Upload folder using huggingface_hub
Browse files- config.json +29 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +51 -0
- tokenizer_config.json +56 -0
- trainer_state.json +1177 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/codebert-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"output_past": true,
|
| 21 |
+
"pad_token_id": 1,
|
| 22 |
+
"position_embedding_type": "absolute",
|
| 23 |
+
"problem_type": "single_label_classification",
|
| 24 |
+
"torch_dtype": "float32",
|
| 25 |
+
"transformers_version": "4.35.0",
|
| 26 |
+
"type_vocab_size": 1,
|
| 27 |
+
"use_cache": true,
|
| 28 |
+
"vocab_size": 50265
|
| 29 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:963ca2cad6743e0cc5a515566f3f51f44995fc1bfb970d7d13eb8d5af0f9eb2b
|
| 3 |
+
size 498612824
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ccab0b51be41c1782e43ea6e5110e623ec3ad7c7068f4ba82ac8cc03facdc45a
|
| 3 |
+
size 997345530
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0cd62c0c650c6428068d474cf57bc47558b2fa2df88e4b7c5b24cdf7bfee963c
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a0b550340cbd7146c8b1f6c1e8f028def75fe29465aa33610cf9c8580f36c620
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": true,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": true,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": true,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": true,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "<s>",
|
| 46 |
+
"clean_up_tokenization_spaces": true,
|
| 47 |
+
"cls_token": "<s>",
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"errors": "replace",
|
| 50 |
+
"mask_token": "<mask>",
|
| 51 |
+
"model_max_length": 512,
|
| 52 |
+
"pad_token": "<pad>",
|
| 53 |
+
"sep_token": "</s>",
|
| 54 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 55 |
+
"unk_token": "<unk>"
|
| 56 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1177 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.374613881111145,
|
| 3 |
+
"best_model_checkpoint": "/leonardo_work/IscrC_AGENT/PROFES2025/results/balanced/oss_large/checkpoint-1234",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1851,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.9967585089141006e-05,
|
| 14 |
+
"loss": 0.6786,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 1.993517017828201e-05,
|
| 20 |
+
"loss": 0.6829,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 1.9902755267423015e-05,
|
| 26 |
+
"loss": 0.6582,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.06,
|
| 31 |
+
"learning_rate": 1.9870340356564023e-05,
|
| 32 |
+
"loss": 0.6107,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.08,
|
| 37 |
+
"learning_rate": 1.9837925445705027e-05,
|
| 38 |
+
"loss": 0.6041,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.1,
|
| 43 |
+
"learning_rate": 1.980551053484603e-05,
|
| 44 |
+
"loss": 0.6188,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11,
|
| 49 |
+
"learning_rate": 1.9773095623987036e-05,
|
| 50 |
+
"loss": 0.553,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.13,
|
| 55 |
+
"learning_rate": 1.974068071312804e-05,
|
| 56 |
+
"loss": 0.5662,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.15,
|
| 61 |
+
"learning_rate": 1.9708265802269045e-05,
|
| 62 |
+
"loss": 0.5633,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.16,
|
| 67 |
+
"learning_rate": 1.9675850891410053e-05,
|
| 68 |
+
"loss": 0.591,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.18,
|
| 73 |
+
"learning_rate": 1.9643435980551057e-05,
|
| 74 |
+
"loss": 0.5583,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.19,
|
| 79 |
+
"learning_rate": 1.961102106969206e-05,
|
| 80 |
+
"loss": 0.5978,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.21,
|
| 85 |
+
"learning_rate": 1.9578606158833066e-05,
|
| 86 |
+
"loss": 0.5829,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23,
|
| 91 |
+
"learning_rate": 1.954619124797407e-05,
|
| 92 |
+
"loss": 0.5787,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.24,
|
| 97 |
+
"learning_rate": 1.9513776337115075e-05,
|
| 98 |
+
"loss": 0.6376,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.26,
|
| 103 |
+
"learning_rate": 1.948136142625608e-05,
|
| 104 |
+
"loss": 0.5732,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.28,
|
| 109 |
+
"learning_rate": 1.9448946515397084e-05,
|
| 110 |
+
"loss": 0.5465,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.29,
|
| 115 |
+
"learning_rate": 1.9416531604538088e-05,
|
| 116 |
+
"loss": 0.5513,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.31,
|
| 121 |
+
"learning_rate": 1.9384116693679093e-05,
|
| 122 |
+
"loss": 0.6165,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.32,
|
| 127 |
+
"learning_rate": 1.9351701782820097e-05,
|
| 128 |
+
"loss": 0.6196,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.34,
|
| 133 |
+
"learning_rate": 1.93192868719611e-05,
|
| 134 |
+
"loss": 0.5344,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.36,
|
| 139 |
+
"learning_rate": 1.928687196110211e-05,
|
| 140 |
+
"loss": 0.6042,
|
| 141 |
+
"step": 220
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.37,
|
| 145 |
+
"learning_rate": 1.9254457050243114e-05,
|
| 146 |
+
"loss": 0.5838,
|
| 147 |
+
"step": 230
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.39,
|
| 151 |
+
"learning_rate": 1.9222042139384118e-05,
|
| 152 |
+
"loss": 0.4611,
|
| 153 |
+
"step": 240
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.41,
|
| 157 |
+
"learning_rate": 1.9189627228525122e-05,
|
| 158 |
+
"loss": 0.5259,
|
| 159 |
+
"step": 250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.42,
|
| 163 |
+
"learning_rate": 1.9157212317666127e-05,
|
| 164 |
+
"loss": 0.5705,
|
| 165 |
+
"step": 260
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.44,
|
| 169 |
+
"learning_rate": 1.912479740680713e-05,
|
| 170 |
+
"loss": 0.5051,
|
| 171 |
+
"step": 270
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.45,
|
| 175 |
+
"learning_rate": 1.909238249594814e-05,
|
| 176 |
+
"loss": 0.5085,
|
| 177 |
+
"step": 280
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.47,
|
| 181 |
+
"learning_rate": 1.9059967585089144e-05,
|
| 182 |
+
"loss": 0.5847,
|
| 183 |
+
"step": 290
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.49,
|
| 187 |
+
"learning_rate": 1.9027552674230148e-05,
|
| 188 |
+
"loss": 0.5183,
|
| 189 |
+
"step": 300
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.5,
|
| 193 |
+
"learning_rate": 1.8995137763371152e-05,
|
| 194 |
+
"loss": 0.4574,
|
| 195 |
+
"step": 310
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.52,
|
| 199 |
+
"learning_rate": 1.8962722852512157e-05,
|
| 200 |
+
"loss": 0.5197,
|
| 201 |
+
"step": 320
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.53,
|
| 205 |
+
"learning_rate": 1.893030794165316e-05,
|
| 206 |
+
"loss": 0.5033,
|
| 207 |
+
"step": 330
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.55,
|
| 211 |
+
"learning_rate": 1.889789303079417e-05,
|
| 212 |
+
"loss": 0.5166,
|
| 213 |
+
"step": 340
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.57,
|
| 217 |
+
"learning_rate": 1.8865478119935173e-05,
|
| 218 |
+
"loss": 0.5017,
|
| 219 |
+
"step": 350
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.58,
|
| 223 |
+
"learning_rate": 1.8833063209076178e-05,
|
| 224 |
+
"loss": 0.501,
|
| 225 |
+
"step": 360
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.6,
|
| 229 |
+
"learning_rate": 1.8800648298217182e-05,
|
| 230 |
+
"loss": 0.491,
|
| 231 |
+
"step": 370
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.62,
|
| 235 |
+
"learning_rate": 1.8768233387358187e-05,
|
| 236 |
+
"loss": 0.463,
|
| 237 |
+
"step": 380
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.63,
|
| 241 |
+
"learning_rate": 1.873581847649919e-05,
|
| 242 |
+
"loss": 0.6343,
|
| 243 |
+
"step": 390
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.65,
|
| 247 |
+
"learning_rate": 1.8703403565640196e-05,
|
| 248 |
+
"loss": 0.4775,
|
| 249 |
+
"step": 400
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.66,
|
| 253 |
+
"learning_rate": 1.86709886547812e-05,
|
| 254 |
+
"loss": 0.5991,
|
| 255 |
+
"step": 410
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.68,
|
| 259 |
+
"learning_rate": 1.8638573743922204e-05,
|
| 260 |
+
"loss": 0.5008,
|
| 261 |
+
"step": 420
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7,
|
| 265 |
+
"learning_rate": 1.860615883306321e-05,
|
| 266 |
+
"loss": 0.4652,
|
| 267 |
+
"step": 430
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.71,
|
| 271 |
+
"learning_rate": 1.8573743922204213e-05,
|
| 272 |
+
"loss": 0.434,
|
| 273 |
+
"step": 440
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.73,
|
| 277 |
+
"learning_rate": 1.8541329011345218e-05,
|
| 278 |
+
"loss": 0.5431,
|
| 279 |
+
"step": 450
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.75,
|
| 283 |
+
"learning_rate": 1.8508914100486226e-05,
|
| 284 |
+
"loss": 0.5627,
|
| 285 |
+
"step": 460
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.76,
|
| 289 |
+
"learning_rate": 1.847649918962723e-05,
|
| 290 |
+
"loss": 0.5589,
|
| 291 |
+
"step": 470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.78,
|
| 295 |
+
"learning_rate": 1.8444084278768234e-05,
|
| 296 |
+
"loss": 0.4425,
|
| 297 |
+
"step": 480
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.79,
|
| 301 |
+
"learning_rate": 1.841166936790924e-05,
|
| 302 |
+
"loss": 0.5187,
|
| 303 |
+
"step": 490
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.81,
|
| 307 |
+
"learning_rate": 1.8379254457050243e-05,
|
| 308 |
+
"loss": 0.4464,
|
| 309 |
+
"step": 500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.83,
|
| 313 |
+
"learning_rate": 1.8346839546191248e-05,
|
| 314 |
+
"loss": 0.4694,
|
| 315 |
+
"step": 510
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.84,
|
| 319 |
+
"learning_rate": 1.8314424635332255e-05,
|
| 320 |
+
"loss": 0.4667,
|
| 321 |
+
"step": 520
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.86,
|
| 325 |
+
"learning_rate": 1.828200972447326e-05,
|
| 326 |
+
"loss": 0.4848,
|
| 327 |
+
"step": 530
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.88,
|
| 331 |
+
"learning_rate": 1.8249594813614264e-05,
|
| 332 |
+
"loss": 0.3969,
|
| 333 |
+
"step": 540
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.89,
|
| 337 |
+
"learning_rate": 1.821717990275527e-05,
|
| 338 |
+
"loss": 0.4058,
|
| 339 |
+
"step": 550
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.91,
|
| 343 |
+
"learning_rate": 1.8184764991896273e-05,
|
| 344 |
+
"loss": 0.4945,
|
| 345 |
+
"step": 560
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.92,
|
| 349 |
+
"learning_rate": 1.8152350081037278e-05,
|
| 350 |
+
"loss": 0.4746,
|
| 351 |
+
"step": 570
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.94,
|
| 355 |
+
"learning_rate": 1.8119935170178285e-05,
|
| 356 |
+
"loss": 0.5074,
|
| 357 |
+
"step": 580
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.96,
|
| 361 |
+
"learning_rate": 1.808752025931929e-05,
|
| 362 |
+
"loss": 0.5437,
|
| 363 |
+
"step": 590
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.97,
|
| 367 |
+
"learning_rate": 1.8055105348460294e-05,
|
| 368 |
+
"loss": 0.4835,
|
| 369 |
+
"step": 600
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.99,
|
| 373 |
+
"learning_rate": 1.80226904376013e-05,
|
| 374 |
+
"loss": 0.5425,
|
| 375 |
+
"step": 610
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.0,
|
| 379 |
+
"eval_FN": 1004,
|
| 380 |
+
"eval_FP": 3324,
|
| 381 |
+
"eval_TN": 12435,
|
| 382 |
+
"eval_TP": 3902,
|
| 383 |
+
"eval_accuracy": 0.7905637551415436,
|
| 384 |
+
"eval_f1": 0.6432575008242664,
|
| 385 |
+
"eval_loss": 0.4139794707298279,
|
| 386 |
+
"eval_precision": 0.5399944644339884,
|
| 387 |
+
"eval_recall": 0.795352629433347,
|
| 388 |
+
"eval_runtime": 142.1799,
|
| 389 |
+
"eval_samples_per_second": 145.344,
|
| 390 |
+
"eval_steps_per_second": 9.087,
|
| 391 |
+
"step": 617
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.0,
|
| 395 |
+
"learning_rate": 1.7990275526742303e-05,
|
| 396 |
+
"loss": 0.3872,
|
| 397 |
+
"step": 620
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 1.02,
|
| 401 |
+
"learning_rate": 1.7957860615883308e-05,
|
| 402 |
+
"loss": 0.3541,
|
| 403 |
+
"step": 630
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 1.04,
|
| 407 |
+
"learning_rate": 1.7925445705024312e-05,
|
| 408 |
+
"loss": 0.3185,
|
| 409 |
+
"step": 640
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.05,
|
| 413 |
+
"learning_rate": 1.7893030794165316e-05,
|
| 414 |
+
"loss": 0.4364,
|
| 415 |
+
"step": 650
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.07,
|
| 419 |
+
"learning_rate": 1.7860615883306324e-05,
|
| 420 |
+
"loss": 0.4102,
|
| 421 |
+
"step": 660
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 1.09,
|
| 425 |
+
"learning_rate": 1.782820097244733e-05,
|
| 426 |
+
"loss": 0.3797,
|
| 427 |
+
"step": 670
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 1.1,
|
| 431 |
+
"learning_rate": 1.7795786061588333e-05,
|
| 432 |
+
"loss": 0.3964,
|
| 433 |
+
"step": 680
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 1.12,
|
| 437 |
+
"learning_rate": 1.7763371150729338e-05,
|
| 438 |
+
"loss": 0.4649,
|
| 439 |
+
"step": 690
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 1.13,
|
| 443 |
+
"learning_rate": 1.7730956239870342e-05,
|
| 444 |
+
"loss": 0.4855,
|
| 445 |
+
"step": 700
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 1.15,
|
| 449 |
+
"learning_rate": 1.7698541329011346e-05,
|
| 450 |
+
"loss": 0.4026,
|
| 451 |
+
"step": 710
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.17,
|
| 455 |
+
"learning_rate": 1.766612641815235e-05,
|
| 456 |
+
"loss": 0.3695,
|
| 457 |
+
"step": 720
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.18,
|
| 461 |
+
"learning_rate": 1.7633711507293355e-05,
|
| 462 |
+
"loss": 0.4194,
|
| 463 |
+
"step": 730
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 1.2,
|
| 467 |
+
"learning_rate": 1.760129659643436e-05,
|
| 468 |
+
"loss": 0.409,
|
| 469 |
+
"step": 740
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 1.22,
|
| 473 |
+
"learning_rate": 1.7568881685575364e-05,
|
| 474 |
+
"loss": 0.3952,
|
| 475 |
+
"step": 750
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 1.23,
|
| 479 |
+
"learning_rate": 1.7536466774716372e-05,
|
| 480 |
+
"loss": 0.4426,
|
| 481 |
+
"step": 760
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 1.25,
|
| 485 |
+
"learning_rate": 1.7504051863857376e-05,
|
| 486 |
+
"loss": 0.333,
|
| 487 |
+
"step": 770
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 1.26,
|
| 491 |
+
"learning_rate": 1.747163695299838e-05,
|
| 492 |
+
"loss": 0.4531,
|
| 493 |
+
"step": 780
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.28,
|
| 497 |
+
"learning_rate": 1.7439222042139385e-05,
|
| 498 |
+
"loss": 0.4094,
|
| 499 |
+
"step": 790
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.3,
|
| 503 |
+
"learning_rate": 1.740680713128039e-05,
|
| 504 |
+
"loss": 0.4931,
|
| 505 |
+
"step": 800
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.31,
|
| 509 |
+
"learning_rate": 1.7374392220421394e-05,
|
| 510 |
+
"loss": 0.4147,
|
| 511 |
+
"step": 810
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 1.33,
|
| 515 |
+
"learning_rate": 1.7341977309562402e-05,
|
| 516 |
+
"loss": 0.3243,
|
| 517 |
+
"step": 820
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.35,
|
| 521 |
+
"learning_rate": 1.7309562398703406e-05,
|
| 522 |
+
"loss": 0.3881,
|
| 523 |
+
"step": 830
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 1.36,
|
| 527 |
+
"learning_rate": 1.727714748784441e-05,
|
| 528 |
+
"loss": 0.4175,
|
| 529 |
+
"step": 840
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 1.38,
|
| 533 |
+
"learning_rate": 1.7244732576985415e-05,
|
| 534 |
+
"loss": 0.4215,
|
| 535 |
+
"step": 850
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.39,
|
| 539 |
+
"learning_rate": 1.721231766612642e-05,
|
| 540 |
+
"loss": 0.4549,
|
| 541 |
+
"step": 860
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.41,
|
| 545 |
+
"learning_rate": 1.7179902755267424e-05,
|
| 546 |
+
"loss": 0.4592,
|
| 547 |
+
"step": 870
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.43,
|
| 551 |
+
"learning_rate": 1.7147487844408432e-05,
|
| 552 |
+
"loss": 0.4325,
|
| 553 |
+
"step": 880
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.44,
|
| 557 |
+
"learning_rate": 1.7115072933549436e-05,
|
| 558 |
+
"loss": 0.3814,
|
| 559 |
+
"step": 890
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.46,
|
| 563 |
+
"learning_rate": 1.708265802269044e-05,
|
| 564 |
+
"loss": 0.4048,
|
| 565 |
+
"step": 900
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.47,
|
| 569 |
+
"learning_rate": 1.7050243111831445e-05,
|
| 570 |
+
"loss": 0.375,
|
| 571 |
+
"step": 910
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.49,
|
| 575 |
+
"learning_rate": 1.701782820097245e-05,
|
| 576 |
+
"loss": 0.3827,
|
| 577 |
+
"step": 920
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.51,
|
| 581 |
+
"learning_rate": 1.6985413290113454e-05,
|
| 582 |
+
"loss": 0.3957,
|
| 583 |
+
"step": 930
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.52,
|
| 587 |
+
"learning_rate": 1.695299837925446e-05,
|
| 588 |
+
"loss": 0.4487,
|
| 589 |
+
"step": 940
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.54,
|
| 593 |
+
"learning_rate": 1.6920583468395463e-05,
|
| 594 |
+
"loss": 0.4007,
|
| 595 |
+
"step": 950
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.56,
|
| 599 |
+
"learning_rate": 1.6888168557536467e-05,
|
| 600 |
+
"loss": 0.4355,
|
| 601 |
+
"step": 960
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.57,
|
| 605 |
+
"learning_rate": 1.685575364667747e-05,
|
| 606 |
+
"loss": 0.4302,
|
| 607 |
+
"step": 970
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 1.59,
|
| 611 |
+
"learning_rate": 1.6823338735818476e-05,
|
| 612 |
+
"loss": 0.4294,
|
| 613 |
+
"step": 980
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.6,
|
| 617 |
+
"learning_rate": 1.679092382495948e-05,
|
| 618 |
+
"loss": 0.4039,
|
| 619 |
+
"step": 990
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.62,
|
| 623 |
+
"learning_rate": 1.6758508914100488e-05,
|
| 624 |
+
"loss": 0.4001,
|
| 625 |
+
"step": 1000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.64,
|
| 629 |
+
"learning_rate": 1.6726094003241493e-05,
|
| 630 |
+
"loss": 0.4316,
|
| 631 |
+
"step": 1010
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.65,
|
| 635 |
+
"learning_rate": 1.6693679092382497e-05,
|
| 636 |
+
"loss": 0.4332,
|
| 637 |
+
"step": 1020
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.67,
|
| 641 |
+
"learning_rate": 1.66612641815235e-05,
|
| 642 |
+
"loss": 0.4815,
|
| 643 |
+
"step": 1030
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.69,
|
| 647 |
+
"learning_rate": 1.6628849270664506e-05,
|
| 648 |
+
"loss": 0.3744,
|
| 649 |
+
"step": 1040
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.7,
|
| 653 |
+
"learning_rate": 1.659643435980551e-05,
|
| 654 |
+
"loss": 0.4437,
|
| 655 |
+
"step": 1050
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.72,
|
| 659 |
+
"learning_rate": 1.6564019448946518e-05,
|
| 660 |
+
"loss": 0.4247,
|
| 661 |
+
"step": 1060
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.73,
|
| 665 |
+
"learning_rate": 1.6531604538087523e-05,
|
| 666 |
+
"loss": 0.4649,
|
| 667 |
+
"step": 1070
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.75,
|
| 671 |
+
"learning_rate": 1.6499189627228527e-05,
|
| 672 |
+
"loss": 0.4461,
|
| 673 |
+
"step": 1080
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.77,
|
| 677 |
+
"learning_rate": 1.646677471636953e-05,
|
| 678 |
+
"loss": 0.5386,
|
| 679 |
+
"step": 1090
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.78,
|
| 683 |
+
"learning_rate": 1.6434359805510536e-05,
|
| 684 |
+
"loss": 0.3934,
|
| 685 |
+
"step": 1100
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.8,
|
| 689 |
+
"learning_rate": 1.640194489465154e-05,
|
| 690 |
+
"loss": 0.4158,
|
| 691 |
+
"step": 1110
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 1.82,
|
| 695 |
+
"learning_rate": 1.6369529983792548e-05,
|
| 696 |
+
"loss": 0.2754,
|
| 697 |
+
"step": 1120
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.83,
|
| 701 |
+
"learning_rate": 1.6337115072933553e-05,
|
| 702 |
+
"loss": 0.4166,
|
| 703 |
+
"step": 1130
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.85,
|
| 707 |
+
"learning_rate": 1.6304700162074557e-05,
|
| 708 |
+
"loss": 0.3605,
|
| 709 |
+
"step": 1140
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 1.86,
|
| 713 |
+
"learning_rate": 1.627228525121556e-05,
|
| 714 |
+
"loss": 0.4361,
|
| 715 |
+
"step": 1150
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.88,
|
| 719 |
+
"learning_rate": 1.6239870340356566e-05,
|
| 720 |
+
"loss": 0.4436,
|
| 721 |
+
"step": 1160
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.9,
|
| 725 |
+
"learning_rate": 1.620745542949757e-05,
|
| 726 |
+
"loss": 0.4006,
|
| 727 |
+
"step": 1170
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.91,
|
| 731 |
+
"learning_rate": 1.6175040518638575e-05,
|
| 732 |
+
"loss": 0.2996,
|
| 733 |
+
"step": 1180
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.93,
|
| 737 |
+
"learning_rate": 1.614262560777958e-05,
|
| 738 |
+
"loss": 0.4593,
|
| 739 |
+
"step": 1190
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.94,
|
| 743 |
+
"learning_rate": 1.6110210696920584e-05,
|
| 744 |
+
"loss": 0.4633,
|
| 745 |
+
"step": 1200
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.96,
|
| 749 |
+
"learning_rate": 1.6077795786061588e-05,
|
| 750 |
+
"loss": 0.3404,
|
| 751 |
+
"step": 1210
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.98,
|
| 755 |
+
"learning_rate": 1.6045380875202596e-05,
|
| 756 |
+
"loss": 0.3477,
|
| 757 |
+
"step": 1220
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.99,
|
| 761 |
+
"learning_rate": 1.60129659643436e-05,
|
| 762 |
+
"loss": 0.4074,
|
| 763 |
+
"step": 1230
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 2.0,
|
| 767 |
+
"eval_FN": 1412,
|
| 768 |
+
"eval_FP": 1850,
|
| 769 |
+
"eval_TN": 13909,
|
| 770 |
+
"eval_TP": 3494,
|
| 771 |
+
"eval_accuracy": 0.8421485603677716,
|
| 772 |
+
"eval_f1": 0.6817560975609757,
|
| 773 |
+
"eval_loss": 0.374613881111145,
|
| 774 |
+
"eval_precision": 0.6538173652694611,
|
| 775 |
+
"eval_recall": 0.7121891561353445,
|
| 776 |
+
"eval_runtime": 142.3346,
|
| 777 |
+
"eval_samples_per_second": 145.186,
|
| 778 |
+
"eval_steps_per_second": 9.077,
|
| 779 |
+
"step": 1234
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.01,
|
| 783 |
+
"learning_rate": 1.5980551053484605e-05,
|
| 784 |
+
"loss": 0.3298,
|
| 785 |
+
"step": 1240
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 2.03,
|
| 789 |
+
"learning_rate": 1.594813614262561e-05,
|
| 790 |
+
"loss": 0.2898,
|
| 791 |
+
"step": 1250
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 2.04,
|
| 795 |
+
"learning_rate": 1.5915721231766613e-05,
|
| 796 |
+
"loss": 0.2781,
|
| 797 |
+
"step": 1260
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 2.06,
|
| 801 |
+
"learning_rate": 1.5883306320907618e-05,
|
| 802 |
+
"loss": 0.4103,
|
| 803 |
+
"step": 1270
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 2.07,
|
| 807 |
+
"learning_rate": 1.5850891410048622e-05,
|
| 808 |
+
"loss": 0.2744,
|
| 809 |
+
"step": 1280
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.09,
|
| 813 |
+
"learning_rate": 1.5818476499189627e-05,
|
| 814 |
+
"loss": 0.2781,
|
| 815 |
+
"step": 1290
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 2.11,
|
| 819 |
+
"learning_rate": 1.578606158833063e-05,
|
| 820 |
+
"loss": 0.2509,
|
| 821 |
+
"step": 1300
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 2.12,
|
| 825 |
+
"learning_rate": 1.575364667747164e-05,
|
| 826 |
+
"loss": 0.2638,
|
| 827 |
+
"step": 1310
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 2.14,
|
| 831 |
+
"learning_rate": 1.5721231766612643e-05,
|
| 832 |
+
"loss": 0.4286,
|
| 833 |
+
"step": 1320
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 2.16,
|
| 837 |
+
"learning_rate": 1.5688816855753648e-05,
|
| 838 |
+
"loss": 0.3699,
|
| 839 |
+
"step": 1330
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 2.17,
|
| 843 |
+
"learning_rate": 1.5656401944894652e-05,
|
| 844 |
+
"loss": 0.3562,
|
| 845 |
+
"step": 1340
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 2.19,
|
| 849 |
+
"learning_rate": 1.5623987034035657e-05,
|
| 850 |
+
"loss": 0.4909,
|
| 851 |
+
"step": 1350
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 2.2,
|
| 855 |
+
"learning_rate": 1.559157212317666e-05,
|
| 856 |
+
"loss": 0.303,
|
| 857 |
+
"step": 1360
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.22,
|
| 861 |
+
"learning_rate": 1.555915721231767e-05,
|
| 862 |
+
"loss": 0.2468,
|
| 863 |
+
"step": 1370
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 2.24,
|
| 867 |
+
"learning_rate": 1.5526742301458673e-05,
|
| 868 |
+
"loss": 0.3632,
|
| 869 |
+
"step": 1380
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 2.25,
|
| 873 |
+
"learning_rate": 1.5494327390599678e-05,
|
| 874 |
+
"loss": 0.2486,
|
| 875 |
+
"step": 1390
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 2.27,
|
| 879 |
+
"learning_rate": 1.5461912479740682e-05,
|
| 880 |
+
"loss": 0.2667,
|
| 881 |
+
"step": 1400
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 2.29,
|
| 885 |
+
"learning_rate": 1.5429497568881687e-05,
|
| 886 |
+
"loss": 0.205,
|
| 887 |
+
"step": 1410
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 2.3,
|
| 891 |
+
"learning_rate": 1.539708265802269e-05,
|
| 892 |
+
"loss": 0.2899,
|
| 893 |
+
"step": 1420
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 2.32,
|
| 897 |
+
"learning_rate": 1.53646677471637e-05,
|
| 898 |
+
"loss": 0.3234,
|
| 899 |
+
"step": 1430
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 2.33,
|
| 903 |
+
"learning_rate": 1.5332252836304703e-05,
|
| 904 |
+
"loss": 0.2796,
|
| 905 |
+
"step": 1440
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.35,
|
| 909 |
+
"learning_rate": 1.5299837925445708e-05,
|
| 910 |
+
"loss": 0.3992,
|
| 911 |
+
"step": 1450
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 2.37,
|
| 915 |
+
"learning_rate": 1.5267423014586712e-05,
|
| 916 |
+
"loss": 0.3905,
|
| 917 |
+
"step": 1460
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 2.38,
|
| 921 |
+
"learning_rate": 1.5235008103727715e-05,
|
| 922 |
+
"loss": 0.3359,
|
| 923 |
+
"step": 1470
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 2.4,
|
| 927 |
+
"learning_rate": 1.520259319286872e-05,
|
| 928 |
+
"loss": 0.3915,
|
| 929 |
+
"step": 1480
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 2.41,
|
| 933 |
+
"learning_rate": 1.5170178282009727e-05,
|
| 934 |
+
"loss": 0.316,
|
| 935 |
+
"step": 1490
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 2.43,
|
| 939 |
+
"learning_rate": 1.5137763371150732e-05,
|
| 940 |
+
"loss": 0.4686,
|
| 941 |
+
"step": 1500
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 2.45,
|
| 945 |
+
"learning_rate": 1.5105348460291736e-05,
|
| 946 |
+
"loss": 0.4413,
|
| 947 |
+
"step": 1510
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 2.46,
|
| 951 |
+
"learning_rate": 1.507293354943274e-05,
|
| 952 |
+
"loss": 0.2946,
|
| 953 |
+
"step": 1520
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 2.48,
|
| 957 |
+
"learning_rate": 1.5040518638573745e-05,
|
| 958 |
+
"loss": 0.3483,
|
| 959 |
+
"step": 1530
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 2.5,
|
| 963 |
+
"learning_rate": 1.500810372771475e-05,
|
| 964 |
+
"loss": 0.3222,
|
| 965 |
+
"step": 1540
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 2.51,
|
| 969 |
+
"learning_rate": 1.4975688816855755e-05,
|
| 970 |
+
"loss": 0.3557,
|
| 971 |
+
"step": 1550
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 2.53,
|
| 975 |
+
"learning_rate": 1.494327390599676e-05,
|
| 976 |
+
"loss": 0.2867,
|
| 977 |
+
"step": 1560
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 2.54,
|
| 981 |
+
"learning_rate": 1.4910858995137764e-05,
|
| 982 |
+
"loss": 0.2718,
|
| 983 |
+
"step": 1570
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 2.56,
|
| 987 |
+
"learning_rate": 1.4878444084278769e-05,
|
| 988 |
+
"loss": 0.265,
|
| 989 |
+
"step": 1580
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 2.58,
|
| 993 |
+
"learning_rate": 1.4846029173419773e-05,
|
| 994 |
+
"loss": 0.3511,
|
| 995 |
+
"step": 1590
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 2.59,
|
| 999 |
+
"learning_rate": 1.4813614262560778e-05,
|
| 1000 |
+
"loss": 0.3416,
|
| 1001 |
+
"step": 1600
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 2.61,
|
| 1005 |
+
"learning_rate": 1.4781199351701785e-05,
|
| 1006 |
+
"loss": 0.3477,
|
| 1007 |
+
"step": 1610
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 2.63,
|
| 1011 |
+
"learning_rate": 1.474878444084279e-05,
|
| 1012 |
+
"loss": 0.2636,
|
| 1013 |
+
"step": 1620
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 2.64,
|
| 1017 |
+
"learning_rate": 1.4716369529983794e-05,
|
| 1018 |
+
"loss": 0.3245,
|
| 1019 |
+
"step": 1630
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 2.66,
|
| 1023 |
+
"learning_rate": 1.4683954619124799e-05,
|
| 1024 |
+
"loss": 0.3413,
|
| 1025 |
+
"step": 1640
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 2.67,
|
| 1029 |
+
"learning_rate": 1.4651539708265803e-05,
|
| 1030 |
+
"loss": 0.3294,
|
| 1031 |
+
"step": 1650
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 2.69,
|
| 1035 |
+
"learning_rate": 1.4619124797406807e-05,
|
| 1036 |
+
"loss": 0.3222,
|
| 1037 |
+
"step": 1660
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 2.71,
|
| 1041 |
+
"learning_rate": 1.4586709886547814e-05,
|
| 1042 |
+
"loss": 0.2559,
|
| 1043 |
+
"step": 1670
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 2.72,
|
| 1047 |
+
"learning_rate": 1.4554294975688818e-05,
|
| 1048 |
+
"loss": 0.3845,
|
| 1049 |
+
"step": 1680
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 2.74,
|
| 1053 |
+
"learning_rate": 1.4521880064829822e-05,
|
| 1054 |
+
"loss": 0.2523,
|
| 1055 |
+
"step": 1690
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 2.76,
|
| 1059 |
+
"learning_rate": 1.4489465153970827e-05,
|
| 1060 |
+
"loss": 0.2542,
|
| 1061 |
+
"step": 1700
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.77,
|
| 1065 |
+
"learning_rate": 1.4457050243111831e-05,
|
| 1066 |
+
"loss": 0.2859,
|
| 1067 |
+
"step": 1710
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 2.79,
|
| 1071 |
+
"learning_rate": 1.4424635332252836e-05,
|
| 1072 |
+
"loss": 0.4221,
|
| 1073 |
+
"step": 1720
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.8,
|
| 1077 |
+
"learning_rate": 1.4392220421393844e-05,
|
| 1078 |
+
"loss": 0.3414,
|
| 1079 |
+
"step": 1730
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 2.82,
|
| 1083 |
+
"learning_rate": 1.4359805510534848e-05,
|
| 1084 |
+
"loss": 0.2738,
|
| 1085 |
+
"step": 1740
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 2.84,
|
| 1089 |
+
"learning_rate": 1.4327390599675852e-05,
|
| 1090 |
+
"loss": 0.2955,
|
| 1091 |
+
"step": 1750
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.85,
|
| 1095 |
+
"learning_rate": 1.4294975688816857e-05,
|
| 1096 |
+
"loss": 0.315,
|
| 1097 |
+
"step": 1760
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.87,
|
| 1101 |
+
"learning_rate": 1.4262560777957861e-05,
|
| 1102 |
+
"loss": 0.2628,
|
| 1103 |
+
"step": 1770
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 2.88,
|
| 1107 |
+
"learning_rate": 1.4230145867098866e-05,
|
| 1108 |
+
"loss": 0.2731,
|
| 1109 |
+
"step": 1780
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 2.9,
|
| 1113 |
+
"learning_rate": 1.4197730956239872e-05,
|
| 1114 |
+
"loss": 0.3792,
|
| 1115 |
+
"step": 1790
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.92,
|
| 1119 |
+
"learning_rate": 1.4165316045380876e-05,
|
| 1120 |
+
"loss": 0.3875,
|
| 1121 |
+
"step": 1800
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 2.93,
|
| 1125 |
+
"learning_rate": 1.413290113452188e-05,
|
| 1126 |
+
"loss": 0.2959,
|
| 1127 |
+
"step": 1810
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 2.95,
|
| 1131 |
+
"learning_rate": 1.4100486223662885e-05,
|
| 1132 |
+
"loss": 0.2504,
|
| 1133 |
+
"step": 1820
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.97,
|
| 1137 |
+
"learning_rate": 1.4068071312803891e-05,
|
| 1138 |
+
"loss": 0.3684,
|
| 1139 |
+
"step": 1830
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.98,
|
| 1143 |
+
"learning_rate": 1.4035656401944896e-05,
|
| 1144 |
+
"loss": 0.3581,
|
| 1145 |
+
"step": 1840
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 3.0,
|
| 1149 |
+
"learning_rate": 1.4003241491085902e-05,
|
| 1150 |
+
"loss": 0.3028,
|
| 1151 |
+
"step": 1850
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 3.0,
|
| 1155 |
+
"eval_FN": 932,
|
| 1156 |
+
"eval_FP": 2492,
|
| 1157 |
+
"eval_TN": 13267,
|
| 1158 |
+
"eval_TP": 3974,
|
| 1159 |
+
"eval_accuracy": 0.8343092184853618,
|
| 1160 |
+
"eval_f1": 0.6989096025325361,
|
| 1161 |
+
"eval_loss": 0.39430731534957886,
|
| 1162 |
+
"eval_precision": 0.6145994432415713,
|
| 1163 |
+
"eval_recall": 0.8100285364859355,
|
| 1164 |
+
"eval_runtime": 142.2921,
|
| 1165 |
+
"eval_samples_per_second": 145.229,
|
| 1166 |
+
"eval_steps_per_second": 9.08,
|
| 1167 |
+
"step": 1851
|
| 1168 |
+
}
|
| 1169 |
+
],
|
| 1170 |
+
"logging_steps": 10,
|
| 1171 |
+
"max_steps": 6170,
|
| 1172 |
+
"num_train_epochs": 10,
|
| 1173 |
+
"save_steps": 500,
|
| 1174 |
+
"total_flos": 7789139682877440.0,
|
| 1175 |
+
"trial_name": null,
|
| 1176 |
+
"trial_params": null
|
| 1177 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:973562431ff672764aee4ed9cd57ccd324cd788c53c8f3bc8094aaabb05bea7b
|
| 3 |
+
size 4664
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|