Commit
•
38dc20a
1
Parent(s):
4b86bd4
Upload 11 files
Browse files- config.json +45 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +57 -0
- trainer_state.json +841 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "NbAiLab/nb-bert-large",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 1024,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0",
|
| 14 |
+
"1": "LABEL_1",
|
| 15 |
+
"2": "LABEL_2",
|
| 16 |
+
"3": "LABEL_3",
|
| 17 |
+
"4": "LABEL_4",
|
| 18 |
+
"5": "LABEL_5",
|
| 19 |
+
"6": "LABEL_6"
|
| 20 |
+
},
|
| 21 |
+
"initializer_range": 0.02,
|
| 22 |
+
"intermediate_size": 4096,
|
| 23 |
+
"label2id": {
|
| 24 |
+
"LABEL_0": 0,
|
| 25 |
+
"LABEL_1": 1,
|
| 26 |
+
"LABEL_2": 2,
|
| 27 |
+
"LABEL_3": 3,
|
| 28 |
+
"LABEL_4": 4,
|
| 29 |
+
"LABEL_5": 5,
|
| 30 |
+
"LABEL_6": 6
|
| 31 |
+
},
|
| 32 |
+
"layer_norm_eps": 1e-12,
|
| 33 |
+
"max_position_embeddings": 512,
|
| 34 |
+
"model_type": "bert",
|
| 35 |
+
"num_attention_heads": 16,
|
| 36 |
+
"num_hidden_layers": 24,
|
| 37 |
+
"pad_token_id": 0,
|
| 38 |
+
"position_embedding_type": "absolute",
|
| 39 |
+
"problem_type": "single_label_classification",
|
| 40 |
+
"torch_dtype": "float32",
|
| 41 |
+
"transformers_version": "4.40.0",
|
| 42 |
+
"type_vocab_size": 2,
|
| 43 |
+
"use_cache": true,
|
| 44 |
+
"vocab_size": 50000
|
| 45 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2364077fd4ca53d4a689c7f010523950e4f8b184caebf77203dd5991fb2ebdce
|
| 3 |
+
size 1420425204
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:084f35c82b56101e12303ac699d49a30789ad039323b9f6c5abf4f06f4f2f8e4
|
| 3 |
+
size 2841074157
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5f78c04db25254fcb28b4b8f75b8844280da67041351e4b57fa36d0916fb95a3
|
| 3 |
+
size 13990
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3fae0b86254a3f75f212ad9defa917a452476f23140d38f4aae369e916295c09
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"501": {
|
| 4 |
+
"content": "[CLS]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"502": {
|
| 12 |
+
"content": "[MASK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"503": {
|
| 20 |
+
"content": "[PAD]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"504": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"505": {
|
| 36 |
+
"content": "[UNK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": false,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,841 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.390835579514825,
|
| 5 |
+
"eval_steps": 30,
|
| 6 |
+
"global_step": 2000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.08086253369272237,
|
| 13 |
+
"eval_accuracy": 0.34545454545454546,
|
| 14 |
+
"eval_f1": 0.08213880790411401,
|
| 15 |
+
"eval_loss": 1.7152690887451172,
|
| 16 |
+
"eval_precision": 0.12064459930313588,
|
| 17 |
+
"eval_recall": 0.14761904761904762,
|
| 18 |
+
"eval_runtime": 13.6508,
|
| 19 |
+
"eval_samples_per_second": 24.174,
|
| 20 |
+
"eval_steps_per_second": 3.077,
|
| 21 |
+
"step": 30
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.16172506738544473,
|
| 25 |
+
"eval_accuracy": 0.4393939393939394,
|
| 26 |
+
"eval_f1": 0.21690145794050794,
|
| 27 |
+
"eval_loss": 1.5937703847885132,
|
| 28 |
+
"eval_precision": 0.32583090861779385,
|
| 29 |
+
"eval_recall": 0.25099255283137484,
|
| 30 |
+
"eval_runtime": 13.1278,
|
| 31 |
+
"eval_samples_per_second": 25.137,
|
| 32 |
+
"eval_steps_per_second": 3.199,
|
| 33 |
+
"step": 60
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.24258760107816713,
|
| 37 |
+
"eval_accuracy": 0.38181818181818183,
|
| 38 |
+
"eval_f1": 0.13698374256214413,
|
| 39 |
+
"eval_loss": 1.5304551124572754,
|
| 40 |
+
"eval_precision": 0.23018433179723502,
|
| 41 |
+
"eval_recall": 0.17921977437427894,
|
| 42 |
+
"eval_runtime": 13.0828,
|
| 43 |
+
"eval_samples_per_second": 25.224,
|
| 44 |
+
"eval_steps_per_second": 3.21,
|
| 45 |
+
"step": 90
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.32345013477088946,
|
| 49 |
+
"eval_accuracy": 0.37575757575757573,
|
| 50 |
+
"eval_f1": 0.13802749957562382,
|
| 51 |
+
"eval_loss": 1.6096415519714355,
|
| 52 |
+
"eval_precision": 0.1890046674322801,
|
| 53 |
+
"eval_recall": 0.18045360027080612,
|
| 54 |
+
"eval_runtime": 13.3288,
|
| 55 |
+
"eval_samples_per_second": 24.758,
|
| 56 |
+
"eval_steps_per_second": 3.151,
|
| 57 |
+
"step": 120
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.40431266846361186,
|
| 61 |
+
"eval_accuracy": 0.45151515151515154,
|
| 62 |
+
"eval_f1": 0.23452986300450943,
|
| 63 |
+
"eval_loss": 1.4274822473526,
|
| 64 |
+
"eval_precision": 0.38950040524911433,
|
| 65 |
+
"eval_recall": 0.27641592488584005,
|
| 66 |
+
"eval_runtime": 13.4187,
|
| 67 |
+
"eval_samples_per_second": 24.593,
|
| 68 |
+
"eval_steps_per_second": 3.13,
|
| 69 |
+
"step": 150
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.48517520215633425,
|
| 73 |
+
"eval_accuracy": 0.5424242424242425,
|
| 74 |
+
"eval_f1": 0.327924405270561,
|
| 75 |
+
"eval_loss": 1.2386934757232666,
|
| 76 |
+
"eval_precision": 0.35548957287833166,
|
| 77 |
+
"eval_recall": 0.35121955898274004,
|
| 78 |
+
"eval_runtime": 14.6075,
|
| 79 |
+
"eval_samples_per_second": 22.591,
|
| 80 |
+
"eval_steps_per_second": 2.875,
|
| 81 |
+
"step": 180
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.5660377358490566,
|
| 85 |
+
"eval_accuracy": 0.41818181818181815,
|
| 86 |
+
"eval_f1": 0.2642022704662224,
|
| 87 |
+
"eval_loss": 1.7232695817947388,
|
| 88 |
+
"eval_precision": 0.440507406905074,
|
| 89 |
+
"eval_recall": 0.29508974215683664,
|
| 90 |
+
"eval_runtime": 14.2568,
|
| 91 |
+
"eval_samples_per_second": 23.147,
|
| 92 |
+
"eval_steps_per_second": 2.946,
|
| 93 |
+
"step": 210
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.6469002695417789,
|
| 97 |
+
"eval_accuracy": 0.5575757575757576,
|
| 98 |
+
"eval_f1": 0.36732006348386254,
|
| 99 |
+
"eval_loss": 1.1753942966461182,
|
| 100 |
+
"eval_precision": 0.43188775510204075,
|
| 101 |
+
"eval_recall": 0.36647921681366585,
|
| 102 |
+
"eval_runtime": 13.3212,
|
| 103 |
+
"eval_samples_per_second": 24.772,
|
| 104 |
+
"eval_steps_per_second": 3.153,
|
| 105 |
+
"step": 240
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.7277628032345014,
|
| 109 |
+
"eval_accuracy": 0.5393939393939394,
|
| 110 |
+
"eval_f1": 0.32937378728076405,
|
| 111 |
+
"eval_loss": 1.2770304679870605,
|
| 112 |
+
"eval_precision": 0.39731428747888,
|
| 113 |
+
"eval_recall": 0.32973768555006505,
|
| 114 |
+
"eval_runtime": 13.2685,
|
| 115 |
+
"eval_samples_per_second": 24.871,
|
| 116 |
+
"eval_steps_per_second": 3.165,
|
| 117 |
+
"step": 270
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.8086253369272237,
|
| 121 |
+
"eval_accuracy": 0.5606060606060606,
|
| 122 |
+
"eval_f1": 0.3439473243062972,
|
| 123 |
+
"eval_loss": 1.2049232721328735,
|
| 124 |
+
"eval_precision": 0.4835807771246264,
|
| 125 |
+
"eval_recall": 0.33995508230465793,
|
| 126 |
+
"eval_runtime": 15.2882,
|
| 127 |
+
"eval_samples_per_second": 21.585,
|
| 128 |
+
"eval_steps_per_second": 2.747,
|
| 129 |
+
"step": 300
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.889487870619946,
|
| 133 |
+
"eval_accuracy": 0.5666666666666667,
|
| 134 |
+
"eval_f1": 0.394947488354592,
|
| 135 |
+
"eval_loss": 1.108344554901123,
|
| 136 |
+
"eval_precision": 0.5166598073288214,
|
| 137 |
+
"eval_recall": 0.3978135018724203,
|
| 138 |
+
"eval_runtime": 14.5889,
|
| 139 |
+
"eval_samples_per_second": 22.62,
|
| 140 |
+
"eval_steps_per_second": 2.879,
|
| 141 |
+
"step": 330
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.9703504043126685,
|
| 145 |
+
"eval_accuracy": 0.6121212121212121,
|
| 146 |
+
"eval_f1": 0.428135738050604,
|
| 147 |
+
"eval_loss": 1.141391634941101,
|
| 148 |
+
"eval_precision": 0.6312669683257919,
|
| 149 |
+
"eval_recall": 0.40262976955895236,
|
| 150 |
+
"eval_runtime": 12.9679,
|
| 151 |
+
"eval_samples_per_second": 25.447,
|
| 152 |
+
"eval_steps_per_second": 3.239,
|
| 153 |
+
"step": 360
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.0512129380053907,
|
| 157 |
+
"eval_accuracy": 0.603030303030303,
|
| 158 |
+
"eval_f1": 0.4038251382972519,
|
| 159 |
+
"eval_loss": 1.116855263710022,
|
| 160 |
+
"eval_precision": 0.4923091038697975,
|
| 161 |
+
"eval_recall": 0.41548484912495504,
|
| 162 |
+
"eval_runtime": 13.6567,
|
| 163 |
+
"eval_samples_per_second": 24.164,
|
| 164 |
+
"eval_steps_per_second": 3.075,
|
| 165 |
+
"step": 390
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.1320754716981132,
|
| 169 |
+
"eval_accuracy": 0.5757575757575758,
|
| 170 |
+
"eval_f1": 0.41737552463332295,
|
| 171 |
+
"eval_loss": 1.13466477394104,
|
| 172 |
+
"eval_precision": 0.6076030412164866,
|
| 173 |
+
"eval_recall": 0.4208784549569585,
|
| 174 |
+
"eval_runtime": 12.847,
|
| 175 |
+
"eval_samples_per_second": 25.687,
|
| 176 |
+
"eval_steps_per_second": 3.269,
|
| 177 |
+
"step": 420
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.2129380053908356,
|
| 181 |
+
"eval_accuracy": 0.603030303030303,
|
| 182 |
+
"eval_f1": 0.43365687052323004,
|
| 183 |
+
"eval_loss": 1.1254369020462036,
|
| 184 |
+
"eval_precision": 0.5694631128061721,
|
| 185 |
+
"eval_recall": 0.41092089239472196,
|
| 186 |
+
"eval_runtime": 12.8187,
|
| 187 |
+
"eval_samples_per_second": 25.744,
|
| 188 |
+
"eval_steps_per_second": 3.276,
|
| 189 |
+
"step": 450
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.2938005390835579,
|
| 193 |
+
"eval_accuracy": 0.5696969696969697,
|
| 194 |
+
"eval_f1": 0.3965291918438572,
|
| 195 |
+
"eval_loss": 1.2160786390304565,
|
| 196 |
+
"eval_precision": 0.4828996768066082,
|
| 197 |
+
"eval_recall": 0.38449281792865353,
|
| 198 |
+
"eval_runtime": 14.7332,
|
| 199 |
+
"eval_samples_per_second": 22.398,
|
| 200 |
+
"eval_steps_per_second": 2.851,
|
| 201 |
+
"step": 480
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.3477088948787062,
|
| 205 |
+
"grad_norm": 18.155202865600586,
|
| 206 |
+
"learning_rate": 3.876909254267745e-05,
|
| 207 |
+
"loss": 1.3924,
|
| 208 |
+
"step": 500
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 1.3746630727762803,
|
| 212 |
+
"eval_accuracy": 0.5424242424242425,
|
| 213 |
+
"eval_f1": 0.4319260406224048,
|
| 214 |
+
"eval_loss": 1.1475260257720947,
|
| 215 |
+
"eval_precision": 0.45103525254505755,
|
| 216 |
+
"eval_recall": 0.4340482151274285,
|
| 217 |
+
"eval_runtime": 32.076,
|
| 218 |
+
"eval_samples_per_second": 10.288,
|
| 219 |
+
"eval_steps_per_second": 1.309,
|
| 220 |
+
"step": 510
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 1.4555256064690028,
|
| 224 |
+
"eval_accuracy": 0.5878787878787879,
|
| 225 |
+
"eval_f1": 0.3939948328709428,
|
| 226 |
+
"eval_loss": 1.1194807291030884,
|
| 227 |
+
"eval_precision": 0.608180480837931,
|
| 228 |
+
"eval_recall": 0.392683155449795,
|
| 229 |
+
"eval_runtime": 110.8563,
|
| 230 |
+
"eval_samples_per_second": 2.977,
|
| 231 |
+
"eval_steps_per_second": 0.379,
|
| 232 |
+
"step": 540
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 1.536388140161725,
|
| 236 |
+
"eval_accuracy": 0.5818181818181818,
|
| 237 |
+
"eval_f1": 0.4490235853265899,
|
| 238 |
+
"eval_loss": 1.1963447332382202,
|
| 239 |
+
"eval_precision": 0.5116256986573146,
|
| 240 |
+
"eval_recall": 0.4377877348363712,
|
| 241 |
+
"eval_runtime": 64.5493,
|
| 242 |
+
"eval_samples_per_second": 5.112,
|
| 243 |
+
"eval_steps_per_second": 0.651,
|
| 244 |
+
"step": 570
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 1.6172506738544474,
|
| 248 |
+
"eval_accuracy": 0.6060606060606061,
|
| 249 |
+
"eval_f1": 0.44688978798681084,
|
| 250 |
+
"eval_loss": 1.1052302122116089,
|
| 251 |
+
"eval_precision": 0.5367127769479213,
|
| 252 |
+
"eval_recall": 0.4296217856765177,
|
| 253 |
+
"eval_runtime": 71.5496,
|
| 254 |
+
"eval_samples_per_second": 4.612,
|
| 255 |
+
"eval_steps_per_second": 0.587,
|
| 256 |
+
"step": 600
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.6981132075471699,
|
| 260 |
+
"eval_accuracy": 0.5848484848484848,
|
| 261 |
+
"eval_f1": 0.46957050415643353,
|
| 262 |
+
"eval_loss": 1.108028769493103,
|
| 263 |
+
"eval_precision": 0.572241929381981,
|
| 264 |
+
"eval_recall": 0.4537304361700488,
|
| 265 |
+
"eval_runtime": 12.7244,
|
| 266 |
+
"eval_samples_per_second": 25.935,
|
| 267 |
+
"eval_steps_per_second": 3.301,
|
| 268 |
+
"step": 630
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.778975741239892,
|
| 272 |
+
"eval_accuracy": 0.6212121212121212,
|
| 273 |
+
"eval_f1": 0.47748925036798145,
|
| 274 |
+
"eval_loss": 1.033095359802246,
|
| 275 |
+
"eval_precision": 0.5650004027150538,
|
| 276 |
+
"eval_recall": 0.4613483358977623,
|
| 277 |
+
"eval_runtime": 13.0381,
|
| 278 |
+
"eval_samples_per_second": 25.31,
|
| 279 |
+
"eval_steps_per_second": 3.221,
|
| 280 |
+
"step": 660
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 1.8598382749326146,
|
| 284 |
+
"eval_accuracy": 0.6242424242424243,
|
| 285 |
+
"eval_f1": 0.4612218625033871,
|
| 286 |
+
"eval_loss": 1.0287344455718994,
|
| 287 |
+
"eval_precision": 0.5787994888332474,
|
| 288 |
+
"eval_recall": 0.4401355279945231,
|
| 289 |
+
"eval_runtime": 14.0033,
|
| 290 |
+
"eval_samples_per_second": 23.566,
|
| 291 |
+
"eval_steps_per_second": 2.999,
|
| 292 |
+
"step": 690
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 1.940700808625337,
|
| 296 |
+
"eval_accuracy": 0.6,
|
| 297 |
+
"eval_f1": 0.43355058251543566,
|
| 298 |
+
"eval_loss": 1.0855339765548706,
|
| 299 |
+
"eval_precision": 0.5482738467345629,
|
| 300 |
+
"eval_recall": 0.407901742703353,
|
| 301 |
+
"eval_runtime": 14.8086,
|
| 302 |
+
"eval_samples_per_second": 22.284,
|
| 303 |
+
"eval_steps_per_second": 2.836,
|
| 304 |
+
"step": 720
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 2.0215633423180592,
|
| 308 |
+
"eval_accuracy": 0.6060606060606061,
|
| 309 |
+
"eval_f1": 0.4278903186584704,
|
| 310 |
+
"eval_loss": 1.0721008777618408,
|
| 311 |
+
"eval_precision": 0.6261473906210748,
|
| 312 |
+
"eval_recall": 0.400766872593428,
|
| 313 |
+
"eval_runtime": 13.33,
|
| 314 |
+
"eval_samples_per_second": 24.756,
|
| 315 |
+
"eval_steps_per_second": 3.151,
|
| 316 |
+
"step": 750
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 2.1024258760107815,
|
| 320 |
+
"eval_accuracy": 0.5818181818181818,
|
| 321 |
+
"eval_f1": 0.46279487414189074,
|
| 322 |
+
"eval_loss": 1.0543911457061768,
|
| 323 |
+
"eval_precision": 0.4927064392103887,
|
| 324 |
+
"eval_recall": 0.4551381887414431,
|
| 325 |
+
"eval_runtime": 13.3061,
|
| 326 |
+
"eval_samples_per_second": 24.801,
|
| 327 |
+
"eval_steps_per_second": 3.156,
|
| 328 |
+
"step": 780
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 2.183288409703504,
|
| 332 |
+
"eval_accuracy": 0.6,
|
| 333 |
+
"eval_f1": 0.44787598241263715,
|
| 334 |
+
"eval_loss": 1.1306463479995728,
|
| 335 |
+
"eval_precision": 0.5029518040278441,
|
| 336 |
+
"eval_recall": 0.42696476597535643,
|
| 337 |
+
"eval_runtime": 14.7754,
|
| 338 |
+
"eval_samples_per_second": 22.334,
|
| 339 |
+
"eval_steps_per_second": 2.843,
|
| 340 |
+
"step": 810
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.2641509433962264,
|
| 344 |
+
"eval_accuracy": 0.6151515151515151,
|
| 345 |
+
"eval_f1": 0.47831287112215076,
|
| 346 |
+
"eval_loss": 1.1025209426879883,
|
| 347 |
+
"eval_precision": 0.5683183097103888,
|
| 348 |
+
"eval_recall": 0.45416730293429747,
|
| 349 |
+
"eval_runtime": 14.712,
|
| 350 |
+
"eval_samples_per_second": 22.431,
|
| 351 |
+
"eval_steps_per_second": 2.855,
|
| 352 |
+
"step": 840
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.3450134770889486,
|
| 356 |
+
"eval_accuracy": 0.5848484848484848,
|
| 357 |
+
"eval_f1": 0.4612687851809403,
|
| 358 |
+
"eval_loss": 1.1565966606140137,
|
| 359 |
+
"eval_precision": 0.5073562417803803,
|
| 360 |
+
"eval_recall": 0.4593010520365005,
|
| 361 |
+
"eval_runtime": 13.517,
|
| 362 |
+
"eval_samples_per_second": 24.414,
|
| 363 |
+
"eval_steps_per_second": 3.107,
|
| 364 |
+
"step": 870
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 2.4258760107816713,
|
| 368 |
+
"eval_accuracy": 0.5636363636363636,
|
| 369 |
+
"eval_f1": 0.45972084178362305,
|
| 370 |
+
"eval_loss": 1.2057921886444092,
|
| 371 |
+
"eval_precision": 0.5695289983467323,
|
| 372 |
+
"eval_recall": 0.461413466501444,
|
| 373 |
+
"eval_runtime": 13.0432,
|
| 374 |
+
"eval_samples_per_second": 25.301,
|
| 375 |
+
"eval_steps_per_second": 3.22,
|
| 376 |
+
"step": 900
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 2.5067385444743935,
|
| 380 |
+
"eval_accuracy": 0.6090909090909091,
|
| 381 |
+
"eval_f1": 0.4628685133833875,
|
| 382 |
+
"eval_loss": 1.1381313800811768,
|
| 383 |
+
"eval_precision": 0.5632985257985258,
|
| 384 |
+
"eval_recall": 0.4583315282026837,
|
| 385 |
+
"eval_runtime": 14.1243,
|
| 386 |
+
"eval_samples_per_second": 23.364,
|
| 387 |
+
"eval_steps_per_second": 2.974,
|
| 388 |
+
"step": 930
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 2.5876010781671157,
|
| 392 |
+
"eval_accuracy": 0.6242424242424243,
|
| 393 |
+
"eval_f1": 0.4615624143566702,
|
| 394 |
+
"eval_loss": 1.0724554061889648,
|
| 395 |
+
"eval_precision": 0.5741660836493663,
|
| 396 |
+
"eval_recall": 0.4438527652883962,
|
| 397 |
+
"eval_runtime": 14.1621,
|
| 398 |
+
"eval_samples_per_second": 23.302,
|
| 399 |
+
"eval_steps_per_second": 2.966,
|
| 400 |
+
"step": 960
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 2.6684636118598384,
|
| 404 |
+
"eval_accuracy": 0.6060606060606061,
|
| 405 |
+
"eval_f1": 0.47087755462465813,
|
| 406 |
+
"eval_loss": 1.135140299797058,
|
| 407 |
+
"eval_precision": 0.5209651599354117,
|
| 408 |
+
"eval_recall": 0.45427047745523014,
|
| 409 |
+
"eval_runtime": 14.0682,
|
| 410 |
+
"eval_samples_per_second": 23.457,
|
| 411 |
+
"eval_steps_per_second": 2.985,
|
| 412 |
+
"step": 990
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 2.6954177897574123,
|
| 416 |
+
"grad_norm": 7.412640571594238,
|
| 417 |
+
"learning_rate": 2.75381850853549e-05,
|
| 418 |
+
"loss": 0.9741,
|
| 419 |
+
"step": 1000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.7493261455525606,
|
| 423 |
+
"eval_accuracy": 0.6242424242424243,
|
| 424 |
+
"eval_f1": 0.4549752385442883,
|
| 425 |
+
"eval_loss": 0.9898973703384399,
|
| 426 |
+
"eval_precision": 0.5431213833871716,
|
| 427 |
+
"eval_recall": 0.44127615017259786,
|
| 428 |
+
"eval_runtime": 113.2331,
|
| 429 |
+
"eval_samples_per_second": 2.914,
|
| 430 |
+
"eval_steps_per_second": 0.371,
|
| 431 |
+
"step": 1020
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 2.830188679245283,
|
| 435 |
+
"eval_accuracy": 0.6212121212121212,
|
| 436 |
+
"eval_f1": 0.477029569791836,
|
| 437 |
+
"eval_loss": 1.1227794885635376,
|
| 438 |
+
"eval_precision": 0.5355122892068706,
|
| 439 |
+
"eval_recall": 0.4702582406289971,
|
| 440 |
+
"eval_runtime": 22.5487,
|
| 441 |
+
"eval_samples_per_second": 14.635,
|
| 442 |
+
"eval_steps_per_second": 1.863,
|
| 443 |
+
"step": 1050
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 2.9110512129380055,
|
| 447 |
+
"eval_accuracy": 0.6272727272727273,
|
| 448 |
+
"eval_f1": 0.4661588293248819,
|
| 449 |
+
"eval_loss": 1.0796787738800049,
|
| 450 |
+
"eval_precision": 0.5698680297271846,
|
| 451 |
+
"eval_recall": 0.4402755664195842,
|
| 452 |
+
"eval_runtime": 142.0116,
|
| 453 |
+
"eval_samples_per_second": 2.324,
|
| 454 |
+
"eval_steps_per_second": 0.296,
|
| 455 |
+
"step": 1080
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 2.9919137466307277,
|
| 459 |
+
"eval_accuracy": 0.6424242424242425,
|
| 460 |
+
"eval_f1": 0.49676860579479715,
|
| 461 |
+
"eval_loss": 1.0576601028442383,
|
| 462 |
+
"eval_precision": 0.5889560125125319,
|
| 463 |
+
"eval_recall": 0.48000300606800994,
|
| 464 |
+
"eval_runtime": 25.0835,
|
| 465 |
+
"eval_samples_per_second": 13.156,
|
| 466 |
+
"eval_steps_per_second": 1.674,
|
| 467 |
+
"step": 1110
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 3.07277628032345,
|
| 471 |
+
"eval_accuracy": 0.6333333333333333,
|
| 472 |
+
"eval_f1": 0.5076311071541079,
|
| 473 |
+
"eval_loss": 1.1107242107391357,
|
| 474 |
+
"eval_precision": 0.5412034459392385,
|
| 475 |
+
"eval_recall": 0.4934487547962528,
|
| 476 |
+
"eval_runtime": 13.978,
|
| 477 |
+
"eval_samples_per_second": 23.608,
|
| 478 |
+
"eval_steps_per_second": 3.005,
|
| 479 |
+
"step": 1140
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 3.1536388140161726,
|
| 483 |
+
"eval_accuracy": 0.6515151515151515,
|
| 484 |
+
"eval_f1": 0.5186595057636214,
|
| 485 |
+
"eval_loss": 1.1433833837509155,
|
| 486 |
+
"eval_precision": 0.5322955893403752,
|
| 487 |
+
"eval_recall": 0.5121777726730862,
|
| 488 |
+
"eval_runtime": 14.8138,
|
| 489 |
+
"eval_samples_per_second": 22.277,
|
| 490 |
+
"eval_steps_per_second": 2.835,
|
| 491 |
+
"step": 1170
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 3.234501347708895,
|
| 495 |
+
"eval_accuracy": 0.6303030303030303,
|
| 496 |
+
"eval_f1": 0.4818021563155342,
|
| 497 |
+
"eval_loss": 1.2040390968322754,
|
| 498 |
+
"eval_precision": 0.5245503034019318,
|
| 499 |
+
"eval_recall": 0.4790353988658817,
|
| 500 |
+
"eval_runtime": 14.5347,
|
| 501 |
+
"eval_samples_per_second": 22.704,
|
| 502 |
+
"eval_steps_per_second": 2.89,
|
| 503 |
+
"step": 1200
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 3.315363881401617,
|
| 507 |
+
"eval_accuracy": 0.6303030303030303,
|
| 508 |
+
"eval_f1": 0.5005894502956657,
|
| 509 |
+
"eval_loss": 1.106711983680725,
|
| 510 |
+
"eval_precision": 0.5347839031818471,
|
| 511 |
+
"eval_recall": 0.49606350872380334,
|
| 512 |
+
"eval_runtime": 14.5181,
|
| 513 |
+
"eval_samples_per_second": 22.73,
|
| 514 |
+
"eval_steps_per_second": 2.893,
|
| 515 |
+
"step": 1230
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.3962264150943398,
|
| 519 |
+
"eval_accuracy": 0.6363636363636364,
|
| 520 |
+
"eval_f1": 0.49230222929692463,
|
| 521 |
+
"eval_loss": 1.100297451019287,
|
| 522 |
+
"eval_precision": 0.5613737992397406,
|
| 523 |
+
"eval_recall": 0.48222645824949406,
|
| 524 |
+
"eval_runtime": 14.6402,
|
| 525 |
+
"eval_samples_per_second": 22.541,
|
| 526 |
+
"eval_steps_per_second": 2.869,
|
| 527 |
+
"step": 1260
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.477088948787062,
|
| 531 |
+
"eval_accuracy": 0.603030303030303,
|
| 532 |
+
"eval_f1": 0.48380416653626124,
|
| 533 |
+
"eval_loss": 1.1756219863891602,
|
| 534 |
+
"eval_precision": 0.5033237551313431,
|
| 535 |
+
"eval_recall": 0.4908017943866942,
|
| 536 |
+
"eval_runtime": 13.454,
|
| 537 |
+
"eval_samples_per_second": 24.528,
|
| 538 |
+
"eval_steps_per_second": 3.122,
|
| 539 |
+
"step": 1290
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.557951482479784,
|
| 543 |
+
"eval_accuracy": 0.6060606060606061,
|
| 544 |
+
"eval_f1": 0.4764702057860691,
|
| 545 |
+
"eval_loss": 1.1003729104995728,
|
| 546 |
+
"eval_precision": 0.5288839860268432,
|
| 547 |
+
"eval_recall": 0.4590755558938861,
|
| 548 |
+
"eval_runtime": 12.5627,
|
| 549 |
+
"eval_samples_per_second": 26.268,
|
| 550 |
+
"eval_steps_per_second": 3.343,
|
| 551 |
+
"step": 1320
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.638814016172507,
|
| 555 |
+
"eval_accuracy": 0.593939393939394,
|
| 556 |
+
"eval_f1": 0.487375493221817,
|
| 557 |
+
"eval_loss": 1.086908221244812,
|
| 558 |
+
"eval_precision": 0.48992444445192695,
|
| 559 |
+
"eval_recall": 0.5073454405336776,
|
| 560 |
+
"eval_runtime": 12.6831,
|
| 561 |
+
"eval_samples_per_second": 26.019,
|
| 562 |
+
"eval_steps_per_second": 3.311,
|
| 563 |
+
"step": 1350
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 3.719676549865229,
|
| 567 |
+
"eval_accuracy": 0.5818181818181818,
|
| 568 |
+
"eval_f1": 0.4671306646522057,
|
| 569 |
+
"eval_loss": 1.1851921081542969,
|
| 570 |
+
"eval_precision": 0.5193033472658716,
|
| 571 |
+
"eval_recall": 0.46147135451955557,
|
| 572 |
+
"eval_runtime": 13.4674,
|
| 573 |
+
"eval_samples_per_second": 24.504,
|
| 574 |
+
"eval_steps_per_second": 3.119,
|
| 575 |
+
"step": 1380
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.8005390835579513,
|
| 579 |
+
"eval_accuracy": 0.6272727272727273,
|
| 580 |
+
"eval_f1": 0.4738897906354585,
|
| 581 |
+
"eval_loss": 1.1233046054840088,
|
| 582 |
+
"eval_precision": 0.5308286308286309,
|
| 583 |
+
"eval_recall": 0.45352857445191025,
|
| 584 |
+
"eval_runtime": 14.4652,
|
| 585 |
+
"eval_samples_per_second": 22.813,
|
| 586 |
+
"eval_steps_per_second": 2.904,
|
| 587 |
+
"step": 1410
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.881401617250674,
|
| 591 |
+
"eval_accuracy": 0.6333333333333333,
|
| 592 |
+
"eval_f1": 0.49797095491412685,
|
| 593 |
+
"eval_loss": 1.0570012331008911,
|
| 594 |
+
"eval_precision": 0.5228267459278104,
|
| 595 |
+
"eval_recall": 0.48949827535886625,
|
| 596 |
+
"eval_runtime": 13.7881,
|
| 597 |
+
"eval_samples_per_second": 23.934,
|
| 598 |
+
"eval_steps_per_second": 3.046,
|
| 599 |
+
"step": 1440
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.9622641509433962,
|
| 603 |
+
"eval_accuracy": 0.5696969696969697,
|
| 604 |
+
"eval_f1": 0.436851400958813,
|
| 605 |
+
"eval_loss": 1.196516513824463,
|
| 606 |
+
"eval_precision": 0.4607919084994466,
|
| 607 |
+
"eval_recall": 0.44320607800419104,
|
| 608 |
+
"eval_runtime": 13.9502,
|
| 609 |
+
"eval_samples_per_second": 23.656,
|
| 610 |
+
"eval_steps_per_second": 3.011,
|
| 611 |
+
"step": 1470
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 4.0431266846361185,
|
| 615 |
+
"grad_norm": 4.396773815155029,
|
| 616 |
+
"learning_rate": 1.6307277628032348e-05,
|
| 617 |
+
"loss": 0.6945,
|
| 618 |
+
"step": 1500
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 4.0431266846361185,
|
| 622 |
+
"eval_accuracy": 0.6242424242424243,
|
| 623 |
+
"eval_f1": 0.48252518710343445,
|
| 624 |
+
"eval_loss": 1.0697778463363647,
|
| 625 |
+
"eval_precision": 0.529925883396837,
|
| 626 |
+
"eval_recall": 0.4736448771051681,
|
| 627 |
+
"eval_runtime": 14.005,
|
| 628 |
+
"eval_samples_per_second": 23.563,
|
| 629 |
+
"eval_steps_per_second": 2.999,
|
| 630 |
+
"step": 1500
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 4.123989218328841,
|
| 634 |
+
"eval_accuracy": 0.6242424242424243,
|
| 635 |
+
"eval_f1": 0.49720491179937104,
|
| 636 |
+
"eval_loss": 1.2521579265594482,
|
| 637 |
+
"eval_precision": 0.5386641378880501,
|
| 638 |
+
"eval_recall": 0.49817853705778303,
|
| 639 |
+
"eval_runtime": 115.3902,
|
| 640 |
+
"eval_samples_per_second": 2.86,
|
| 641 |
+
"eval_steps_per_second": 0.364,
|
| 642 |
+
"step": 1530
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 4.204851752021563,
|
| 646 |
+
"eval_accuracy": 0.6181818181818182,
|
| 647 |
+
"eval_f1": 0.47697868890738926,
|
| 648 |
+
"eval_loss": 1.4063215255737305,
|
| 649 |
+
"eval_precision": 0.4929803350172045,
|
| 650 |
+
"eval_recall": 0.4728290050243222,
|
| 651 |
+
"eval_runtime": 101.5928,
|
| 652 |
+
"eval_samples_per_second": 3.248,
|
| 653 |
+
"eval_steps_per_second": 0.413,
|
| 654 |
+
"step": 1560
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 4.285714285714286,
|
| 658 |
+
"eval_accuracy": 0.6212121212121212,
|
| 659 |
+
"eval_f1": 0.4993387370778744,
|
| 660 |
+
"eval_loss": 1.349918007850647,
|
| 661 |
+
"eval_precision": 0.49880570409982167,
|
| 662 |
+
"eval_recall": 0.5059772713055806,
|
| 663 |
+
"eval_runtime": 115.6893,
|
| 664 |
+
"eval_samples_per_second": 2.852,
|
| 665 |
+
"eval_steps_per_second": 0.363,
|
| 666 |
+
"step": 1590
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 4.366576819407008,
|
| 670 |
+
"eval_accuracy": 0.6,
|
| 671 |
+
"eval_f1": 0.46437393231058216,
|
| 672 |
+
"eval_loss": 1.340740442276001,
|
| 673 |
+
"eval_precision": 0.47078319079488345,
|
| 674 |
+
"eval_recall": 0.46173072420332834,
|
| 675 |
+
"eval_runtime": 13.0095,
|
| 676 |
+
"eval_samples_per_second": 25.366,
|
| 677 |
+
"eval_steps_per_second": 3.228,
|
| 678 |
+
"step": 1620
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 4.44743935309973,
|
| 682 |
+
"eval_accuracy": 0.6393939393939394,
|
| 683 |
+
"eval_f1": 0.4956975851278201,
|
| 684 |
+
"eval_loss": 1.4181333780288696,
|
| 685 |
+
"eval_precision": 0.5323631437686738,
|
| 686 |
+
"eval_recall": 0.48370483391830105,
|
| 687 |
+
"eval_runtime": 15.6022,
|
| 688 |
+
"eval_samples_per_second": 21.151,
|
| 689 |
+
"eval_steps_per_second": 2.692,
|
| 690 |
+
"step": 1650
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 4.528301886792453,
|
| 694 |
+
"eval_accuracy": 0.6212121212121212,
|
| 695 |
+
"eval_f1": 0.48744288293160476,
|
| 696 |
+
"eval_loss": 1.4648175239562988,
|
| 697 |
+
"eval_precision": 0.5288562135045842,
|
| 698 |
+
"eval_recall": 0.484239463132841,
|
| 699 |
+
"eval_runtime": 14.8896,
|
| 700 |
+
"eval_samples_per_second": 22.163,
|
| 701 |
+
"eval_steps_per_second": 2.821,
|
| 702 |
+
"step": 1680
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 4.609164420485175,
|
| 706 |
+
"eval_accuracy": 0.6363636363636364,
|
| 707 |
+
"eval_f1": 0.49874561310412674,
|
| 708 |
+
"eval_loss": 1.3795430660247803,
|
| 709 |
+
"eval_precision": 0.515957965289801,
|
| 710 |
+
"eval_recall": 0.49555231275739653,
|
| 711 |
+
"eval_runtime": 15.2135,
|
| 712 |
+
"eval_samples_per_second": 21.691,
|
| 713 |
+
"eval_steps_per_second": 2.761,
|
| 714 |
+
"step": 1710
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 4.690026954177897,
|
| 718 |
+
"eval_accuracy": 0.6545454545454545,
|
| 719 |
+
"eval_f1": 0.5072945393065025,
|
| 720 |
+
"eval_loss": 1.255421757698059,
|
| 721 |
+
"eval_precision": 0.527637085699966,
|
| 722 |
+
"eval_recall": 0.5017196060314337,
|
| 723 |
+
"eval_runtime": 13.3432,
|
| 724 |
+
"eval_samples_per_second": 24.732,
|
| 725 |
+
"eval_steps_per_second": 3.148,
|
| 726 |
+
"step": 1740
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 4.77088948787062,
|
| 730 |
+
"eval_accuracy": 0.6181818181818182,
|
| 731 |
+
"eval_f1": 0.47364322391678987,
|
| 732 |
+
"eval_loss": 1.2744801044464111,
|
| 733 |
+
"eval_precision": 0.5261603756546867,
|
| 734 |
+
"eval_recall": 0.4762025977768268,
|
| 735 |
+
"eval_runtime": 13.3766,
|
| 736 |
+
"eval_samples_per_second": 24.67,
|
| 737 |
+
"eval_steps_per_second": 3.14,
|
| 738 |
+
"step": 1770
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 4.8517520215633425,
|
| 742 |
+
"eval_accuracy": 0.6272727272727273,
|
| 743 |
+
"eval_f1": 0.5006190896494088,
|
| 744 |
+
"eval_loss": 1.2532024383544922,
|
| 745 |
+
"eval_precision": 0.5347134093222614,
|
| 746 |
+
"eval_recall": 0.4860070789020448,
|
| 747 |
+
"eval_runtime": 13.4318,
|
| 748 |
+
"eval_samples_per_second": 24.569,
|
| 749 |
+
"eval_steps_per_second": 3.127,
|
| 750 |
+
"step": 1800
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 4.932614555256064,
|
| 754 |
+
"eval_accuracy": 0.6060606060606061,
|
| 755 |
+
"eval_f1": 0.47951387194587414,
|
| 756 |
+
"eval_loss": 1.2477487325668335,
|
| 757 |
+
"eval_precision": 0.4841348163818714,
|
| 758 |
+
"eval_recall": 0.4813308565204252,
|
| 759 |
+
"eval_runtime": 13.4442,
|
| 760 |
+
"eval_samples_per_second": 24.546,
|
| 761 |
+
"eval_steps_per_second": 3.124,
|
| 762 |
+
"step": 1830
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 5.013477088948787,
|
| 766 |
+
"eval_accuracy": 0.6393939393939394,
|
| 767 |
+
"eval_f1": 0.4989217275259948,
|
| 768 |
+
"eval_loss": 1.2193477153778076,
|
| 769 |
+
"eval_precision": 0.5258493064306811,
|
| 770 |
+
"eval_recall": 0.4869117013024005,
|
| 771 |
+
"eval_runtime": 13.4387,
|
| 772 |
+
"eval_samples_per_second": 24.556,
|
| 773 |
+
"eval_steps_per_second": 3.125,
|
| 774 |
+
"step": 1860
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 5.09433962264151,
|
| 778 |
+
"eval_accuracy": 0.6424242424242425,
|
| 779 |
+
"eval_f1": 0.5115200267534169,
|
| 780 |
+
"eval_loss": 1.381670355796814,
|
| 781 |
+
"eval_precision": 0.5336989009157052,
|
| 782 |
+
"eval_recall": 0.5001792056559704,
|
| 783 |
+
"eval_runtime": 14.4622,
|
| 784 |
+
"eval_samples_per_second": 22.818,
|
| 785 |
+
"eval_steps_per_second": 2.904,
|
| 786 |
+
"step": 1890
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 5.175202156334231,
|
| 790 |
+
"eval_accuracy": 0.6272727272727273,
|
| 791 |
+
"eval_f1": 0.49956294294308473,
|
| 792 |
+
"eval_loss": 1.4055507183074951,
|
| 793 |
+
"eval_precision": 0.510424264019138,
|
| 794 |
+
"eval_recall": 0.49196977915696183,
|
| 795 |
+
"eval_runtime": 13.4719,
|
| 796 |
+
"eval_samples_per_second": 24.496,
|
| 797 |
+
"eval_steps_per_second": 3.118,
|
| 798 |
+
"step": 1920
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 5.256064690026954,
|
| 802 |
+
"eval_accuracy": 0.6303030303030303,
|
| 803 |
+
"eval_f1": 0.49657096716171206,
|
| 804 |
+
"eval_loss": 1.4433475732803345,
|
| 805 |
+
"eval_precision": 0.5023020702533887,
|
| 806 |
+
"eval_recall": 0.49618528944682805,
|
| 807 |
+
"eval_runtime": 12.5628,
|
| 808 |
+
"eval_samples_per_second": 26.268,
|
| 809 |
+
"eval_steps_per_second": 3.343,
|
| 810 |
+
"step": 1950
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 5.336927223719677,
|
| 814 |
+
"eval_accuracy": 0.6212121212121212,
|
| 815 |
+
"eval_f1": 0.4846771824827281,
|
| 816 |
+
"eval_loss": 1.4632567167282104,
|
| 817 |
+
"eval_precision": 0.49600215736518566,
|
| 818 |
+
"eval_recall": 0.4812536923450428,
|
| 819 |
+
"eval_runtime": 12.7703,
|
| 820 |
+
"eval_samples_per_second": 25.841,
|
| 821 |
+
"eval_steps_per_second": 3.289,
|
| 822 |
+
"step": 1980
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 5.390835579514825,
|
| 826 |
+
"grad_norm": 10.319731712341309,
|
| 827 |
+
"learning_rate": 5.0763701707097935e-06,
|
| 828 |
+
"loss": 0.4025,
|
| 829 |
+
"step": 2000
|
| 830 |
+
}
|
| 831 |
+
],
|
| 832 |
+
"logging_steps": 500,
|
| 833 |
+
"max_steps": 2226,
|
| 834 |
+
"num_input_tokens_seen": 0,
|
| 835 |
+
"num_train_epochs": 6,
|
| 836 |
+
"save_steps": 500,
|
| 837 |
+
"total_flos": 9271286495496432.0,
|
| 838 |
+
"train_batch_size": 8,
|
| 839 |
+
"trial_name": null,
|
| 840 |
+
"trial_params": null
|
| 841 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:802688dde848bb24ec73ba5b6be61944f4eb3d665acc269a57afb8710abb6b87
|
| 3 |
+
size 4984
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|