Upload folder using huggingface_hub
Browse files- training_checkpoints/checkpoint-4704/config.json +40 -0
- training_checkpoints/checkpoint-4704/merges.txt +0 -0
- training_checkpoints/checkpoint-4704/model.safetensors +3 -0
- training_checkpoints/checkpoint-4704/optimizer.pt +3 -0
- training_checkpoints/checkpoint-4704/rng_state.pth +3 -0
- training_checkpoints/checkpoint-4704/scheduler.pt +3 -0
- training_checkpoints/checkpoint-4704/special_tokens_map.json +15 -0
- training_checkpoints/checkpoint-4704/tokenizer.json +0 -0
- training_checkpoints/checkpoint-4704/tokenizer_config.json +57 -0
- training_checkpoints/checkpoint-4704/trainer_state.json +546 -0
- training_checkpoints/checkpoint-4704/training_args.bin +3 -0
- training_checkpoints/checkpoint-4704/vocab.json +0 -0
training_checkpoints/checkpoint-4704/config.json
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "FacebookAI/roberta-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 |
+
"id2label": {
|
| 14 |
+
"0": "positive",
|
| 15 |
+
"1": "negative",
|
| 16 |
+
"2": "neutral",
|
| 17 |
+
"3": "rel_by_party"
|
| 18 |
+
},
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 3072,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"negative": 1,
|
| 23 |
+
"neutral": 2,
|
| 24 |
+
"positive": 0,
|
| 25 |
+
"rel_by_party": 3
|
| 26 |
+
},
|
| 27 |
+
"layer_norm_eps": 1e-05,
|
| 28 |
+
"max_position_embeddings": 514,
|
| 29 |
+
"model_type": "roberta",
|
| 30 |
+
"num_attention_heads": 12,
|
| 31 |
+
"num_hidden_layers": 12,
|
| 32 |
+
"pad_token_id": 1,
|
| 33 |
+
"position_embedding_type": "absolute",
|
| 34 |
+
"problem_type": "single_label_classification",
|
| 35 |
+
"torch_dtype": "float32",
|
| 36 |
+
"transformers_version": "4.40.2",
|
| 37 |
+
"type_vocab_size": 1,
|
| 38 |
+
"use_cache": true,
|
| 39 |
+
"vocab_size": 50265
|
| 40 |
+
}
|
training_checkpoints/checkpoint-4704/merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
training_checkpoints/checkpoint-4704/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b4649d09eeefd0a31cd5e311000d912bfcbf232cc37574aa79bc1d44cd462ac
|
| 3 |
+
size 498618976
|
training_checkpoints/checkpoint-4704/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e5ef26858bfe65dc76c7c1418b5c6d056fff9840d800a3b4668d630a06107d1
|
| 3 |
+
size 997357818
|
training_checkpoints/checkpoint-4704/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:501ba3dbdb6e47563e860a8261dda92259b82573721aaa97a10c57861840937c
|
| 3 |
+
size 14244
|
training_checkpoints/checkpoint-4704/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33afda840378640c5abdb5d9753703bb18c86c79fe88558db9715cc64862d7d9
|
| 3 |
+
size 1064
|
training_checkpoints/checkpoint-4704/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 |
+
}
|
training_checkpoints/checkpoint-4704/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
training_checkpoints/checkpoint-4704/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 |
+
}
|
training_checkpoints/checkpoint-4704/trainer_state.json
ADDED
|
@@ -0,0 +1,546 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6406636359280463,
|
| 3 |
+
"best_model_checkpoint": "case-analysis-roberta-base/checkpoint-4704",
|
| 4 |
+
"epoch": 21.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4704,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.7527839643652561,
|
| 14 |
+
"eval_f1_macro": 0.5389063030773956,
|
| 15 |
+
"eval_f1_micro": 0.7527839643652561,
|
| 16 |
+
"eval_f1_weighted": 0.7300520074128541,
|
| 17 |
+
"eval_loss": 0.8771109580993652,
|
| 18 |
+
"eval_macro_fpr": 0.13144039552419665,
|
| 19 |
+
"eval_macro_sensitivity": 0.5401895401895401,
|
| 20 |
+
"eval_macro_specificity": 0.8833499134100214,
|
| 21 |
+
"eval_precision": 0.7119932520850578,
|
| 22 |
+
"eval_precision_macro": 0.5403637213518124,
|
| 23 |
+
"eval_recall": 0.7527839643652561,
|
| 24 |
+
"eval_recall_macro": 0.5401895401895401,
|
| 25 |
+
"eval_runtime": 6.9264,
|
| 26 |
+
"eval_samples_per_second": 64.825,
|
| 27 |
+
"eval_steps_per_second": 8.229,
|
| 28 |
+
"eval_weighted_fpr": 0.09866666666666667,
|
| 29 |
+
"eval_weighted_sensitivity": 0.7527839643652561,
|
| 30 |
+
"eval_weighted_specificity": 0.7806156892748298,
|
| 31 |
+
"step": 224
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 2.0,
|
| 35 |
+
"eval_accuracy": 0.7728285077951003,
|
| 36 |
+
"eval_f1_macro": 0.5712387529890991,
|
| 37 |
+
"eval_f1_micro": 0.7728285077951002,
|
| 38 |
+
"eval_f1_weighted": 0.7555182045354769,
|
| 39 |
+
"eval_loss": 0.7935857176780701,
|
| 40 |
+
"eval_macro_fpr": 0.10804320901264429,
|
| 41 |
+
"eval_macro_sensitivity": 0.59374376286141,
|
| 42 |
+
"eval_macro_specificity": 0.9046495531497725,
|
| 43 |
+
"eval_precision": 0.7420199260104934,
|
| 44 |
+
"eval_precision_macro": 0.5529178338001868,
|
| 45 |
+
"eval_recall": 0.7728285077951003,
|
| 46 |
+
"eval_recall_macro": 0.59374376286141,
|
| 47 |
+
"eval_runtime": 33.5269,
|
| 48 |
+
"eval_samples_per_second": 13.392,
|
| 49 |
+
"eval_steps_per_second": 1.7,
|
| 50 |
+
"eval_weighted_fpr": 0.08923884514435695,
|
| 51 |
+
"eval_weighted_sensitivity": 0.7728285077951003,
|
| 52 |
+
"eval_weighted_specificity": 0.8457697048039892,
|
| 53 |
+
"step": 448
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 2.232142857142857,
|
| 57 |
+
"grad_norm": 10.557605743408203,
|
| 58 |
+
"learning_rate": 4.631696428571429e-05,
|
| 59 |
+
"loss": 0.8855,
|
| 60 |
+
"step": 500
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 3.0,
|
| 64 |
+
"eval_accuracy": 0.7305122494432071,
|
| 65 |
+
"eval_f1_macro": 0.5313357935186093,
|
| 66 |
+
"eval_f1_micro": 0.730512249443207,
|
| 67 |
+
"eval_f1_weighted": 0.7208063173961943,
|
| 68 |
+
"eval_loss": 0.8126917481422424,
|
| 69 |
+
"eval_macro_fpr": 0.1288106780460812,
|
| 70 |
+
"eval_macro_sensitivity": 0.5599513761278467,
|
| 71 |
+
"eval_macro_specificity": 0.8878517309481255,
|
| 72 |
+
"eval_precision": 0.7320572933103704,
|
| 73 |
+
"eval_precision_macro": 0.5335863697705803,
|
| 74 |
+
"eval_recall": 0.7305122494432071,
|
| 75 |
+
"eval_recall_macro": 0.5599513761278467,
|
| 76 |
+
"eval_runtime": 35.5026,
|
| 77 |
+
"eval_samples_per_second": 12.647,
|
| 78 |
+
"eval_steps_per_second": 1.606,
|
| 79 |
+
"eval_weighted_fpr": 0.10950226244343891,
|
| 80 |
+
"eval_weighted_sensitivity": 0.7305122494432071,
|
| 81 |
+
"eval_weighted_specificity": 0.8208946743492954,
|
| 82 |
+
"step": 672
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 4.0,
|
| 86 |
+
"eval_accuracy": 0.779510022271715,
|
| 87 |
+
"eval_f1_macro": 0.5605264699901076,
|
| 88 |
+
"eval_f1_micro": 0.779510022271715,
|
| 89 |
+
"eval_f1_weighted": 0.756140232373208,
|
| 90 |
+
"eval_loss": 1.0185617208480835,
|
| 91 |
+
"eval_macro_fpr": 0.11840141313918778,
|
| 92 |
+
"eval_macro_sensitivity": 0.565439197792139,
|
| 93 |
+
"eval_macro_specificity": 0.8949862362922303,
|
| 94 |
+
"eval_precision": 0.7503024658142448,
|
| 95 |
+
"eval_precision_macro": 0.5721560130010834,
|
| 96 |
+
"eval_recall": 0.779510022271715,
|
| 97 |
+
"eval_recall_macro": 0.565439197792139,
|
| 98 |
+
"eval_runtime": 42.1679,
|
| 99 |
+
"eval_samples_per_second": 10.648,
|
| 100 |
+
"eval_steps_per_second": 1.352,
|
| 101 |
+
"eval_weighted_fpr": 0.08616187989556136,
|
| 102 |
+
"eval_weighted_sensitivity": 0.779510022271715,
|
| 103 |
+
"eval_weighted_specificity": 0.800434922897206,
|
| 104 |
+
"step": 896
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 4.464285714285714,
|
| 108 |
+
"grad_norm": 4.521566867828369,
|
| 109 |
+
"learning_rate": 4.260416666666667e-05,
|
| 110 |
+
"loss": 0.5551,
|
| 111 |
+
"step": 1000
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 5.0,
|
| 115 |
+
"eval_accuracy": 0.8084632516703786,
|
| 116 |
+
"eval_f1_macro": 0.5892045454545454,
|
| 117 |
+
"eval_f1_micro": 0.8084632516703786,
|
| 118 |
+
"eval_f1_weighted": 0.7866926503340758,
|
| 119 |
+
"eval_loss": 0.7591320276260376,
|
| 120 |
+
"eval_macro_fpr": 0.09875084526504571,
|
| 121 |
+
"eval_macro_sensitivity": 0.5962627065568242,
|
| 122 |
+
"eval_macro_specificity": 0.91148202322346,
|
| 123 |
+
"eval_precision": 0.7673786757699211,
|
| 124 |
+
"eval_precision_macro": 0.583308107932168,
|
| 125 |
+
"eval_recall": 0.8084632516703786,
|
| 126 |
+
"eval_recall_macro": 0.5962627065568242,
|
| 127 |
+
"eval_runtime": 42.7116,
|
| 128 |
+
"eval_samples_per_second": 10.512,
|
| 129 |
+
"eval_steps_per_second": 1.335,
|
| 130 |
+
"eval_weighted_fpr": 0.07319148936170213,
|
| 131 |
+
"eval_weighted_sensitivity": 0.8084632516703786,
|
| 132 |
+
"eval_weighted_specificity": 0.8374648412234613,
|
| 133 |
+
"step": 1120
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 6.0,
|
| 137 |
+
"eval_accuracy": 0.8173719376391982,
|
| 138 |
+
"eval_f1_macro": 0.603048355342882,
|
| 139 |
+
"eval_f1_micro": 0.8173719376391981,
|
| 140 |
+
"eval_f1_weighted": 0.7967392787416411,
|
| 141 |
+
"eval_loss": 0.9521613717079163,
|
| 142 |
+
"eval_macro_fpr": 0.09667046583838863,
|
| 143 |
+
"eval_macro_sensitivity": 0.5987868120221062,
|
| 144 |
+
"eval_macro_specificity": 0.9117780208671373,
|
| 145 |
+
"eval_precision": 0.7816269250029019,
|
| 146 |
+
"eval_precision_macro": 0.6116760996647348,
|
| 147 |
+
"eval_recall": 0.8173719376391982,
|
| 148 |
+
"eval_recall_macro": 0.5987868120221062,
|
| 149 |
+
"eval_runtime": 40.3104,
|
| 150 |
+
"eval_samples_per_second": 11.139,
|
| 151 |
+
"eval_steps_per_second": 1.414,
|
| 152 |
+
"eval_weighted_fpr": 0.06931530008453085,
|
| 153 |
+
"eval_weighted_sensitivity": 0.8173719376391982,
|
| 154 |
+
"eval_weighted_specificity": 0.8297401458293509,
|
| 155 |
+
"step": 1344
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 6.696428571428571,
|
| 159 |
+
"grad_norm": 0.38720017671585083,
|
| 160 |
+
"learning_rate": 3.888392857142857e-05,
|
| 161 |
+
"loss": 0.386,
|
| 162 |
+
"step": 1500
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 7.0,
|
| 166 |
+
"eval_accuracy": 0.7706013363028953,
|
| 167 |
+
"eval_f1_macro": 0.5781817655906514,
|
| 168 |
+
"eval_f1_micro": 0.7706013363028953,
|
| 169 |
+
"eval_f1_weighted": 0.7656335044104077,
|
| 170 |
+
"eval_loss": 1.0569394826889038,
|
| 171 |
+
"eval_macro_fpr": 0.10894725594931005,
|
| 172 |
+
"eval_macro_sensitivity": 0.5857963578551814,
|
| 173 |
+
"eval_macro_specificity": 0.9056966184410946,
|
| 174 |
+
"eval_precision": 0.7610212023939332,
|
| 175 |
+
"eval_precision_macro": 0.5709751150784568,
|
| 176 |
+
"eval_recall": 0.7706013363028953,
|
| 177 |
+
"eval_recall_macro": 0.5857963578551814,
|
| 178 |
+
"eval_runtime": 36.0908,
|
| 179 |
+
"eval_samples_per_second": 12.441,
|
| 180 |
+
"eval_steps_per_second": 1.579,
|
| 181 |
+
"eval_weighted_fpr": 0.09027169149868536,
|
| 182 |
+
"eval_weighted_sensitivity": 0.7706013363028953,
|
| 183 |
+
"eval_weighted_specificity": 0.852185137461483,
|
| 184 |
+
"step": 1568
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 8.0,
|
| 188 |
+
"eval_accuracy": 0.7572383073496659,
|
| 189 |
+
"eval_f1_macro": 0.6161842736352976,
|
| 190 |
+
"eval_f1_micro": 0.7572383073496659,
|
| 191 |
+
"eval_f1_weighted": 0.7715266654127986,
|
| 192 |
+
"eval_loss": 1.1956746578216553,
|
| 193 |
+
"eval_macro_fpr": 0.10518103765035827,
|
| 194 |
+
"eval_macro_sensitivity": 0.6263515234103469,
|
| 195 |
+
"eval_macro_specificity": 0.9119340321339534,
|
| 196 |
+
"eval_precision": 0.7917843915207833,
|
| 197 |
+
"eval_precision_macro": 0.6175221125829533,
|
| 198 |
+
"eval_recall": 0.7572383073496659,
|
| 199 |
+
"eval_recall_macro": 0.6263515234103469,
|
| 200 |
+
"eval_runtime": 39.6918,
|
| 201 |
+
"eval_samples_per_second": 11.312,
|
| 202 |
+
"eval_steps_per_second": 1.436,
|
| 203 |
+
"eval_weighted_fpr": 0.09654561558901682,
|
| 204 |
+
"eval_weighted_sensitivity": 0.7572383073496659,
|
| 205 |
+
"eval_weighted_specificity": 0.8904978211861475,
|
| 206 |
+
"step": 1792
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 8.928571428571429,
|
| 210 |
+
"grad_norm": 0.17783766984939575,
|
| 211 |
+
"learning_rate": 3.516369047619048e-05,
|
| 212 |
+
"loss": 0.2709,
|
| 213 |
+
"step": 2000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 9.0,
|
| 217 |
+
"eval_accuracy": 0.7728285077951003,
|
| 218 |
+
"eval_f1_macro": 0.6264265498194235,
|
| 219 |
+
"eval_f1_micro": 0.7728285077951002,
|
| 220 |
+
"eval_f1_weighted": 0.7773023130906183,
|
| 221 |
+
"eval_loss": 1.2091724872589111,
|
| 222 |
+
"eval_macro_fpr": 0.10207367511358464,
|
| 223 |
+
"eval_macro_sensitivity": 0.6301385419032478,
|
| 224 |
+
"eval_macro_specificity": 0.9119745284052109,
|
| 225 |
+
"eval_precision": 0.7896658353037499,
|
| 226 |
+
"eval_precision_macro": 0.6331069825924028,
|
| 227 |
+
"eval_recall": 0.7728285077951003,
|
| 228 |
+
"eval_recall_macro": 0.6301385419032478,
|
| 229 |
+
"eval_runtime": 39.549,
|
| 230 |
+
"eval_samples_per_second": 11.353,
|
| 231 |
+
"eval_steps_per_second": 1.441,
|
| 232 |
+
"eval_weighted_fpr": 0.08923884514435695,
|
| 233 |
+
"eval_weighted_sensitivity": 0.7728285077951003,
|
| 234 |
+
"eval_weighted_specificity": 0.8750696058257429,
|
| 235 |
+
"step": 2016
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 10.0,
|
| 239 |
+
"eval_accuracy": 0.7706013363028953,
|
| 240 |
+
"eval_f1_macro": 0.6072269083290167,
|
| 241 |
+
"eval_f1_micro": 0.7706013363028953,
|
| 242 |
+
"eval_f1_weighted": 0.7728190961011124,
|
| 243 |
+
"eval_loss": 1.3829654455184937,
|
| 244 |
+
"eval_macro_fpr": 0.10942591948269315,
|
| 245 |
+
"eval_macro_sensitivity": 0.6073347029229382,
|
| 246 |
+
"eval_macro_specificity": 0.9042606300967866,
|
| 247 |
+
"eval_precision": 0.7782484840467806,
|
| 248 |
+
"eval_precision_macro": 0.611192146322276,
|
| 249 |
+
"eval_recall": 0.7706013363028953,
|
| 250 |
+
"eval_recall_macro": 0.6073347029229382,
|
| 251 |
+
"eval_runtime": 36.3346,
|
| 252 |
+
"eval_samples_per_second": 12.357,
|
| 253 |
+
"eval_steps_per_second": 1.569,
|
| 254 |
+
"eval_weighted_fpr": 0.09027169149868536,
|
| 255 |
+
"eval_weighted_sensitivity": 0.7706013363028953,
|
| 256 |
+
"eval_weighted_specificity": 0.8464411840842511,
|
| 257 |
+
"step": 2240
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 11.0,
|
| 261 |
+
"eval_accuracy": 0.7817371937639198,
|
| 262 |
+
"eval_f1_macro": 0.6076643977953123,
|
| 263 |
+
"eval_f1_micro": 0.7817371937639198,
|
| 264 |
+
"eval_f1_weighted": 0.7855527916255841,
|
| 265 |
+
"eval_loss": 1.4518017768859863,
|
| 266 |
+
"eval_macro_fpr": 0.10298989044794646,
|
| 267 |
+
"eval_macro_sensitivity": 0.6059220985691574,
|
| 268 |
+
"eval_macro_specificity": 0.910581472510156,
|
| 269 |
+
"eval_precision": 0.7944351522866733,
|
| 270 |
+
"eval_precision_macro": 0.6156807945806044,
|
| 271 |
+
"eval_recall": 0.7817371937639198,
|
| 272 |
+
"eval_recall_macro": 0.6059220985691574,
|
| 273 |
+
"eval_runtime": 37.6161,
|
| 274 |
+
"eval_samples_per_second": 11.936,
|
| 275 |
+
"eval_steps_per_second": 1.515,
|
| 276 |
+
"eval_weighted_fpr": 0.08514335360556038,
|
| 277 |
+
"eval_weighted_sensitivity": 0.7817371937639198,
|
| 278 |
+
"eval_weighted_specificity": 0.8605886962767039,
|
| 279 |
+
"step": 2464
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 11.160714285714286,
|
| 283 |
+
"grad_norm": 0.04213592782616615,
|
| 284 |
+
"learning_rate": 3.1443452380952384e-05,
|
| 285 |
+
"loss": 0.1837,
|
| 286 |
+
"step": 2500
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 12.0,
|
| 290 |
+
"eval_accuracy": 0.7683741648106904,
|
| 291 |
+
"eval_f1_macro": 0.6022316813621161,
|
| 292 |
+
"eval_f1_micro": 0.7683741648106904,
|
| 293 |
+
"eval_f1_weighted": 0.7725571842701697,
|
| 294 |
+
"eval_loss": 1.5282827615737915,
|
| 295 |
+
"eval_macro_fpr": 0.10584759106468998,
|
| 296 |
+
"eval_macro_sensitivity": 0.6002802841038135,
|
| 297 |
+
"eval_macro_specificity": 0.9096129114606906,
|
| 298 |
+
"eval_precision": 0.7839702071058822,
|
| 299 |
+
"eval_precision_macro": 0.6142985139152926,
|
| 300 |
+
"eval_recall": 0.7683741648106904,
|
| 301 |
+
"eval_recall_macro": 0.6002802841038135,
|
| 302 |
+
"eval_runtime": 43.2739,
|
| 303 |
+
"eval_samples_per_second": 10.376,
|
| 304 |
+
"eval_steps_per_second": 1.317,
|
| 305 |
+
"eval_weighted_fpr": 0.0913081650570676,
|
| 306 |
+
"eval_weighted_sensitivity": 0.7683741648106904,
|
| 307 |
+
"eval_weighted_specificity": 0.8700774810320715,
|
| 308 |
+
"step": 2688
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 13.0,
|
| 312 |
+
"eval_accuracy": 0.7817371937639198,
|
| 313 |
+
"eval_f1_macro": 0.6331842691361635,
|
| 314 |
+
"eval_f1_micro": 0.7817371937639198,
|
| 315 |
+
"eval_f1_weighted": 0.7847832997246655,
|
| 316 |
+
"eval_loss": 1.5136499404907227,
|
| 317 |
+
"eval_macro_fpr": 0.09788769858953361,
|
| 318 |
+
"eval_macro_sensitivity": 0.6472214192802428,
|
| 319 |
+
"eval_macro_specificity": 0.9137489579374409,
|
| 320 |
+
"eval_precision": 0.7907383501868388,
|
| 321 |
+
"eval_precision_macro": 0.6230847265054374,
|
| 322 |
+
"eval_recall": 0.7817371937639198,
|
| 323 |
+
"eval_recall_macro": 0.6472214192802428,
|
| 324 |
+
"eval_runtime": 37.6374,
|
| 325 |
+
"eval_samples_per_second": 11.93,
|
| 326 |
+
"eval_steps_per_second": 1.514,
|
| 327 |
+
"eval_weighted_fpr": 0.08514335360556038,
|
| 328 |
+
"eval_weighted_sensitivity": 0.7817371937639198,
|
| 329 |
+
"eval_weighted_specificity": 0.8732586379858437,
|
| 330 |
+
"step": 2912
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 13.392857142857142,
|
| 334 |
+
"grad_norm": 0.052352067083120346,
|
| 335 |
+
"learning_rate": 2.7723214285714287e-05,
|
| 336 |
+
"loss": 0.1212,
|
| 337 |
+
"step": 3000
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 14.0,
|
| 341 |
+
"eval_accuracy": 0.7505567928730512,
|
| 342 |
+
"eval_f1_macro": 0.6326679794724908,
|
| 343 |
+
"eval_f1_micro": 0.7505567928730512,
|
| 344 |
+
"eval_f1_weighted": 0.7763683079087403,
|
| 345 |
+
"eval_loss": 1.6569428443908691,
|
| 346 |
+
"eval_macro_fpr": 0.10392968597226783,
|
| 347 |
+
"eval_macro_sensitivity": 0.6498986572515985,
|
| 348 |
+
"eval_macro_specificity": 0.910418612538819,
|
| 349 |
+
"eval_precision": 0.8137737990649265,
|
| 350 |
+
"eval_precision_macro": 0.6379744028295854,
|
| 351 |
+
"eval_recall": 0.7505567928730512,
|
| 352 |
+
"eval_recall_macro": 0.6498986572515985,
|
| 353 |
+
"eval_runtime": 39.2148,
|
| 354 |
+
"eval_samples_per_second": 11.45,
|
| 355 |
+
"eval_steps_per_second": 1.454,
|
| 356 |
+
"eval_weighted_fpr": 0.09973285841495994,
|
| 357 |
+
"eval_weighted_sensitivity": 0.7505567928730512,
|
| 358 |
+
"eval_weighted_specificity": 0.8911176572822244,
|
| 359 |
+
"step": 3136
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 15.0,
|
| 363 |
+
"eval_accuracy": 0.7661469933184856,
|
| 364 |
+
"eval_f1_macro": 0.6067507092613289,
|
| 365 |
+
"eval_f1_micro": 0.7661469933184856,
|
| 366 |
+
"eval_f1_weighted": 0.7669334720264783,
|
| 367 |
+
"eval_loss": 1.5304632186889648,
|
| 368 |
+
"eval_macro_fpr": 0.1054152641675519,
|
| 369 |
+
"eval_macro_sensitivity": 0.6202614952614952,
|
| 370 |
+
"eval_macro_specificity": 0.9092878271223507,
|
| 371 |
+
"eval_precision": 0.7714273125569682,
|
| 372 |
+
"eval_precision_macro": 0.5964778042531244,
|
| 373 |
+
"eval_recall": 0.7661469933184856,
|
| 374 |
+
"eval_recall_macro": 0.6202614952614952,
|
| 375 |
+
"eval_runtime": 36.2301,
|
| 376 |
+
"eval_samples_per_second": 12.393,
|
| 377 |
+
"eval_steps_per_second": 1.573,
|
| 378 |
+
"eval_weighted_fpr": 0.09234828496042216,
|
| 379 |
+
"eval_weighted_sensitivity": 0.7661469933184856,
|
| 380 |
+
"eval_weighted_specificity": 0.8710043151709171,
|
| 381 |
+
"step": 3360
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 15.625,
|
| 385 |
+
"grad_norm": 0.012736031785607338,
|
| 386 |
+
"learning_rate": 2.4002976190476193e-05,
|
| 387 |
+
"loss": 0.0793,
|
| 388 |
+
"step": 3500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 16.0,
|
| 392 |
+
"eval_accuracy": 0.799554565701559,
|
| 393 |
+
"eval_f1_macro": 0.6084585912172119,
|
| 394 |
+
"eval_f1_micro": 0.799554565701559,
|
| 395 |
+
"eval_f1_weighted": 0.7932752942813613,
|
| 396 |
+
"eval_loss": 1.493147611618042,
|
| 397 |
+
"eval_macro_fpr": 0.0947020823003073,
|
| 398 |
+
"eval_macro_sensitivity": 0.6192910825263767,
|
| 399 |
+
"eval_macro_specificity": 0.9155234182146221,
|
| 400 |
+
"eval_precision": 0.7896342753264775,
|
| 401 |
+
"eval_precision_macro": 0.6015557680872304,
|
| 402 |
+
"eval_recall": 0.799554565701559,
|
| 403 |
+
"eval_recall_macro": 0.6192910825263767,
|
| 404 |
+
"eval_runtime": 41.0154,
|
| 405 |
+
"eval_samples_per_second": 10.947,
|
| 406 |
+
"eval_steps_per_second": 1.39,
|
| 407 |
+
"eval_weighted_fpr": 0.07712082262210797,
|
| 408 |
+
"eval_weighted_sensitivity": 0.799554565701559,
|
| 409 |
+
"eval_weighted_specificity": 0.8625391071569296,
|
| 410 |
+
"step": 3584
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 17.0,
|
| 414 |
+
"eval_accuracy": 0.8017817371937639,
|
| 415 |
+
"eval_f1_macro": 0.6132058230219714,
|
| 416 |
+
"eval_f1_micro": 0.8017817371937639,
|
| 417 |
+
"eval_f1_weighted": 0.7958283353419163,
|
| 418 |
+
"eval_loss": 1.4581812620162964,
|
| 419 |
+
"eval_macro_fpr": 0.09631731021801944,
|
| 420 |
+
"eval_macro_sensitivity": 0.6131253484194661,
|
| 421 |
+
"eval_macro_specificity": 0.9135230515715727,
|
| 422 |
+
"eval_precision": 0.7910538038177782,
|
| 423 |
+
"eval_precision_macro": 0.614274531356077,
|
| 424 |
+
"eval_recall": 0.8017817371937639,
|
| 425 |
+
"eval_recall_macro": 0.6131253484194661,
|
| 426 |
+
"eval_runtime": 38.6901,
|
| 427 |
+
"eval_samples_per_second": 11.605,
|
| 428 |
+
"eval_steps_per_second": 1.473,
|
| 429 |
+
"eval_weighted_fpr": 0.07613344739093242,
|
| 430 |
+
"eval_weighted_sensitivity": 0.8017817371937639,
|
| 431 |
+
"eval_weighted_specificity": 0.8523104690925272,
|
| 432 |
+
"step": 3808
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 17.857142857142858,
|
| 436 |
+
"grad_norm": 0.029111037030816078,
|
| 437 |
+
"learning_rate": 2.0282738095238096e-05,
|
| 438 |
+
"loss": 0.0473,
|
| 439 |
+
"step": 4000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 18.0,
|
| 443 |
+
"eval_accuracy": 0.779510022271715,
|
| 444 |
+
"eval_f1_macro": 0.6223546079097706,
|
| 445 |
+
"eval_f1_micro": 0.779510022271715,
|
| 446 |
+
"eval_f1_weighted": 0.7842607118763782,
|
| 447 |
+
"eval_loss": 1.6771754026412964,
|
| 448 |
+
"eval_macro_fpr": 0.0990010834451406,
|
| 449 |
+
"eval_macro_sensitivity": 0.6342232224585166,
|
| 450 |
+
"eval_macro_specificity": 0.9134334398466122,
|
| 451 |
+
"eval_precision": 0.7923672038744829,
|
| 452 |
+
"eval_precision_macro": 0.615368510122067,
|
| 453 |
+
"eval_recall": 0.779510022271715,
|
| 454 |
+
"eval_recall_macro": 0.6342232224585166,
|
| 455 |
+
"eval_runtime": 46.3028,
|
| 456 |
+
"eval_samples_per_second": 9.697,
|
| 457 |
+
"eval_steps_per_second": 1.231,
|
| 458 |
+
"eval_weighted_fpr": 0.08616187989556136,
|
| 459 |
+
"eval_weighted_sensitivity": 0.779510022271715,
|
| 460 |
+
"eval_weighted_specificity": 0.8742237371147337,
|
| 461 |
+
"step": 4032
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 19.0,
|
| 465 |
+
"eval_accuracy": 0.7928730512249443,
|
| 466 |
+
"eval_f1_macro": 0.634785770715968,
|
| 467 |
+
"eval_f1_micro": 0.7928730512249443,
|
| 468 |
+
"eval_f1_weighted": 0.7892474652218282,
|
| 469 |
+
"eval_loss": 1.5706700086593628,
|
| 470 |
+
"eval_macro_fpr": 0.09660121866045568,
|
| 471 |
+
"eval_macro_sensitivity": 0.6339238398061927,
|
| 472 |
+
"eval_macro_specificity": 0.9148629938940258,
|
| 473 |
+
"eval_precision": 0.7889745868395741,
|
| 474 |
+
"eval_precision_macro": 0.6408528928812904,
|
| 475 |
+
"eval_recall": 0.7928730512249443,
|
| 476 |
+
"eval_recall_macro": 0.6339238398061927,
|
| 477 |
+
"eval_runtime": 36.9538,
|
| 478 |
+
"eval_samples_per_second": 12.15,
|
| 479 |
+
"eval_steps_per_second": 1.542,
|
| 480 |
+
"eval_weighted_fpr": 0.08010335917312661,
|
| 481 |
+
"eval_weighted_sensitivity": 0.7928730512249443,
|
| 482 |
+
"eval_weighted_specificity": 0.8665789243511589,
|
| 483 |
+
"step": 4256
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 20.0,
|
| 487 |
+
"eval_accuracy": 0.8017817371937639,
|
| 488 |
+
"eval_f1_macro": 0.6355024390764469,
|
| 489 |
+
"eval_f1_micro": 0.8017817371937639,
|
| 490 |
+
"eval_f1_weighted": 0.8072864786995735,
|
| 491 |
+
"eval_loss": 1.4890631437301636,
|
| 492 |
+
"eval_macro_fpr": 0.09155341651475672,
|
| 493 |
+
"eval_macro_sensitivity": 0.6284256946021651,
|
| 494 |
+
"eval_macro_specificity": 0.9196350304544231,
|
| 495 |
+
"eval_precision": 0.813571143284143,
|
| 496 |
+
"eval_precision_macro": 0.6440783508741144,
|
| 497 |
+
"eval_recall": 0.8017817371937639,
|
| 498 |
+
"eval_recall_macro": 0.6284256946021651,
|
| 499 |
+
"eval_runtime": 39.4976,
|
| 500 |
+
"eval_samples_per_second": 11.368,
|
| 501 |
+
"eval_steps_per_second": 1.443,
|
| 502 |
+
"eval_weighted_fpr": 0.07613344739093242,
|
| 503 |
+
"eval_weighted_sensitivity": 0.8017817371937639,
|
| 504 |
+
"eval_weighted_specificity": 0.876758384623928,
|
| 505 |
+
"step": 4480
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 20.089285714285715,
|
| 509 |
+
"grad_norm": 0.062061864882707596,
|
| 510 |
+
"learning_rate": 1.65625e-05,
|
| 511 |
+
"loss": 0.0476,
|
| 512 |
+
"step": 4500
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 21.0,
|
| 516 |
+
"eval_accuracy": 0.8062360801781737,
|
| 517 |
+
"eval_f1_macro": 0.6406636359280463,
|
| 518 |
+
"eval_f1_micro": 0.8062360801781738,
|
| 519 |
+
"eval_f1_weighted": 0.8116960978838106,
|
| 520 |
+
"eval_loss": 1.506375789642334,
|
| 521 |
+
"eval_macro_fpr": 0.08956088844408511,
|
| 522 |
+
"eval_macro_sensitivity": 0.631990756990757,
|
| 523 |
+
"eval_macro_specificity": 0.9204129046853009,
|
| 524 |
+
"eval_precision": 0.8180661383650976,
|
| 525 |
+
"eval_precision_macro": 0.6510963251330297,
|
| 526 |
+
"eval_recall": 0.8062360801781737,
|
| 527 |
+
"eval_recall_macro": 0.631990756990757,
|
| 528 |
+
"eval_runtime": 36.7972,
|
| 529 |
+
"eval_samples_per_second": 12.202,
|
| 530 |
+
"eval_steps_per_second": 1.549,
|
| 531 |
+
"eval_weighted_fpr": 0.0741687979539642,
|
| 532 |
+
"eval_weighted_sensitivity": 0.8062360801781737,
|
| 533 |
+
"eval_weighted_specificity": 0.8754155385630296,
|
| 534 |
+
"step": 4704
|
| 535 |
+
}
|
| 536 |
+
],
|
| 537 |
+
"logging_steps": 500,
|
| 538 |
+
"max_steps": 6720,
|
| 539 |
+
"num_input_tokens_seen": 0,
|
| 540 |
+
"num_train_epochs": 30,
|
| 541 |
+
"save_steps": 500,
|
| 542 |
+
"total_flos": 9901573036572672.0,
|
| 543 |
+
"train_batch_size": 8,
|
| 544 |
+
"trial_name": null,
|
| 545 |
+
"trial_params": null
|
| 546 |
+
}
|
training_checkpoints/checkpoint-4704/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8855967b7a0e1809c878c330ee1e252a00739f2113c0de07ba125067993b5258
|
| 3 |
+
size 5048
|
training_checkpoints/checkpoint-4704/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|