Upload folder using huggingface_hub
Browse files- config.json +56 -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 +15 -0
- tokenizer.json +0 -0
- tokenizer_config.json +57 -0
- trainer_state.json +1044 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "FacebookAI/roberta-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForMultiLabelClassification"
|
| 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 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0",
|
| 15 |
+
"1": "LABEL_1",
|
| 16 |
+
"2": "LABEL_2",
|
| 17 |
+
"3": "LABEL_3",
|
| 18 |
+
"4": "LABEL_4",
|
| 19 |
+
"5": "LABEL_5",
|
| 20 |
+
"6": "LABEL_6",
|
| 21 |
+
"7": "LABEL_7",
|
| 22 |
+
"8": "LABEL_8",
|
| 23 |
+
"9": "LABEL_9",
|
| 24 |
+
"10": "LABEL_10",
|
| 25 |
+
"11": "LABEL_11"
|
| 26 |
+
},
|
| 27 |
+
"initializer_range": 0.02,
|
| 28 |
+
"intermediate_size": 3072,
|
| 29 |
+
"label2id": {
|
| 30 |
+
"LABEL_0": 0,
|
| 31 |
+
"LABEL_1": 1,
|
| 32 |
+
"LABEL_10": 10,
|
| 33 |
+
"LABEL_11": 11,
|
| 34 |
+
"LABEL_2": 2,
|
| 35 |
+
"LABEL_3": 3,
|
| 36 |
+
"LABEL_4": 4,
|
| 37 |
+
"LABEL_5": 5,
|
| 38 |
+
"LABEL_6": 6,
|
| 39 |
+
"LABEL_7": 7,
|
| 40 |
+
"LABEL_8": 8,
|
| 41 |
+
"LABEL_9": 9
|
| 42 |
+
},
|
| 43 |
+
"layer_norm_eps": 1e-05,
|
| 44 |
+
"max_position_embeddings": 514,
|
| 45 |
+
"model_type": "roberta",
|
| 46 |
+
"num_attention_heads": 12,
|
| 47 |
+
"num_hidden_layers": 12,
|
| 48 |
+
"pad_token_id": 1,
|
| 49 |
+
"position_embedding_type": "absolute",
|
| 50 |
+
"problem_type": "multi_label_classification",
|
| 51 |
+
"torch_dtype": "float32",
|
| 52 |
+
"transformers_version": "4.39.1",
|
| 53 |
+
"type_vocab_size": 1,
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 50265
|
| 56 |
+
}
|
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 sha256:13a9dd273db43352f069a46dd89183458bc0a77e9d32d6d9e5738ef563b0eab6
|
| 3 |
+
size 498643584
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e344d9323361d88b42a016e9a14ad7117654c6b6d44c2fc379c68414ad86ddf8
|
| 3 |
+
size 997406970
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ff03cb5a7c5b7b58ae9a65bf9107d0599b8dcfcb859e65f1c02740140e3c776
|
| 3 |
+
size 14180
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f718d77727b36397c11f5e43219dde8a8f40ff1ec9d4e6fe982f01e39dce83db
|
| 3 |
+
size 1064
|
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.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"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 |
+
"trim_offsets": true,
|
| 56 |
+
"unk_token": "<unk>"
|
| 57 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1044 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9175407198197038,
|
| 3 |
+
"best_model_checkpoint": "./saved_models/roberta_sbdh_gpt4_v2_0/checkpoint-792",
|
| 4 |
+
"epoch": 33.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 792,
|
| 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 |
+
"grad_norm": 1.3806931972503662,
|
| 14 |
+
"learning_rate": 2.5e-06,
|
| 15 |
+
"loss": 0.6134,
|
| 16 |
+
"step": 24
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_acc_macro": 0.10215572903385978,
|
| 21 |
+
"eval_acc_micro": 0.17855121675155125,
|
| 22 |
+
"eval_auc_macro": 0.6100041638713678,
|
| 23 |
+
"eval_auc_micro": 0.6464031657743947,
|
| 24 |
+
"eval_f1_at_5": 0.2718421152169764,
|
| 25 |
+
"eval_f1_at_8": 0.22628042555556532,
|
| 26 |
+
"eval_f1_macro": 0.1715749294146787,
|
| 27 |
+
"eval_f1_micro": 0.30300120048017753,
|
| 28 |
+
"eval_loss": 0.4604409337043762,
|
| 29 |
+
"eval_prec_at_5": 0.17442922374429226,
|
| 30 |
+
"eval_prec_at_8": 0.1329908675799087,
|
| 31 |
+
"eval_prec_macro": 0.1401339887841245,
|
| 32 |
+
"eval_prec_micro": 0.2132477188239198,
|
| 33 |
+
"eval_rec_at_5": 0.6156773211567731,
|
| 34 |
+
"eval_rec_at_8": 0.7579908675799086,
|
| 35 |
+
"eval_rec_macro": 0.36147047478751165,
|
| 36 |
+
"eval_rec_micro": 0.5232172470978007,
|
| 37 |
+
"eval_runtime": 3.307,
|
| 38 |
+
"eval_samples_per_second": 264.896,
|
| 39 |
+
"eval_steps_per_second": 33.263,
|
| 40 |
+
"step": 24
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 2.0,
|
| 44 |
+
"grad_norm": 0.42557960748672485,
|
| 45 |
+
"learning_rate": 5e-06,
|
| 46 |
+
"loss": 0.4178,
|
| 47 |
+
"step": 48
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 2.0,
|
| 51 |
+
"eval_acc_macro": 0.054161895439312095,
|
| 52 |
+
"eval_acc_micro": 0.18999494694289085,
|
| 53 |
+
"eval_auc_macro": 0.777689574836493,
|
| 54 |
+
"eval_auc_micro": 0.6837933158193559,
|
| 55 |
+
"eval_f1_at_5": 0.2980163960623497,
|
| 56 |
+
"eval_f1_at_8": 0.2355215334598575,
|
| 57 |
+
"eval_f1_macro": 0.08160871746824805,
|
| 58 |
+
"eval_f1_micro": 0.31932059447980304,
|
| 59 |
+
"eval_loss": 0.35270196199417114,
|
| 60 |
+
"eval_prec_at_5": 0.19109589041095887,
|
| 61 |
+
"eval_prec_at_8": 0.13855593607305935,
|
| 62 |
+
"eval_prec_macro": 0.057278901143779114,
|
| 63 |
+
"eval_prec_micro": 0.32724107919927525,
|
| 64 |
+
"eval_rec_at_5": 0.6765601217656013,
|
| 65 |
+
"eval_rec_at_8": 0.784627092846271,
|
| 66 |
+
"eval_rec_macro": 0.14846491228063083,
|
| 67 |
+
"eval_rec_micro": 0.3117744610281665,
|
| 68 |
+
"eval_runtime": 3.2254,
|
| 69 |
+
"eval_samples_per_second": 271.594,
|
| 70 |
+
"eval_steps_per_second": 34.104,
|
| 71 |
+
"step": 48
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 3.0,
|
| 75 |
+
"grad_norm": 0.21710661053657532,
|
| 76 |
+
"learning_rate": 7.500000000000001e-06,
|
| 77 |
+
"loss": 0.3548,
|
| 78 |
+
"step": 72
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 3.0,
|
| 82 |
+
"eval_acc_macro": 0.08172631395735021,
|
| 83 |
+
"eval_acc_micro": 0.2394548994159481,
|
| 84 |
+
"eval_auc_macro": 0.8898708262347639,
|
| 85 |
+
"eval_auc_micro": 0.7531371635981566,
|
| 86 |
+
"eval_f1_at_5": 0.3273134867709266,
|
| 87 |
+
"eval_f1_at_8": 0.2656356996119607,
|
| 88 |
+
"eval_f1_macro": 0.10962566844915878,
|
| 89 |
+
"eval_f1_micro": 0.3863874345549333,
|
| 90 |
+
"eval_loss": 0.3343456983566284,
|
| 91 |
+
"eval_prec_at_5": 0.20958904109589044,
|
| 92 |
+
"eval_prec_at_8": 0.15625,
|
| 93 |
+
"eval_prec_macro": 0.08775419624473717,
|
| 94 |
+
"eval_prec_micro": 0.5241477272726528,
|
| 95 |
+
"eval_rec_at_5": 0.746765601217656,
|
| 96 |
+
"eval_rec_at_8": 0.8856544901065448,
|
| 97 |
+
"eval_rec_macro": 0.14795321637419723,
|
| 98 |
+
"eval_rec_micro": 0.30597014925370597,
|
| 99 |
+
"eval_runtime": 3.2799,
|
| 100 |
+
"eval_samples_per_second": 267.084,
|
| 101 |
+
"eval_steps_per_second": 33.538,
|
| 102 |
+
"step": 72
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 4.0,
|
| 106 |
+
"grad_norm": 0.26009050011634827,
|
| 107 |
+
"learning_rate": 1e-05,
|
| 108 |
+
"loss": 0.3315,
|
| 109 |
+
"step": 96
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 4.0,
|
| 113 |
+
"eval_acc_macro": 0.2205899892707659,
|
| 114 |
+
"eval_acc_micro": 0.31238003838770095,
|
| 115 |
+
"eval_auc_macro": 0.9150780627599414,
|
| 116 |
+
"eval_auc_micro": 0.8412487048958944,
|
| 117 |
+
"eval_f1_at_5": 0.36996556565347266,
|
| 118 |
+
"eval_f1_at_8": 0.27616272202971137,
|
| 119 |
+
"eval_f1_macro": 0.32075046434231563,
|
| 120 |
+
"eval_f1_micro": 0.4760511882997823,
|
| 121 |
+
"eval_loss": 0.30232226848602295,
|
| 122 |
+
"eval_prec_at_5": 0.23675799086757995,
|
| 123 |
+
"eval_prec_at_8": 0.16238584474885845,
|
| 124 |
+
"eval_prec_macro": 0.534140737674402,
|
| 125 |
+
"eval_prec_micro": 0.4257684761281605,
|
| 126 |
+
"eval_rec_at_5": 0.8458904109589042,
|
| 127 |
+
"eval_rec_at_8": 0.9225646879756468,
|
| 128 |
+
"eval_rec_macro": 0.37458794170416176,
|
| 129 |
+
"eval_rec_micro": 0.5398009950248308,
|
| 130 |
+
"eval_runtime": 3.2394,
|
| 131 |
+
"eval_samples_per_second": 270.421,
|
| 132 |
+
"eval_steps_per_second": 33.957,
|
| 133 |
+
"step": 96
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 5.0,
|
| 137 |
+
"grad_norm": 0.37228670716285706,
|
| 138 |
+
"learning_rate": 9.722222222222223e-06,
|
| 139 |
+
"loss": 0.2947,
|
| 140 |
+
"step": 120
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 5.0,
|
| 144 |
+
"eval_acc_macro": 0.4283107268415045,
|
| 145 |
+
"eval_acc_micro": 0.5237793278376333,
|
| 146 |
+
"eval_auc_macro": 0.944198889889774,
|
| 147 |
+
"eval_auc_micro": 0.9267578754982164,
|
| 148 |
+
"eval_f1_at_5": 0.4096518568367693,
|
| 149 |
+
"eval_f1_at_8": 0.29017455585500507,
|
| 150 |
+
"eval_f1_macro": 0.539018158110013,
|
| 151 |
+
"eval_f1_micro": 0.6874739908447202,
|
| 152 |
+
"eval_loss": 0.25655001401901245,
|
| 153 |
+
"eval_prec_at_5": 0.2618721461187215,
|
| 154 |
+
"eval_prec_at_8": 0.17051940639269406,
|
| 155 |
+
"eval_prec_macro": 0.5661692193127564,
|
| 156 |
+
"eval_prec_micro": 0.6900584795321061,
|
| 157 |
+
"eval_rec_at_5": 0.9402587519025875,
|
| 158 |
+
"eval_rec_at_8": 0.9727929984779301,
|
| 159 |
+
"eval_rec_macro": 0.564018051489352,
|
| 160 |
+
"eval_rec_micro": 0.6849087893863445,
|
| 161 |
+
"eval_runtime": 3.287,
|
| 162 |
+
"eval_samples_per_second": 266.502,
|
| 163 |
+
"eval_steps_per_second": 33.465,
|
| 164 |
+
"step": 120
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 6.0,
|
| 168 |
+
"grad_norm": 0.2922113835811615,
|
| 169 |
+
"learning_rate": 9.444444444444445e-06,
|
| 170 |
+
"loss": 0.255,
|
| 171 |
+
"step": 144
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 6.0,
|
| 175 |
+
"eval_acc_macro": 0.5136685674094098,
|
| 176 |
+
"eval_acc_micro": 0.6013651877132695,
|
| 177 |
+
"eval_auc_macro": 0.9633338119361814,
|
| 178 |
+
"eval_auc_micro": 0.9605888281923002,
|
| 179 |
+
"eval_f1_at_5": 0.42005389856494574,
|
| 180 |
+
"eval_f1_at_8": 0.29155799132187266,
|
| 181 |
+
"eval_f1_macro": 0.6233454431290059,
|
| 182 |
+
"eval_f1_micro": 0.7510656436486999,
|
| 183 |
+
"eval_loss": 0.21982233226299286,
|
| 184 |
+
"eval_prec_at_5": 0.2687214611872146,
|
| 185 |
+
"eval_prec_at_8": 0.1713755707762557,
|
| 186 |
+
"eval_prec_macro": 0.7586900162484641,
|
| 187 |
+
"eval_prec_micro": 0.7728070175437918,
|
| 188 |
+
"eval_rec_at_5": 0.9615677321156773,
|
| 189 |
+
"eval_rec_at_8": 0.976027397260274,
|
| 190 |
+
"eval_rec_macro": 0.6250160961046504,
|
| 191 |
+
"eval_rec_micro": 0.7305140961856774,
|
| 192 |
+
"eval_runtime": 3.2407,
|
| 193 |
+
"eval_samples_per_second": 270.308,
|
| 194 |
+
"eval_steps_per_second": 33.943,
|
| 195 |
+
"step": 144
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 7.0,
|
| 199 |
+
"grad_norm": 0.6698095202445984,
|
| 200 |
+
"learning_rate": 9.166666666666666e-06,
|
| 201 |
+
"loss": 0.2228,
|
| 202 |
+
"step": 168
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 7.0,
|
| 206 |
+
"eval_acc_macro": 0.6303220555404457,
|
| 207 |
+
"eval_acc_micro": 0.6847290640393606,
|
| 208 |
+
"eval_auc_macro": 0.9734325874763358,
|
| 209 |
+
"eval_auc_micro": 0.9748103810769206,
|
| 210 |
+
"eval_f1_at_5": 0.42340560395061977,
|
| 211 |
+
"eval_f1_at_8": 0.2920219218917755,
|
| 212 |
+
"eval_f1_macro": 0.7458516061176264,
|
| 213 |
+
"eval_f1_micro": 0.8128654970759555,
|
| 214 |
+
"eval_loss": 0.19146637618541718,
|
| 215 |
+
"eval_prec_at_5": 0.27100456621004565,
|
| 216 |
+
"eval_prec_at_8": 0.1716609589041096,
|
| 217 |
+
"eval_prec_macro": 0.86104613287547,
|
| 218 |
+
"eval_prec_micro": 0.8190235690235,
|
| 219 |
+
"eval_rec_at_5": 0.9674657534246576,
|
| 220 |
+
"eval_rec_at_8": 0.9771689497716894,
|
| 221 |
+
"eval_rec_macro": 0.7345526249297327,
|
| 222 |
+
"eval_rec_micro": 0.806799336650016,
|
| 223 |
+
"eval_runtime": 3.3221,
|
| 224 |
+
"eval_samples_per_second": 263.692,
|
| 225 |
+
"eval_steps_per_second": 33.112,
|
| 226 |
+
"step": 168
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 8.0,
|
| 230 |
+
"grad_norm": 0.3862050771713257,
|
| 231 |
+
"learning_rate": 8.888888888888888e-06,
|
| 232 |
+
"loss": 0.1968,
|
| 233 |
+
"step": 192
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 8.0,
|
| 237 |
+
"eval_acc_macro": 0.6911445121462139,
|
| 238 |
+
"eval_acc_micro": 0.7442922374428657,
|
| 239 |
+
"eval_auc_macro": 0.9794839519003108,
|
| 240 |
+
"eval_auc_micro": 0.9822416144793619,
|
| 241 |
+
"eval_f1_at_5": 0.4247570125938748,
|
| 242 |
+
"eval_f1_at_8": 0.291789958878004,
|
| 243 |
+
"eval_f1_macro": 0.7968110573167572,
|
| 244 |
+
"eval_f1_micro": 0.853403141361182,
|
| 245 |
+
"eval_loss": 0.16935397684574127,
|
| 246 |
+
"eval_prec_at_5": 0.2719178082191781,
|
| 247 |
+
"eval_prec_at_8": 0.17151826484018265,
|
| 248 |
+
"eval_prec_macro": 0.9142914075642903,
|
| 249 |
+
"eval_prec_micro": 0.9005524861877623,
|
| 250 |
+
"eval_rec_at_5": 0.9699391171993911,
|
| 251 |
+
"eval_rec_at_8": 0.9765981735159818,
|
| 252 |
+
"eval_rec_macro": 0.7524753915565773,
|
| 253 |
+
"eval_rec_micro": 0.8109452736317735,
|
| 254 |
+
"eval_runtime": 3.3593,
|
| 255 |
+
"eval_samples_per_second": 260.765,
|
| 256 |
+
"eval_steps_per_second": 32.745,
|
| 257 |
+
"step": 192
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 9.0,
|
| 261 |
+
"grad_norm": 0.3425958752632141,
|
| 262 |
+
"learning_rate": 8.611111111111112e-06,
|
| 263 |
+
"loss": 0.1756,
|
| 264 |
+
"step": 216
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 9.0,
|
| 268 |
+
"eval_acc_macro": 0.7538851024227721,
|
| 269 |
+
"eval_acc_micro": 0.7733711048158092,
|
| 270 |
+
"eval_auc_macro": 0.984468975401199,
|
| 271 |
+
"eval_auc_micro": 0.9869544212457306,
|
| 272 |
+
"eval_f1_at_5": 0.42679308597884263,
|
| 273 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 274 |
+
"eval_f1_macro": 0.8539710926649278,
|
| 275 |
+
"eval_f1_micro": 0.8722044728433809,
|
| 276 |
+
"eval_loss": 0.15195928514003754,
|
| 277 |
+
"eval_prec_at_5": 0.2732876712328767,
|
| 278 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 279 |
+
"eval_prec_macro": 0.8477325049832564,
|
| 280 |
+
"eval_prec_micro": 0.8412942989213528,
|
| 281 |
+
"eval_rec_at_5": 0.973744292237443,
|
| 282 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 283 |
+
"eval_rec_macro": 0.8733258072194854,
|
| 284 |
+
"eval_rec_micro": 0.9054726368158453,
|
| 285 |
+
"eval_runtime": 3.3296,
|
| 286 |
+
"eval_samples_per_second": 263.093,
|
| 287 |
+
"eval_steps_per_second": 33.037,
|
| 288 |
+
"step": 216
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 10.0,
|
| 292 |
+
"grad_norm": 0.3327239453792572,
|
| 293 |
+
"learning_rate": 8.333333333333334e-06,
|
| 294 |
+
"loss": 0.1578,
|
| 295 |
+
"step": 240
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 10.0,
|
| 299 |
+
"eval_acc_macro": 0.7658814025380632,
|
| 300 |
+
"eval_acc_micro": 0.7931292008961319,
|
| 301 |
+
"eval_auc_macro": 0.9863575312994878,
|
| 302 |
+
"eval_auc_micro": 0.9888696784007464,
|
| 303 |
+
"eval_f1_at_5": 0.42745954575018524,
|
| 304 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 305 |
+
"eval_f1_macro": 0.8608080570922364,
|
| 306 |
+
"eval_f1_micro": 0.8846314035817673,
|
| 307 |
+
"eval_loss": 0.13785400986671448,
|
| 308 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 309 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 310 |
+
"eval_prec_macro": 0.8916528884446233,
|
| 311 |
+
"eval_prec_micro": 0.8887029288702185,
|
| 312 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 313 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 314 |
+
"eval_rec_macro": 0.8473101584524304,
|
| 315 |
+
"eval_rec_micro": 0.8805970149253001,
|
| 316 |
+
"eval_runtime": 3.3169,
|
| 317 |
+
"eval_samples_per_second": 264.104,
|
| 318 |
+
"eval_steps_per_second": 33.164,
|
| 319 |
+
"step": 240
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 11.0,
|
| 323 |
+
"grad_norm": 0.4798758029937744,
|
| 324 |
+
"learning_rate": 8.055555555555557e-06,
|
| 325 |
+
"loss": 0.1437,
|
| 326 |
+
"step": 264
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 11.0,
|
| 330 |
+
"eval_acc_macro": 0.7817093750541505,
|
| 331 |
+
"eval_acc_micro": 0.8088347296267472,
|
| 332 |
+
"eval_auc_macro": 0.987966835700279,
|
| 333 |
+
"eval_auc_micro": 0.990288011194119,
|
| 334 |
+
"eval_f1_at_5": 0.4281808547568474,
|
| 335 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 336 |
+
"eval_f1_macro": 0.8718351509550711,
|
| 337 |
+
"eval_f1_micro": 0.8943157894736088,
|
| 338 |
+
"eval_loss": 0.1261390894651413,
|
| 339 |
+
"eval_prec_at_5": 0.2742009132420091,
|
| 340 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 341 |
+
"eval_prec_macro": 0.9059016261714952,
|
| 342 |
+
"eval_prec_micro": 0.9084687767321721,
|
| 343 |
+
"eval_rec_at_5": 0.9765981735159818,
|
| 344 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 345 |
+
"eval_rec_macro": 0.852231468930321,
|
| 346 |
+
"eval_rec_micro": 0.8805970149253001,
|
| 347 |
+
"eval_runtime": 3.1998,
|
| 348 |
+
"eval_samples_per_second": 273.768,
|
| 349 |
+
"eval_steps_per_second": 34.377,
|
| 350 |
+
"step": 264
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 12.0,
|
| 354 |
+
"grad_norm": 0.5915816426277161,
|
| 355 |
+
"learning_rate": 7.77777777777778e-06,
|
| 356 |
+
"loss": 0.1315,
|
| 357 |
+
"step": 288
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 12.0,
|
| 361 |
+
"eval_acc_macro": 0.8067338258553937,
|
| 362 |
+
"eval_acc_micro": 0.8225075528700285,
|
| 363 |
+
"eval_auc_macro": 0.9890198593241079,
|
| 364 |
+
"eval_auc_micro": 0.9912727714675422,
|
| 365 |
+
"eval_f1_at_5": 0.42745954575018524,
|
| 366 |
+
"eval_f1_at_8": 0.2925113305987028,
|
| 367 |
+
"eval_f1_macro": 0.8889792204408117,
|
| 368 |
+
"eval_f1_micro": 0.9026108578532198,
|
| 369 |
+
"eval_loss": 0.11689846962690353,
|
| 370 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 371 |
+
"eval_prec_at_8": 0.17194634703196346,
|
| 372 |
+
"eval_prec_macro": 0.8957261916294993,
|
| 373 |
+
"eval_prec_micro": 0.9022369511184007,
|
| 374 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 375 |
+
"eval_rec_at_8": 0.9788812785388128,
|
| 376 |
+
"eval_rec_macro": 0.8884178007276748,
|
| 377 |
+
"eval_rec_micro": 0.9029850746267908,
|
| 378 |
+
"eval_runtime": 3.2003,
|
| 379 |
+
"eval_samples_per_second": 273.728,
|
| 380 |
+
"eval_steps_per_second": 34.372,
|
| 381 |
+
"step": 288
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 13.0,
|
| 385 |
+
"grad_norm": 0.492089182138443,
|
| 386 |
+
"learning_rate": 7.500000000000001e-06,
|
| 387 |
+
"loss": 0.1215,
|
| 388 |
+
"step": 312
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 13.0,
|
| 392 |
+
"eval_acc_macro": 0.8126972127948302,
|
| 393 |
+
"eval_acc_micro": 0.8267363704256291,
|
| 394 |
+
"eval_auc_macro": 0.9893756572609064,
|
| 395 |
+
"eval_auc_micro": 0.9916252607583189,
|
| 396 |
+
"eval_f1_at_5": 0.42745954575018524,
|
| 397 |
+
"eval_f1_at_8": 0.2925113305987028,
|
| 398 |
+
"eval_f1_macro": 0.8933532206974574,
|
| 399 |
+
"eval_f1_micro": 0.9051512673752325,
|
| 400 |
+
"eval_loss": 0.10864967107772827,
|
| 401 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 402 |
+
"eval_prec_at_8": 0.17194634703196346,
|
| 403 |
+
"eval_prec_macro": 0.8891547780046246,
|
| 404 |
+
"eval_prec_micro": 0.892741935483799,
|
| 405 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 406 |
+
"eval_rec_at_8": 0.9788812785388128,
|
| 407 |
+
"eval_rec_macro": 0.9014409945020224,
|
| 408 |
+
"eval_rec_micro": 0.9179104477611179,
|
| 409 |
+
"eval_runtime": 3.3389,
|
| 410 |
+
"eval_samples_per_second": 262.36,
|
| 411 |
+
"eval_steps_per_second": 32.945,
|
| 412 |
+
"step": 312
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 14.0,
|
| 416 |
+
"grad_norm": 0.3663789927959442,
|
| 417 |
+
"learning_rate": 7.222222222222223e-06,
|
| 418 |
+
"loss": 0.1129,
|
| 419 |
+
"step": 336
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 14.0,
|
| 423 |
+
"eval_acc_macro": 0.8228971770857018,
|
| 424 |
+
"eval_acc_micro": 0.8329588014980649,
|
| 425 |
+
"eval_auc_macro": 0.9899641824346502,
|
| 426 |
+
"eval_auc_micro": 0.9921304716477787,
|
| 427 |
+
"eval_f1_at_5": 0.4281991387591175,
|
| 428 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 429 |
+
"eval_f1_macro": 0.9004325486077462,
|
| 430 |
+
"eval_f1_micro": 0.9088680016345804,
|
| 431 |
+
"eval_loss": 0.10234559327363968,
|
| 432 |
+
"eval_prec_at_5": 0.2742009132420091,
|
| 433 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 434 |
+
"eval_prec_macro": 0.890683762800314,
|
| 435 |
+
"eval_prec_micro": 0.8960515713133846,
|
| 436 |
+
"eval_rec_at_5": 0.9767884322678843,
|
| 437 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 438 |
+
"eval_rec_macro": 0.9120840636750266,
|
| 439 |
+
"eval_rec_micro": 0.9220563847428754,
|
| 440 |
+
"eval_runtime": 3.2169,
|
| 441 |
+
"eval_samples_per_second": 272.316,
|
| 442 |
+
"eval_steps_per_second": 34.195,
|
| 443 |
+
"step": 336
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 15.0,
|
| 447 |
+
"grad_norm": 0.37758299708366394,
|
| 448 |
+
"learning_rate": 6.944444444444445e-06,
|
| 449 |
+
"loss": 0.1046,
|
| 450 |
+
"step": 360
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 15.0,
|
| 454 |
+
"eval_acc_macro": 0.8302977681001992,
|
| 455 |
+
"eval_acc_micro": 0.8413059984813333,
|
| 456 |
+
"eval_auc_macro": 0.9901671140838944,
|
| 457 |
+
"eval_auc_micro": 0.9919789083809407,
|
| 458 |
+
"eval_f1_at_5": 0.42784762619628447,
|
| 459 |
+
"eval_f1_at_8": 0.2925113305987028,
|
| 460 |
+
"eval_f1_macro": 0.9052032341883497,
|
| 461 |
+
"eval_f1_micro": 0.9138144329896153,
|
| 462 |
+
"eval_loss": 0.09653711318969727,
|
| 463 |
+
"eval_prec_at_5": 0.273972602739726,
|
| 464 |
+
"eval_prec_at_8": 0.17194634703196346,
|
| 465 |
+
"eval_prec_macro": 0.9028285099698952,
|
| 466 |
+
"eval_prec_micro": 0.9089417555372511,
|
| 467 |
+
"eval_rec_at_5": 0.976027397260274,
|
| 468 |
+
"eval_rec_at_8": 0.9788812785388128,
|
| 469 |
+
"eval_rec_macro": 0.9099411755814296,
|
| 470 |
+
"eval_rec_micro": 0.9187396351574694,
|
| 471 |
+
"eval_runtime": 3.2472,
|
| 472 |
+
"eval_samples_per_second": 269.767,
|
| 473 |
+
"eval_steps_per_second": 33.875,
|
| 474 |
+
"step": 360
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 16.0,
|
| 478 |
+
"grad_norm": 0.4621961712837219,
|
| 479 |
+
"learning_rate": 6.666666666666667e-06,
|
| 480 |
+
"loss": 0.0979,
|
| 481 |
+
"step": 384
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 16.0,
|
| 485 |
+
"eval_acc_macro": 0.8273587016423724,
|
| 486 |
+
"eval_acc_micro": 0.8396299151888327,
|
| 487 |
+
"eval_auc_macro": 0.989802544326232,
|
| 488 |
+
"eval_auc_micro": 0.9918006143792109,
|
| 489 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 490 |
+
"eval_f1_at_8": 0.2920473967500829,
|
| 491 |
+
"eval_f1_macro": 0.9028525753108699,
|
| 492 |
+
"eval_f1_micro": 0.9128248113997558,
|
| 493 |
+
"eval_loss": 0.09215801954269409,
|
| 494 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 495 |
+
"eval_prec_at_8": 0.1716609589041096,
|
| 496 |
+
"eval_prec_macro": 0.9158518703241821,
|
| 497 |
+
"eval_prec_micro": 0.9228813559321252,
|
| 498 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 499 |
+
"eval_rec_at_8": 0.9777397260273972,
|
| 500 |
+
"eval_rec_macro": 0.8926981936861532,
|
| 501 |
+
"eval_rec_micro": 0.9029850746267908,
|
| 502 |
+
"eval_runtime": 3.233,
|
| 503 |
+
"eval_samples_per_second": 270.955,
|
| 504 |
+
"eval_steps_per_second": 34.024,
|
| 505 |
+
"step": 384
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 17.0,
|
| 509 |
+
"grad_norm": 0.4058220386505127,
|
| 510 |
+
"learning_rate": 6.3888888888888885e-06,
|
| 511 |
+
"loss": 0.092,
|
| 512 |
+
"step": 408
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 17.0,
|
| 516 |
+
"eval_acc_macro": 0.8347539263761662,
|
| 517 |
+
"eval_acc_micro": 0.8474708171205566,
|
| 518 |
+
"eval_auc_macro": 0.9899894858470043,
|
| 519 |
+
"eval_auc_micro": 0.9920498339308542,
|
| 520 |
+
"eval_f1_at_5": 0.4275326584009282,
|
| 521 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 522 |
+
"eval_f1_macro": 0.9071161229466114,
|
| 523 |
+
"eval_f1_micro": 0.9174389216511443,
|
| 524 |
+
"eval_loss": 0.08766299486160278,
|
| 525 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 526 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 527 |
+
"eval_prec_macro": 0.9264703324172373,
|
| 528 |
+
"eval_prec_micro": 0.9323630136985502,
|
| 529 |
+
"eval_rec_at_5": 0.9756468797564688,
|
| 530 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 531 |
+
"eval_rec_macro": 0.8904830078050127,
|
| 532 |
+
"eval_rec_micro": 0.9029850746267908,
|
| 533 |
+
"eval_runtime": 3.2006,
|
| 534 |
+
"eval_samples_per_second": 273.696,
|
| 535 |
+
"eval_steps_per_second": 34.368,
|
| 536 |
+
"step": 408
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 18.0,
|
| 540 |
+
"grad_norm": 0.4305408000946045,
|
| 541 |
+
"learning_rate": 6.111111111111112e-06,
|
| 542 |
+
"loss": 0.0863,
|
| 543 |
+
"step": 432
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 18.0,
|
| 547 |
+
"eval_acc_macro": 0.8361202794579444,
|
| 548 |
+
"eval_acc_micro": 0.8485316846985433,
|
| 549 |
+
"eval_auc_macro": 0.9904755972939142,
|
| 550 |
+
"eval_auc_micro": 0.9926707889023967,
|
| 551 |
+
"eval_f1_at_5": 0.4278476261962846,
|
| 552 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 553 |
+
"eval_f1_macro": 0.9082261738031537,
|
| 554 |
+
"eval_f1_micro": 0.9180602006688194,
|
| 555 |
+
"eval_loss": 0.08403860032558441,
|
| 556 |
+
"eval_prec_at_5": 0.27397260273972607,
|
| 557 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 558 |
+
"eval_prec_macro": 0.9194984843968411,
|
| 559 |
+
"eval_prec_micro": 0.9258010118043064,
|
| 560 |
+
"eval_rec_at_5": 0.976027397260274,
|
| 561 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 562 |
+
"eval_rec_macro": 0.8989752524534329,
|
| 563 |
+
"eval_rec_micro": 0.9104477611939543,
|
| 564 |
+
"eval_runtime": 3.2293,
|
| 565 |
+
"eval_samples_per_second": 271.27,
|
| 566 |
+
"eval_steps_per_second": 34.064,
|
| 567 |
+
"step": 432
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 19.0,
|
| 571 |
+
"grad_norm": 0.3943960964679718,
|
| 572 |
+
"learning_rate": 5.833333333333334e-06,
|
| 573 |
+
"loss": 0.0814,
|
| 574 |
+
"step": 456
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 19.0,
|
| 578 |
+
"eval_acc_macro": 0.839106311750719,
|
| 579 |
+
"eval_acc_micro": 0.8515684774291621,
|
| 580 |
+
"eval_auc_macro": 0.990108145855512,
|
| 581 |
+
"eval_auc_micro": 0.992045690666946,
|
| 582 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 583 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 584 |
+
"eval_f1_macro": 0.9099447887675026,
|
| 585 |
+
"eval_f1_micro": 0.9198347107437256,
|
| 586 |
+
"eval_loss": 0.08095283061265945,
|
| 587 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 588 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 589 |
+
"eval_prec_macro": 0.9089344385619288,
|
| 590 |
+
"eval_prec_micro": 0.9168039538714237,
|
| 591 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 592 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 593 |
+
"eval_rec_macro": 0.912673576698828,
|
| 594 |
+
"eval_rec_micro": 0.922885572139227,
|
| 595 |
+
"eval_runtime": 3.2592,
|
| 596 |
+
"eval_samples_per_second": 268.78,
|
| 597 |
+
"eval_steps_per_second": 33.751,
|
| 598 |
+
"step": 456
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 20.0,
|
| 602 |
+
"grad_norm": 0.40355971455574036,
|
| 603 |
+
"learning_rate": 5.555555555555557e-06,
|
| 604 |
+
"loss": 0.077,
|
| 605 |
+
"step": 480
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 20.0,
|
| 609 |
+
"eval_acc_macro": 0.841293209326155,
|
| 610 |
+
"eval_acc_micro": 0.8523335883702484,
|
| 611 |
+
"eval_auc_macro": 0.9901090394728352,
|
| 612 |
+
"eval_auc_micro": 0.9921565786655233,
|
| 613 |
+
"eval_f1_at_5": 0.42712632039330145,
|
| 614 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 615 |
+
"eval_f1_macro": 0.9118247332403943,
|
| 616 |
+
"eval_f1_micro": 0.9202808756711341,
|
| 617 |
+
"eval_loss": 0.07820397615432739,
|
| 618 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 619 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 620 |
+
"eval_prec_macro": 0.9128408681170823,
|
| 621 |
+
"eval_prec_micro": 0.9168724279834636,
|
| 622 |
+
"eval_rec_at_5": 0.9743150684931506,
|
| 623 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 624 |
+
"eval_rec_macro": 0.913025214091804,
|
| 625 |
+
"eval_rec_micro": 0.9237147595355785,
|
| 626 |
+
"eval_runtime": 3.335,
|
| 627 |
+
"eval_samples_per_second": 262.673,
|
| 628 |
+
"eval_steps_per_second": 32.984,
|
| 629 |
+
"step": 480
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 21.0,
|
| 633 |
+
"grad_norm": 0.4673054814338684,
|
| 634 |
+
"learning_rate": 5.2777777777777785e-06,
|
| 635 |
+
"loss": 0.0732,
|
| 636 |
+
"step": 504
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 21.0,
|
| 640 |
+
"eval_acc_macro": 0.8435745126219377,
|
| 641 |
+
"eval_acc_micro": 0.8546017014693847,
|
| 642 |
+
"eval_auc_macro": 0.990212670611912,
|
| 643 |
+
"eval_auc_micro": 0.9923085874446093,
|
| 644 |
+
"eval_f1_at_5": 0.4278476261962846,
|
| 645 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 646 |
+
"eval_f1_macro": 0.9131099703856408,
|
| 647 |
+
"eval_f1_micro": 0.9216013344452942,
|
| 648 |
+
"eval_loss": 0.07635616511106491,
|
| 649 |
+
"eval_prec_at_5": 0.27397260273972607,
|
| 650 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 651 |
+
"eval_prec_macro": 0.9220960220647866,
|
| 652 |
+
"eval_prec_micro": 0.9270134228187141,
|
| 653 |
+
"eval_rec_at_5": 0.976027397260274,
|
| 654 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 655 |
+
"eval_rec_macro": 0.9063401498441861,
|
| 656 |
+
"eval_rec_micro": 0.9162520729684148,
|
| 657 |
+
"eval_runtime": 3.335,
|
| 658 |
+
"eval_samples_per_second": 262.666,
|
| 659 |
+
"eval_steps_per_second": 32.983,
|
| 660 |
+
"step": 504
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 22.0,
|
| 664 |
+
"grad_norm": 0.4005274772644043,
|
| 665 |
+
"learning_rate": 5e-06,
|
| 666 |
+
"loss": 0.0693,
|
| 667 |
+
"step": 528
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 22.0,
|
| 671 |
+
"eval_acc_macro": 0.8455313120389342,
|
| 672 |
+
"eval_acc_micro": 0.8565950920244741,
|
| 673 |
+
"eval_auc_macro": 0.9901732890859408,
|
| 674 |
+
"eval_auc_micro": 0.9921753792823973,
|
| 675 |
+
"eval_f1_at_5": 0.42745954575018524,
|
| 676 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 677 |
+
"eval_f1_macro": 0.9143994666596676,
|
| 678 |
+
"eval_f1_micro": 0.9227591904171067,
|
| 679 |
+
"eval_loss": 0.07342522591352463,
|
| 680 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 681 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 682 |
+
"eval_prec_macro": 0.9139300975502155,
|
| 683 |
+
"eval_prec_micro": 0.9193415637859326,
|
| 684 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 685 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 686 |
+
"eval_rec_macro": 0.9161172406072337,
|
| 687 |
+
"eval_rec_micro": 0.9262023217246329,
|
| 688 |
+
"eval_runtime": 3.2711,
|
| 689 |
+
"eval_samples_per_second": 267.802,
|
| 690 |
+
"eval_steps_per_second": 33.628,
|
| 691 |
+
"step": 528
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 23.0,
|
| 695 |
+
"grad_norm": 0.45327135920524597,
|
| 696 |
+
"learning_rate": 4.722222222222222e-06,
|
| 697 |
+
"loss": 0.0661,
|
| 698 |
+
"step": 552
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 23.0,
|
| 702 |
+
"eval_acc_macro": 0.8467210056386675,
|
| 703 |
+
"eval_acc_micro": 0.8575827559660617,
|
| 704 |
+
"eval_auc_macro": 0.990201788357871,
|
| 705 |
+
"eval_auc_micro": 0.9922312465183218,
|
| 706 |
+
"eval_f1_at_5": 0.42745954575018524,
|
| 707 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 708 |
+
"eval_f1_macro": 0.9148197613459009,
|
| 709 |
+
"eval_f1_micro": 0.9233319519269851,
|
| 710 |
+
"eval_loss": 0.07164816558361053,
|
| 711 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 712 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 713 |
+
"eval_prec_macro": 0.9172647910903643,
|
| 714 |
+
"eval_prec_micro": 0.9229494614746543,
|
| 715 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 716 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 717 |
+
"eval_rec_macro": 0.9139228248292297,
|
| 718 |
+
"eval_rec_micro": 0.9237147595355785,
|
| 719 |
+
"eval_runtime": 3.2794,
|
| 720 |
+
"eval_samples_per_second": 267.12,
|
| 721 |
+
"eval_steps_per_second": 33.542,
|
| 722 |
+
"step": 552
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 24.0,
|
| 726 |
+
"grad_norm": 0.44867074489593506,
|
| 727 |
+
"learning_rate": 4.444444444444444e-06,
|
| 728 |
+
"loss": 0.0626,
|
| 729 |
+
"step": 576
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 24.0,
|
| 733 |
+
"eval_acc_macro": 0.8443607544183838,
|
| 734 |
+
"eval_acc_micro": 0.8557098765431438,
|
| 735 |
+
"eval_auc_macro": 0.9902370599717555,
|
| 736 |
+
"eval_auc_micro": 0.992274817616196,
|
| 737 |
+
"eval_f1_at_5": 0.42712632039330145,
|
| 738 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 739 |
+
"eval_f1_macro": 0.9135513991863947,
|
| 740 |
+
"eval_f1_micro": 0.9222453222452455,
|
| 741 |
+
"eval_loss": 0.07073331624269485,
|
| 742 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 743 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 744 |
+
"eval_prec_macro": 0.9185147069172412,
|
| 745 |
+
"eval_prec_micro": 0.9249374478731505,
|
| 746 |
+
"eval_rec_at_5": 0.9743150684931506,
|
| 747 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 748 |
+
"eval_rec_macro": 0.9094868029221347,
|
| 749 |
+
"eval_rec_micro": 0.9195688225538209,
|
| 750 |
+
"eval_runtime": 3.2386,
|
| 751 |
+
"eval_samples_per_second": 270.487,
|
| 752 |
+
"eval_steps_per_second": 33.965,
|
| 753 |
+
"step": 576
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 25.0,
|
| 757 |
+
"grad_norm": 0.509304940700531,
|
| 758 |
+
"learning_rate": 4.166666666666667e-06,
|
| 759 |
+
"loss": 0.06,
|
| 760 |
+
"step": 600
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 25.0,
|
| 764 |
+
"eval_acc_macro": 0.8405381328154341,
|
| 765 |
+
"eval_acc_micro": 0.8539238539237876,
|
| 766 |
+
"eval_auc_macro": 0.9899142306185617,
|
| 767 |
+
"eval_auc_micro": 0.9919983327149624,
|
| 768 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 769 |
+
"eval_f1_at_8": 0.2920473967500829,
|
| 770 |
+
"eval_f1_macro": 0.91134605843459,
|
| 771 |
+
"eval_f1_micro": 0.9212070410728481,
|
| 772 |
+
"eval_loss": 0.06922697275876999,
|
| 773 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 774 |
+
"eval_prec_at_8": 0.1716609589041096,
|
| 775 |
+
"eval_prec_macro": 0.9243865602122843,
|
| 776 |
+
"eval_prec_micro": 0.9313559322033109,
|
| 777 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 778 |
+
"eval_rec_at_8": 0.9777397260273972,
|
| 779 |
+
"eval_rec_macro": 0.8997359146591782,
|
| 780 |
+
"eval_rec_micro": 0.9112769485903058,
|
| 781 |
+
"eval_runtime": 3.3027,
|
| 782 |
+
"eval_samples_per_second": 265.24,
|
| 783 |
+
"eval_steps_per_second": 33.306,
|
| 784 |
+
"step": 600
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 26.0,
|
| 788 |
+
"grad_norm": 0.5278392434120178,
|
| 789 |
+
"learning_rate": 3.88888888888889e-06,
|
| 790 |
+
"loss": 0.0575,
|
| 791 |
+
"step": 624
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 26.0,
|
| 795 |
+
"eval_acc_macro": 0.8460307636339017,
|
| 796 |
+
"eval_acc_micro": 0.8563664596272627,
|
| 797 |
+
"eval_auc_macro": 0.9901957256905628,
|
| 798 |
+
"eval_auc_micro": 0.9919254469111567,
|
| 799 |
+
"eval_f1_at_5": 0.42718114191994605,
|
| 800 |
+
"eval_f1_at_8": 0.2925113305987028,
|
| 801 |
+
"eval_f1_macro": 0.9146326110180735,
|
| 802 |
+
"eval_f1_micro": 0.9226265161019722,
|
| 803 |
+
"eval_loss": 0.06812591105699539,
|
| 804 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 805 |
+
"eval_prec_at_8": 0.17194634703196346,
|
| 806 |
+
"eval_prec_macro": 0.9259544901379119,
|
| 807 |
+
"eval_prec_micro": 0.9308016877636345,
|
| 808 |
+
"eval_rec_at_5": 0.9748858447488584,
|
| 809 |
+
"eval_rec_at_8": 0.9788812785388128,
|
| 810 |
+
"eval_rec_macro": 0.9052694584097175,
|
| 811 |
+
"eval_rec_micro": 0.9145936981757119,
|
| 812 |
+
"eval_runtime": 3.238,
|
| 813 |
+
"eval_samples_per_second": 270.536,
|
| 814 |
+
"eval_steps_per_second": 33.971,
|
| 815 |
+
"step": 624
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 27.0,
|
| 819 |
+
"grad_norm": 0.46283113956451416,
|
| 820 |
+
"learning_rate": 3.6111111111111115e-06,
|
| 821 |
+
"loss": 0.0547,
|
| 822 |
+
"step": 648
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 27.0,
|
| 826 |
+
"eval_acc_macro": 0.8437978381816325,
|
| 827 |
+
"eval_acc_micro": 0.8569194683345694,
|
| 828 |
+
"eval_auc_macro": 0.9900672235373739,
|
| 829 |
+
"eval_auc_micro": 0.9922836387587103,
|
| 830 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 831 |
+
"eval_f1_at_8": 0.2922793659426448,
|
| 832 |
+
"eval_f1_macro": 0.9133228707359765,
|
| 833 |
+
"eval_f1_micro": 0.9229473684209749,
|
| 834 |
+
"eval_loss": 0.06721309572458267,
|
| 835 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 836 |
+
"eval_prec_at_8": 0.17180365296803654,
|
| 837 |
+
"eval_prec_macro": 0.9302491335296357,
|
| 838 |
+
"eval_prec_micro": 0.9375534644994921,
|
| 839 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 840 |
+
"eval_rec_at_8": 0.978310502283105,
|
| 841 |
+
"eval_rec_macro": 0.8981640840687014,
|
| 842 |
+
"eval_rec_micro": 0.9087893864012513,
|
| 843 |
+
"eval_runtime": 3.2966,
|
| 844 |
+
"eval_samples_per_second": 265.728,
|
| 845 |
+
"eval_steps_per_second": 33.368,
|
| 846 |
+
"step": 648
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 28.0,
|
| 850 |
+
"grad_norm": 0.5683040022850037,
|
| 851 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 852 |
+
"loss": 0.0523,
|
| 853 |
+
"step": 672
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 28.0,
|
| 857 |
+
"eval_acc_macro": 0.847085539786285,
|
| 858 |
+
"eval_acc_micro": 0.8558139534883057,
|
| 859 |
+
"eval_auc_macro": 0.9896638187547525,
|
| 860 |
+
"eval_auc_micro": 0.99160271783856,
|
| 861 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 862 |
+
"eval_f1_at_8": 0.2915834447216144,
|
| 863 |
+
"eval_f1_macro": 0.9152884545582967,
|
| 864 |
+
"eval_f1_micro": 0.9223057644109505,
|
| 865 |
+
"eval_loss": 0.06634338945150375,
|
| 866 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 867 |
+
"eval_prec_at_8": 0.1713755707762557,
|
| 868 |
+
"eval_prec_macro": 0.9276311524633641,
|
| 869 |
+
"eval_prec_micro": 0.929292929292851,
|
| 870 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 871 |
+
"eval_rec_at_8": 0.9765981735159818,
|
| 872 |
+
"eval_rec_macro": 0.9050852833749707,
|
| 873 |
+
"eval_rec_micro": 0.9154228855720634,
|
| 874 |
+
"eval_runtime": 3.3379,
|
| 875 |
+
"eval_samples_per_second": 262.442,
|
| 876 |
+
"eval_steps_per_second": 32.955,
|
| 877 |
+
"step": 672
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 29.0,
|
| 881 |
+
"grad_norm": 0.5064497590065002,
|
| 882 |
+
"learning_rate": 3.055555555555556e-06,
|
| 883 |
+
"loss": 0.0507,
|
| 884 |
+
"step": 696
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 29.0,
|
| 888 |
+
"eval_acc_macro": 0.8424595332709467,
|
| 889 |
+
"eval_acc_micro": 0.8554687499999332,
|
| 890 |
+
"eval_auc_macro": 0.9903800574812361,
|
| 891 |
+
"eval_auc_micro": 0.9924352911279979,
|
| 892 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 893 |
+
"eval_f1_at_8": 0.2918154230125642,
|
| 894 |
+
"eval_f1_macro": 0.9125139257396052,
|
| 895 |
+
"eval_f1_micro": 0.9221052631578169,
|
| 896 |
+
"eval_loss": 0.06477358192205429,
|
| 897 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 898 |
+
"eval_prec_at_8": 0.17151826484018265,
|
| 899 |
+
"eval_prec_macro": 0.9271042210984995,
|
| 900 |
+
"eval_prec_micro": 0.9366980325063355,
|
| 901 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 902 |
+
"eval_rec_at_8": 0.9771689497716894,
|
| 903 |
+
"eval_rec_macro": 0.8991646180232234,
|
| 904 |
+
"eval_rec_micro": 0.9079601990048998,
|
| 905 |
+
"eval_runtime": 3.3033,
|
| 906 |
+
"eval_samples_per_second": 265.185,
|
| 907 |
+
"eval_steps_per_second": 33.3,
|
| 908 |
+
"step": 696
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 30.0,
|
| 912 |
+
"grad_norm": 0.5206765532493591,
|
| 913 |
+
"learning_rate": 2.7777777777777783e-06,
|
| 914 |
+
"loss": 0.0484,
|
| 915 |
+
"step": 720
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 30.0,
|
| 919 |
+
"eval_acc_macro": 0.8482879664337308,
|
| 920 |
+
"eval_acc_micro": 0.8564920273347869,
|
| 921 |
+
"eval_auc_macro": 0.9902796769975294,
|
| 922 |
+
"eval_auc_micro": 0.9923823642729116,
|
| 923 |
+
"eval_f1_at_5": 0.42712632039330145,
|
| 924 |
+
"eval_f1_at_8": 0.2925113305987028,
|
| 925 |
+
"eval_f1_macro": 0.9155214146126874,
|
| 926 |
+
"eval_f1_micro": 0.9226993865029919,
|
| 927 |
+
"eval_loss": 0.06509387493133545,
|
| 928 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 929 |
+
"eval_prec_at_8": 0.17194634703196346,
|
| 930 |
+
"eval_prec_macro": 0.9025052330017584,
|
| 931 |
+
"eval_prec_micro": 0.9104116222759555,
|
| 932 |
+
"eval_rec_at_5": 0.9743150684931506,
|
| 933 |
+
"eval_rec_at_8": 0.9788812785388128,
|
| 934 |
+
"eval_rec_macro": 0.9295157618718927,
|
| 935 |
+
"eval_rec_micro": 0.9353233830844995,
|
| 936 |
+
"eval_runtime": 3.2722,
|
| 937 |
+
"eval_samples_per_second": 267.709,
|
| 938 |
+
"eval_steps_per_second": 33.616,
|
| 939 |
+
"step": 720
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 31.0,
|
| 943 |
+
"grad_norm": 0.5268795490264893,
|
| 944 |
+
"learning_rate": 2.5e-06,
|
| 945 |
+
"loss": 0.0456,
|
| 946 |
+
"step": 744
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 31.0,
|
| 950 |
+
"eval_acc_macro": 0.847691548444017,
|
| 951 |
+
"eval_acc_micro": 0.8558421851289183,
|
| 952 |
+
"eval_auc_macro": 0.9898163244235566,
|
| 953 |
+
"eval_auc_micro": 0.9917076805242361,
|
| 954 |
+
"eval_f1_at_5": 0.42751438858977425,
|
| 955 |
+
"eval_f1_at_8": 0.2918154230125642,
|
| 956 |
+
"eval_f1_macro": 0.9153671638721068,
|
| 957 |
+
"eval_f1_micro": 0.9223221586262532,
|
| 958 |
+
"eval_loss": 0.0645858645439148,
|
| 959 |
+
"eval_prec_at_5": 0.27374429223744295,
|
| 960 |
+
"eval_prec_at_8": 0.17151826484018265,
|
| 961 |
+
"eval_prec_macro": 0.9015493618181559,
|
| 962 |
+
"eval_prec_micro": 0.9096774193547653,
|
| 963 |
+
"eval_rec_at_5": 0.9754566210045662,
|
| 964 |
+
"eval_rec_at_8": 0.9771689497716894,
|
| 965 |
+
"eval_rec_macro": 0.9300094432719982,
|
| 966 |
+
"eval_rec_micro": 0.9353233830844995,
|
| 967 |
+
"eval_runtime": 3.3402,
|
| 968 |
+
"eval_samples_per_second": 262.263,
|
| 969 |
+
"eval_steps_per_second": 32.933,
|
| 970 |
+
"step": 744
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 32.0,
|
| 974 |
+
"grad_norm": 0.4488939046859741,
|
| 975 |
+
"learning_rate": 2.222222222222222e-06,
|
| 976 |
+
"loss": 0.0446,
|
| 977 |
+
"step": 768
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 32.0,
|
| 981 |
+
"eval_acc_macro": 0.8474584233282965,
|
| 982 |
+
"eval_acc_micro": 0.8549848942597541,
|
| 983 |
+
"eval_auc_macro": 0.9896580840322796,
|
| 984 |
+
"eval_auc_micro": 0.9914708462130924,
|
| 985 |
+
"eval_f1_at_5": 0.42712632039330145,
|
| 986 |
+
"eval_f1_at_8": 0.29135146186873845,
|
| 987 |
+
"eval_f1_macro": 0.9153561345178298,
|
| 988 |
+
"eval_f1_micro": 0.9218241042344526,
|
| 989 |
+
"eval_loss": 0.06430496275424957,
|
| 990 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 991 |
+
"eval_prec_at_8": 0.17123287671232876,
|
| 992 |
+
"eval_prec_macro": 0.8975121306620646,
|
| 993 |
+
"eval_prec_micro": 0.9055999999999276,
|
| 994 |
+
"eval_rec_at_5": 0.9743150684931506,
|
| 995 |
+
"eval_rec_at_8": 0.976027397260274,
|
| 996 |
+
"eval_rec_macro": 0.9345814562185115,
|
| 997 |
+
"eval_rec_micro": 0.9386401326699055,
|
| 998 |
+
"eval_runtime": 3.2443,
|
| 999 |
+
"eval_samples_per_second": 270.009,
|
| 1000 |
+
"eval_steps_per_second": 33.905,
|
| 1001 |
+
"step": 768
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 33.0,
|
| 1005 |
+
"grad_norm": 0.5452151298522949,
|
| 1006 |
+
"learning_rate": 1.944444444444445e-06,
|
| 1007 |
+
"loss": 0.0425,
|
| 1008 |
+
"step": 792
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 33.0,
|
| 1012 |
+
"eval_acc_macro": 0.8512273558913187,
|
| 1013 |
+
"eval_acc_micro": 0.8594224924011504,
|
| 1014 |
+
"eval_auc_macro": 0.9896851270268344,
|
| 1015 |
+
"eval_auc_micro": 0.9917295997268475,
|
| 1016 |
+
"eval_f1_at_5": 0.42712632039330145,
|
| 1017 |
+
"eval_f1_at_8": 0.2918154230125642,
|
| 1018 |
+
"eval_f1_macro": 0.9175407198197038,
|
| 1019 |
+
"eval_f1_micro": 0.9243972210869698,
|
| 1020 |
+
"eval_loss": 0.06364509463310242,
|
| 1021 |
+
"eval_prec_at_5": 0.27351598173515984,
|
| 1022 |
+
"eval_prec_at_8": 0.17151826484018265,
|
| 1023 |
+
"eval_prec_macro": 0.9039169093915104,
|
| 1024 |
+
"eval_prec_micro": 0.9113618049958975,
|
| 1025 |
+
"eval_rec_at_5": 0.9743150684931506,
|
| 1026 |
+
"eval_rec_at_8": 0.9771689497716894,
|
| 1027 |
+
"eval_rec_macro": 0.9318495447404759,
|
| 1028 |
+
"eval_rec_micro": 0.937810945273554,
|
| 1029 |
+
"eval_runtime": 3.2885,
|
| 1030 |
+
"eval_samples_per_second": 266.382,
|
| 1031 |
+
"eval_steps_per_second": 33.45,
|
| 1032 |
+
"step": 792
|
| 1033 |
+
}
|
| 1034 |
+
],
|
| 1035 |
+
"logging_steps": 500,
|
| 1036 |
+
"max_steps": 960,
|
| 1037 |
+
"num_input_tokens_seen": 0,
|
| 1038 |
+
"num_train_epochs": 40,
|
| 1039 |
+
"save_steps": 500,
|
| 1040 |
+
"total_flos": 1.332040372236288e+16,
|
| 1041 |
+
"train_batch_size": 32,
|
| 1042 |
+
"trial_name": null,
|
| 1043 |
+
"trial_params": null
|
| 1044 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12f80262b2e3bdb2ae4247f0dffc9ed3250ed0506935e0e564a5db2a7e4491d4
|
| 3 |
+
size 5048
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|