Upload 6 files
Browse files- config.json +69 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +955 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "VCNC/bert_piezas",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "LABEL_0",
|
| 13 |
+
"1": "LABEL_1",
|
| 14 |
+
"2": "LABEL_2",
|
| 15 |
+
"3": "LABEL_3",
|
| 16 |
+
"4": "LABEL_4",
|
| 17 |
+
"5": "LABEL_5",
|
| 18 |
+
"6": "LABEL_6",
|
| 19 |
+
"7": "LABEL_7",
|
| 20 |
+
"8": "LABEL_8",
|
| 21 |
+
"9": "LABEL_9",
|
| 22 |
+
"10": "LABEL_10",
|
| 23 |
+
"11": "LABEL_11",
|
| 24 |
+
"12": "LABEL_12",
|
| 25 |
+
"13": "LABEL_13",
|
| 26 |
+
"14": "LABEL_14",
|
| 27 |
+
"15": "LABEL_15",
|
| 28 |
+
"16": "LABEL_16",
|
| 29 |
+
"17": "LABEL_17",
|
| 30 |
+
"18": "LABEL_18",
|
| 31 |
+
"19": "LABEL_19"
|
| 32 |
+
},
|
| 33 |
+
"initializer_range": 0.02,
|
| 34 |
+
"intermediate_size": 3072,
|
| 35 |
+
"label2id": {
|
| 36 |
+
"LABEL_0": 0,
|
| 37 |
+
"LABEL_1": 1,
|
| 38 |
+
"LABEL_10": 10,
|
| 39 |
+
"LABEL_11": 11,
|
| 40 |
+
"LABEL_12": 12,
|
| 41 |
+
"LABEL_13": 13,
|
| 42 |
+
"LABEL_14": 14,
|
| 43 |
+
"LABEL_15": 15,
|
| 44 |
+
"LABEL_16": 16,
|
| 45 |
+
"LABEL_17": 17,
|
| 46 |
+
"LABEL_18": 18,
|
| 47 |
+
"LABEL_19": 19,
|
| 48 |
+
"LABEL_2": 2,
|
| 49 |
+
"LABEL_3": 3,
|
| 50 |
+
"LABEL_4": 4,
|
| 51 |
+
"LABEL_5": 5,
|
| 52 |
+
"LABEL_6": 6,
|
| 53 |
+
"LABEL_7": 7,
|
| 54 |
+
"LABEL_8": 8,
|
| 55 |
+
"LABEL_9": 9
|
| 56 |
+
},
|
| 57 |
+
"layer_norm_eps": 1e-12,
|
| 58 |
+
"max_position_embeddings": 1024,
|
| 59 |
+
"model_type": "bert",
|
| 60 |
+
"num_attention_heads": 12,
|
| 61 |
+
"num_hidden_layers": 12,
|
| 62 |
+
"pad_token_id": 13,
|
| 63 |
+
"position_embedding_type": "absolute",
|
| 64 |
+
"torch_dtype": "float32",
|
| 65 |
+
"transformers_version": "4.33.1",
|
| 66 |
+
"type_vocab_size": 2,
|
| 67 |
+
"use_cache": true,
|
| 68 |
+
"vocab_size": 21
|
| 69 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5f0c2c25f31c716c9094ed5f66a20de3d464e8d1fce15527ad7d4908895d277a
|
| 3 |
+
size 345932593
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69f1ecf64fea95b4471cfc6855a05cef4190420533ad9e5047ad842898cea5af
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a10d618e59605b6eb6ac6c951b30aac00c12892ef101ca96d77006c6cc9f8cdc
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,955 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 16.69195751138088,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 66000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.13,
|
| 13 |
+
"learning_rate": 2.9776845487815763e-05,
|
| 14 |
+
"loss": 0.169,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.25,
|
| 19 |
+
"learning_rate": 2.955369097563153e-05,
|
| 20 |
+
"loss": 0.1725,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.38,
|
| 25 |
+
"learning_rate": 2.9330536463447292e-05,
|
| 26 |
+
"loss": 0.169,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.51,
|
| 31 |
+
"learning_rate": 2.910738195126306e-05,
|
| 32 |
+
"loss": 0.1688,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.63,
|
| 37 |
+
"learning_rate": 2.8884227439078818e-05,
|
| 38 |
+
"loss": 0.1694,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.76,
|
| 43 |
+
"learning_rate": 2.866107292689458e-05,
|
| 44 |
+
"loss": 0.1682,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.89,
|
| 49 |
+
"learning_rate": 2.8437918414710347e-05,
|
| 50 |
+
"loss": 0.1686,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 1.0,
|
| 55 |
+
"eval_accuracy": 0.4347826086956522,
|
| 56 |
+
"eval_loss": 0.10843166708946228,
|
| 57 |
+
"eval_runtime": 17.9782,
|
| 58 |
+
"eval_samples_per_second": 55.011,
|
| 59 |
+
"eval_steps_per_second": 2.336,
|
| 60 |
+
"step": 3954
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.01,
|
| 64 |
+
"learning_rate": 2.821476390252611e-05,
|
| 65 |
+
"loss": 0.1685,
|
| 66 |
+
"step": 4000
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.14,
|
| 70 |
+
"learning_rate": 2.7991609390341872e-05,
|
| 71 |
+
"loss": 0.1695,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.26,
|
| 76 |
+
"learning_rate": 2.7768454878157638e-05,
|
| 77 |
+
"loss": 0.1727,
|
| 78 |
+
"step": 5000
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.39,
|
| 82 |
+
"learning_rate": 2.75453003659734e-05,
|
| 83 |
+
"loss": 0.166,
|
| 84 |
+
"step": 5500
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.52,
|
| 88 |
+
"learning_rate": 2.7322145853789167e-05,
|
| 89 |
+
"loss": 0.1673,
|
| 90 |
+
"step": 6000
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.64,
|
| 94 |
+
"learning_rate": 2.709899134160493e-05,
|
| 95 |
+
"loss": 0.1697,
|
| 96 |
+
"step": 6500
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.77,
|
| 100 |
+
"learning_rate": 2.687583682942069e-05,
|
| 101 |
+
"loss": 0.1662,
|
| 102 |
+
"step": 7000
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.9,
|
| 106 |
+
"learning_rate": 2.6652682317236455e-05,
|
| 107 |
+
"loss": 0.1683,
|
| 108 |
+
"step": 7500
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"eval_accuracy": 0.47927199191102127,
|
| 113 |
+
"eval_loss": 0.10892220586538315,
|
| 114 |
+
"eval_runtime": 17.9026,
|
| 115 |
+
"eval_samples_per_second": 55.243,
|
| 116 |
+
"eval_steps_per_second": 2.346,
|
| 117 |
+
"step": 7908
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 2.02,
|
| 121 |
+
"learning_rate": 2.6429527805052218e-05,
|
| 122 |
+
"loss": 0.1704,
|
| 123 |
+
"step": 8000
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.15,
|
| 127 |
+
"learning_rate": 2.6206373292867984e-05,
|
| 128 |
+
"loss": 0.1679,
|
| 129 |
+
"step": 8500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 2.28,
|
| 133 |
+
"learning_rate": 2.5983218780683747e-05,
|
| 134 |
+
"loss": 0.1686,
|
| 135 |
+
"step": 9000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 2.4,
|
| 139 |
+
"learning_rate": 2.576006426849951e-05,
|
| 140 |
+
"loss": 0.1667,
|
| 141 |
+
"step": 9500
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 2.53,
|
| 145 |
+
"learning_rate": 2.5536909756315275e-05,
|
| 146 |
+
"loss": 0.1674,
|
| 147 |
+
"step": 10000
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 2.66,
|
| 151 |
+
"learning_rate": 2.5313755244131038e-05,
|
| 152 |
+
"loss": 0.1715,
|
| 153 |
+
"step": 10500
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.78,
|
| 157 |
+
"learning_rate": 2.50906007319468e-05,
|
| 158 |
+
"loss": 0.1679,
|
| 159 |
+
"step": 11000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.91,
|
| 163 |
+
"learning_rate": 2.4867446219762564e-05,
|
| 164 |
+
"loss": 0.1667,
|
| 165 |
+
"step": 11500
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.4347826086956522,
|
| 170 |
+
"eval_loss": 0.10988607257604599,
|
| 171 |
+
"eval_runtime": 17.9778,
|
| 172 |
+
"eval_samples_per_second": 55.012,
|
| 173 |
+
"eval_steps_per_second": 2.336,
|
| 174 |
+
"step": 11862
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 3.03,
|
| 178 |
+
"learning_rate": 2.4644291707578326e-05,
|
| 179 |
+
"loss": 0.1695,
|
| 180 |
+
"step": 12000
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 3.16,
|
| 184 |
+
"learning_rate": 2.4421137195394092e-05,
|
| 185 |
+
"loss": 0.168,
|
| 186 |
+
"step": 12500
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.29,
|
| 190 |
+
"learning_rate": 2.4197982683209855e-05,
|
| 191 |
+
"loss": 0.1663,
|
| 192 |
+
"step": 13000
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 3.41,
|
| 196 |
+
"learning_rate": 2.3974828171025618e-05,
|
| 197 |
+
"loss": 0.1667,
|
| 198 |
+
"step": 13500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 3.54,
|
| 202 |
+
"learning_rate": 2.3751673658841384e-05,
|
| 203 |
+
"loss": 0.1682,
|
| 204 |
+
"step": 14000
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.67,
|
| 208 |
+
"learning_rate": 2.3528519146657147e-05,
|
| 209 |
+
"loss": 0.1672,
|
| 210 |
+
"step": 14500
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 3.79,
|
| 214 |
+
"learning_rate": 2.330536463447291e-05,
|
| 215 |
+
"loss": 0.1701,
|
| 216 |
+
"step": 15000
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 3.92,
|
| 220 |
+
"learning_rate": 2.3082210122288672e-05,
|
| 221 |
+
"loss": 0.171,
|
| 222 |
+
"step": 15500
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.0,
|
| 226 |
+
"eval_accuracy": 0.47927199191102127,
|
| 227 |
+
"eval_loss": 0.10811471939086914,
|
| 228 |
+
"eval_runtime": 17.6966,
|
| 229 |
+
"eval_samples_per_second": 55.887,
|
| 230 |
+
"eval_steps_per_second": 2.373,
|
| 231 |
+
"step": 15816
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 4.05,
|
| 235 |
+
"learning_rate": 2.2859055610104435e-05,
|
| 236 |
+
"loss": 0.1682,
|
| 237 |
+
"step": 16000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 4.17,
|
| 241 |
+
"learning_rate": 2.26359010979202e-05,
|
| 242 |
+
"loss": 0.1706,
|
| 243 |
+
"step": 16500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.3,
|
| 247 |
+
"learning_rate": 2.2412746585735964e-05,
|
| 248 |
+
"loss": 0.1687,
|
| 249 |
+
"step": 17000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.43,
|
| 253 |
+
"learning_rate": 2.2189592073551726e-05,
|
| 254 |
+
"loss": 0.1664,
|
| 255 |
+
"step": 17500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 4.55,
|
| 259 |
+
"learning_rate": 2.1966437561367493e-05,
|
| 260 |
+
"loss": 0.168,
|
| 261 |
+
"step": 18000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 4.68,
|
| 265 |
+
"learning_rate": 2.1743283049183255e-05,
|
| 266 |
+
"loss": 0.1681,
|
| 267 |
+
"step": 18500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 4.81,
|
| 271 |
+
"learning_rate": 2.152012853699902e-05,
|
| 272 |
+
"loss": 0.1697,
|
| 273 |
+
"step": 19000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 4.93,
|
| 277 |
+
"learning_rate": 2.1296974024814784e-05,
|
| 278 |
+
"loss": 0.1681,
|
| 279 |
+
"step": 19500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 5.0,
|
| 283 |
+
"eval_accuracy": 0.47927199191102127,
|
| 284 |
+
"eval_loss": 0.10968898236751556,
|
| 285 |
+
"eval_runtime": 18.0538,
|
| 286 |
+
"eval_samples_per_second": 54.781,
|
| 287 |
+
"eval_steps_per_second": 2.326,
|
| 288 |
+
"step": 19770
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.06,
|
| 292 |
+
"learning_rate": 2.1073819512630543e-05,
|
| 293 |
+
"loss": 0.1684,
|
| 294 |
+
"step": 20000
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 5.18,
|
| 298 |
+
"learning_rate": 2.085066500044631e-05,
|
| 299 |
+
"loss": 0.1692,
|
| 300 |
+
"step": 20500
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 5.31,
|
| 304 |
+
"learning_rate": 2.0627510488262072e-05,
|
| 305 |
+
"loss": 0.1689,
|
| 306 |
+
"step": 21000
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.44,
|
| 310 |
+
"learning_rate": 2.0404355976077835e-05,
|
| 311 |
+
"loss": 0.1694,
|
| 312 |
+
"step": 21500
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 5.56,
|
| 316 |
+
"learning_rate": 2.01812014638936e-05,
|
| 317 |
+
"loss": 0.1699,
|
| 318 |
+
"step": 22000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 5.69,
|
| 322 |
+
"learning_rate": 1.9958046951709364e-05,
|
| 323 |
+
"loss": 0.1705,
|
| 324 |
+
"step": 22500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.82,
|
| 328 |
+
"learning_rate": 1.973489243952513e-05,
|
| 329 |
+
"loss": 0.1661,
|
| 330 |
+
"step": 23000
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 5.94,
|
| 334 |
+
"learning_rate": 1.9511737927340893e-05,
|
| 335 |
+
"loss": 0.1646,
|
| 336 |
+
"step": 23500
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 6.0,
|
| 340 |
+
"eval_accuracy": 0.4347826086956522,
|
| 341 |
+
"eval_loss": 0.11061158031225204,
|
| 342 |
+
"eval_runtime": 17.7602,
|
| 343 |
+
"eval_samples_per_second": 55.686,
|
| 344 |
+
"eval_steps_per_second": 2.365,
|
| 345 |
+
"step": 23724
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.07,
|
| 349 |
+
"learning_rate": 1.9288583415156655e-05,
|
| 350 |
+
"loss": 0.1663,
|
| 351 |
+
"step": 24000
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 6.2,
|
| 355 |
+
"learning_rate": 1.9065428902972418e-05,
|
| 356 |
+
"loss": 0.1677,
|
| 357 |
+
"step": 24500
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 6.32,
|
| 361 |
+
"learning_rate": 1.884227439078818e-05,
|
| 362 |
+
"loss": 0.1688,
|
| 363 |
+
"step": 25000
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.45,
|
| 367 |
+
"learning_rate": 1.8619119878603947e-05,
|
| 368 |
+
"loss": 0.1692,
|
| 369 |
+
"step": 25500
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 6.58,
|
| 373 |
+
"learning_rate": 1.839596536641971e-05,
|
| 374 |
+
"loss": 0.1666,
|
| 375 |
+
"step": 26000
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 6.7,
|
| 379 |
+
"learning_rate": 1.8172810854235472e-05,
|
| 380 |
+
"loss": 0.168,
|
| 381 |
+
"step": 26500
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 6.83,
|
| 385 |
+
"learning_rate": 1.794965634205124e-05,
|
| 386 |
+
"loss": 0.1704,
|
| 387 |
+
"step": 27000
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.95,
|
| 391 |
+
"learning_rate": 1.7726501829867e-05,
|
| 392 |
+
"loss": 0.1679,
|
| 393 |
+
"step": 27500
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 7.0,
|
| 397 |
+
"eval_accuracy": 0.47927199191102127,
|
| 398 |
+
"eval_loss": 0.10994414240121841,
|
| 399 |
+
"eval_runtime": 17.7058,
|
| 400 |
+
"eval_samples_per_second": 55.858,
|
| 401 |
+
"eval_steps_per_second": 2.372,
|
| 402 |
+
"step": 27678
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.08,
|
| 406 |
+
"learning_rate": 1.7503347317682764e-05,
|
| 407 |
+
"loss": 0.168,
|
| 408 |
+
"step": 28000
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.21,
|
| 412 |
+
"learning_rate": 1.728019280549853e-05,
|
| 413 |
+
"loss": 0.1657,
|
| 414 |
+
"step": 28500
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.33,
|
| 418 |
+
"learning_rate": 1.705703829331429e-05,
|
| 419 |
+
"loss": 0.1688,
|
| 420 |
+
"step": 29000
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.46,
|
| 424 |
+
"learning_rate": 1.6833883781130055e-05,
|
| 425 |
+
"loss": 0.1707,
|
| 426 |
+
"step": 29500
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.59,
|
| 430 |
+
"learning_rate": 1.6610729268945818e-05,
|
| 431 |
+
"loss": 0.1676,
|
| 432 |
+
"step": 30000
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.71,
|
| 436 |
+
"learning_rate": 1.638757475676158e-05,
|
| 437 |
+
"loss": 0.1696,
|
| 438 |
+
"step": 30500
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.84,
|
| 442 |
+
"learning_rate": 1.6164420244577347e-05,
|
| 443 |
+
"loss": 0.1701,
|
| 444 |
+
"step": 31000
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.97,
|
| 448 |
+
"learning_rate": 1.594126573239311e-05,
|
| 449 |
+
"loss": 0.1645,
|
| 450 |
+
"step": 31500
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 8.0,
|
| 454 |
+
"eval_accuracy": 0.47927199191102127,
|
| 455 |
+
"eval_loss": 0.10965029150247574,
|
| 456 |
+
"eval_runtime": 17.9019,
|
| 457 |
+
"eval_samples_per_second": 55.246,
|
| 458 |
+
"eval_steps_per_second": 2.346,
|
| 459 |
+
"step": 31632
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 8.09,
|
| 463 |
+
"learning_rate": 1.5718111220208872e-05,
|
| 464 |
+
"loss": 0.1701,
|
| 465 |
+
"step": 32000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.22,
|
| 469 |
+
"learning_rate": 1.549495670802464e-05,
|
| 470 |
+
"loss": 0.1669,
|
| 471 |
+
"step": 32500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 8.35,
|
| 475 |
+
"learning_rate": 1.52718021958404e-05,
|
| 476 |
+
"loss": 0.1664,
|
| 477 |
+
"step": 33000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 8.47,
|
| 481 |
+
"learning_rate": 1.5048647683656162e-05,
|
| 482 |
+
"loss": 0.1688,
|
| 483 |
+
"step": 33500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.6,
|
| 487 |
+
"learning_rate": 1.4825493171471928e-05,
|
| 488 |
+
"loss": 0.1684,
|
| 489 |
+
"step": 34000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 8.73,
|
| 493 |
+
"learning_rate": 1.4602338659287691e-05,
|
| 494 |
+
"loss": 0.1702,
|
| 495 |
+
"step": 34500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 8.85,
|
| 499 |
+
"learning_rate": 1.4379184147103455e-05,
|
| 500 |
+
"loss": 0.1644,
|
| 501 |
+
"step": 35000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 8.98,
|
| 505 |
+
"learning_rate": 1.4156029634919218e-05,
|
| 506 |
+
"loss": 0.1694,
|
| 507 |
+
"step": 35500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 9.0,
|
| 511 |
+
"eval_accuracy": 0.47927199191102127,
|
| 512 |
+
"eval_loss": 0.10830555856227875,
|
| 513 |
+
"eval_runtime": 17.853,
|
| 514 |
+
"eval_samples_per_second": 55.397,
|
| 515 |
+
"eval_steps_per_second": 2.353,
|
| 516 |
+
"step": 35586
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 9.1,
|
| 520 |
+
"learning_rate": 1.3932875122734983e-05,
|
| 521 |
+
"loss": 0.1685,
|
| 522 |
+
"step": 36000
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 9.23,
|
| 526 |
+
"learning_rate": 1.3709720610550745e-05,
|
| 527 |
+
"loss": 0.1672,
|
| 528 |
+
"step": 36500
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 9.36,
|
| 532 |
+
"learning_rate": 1.348656609836651e-05,
|
| 533 |
+
"loss": 0.1703,
|
| 534 |
+
"step": 37000
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 9.48,
|
| 538 |
+
"learning_rate": 1.3263411586182272e-05,
|
| 539 |
+
"loss": 0.1712,
|
| 540 |
+
"step": 37500
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 9.61,
|
| 544 |
+
"learning_rate": 1.3040257073998037e-05,
|
| 545 |
+
"loss": 0.1691,
|
| 546 |
+
"step": 38000
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 9.74,
|
| 550 |
+
"learning_rate": 1.2817102561813801e-05,
|
| 551 |
+
"loss": 0.1662,
|
| 552 |
+
"step": 38500
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 9.86,
|
| 556 |
+
"learning_rate": 1.2593948049629564e-05,
|
| 557 |
+
"loss": 0.1667,
|
| 558 |
+
"step": 39000
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 9.99,
|
| 562 |
+
"learning_rate": 1.2370793537445327e-05,
|
| 563 |
+
"loss": 0.1637,
|
| 564 |
+
"step": 39500
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 10.0,
|
| 568 |
+
"eval_accuracy": 0.47927199191102127,
|
| 569 |
+
"eval_loss": 0.10983388870954514,
|
| 570 |
+
"eval_runtime": 18.0727,
|
| 571 |
+
"eval_samples_per_second": 54.723,
|
| 572 |
+
"eval_steps_per_second": 2.324,
|
| 573 |
+
"step": 39540
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 10.12,
|
| 577 |
+
"learning_rate": 1.2147639025261091e-05,
|
| 578 |
+
"loss": 0.1646,
|
| 579 |
+
"step": 40000
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 10.24,
|
| 583 |
+
"learning_rate": 1.1924484513076856e-05,
|
| 584 |
+
"loss": 0.1673,
|
| 585 |
+
"step": 40500
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 10.37,
|
| 589 |
+
"learning_rate": 1.1701330000892618e-05,
|
| 590 |
+
"loss": 0.1681,
|
| 591 |
+
"step": 41000
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 10.5,
|
| 595 |
+
"learning_rate": 1.1478175488708381e-05,
|
| 596 |
+
"loss": 0.1673,
|
| 597 |
+
"step": 41500
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 10.62,
|
| 601 |
+
"learning_rate": 1.1255020976524145e-05,
|
| 602 |
+
"loss": 0.1676,
|
| 603 |
+
"step": 42000
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 10.75,
|
| 607 |
+
"learning_rate": 1.103186646433991e-05,
|
| 608 |
+
"loss": 0.1679,
|
| 609 |
+
"step": 42500
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 10.88,
|
| 613 |
+
"learning_rate": 1.0808711952155672e-05,
|
| 614 |
+
"loss": 0.1682,
|
| 615 |
+
"step": 43000
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 11.0,
|
| 619 |
+
"eval_accuracy": 0.4246713852376138,
|
| 620 |
+
"eval_loss": 0.10815968364477158,
|
| 621 |
+
"eval_runtime": 17.8822,
|
| 622 |
+
"eval_samples_per_second": 55.306,
|
| 623 |
+
"eval_steps_per_second": 2.349,
|
| 624 |
+
"step": 43494
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 11.0,
|
| 628 |
+
"learning_rate": 1.0585557439971437e-05,
|
| 629 |
+
"loss": 0.1714,
|
| 630 |
+
"step": 43500
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 11.13,
|
| 634 |
+
"learning_rate": 1.03624029277872e-05,
|
| 635 |
+
"loss": 0.1644,
|
| 636 |
+
"step": 44000
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 11.25,
|
| 640 |
+
"learning_rate": 1.0139248415602964e-05,
|
| 641 |
+
"loss": 0.1675,
|
| 642 |
+
"step": 44500
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 11.38,
|
| 646 |
+
"learning_rate": 9.916093903418728e-06,
|
| 647 |
+
"loss": 0.1655,
|
| 648 |
+
"step": 45000
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 11.51,
|
| 652 |
+
"learning_rate": 9.692939391234491e-06,
|
| 653 |
+
"loss": 0.1697,
|
| 654 |
+
"step": 45500
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 11.63,
|
| 658 |
+
"learning_rate": 9.469784879050254e-06,
|
| 659 |
+
"loss": 0.1682,
|
| 660 |
+
"step": 46000
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 11.76,
|
| 664 |
+
"learning_rate": 9.246630366866018e-06,
|
| 665 |
+
"loss": 0.1674,
|
| 666 |
+
"step": 46500
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 11.89,
|
| 670 |
+
"learning_rate": 9.023475854681783e-06,
|
| 671 |
+
"loss": 0.1706,
|
| 672 |
+
"step": 47000
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 12.0,
|
| 676 |
+
"eval_accuracy": 0.47927199191102127,
|
| 677 |
+
"eval_loss": 0.10891950130462646,
|
| 678 |
+
"eval_runtime": 17.7023,
|
| 679 |
+
"eval_samples_per_second": 55.868,
|
| 680 |
+
"eval_steps_per_second": 2.373,
|
| 681 |
+
"step": 47448
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 12.01,
|
| 685 |
+
"learning_rate": 8.800321342497545e-06,
|
| 686 |
+
"loss": 0.1684,
|
| 687 |
+
"step": 47500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 12.14,
|
| 691 |
+
"learning_rate": 8.577166830313308e-06,
|
| 692 |
+
"loss": 0.1662,
|
| 693 |
+
"step": 48000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 12.27,
|
| 697 |
+
"learning_rate": 8.354012318129073e-06,
|
| 698 |
+
"loss": 0.1674,
|
| 699 |
+
"step": 48500
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 12.39,
|
| 703 |
+
"learning_rate": 8.130857805944837e-06,
|
| 704 |
+
"loss": 0.1662,
|
| 705 |
+
"step": 49000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 12.52,
|
| 709 |
+
"learning_rate": 7.907703293760601e-06,
|
| 710 |
+
"loss": 0.1681,
|
| 711 |
+
"step": 49500
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 12.65,
|
| 715 |
+
"learning_rate": 7.684548781576362e-06,
|
| 716 |
+
"loss": 0.1701,
|
| 717 |
+
"step": 50000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 12.77,
|
| 721 |
+
"learning_rate": 7.461394269392127e-06,
|
| 722 |
+
"loss": 0.1669,
|
| 723 |
+
"step": 50500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 12.9,
|
| 727 |
+
"learning_rate": 7.238239757207891e-06,
|
| 728 |
+
"loss": 0.1689,
|
| 729 |
+
"step": 51000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 13.0,
|
| 733 |
+
"eval_accuracy": 0.47927199191102127,
|
| 734 |
+
"eval_loss": 0.1080804243683815,
|
| 735 |
+
"eval_runtime": 18.0476,
|
| 736 |
+
"eval_samples_per_second": 54.799,
|
| 737 |
+
"eval_steps_per_second": 2.327,
|
| 738 |
+
"step": 51402
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 13.02,
|
| 742 |
+
"learning_rate": 7.015085245023655e-06,
|
| 743 |
+
"loss": 0.1697,
|
| 744 |
+
"step": 51500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 13.15,
|
| 748 |
+
"learning_rate": 6.791930732839418e-06,
|
| 749 |
+
"loss": 0.1666,
|
| 750 |
+
"step": 52000
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 13.28,
|
| 754 |
+
"learning_rate": 6.568776220655182e-06,
|
| 755 |
+
"loss": 0.1695,
|
| 756 |
+
"step": 52500
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 13.4,
|
| 760 |
+
"learning_rate": 6.3456217084709455e-06,
|
| 761 |
+
"loss": 0.1689,
|
| 762 |
+
"step": 53000
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 13.53,
|
| 766 |
+
"learning_rate": 6.122467196286709e-06,
|
| 767 |
+
"loss": 0.1683,
|
| 768 |
+
"step": 53500
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 13.66,
|
| 772 |
+
"learning_rate": 5.899312684102473e-06,
|
| 773 |
+
"loss": 0.1645,
|
| 774 |
+
"step": 54000
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 13.78,
|
| 778 |
+
"learning_rate": 5.676158171918236e-06,
|
| 779 |
+
"loss": 0.1666,
|
| 780 |
+
"step": 54500
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 13.91,
|
| 784 |
+
"learning_rate": 5.453003659734e-06,
|
| 785 |
+
"loss": 0.1677,
|
| 786 |
+
"step": 55000
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 14.0,
|
| 790 |
+
"eval_accuracy": 0.47927199191102127,
|
| 791 |
+
"eval_loss": 0.10935764014720917,
|
| 792 |
+
"eval_runtime": 17.849,
|
| 793 |
+
"eval_samples_per_second": 55.409,
|
| 794 |
+
"eval_steps_per_second": 2.353,
|
| 795 |
+
"step": 55356
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 14.04,
|
| 799 |
+
"learning_rate": 5.229849147549764e-06,
|
| 800 |
+
"loss": 0.1659,
|
| 801 |
+
"step": 55500
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 14.16,
|
| 805 |
+
"learning_rate": 5.006694635365527e-06,
|
| 806 |
+
"loss": 0.1641,
|
| 807 |
+
"step": 56000
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 14.29,
|
| 811 |
+
"learning_rate": 4.783540123181291e-06,
|
| 812 |
+
"loss": 0.1692,
|
| 813 |
+
"step": 56500
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 14.42,
|
| 817 |
+
"learning_rate": 4.560385610997054e-06,
|
| 818 |
+
"loss": 0.1699,
|
| 819 |
+
"step": 57000
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 14.54,
|
| 823 |
+
"learning_rate": 4.3372310988128184e-06,
|
| 824 |
+
"loss": 0.1687,
|
| 825 |
+
"step": 57500
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 14.67,
|
| 829 |
+
"learning_rate": 4.114076586628581e-06,
|
| 830 |
+
"loss": 0.1659,
|
| 831 |
+
"step": 58000
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 14.8,
|
| 835 |
+
"learning_rate": 3.8909220744443456e-06,
|
| 836 |
+
"loss": 0.1651,
|
| 837 |
+
"step": 58500
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 14.92,
|
| 841 |
+
"learning_rate": 3.667767562260109e-06,
|
| 842 |
+
"loss": 0.1684,
|
| 843 |
+
"step": 59000
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 15.0,
|
| 847 |
+
"eval_accuracy": 0.47927199191102127,
|
| 848 |
+
"eval_loss": 0.10881149023771286,
|
| 849 |
+
"eval_runtime": 17.735,
|
| 850 |
+
"eval_samples_per_second": 55.765,
|
| 851 |
+
"eval_steps_per_second": 2.368,
|
| 852 |
+
"step": 59310
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 15.05,
|
| 856 |
+
"learning_rate": 3.4446130500758727e-06,
|
| 857 |
+
"loss": 0.1695,
|
| 858 |
+
"step": 59500
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 15.17,
|
| 862 |
+
"learning_rate": 3.2214585378916363e-06,
|
| 863 |
+
"loss": 0.1679,
|
| 864 |
+
"step": 60000
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 15.3,
|
| 868 |
+
"learning_rate": 2.9983040257074e-06,
|
| 869 |
+
"loss": 0.1677,
|
| 870 |
+
"step": 60500
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 15.43,
|
| 874 |
+
"learning_rate": 2.7751495135231634e-06,
|
| 875 |
+
"loss": 0.1656,
|
| 876 |
+
"step": 61000
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 15.55,
|
| 880 |
+
"learning_rate": 2.551995001338927e-06,
|
| 881 |
+
"loss": 0.1675,
|
| 882 |
+
"step": 61500
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 15.68,
|
| 886 |
+
"learning_rate": 2.3288404891546905e-06,
|
| 887 |
+
"loss": 0.1706,
|
| 888 |
+
"step": 62000
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 15.81,
|
| 892 |
+
"learning_rate": 2.105685976970454e-06,
|
| 893 |
+
"loss": 0.1653,
|
| 894 |
+
"step": 62500
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 15.93,
|
| 898 |
+
"learning_rate": 1.8825314647862179e-06,
|
| 899 |
+
"loss": 0.1686,
|
| 900 |
+
"step": 63000
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 16.0,
|
| 904 |
+
"eval_accuracy": 0.47927199191102127,
|
| 905 |
+
"eval_loss": 0.10850092023611069,
|
| 906 |
+
"eval_runtime": 18.097,
|
| 907 |
+
"eval_samples_per_second": 54.65,
|
| 908 |
+
"eval_steps_per_second": 2.321,
|
| 909 |
+
"step": 63264
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 16.06,
|
| 913 |
+
"learning_rate": 1.6593769526019816e-06,
|
| 914 |
+
"loss": 0.1642,
|
| 915 |
+
"step": 63500
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 16.19,
|
| 919 |
+
"learning_rate": 1.4362224404177452e-06,
|
| 920 |
+
"loss": 0.1664,
|
| 921 |
+
"step": 64000
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 16.31,
|
| 925 |
+
"learning_rate": 1.213067928233509e-06,
|
| 926 |
+
"loss": 0.1655,
|
| 927 |
+
"step": 64500
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 16.44,
|
| 931 |
+
"learning_rate": 9.899134160492725e-07,
|
| 932 |
+
"loss": 0.1693,
|
| 933 |
+
"step": 65000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 16.57,
|
| 937 |
+
"learning_rate": 7.667589038650362e-07,
|
| 938 |
+
"loss": 0.1696,
|
| 939 |
+
"step": 65500
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 16.69,
|
| 943 |
+
"learning_rate": 5.436043916807998e-07,
|
| 944 |
+
"loss": 0.1644,
|
| 945 |
+
"step": 66000
|
| 946 |
+
}
|
| 947 |
+
],
|
| 948 |
+
"logging_steps": 500,
|
| 949 |
+
"max_steps": 67218,
|
| 950 |
+
"num_train_epochs": 17,
|
| 951 |
+
"save_steps": 1500,
|
| 952 |
+
"total_flos": 9939187287216000.0,
|
| 953 |
+
"trial_name": null,
|
| 954 |
+
"trial_params": null
|
| 955 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e31edcdc38b576fb55b071c4c6d489d083b8fb136746deb0915b8c9820558530
|
| 3 |
+
size 4027
|