Upload folder using huggingface_hub
Browse files- config.json +34 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +13 -0
- trainer_state.json +753 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "bert-base-uncased",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "NEGATIVE",
|
| 14 |
+
"1": "POSITIVE"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 3072,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"NEGATIVE": 0,
|
| 20 |
+
"POSITIVE": 1
|
| 21 |
+
},
|
| 22 |
+
"layer_norm_eps": 1e-12,
|
| 23 |
+
"max_position_embeddings": 512,
|
| 24 |
+
"model_type": "bert",
|
| 25 |
+
"num_attention_heads": 12,
|
| 26 |
+
"num_hidden_layers": 12,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.28.0.dev0",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 30522
|
| 34 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f81199af534cde8c67b860d7edaff323e0be85af284a0fbbff319f2570145ee
|
| 3 |
+
size 875988101
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:87e00ea5d4f2be8d0fdf5de58eefdcdcb32aeae130f361e4554b2912abf49c41
|
| 3 |
+
size 438007925
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0893551351a60eae0177f37d8f192dc96139787f0a29e2546f8c90e33e13385
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4171b430d5c3ba093b3799f5b5633038eeedea420f2e0b69e2e319a304183e1b
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"do_lower_case": true,
|
| 4 |
+
"mask_token": "[MASK]",
|
| 5 |
+
"model_max_length": 512,
|
| 6 |
+
"pad_token": "[PAD]",
|
| 7 |
+
"sep_token": "[SEP]",
|
| 8 |
+
"special_tokens_map_file": null,
|
| 9 |
+
"strip_accents": null,
|
| 10 |
+
"tokenize_chinese_chars": true,
|
| 11 |
+
"tokenizer_class": "BertTokenizer",
|
| 12 |
+
"unk_token": "[UNK]"
|
| 13 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,753 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 350,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0,
|
| 12 |
+
"step": 0,
|
| 13 |
+
"train_loss": 0.9102265238761902
|
| 14 |
+
},
|
| 15 |
+
{
|
| 16 |
+
"epoch": 0,
|
| 17 |
+
"step": 0,
|
| 18 |
+
"train_loss": 0.7791410684585571
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 0.29,
|
| 22 |
+
"learning_rate": 1.942857142857143e-05,
|
| 23 |
+
"loss": 0.6171,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.29,
|
| 28 |
+
"step": 20,
|
| 29 |
+
"train_loss": 0.9948083162307739
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.29,
|
| 33 |
+
"step": 20,
|
| 34 |
+
"train_loss": 0.43645456433296204
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.57,
|
| 38 |
+
"learning_rate": 1.885714285714286e-05,
|
| 39 |
+
"loss": 0.5985,
|
| 40 |
+
"step": 40
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 0.57,
|
| 44 |
+
"step": 40,
|
| 45 |
+
"train_loss": 0.6493304967880249
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.57,
|
| 49 |
+
"step": 40,
|
| 50 |
+
"train_loss": 0.44140735268592834
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.86,
|
| 54 |
+
"learning_rate": 1.8285714285714288e-05,
|
| 55 |
+
"loss": 0.5388,
|
| 56 |
+
"step": 60
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.86,
|
| 60 |
+
"step": 60,
|
| 61 |
+
"train_loss": 0.5935375690460205
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.86,
|
| 65 |
+
"step": 60,
|
| 66 |
+
"train_loss": 0.7153045535087585
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.0,
|
| 70 |
+
"eval_accuracy": {
|
| 71 |
+
"accuracy": 0.8339285714285715
|
| 72 |
+
},
|
| 73 |
+
"eval_auc": 0.7155102040816327,
|
| 74 |
+
"eval_f1": {
|
| 75 |
+
"f1": 0.3404255319148936
|
| 76 |
+
},
|
| 77 |
+
"eval_loss": 0.6475747227668762,
|
| 78 |
+
"eval_precision": {
|
| 79 |
+
"precision": 0.3380281690140845
|
| 80 |
+
},
|
| 81 |
+
"eval_recall": {
|
| 82 |
+
"recall": 0.34285714285714286
|
| 83 |
+
},
|
| 84 |
+
"eval_runtime": 0.6529,
|
| 85 |
+
"eval_samples_per_second": 857.718,
|
| 86 |
+
"eval_steps_per_second": 53.607,
|
| 87 |
+
"step": 70
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.14,
|
| 91 |
+
"learning_rate": 1.7714285714285717e-05,
|
| 92 |
+
"loss": 0.5155,
|
| 93 |
+
"step": 80
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.14,
|
| 97 |
+
"step": 80,
|
| 98 |
+
"train_loss": 0.30939409136772156
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.14,
|
| 102 |
+
"step": 80,
|
| 103 |
+
"train_loss": 0.46914660930633545
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 1.43,
|
| 107 |
+
"learning_rate": 1.7142857142857142e-05,
|
| 108 |
+
"loss": 0.4106,
|
| 109 |
+
"step": 100
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.43,
|
| 113 |
+
"step": 100,
|
| 114 |
+
"train_loss": 0.5445544719696045
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 1.43,
|
| 118 |
+
"step": 100,
|
| 119 |
+
"train_loss": 0.2606019377708435
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.71,
|
| 123 |
+
"learning_rate": 1.6571428571428574e-05,
|
| 124 |
+
"loss": 0.4368,
|
| 125 |
+
"step": 120
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.71,
|
| 129 |
+
"step": 120,
|
| 130 |
+
"train_loss": 0.3977287709712982
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.71,
|
| 134 |
+
"step": 120,
|
| 135 |
+
"train_loss": 0.27664005756378174
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 2.0,
|
| 139 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 140 |
+
"loss": 0.3439,
|
| 141 |
+
"step": 140
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 2.0,
|
| 145 |
+
"step": 140,
|
| 146 |
+
"train_loss": 0.281027227640152
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.0,
|
| 150 |
+
"step": 140,
|
| 151 |
+
"train_loss": 0.8136294484138489
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.0,
|
| 155 |
+
"step": 140,
|
| 156 |
+
"train_loss": 0.12487435340881348
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 2.0,
|
| 160 |
+
"step": 140,
|
| 161 |
+
"train_loss": 0.4965817332267761
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 2.0,
|
| 165 |
+
"step": 140,
|
| 166 |
+
"train_loss": 0.10273457318544388
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 2.0,
|
| 170 |
+
"step": 140,
|
| 171 |
+
"train_loss": 0.4042325019836426
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.0,
|
| 175 |
+
"step": 140,
|
| 176 |
+
"train_loss": 0.18423768877983093
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.0,
|
| 180 |
+
"step": 140,
|
| 181 |
+
"train_loss": 0.2865528464317322
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 2.0,
|
| 185 |
+
"step": 140,
|
| 186 |
+
"train_loss": 0.6649780869483948
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.0,
|
| 190 |
+
"step": 140,
|
| 191 |
+
"train_loss": 0.187837615609169
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.0,
|
| 195 |
+
"step": 140,
|
| 196 |
+
"train_loss": 0.5332860946655273
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 2.0,
|
| 200 |
+
"step": 140,
|
| 201 |
+
"train_loss": 1.3017665147781372
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.0,
|
| 205 |
+
"step": 140,
|
| 206 |
+
"train_loss": 0.17419536411762238
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.0,
|
| 210 |
+
"step": 140,
|
| 211 |
+
"train_loss": 0.21602007746696472
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 2.0,
|
| 215 |
+
"step": 140,
|
| 216 |
+
"train_loss": 0.19301216304302216
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 2.0,
|
| 220 |
+
"step": 140,
|
| 221 |
+
"train_loss": 0.7078395485877991
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.0,
|
| 225 |
+
"step": 140,
|
| 226 |
+
"train_loss": 1.2670217752456665
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 2.0,
|
| 230 |
+
"step": 140,
|
| 231 |
+
"train_loss": 1.3781671524047852
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 2.0,
|
| 235 |
+
"step": 140,
|
| 236 |
+
"train_loss": 1.1845088005065918
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.0,
|
| 240 |
+
"step": 140,
|
| 241 |
+
"train_loss": 1.5374538898468018
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 2.0,
|
| 245 |
+
"step": 140,
|
| 246 |
+
"train_loss": 0.8708707690238953
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 2.0,
|
| 250 |
+
"step": 140,
|
| 251 |
+
"train_loss": 0.8791667222976685
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 2.0,
|
| 255 |
+
"step": 140,
|
| 256 |
+
"train_loss": 0.30802324414253235
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.0,
|
| 260 |
+
"step": 140,
|
| 261 |
+
"train_loss": 1.2370085716247559
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 2.0,
|
| 265 |
+
"step": 140,
|
| 266 |
+
"train_loss": 1.2875761985778809
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.0,
|
| 270 |
+
"step": 140,
|
| 271 |
+
"train_loss": 1.280977487564087
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 2.0,
|
| 275 |
+
"step": 140,
|
| 276 |
+
"train_loss": 1.4434828758239746
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 2.0,
|
| 280 |
+
"step": 140,
|
| 281 |
+
"train_loss": 0.9846087098121643
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 2.0,
|
| 285 |
+
"step": 140,
|
| 286 |
+
"train_loss": 0.3032301068305969
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 2.0,
|
| 290 |
+
"step": 140,
|
| 291 |
+
"train_loss": 0.22137752175331116
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 2.0,
|
| 295 |
+
"step": 140,
|
| 296 |
+
"train_loss": 0.8634898066520691
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.0,
|
| 300 |
+
"step": 140,
|
| 301 |
+
"train_loss": 1.080783486366272
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 2.0,
|
| 305 |
+
"step": 140,
|
| 306 |
+
"train_loss": 0.8349682092666626
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 2.0,
|
| 310 |
+
"step": 140,
|
| 311 |
+
"train_loss": 0.6795739531517029
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 2.0,
|
| 315 |
+
"step": 140,
|
| 316 |
+
"train_loss": 0.676697850227356
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 2.0,
|
| 320 |
+
"eval_accuracy": {
|
| 321 |
+
"accuracy": 0.8428571428571429
|
| 322 |
+
},
|
| 323 |
+
"eval_auc": 0.7470845481049562,
|
| 324 |
+
"eval_f1": {
|
| 325 |
+
"f1": 0.3802816901408451
|
| 326 |
+
},
|
| 327 |
+
"eval_loss": 0.7140511870384216,
|
| 328 |
+
"eval_precision": {
|
| 329 |
+
"precision": 0.375
|
| 330 |
+
},
|
| 331 |
+
"eval_recall": {
|
| 332 |
+
"recall": 0.38571428571428573
|
| 333 |
+
},
|
| 334 |
+
"eval_runtime": 0.8142,
|
| 335 |
+
"eval_samples_per_second": 687.816,
|
| 336 |
+
"eval_steps_per_second": 42.989,
|
| 337 |
+
"step": 140
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 2.0,
|
| 341 |
+
"step": 140,
|
| 342 |
+
"train_loss": 0.18269318342208862
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 2.0,
|
| 346 |
+
"step": 140,
|
| 347 |
+
"train_loss": 0.16265031695365906
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.29,
|
| 351 |
+
"learning_rate": 1.542857142857143e-05,
|
| 352 |
+
"loss": 0.22,
|
| 353 |
+
"step": 160
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.29,
|
| 357 |
+
"step": 160,
|
| 358 |
+
"train_loss": 0.15252527594566345
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 2.29,
|
| 362 |
+
"step": 160,
|
| 363 |
+
"train_loss": 0.26980623602867126
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 2.57,
|
| 367 |
+
"learning_rate": 1.4857142857142858e-05,
|
| 368 |
+
"loss": 0.2508,
|
| 369 |
+
"step": 180
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.57,
|
| 373 |
+
"step": 180,
|
| 374 |
+
"train_loss": 0.1389356404542923
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.57,
|
| 378 |
+
"step": 180,
|
| 379 |
+
"train_loss": 0.07751139253377914
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 2.86,
|
| 383 |
+
"learning_rate": 1.4285714285714287e-05,
|
| 384 |
+
"loss": 0.2149,
|
| 385 |
+
"step": 200
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 2.86,
|
| 389 |
+
"step": 200,
|
| 390 |
+
"train_loss": 0.13591702282428741
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 2.86,
|
| 394 |
+
"step": 200,
|
| 395 |
+
"train_loss": 0.04827806353569031
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 3.0,
|
| 399 |
+
"eval_accuracy": {
|
| 400 |
+
"accuracy": 0.8232142857142857
|
| 401 |
+
},
|
| 402 |
+
"eval_auc": 0.7547521865889213,
|
| 403 |
+
"eval_f1": {
|
| 404 |
+
"f1": 0.3926380368098159
|
| 405 |
+
},
|
| 406 |
+
"eval_loss": 0.9247345924377441,
|
| 407 |
+
"eval_precision": {
|
| 408 |
+
"precision": 0.34408602150537637
|
| 409 |
+
},
|
| 410 |
+
"eval_recall": {
|
| 411 |
+
"recall": 0.45714285714285713
|
| 412 |
+
},
|
| 413 |
+
"eval_runtime": 0.6565,
|
| 414 |
+
"eval_samples_per_second": 852.978,
|
| 415 |
+
"eval_steps_per_second": 53.311,
|
| 416 |
+
"step": 210
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 3.14,
|
| 420 |
+
"learning_rate": 1.3714285714285716e-05,
|
| 421 |
+
"loss": 0.2083,
|
| 422 |
+
"step": 220
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 3.14,
|
| 426 |
+
"step": 220,
|
| 427 |
+
"train_loss": 0.08572366833686829
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 3.14,
|
| 431 |
+
"step": 220,
|
| 432 |
+
"train_loss": 0.16466383635997772
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 3.43,
|
| 436 |
+
"learning_rate": 1.3142857142857145e-05,
|
| 437 |
+
"loss": 0.1521,
|
| 438 |
+
"step": 240
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 3.43,
|
| 442 |
+
"step": 240,
|
| 443 |
+
"train_loss": 0.4424760043621063
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 3.43,
|
| 447 |
+
"step": 240,
|
| 448 |
+
"train_loss": 0.0380471907556057
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 3.71,
|
| 452 |
+
"learning_rate": 1.2571428571428572e-05,
|
| 453 |
+
"loss": 0.1872,
|
| 454 |
+
"step": 260
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 3.71,
|
| 458 |
+
"step": 260,
|
| 459 |
+
"train_loss": 0.4089410901069641
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 3.71,
|
| 463 |
+
"step": 260,
|
| 464 |
+
"train_loss": 0.2289683222770691
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.0,
|
| 468 |
+
"learning_rate": 1.2e-05,
|
| 469 |
+
"loss": 0.1486,
|
| 470 |
+
"step": 280
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.0,
|
| 474 |
+
"step": 280,
|
| 475 |
+
"train_loss": 0.4819692373275757
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 4.0,
|
| 479 |
+
"step": 280,
|
| 480 |
+
"train_loss": 1.6059560775756836
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 4.0,
|
| 484 |
+
"step": 280,
|
| 485 |
+
"train_loss": 0.02861696481704712
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 4.0,
|
| 489 |
+
"step": 280,
|
| 490 |
+
"train_loss": 0.6295557618141174
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 4.0,
|
| 494 |
+
"step": 280,
|
| 495 |
+
"train_loss": 0.011990266852080822
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 4.0,
|
| 499 |
+
"step": 280,
|
| 500 |
+
"train_loss": 0.8871378302574158
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 4.0,
|
| 504 |
+
"step": 280,
|
| 505 |
+
"train_loss": 0.96287602186203
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 4.0,
|
| 509 |
+
"step": 280,
|
| 510 |
+
"train_loss": 0.368556946516037
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 4.0,
|
| 514 |
+
"step": 280,
|
| 515 |
+
"train_loss": 0.9895154237747192
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 4.0,
|
| 519 |
+
"step": 280,
|
| 520 |
+
"train_loss": 0.11174698173999786
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 4.0,
|
| 524 |
+
"step": 280,
|
| 525 |
+
"train_loss": 0.6365806460380554
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 4.0,
|
| 529 |
+
"step": 280,
|
| 530 |
+
"train_loss": 1.9894587993621826
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 4.0,
|
| 534 |
+
"step": 280,
|
| 535 |
+
"train_loss": 0.27545443177223206
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 4.0,
|
| 539 |
+
"step": 280,
|
| 540 |
+
"train_loss": 0.15115408599376678
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 4.0,
|
| 544 |
+
"step": 280,
|
| 545 |
+
"train_loss": 0.1443568766117096
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 4.0,
|
| 549 |
+
"step": 280,
|
| 550 |
+
"train_loss": 0.6999047994613647
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 4.0,
|
| 554 |
+
"step": 280,
|
| 555 |
+
"train_loss": 1.7545350790023804
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 4.0,
|
| 559 |
+
"step": 280,
|
| 560 |
+
"train_loss": 2.3427138328552246
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 4.0,
|
| 564 |
+
"step": 280,
|
| 565 |
+
"train_loss": 2.2553696632385254
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 4.0,
|
| 569 |
+
"step": 280,
|
| 570 |
+
"train_loss": 2.8793208599090576
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 4.0,
|
| 574 |
+
"step": 280,
|
| 575 |
+
"train_loss": 1.293062686920166
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 4.0,
|
| 579 |
+
"step": 280,
|
| 580 |
+
"train_loss": 1.2954611778259277
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 4.0,
|
| 584 |
+
"step": 280,
|
| 585 |
+
"train_loss": 0.26605361700057983
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 4.0,
|
| 589 |
+
"step": 280,
|
| 590 |
+
"train_loss": 2.2778561115264893
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 4.0,
|
| 594 |
+
"step": 280,
|
| 595 |
+
"train_loss": 2.4143030643463135
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 4.0,
|
| 599 |
+
"step": 280,
|
| 600 |
+
"train_loss": 2.61956524848938
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 4.0,
|
| 604 |
+
"step": 280,
|
| 605 |
+
"train_loss": 2.9289610385894775
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 4.0,
|
| 609 |
+
"step": 280,
|
| 610 |
+
"train_loss": 1.7221819162368774
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 4.0,
|
| 614 |
+
"step": 280,
|
| 615 |
+
"train_loss": 0.3664304316043854
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 4.0,
|
| 619 |
+
"step": 280,
|
| 620 |
+
"train_loss": 0.2494506984949112
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 4.0,
|
| 624 |
+
"step": 280,
|
| 625 |
+
"train_loss": 1.011549472808838
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 4.0,
|
| 629 |
+
"step": 280,
|
| 630 |
+
"train_loss": 1.9930472373962402
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 4.0,
|
| 634 |
+
"step": 280,
|
| 635 |
+
"train_loss": 0.8549365997314453
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 4.0,
|
| 639 |
+
"step": 280,
|
| 640 |
+
"train_loss": 1.246674656867981
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 4.0,
|
| 644 |
+
"step": 280,
|
| 645 |
+
"train_loss": 1.2460875511169434
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 4.0,
|
| 649 |
+
"eval_accuracy": {
|
| 650 |
+
"accuracy": 0.8375
|
| 651 |
+
},
|
| 652 |
+
"eval_auc": 0.7523032069970845,
|
| 653 |
+
"eval_f1": {
|
| 654 |
+
"f1": 0.35460992907801414
|
| 655 |
+
},
|
| 656 |
+
"eval_loss": 1.1712112426757812,
|
| 657 |
+
"eval_precision": {
|
| 658 |
+
"precision": 0.352112676056338
|
| 659 |
+
},
|
| 660 |
+
"eval_recall": {
|
| 661 |
+
"recall": 0.35714285714285715
|
| 662 |
+
},
|
| 663 |
+
"eval_runtime": 0.8151,
|
| 664 |
+
"eval_samples_per_second": 687.069,
|
| 665 |
+
"eval_steps_per_second": 42.942,
|
| 666 |
+
"step": 280
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 4.0,
|
| 670 |
+
"step": 280,
|
| 671 |
+
"train_loss": 0.1625460833311081
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 4.0,
|
| 675 |
+
"step": 280,
|
| 676 |
+
"train_loss": 0.25639742612838745
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 4.29,
|
| 680 |
+
"learning_rate": 1.1428571428571429e-05,
|
| 681 |
+
"loss": 0.1546,
|
| 682 |
+
"step": 300
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 4.29,
|
| 686 |
+
"step": 300,
|
| 687 |
+
"train_loss": 0.030452944338321686
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 4.29,
|
| 691 |
+
"step": 300,
|
| 692 |
+
"train_loss": 0.010571416467428207
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.57,
|
| 696 |
+
"learning_rate": 1.0857142857142858e-05,
|
| 697 |
+
"loss": 0.1151,
|
| 698 |
+
"step": 320
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.57,
|
| 702 |
+
"step": 320,
|
| 703 |
+
"train_loss": 0.11510075628757477
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 4.57,
|
| 707 |
+
"step": 320,
|
| 708 |
+
"train_loss": 0.19758647680282593
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 4.86,
|
| 712 |
+
"learning_rate": 1.0285714285714285e-05,
|
| 713 |
+
"loss": 0.1334,
|
| 714 |
+
"step": 340
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 4.86,
|
| 718 |
+
"step": 340,
|
| 719 |
+
"train_loss": 0.25014737248420715
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 4.86,
|
| 723 |
+
"step": 340,
|
| 724 |
+
"train_loss": 0.019543316215276718
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 5.0,
|
| 728 |
+
"eval_accuracy": {
|
| 729 |
+
"accuracy": 0.8482142857142857
|
| 730 |
+
},
|
| 731 |
+
"eval_auc": 0.7345772594752188,
|
| 732 |
+
"eval_f1": {
|
| 733 |
+
"f1": 0.32
|
| 734 |
+
},
|
| 735 |
+
"eval_loss": 1.4309738874435425,
|
| 736 |
+
"eval_precision": {
|
| 737 |
+
"precision": 0.36363636363636365
|
| 738 |
+
},
|
| 739 |
+
"eval_recall": {
|
| 740 |
+
"recall": 0.2857142857142857
|
| 741 |
+
},
|
| 742 |
+
"eval_runtime": 0.6857,
|
| 743 |
+
"eval_samples_per_second": 816.675,
|
| 744 |
+
"eval_steps_per_second": 51.042,
|
| 745 |
+
"step": 350
|
| 746 |
+
}
|
| 747 |
+
],
|
| 748 |
+
"max_steps": 700,
|
| 749 |
+
"num_train_epochs": 10,
|
| 750 |
+
"total_flos": 161884215698040.0,
|
| 751 |
+
"trial_name": null,
|
| 752 |
+
"trial_params": null
|
| 753 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c737ac861b9e676a56c2fc6f6cdaaf889f11880182da85c3f6483a9249bc52e
|
| 3 |
+
size 3579
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|