upload nace2_level0 model
Browse files- config.json +84 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- trainer_state.json +816 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "../models/roberta-large/",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"finetuning_task": "nace",
|
| 10 |
+
"gradient_checkpointing": false,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 1024,
|
| 14 |
+
"id2label": {
|
| 15 |
+
"0": "10",
|
| 16 |
+
"1": "11",
|
| 17 |
+
"2": "12",
|
| 18 |
+
"3": "13",
|
| 19 |
+
"4": "14",
|
| 20 |
+
"5": "15",
|
| 21 |
+
"6": "16",
|
| 22 |
+
"7": "17",
|
| 23 |
+
"8": "18",
|
| 24 |
+
"9": "19",
|
| 25 |
+
"10": "20",
|
| 26 |
+
"11": "21",
|
| 27 |
+
"12": "22",
|
| 28 |
+
"13": "23",
|
| 29 |
+
"14": "24",
|
| 30 |
+
"15": "25",
|
| 31 |
+
"16": "26",
|
| 32 |
+
"17": "27",
|
| 33 |
+
"18": "28",
|
| 34 |
+
"19": "29",
|
| 35 |
+
"20": "30",
|
| 36 |
+
"21": "31",
|
| 37 |
+
"22": "32",
|
| 38 |
+
"23": "42",
|
| 39 |
+
"24": "43",
|
| 40 |
+
"25": "62"
|
| 41 |
+
},
|
| 42 |
+
"initializer_range": 0.02,
|
| 43 |
+
"intermediate_size": 4096,
|
| 44 |
+
"label2id": {
|
| 45 |
+
"10": 0,
|
| 46 |
+
"11": 1,
|
| 47 |
+
"12": 2,
|
| 48 |
+
"13": 3,
|
| 49 |
+
"14": 4,
|
| 50 |
+
"15": 5,
|
| 51 |
+
"16": 6,
|
| 52 |
+
"17": 7,
|
| 53 |
+
"18": 8,
|
| 54 |
+
"19": 9,
|
| 55 |
+
"20": 10,
|
| 56 |
+
"21": 11,
|
| 57 |
+
"22": 12,
|
| 58 |
+
"23": 13,
|
| 59 |
+
"24": 14,
|
| 60 |
+
"25": 15,
|
| 61 |
+
"26": 16,
|
| 62 |
+
"27": 17,
|
| 63 |
+
"28": 18,
|
| 64 |
+
"29": 19,
|
| 65 |
+
"30": 20,
|
| 66 |
+
"31": 21,
|
| 67 |
+
"32": 22,
|
| 68 |
+
"42": 23,
|
| 69 |
+
"43": 24,
|
| 70 |
+
"62": 25
|
| 71 |
+
},
|
| 72 |
+
"layer_norm_eps": 1e-05,
|
| 73 |
+
"max_position_embeddings": 514,
|
| 74 |
+
"model_type": "roberta",
|
| 75 |
+
"num_attention_heads": 16,
|
| 76 |
+
"num_hidden_layers": 24,
|
| 77 |
+
"pad_token_id": 1,
|
| 78 |
+
"position_embedding_type": "absolute",
|
| 79 |
+
"problem_type": "multi_label_classification",
|
| 80 |
+
"transformers_version": "4.6.1",
|
| 81 |
+
"type_vocab_size": 1,
|
| 82 |
+
"use_cache": true,
|
| 83 |
+
"vocab_size": 50265
|
| 84 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8576ce97fdbc776150b9c4293fa2ffa0e2f1a2ce621b67ae8ec85a55224239a5
|
| 3 |
+
size 1421709677
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02799ae29d8e97c746469176f763b32351e4563ab92507ce57bd575981b0b31a
|
| 3 |
+
size 15523
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "special_tokens_map_file": null, "name_or_path": "../models/roberta-large/"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7953262197881779,
|
| 3 |
+
"best_model_checkpoint": "./output//roberta-large_nace_5__5e-6_0.01_0.06_07-12-22_21-12/checkpoint-98000",
|
| 4 |
+
"epoch": 0.08573862323490967,
|
| 5 |
+
"global_step": 100000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 2.8579512490675938e-08,
|
| 13 |
+
"loss": 0.6423,
|
| 14 |
+
"step": 2000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"eval_accuracy": 0.00020930884194819268,
|
| 19 |
+
"eval_f1": 0.017752777264977686,
|
| 20 |
+
"eval_loss": 0.45688706636428833,
|
| 21 |
+
"eval_roc_auc": 0.4999505258927197,
|
| 22 |
+
"eval_runtime": 8888.0353,
|
| 23 |
+
"eval_samples_per_second": 33.327,
|
| 24 |
+
"step": 2000
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.0,
|
| 28 |
+
"learning_rate": 5.7159024981351875e-08,
|
| 29 |
+
"loss": 0.3833,
|
| 30 |
+
"step": 4000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0,
|
| 34 |
+
"eval_accuracy": 0.0,
|
| 35 |
+
"eval_f1": 0.0074795009434814345,
|
| 36 |
+
"eval_loss": 0.2739907205104828,
|
| 37 |
+
"eval_roc_auc": 0.4998445851936328,
|
| 38 |
+
"eval_runtime": 8865.9827,
|
| 39 |
+
"eval_samples_per_second": 33.41,
|
| 40 |
+
"step": 4000
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.01,
|
| 44 |
+
"learning_rate": 8.573853747202781e-08,
|
| 45 |
+
"loss": 0.2538,
|
| 46 |
+
"step": 6000
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.01,
|
| 50 |
+
"eval_accuracy": 0.00013166201348354057,
|
| 51 |
+
"eval_f1": 0.0004511461766688428,
|
| 52 |
+
"eval_loss": 0.2050675004720688,
|
| 53 |
+
"eval_roc_auc": 0.5000627296722496,
|
| 54 |
+
"eval_runtime": 8869.0926,
|
| 55 |
+
"eval_samples_per_second": 33.398,
|
| 56 |
+
"step": 6000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.01,
|
| 60 |
+
"learning_rate": 1.1431804996270375e-07,
|
| 61 |
+
"loss": 0.1998,
|
| 62 |
+
"step": 8000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.01,
|
| 66 |
+
"eval_accuracy": 3.3759490636805274e-06,
|
| 67 |
+
"eval_f1": 5.318554842278257e-06,
|
| 68 |
+
"eval_loss": 0.17527928948402405,
|
| 69 |
+
"eval_roc_auc": 0.4999982583562919,
|
| 70 |
+
"eval_runtime": 8867.7398,
|
| 71 |
+
"eval_samples_per_second": 33.403,
|
| 72 |
+
"step": 8000
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.01,
|
| 76 |
+
"learning_rate": 1.428975624533797e-07,
|
| 77 |
+
"loss": 0.1693,
|
| 78 |
+
"step": 10000
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.01,
|
| 82 |
+
"eval_accuracy": 0.18926583235712138,
|
| 83 |
+
"eval_f1": 0.310704534917664,
|
| 84 |
+
"eval_loss": 0.15055961906909943,
|
| 85 |
+
"eval_roc_auc": 0.5950524118809889,
|
| 86 |
+
"eval_runtime": 8868.9495,
|
| 87 |
+
"eval_samples_per_second": 33.399,
|
| 88 |
+
"step": 10000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.01,
|
| 92 |
+
"learning_rate": 1.7147707494405562e-07,
|
| 93 |
+
"loss": 0.1457,
|
| 94 |
+
"step": 12000
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.01,
|
| 98 |
+
"eval_accuracy": 0.3331555333493128,
|
| 99 |
+
"eval_f1": 0.4681315118558145,
|
| 100 |
+
"eval_loss": 0.13044112920761108,
|
| 101 |
+
"eval_roc_auc": 0.6662536331303048,
|
| 102 |
+
"eval_runtime": 8874.2821,
|
| 103 |
+
"eval_samples_per_second": 33.379,
|
| 104 |
+
"step": 12000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.01,
|
| 108 |
+
"learning_rate": 2.0005658743473155e-07,
|
| 109 |
+
"loss": 0.127,
|
| 110 |
+
"step": 14000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.01,
|
| 114 |
+
"eval_accuracy": 0.41038374413006856,
|
| 115 |
+
"eval_f1": 0.5463731063419305,
|
| 116 |
+
"eval_loss": 0.11502532660961151,
|
| 117 |
+
"eval_roc_auc": 0.7051608184740583,
|
| 118 |
+
"eval_runtime": 8870.8129,
|
| 119 |
+
"eval_samples_per_second": 33.392,
|
| 120 |
+
"step": 14000
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.01,
|
| 124 |
+
"learning_rate": 2.286360999254075e-07,
|
| 125 |
+
"loss": 0.1132,
|
| 126 |
+
"step": 16000
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.01,
|
| 130 |
+
"eval_accuracy": 0.4763227812418766,
|
| 131 |
+
"eval_f1": 0.6070642556360302,
|
| 132 |
+
"eval_loss": 0.10388284176588058,
|
| 133 |
+
"eval_roc_auc": 0.7403745042255975,
|
| 134 |
+
"eval_runtime": 8872.294,
|
| 135 |
+
"eval_samples_per_second": 33.386,
|
| 136 |
+
"step": 16000
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.02,
|
| 140 |
+
"learning_rate": 2.5721561241608343e-07,
|
| 141 |
+
"loss": 0.1032,
|
| 142 |
+
"step": 18000
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.02,
|
| 146 |
+
"eval_accuracy": 0.5087960352854196,
|
| 147 |
+
"eval_f1": 0.6375540199236641,
|
| 148 |
+
"eval_loss": 0.09574878960847855,
|
| 149 |
+
"eval_roc_auc": 0.7589272821999232,
|
| 150 |
+
"eval_runtime": 8879.9353,
|
| 151 |
+
"eval_samples_per_second": 33.358,
|
| 152 |
+
"step": 18000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.02,
|
| 156 |
+
"learning_rate": 2.857951249067594e-07,
|
| 157 |
+
"loss": 0.0959,
|
| 158 |
+
"step": 20000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.02,
|
| 162 |
+
"eval_accuracy": 0.5288154132330452,
|
| 163 |
+
"eval_f1": 0.6573253473962186,
|
| 164 |
+
"eval_loss": 0.08954177796840668,
|
| 165 |
+
"eval_roc_auc": 0.7715952535178425,
|
| 166 |
+
"eval_runtime": 8893.417,
|
| 167 |
+
"eval_samples_per_second": 33.307,
|
| 168 |
+
"step": 20000
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 0.02,
|
| 172 |
+
"learning_rate": 3.1437463739743534e-07,
|
| 173 |
+
"loss": 0.0899,
|
| 174 |
+
"step": 22000
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.02,
|
| 178 |
+
"eval_accuracy": 0.5462656939432098,
|
| 179 |
+
"eval_f1": 0.6785063497828563,
|
| 180 |
+
"eval_loss": 0.08448488265275955,
|
| 181 |
+
"eval_roc_auc": 0.7872646974276402,
|
| 182 |
+
"eval_runtime": 8886.3548,
|
| 183 |
+
"eval_samples_per_second": 33.333,
|
| 184 |
+
"step": 22000
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.02,
|
| 188 |
+
"learning_rate": 3.4295414988811124e-07,
|
| 189 |
+
"loss": 0.085,
|
| 190 |
+
"step": 24000
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 0.02,
|
| 194 |
+
"eval_accuracy": 0.5589457586263938,
|
| 195 |
+
"eval_f1": 0.6890862787276392,
|
| 196 |
+
"eval_loss": 0.08046474307775497,
|
| 197 |
+
"eval_roc_auc": 0.793202863995178,
|
| 198 |
+
"eval_runtime": 8888.5138,
|
| 199 |
+
"eval_samples_per_second": 33.325,
|
| 200 |
+
"step": 24000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.02,
|
| 204 |
+
"learning_rate": 3.715336623787872e-07,
|
| 205 |
+
"loss": 0.0807,
|
| 206 |
+
"step": 26000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.02,
|
| 210 |
+
"eval_accuracy": 0.5779354721095968,
|
| 211 |
+
"eval_f1": 0.7085340529054537,
|
| 212 |
+
"eval_loss": 0.07694748789072037,
|
| 213 |
+
"eval_roc_auc": 0.8086207353956607,
|
| 214 |
+
"eval_runtime": 9108.0018,
|
| 215 |
+
"eval_samples_per_second": 32.522,
|
| 216 |
+
"step": 26000
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 0.02,
|
| 220 |
+
"learning_rate": 4.001131748694631e-07,
|
| 221 |
+
"loss": 0.0778,
|
| 222 |
+
"step": 28000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.02,
|
| 226 |
+
"eval_accuracy": 0.5849236866714155,
|
| 227 |
+
"eval_f1": 0.7155799872243565,
|
| 228 |
+
"eval_loss": 0.07436001300811768,
|
| 229 |
+
"eval_roc_auc": 0.8137322286922238,
|
| 230 |
+
"eval_runtime": 8902.3443,
|
| 231 |
+
"eval_samples_per_second": 33.274,
|
| 232 |
+
"step": 28000
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 0.03,
|
| 236 |
+
"learning_rate": 4.28692687360139e-07,
|
| 237 |
+
"loss": 0.0753,
|
| 238 |
+
"step": 30000
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 0.03,
|
| 242 |
+
"eval_accuracy": 0.5928402872257463,
|
| 243 |
+
"eval_f1": 0.7241294650780317,
|
| 244 |
+
"eval_loss": 0.07221980392932892,
|
| 245 |
+
"eval_roc_auc": 0.8217651733984279,
|
| 246 |
+
"eval_runtime": 8918.1851,
|
| 247 |
+
"eval_samples_per_second": 33.214,
|
| 248 |
+
"step": 30000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.03,
|
| 252 |
+
"learning_rate": 4.57272199850815e-07,
|
| 253 |
+
"loss": 0.0725,
|
| 254 |
+
"step": 32000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.03,
|
| 258 |
+
"eval_accuracy": 0.5951764439778132,
|
| 259 |
+
"eval_f1": 0.7261386622802142,
|
| 260 |
+
"eval_loss": 0.07057042419910431,
|
| 261 |
+
"eval_roc_auc": 0.822130038839097,
|
| 262 |
+
"eval_runtime": 8995.4049,
|
| 263 |
+
"eval_samples_per_second": 32.929,
|
| 264 |
+
"step": 32000
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.03,
|
| 268 |
+
"learning_rate": 4.858517123414909e-07,
|
| 269 |
+
"loss": 0.0711,
|
| 270 |
+
"step": 34000
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.03,
|
| 274 |
+
"eval_accuracy": 0.6020971395583583,
|
| 275 |
+
"eval_f1": 0.7323235836003227,
|
| 276 |
+
"eval_loss": 0.06848303228616714,
|
| 277 |
+
"eval_roc_auc": 0.8254645532099296,
|
| 278 |
+
"eval_runtime": 9451.4989,
|
| 279 |
+
"eval_samples_per_second": 31.34,
|
| 280 |
+
"step": 34000
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.03,
|
| 284 |
+
"learning_rate": 5.144312248321669e-07,
|
| 285 |
+
"loss": 0.0696,
|
| 286 |
+
"step": 36000
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 0.03,
|
| 290 |
+
"eval_accuracy": 0.6094972199059461,
|
| 291 |
+
"eval_f1": 0.7393677046231971,
|
| 292 |
+
"eval_loss": 0.06683486700057983,
|
| 293 |
+
"eval_roc_auc": 0.8309554684971688,
|
| 294 |
+
"eval_runtime": 10018.5624,
|
| 295 |
+
"eval_samples_per_second": 29.566,
|
| 296 |
+
"step": 36000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.03,
|
| 300 |
+
"learning_rate": 5.430107373228428e-07,
|
| 301 |
+
"loss": 0.0676,
|
| 302 |
+
"step": 38000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.03,
|
| 306 |
+
"eval_accuracy": 0.6112223298774868,
|
| 307 |
+
"eval_f1": 0.7425602873268343,
|
| 308 |
+
"eval_loss": 0.06597442924976349,
|
| 309 |
+
"eval_roc_auc": 0.8349447666298531,
|
| 310 |
+
"eval_runtime": 11692.7897,
|
| 311 |
+
"eval_samples_per_second": 25.333,
|
| 312 |
+
"step": 38000
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.03,
|
| 316 |
+
"learning_rate": 5.715902498135188e-07,
|
| 317 |
+
"loss": 0.0663,
|
| 318 |
+
"step": 40000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.03,
|
| 322 |
+
"eval_accuracy": 0.6157629813681371,
|
| 323 |
+
"eval_f1": 0.7460401136089835,
|
| 324 |
+
"eval_loss": 0.06471637636423111,
|
| 325 |
+
"eval_roc_auc": 0.8360501874414529,
|
| 326 |
+
"eval_runtime": 8963.8102,
|
| 327 |
+
"eval_samples_per_second": 33.045,
|
| 328 |
+
"step": 40000
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 0.04,
|
| 332 |
+
"learning_rate": 6.001697623041946e-07,
|
| 333 |
+
"loss": 0.0647,
|
| 334 |
+
"step": 42000
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 0.04,
|
| 338 |
+
"eval_accuracy": 0.6194090063569121,
|
| 339 |
+
"eval_f1": 0.7492256354462544,
|
| 340 |
+
"eval_loss": 0.06362345069646835,
|
| 341 |
+
"eval_roc_auc": 0.8375523158859597,
|
| 342 |
+
"eval_runtime": 8891.9369,
|
| 343 |
+
"eval_samples_per_second": 33.313,
|
| 344 |
+
"step": 42000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.04,
|
| 348 |
+
"learning_rate": 6.287492747948707e-07,
|
| 349 |
+
"loss": 0.0642,
|
| 350 |
+
"step": 44000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.04,
|
| 354 |
+
"eval_accuracy": 0.6211577479718986,
|
| 355 |
+
"eval_f1": 0.7510341371183263,
|
| 356 |
+
"eval_loss": 0.06275659799575806,
|
| 357 |
+
"eval_roc_auc": 0.8382498982808346,
|
| 358 |
+
"eval_runtime": 8889.7383,
|
| 359 |
+
"eval_samples_per_second": 33.321,
|
| 360 |
+
"step": 44000
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.04,
|
| 364 |
+
"learning_rate": 6.573287872855465e-07,
|
| 365 |
+
"loss": 0.0633,
|
| 366 |
+
"step": 46000
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.04,
|
| 370 |
+
"eval_accuracy": 0.6240340565741544,
|
| 371 |
+
"eval_f1": 0.7552442670824074,
|
| 372 |
+
"eval_loss": 0.061649855226278305,
|
| 373 |
+
"eval_roc_auc": 0.8417446851046798,
|
| 374 |
+
"eval_runtime": 9047.8013,
|
| 375 |
+
"eval_samples_per_second": 32.739,
|
| 376 |
+
"step": 46000
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 0.04,
|
| 380 |
+
"learning_rate": 6.859082997762225e-07,
|
| 381 |
+
"loss": 0.0623,
|
| 382 |
+
"step": 48000
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 0.04,
|
| 386 |
+
"eval_accuracy": 0.6283518954266019,
|
| 387 |
+
"eval_f1": 0.7614445531010631,
|
| 388 |
+
"eval_loss": 0.06095787137746811,
|
| 389 |
+
"eval_roc_auc": 0.8492713317355323,
|
| 390 |
+
"eval_runtime": 8895.4655,
|
| 391 |
+
"eval_samples_per_second": 33.299,
|
| 392 |
+
"step": 48000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.04,
|
| 396 |
+
"learning_rate": 7.144878122668984e-07,
|
| 397 |
+
"loss": 0.0617,
|
| 398 |
+
"step": 50000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.04,
|
| 402 |
+
"eval_accuracy": 0.6315151596992704,
|
| 403 |
+
"eval_f1": 0.7623575299697047,
|
| 404 |
+
"eval_loss": 0.059979602694511414,
|
| 405 |
+
"eval_roc_auc": 0.8468318365424115,
|
| 406 |
+
"eval_runtime": 8905.8909,
|
| 407 |
+
"eval_samples_per_second": 33.26,
|
| 408 |
+
"step": 50000
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.04,
|
| 412 |
+
"learning_rate": 7.430673247575744e-07,
|
| 413 |
+
"loss": 0.0603,
|
| 414 |
+
"step": 52000
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.04,
|
| 418 |
+
"eval_accuracy": 0.6323929064558274,
|
| 419 |
+
"eval_f1": 0.7646658582301569,
|
| 420 |
+
"eval_loss": 0.0595267117023468,
|
| 421 |
+
"eval_roc_auc": 0.8497986956155574,
|
| 422 |
+
"eval_runtime": 8916.2146,
|
| 423 |
+
"eval_samples_per_second": 33.222,
|
| 424 |
+
"step": 52000
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 0.05,
|
| 428 |
+
"learning_rate": 7.716468372482502e-07,
|
| 429 |
+
"loss": 0.0605,
|
| 430 |
+
"step": 54000
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.05,
|
| 434 |
+
"eval_accuracy": 0.6345197543659461,
|
| 435 |
+
"eval_f1": 0.7675632990904457,
|
| 436 |
+
"eval_loss": 0.05862889438867569,
|
| 437 |
+
"eval_roc_auc": 0.8518276428326589,
|
| 438 |
+
"eval_runtime": 8909.536,
|
| 439 |
+
"eval_samples_per_second": 33.247,
|
| 440 |
+
"step": 54000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.05,
|
| 444 |
+
"learning_rate": 8.002263497389262e-07,
|
| 445 |
+
"loss": 0.0594,
|
| 446 |
+
"step": 56000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.05,
|
| 450 |
+
"eval_accuracy": 0.6367647604932937,
|
| 451 |
+
"eval_f1": 0.7709907906594025,
|
| 452 |
+
"eval_loss": 0.05837056040763855,
|
| 453 |
+
"eval_roc_auc": 0.8562941632344473,
|
| 454 |
+
"eval_runtime": 8903.6442,
|
| 455 |
+
"eval_samples_per_second": 33.269,
|
| 456 |
+
"step": 56000
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 0.05,
|
| 460 |
+
"learning_rate": 8.288058622296021e-07,
|
| 461 |
+
"loss": 0.0592,
|
| 462 |
+
"step": 58000
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 0.05,
|
| 466 |
+
"eval_accuracy": 0.6382434261831857,
|
| 467 |
+
"eval_f1": 0.7715028842192007,
|
| 468 |
+
"eval_loss": 0.05767415836453438,
|
| 469 |
+
"eval_roc_auc": 0.8549143755897282,
|
| 470 |
+
"eval_runtime": 8915.2586,
|
| 471 |
+
"eval_samples_per_second": 33.225,
|
| 472 |
+
"step": 58000
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.05,
|
| 476 |
+
"learning_rate": 8.57385374720278e-07,
|
| 477 |
+
"loss": 0.0585,
|
| 478 |
+
"step": 60000
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.05,
|
| 482 |
+
"eval_accuracy": 0.6385371337517259,
|
| 483 |
+
"eval_f1": 0.7683973772822023,
|
| 484 |
+
"eval_loss": 0.057559411972761154,
|
| 485 |
+
"eval_roc_auc": 0.8502475677045068,
|
| 486 |
+
"eval_runtime": 8911.1959,
|
| 487 |
+
"eval_samples_per_second": 33.241,
|
| 488 |
+
"step": 60000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.05,
|
| 492 |
+
"learning_rate": 8.85964887210954e-07,
|
| 493 |
+
"loss": 0.0578,
|
| 494 |
+
"step": 62000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.05,
|
| 498 |
+
"eval_accuracy": 0.643915020610169,
|
| 499 |
+
"eval_f1": 0.7735556679913914,
|
| 500 |
+
"eval_loss": 0.056609489023685455,
|
| 501 |
+
"eval_roc_auc": 0.8534382569874794,
|
| 502 |
+
"eval_runtime": 8924.5002,
|
| 503 |
+
"eval_samples_per_second": 33.191,
|
| 504 |
+
"step": 62000
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.05,
|
| 508 |
+
"learning_rate": 9.1454439970163e-07,
|
| 509 |
+
"loss": 0.0572,
|
| 510 |
+
"step": 64000
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 0.05,
|
| 514 |
+
"eval_accuracy": 0.6444484205622305,
|
| 515 |
+
"eval_f1": 0.7776911375576832,
|
| 516 |
+
"eval_loss": 0.05652829632163048,
|
| 517 |
+
"eval_roc_auc": 0.8603552529596437,
|
| 518 |
+
"eval_runtime": 8931.0301,
|
| 519 |
+
"eval_samples_per_second": 33.167,
|
| 520 |
+
"step": 64000
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.06,
|
| 524 |
+
"learning_rate": 9.43123912192306e-07,
|
| 525 |
+
"loss": 0.0571,
|
| 526 |
+
"step": 66000
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 0.06,
|
| 530 |
+
"eval_accuracy": 0.6454882128738442,
|
| 531 |
+
"eval_f1": 0.7742083504686046,
|
| 532 |
+
"eval_loss": 0.05627002194523811,
|
| 533 |
+
"eval_roc_auc": 0.8534170877970823,
|
| 534 |
+
"eval_runtime": 8935.7869,
|
| 535 |
+
"eval_samples_per_second": 33.149,
|
| 536 |
+
"step": 66000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.06,
|
| 540 |
+
"learning_rate": 9.717034246829818e-07,
|
| 541 |
+
"loss": 0.0563,
|
| 542 |
+
"step": 68000
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.06,
|
| 546 |
+
"eval_accuracy": 0.6462410495150449,
|
| 547 |
+
"eval_f1": 0.7776261820301417,
|
| 548 |
+
"eval_loss": 0.05584974214434624,
|
| 549 |
+
"eval_roc_auc": 0.8583744396413601,
|
| 550 |
+
"eval_runtime": 8945.9116,
|
| 551 |
+
"eval_samples_per_second": 33.112,
|
| 552 |
+
"step": 68000
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 0.06,
|
| 556 |
+
"learning_rate": 1.0002829371736579e-06,
|
| 557 |
+
"loss": 0.0562,
|
| 558 |
+
"step": 70000
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 0.06,
|
| 562 |
+
"eval_accuracy": 0.6459709735899505,
|
| 563 |
+
"eval_f1": 0.7794595088938681,
|
| 564 |
+
"eval_loss": 0.05508234724402428,
|
| 565 |
+
"eval_roc_auc": 0.8591985509814772,
|
| 566 |
+
"eval_runtime": 8944.2225,
|
| 567 |
+
"eval_samples_per_second": 33.118,
|
| 568 |
+
"step": 70000
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 0.06,
|
| 572 |
+
"learning_rate": 1.0288624496643337e-06,
|
| 573 |
+
"loss": 0.0554,
|
| 574 |
+
"step": 72000
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 0.06,
|
| 578 |
+
"eval_accuracy": 0.6490059517981993,
|
| 579 |
+
"eval_f1": 0.7775233968578572,
|
| 580 |
+
"eval_loss": 0.05504510924220085,
|
| 581 |
+
"eval_roc_auc": 0.8542740144934026,
|
| 582 |
+
"eval_runtime": 8953.1995,
|
| 583 |
+
"eval_samples_per_second": 33.085,
|
| 584 |
+
"step": 72000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.06,
|
| 588 |
+
"learning_rate": 1.0574419621550096e-06,
|
| 589 |
+
"loss": 0.0552,
|
| 590 |
+
"step": 74000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.06,
|
| 594 |
+
"eval_accuracy": 0.6509471225098156,
|
| 595 |
+
"eval_f1": 0.7816832192367912,
|
| 596 |
+
"eval_loss": 0.054713618010282516,
|
| 597 |
+
"eval_roc_auc": 0.86037433954253,
|
| 598 |
+
"eval_runtime": 8945.5727,
|
| 599 |
+
"eval_samples_per_second": 33.113,
|
| 600 |
+
"step": 74000
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.07,
|
| 604 |
+
"learning_rate": 1.0860214746456856e-06,
|
| 605 |
+
"loss": 0.0547,
|
| 606 |
+
"step": 76000
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.07,
|
| 610 |
+
"eval_accuracy": 0.6510315212364076,
|
| 611 |
+
"eval_f1": 0.783673992804474,
|
| 612 |
+
"eval_loss": 0.05407635122537613,
|
| 613 |
+
"eval_roc_auc": 0.8622274781842953,
|
| 614 |
+
"eval_runtime": 8946.7264,
|
| 615 |
+
"eval_samples_per_second": 33.109,
|
| 616 |
+
"step": 76000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.07,
|
| 620 |
+
"learning_rate": 1.1146009871363615e-06,
|
| 621 |
+
"loss": 0.0547,
|
| 622 |
+
"step": 78000
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 0.07,
|
| 626 |
+
"eval_accuracy": 0.6545526361098264,
|
| 627 |
+
"eval_f1": 0.7855580273587336,
|
| 628 |
+
"eval_loss": 0.053795717656612396,
|
| 629 |
+
"eval_roc_auc": 0.863520124962409,
|
| 630 |
+
"eval_runtime": 8947.2041,
|
| 631 |
+
"eval_samples_per_second": 33.107,
|
| 632 |
+
"step": 78000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.07,
|
| 636 |
+
"learning_rate": 1.1431804996270375e-06,
|
| 637 |
+
"loss": 0.0539,
|
| 638 |
+
"step": 80000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.07,
|
| 642 |
+
"eval_accuracy": 0.6558962638371713,
|
| 643 |
+
"eval_f1": 0.7874432731515895,
|
| 644 |
+
"eval_loss": 0.053422119468450546,
|
| 645 |
+
"eval_roc_auc": 0.8649282174633808,
|
| 646 |
+
"eval_runtime": 8946.6771,
|
| 647 |
+
"eval_samples_per_second": 33.109,
|
| 648 |
+
"step": 80000
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.07,
|
| 652 |
+
"learning_rate": 1.1717600121177134e-06,
|
| 653 |
+
"loss": 0.0541,
|
| 654 |
+
"step": 82000
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.07,
|
| 658 |
+
"eval_accuracy": 0.654451357637916,
|
| 659 |
+
"eval_f1": 0.7878992158367831,
|
| 660 |
+
"eval_loss": 0.05306807905435562,
|
| 661 |
+
"eval_roc_auc": 0.8663645266466831,
|
| 662 |
+
"eval_runtime": 8946.194,
|
| 663 |
+
"eval_samples_per_second": 33.111,
|
| 664 |
+
"step": 82000
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 0.07,
|
| 668 |
+
"learning_rate": 1.2003395246083892e-06,
|
| 669 |
+
"loss": 0.0534,
|
| 670 |
+
"step": 84000
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 0.07,
|
| 674 |
+
"eval_accuracy": 0.6543129437263051,
|
| 675 |
+
"eval_f1": 0.7862359590552527,
|
| 676 |
+
"eval_loss": 0.05337332561612129,
|
| 677 |
+
"eval_roc_auc": 0.8637659543017016,
|
| 678 |
+
"eval_runtime": 8946.5504,
|
| 679 |
+
"eval_samples_per_second": 33.109,
|
| 680 |
+
"step": 84000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.07,
|
| 684 |
+
"learning_rate": 1.2289190370990653e-06,
|
| 685 |
+
"loss": 0.0533,
|
| 686 |
+
"step": 86000
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.07,
|
| 690 |
+
"eval_accuracy": 0.6579623446641437,
|
| 691 |
+
"eval_f1": 0.789641153558776,
|
| 692 |
+
"eval_loss": 0.05264722555875778,
|
| 693 |
+
"eval_roc_auc": 0.8661964026287946,
|
| 694 |
+
"eval_runtime": 8946.6953,
|
| 695 |
+
"eval_samples_per_second": 33.109,
|
| 696 |
+
"step": 86000
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.08,
|
| 700 |
+
"learning_rate": 1.2574985495897414e-06,
|
| 701 |
+
"loss": 0.0536,
|
| 702 |
+
"step": 88000
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.08,
|
| 706 |
+
"eval_accuracy": 0.6578779459375517,
|
| 707 |
+
"eval_f1": 0.7904054501316083,
|
| 708 |
+
"eval_loss": 0.05252569913864136,
|
| 709 |
+
"eval_roc_auc": 0.8665749344562018,
|
| 710 |
+
"eval_runtime": 8946.467,
|
| 711 |
+
"eval_samples_per_second": 33.109,
|
| 712 |
+
"step": 88000
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 0.08,
|
| 716 |
+
"learning_rate": 1.2860780620804172e-06,
|
| 717 |
+
"loss": 0.0529,
|
| 718 |
+
"step": 90000
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 0.08,
|
| 722 |
+
"eval_accuracy": 0.6560279258506547,
|
| 723 |
+
"eval_f1": 0.7903261630505988,
|
| 724 |
+
"eval_loss": 0.052733857184648514,
|
| 725 |
+
"eval_roc_auc": 0.8701088706242622,
|
| 726 |
+
"eval_runtime": 8948.9618,
|
| 727 |
+
"eval_samples_per_second": 33.1,
|
| 728 |
+
"step": 90000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.08,
|
| 732 |
+
"learning_rate": 1.314657574571093e-06,
|
| 733 |
+
"loss": 0.0525,
|
| 734 |
+
"step": 92000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.08,
|
| 738 |
+
"eval_accuracy": 0.6622430480768906,
|
| 739 |
+
"eval_f1": 0.7920309543792765,
|
| 740 |
+
"eval_loss": 0.0518718883395195,
|
| 741 |
+
"eval_roc_auc": 0.8668831436253378,
|
| 742 |
+
"eval_runtime": 8946.4377,
|
| 743 |
+
"eval_samples_per_second": 33.11,
|
| 744 |
+
"step": 92000
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.08,
|
| 748 |
+
"learning_rate": 1.343237087061769e-06,
|
| 749 |
+
"loss": 0.0522,
|
| 750 |
+
"step": 94000
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 0.08,
|
| 754 |
+
"eval_accuracy": 0.6575808624199478,
|
| 755 |
+
"eval_f1": 0.7916439543928127,
|
| 756 |
+
"eval_loss": 0.052168361842632294,
|
| 757 |
+
"eval_roc_auc": 0.8691981685442582,
|
| 758 |
+
"eval_runtime": 8948.4665,
|
| 759 |
+
"eval_samples_per_second": 33.102,
|
| 760 |
+
"step": 94000
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.08,
|
| 764 |
+
"learning_rate": 1.371816599552445e-06,
|
| 765 |
+
"loss": 0.0518,
|
| 766 |
+
"step": 96000
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.08,
|
| 770 |
+
"eval_accuracy": 0.6617164000229565,
|
| 771 |
+
"eval_f1": 0.7952230824994735,
|
| 772 |
+
"eval_loss": 0.051213569939136505,
|
| 773 |
+
"eval_roc_auc": 0.8707353068301484,
|
| 774 |
+
"eval_runtime": 8946.6417,
|
| 775 |
+
"eval_samples_per_second": 33.109,
|
| 776 |
+
"step": 96000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.08,
|
| 780 |
+
"learning_rate": 1.4003961120431208e-06,
|
| 781 |
+
"loss": 0.0515,
|
| 782 |
+
"step": 98000
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.08,
|
| 786 |
+
"eval_accuracy": 0.6638094884424384,
|
| 787 |
+
"eval_f1": 0.7953262197881779,
|
| 788 |
+
"eval_loss": 0.05120311677455902,
|
| 789 |
+
"eval_roc_auc": 0.8704717586188649,
|
| 790 |
+
"eval_runtime": 8948.3566,
|
| 791 |
+
"eval_samples_per_second": 33.103,
|
| 792 |
+
"step": 98000
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 0.09,
|
| 796 |
+
"learning_rate": 1.4289756245337969e-06,
|
| 797 |
+
"loss": 0.0515,
|
| 798 |
+
"step": 100000
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.09,
|
| 802 |
+
"eval_accuracy": 0.6594410103540358,
|
| 803 |
+
"eval_f1": 0.7952893225951836,
|
| 804 |
+
"eval_loss": 0.05145873501896858,
|
| 805 |
+
"eval_roc_auc": 0.8740886532218961,
|
| 806 |
+
"eval_runtime": 8946.2483,
|
| 807 |
+
"eval_samples_per_second": 33.11,
|
| 808 |
+
"step": 100000
|
| 809 |
+
}
|
| 810 |
+
],
|
| 811 |
+
"max_steps": 5831675,
|
| 812 |
+
"num_train_epochs": 5,
|
| 813 |
+
"total_flos": 6.9871808151552e+16,
|
| 814 |
+
"trial_name": null,
|
| 815 |
+
"trial_params": null
|
| 816 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d73543aa5084a4ad0602668eb3ca05008e8ca3a960941c767e96da96ae38fd8e
|
| 3 |
+
size 2479
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|