Updated model for regularization testing
Browse files- config.json +25 -37
- model.safetensors +3 -0
- optimizer.pt +2 -2
- rng_state.pth +2 -2
- scheduler.pt +2 -2
- tokenizer.json +10 -10
- tokenizer_config.json +43 -2
- trainer_state.json +901 -53
- training_args.bin +2 -2
config.json
CHANGED
|
@@ -9,45 +9,33 @@
|
|
| 9 |
"dropout": 0.1,
|
| 10 |
"hidden_dim": 3072,
|
| 11 |
"id2label": {
|
| 12 |
-
"0": "
|
| 13 |
-
"1": "
|
| 14 |
-
"2": "
|
| 15 |
-
"3": "
|
| 16 |
-
"4": "
|
| 17 |
-
"5": "
|
| 18 |
-
"6": "
|
| 19 |
-
"7": "
|
| 20 |
-
"8": "
|
| 21 |
-
"9": "
|
| 22 |
-
"10": "
|
| 23 |
-
"11": "
|
| 24 |
-
"12": "ACTION_Physical_Material",
|
| 25 |
-
"13": "ACTION_Technical_Specialist",
|
| 26 |
-
"14": "ACTION_Divination",
|
| 27 |
-
"15": "ACTION_Shaman_Medium_Healer",
|
| 28 |
-
"16": "ACTION_Priest_High_Religion",
|
| 29 |
-
"17": "ACTION_Other"
|
| 30 |
},
|
| 31 |
"initializer_range": 0.02,
|
| 32 |
"label2id": {
|
| 33 |
-
"
|
| 34 |
-
"
|
| 35 |
-
"
|
| 36 |
-
"
|
| 37 |
-
"
|
| 38 |
-
"
|
| 39 |
-
"
|
| 40 |
-
"
|
| 41 |
-
"
|
| 42 |
-
"
|
| 43 |
-
"
|
| 44 |
-
"
|
| 45 |
-
"CAUSE_Spirits_Gods": 7,
|
| 46 |
-
"CAUSE_Witchcraft_Sorcery": 8,
|
| 47 |
-
"EVENT": 0,
|
| 48 |
-
"EVENT_Accident": 2,
|
| 49 |
-
"EVENT_Illness": 1,
|
| 50 |
-
"EVENT_Other": 3
|
| 51 |
},
|
| 52 |
"max_position_embeddings": 512,
|
| 53 |
"model_type": "distilbert",
|
|
@@ -60,6 +48,6 @@
|
|
| 60 |
"sinusoidal_pos_embds": false,
|
| 61 |
"tie_weights_": true,
|
| 62 |
"torch_dtype": "float32",
|
| 63 |
-
"transformers_version": "4.
|
| 64 |
"vocab_size": 30522
|
| 65 |
}
|
|
|
|
| 9 |
"dropout": 0.1,
|
| 10 |
"hidden_dim": 3072,
|
| 11 |
"id2label": {
|
| 12 |
+
"0": "EVENT_Illness",
|
| 13 |
+
"1": "EVENT_Accident",
|
| 14 |
+
"2": "EVENT_Other",
|
| 15 |
+
"3": "CAUSE_Material_Physical",
|
| 16 |
+
"4": "CAUSE_Spirits_Gods",
|
| 17 |
+
"5": "CAUSE_Witchcraft_Sorcery",
|
| 18 |
+
"6": "CAUSE_Rule_Violation_Taboo",
|
| 19 |
+
"7": "ACTION_Physical_Material",
|
| 20 |
+
"8": "ACTION_Technical_Specialist",
|
| 21 |
+
"9": "ACTION_Divination",
|
| 22 |
+
"10": "ACTION_Shaman_Medium_Healer",
|
| 23 |
+
"11": "ACTION_Priest_High_Religion"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
},
|
| 25 |
"initializer_range": 0.02,
|
| 26 |
"label2id": {
|
| 27 |
+
"ACTION_Divination": 9,
|
| 28 |
+
"ACTION_Physical_Material": 7,
|
| 29 |
+
"ACTION_Priest_High_Religion": 11,
|
| 30 |
+
"ACTION_Shaman_Medium_Healer": 10,
|
| 31 |
+
"ACTION_Technical_Specialist": 8,
|
| 32 |
+
"CAUSE_Material_Physical": 3,
|
| 33 |
+
"CAUSE_Rule_Violation_Taboo": 6,
|
| 34 |
+
"CAUSE_Spirits_Gods": 4,
|
| 35 |
+
"CAUSE_Witchcraft_Sorcery": 5,
|
| 36 |
+
"EVENT_Accident": 1,
|
| 37 |
+
"EVENT_Illness": 0,
|
| 38 |
+
"EVENT_Other": 2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
},
|
| 40 |
"max_position_embeddings": 512,
|
| 41 |
"model_type": "distilbert",
|
|
|
|
| 48 |
"sinusoidal_pos_embds": false,
|
| 49 |
"tie_weights_": true,
|
| 50 |
"torch_dtype": "float32",
|
| 51 |
+
"transformers_version": "4.41.1",
|
| 52 |
"vocab_size": 30522
|
| 53 |
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:610010ff6d1820ecda9ccc984dc14b78e65a98750da221e9da71dd4100294c76
|
| 3 |
+
size 267863328
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:097a9d2dbbb8bd734adf115a34b696de90b89b7e310a93e9fb3c58bd2e7ddceb
|
| 3 |
+
size 535788730
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88ff5bb36b8e29678107a9fe193a35ed023fde1c7b0d9c95341d37f70c2cdbea
|
| 3 |
+
size 14244
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10b3b185fea1205fbfcb56f550d303bc0b6cb7d325de973a2d3dc5d3e34c96cf
|
| 3 |
+
size 1064
|
tokenizer.json
CHANGED
|
@@ -10,48 +10,48 @@
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
| 13 |
-
"special": true,
|
| 14 |
"content": "[PAD]",
|
| 15 |
"single_word": false,
|
| 16 |
"lstrip": false,
|
| 17 |
"rstrip": false,
|
| 18 |
-
"normalized": false
|
|
|
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 100,
|
| 22 |
-
"special": true,
|
| 23 |
"content": "[UNK]",
|
| 24 |
"single_word": false,
|
| 25 |
"lstrip": false,
|
| 26 |
"rstrip": false,
|
| 27 |
-
"normalized": false
|
|
|
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 101,
|
| 31 |
-
"special": true,
|
| 32 |
"content": "[CLS]",
|
| 33 |
"single_word": false,
|
| 34 |
"lstrip": false,
|
| 35 |
"rstrip": false,
|
| 36 |
-
"normalized": false
|
|
|
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 102,
|
| 40 |
-
"special": true,
|
| 41 |
"content": "[SEP]",
|
| 42 |
"single_word": false,
|
| 43 |
"lstrip": false,
|
| 44 |
"rstrip": false,
|
| 45 |
-
"normalized": false
|
|
|
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 103,
|
| 49 |
-
"special": true,
|
| 50 |
"content": "[MASK]",
|
| 51 |
"single_word": false,
|
| 52 |
"lstrip": false,
|
| 53 |
"rstrip": false,
|
| 54 |
-
"normalized": false
|
|
|
|
| 55 |
}
|
| 56 |
],
|
| 57 |
"normalizer": {
|
|
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
|
|
|
| 13 |
"content": "[PAD]",
|
| 14 |
"single_word": false,
|
| 15 |
"lstrip": false,
|
| 16 |
"rstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"special": true
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 100,
|
|
|
|
| 22 |
"content": "[UNK]",
|
| 23 |
"single_word": false,
|
| 24 |
"lstrip": false,
|
| 25 |
"rstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"special": true
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 101,
|
|
|
|
| 31 |
"content": "[CLS]",
|
| 32 |
"single_word": false,
|
| 33 |
"lstrip": false,
|
| 34 |
"rstrip": false,
|
| 35 |
+
"normalized": false,
|
| 36 |
+
"special": true
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 102,
|
|
|
|
| 40 |
"content": "[SEP]",
|
| 41 |
"single_word": false,
|
| 42 |
"lstrip": false,
|
| 43 |
"rstrip": false,
|
| 44 |
+
"normalized": false,
|
| 45 |
+
"special": true
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 103,
|
|
|
|
| 49 |
"content": "[MASK]",
|
| 50 |
"single_word": false,
|
| 51 |
"lstrip": false,
|
| 52 |
"rstrip": false,
|
| 53 |
+
"normalized": false,
|
| 54 |
+
"special": true
|
| 55 |
}
|
| 56 |
],
|
| 57 |
"normalizer": {
|
tokenizer_config.json
CHANGED
|
@@ -1,12 +1,53 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"cls_token": "[CLS]",
|
| 3 |
"do_lower_case": true,
|
| 4 |
"mask_token": "[MASK]",
|
| 5 |
"model_max_length": 512,
|
| 6 |
-
"name_or_path": "distilbert-base-uncased",
|
| 7 |
"pad_token": "[PAD]",
|
| 8 |
"sep_token": "[SEP]",
|
| 9 |
-
"special_tokens_map_file": null,
|
| 10 |
"strip_accents": null,
|
| 11 |
"tokenize_chinese_chars": true,
|
| 12 |
"tokenizer_class": "DistilBertTokenizer",
|
|
|
|
| 1 |
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
"cls_token": "[CLS]",
|
| 46 |
"do_lower_case": true,
|
| 47 |
"mask_token": "[MASK]",
|
| 48 |
"model_max_length": 512,
|
|
|
|
| 49 |
"pad_token": "[PAD]",
|
| 50 |
"sep_token": "[SEP]",
|
|
|
|
| 51 |
"strip_accents": null,
|
| 52 |
"tokenize_chinese_chars": true,
|
| 53 |
"tokenizer_class": "DistilBertTokenizer",
|
trainer_state.json
CHANGED
|
@@ -1,77 +1,925 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "
|
| 4 |
-
"epoch":
|
| 5 |
-
"
|
|
|
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
{
|
| 11 |
"epoch": 1.0,
|
| 12 |
-
"eval_accuracy": 0.
|
| 13 |
-
"eval_f1": 0.
|
| 14 |
-
"eval_loss": 0.
|
| 15 |
-
"eval_roc_auc": 0.
|
| 16 |
-
"eval_runtime":
|
| 17 |
-
"eval_samples_per_second":
|
| 18 |
-
"eval_steps_per_second":
|
| 19 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
},
|
| 21 |
{
|
| 22 |
"epoch": 2.0,
|
| 23 |
-
"eval_accuracy": 0.
|
| 24 |
-
"eval_f1": 0.
|
| 25 |
-
"eval_loss": 0.
|
| 26 |
-
"eval_roc_auc": 0.
|
| 27 |
-
"eval_runtime":
|
| 28 |
-
"eval_samples_per_second":
|
| 29 |
-
"eval_steps_per_second":
|
| 30 |
-
"step":
|
| 31 |
-
},
|
| 32 |
-
{
|
| 33 |
-
"epoch": 2.
|
| 34 |
-
"
|
| 35 |
-
"
|
| 36 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"epoch": 3.0,
|
| 40 |
-
"eval_accuracy": 0.
|
| 41 |
-
"eval_f1": 0.
|
| 42 |
-
"eval_loss": 0.
|
| 43 |
-
"eval_roc_auc": 0.
|
| 44 |
-
"eval_runtime":
|
| 45 |
-
"eval_samples_per_second":
|
| 46 |
-
"eval_steps_per_second":
|
| 47 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
},
|
| 49 |
{
|
| 50 |
"epoch": 4.0,
|
| 51 |
-
"eval_accuracy": 0.
|
| 52 |
-
"eval_f1": 0.
|
| 53 |
-
"eval_loss": 0.
|
| 54 |
-
"eval_roc_auc": 0.
|
| 55 |
-
"eval_runtime":
|
| 56 |
-
"eval_samples_per_second":
|
| 57 |
-
"eval_steps_per_second":
|
| 58 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 5.0,
|
| 62 |
-
"eval_accuracy": 0.
|
| 63 |
-
"eval_f1": 0.
|
| 64 |
-
"eval_loss": 0.
|
| 65 |
-
"eval_roc_auc": 0.
|
| 66 |
-
"eval_runtime":
|
| 67 |
-
"eval_samples_per_second":
|
| 68 |
-
"eval_steps_per_second":
|
| 69 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 70 |
}
|
| 71 |
],
|
| 72 |
-
"
|
| 73 |
-
"
|
| 74 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
"trial_name": null,
|
| 76 |
"trial_params": null
|
| 77 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.6474642162926313,
|
| 3 |
+
"best_model_checkpoint": "drive/MyDrive/NLP_HRAF//Models/HRAF_MultiLabel_SubClasses_Kfolds/Model_3_LearningRates/Learning_Rate_2e-05_fold_1/checkpoint-10790",
|
| 4 |
+
"epoch": 13.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 10790,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.12048192771084337,
|
| 13 |
+
"grad_norm": 0.7108750343322754,
|
| 14 |
+
"learning_rate": 1.9839357429718877e-05,
|
| 15 |
+
"loss": 0.4589,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.24096385542168675,
|
| 20 |
+
"grad_norm": 0.7625552415847778,
|
| 21 |
+
"learning_rate": 1.967871485943775e-05,
|
| 22 |
+
"loss": 0.3592,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.3614457831325301,
|
| 27 |
+
"grad_norm": 0.8666885495185852,
|
| 28 |
+
"learning_rate": 1.951807228915663e-05,
|
| 29 |
+
"loss": 0.3393,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.4819277108433735,
|
| 34 |
+
"grad_norm": 0.9493631720542908,
|
| 35 |
+
"learning_rate": 1.9357429718875505e-05,
|
| 36 |
+
"loss": 0.3154,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.6024096385542169,
|
| 41 |
+
"grad_norm": 0.8787522315979004,
|
| 42 |
+
"learning_rate": 1.9196787148594377e-05,
|
| 43 |
+
"loss": 0.3001,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.7228915662650602,
|
| 48 |
+
"grad_norm": 0.9206348061561584,
|
| 49 |
+
"learning_rate": 1.9036144578313255e-05,
|
| 50 |
+
"loss": 0.2905,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.8433734939759037,
|
| 55 |
+
"grad_norm": 1.008158564567566,
|
| 56 |
+
"learning_rate": 1.887550200803213e-05,
|
| 57 |
+
"loss": 0.2877,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.963855421686747,
|
| 62 |
+
"grad_norm": 1.307525873184204,
|
| 63 |
+
"learning_rate": 1.8714859437751005e-05,
|
| 64 |
+
"loss": 0.2888,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
{
|
| 68 |
"epoch": 1.0,
|
| 69 |
+
"eval_accuracy": 0.36347197106690776,
|
| 70 |
+
"eval_f1": 0.5040954832670255,
|
| 71 |
+
"eval_loss": 0.2610304355621338,
|
| 72 |
+
"eval_roc_auc": 0.67926006705383,
|
| 73 |
+
"eval_runtime": 20.823,
|
| 74 |
+
"eval_samples_per_second": 79.671,
|
| 75 |
+
"eval_steps_per_second": 9.989,
|
| 76 |
+
"step": 830
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 1.0843373493975903,
|
| 80 |
+
"grad_norm": 1.141208529472351,
|
| 81 |
+
"learning_rate": 1.855421686746988e-05,
|
| 82 |
+
"loss": 0.269,
|
| 83 |
+
"step": 900
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.2048192771084336,
|
| 87 |
+
"grad_norm": 0.8844193816184998,
|
| 88 |
+
"learning_rate": 1.8393574297188755e-05,
|
| 89 |
+
"loss": 0.2579,
|
| 90 |
+
"step": 1000
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.3253012048192772,
|
| 94 |
+
"grad_norm": 1.1616642475128174,
|
| 95 |
+
"learning_rate": 1.8232931726907634e-05,
|
| 96 |
+
"loss": 0.2549,
|
| 97 |
+
"step": 1100
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 1.4457831325301205,
|
| 101 |
+
"grad_norm": 1.3445640802383423,
|
| 102 |
+
"learning_rate": 1.807228915662651e-05,
|
| 103 |
+
"loss": 0.2551,
|
| 104 |
+
"step": 1200
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.5662650602409638,
|
| 108 |
+
"grad_norm": 1.131273865699768,
|
| 109 |
+
"learning_rate": 1.7911646586345384e-05,
|
| 110 |
+
"loss": 0.2672,
|
| 111 |
+
"step": 1300
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 1.6867469879518073,
|
| 115 |
+
"grad_norm": 1.3358525037765503,
|
| 116 |
+
"learning_rate": 1.775100401606426e-05,
|
| 117 |
+
"loss": 0.2454,
|
| 118 |
+
"step": 1400
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 1.8072289156626506,
|
| 122 |
+
"grad_norm": 1.2100324630737305,
|
| 123 |
+
"learning_rate": 1.7590361445783134e-05,
|
| 124 |
+
"loss": 0.2364,
|
| 125 |
+
"step": 1500
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.927710843373494,
|
| 129 |
+
"grad_norm": 1.2214767932891846,
|
| 130 |
+
"learning_rate": 1.742971887550201e-05,
|
| 131 |
+
"loss": 0.2409,
|
| 132 |
+
"step": 1600
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 2.0,
|
| 136 |
+
"eval_accuracy": 0.3767329716696805,
|
| 137 |
+
"eval_f1": 0.6076237435994689,
|
| 138 |
+
"eval_loss": 0.24850943684577942,
|
| 139 |
+
"eval_roc_auc": 0.7578414834576195,
|
| 140 |
+
"eval_runtime": 21.0346,
|
| 141 |
+
"eval_samples_per_second": 78.87,
|
| 142 |
+
"eval_steps_per_second": 9.888,
|
| 143 |
+
"step": 1660
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 2.0481927710843375,
|
| 147 |
+
"grad_norm": 0.8881352543830872,
|
| 148 |
+
"learning_rate": 1.7269076305220884e-05,
|
| 149 |
+
"loss": 0.2341,
|
| 150 |
+
"step": 1700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 2.1686746987951806,
|
| 154 |
+
"grad_norm": 1.4135143756866455,
|
| 155 |
+
"learning_rate": 1.710843373493976e-05,
|
| 156 |
+
"loss": 0.2119,
|
| 157 |
+
"step": 1800
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 2.289156626506024,
|
| 161 |
+
"grad_norm": 1.291266918182373,
|
| 162 |
+
"learning_rate": 1.6947791164658637e-05,
|
| 163 |
+
"loss": 0.2074,
|
| 164 |
+
"step": 1900
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.4096385542168672,
|
| 168 |
+
"grad_norm": 1.311950445175171,
|
| 169 |
+
"learning_rate": 1.6787148594377512e-05,
|
| 170 |
+
"loss": 0.2063,
|
| 171 |
+
"step": 2000
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.5301204819277108,
|
| 175 |
+
"grad_norm": 1.6564782857894897,
|
| 176 |
+
"learning_rate": 1.6626506024096387e-05,
|
| 177 |
+
"loss": 0.2169,
|
| 178 |
+
"step": 2100
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 2.6506024096385543,
|
| 182 |
+
"grad_norm": 2.071871757507324,
|
| 183 |
+
"learning_rate": 1.6465863453815262e-05,
|
| 184 |
+
"loss": 0.2041,
|
| 185 |
+
"step": 2200
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.7710843373493974,
|
| 189 |
+
"grad_norm": 1.6373318433761597,
|
| 190 |
+
"learning_rate": 1.6305220883534137e-05,
|
| 191 |
+
"loss": 0.2149,
|
| 192 |
+
"step": 2300
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.891566265060241,
|
| 196 |
+
"grad_norm": 1.5182716846466064,
|
| 197 |
+
"learning_rate": 1.6144578313253015e-05,
|
| 198 |
+
"loss": 0.2057,
|
| 199 |
+
"step": 2400
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"epoch": 3.0,
|
| 203 |
+
"eval_accuracy": 0.3821579264617239,
|
| 204 |
+
"eval_f1": 0.6133793367835921,
|
| 205 |
+
"eval_loss": 0.24218665063381195,
|
| 206 |
+
"eval_roc_auc": 0.7590708449600018,
|
| 207 |
+
"eval_runtime": 20.7829,
|
| 208 |
+
"eval_samples_per_second": 79.825,
|
| 209 |
+
"eval_steps_per_second": 10.008,
|
| 210 |
+
"step": 2490
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 3.0120481927710845,
|
| 214 |
+
"grad_norm": 1.7013568878173828,
|
| 215 |
+
"learning_rate": 1.5983935742971887e-05,
|
| 216 |
+
"loss": 0.202,
|
| 217 |
+
"step": 2500
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 3.1325301204819276,
|
| 221 |
+
"grad_norm": 1.2363818883895874,
|
| 222 |
+
"learning_rate": 1.5823293172690762e-05,
|
| 223 |
+
"loss": 0.1797,
|
| 224 |
+
"step": 2600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.253012048192771,
|
| 228 |
+
"grad_norm": 2.084066390991211,
|
| 229 |
+
"learning_rate": 1.566265060240964e-05,
|
| 230 |
+
"loss": 0.1788,
|
| 231 |
+
"step": 2700
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 3.3734939759036147,
|
| 235 |
+
"grad_norm": 1.1712193489074707,
|
| 236 |
+
"learning_rate": 1.5502008032128516e-05,
|
| 237 |
+
"loss": 0.169,
|
| 238 |
+
"step": 2800
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 3.4939759036144578,
|
| 242 |
+
"grad_norm": 1.9433542490005493,
|
| 243 |
+
"learning_rate": 1.534136546184739e-05,
|
| 244 |
+
"loss": 0.1712,
|
| 245 |
+
"step": 2900
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 3.6144578313253013,
|
| 249 |
+
"grad_norm": 2.300347328186035,
|
| 250 |
+
"learning_rate": 1.5180722891566266e-05,
|
| 251 |
+
"loss": 0.1749,
|
| 252 |
+
"step": 3000
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 3.734939759036145,
|
| 256 |
+
"grad_norm": 1.4740608930587769,
|
| 257 |
+
"learning_rate": 1.5020080321285142e-05,
|
| 258 |
+
"loss": 0.1736,
|
| 259 |
+
"step": 3100
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 3.855421686746988,
|
| 263 |
+
"grad_norm": 1.7682331800460815,
|
| 264 |
+
"learning_rate": 1.4859437751004017e-05,
|
| 265 |
+
"loss": 0.1635,
|
| 266 |
+
"step": 3200
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 3.9759036144578315,
|
| 270 |
+
"grad_norm": 1.4352338314056396,
|
| 271 |
+
"learning_rate": 1.4698795180722894e-05,
|
| 272 |
+
"loss": 0.1587,
|
| 273 |
+
"step": 3300
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 4.0,
|
| 277 |
+
"eval_accuracy": 0.38276069921639544,
|
| 278 |
+
"eval_f1": 0.6340291889894697,
|
| 279 |
+
"eval_loss": 0.2462325543165207,
|
| 280 |
+
"eval_roc_auc": 0.7771644800479557,
|
| 281 |
+
"eval_runtime": 21.0125,
|
| 282 |
+
"eval_samples_per_second": 78.953,
|
| 283 |
+
"eval_steps_per_second": 9.899,
|
| 284 |
+
"step": 3320
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 4.096385542168675,
|
| 288 |
+
"grad_norm": 3.4992010593414307,
|
| 289 |
+
"learning_rate": 1.4538152610441769e-05,
|
| 290 |
+
"loss": 0.148,
|
| 291 |
+
"step": 3400
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 4.216867469879518,
|
| 295 |
+
"grad_norm": 1.555429458618164,
|
| 296 |
+
"learning_rate": 1.4377510040160642e-05,
|
| 297 |
+
"loss": 0.1397,
|
| 298 |
+
"step": 3500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 4.337349397590361,
|
| 302 |
+
"grad_norm": 2.655567169189453,
|
| 303 |
+
"learning_rate": 1.4216867469879519e-05,
|
| 304 |
+
"loss": 0.1412,
|
| 305 |
+
"step": 3600
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 4.457831325301205,
|
| 309 |
+
"grad_norm": 1.797630786895752,
|
| 310 |
+
"learning_rate": 1.4056224899598394e-05,
|
| 311 |
+
"loss": 0.1419,
|
| 312 |
+
"step": 3700
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 4.578313253012048,
|
| 316 |
+
"grad_norm": 1.2415262460708618,
|
| 317 |
+
"learning_rate": 1.3895582329317269e-05,
|
| 318 |
+
"loss": 0.1405,
|
| 319 |
+
"step": 3800
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 4.698795180722891,
|
| 323 |
+
"grad_norm": 1.4111042022705078,
|
| 324 |
+
"learning_rate": 1.3734939759036146e-05,
|
| 325 |
+
"loss": 0.1349,
|
| 326 |
+
"step": 3900
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 4.8192771084337345,
|
| 330 |
+
"grad_norm": 2.2596189975738525,
|
| 331 |
+
"learning_rate": 1.357429718875502e-05,
|
| 332 |
+
"loss": 0.1291,
|
| 333 |
+
"step": 4000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 4.9397590361445785,
|
| 337 |
+
"grad_norm": 2.264066457748413,
|
| 338 |
+
"learning_rate": 1.3413654618473897e-05,
|
| 339 |
+
"loss": 0.1412,
|
| 340 |
+
"step": 4100
|
| 341 |
},
|
| 342 |
{
|
| 343 |
"epoch": 5.0,
|
| 344 |
+
"eval_accuracy": 0.38637733574442434,
|
| 345 |
+
"eval_f1": 0.627102627102627,
|
| 346 |
+
"eval_loss": 0.25749197602272034,
|
| 347 |
+
"eval_roc_auc": 0.7690261217645041,
|
| 348 |
+
"eval_runtime": 20.8914,
|
| 349 |
+
"eval_samples_per_second": 79.411,
|
| 350 |
+
"eval_steps_per_second": 9.956,
|
| 351 |
+
"step": 4150
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 5.0602409638554215,
|
| 355 |
+
"grad_norm": 1.1299407482147217,
|
| 356 |
+
"learning_rate": 1.3253012048192772e-05,
|
| 357 |
+
"loss": 0.1223,
|
| 358 |
+
"step": 4200
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 5.180722891566265,
|
| 362 |
+
"grad_norm": 1.1912181377410889,
|
| 363 |
+
"learning_rate": 1.309236947791165e-05,
|
| 364 |
+
"loss": 0.1112,
|
| 365 |
+
"step": 4300
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 5.301204819277109,
|
| 369 |
+
"grad_norm": 2.00722074508667,
|
| 370 |
+
"learning_rate": 1.2931726907630524e-05,
|
| 371 |
+
"loss": 0.1035,
|
| 372 |
+
"step": 4400
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 5.421686746987952,
|
| 376 |
+
"grad_norm": 1.543757677078247,
|
| 377 |
+
"learning_rate": 1.2771084337349398e-05,
|
| 378 |
+
"loss": 0.1082,
|
| 379 |
+
"step": 4500
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 5.542168674698795,
|
| 383 |
+
"grad_norm": 1.6844489574432373,
|
| 384 |
+
"learning_rate": 1.2610441767068273e-05,
|
| 385 |
+
"loss": 0.1113,
|
| 386 |
+
"step": 4600
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 5.662650602409639,
|
| 390 |
+
"grad_norm": 2.3987362384796143,
|
| 391 |
+
"learning_rate": 1.244979919678715e-05,
|
| 392 |
+
"loss": 0.1054,
|
| 393 |
+
"step": 4700
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 5.783132530120482,
|
| 397 |
+
"grad_norm": 1.196558952331543,
|
| 398 |
+
"learning_rate": 1.2289156626506024e-05,
|
| 399 |
+
"loss": 0.1076,
|
| 400 |
+
"step": 4800
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 5.903614457831325,
|
| 404 |
+
"grad_norm": 1.7159672975540161,
|
| 405 |
+
"learning_rate": 1.2128514056224901e-05,
|
| 406 |
+
"loss": 0.1129,
|
| 407 |
+
"step": 4900
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 6.0,
|
| 411 |
+
"eval_accuracy": 0.38396624472573837,
|
| 412 |
+
"eval_f1": 0.6401569186875892,
|
| 413 |
+
"eval_loss": 0.28615912795066833,
|
| 414 |
+
"eval_roc_auc": 0.7876849584660104,
|
| 415 |
+
"eval_runtime": 21.0174,
|
| 416 |
+
"eval_samples_per_second": 78.935,
|
| 417 |
+
"eval_steps_per_second": 9.897,
|
| 418 |
+
"step": 4980
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 6.024096385542169,
|
| 422 |
+
"grad_norm": 1.7499059438705444,
|
| 423 |
+
"learning_rate": 1.1967871485943776e-05,
|
| 424 |
+
"loss": 0.101,
|
| 425 |
+
"step": 5000
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 6.144578313253012,
|
| 429 |
+
"grad_norm": 1.2233022451400757,
|
| 430 |
+
"learning_rate": 1.1807228915662651e-05,
|
| 431 |
+
"loss": 0.0835,
|
| 432 |
+
"step": 5100
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 6.265060240963855,
|
| 436 |
+
"grad_norm": 1.219558596611023,
|
| 437 |
+
"learning_rate": 1.1646586345381528e-05,
|
| 438 |
+
"loss": 0.082,
|
| 439 |
+
"step": 5200
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 6.385542168674699,
|
| 443 |
+
"grad_norm": 2.4673707485198975,
|
| 444 |
+
"learning_rate": 1.1485943775100403e-05,
|
| 445 |
+
"loss": 0.0808,
|
| 446 |
+
"step": 5300
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 6.506024096385542,
|
| 450 |
+
"grad_norm": 2.749701738357544,
|
| 451 |
+
"learning_rate": 1.132530120481928e-05,
|
| 452 |
+
"loss": 0.0891,
|
| 453 |
+
"step": 5400
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 6.626506024096385,
|
| 457 |
+
"grad_norm": 2.653024196624756,
|
| 458 |
+
"learning_rate": 1.1164658634538153e-05,
|
| 459 |
+
"loss": 0.0895,
|
| 460 |
+
"step": 5500
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 6.746987951807229,
|
| 464 |
+
"grad_norm": 1.81606924533844,
|
| 465 |
+
"learning_rate": 1.1004016064257028e-05,
|
| 466 |
+
"loss": 0.0827,
|
| 467 |
+
"step": 5600
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 6.867469879518072,
|
| 471 |
+
"grad_norm": 2.775585174560547,
|
| 472 |
+
"learning_rate": 1.0843373493975904e-05,
|
| 473 |
+
"loss": 0.0872,
|
| 474 |
+
"step": 5700
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 6.9879518072289155,
|
| 478 |
+
"grad_norm": 3.0529415607452393,
|
| 479 |
+
"learning_rate": 1.068273092369478e-05,
|
| 480 |
+
"loss": 0.0754,
|
| 481 |
+
"step": 5800
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 7.0,
|
| 485 |
+
"eval_accuracy": 0.3899939722724533,
|
| 486 |
+
"eval_f1": 0.6352293577981651,
|
| 487 |
+
"eval_loss": 0.30027899146080017,
|
| 488 |
+
"eval_roc_auc": 0.7791627807894616,
|
| 489 |
+
"eval_runtime": 20.9855,
|
| 490 |
+
"eval_samples_per_second": 79.054,
|
| 491 |
+
"eval_steps_per_second": 9.912,
|
| 492 |
+
"step": 5810
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 7.108433734939759,
|
| 496 |
+
"grad_norm": 1.6993205547332764,
|
| 497 |
+
"learning_rate": 1.0522088353413654e-05,
|
| 498 |
+
"loss": 0.0723,
|
| 499 |
+
"step": 5900
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 7.228915662650603,
|
| 503 |
+
"grad_norm": 2.1551597118377686,
|
| 504 |
+
"learning_rate": 1.0361445783132531e-05,
|
| 505 |
+
"loss": 0.0685,
|
| 506 |
+
"step": 6000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 7.349397590361446,
|
| 510 |
+
"grad_norm": 1.4902188777923584,
|
| 511 |
+
"learning_rate": 1.0200803212851406e-05,
|
| 512 |
+
"loss": 0.0664,
|
| 513 |
+
"step": 6100
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 7.469879518072289,
|
| 517 |
+
"grad_norm": 0.829651951789856,
|
| 518 |
+
"learning_rate": 1.0040160642570283e-05,
|
| 519 |
+
"loss": 0.0662,
|
| 520 |
+
"step": 6200
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 7.590361445783133,
|
| 524 |
+
"grad_norm": 2.141355037689209,
|
| 525 |
+
"learning_rate": 9.879518072289156e-06,
|
| 526 |
+
"loss": 0.0654,
|
| 527 |
+
"step": 6300
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 7.710843373493976,
|
| 531 |
+
"grad_norm": 0.7993516325950623,
|
| 532 |
+
"learning_rate": 9.718875502008033e-06,
|
| 533 |
+
"loss": 0.0669,
|
| 534 |
+
"step": 6400
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 7.831325301204819,
|
| 538 |
+
"grad_norm": 0.8961694836616516,
|
| 539 |
+
"learning_rate": 9.558232931726908e-06,
|
| 540 |
+
"loss": 0.0677,
|
| 541 |
+
"step": 6500
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 7.951807228915663,
|
| 545 |
+
"grad_norm": 1.8866826295852661,
|
| 546 |
+
"learning_rate": 9.397590361445785e-06,
|
| 547 |
+
"loss": 0.058,
|
| 548 |
+
"step": 6600
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 8.0,
|
| 552 |
+
"eval_accuracy": 0.3887884267631103,
|
| 553 |
+
"eval_f1": 0.6461868190748349,
|
| 554 |
+
"eval_loss": 0.3176693618297577,
|
| 555 |
+
"eval_roc_auc": 0.790825180731994,
|
| 556 |
+
"eval_runtime": 20.9796,
|
| 557 |
+
"eval_samples_per_second": 79.077,
|
| 558 |
+
"eval_steps_per_second": 9.914,
|
| 559 |
+
"step": 6640
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 8.072289156626505,
|
| 563 |
+
"grad_norm": 0.4812127947807312,
|
| 564 |
+
"learning_rate": 9.23694779116466e-06,
|
| 565 |
+
"loss": 0.0543,
|
| 566 |
+
"step": 6700
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 8.19277108433735,
|
| 570 |
+
"grad_norm": 1.1170074939727783,
|
| 571 |
+
"learning_rate": 9.076305220883535e-06,
|
| 572 |
+
"loss": 0.0534,
|
| 573 |
+
"step": 6800
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 8.313253012048193,
|
| 577 |
+
"grad_norm": 2.044552803039551,
|
| 578 |
+
"learning_rate": 8.91566265060241e-06,
|
| 579 |
+
"loss": 0.0541,
|
| 580 |
+
"step": 6900
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 8.433734939759036,
|
| 584 |
+
"grad_norm": 2.3580517768859863,
|
| 585 |
+
"learning_rate": 8.755020080321286e-06,
|
| 586 |
+
"loss": 0.046,
|
| 587 |
+
"step": 7000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 8.55421686746988,
|
| 591 |
+
"grad_norm": 2.568995952606201,
|
| 592 |
+
"learning_rate": 8.594377510040161e-06,
|
| 593 |
+
"loss": 0.0516,
|
| 594 |
+
"step": 7100
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 8.674698795180722,
|
| 598 |
+
"grad_norm": 0.7591239213943481,
|
| 599 |
+
"learning_rate": 8.433734939759038e-06,
|
| 600 |
+
"loss": 0.0503,
|
| 601 |
+
"step": 7200
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 8.795180722891565,
|
| 605 |
+
"grad_norm": 1.0098503828048706,
|
| 606 |
+
"learning_rate": 8.273092369477911e-06,
|
| 607 |
+
"loss": 0.0452,
|
| 608 |
+
"step": 7300
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 8.91566265060241,
|
| 612 |
+
"grad_norm": 2.4211244583129883,
|
| 613 |
+
"learning_rate": 8.112449799196788e-06,
|
| 614 |
+
"loss": 0.0611,
|
| 615 |
+
"step": 7400
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 9.0,
|
| 619 |
+
"eval_accuracy": 0.38396624472573837,
|
| 620 |
+
"eval_f1": 0.6377283414722372,
|
| 621 |
+
"eval_loss": 0.33842870593070984,
|
| 622 |
+
"eval_roc_auc": 0.783423869627736,
|
| 623 |
+
"eval_runtime": 20.9102,
|
| 624 |
+
"eval_samples_per_second": 79.339,
|
| 625 |
+
"eval_steps_per_second": 9.947,
|
| 626 |
+
"step": 7470
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 9.036144578313253,
|
| 630 |
+
"grad_norm": 1.7786929607391357,
|
| 631 |
+
"learning_rate": 7.951807228915663e-06,
|
| 632 |
+
"loss": 0.0467,
|
| 633 |
+
"step": 7500
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 9.156626506024097,
|
| 637 |
+
"grad_norm": 0.9288263916969299,
|
| 638 |
+
"learning_rate": 7.79116465863454e-06,
|
| 639 |
+
"loss": 0.0358,
|
| 640 |
+
"step": 7600
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 9.27710843373494,
|
| 644 |
+
"grad_norm": 1.6899335384368896,
|
| 645 |
+
"learning_rate": 7.630522088353415e-06,
|
| 646 |
+
"loss": 0.0403,
|
| 647 |
+
"step": 7700
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 9.397590361445783,
|
| 651 |
+
"grad_norm": 0.633351743221283,
|
| 652 |
+
"learning_rate": 7.469879518072289e-06,
|
| 653 |
+
"loss": 0.0409,
|
| 654 |
+
"step": 7800
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 9.518072289156626,
|
| 658 |
+
"grad_norm": 1.880730152130127,
|
| 659 |
+
"learning_rate": 7.309236947791165e-06,
|
| 660 |
+
"loss": 0.04,
|
| 661 |
+
"step": 7900
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 9.638554216867469,
|
| 665 |
+
"grad_norm": 0.7761407494544983,
|
| 666 |
+
"learning_rate": 7.148594377510041e-06,
|
| 667 |
+
"loss": 0.0416,
|
| 668 |
+
"step": 8000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 9.759036144578314,
|
| 672 |
+
"grad_norm": 1.7540706396102905,
|
| 673 |
+
"learning_rate": 6.987951807228917e-06,
|
| 674 |
+
"loss": 0.0383,
|
| 675 |
+
"step": 8100
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 9.879518072289157,
|
| 679 |
+
"grad_norm": 1.621785044670105,
|
| 680 |
+
"learning_rate": 6.8273092369477925e-06,
|
| 681 |
+
"loss": 0.0399,
|
| 682 |
+
"step": 8200
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 10.0,
|
| 686 |
+
"grad_norm": 4.052353382110596,
|
| 687 |
+
"learning_rate": 6.666666666666667e-06,
|
| 688 |
+
"loss": 0.0418,
|
| 689 |
+
"step": 8300
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 10.0,
|
| 693 |
+
"eval_accuracy": 0.38155515370705245,
|
| 694 |
+
"eval_f1": 0.6405322783672002,
|
| 695 |
+
"eval_loss": 0.3592796325683594,
|
| 696 |
+
"eval_roc_auc": 0.7861850408280293,
|
| 697 |
+
"eval_runtime": 20.9788,
|
| 698 |
+
"eval_samples_per_second": 79.08,
|
| 699 |
+
"eval_steps_per_second": 9.915,
|
| 700 |
+
"step": 8300
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 10.120481927710843,
|
| 704 |
+
"grad_norm": 2.524548053741455,
|
| 705 |
+
"learning_rate": 6.5060240963855425e-06,
|
| 706 |
+
"loss": 0.0327,
|
| 707 |
+
"step": 8400
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 10.240963855421686,
|
| 711 |
+
"grad_norm": 2.5789272785186768,
|
| 712 |
+
"learning_rate": 6.345381526104418e-06,
|
| 713 |
+
"loss": 0.0329,
|
| 714 |
+
"step": 8500
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 10.36144578313253,
|
| 718 |
+
"grad_norm": 1.6276124715805054,
|
| 719 |
+
"learning_rate": 6.184738955823294e-06,
|
| 720 |
+
"loss": 0.0327,
|
| 721 |
+
"step": 8600
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 10.481927710843374,
|
| 725 |
+
"grad_norm": 1.2820446491241455,
|
| 726 |
+
"learning_rate": 6.02409638554217e-06,
|
| 727 |
+
"loss": 0.0319,
|
| 728 |
+
"step": 8700
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 10.602409638554217,
|
| 732 |
+
"grad_norm": 0.8200409412384033,
|
| 733 |
+
"learning_rate": 5.863453815261044e-06,
|
| 734 |
+
"loss": 0.0338,
|
| 735 |
+
"step": 8800
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 10.72289156626506,
|
| 739 |
+
"grad_norm": 1.6061540842056274,
|
| 740 |
+
"learning_rate": 5.70281124497992e-06,
|
| 741 |
+
"loss": 0.0311,
|
| 742 |
+
"step": 8900
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 10.843373493975903,
|
| 746 |
+
"grad_norm": 2.1266753673553467,
|
| 747 |
+
"learning_rate": 5.542168674698796e-06,
|
| 748 |
+
"loss": 0.0309,
|
| 749 |
+
"step": 9000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 10.963855421686747,
|
| 753 |
+
"grad_norm": 1.8387219905853271,
|
| 754 |
+
"learning_rate": 5.381526104417672e-06,
|
| 755 |
+
"loss": 0.0338,
|
| 756 |
+
"step": 9100
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 11.0,
|
| 760 |
+
"eval_accuracy": 0.38155515370705245,
|
| 761 |
+
"eval_f1": 0.6427417907769604,
|
| 762 |
+
"eval_loss": 0.37010136246681213,
|
| 763 |
+
"eval_roc_auc": 0.787888262607272,
|
| 764 |
+
"eval_runtime": 20.9246,
|
| 765 |
+
"eval_samples_per_second": 79.285,
|
| 766 |
+
"eval_steps_per_second": 9.94,
|
| 767 |
+
"step": 9130
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 11.08433734939759,
|
| 771 |
+
"grad_norm": 1.1557573080062866,
|
| 772 |
+
"learning_rate": 5.220883534136547e-06,
|
| 773 |
+
"loss": 0.0261,
|
| 774 |
+
"step": 9200
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 11.204819277108435,
|
| 778 |
+
"grad_norm": 0.8962405920028687,
|
| 779 |
+
"learning_rate": 5.060240963855422e-06,
|
| 780 |
+
"loss": 0.0247,
|
| 781 |
+
"step": 9300
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 11.325301204819278,
|
| 785 |
+
"grad_norm": 0.8799346089363098,
|
| 786 |
+
"learning_rate": 4.899598393574298e-06,
|
| 787 |
+
"loss": 0.0255,
|
| 788 |
+
"step": 9400
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 11.44578313253012,
|
| 792 |
+
"grad_norm": 0.8840754628181458,
|
| 793 |
+
"learning_rate": 4.7389558232931736e-06,
|
| 794 |
+
"loss": 0.0266,
|
| 795 |
+
"step": 9500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 11.566265060240964,
|
| 799 |
+
"grad_norm": 1.1683375835418701,
|
| 800 |
+
"learning_rate": 4.578313253012049e-06,
|
| 801 |
+
"loss": 0.0261,
|
| 802 |
+
"step": 9600
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 11.686746987951807,
|
| 806 |
+
"grad_norm": 0.885728120803833,
|
| 807 |
+
"learning_rate": 4.4176706827309244e-06,
|
| 808 |
+
"loss": 0.0257,
|
| 809 |
+
"step": 9700
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 11.80722891566265,
|
| 813 |
+
"grad_norm": 0.2890942394733429,
|
| 814 |
+
"learning_rate": 4.2570281124497995e-06,
|
| 815 |
+
"loss": 0.029,
|
| 816 |
+
"step": 9800
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 11.927710843373493,
|
| 820 |
+
"grad_norm": 2.2087390422821045,
|
| 821 |
+
"learning_rate": 4.096385542168675e-06,
|
| 822 |
+
"loss": 0.0255,
|
| 823 |
+
"step": 9900
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 12.0,
|
| 827 |
+
"eval_accuracy": 0.3845690174804099,
|
| 828 |
+
"eval_f1": 0.6420863309352519,
|
| 829 |
+
"eval_loss": 0.3844664394855499,
|
| 830 |
+
"eval_roc_auc": 0.7870362182333409,
|
| 831 |
+
"eval_runtime": 20.95,
|
| 832 |
+
"eval_samples_per_second": 79.189,
|
| 833 |
+
"eval_steps_per_second": 9.928,
|
| 834 |
+
"step": 9960
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 12.048192771084338,
|
| 838 |
+
"grad_norm": 1.7163885831832886,
|
| 839 |
+
"learning_rate": 3.93574297188755e-06,
|
| 840 |
+
"loss": 0.0261,
|
| 841 |
+
"step": 10000
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 12.168674698795181,
|
| 845 |
+
"grad_norm": 1.0416496992111206,
|
| 846 |
+
"learning_rate": 3.7751004016064258e-06,
|
| 847 |
+
"loss": 0.0215,
|
| 848 |
+
"step": 10100
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 12.289156626506024,
|
| 852 |
+
"grad_norm": 1.3379343748092651,
|
| 853 |
+
"learning_rate": 3.6144578313253016e-06,
|
| 854 |
+
"loss": 0.0223,
|
| 855 |
+
"step": 10200
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 12.409638554216867,
|
| 859 |
+
"grad_norm": 0.658170759677887,
|
| 860 |
+
"learning_rate": 3.453815261044177e-06,
|
| 861 |
+
"loss": 0.0192,
|
| 862 |
+
"step": 10300
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 12.53012048192771,
|
| 866 |
+
"grad_norm": 0.9047495722770691,
|
| 867 |
+
"learning_rate": 3.2931726907630525e-06,
|
| 868 |
+
"loss": 0.0237,
|
| 869 |
+
"step": 10400
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 12.650602409638553,
|
| 873 |
+
"grad_norm": 1.0494842529296875,
|
| 874 |
+
"learning_rate": 3.132530120481928e-06,
|
| 875 |
+
"loss": 0.0193,
|
| 876 |
+
"step": 10500
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 12.771084337349398,
|
| 880 |
+
"grad_norm": 0.32704225182533264,
|
| 881 |
+
"learning_rate": 2.9718875502008034e-06,
|
| 882 |
+
"loss": 0.0208,
|
| 883 |
+
"step": 10600
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 12.891566265060241,
|
| 887 |
+
"grad_norm": 0.49072301387786865,
|
| 888 |
+
"learning_rate": 2.811244979919679e-06,
|
| 889 |
+
"loss": 0.0202,
|
| 890 |
+
"step": 10700
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 13.0,
|
| 894 |
+
"eval_accuracy": 0.38095238095238093,
|
| 895 |
+
"eval_f1": 0.6474642162926313,
|
| 896 |
+
"eval_loss": 0.3947090208530426,
|
| 897 |
+
"eval_roc_auc": 0.7937935574323361,
|
| 898 |
+
"eval_runtime": 21.0756,
|
| 899 |
+
"eval_samples_per_second": 78.716,
|
| 900 |
+
"eval_steps_per_second": 9.869,
|
| 901 |
+
"step": 10790
|
| 902 |
}
|
| 903 |
],
|
| 904 |
+
"logging_steps": 100,
|
| 905 |
+
"max_steps": 12450,
|
| 906 |
+
"num_input_tokens_seen": 0,
|
| 907 |
+
"num_train_epochs": 15,
|
| 908 |
+
"save_steps": 500,
|
| 909 |
+
"stateful_callbacks": {
|
| 910 |
+
"TrainerControl": {
|
| 911 |
+
"args": {
|
| 912 |
+
"should_epoch_stop": false,
|
| 913 |
+
"should_evaluate": false,
|
| 914 |
+
"should_log": false,
|
| 915 |
+
"should_save": true,
|
| 916 |
+
"should_training_stop": false
|
| 917 |
+
},
|
| 918 |
+
"attributes": {}
|
| 919 |
+
}
|
| 920 |
+
},
|
| 921 |
+
"total_flos": 8035320180521232.0,
|
| 922 |
+
"train_batch_size": 8,
|
| 923 |
"trial_name": null,
|
| 924 |
"trial_params": null
|
| 925 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90c00f8b7311babb09b5cfce1bf4c2db61f426d28c044e987c559cbb8c1af657
|
| 3 |
+
size 5304
|