Upload 7 files
Browse files- config.json +58 -0
- pytorch_model.bin +3 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +15 -0
- tokenizer_config.json +55 -0
- trainer_state.json +729 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "../model/cino-small-v2",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"XLMRobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"gradient_checkpointing": false,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"id2label": {
|
| 15 |
+
"0": "Politics",
|
| 16 |
+
"1": "Economics",
|
| 17 |
+
"2": "Education",
|
| 18 |
+
"3": "Tourism",
|
| 19 |
+
"4": "Environment",
|
| 20 |
+
"5": "Language",
|
| 21 |
+
"6": "Literature",
|
| 22 |
+
"7": "Religion",
|
| 23 |
+
"8": "Arts",
|
| 24 |
+
"9": "Medicine",
|
| 25 |
+
"10": "Customs",
|
| 26 |
+
"11": "Instruments"
|
| 27 |
+
},
|
| 28 |
+
"initializer_range": 0.02,
|
| 29 |
+
"intermediate_size": 3072,
|
| 30 |
+
"label2id": {
|
| 31 |
+
"Arts": 8,
|
| 32 |
+
"Customs": 10,
|
| 33 |
+
"Economics": 1,
|
| 34 |
+
"Education": 2,
|
| 35 |
+
"Environment": 4,
|
| 36 |
+
"Instruments": 11,
|
| 37 |
+
"Language": 5,
|
| 38 |
+
"Literature": 6,
|
| 39 |
+
"Medicine": 9,
|
| 40 |
+
"Politics": 0,
|
| 41 |
+
"Religion": 7,
|
| 42 |
+
"Tourism": 3
|
| 43 |
+
},
|
| 44 |
+
"layer_norm_eps": 1e-05,
|
| 45 |
+
"max_position_embeddings": 514,
|
| 46 |
+
"model_type": "xlm-roberta",
|
| 47 |
+
"num_attention_heads": 12,
|
| 48 |
+
"num_hidden_layers": 6,
|
| 49 |
+
"output_past": true,
|
| 50 |
+
"pad_token_id": 1,
|
| 51 |
+
"position_embedding_type": "absolute",
|
| 52 |
+
"problem_type": "single_label_classification",
|
| 53 |
+
"torch_dtype": "float32",
|
| 54 |
+
"transformers_version": "4.36.2",
|
| 55 |
+
"type_vocab_size": 1,
|
| 56 |
+
"use_cache": true,
|
| 57 |
+
"vocab_size": 135359
|
| 58 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36828bf7fd1309375a4dc57652d1148df8c5ce14e1051cfdf1947472d44db87f
|
| 3 |
+
size 589953711
|
sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abec8706178924453be115cd2da858ef32de70ba60d0c10300822a732a868cf7
|
| 3 |
+
size 2814898
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"content": "<mask>",
|
| 7 |
+
"lstrip": true,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"sep_token": "</s>",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 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": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"135358": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"mask_token": "<mask>",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"pad_token": "<pad>",
|
| 51 |
+
"sep_token": "</s>",
|
| 52 |
+
"sp_model_kwargs": {},
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,729 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6838614548013121,
|
| 3 |
+
"best_model_checkpoint": "../saved_model/cino-small-v2_tncc-title_v3/checkpoint-7424",
|
| 4 |
+
"epoch": 40.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 9280,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.49838187702265374,
|
| 14 |
+
"eval_loss": 1.6120452880859375,
|
| 15 |
+
"eval_macro-f1": 0.3080200150906938,
|
| 16 |
+
"eval_macro-precision": 0.4470989896292958,
|
| 17 |
+
"eval_macro-recall": 0.33407156196923765,
|
| 18 |
+
"eval_runtime": 3.5999,
|
| 19 |
+
"eval_samples_per_second": 257.505,
|
| 20 |
+
"eval_steps_per_second": 8.056,
|
| 21 |
+
"eval_weighted-f1": 0.423917385683886,
|
| 22 |
+
"eval_weighted-precision": 0.5166350613267663,
|
| 23 |
+
"eval_weighted-recall": 0.49838187702265374,
|
| 24 |
+
"step": 232
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 2.0,
|
| 28 |
+
"eval_accuracy": 0.6472491909385113,
|
| 29 |
+
"eval_loss": 1.1052128076553345,
|
| 30 |
+
"eval_macro-f1": 0.5457350988248497,
|
| 31 |
+
"eval_macro-precision": 0.6404982007168077,
|
| 32 |
+
"eval_macro-recall": 0.559508539878825,
|
| 33 |
+
"eval_runtime": 3.6454,
|
| 34 |
+
"eval_samples_per_second": 254.294,
|
| 35 |
+
"eval_steps_per_second": 7.955,
|
| 36 |
+
"eval_weighted-f1": 0.6324317601702024,
|
| 37 |
+
"eval_weighted-precision": 0.6658002868686018,
|
| 38 |
+
"eval_weighted-recall": 0.6472491909385113,
|
| 39 |
+
"step": 464
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 2.16,
|
| 43 |
+
"learning_rate": 2.6939655172413796e-05,
|
| 44 |
+
"loss": 1.7073,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 3.0,
|
| 49 |
+
"eval_accuracy": 0.6709816612729234,
|
| 50 |
+
"eval_loss": 1.0001471042633057,
|
| 51 |
+
"eval_macro-f1": 0.6065301675158895,
|
| 52 |
+
"eval_macro-precision": 0.654418913280353,
|
| 53 |
+
"eval_macro-recall": 0.6099139180853289,
|
| 54 |
+
"eval_runtime": 3.6176,
|
| 55 |
+
"eval_samples_per_second": 256.245,
|
| 56 |
+
"eval_steps_per_second": 8.016,
|
| 57 |
+
"eval_weighted-f1": 0.6605809984207812,
|
| 58 |
+
"eval_weighted-precision": 0.6778557170715117,
|
| 59 |
+
"eval_weighted-recall": 0.6709816612729234,
|
| 60 |
+
"step": 696
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.0,
|
| 64 |
+
"eval_accuracy": 0.686084142394822,
|
| 65 |
+
"eval_loss": 0.9795448780059814,
|
| 66 |
+
"eval_macro-f1": 0.645770505361501,
|
| 67 |
+
"eval_macro-precision": 0.6624511940296568,
|
| 68 |
+
"eval_macro-recall": 0.6454397105147797,
|
| 69 |
+
"eval_runtime": 3.6312,
|
| 70 |
+
"eval_samples_per_second": 255.29,
|
| 71 |
+
"eval_steps_per_second": 7.986,
|
| 72 |
+
"eval_weighted-f1": 0.680474769885252,
|
| 73 |
+
"eval_weighted-precision": 0.6882611664755186,
|
| 74 |
+
"eval_weighted-recall": 0.686084142394822,
|
| 75 |
+
"step": 928
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 4.31,
|
| 79 |
+
"learning_rate": 4.9568965517241384e-05,
|
| 80 |
+
"loss": 0.8684,
|
| 81 |
+
"step": 1000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 5.0,
|
| 85 |
+
"eval_accuracy": 0.6947141316073355,
|
| 86 |
+
"eval_loss": 0.9905449748039246,
|
| 87 |
+
"eval_macro-f1": 0.6674507510628628,
|
| 88 |
+
"eval_macro-precision": 0.6625099174020581,
|
| 89 |
+
"eval_macro-recall": 0.6829958649269056,
|
| 90 |
+
"eval_runtime": 3.6247,
|
| 91 |
+
"eval_samples_per_second": 255.745,
|
| 92 |
+
"eval_steps_per_second": 8.001,
|
| 93 |
+
"eval_weighted-f1": 0.6952993996868073,
|
| 94 |
+
"eval_weighted-precision": 0.7036333325841285,
|
| 95 |
+
"eval_weighted-recall": 0.6947141316073355,
|
| 96 |
+
"step": 1160
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 6.0,
|
| 100 |
+
"eval_accuracy": 0.7044228694714132,
|
| 101 |
+
"eval_loss": 1.0819956064224243,
|
| 102 |
+
"eval_macro-f1": 0.6835424949893771,
|
| 103 |
+
"eval_macro-precision": 0.6831322585201048,
|
| 104 |
+
"eval_macro-recall": 0.6897387284921295,
|
| 105 |
+
"eval_runtime": 3.6099,
|
| 106 |
+
"eval_samples_per_second": 256.793,
|
| 107 |
+
"eval_steps_per_second": 8.033,
|
| 108 |
+
"eval_weighted-f1": 0.7063455583024486,
|
| 109 |
+
"eval_weighted-precision": 0.7133818379200453,
|
| 110 |
+
"eval_weighted-recall": 0.7044228694714132,
|
| 111 |
+
"step": 1392
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 6.47,
|
| 115 |
+
"learning_rate": 4.6575670498084294e-05,
|
| 116 |
+
"loss": 0.4904,
|
| 117 |
+
"step": 1500
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 7.0,
|
| 121 |
+
"eval_accuracy": 0.6850053937432579,
|
| 122 |
+
"eval_loss": 1.2797751426696777,
|
| 123 |
+
"eval_macro-f1": 0.6529772810537373,
|
| 124 |
+
"eval_macro-precision": 0.6632776375522283,
|
| 125 |
+
"eval_macro-recall": 0.6621596382156031,
|
| 126 |
+
"eval_runtime": 3.621,
|
| 127 |
+
"eval_samples_per_second": 256.009,
|
| 128 |
+
"eval_steps_per_second": 8.009,
|
| 129 |
+
"eval_weighted-f1": 0.6864725766536225,
|
| 130 |
+
"eval_weighted-precision": 0.7003185776965812,
|
| 131 |
+
"eval_weighted-recall": 0.6850053937432579,
|
| 132 |
+
"step": 1624
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 8.0,
|
| 136 |
+
"eval_accuracy": 0.6796116504854369,
|
| 137 |
+
"eval_loss": 1.3903510570526123,
|
| 138 |
+
"eval_macro-f1": 0.6510499118775824,
|
| 139 |
+
"eval_macro-precision": 0.6651601590963526,
|
| 140 |
+
"eval_macro-recall": 0.6479225440423547,
|
| 141 |
+
"eval_runtime": 3.6152,
|
| 142 |
+
"eval_samples_per_second": 256.415,
|
| 143 |
+
"eval_steps_per_second": 8.022,
|
| 144 |
+
"eval_weighted-f1": 0.6783723538205079,
|
| 145 |
+
"eval_weighted-precision": 0.6881154476659207,
|
| 146 |
+
"eval_weighted-recall": 0.6796116504854369,
|
| 147 |
+
"step": 1856
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 8.62,
|
| 151 |
+
"learning_rate": 4.3582375478927204e-05,
|
| 152 |
+
"loss": 0.2376,
|
| 153 |
+
"step": 2000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 9.0,
|
| 157 |
+
"eval_accuracy": 0.674217907227616,
|
| 158 |
+
"eval_loss": 1.4255810976028442,
|
| 159 |
+
"eval_macro-f1": 0.6424901861691795,
|
| 160 |
+
"eval_macro-precision": 0.6705435579848175,
|
| 161 |
+
"eval_macro-recall": 0.6378243354162482,
|
| 162 |
+
"eval_runtime": 3.6179,
|
| 163 |
+
"eval_samples_per_second": 256.225,
|
| 164 |
+
"eval_steps_per_second": 8.016,
|
| 165 |
+
"eval_weighted-f1": 0.6736702968697759,
|
| 166 |
+
"eval_weighted-precision": 0.6879912323928888,
|
| 167 |
+
"eval_weighted-recall": 0.674217907227616,
|
| 168 |
+
"step": 2088
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 10.0,
|
| 172 |
+
"eval_accuracy": 0.6882416396979504,
|
| 173 |
+
"eval_loss": 1.5364353656768799,
|
| 174 |
+
"eval_macro-f1": 0.6654116450501577,
|
| 175 |
+
"eval_macro-precision": 0.6574262835666708,
|
| 176 |
+
"eval_macro-recall": 0.6801476025508416,
|
| 177 |
+
"eval_runtime": 3.6024,
|
| 178 |
+
"eval_samples_per_second": 257.33,
|
| 179 |
+
"eval_steps_per_second": 8.05,
|
| 180 |
+
"eval_weighted-f1": 0.6895751725299344,
|
| 181 |
+
"eval_weighted-precision": 0.6945151299816443,
|
| 182 |
+
"eval_weighted-recall": 0.6882416396979504,
|
| 183 |
+
"step": 2320
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 10.78,
|
| 187 |
+
"learning_rate": 4.058908045977012e-05,
|
| 188 |
+
"loss": 0.1495,
|
| 189 |
+
"step": 2500
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 11.0,
|
| 193 |
+
"eval_accuracy": 0.6828478964401294,
|
| 194 |
+
"eval_loss": 1.6653459072113037,
|
| 195 |
+
"eval_macro-f1": 0.6512014541822767,
|
| 196 |
+
"eval_macro-precision": 0.6627229410666052,
|
| 197 |
+
"eval_macro-recall": 0.6485507026145255,
|
| 198 |
+
"eval_runtime": 3.5928,
|
| 199 |
+
"eval_samples_per_second": 258.014,
|
| 200 |
+
"eval_steps_per_second": 8.072,
|
| 201 |
+
"eval_weighted-f1": 0.6800769417795847,
|
| 202 |
+
"eval_weighted-precision": 0.6882902408612807,
|
| 203 |
+
"eval_weighted-recall": 0.6828478964401294,
|
| 204 |
+
"step": 2552
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 12.0,
|
| 208 |
+
"eval_accuracy": 0.6839266450916937,
|
| 209 |
+
"eval_loss": 1.6987706422805786,
|
| 210 |
+
"eval_macro-f1": 0.6479532709170456,
|
| 211 |
+
"eval_macro-precision": 0.6580973534101414,
|
| 212 |
+
"eval_macro-recall": 0.6474043097812547,
|
| 213 |
+
"eval_runtime": 3.6004,
|
| 214 |
+
"eval_samples_per_second": 257.471,
|
| 215 |
+
"eval_steps_per_second": 8.055,
|
| 216 |
+
"eval_weighted-f1": 0.6841454877516878,
|
| 217 |
+
"eval_weighted-precision": 0.6902044355228595,
|
| 218 |
+
"eval_weighted-recall": 0.6839266450916937,
|
| 219 |
+
"step": 2784
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 12.93,
|
| 223 |
+
"learning_rate": 3.759578544061303e-05,
|
| 224 |
+
"loss": 0.1058,
|
| 225 |
+
"step": 3000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 13.0,
|
| 229 |
+
"eval_accuracy": 0.6817691477885652,
|
| 230 |
+
"eval_loss": 1.747767448425293,
|
| 231 |
+
"eval_macro-f1": 0.6556363951964088,
|
| 232 |
+
"eval_macro-precision": 0.6700908993068831,
|
| 233 |
+
"eval_macro-recall": 0.6561971975750561,
|
| 234 |
+
"eval_runtime": 3.6016,
|
| 235 |
+
"eval_samples_per_second": 257.383,
|
| 236 |
+
"eval_steps_per_second": 8.052,
|
| 237 |
+
"eval_weighted-f1": 0.6814628702483411,
|
| 238 |
+
"eval_weighted-precision": 0.6867012225910729,
|
| 239 |
+
"eval_weighted-recall": 0.6817691477885652,
|
| 240 |
+
"step": 3016
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 14.0,
|
| 244 |
+
"eval_accuracy": 0.674217907227616,
|
| 245 |
+
"eval_loss": 1.918135166168213,
|
| 246 |
+
"eval_macro-f1": 0.6397376565535934,
|
| 247 |
+
"eval_macro-precision": 0.6549903289689117,
|
| 248 |
+
"eval_macro-recall": 0.6473498155004918,
|
| 249 |
+
"eval_runtime": 3.5879,
|
| 250 |
+
"eval_samples_per_second": 258.368,
|
| 251 |
+
"eval_steps_per_second": 8.083,
|
| 252 |
+
"eval_weighted-f1": 0.6765985742439555,
|
| 253 |
+
"eval_weighted-precision": 0.6901672277432537,
|
| 254 |
+
"eval_weighted-recall": 0.674217907227616,
|
| 255 |
+
"step": 3248
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 15.0,
|
| 259 |
+
"eval_accuracy": 0.6957928802588996,
|
| 260 |
+
"eval_loss": 1.9204330444335938,
|
| 261 |
+
"eval_macro-f1": 0.6658200325646808,
|
| 262 |
+
"eval_macro-precision": 0.6663228409772431,
|
| 263 |
+
"eval_macro-recall": 0.6692964824595501,
|
| 264 |
+
"eval_runtime": 3.5954,
|
| 265 |
+
"eval_samples_per_second": 257.831,
|
| 266 |
+
"eval_steps_per_second": 8.066,
|
| 267 |
+
"eval_weighted-f1": 0.6957701345209284,
|
| 268 |
+
"eval_weighted-precision": 0.69903863020904,
|
| 269 |
+
"eval_weighted-recall": 0.6957928802588996,
|
| 270 |
+
"step": 3480
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 15.09,
|
| 274 |
+
"learning_rate": 3.460249042145594e-05,
|
| 275 |
+
"loss": 0.0775,
|
| 276 |
+
"step": 3500
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 16.0,
|
| 280 |
+
"eval_accuracy": 0.6752966558791802,
|
| 281 |
+
"eval_loss": 1.9924219846725464,
|
| 282 |
+
"eval_macro-f1": 0.6392465953749347,
|
| 283 |
+
"eval_macro-precision": 0.6561467817801693,
|
| 284 |
+
"eval_macro-recall": 0.6402901277480958,
|
| 285 |
+
"eval_runtime": 3.6181,
|
| 286 |
+
"eval_samples_per_second": 256.214,
|
| 287 |
+
"eval_steps_per_second": 8.015,
|
| 288 |
+
"eval_weighted-f1": 0.673631619187917,
|
| 289 |
+
"eval_weighted-precision": 0.6858049526730299,
|
| 290 |
+
"eval_weighted-recall": 0.6752966558791802,
|
| 291 |
+
"step": 3712
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 17.0,
|
| 295 |
+
"eval_accuracy": 0.686084142394822,
|
| 296 |
+
"eval_loss": 2.0478932857513428,
|
| 297 |
+
"eval_macro-f1": 0.6463214194781717,
|
| 298 |
+
"eval_macro-precision": 0.6570967483791249,
|
| 299 |
+
"eval_macro-recall": 0.6503589081998865,
|
| 300 |
+
"eval_runtime": 3.5962,
|
| 301 |
+
"eval_samples_per_second": 257.771,
|
| 302 |
+
"eval_steps_per_second": 8.064,
|
| 303 |
+
"eval_weighted-f1": 0.6852865958011676,
|
| 304 |
+
"eval_weighted-precision": 0.6939600166349821,
|
| 305 |
+
"eval_weighted-recall": 0.686084142394822,
|
| 306 |
+
"step": 3944
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 17.24,
|
| 310 |
+
"learning_rate": 3.160919540229885e-05,
|
| 311 |
+
"loss": 0.0687,
|
| 312 |
+
"step": 4000
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 18.0,
|
| 316 |
+
"eval_accuracy": 0.674217907227616,
|
| 317 |
+
"eval_loss": 2.0522632598876953,
|
| 318 |
+
"eval_macro-f1": 0.6405085398516451,
|
| 319 |
+
"eval_macro-precision": 0.6662302886679464,
|
| 320 |
+
"eval_macro-recall": 0.6393535717017708,
|
| 321 |
+
"eval_runtime": 3.6131,
|
| 322 |
+
"eval_samples_per_second": 256.565,
|
| 323 |
+
"eval_steps_per_second": 8.026,
|
| 324 |
+
"eval_weighted-f1": 0.674060546264426,
|
| 325 |
+
"eval_weighted-precision": 0.6894424888472697,
|
| 326 |
+
"eval_weighted-recall": 0.674217907227616,
|
| 327 |
+
"step": 4176
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 19.0,
|
| 331 |
+
"eval_accuracy": 0.6828478964401294,
|
| 332 |
+
"eval_loss": 1.9823521375656128,
|
| 333 |
+
"eval_macro-f1": 0.651754402283885,
|
| 334 |
+
"eval_macro-precision": 0.6440801367882046,
|
| 335 |
+
"eval_macro-recall": 0.6683888090630882,
|
| 336 |
+
"eval_runtime": 3.6037,
|
| 337 |
+
"eval_samples_per_second": 257.233,
|
| 338 |
+
"eval_steps_per_second": 8.047,
|
| 339 |
+
"eval_weighted-f1": 0.6821812966643793,
|
| 340 |
+
"eval_weighted-precision": 0.68760298706152,
|
| 341 |
+
"eval_weighted-recall": 0.6828478964401294,
|
| 342 |
+
"step": 4408
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 19.4,
|
| 346 |
+
"learning_rate": 2.8615900383141765e-05,
|
| 347 |
+
"loss": 0.0577,
|
| 348 |
+
"step": 4500
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 20.0,
|
| 352 |
+
"eval_accuracy": 0.6914778856526429,
|
| 353 |
+
"eval_loss": 2.0154926776885986,
|
| 354 |
+
"eval_macro-f1": 0.6659900934162544,
|
| 355 |
+
"eval_macro-precision": 0.6712869574383351,
|
| 356 |
+
"eval_macro-recall": 0.6703643770947892,
|
| 357 |
+
"eval_runtime": 3.6164,
|
| 358 |
+
"eval_samples_per_second": 256.332,
|
| 359 |
+
"eval_steps_per_second": 8.019,
|
| 360 |
+
"eval_weighted-f1": 0.6906611756556518,
|
| 361 |
+
"eval_weighted-precision": 0.6985916032921785,
|
| 362 |
+
"eval_weighted-recall": 0.6914778856526429,
|
| 363 |
+
"step": 4640
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 21.0,
|
| 367 |
+
"eval_accuracy": 0.6828478964401294,
|
| 368 |
+
"eval_loss": 2.05501389503479,
|
| 369 |
+
"eval_macro-f1": 0.6486634410090307,
|
| 370 |
+
"eval_macro-precision": 0.6780446089713759,
|
| 371 |
+
"eval_macro-recall": 0.6376704510538888,
|
| 372 |
+
"eval_runtime": 3.5904,
|
| 373 |
+
"eval_samples_per_second": 258.186,
|
| 374 |
+
"eval_steps_per_second": 8.077,
|
| 375 |
+
"eval_weighted-f1": 0.6755183997397308,
|
| 376 |
+
"eval_weighted-precision": 0.6900556731555942,
|
| 377 |
+
"eval_weighted-recall": 0.6828478964401294,
|
| 378 |
+
"step": 4872
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 21.55,
|
| 382 |
+
"learning_rate": 2.5622605363984675e-05,
|
| 383 |
+
"loss": 0.0489,
|
| 384 |
+
"step": 5000
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 22.0,
|
| 388 |
+
"eval_accuracy": 0.6796116504854369,
|
| 389 |
+
"eval_loss": 2.0277795791625977,
|
| 390 |
+
"eval_macro-f1": 0.6505710558420628,
|
| 391 |
+
"eval_macro-precision": 0.6513644147664868,
|
| 392 |
+
"eval_macro-recall": 0.6562986108294033,
|
| 393 |
+
"eval_runtime": 3.5971,
|
| 394 |
+
"eval_samples_per_second": 257.71,
|
| 395 |
+
"eval_steps_per_second": 8.062,
|
| 396 |
+
"eval_weighted-f1": 0.6787275977253558,
|
| 397 |
+
"eval_weighted-precision": 0.6831300629390828,
|
| 398 |
+
"eval_weighted-recall": 0.6796116504854369,
|
| 399 |
+
"step": 5104
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 23.0,
|
| 403 |
+
"eval_accuracy": 0.6828478964401294,
|
| 404 |
+
"eval_loss": 2.033602237701416,
|
| 405 |
+
"eval_macro-f1": 0.6518094813918546,
|
| 406 |
+
"eval_macro-precision": 0.6546940340503233,
|
| 407 |
+
"eval_macro-recall": 0.6544133192355402,
|
| 408 |
+
"eval_runtime": 3.6042,
|
| 409 |
+
"eval_samples_per_second": 257.199,
|
| 410 |
+
"eval_steps_per_second": 8.046,
|
| 411 |
+
"eval_weighted-f1": 0.6812281754499014,
|
| 412 |
+
"eval_weighted-precision": 0.6846668946206854,
|
| 413 |
+
"eval_weighted-recall": 0.6828478964401294,
|
| 414 |
+
"step": 5336
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 23.71,
|
| 418 |
+
"learning_rate": 2.2629310344827588e-05,
|
| 419 |
+
"loss": 0.0431,
|
| 420 |
+
"step": 5500
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 24.0,
|
| 424 |
+
"eval_accuracy": 0.697950377562028,
|
| 425 |
+
"eval_loss": 2.0058302879333496,
|
| 426 |
+
"eval_macro-f1": 0.672432519382824,
|
| 427 |
+
"eval_macro-precision": 0.7031870372780317,
|
| 428 |
+
"eval_macro-recall": 0.6570710914948313,
|
| 429 |
+
"eval_runtime": 3.6053,
|
| 430 |
+
"eval_samples_per_second": 257.123,
|
| 431 |
+
"eval_steps_per_second": 8.044,
|
| 432 |
+
"eval_weighted-f1": 0.6956222916424519,
|
| 433 |
+
"eval_weighted-precision": 0.7050444644388131,
|
| 434 |
+
"eval_weighted-recall": 0.697950377562028,
|
| 435 |
+
"step": 5568
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 25.0,
|
| 439 |
+
"eval_accuracy": 0.6903991370010788,
|
| 440 |
+
"eval_loss": 2.0136592388153076,
|
| 441 |
+
"eval_macro-f1": 0.6593408320415061,
|
| 442 |
+
"eval_macro-precision": 0.6670076069784594,
|
| 443 |
+
"eval_macro-recall": 0.6569151453566795,
|
| 444 |
+
"eval_runtime": 3.6053,
|
| 445 |
+
"eval_samples_per_second": 257.124,
|
| 446 |
+
"eval_steps_per_second": 8.044,
|
| 447 |
+
"eval_weighted-f1": 0.6888694520013063,
|
| 448 |
+
"eval_weighted-precision": 0.6914605561085521,
|
| 449 |
+
"eval_weighted-recall": 0.6903991370010788,
|
| 450 |
+
"step": 5800
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 25.86,
|
| 454 |
+
"learning_rate": 1.9636015325670498e-05,
|
| 455 |
+
"loss": 0.0372,
|
| 456 |
+
"step": 6000
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 26.0,
|
| 460 |
+
"eval_accuracy": 0.6763754045307443,
|
| 461 |
+
"eval_loss": 2.21449875831604,
|
| 462 |
+
"eval_macro-f1": 0.6400140077032914,
|
| 463 |
+
"eval_macro-precision": 0.649984379772116,
|
| 464 |
+
"eval_macro-recall": 0.6499316041741732,
|
| 465 |
+
"eval_runtime": 3.6097,
|
| 466 |
+
"eval_samples_per_second": 256.806,
|
| 467 |
+
"eval_steps_per_second": 8.034,
|
| 468 |
+
"eval_weighted-f1": 0.6791315210465927,
|
| 469 |
+
"eval_weighted-precision": 0.6961390172701661,
|
| 470 |
+
"eval_weighted-recall": 0.6763754045307443,
|
| 471 |
+
"step": 6032
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 27.0,
|
| 475 |
+
"eval_accuracy": 0.6882416396979504,
|
| 476 |
+
"eval_loss": 2.1051621437072754,
|
| 477 |
+
"eval_macro-f1": 0.6563104764431434,
|
| 478 |
+
"eval_macro-precision": 0.6584303947936838,
|
| 479 |
+
"eval_macro-recall": 0.6627288799113591,
|
| 480 |
+
"eval_runtime": 3.5999,
|
| 481 |
+
"eval_samples_per_second": 257.511,
|
| 482 |
+
"eval_steps_per_second": 8.056,
|
| 483 |
+
"eval_weighted-f1": 0.6864978936047862,
|
| 484 |
+
"eval_weighted-precision": 0.6922661220105522,
|
| 485 |
+
"eval_weighted-recall": 0.6882416396979504,
|
| 486 |
+
"step": 6264
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 28.0,
|
| 490 |
+
"eval_accuracy": 0.6947141316073355,
|
| 491 |
+
"eval_loss": 2.096506357192993,
|
| 492 |
+
"eval_macro-f1": 0.6701111894500894,
|
| 493 |
+
"eval_macro-precision": 0.6973353037360721,
|
| 494 |
+
"eval_macro-recall": 0.6569800023874061,
|
| 495 |
+
"eval_runtime": 3.6151,
|
| 496 |
+
"eval_samples_per_second": 256.427,
|
| 497 |
+
"eval_steps_per_second": 8.022,
|
| 498 |
+
"eval_weighted-f1": 0.6888264314204101,
|
| 499 |
+
"eval_weighted-precision": 0.6996849509212928,
|
| 500 |
+
"eval_weighted-recall": 0.6947141316073355,
|
| 501 |
+
"step": 6496
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 28.02,
|
| 505 |
+
"learning_rate": 1.664272030651341e-05,
|
| 506 |
+
"loss": 0.0357,
|
| 507 |
+
"step": 6500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 29.0,
|
| 511 |
+
"eval_accuracy": 0.6947141316073355,
|
| 512 |
+
"eval_loss": 2.094637632369995,
|
| 513 |
+
"eval_macro-f1": 0.6685294866541375,
|
| 514 |
+
"eval_macro-precision": 0.6813823548362715,
|
| 515 |
+
"eval_macro-recall": 0.6641734635067332,
|
| 516 |
+
"eval_runtime": 3.6038,
|
| 517 |
+
"eval_samples_per_second": 257.226,
|
| 518 |
+
"eval_steps_per_second": 8.047,
|
| 519 |
+
"eval_weighted-f1": 0.6943390563150669,
|
| 520 |
+
"eval_weighted-precision": 0.7001493288391278,
|
| 521 |
+
"eval_weighted-recall": 0.6947141316073355,
|
| 522 |
+
"step": 6728
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 30.0,
|
| 526 |
+
"eval_accuracy": 0.6893203883495146,
|
| 527 |
+
"eval_loss": 2.0883684158325195,
|
| 528 |
+
"eval_macro-f1": 0.6602023917903279,
|
| 529 |
+
"eval_macro-precision": 0.6628650635074098,
|
| 530 |
+
"eval_macro-recall": 0.6659307948401801,
|
| 531 |
+
"eval_runtime": 3.6134,
|
| 532 |
+
"eval_samples_per_second": 256.543,
|
| 533 |
+
"eval_steps_per_second": 8.026,
|
| 534 |
+
"eval_weighted-f1": 0.6902134268034101,
|
| 535 |
+
"eval_weighted-precision": 0.6951233806637598,
|
| 536 |
+
"eval_weighted-recall": 0.6893203883495146,
|
| 537 |
+
"step": 6960
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 30.17,
|
| 541 |
+
"learning_rate": 1.3649425287356324e-05,
|
| 542 |
+
"loss": 0.0286,
|
| 543 |
+
"step": 7000
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 31.0,
|
| 547 |
+
"eval_accuracy": 0.703344120819849,
|
| 548 |
+
"eval_loss": 2.126500368118286,
|
| 549 |
+
"eval_macro-f1": 0.6756885174044472,
|
| 550 |
+
"eval_macro-precision": 0.6887882836127188,
|
| 551 |
+
"eval_macro-recall": 0.6701991060577401,
|
| 552 |
+
"eval_runtime": 3.6102,
|
| 553 |
+
"eval_samples_per_second": 256.771,
|
| 554 |
+
"eval_steps_per_second": 8.033,
|
| 555 |
+
"eval_weighted-f1": 0.7025418148884912,
|
| 556 |
+
"eval_weighted-precision": 0.7070891495572698,
|
| 557 |
+
"eval_weighted-recall": 0.703344120819849,
|
| 558 |
+
"step": 7192
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 32.0,
|
| 562 |
+
"eval_accuracy": 0.7022653721682848,
|
| 563 |
+
"eval_loss": 2.110332489013672,
|
| 564 |
+
"eval_macro-f1": 0.6838614548013121,
|
| 565 |
+
"eval_macro-precision": 0.6918280242510296,
|
| 566 |
+
"eval_macro-recall": 0.6819433903345183,
|
| 567 |
+
"eval_runtime": 3.6092,
|
| 568 |
+
"eval_samples_per_second": 256.84,
|
| 569 |
+
"eval_steps_per_second": 8.035,
|
| 570 |
+
"eval_weighted-f1": 0.7015731601193966,
|
| 571 |
+
"eval_weighted-precision": 0.7068600524270214,
|
| 572 |
+
"eval_weighted-recall": 0.7022653721682848,
|
| 573 |
+
"step": 7424
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 32.33,
|
| 577 |
+
"learning_rate": 1.0656130268199234e-05,
|
| 578 |
+
"loss": 0.0248,
|
| 579 |
+
"step": 7500
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 33.0,
|
| 583 |
+
"eval_accuracy": 0.6925566343042071,
|
| 584 |
+
"eval_loss": 2.076742649078369,
|
| 585 |
+
"eval_macro-f1": 0.6677185111141611,
|
| 586 |
+
"eval_macro-precision": 0.6682997232361648,
|
| 587 |
+
"eval_macro-recall": 0.6726620472560384,
|
| 588 |
+
"eval_runtime": 3.6011,
|
| 589 |
+
"eval_samples_per_second": 257.423,
|
| 590 |
+
"eval_steps_per_second": 8.053,
|
| 591 |
+
"eval_weighted-f1": 0.693162535561584,
|
| 592 |
+
"eval_weighted-precision": 0.6976019284855784,
|
| 593 |
+
"eval_weighted-recall": 0.6925566343042071,
|
| 594 |
+
"step": 7656
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 34.0,
|
| 598 |
+
"eval_accuracy": 0.686084142394822,
|
| 599 |
+
"eval_loss": 2.1223480701446533,
|
| 600 |
+
"eval_macro-f1": 0.661874148838316,
|
| 601 |
+
"eval_macro-precision": 0.6610753547813631,
|
| 602 |
+
"eval_macro-recall": 0.6713446322584083,
|
| 603 |
+
"eval_runtime": 3.6215,
|
| 604 |
+
"eval_samples_per_second": 255.972,
|
| 605 |
+
"eval_steps_per_second": 8.008,
|
| 606 |
+
"eval_weighted-f1": 0.6841907777006906,
|
| 607 |
+
"eval_weighted-precision": 0.6894138997369361,
|
| 608 |
+
"eval_weighted-recall": 0.686084142394822,
|
| 609 |
+
"step": 7888
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 34.48,
|
| 613 |
+
"learning_rate": 7.662835249042145e-06,
|
| 614 |
+
"loss": 0.0235,
|
| 615 |
+
"step": 8000
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 35.0,
|
| 619 |
+
"eval_accuracy": 0.6903991370010788,
|
| 620 |
+
"eval_loss": 2.133087635040283,
|
| 621 |
+
"eval_macro-f1": 0.662335562798904,
|
| 622 |
+
"eval_macro-precision": 0.6736920728738612,
|
| 623 |
+
"eval_macro-recall": 0.6597960660569285,
|
| 624 |
+
"eval_runtime": 3.6194,
|
| 625 |
+
"eval_samples_per_second": 256.117,
|
| 626 |
+
"eval_steps_per_second": 8.012,
|
| 627 |
+
"eval_weighted-f1": 0.6874362655396534,
|
| 628 |
+
"eval_weighted-precision": 0.6919778724812079,
|
| 629 |
+
"eval_weighted-recall": 0.6903991370010788,
|
| 630 |
+
"step": 8120
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 36.0,
|
| 634 |
+
"eval_accuracy": 0.697950377562028,
|
| 635 |
+
"eval_loss": 2.122838020324707,
|
| 636 |
+
"eval_macro-f1": 0.670156602236844,
|
| 637 |
+
"eval_macro-precision": 0.6690594327579752,
|
| 638 |
+
"eval_macro-recall": 0.6751260562968397,
|
| 639 |
+
"eval_runtime": 3.6045,
|
| 640 |
+
"eval_samples_per_second": 257.182,
|
| 641 |
+
"eval_steps_per_second": 8.046,
|
| 642 |
+
"eval_weighted-f1": 0.6980890949772001,
|
| 643 |
+
"eval_weighted-precision": 0.7007470226449659,
|
| 644 |
+
"eval_weighted-recall": 0.697950377562028,
|
| 645 |
+
"step": 8352
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 36.64,
|
| 649 |
+
"learning_rate": 4.669540229885057e-06,
|
| 650 |
+
"loss": 0.0211,
|
| 651 |
+
"step": 8500
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 37.0,
|
| 655 |
+
"eval_accuracy": 0.6990291262135923,
|
| 656 |
+
"eval_loss": 2.1283230781555176,
|
| 657 |
+
"eval_macro-f1": 0.6734553818830601,
|
| 658 |
+
"eval_macro-precision": 0.6869153083999092,
|
| 659 |
+
"eval_macro-recall": 0.6662304826991132,
|
| 660 |
+
"eval_runtime": 3.6053,
|
| 661 |
+
"eval_samples_per_second": 257.124,
|
| 662 |
+
"eval_steps_per_second": 8.044,
|
| 663 |
+
"eval_weighted-f1": 0.6984888091028846,
|
| 664 |
+
"eval_weighted-precision": 0.7035914434375581,
|
| 665 |
+
"eval_weighted-recall": 0.6990291262135923,
|
| 666 |
+
"step": 8584
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 38.0,
|
| 670 |
+
"eval_accuracy": 0.6957928802588996,
|
| 671 |
+
"eval_loss": 2.121021270751953,
|
| 672 |
+
"eval_macro-f1": 0.6693193516472609,
|
| 673 |
+
"eval_macro-precision": 0.6852915803739035,
|
| 674 |
+
"eval_macro-recall": 0.6603953493083169,
|
| 675 |
+
"eval_runtime": 3.5959,
|
| 676 |
+
"eval_samples_per_second": 257.795,
|
| 677 |
+
"eval_steps_per_second": 8.065,
|
| 678 |
+
"eval_weighted-f1": 0.6937517174106657,
|
| 679 |
+
"eval_weighted-precision": 0.6980351944602388,
|
| 680 |
+
"eval_weighted-recall": 0.6957928802588996,
|
| 681 |
+
"step": 8816
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 38.79,
|
| 685 |
+
"learning_rate": 1.6762452107279694e-06,
|
| 686 |
+
"loss": 0.0175,
|
| 687 |
+
"step": 9000
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 39.0,
|
| 691 |
+
"eval_accuracy": 0.6957928802588996,
|
| 692 |
+
"eval_loss": 2.125298023223877,
|
| 693 |
+
"eval_macro-f1": 0.670175716053461,
|
| 694 |
+
"eval_macro-precision": 0.6823954080156959,
|
| 695 |
+
"eval_macro-recall": 0.6637507266661784,
|
| 696 |
+
"eval_runtime": 3.6051,
|
| 697 |
+
"eval_samples_per_second": 257.138,
|
| 698 |
+
"eval_steps_per_second": 8.044,
|
| 699 |
+
"eval_weighted-f1": 0.6946838990403312,
|
| 700 |
+
"eval_weighted-precision": 0.6981818094755908,
|
| 701 |
+
"eval_weighted-recall": 0.6957928802588996,
|
| 702 |
+
"step": 9048
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 40.0,
|
| 706 |
+
"eval_accuracy": 0.697950377562028,
|
| 707 |
+
"eval_loss": 2.123837471008301,
|
| 708 |
+
"eval_macro-f1": 0.672888533432289,
|
| 709 |
+
"eval_macro-precision": 0.6846457775029204,
|
| 710 |
+
"eval_macro-recall": 0.6663735016267468,
|
| 711 |
+
"eval_runtime": 3.6082,
|
| 712 |
+
"eval_samples_per_second": 256.916,
|
| 713 |
+
"eval_steps_per_second": 8.037,
|
| 714 |
+
"eval_weighted-f1": 0.6972181365180552,
|
| 715 |
+
"eval_weighted-precision": 0.700475447211827,
|
| 716 |
+
"eval_weighted-recall": 0.697950377562028,
|
| 717 |
+
"step": 9280
|
| 718 |
+
}
|
| 719 |
+
],
|
| 720 |
+
"logging_steps": 500,
|
| 721 |
+
"max_steps": 9280,
|
| 722 |
+
"num_input_tokens_seen": 0,
|
| 723 |
+
"num_train_epochs": 40,
|
| 724 |
+
"save_steps": 500,
|
| 725 |
+
"total_flos": 3.933393471111168e+16,
|
| 726 |
+
"train_batch_size": 32,
|
| 727 |
+
"trial_name": null,
|
| 728 |
+
"trial_params": null
|
| 729 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f37cdc4eef95435e3ae55be109eba3f77575804bcb2e8ed93f021a45cad3f3a
|
| 3 |
+
size 4335
|