576,
|
| 2514 |
+
"grad_norm": 24.15477752685547,
|
| 2515 |
+
"learning_rate": 1.007638072855464e-05,
|
| 2516 |
+
"loss": 3.6327,
|
| 2517 |
+
"step": 248
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.19317300232738557,
|
| 2521 |
+
"grad_norm": 19.688291549682617,
|
| 2522 |
+
"learning_rate": 1.0117508813160987e-05,
|
| 2523 |
+
"loss": 2.9108,
|
| 2524 |
+
"step": 249
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.1939487975174554,
|
| 2528 |
+
"grad_norm": 30.842660903930664,
|
| 2529 |
+
"learning_rate": 1.0158636897767332e-05,
|
| 2530 |
+
"loss": 4.425,
|
| 2531 |
+
"step": 250
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.1947245927075252,
|
| 2535 |
+
"grad_norm": 17.901451110839844,
|
| 2536 |
+
"learning_rate": 1.0199764982373677e-05,
|
| 2537 |
+
"loss": 2.6128,
|
| 2538 |
+
"step": 251
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.19550038789759502,
|
| 2542 |
+
"grad_norm": 19.920259475708008,
|
| 2543 |
+
"learning_rate": 1.0240893066980021e-05,
|
| 2544 |
+
"loss": 3.1872,
|
| 2545 |
+
"step": 252
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.19627618308766487,
|
| 2549 |
+
"grad_norm": 17.547534942626953,
|
| 2550 |
+
"learning_rate": 1.0282021151586368e-05,
|
| 2551 |
+
"loss": 1.9839,
|
| 2552 |
+
"step": 253
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.19705197827773469,
|
| 2556 |
+
"grad_norm": 30.619075775146484,
|
| 2557 |
+
"learning_rate": 1.0323149236192713e-05,
|
| 2558 |
+
"loss": 5.0317,
|
| 2559 |
+
"step": 254
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.1978277734678045,
|
| 2563 |
+
"grad_norm": 16.690624237060547,
|
| 2564 |
+
"learning_rate": 1.0364277320799059e-05,
|
| 2565 |
+
"loss": 1.8456,
|
| 2566 |
+
"step": 255
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.19860356865787432,
|
| 2570 |
+
"grad_norm": 17.084972381591797,
|
| 2571 |
+
"learning_rate": 1.0405405405405406e-05,
|
| 2572 |
+
"loss": 2.7799,
|
| 2573 |
+
"step": 256
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.19937936384794414,
|
| 2577 |
+
"grad_norm": 16.932178497314453,
|
| 2578 |
+
"learning_rate": 1.0446533490011751e-05,
|
| 2579 |
+
"loss": 2.4114,
|
| 2580 |
+
"step": 257
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.20015515903801395,
|
| 2584 |
+
"grad_norm": 14.226729393005371,
|
| 2585 |
+
"learning_rate": 1.0487661574618094e-05,
|
| 2586 |
+
"loss": 1.2619,
|
| 2587 |
+
"step": 258
|
| 2588 |
+
}
|
| 2589 |
+
],
|
| 2590 |
+
"logging_steps": 1,
|
| 2591 |
+
"max_steps": 2578,
|
| 2592 |
+
"num_input_tokens_seen": 0,
|
| 2593 |
+
"num_train_epochs": 2,
|
| 2594 |
+
"save_steps": 258,
|
| 2595 |
+
"stateful_callbacks": {
|
| 2596 |
+
"TrainerControl": {
|
| 2597 |
+
"args": {
|
| 2598 |
+
"should_epoch_stop": false,
|
| 2599 |
+
"should_evaluate": false,
|
| 2600 |
+
"should_log": false,
|
| 2601 |
+
"should_save": true,
|
| 2602 |
+
"should_training_stop": false
|
| 2603 |
+
},
|
| 2604 |
+
"attributes": {}
|
| 2605 |
+
}
|
| 2606 |
+
},
|
| 2607 |
+
"total_flos": 0.0,
|
| 2608 |
+
"train_batch_size": 96,
|
| 2609 |
+
"trial_name": null,
|
| 2610 |
+
"trial_params": null
|
| 2611 |
+
}
|
checkpoint-258/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd87e469d7441c9256c28d82e79e8e3da32741d45a6bc310f3f601255edbdf8d
|
| 3 |
+
size 5752
|