Commit
•
7655b85
1
Parent(s):
a6e3062
Upload folder using huggingface_hub
Browse files- README.md +36 -0
- checkpoint-8454/config.json +48 -0
- checkpoint-8454/model.safetensors +3 -0
- checkpoint-8454/optimizer.pt +3 -0
- checkpoint-8454/rng_state.pth +3 -0
- checkpoint-8454/scheduler.pt +3 -0
- checkpoint-8454/trainer_state.json +2462 -0
- checkpoint-8454/training_args.bin +3 -0
- config.json +48 -0
- model.safetensors +3 -0
- runs/Oct30_19-52-50_r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva/events.out.tfevents.1730317972.r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva.112.0 +2 -2
- runs/Oct30_19-52-50_r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva/events.out.tfevents.1730377445.r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva.112.1 +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +57 -0
- training_args.bin +3 -0
- training_params.json +30 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
tags:
|
| 4 |
+
- autotrain
|
| 5 |
+
- text-classification
|
| 6 |
+
base_model: neuralmind/bert-base-portuguese-cased
|
| 7 |
+
widget:
|
| 8 |
+
- text: "I love AutoTrain"
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Trained Using AutoTrain
|
| 12 |
+
|
| 13 |
+
- Problem type: Text Classification
|
| 14 |
+
|
| 15 |
+
## Validation Metrics
|
| 16 |
+
loss: 0.7668559551239014
|
| 17 |
+
|
| 18 |
+
f1_macro: 0.6934027713968594
|
| 19 |
+
|
| 20 |
+
f1_micro: 0.6885536823425022
|
| 21 |
+
|
| 22 |
+
f1_weighted: 0.6903326985699842
|
| 23 |
+
|
| 24 |
+
precision_macro: 0.7104396368026238
|
| 25 |
+
|
| 26 |
+
precision_micro: 0.6885536823425022
|
| 27 |
+
|
| 28 |
+
precision_weighted: 0.705612321743312
|
| 29 |
+
|
| 30 |
+
recall_macro: 0.6898371166175705
|
| 31 |
+
|
| 32 |
+
recall_micro: 0.6885536823425022
|
| 33 |
+
|
| 34 |
+
recall_weighted: 0.6885536823425022
|
| 35 |
+
|
| 36 |
+
accuracy: 0.6885536823425022
|
checkpoint-8454/config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "neuralmind/bert-base-portuguese-cased",
|
| 3 |
+
"_num_labels": 5,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "1 - INTIMA\u00c7\u00c3O INICIAL",
|
| 15 |
+
"1": "2 - INTIMA\u00c7\u00c3O INSTRU\u00c7\u00c3O",
|
| 16 |
+
"2": "3 - INTIMA\u00c7\u00c3O RECURSAL",
|
| 17 |
+
"3": "4 - INTIMA\u00c7\u00c3O CUMPRIMENTO DE SENTEN\u00c7A",
|
| 18 |
+
"4": "5 - INTIMA\u00c7\u00c3O TR\u00c2NSITO"
|
| 19 |
+
},
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"1 - INTIMA\u00c7\u00c3O INICIAL": 0,
|
| 24 |
+
"2 - INTIMA\u00c7\u00c3O INSTRU\u00c7\u00c3O": 1,
|
| 25 |
+
"3 - INTIMA\u00c7\u00c3O RECURSAL": 2,
|
| 26 |
+
"4 - INTIMA\u00c7\u00c3O CUMPRIMENTO DE SENTEN\u00c7A": 3,
|
| 27 |
+
"5 - INTIMA\u00c7\u00c3O TR\u00c2NSITO": 4
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-12,
|
| 30 |
+
"max_position_embeddings": 512,
|
| 31 |
+
"model_type": "bert",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"output_past": true,
|
| 35 |
+
"pad_token_id": 0,
|
| 36 |
+
"pooler_fc_size": 768,
|
| 37 |
+
"pooler_num_attention_heads": 12,
|
| 38 |
+
"pooler_num_fc_layers": 3,
|
| 39 |
+
"pooler_size_per_head": 128,
|
| 40 |
+
"pooler_type": "first_token_transform",
|
| 41 |
+
"position_embedding_type": "absolute",
|
| 42 |
+
"problem_type": "single_label_classification",
|
| 43 |
+
"torch_dtype": "float32",
|
| 44 |
+
"transformers_version": "4.45.0",
|
| 45 |
+
"type_vocab_size": 2,
|
| 46 |
+
"use_cache": true,
|
| 47 |
+
"vocab_size": 29794
|
| 48 |
+
}
|
checkpoint-8454/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7b5aea858f25e24827711cfbc02c35b160928b6c19e405df1c24d90f5be4a7d
|
| 3 |
+
size 435731452
|
checkpoint-8454/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8500747b1f542fd7722e9ad75be164317058c0f43f1f6ed49a98609035bade2
|
| 3 |
+
size 871578362
|
checkpoint-8454/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9dfee3efe76126be1b59aa57fecfecdae10b2991270f6288b7b3669660418ce
|
| 3 |
+
size 13990
|
checkpoint-8454/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:892595011cd8f92081483aabb2f75366581579d7002ddf5c707401b4c494b888
|
| 3 |
+
size 1064
|
checkpoint-8454/trainer_state.json
ADDED
|
@@ -0,0 +1,2462 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7668559551239014,
|
| 3 |
+
"best_model_checkpoint": "autotrain-k9ag8-b7nm0/checkpoint-8454",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 8454,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.008871540099361249,
|
| 13 |
+
"grad_norm": 7.602209568023682,
|
| 14 |
+
"learning_rate": 1.4775413711583925e-06,
|
| 15 |
+
"loss": 1.6215,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.017743080198722498,
|
| 20 |
+
"grad_norm": 10.643656730651855,
|
| 21 |
+
"learning_rate": 2.955082742316785e-06,
|
| 22 |
+
"loss": 1.6427,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.026614620298083747,
|
| 27 |
+
"grad_norm": 7.053741455078125,
|
| 28 |
+
"learning_rate": 4.432624113475177e-06,
|
| 29 |
+
"loss": 1.614,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.035486160397444996,
|
| 34 |
+
"grad_norm": 6.636758804321289,
|
| 35 |
+
"learning_rate": 5.91016548463357e-06,
|
| 36 |
+
"loss": 1.5967,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.04435770049680625,
|
| 41 |
+
"grad_norm": 6.755153656005859,
|
| 42 |
+
"learning_rate": 7.387706855791962e-06,
|
| 43 |
+
"loss": 1.596,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.053229240596167494,
|
| 48 |
+
"grad_norm": 6.6177239418029785,
|
| 49 |
+
"learning_rate": 8.865248226950355e-06,
|
| 50 |
+
"loss": 1.564,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06210078069552875,
|
| 55 |
+
"grad_norm": 4.62630558013916,
|
| 56 |
+
"learning_rate": 1.0342789598108746e-05,
|
| 57 |
+
"loss": 1.5625,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.07097232079488999,
|
| 62 |
+
"grad_norm": 5.605529308319092,
|
| 63 |
+
"learning_rate": 1.182033096926714e-05,
|
| 64 |
+
"loss": 1.5312,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.07984386089425125,
|
| 69 |
+
"grad_norm": 9.841400146484375,
|
| 70 |
+
"learning_rate": 1.3297872340425532e-05,
|
| 71 |
+
"loss": 1.4957,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0887154009936125,
|
| 76 |
+
"grad_norm": 6.7550458908081055,
|
| 77 |
+
"learning_rate": 1.4775413711583924e-05,
|
| 78 |
+
"loss": 1.4394,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.09758694109297374,
|
| 83 |
+
"grad_norm": 8.143739700317383,
|
| 84 |
+
"learning_rate": 1.6252955082742317e-05,
|
| 85 |
+
"loss": 1.298,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.10645848119233499,
|
| 90 |
+
"grad_norm": 8.003219604492188,
|
| 91 |
+
"learning_rate": 1.773049645390071e-05,
|
| 92 |
+
"loss": 1.2673,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.11533002129169624,
|
| 97 |
+
"grad_norm": 11.118531227111816,
|
| 98 |
+
"learning_rate": 1.92080378250591e-05,
|
| 99 |
+
"loss": 1.3083,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.1242015613910575,
|
| 104 |
+
"grad_norm": 12.492416381835938,
|
| 105 |
+
"learning_rate": 2.0685579196217493e-05,
|
| 106 |
+
"loss": 1.2445,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.13307310149041873,
|
| 111 |
+
"grad_norm": 4.358503818511963,
|
| 112 |
+
"learning_rate": 2.2163120567375885e-05,
|
| 113 |
+
"loss": 1.146,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.14194464158977999,
|
| 118 |
+
"grad_norm": 7.787415504455566,
|
| 119 |
+
"learning_rate": 2.364066193853428e-05,
|
| 120 |
+
"loss": 1.3067,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.15081618168914124,
|
| 125 |
+
"grad_norm": 5.6704912185668945,
|
| 126 |
+
"learning_rate": 2.5118203309692672e-05,
|
| 127 |
+
"loss": 1.2465,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1596877217885025,
|
| 132 |
+
"grad_norm": 6.449095249176025,
|
| 133 |
+
"learning_rate": 2.6595744680851064e-05,
|
| 134 |
+
"loss": 1.144,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16855926188786374,
|
| 139 |
+
"grad_norm": 10.410565376281738,
|
| 140 |
+
"learning_rate": 2.8073286052009455e-05,
|
| 141 |
+
"loss": 1.2432,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.177430801987225,
|
| 146 |
+
"grad_norm": 11.332856178283691,
|
| 147 |
+
"learning_rate": 2.9550827423167847e-05,
|
| 148 |
+
"loss": 1.1153,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.18630234208658622,
|
| 153 |
+
"grad_norm": 5.967296600341797,
|
| 154 |
+
"learning_rate": 3.102836879432624e-05,
|
| 155 |
+
"loss": 1.0873,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.19517388218594747,
|
| 160 |
+
"grad_norm": 8.682112693786621,
|
| 161 |
+
"learning_rate": 3.2505910165484634e-05,
|
| 162 |
+
"loss": 1.1949,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.20404542228530873,
|
| 167 |
+
"grad_norm": 9.76791763305664,
|
| 168 |
+
"learning_rate": 3.3983451536643026e-05,
|
| 169 |
+
"loss": 1.2161,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.21291696238466998,
|
| 174 |
+
"grad_norm": 5.555416107177734,
|
| 175 |
+
"learning_rate": 3.546099290780142e-05,
|
| 176 |
+
"loss": 1.2722,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.22178850248403123,
|
| 181 |
+
"grad_norm": 7.815185546875,
|
| 182 |
+
"learning_rate": 3.693853427895981e-05,
|
| 183 |
+
"loss": 1.0951,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.23066004258339248,
|
| 188 |
+
"grad_norm": 6.9715895652771,
|
| 189 |
+
"learning_rate": 3.84160756501182e-05,
|
| 190 |
+
"loss": 1.1922,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.23953158268275374,
|
| 195 |
+
"grad_norm": 5.526481628417969,
|
| 196 |
+
"learning_rate": 3.9893617021276594e-05,
|
| 197 |
+
"loss": 1.2998,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.248403122782115,
|
| 202 |
+
"grad_norm": 10.200905799865723,
|
| 203 |
+
"learning_rate": 4.1371158392434986e-05,
|
| 204 |
+
"loss": 1.113,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.2572746628814762,
|
| 209 |
+
"grad_norm": 8.1790771484375,
|
| 210 |
+
"learning_rate": 4.284869976359338e-05,
|
| 211 |
+
"loss": 1.0354,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.26614620298083747,
|
| 216 |
+
"grad_norm": 12.212372779846191,
|
| 217 |
+
"learning_rate": 4.432624113475177e-05,
|
| 218 |
+
"loss": 1.1553,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2750177430801987,
|
| 223 |
+
"grad_norm": 7.486617565155029,
|
| 224 |
+
"learning_rate": 4.580378250591017e-05,
|
| 225 |
+
"loss": 1.0368,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.28388928317955997,
|
| 230 |
+
"grad_norm": 5.094996929168701,
|
| 231 |
+
"learning_rate": 4.728132387706856e-05,
|
| 232 |
+
"loss": 1.0357,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2927608232789212,
|
| 237 |
+
"grad_norm": 7.159431457519531,
|
| 238 |
+
"learning_rate": 4.875886524822695e-05,
|
| 239 |
+
"loss": 1.0414,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.3016323633782825,
|
| 244 |
+
"grad_norm": 7.34111213684082,
|
| 245 |
+
"learning_rate": 4.9973711882229235e-05,
|
| 246 |
+
"loss": 1.0777,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.31050390347764373,
|
| 251 |
+
"grad_norm": 10.668549537658691,
|
| 252 |
+
"learning_rate": 4.980941114616194e-05,
|
| 253 |
+
"loss": 1.0795,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.319375443577005,
|
| 258 |
+
"grad_norm": 10.225701332092285,
|
| 259 |
+
"learning_rate": 4.9645110410094644e-05,
|
| 260 |
+
"loss": 1.0581,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.32824698367636623,
|
| 265 |
+
"grad_norm": 10.795150756835938,
|
| 266 |
+
"learning_rate": 4.948080967402734e-05,
|
| 267 |
+
"loss": 1.0562,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.3371185237757275,
|
| 272 |
+
"grad_norm": 5.316555500030518,
|
| 273 |
+
"learning_rate": 4.931650893796004e-05,
|
| 274 |
+
"loss": 0.9869,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.34599006387508874,
|
| 279 |
+
"grad_norm": 5.70430850982666,
|
| 280 |
+
"learning_rate": 4.9152208201892744e-05,
|
| 281 |
+
"loss": 1.0267,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.35486160397445,
|
| 286 |
+
"grad_norm": 8.478789329528809,
|
| 287 |
+
"learning_rate": 4.898790746582545e-05,
|
| 288 |
+
"loss": 1.0044,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3637331440738112,
|
| 293 |
+
"grad_norm": 8.099587440490723,
|
| 294 |
+
"learning_rate": 4.8823606729758154e-05,
|
| 295 |
+
"loss": 0.9652,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.37260468417317244,
|
| 300 |
+
"grad_norm": 7.185507774353027,
|
| 301 |
+
"learning_rate": 4.865930599369085e-05,
|
| 302 |
+
"loss": 0.979,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3814762242725337,
|
| 307 |
+
"grad_norm": 5.971964359283447,
|
| 308 |
+
"learning_rate": 4.8495005257623556e-05,
|
| 309 |
+
"loss": 1.0463,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.39034776437189495,
|
| 314 |
+
"grad_norm": 7.249493598937988,
|
| 315 |
+
"learning_rate": 4.833070452155626e-05,
|
| 316 |
+
"loss": 0.9667,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.3992193044712562,
|
| 321 |
+
"grad_norm": 5.658076763153076,
|
| 322 |
+
"learning_rate": 4.8166403785488965e-05,
|
| 323 |
+
"loss": 1.1338,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.40809084457061745,
|
| 328 |
+
"grad_norm": 9.11898136138916,
|
| 329 |
+
"learning_rate": 4.800210304942166e-05,
|
| 330 |
+
"loss": 1.0147,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.4169623846699787,
|
| 335 |
+
"grad_norm": 6.678395748138428,
|
| 336 |
+
"learning_rate": 4.783780231335437e-05,
|
| 337 |
+
"loss": 0.8906,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.42583392476933996,
|
| 342 |
+
"grad_norm": 6.076191425323486,
|
| 343 |
+
"learning_rate": 4.7673501577287065e-05,
|
| 344 |
+
"loss": 1.0508,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4347054648687012,
|
| 349 |
+
"grad_norm": 4.180179119110107,
|
| 350 |
+
"learning_rate": 4.750920084121977e-05,
|
| 351 |
+
"loss": 0.9826,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.44357700496806246,
|
| 356 |
+
"grad_norm": 7.360926628112793,
|
| 357 |
+
"learning_rate": 4.7344900105152474e-05,
|
| 358 |
+
"loss": 0.9848,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4524485450674237,
|
| 363 |
+
"grad_norm": 4.815805435180664,
|
| 364 |
+
"learning_rate": 4.718059936908518e-05,
|
| 365 |
+
"loss": 0.9799,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.46132008516678497,
|
| 370 |
+
"grad_norm": 6.276717662811279,
|
| 371 |
+
"learning_rate": 4.701629863301788e-05,
|
| 372 |
+
"loss": 1.0078,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4701916252661462,
|
| 377 |
+
"grad_norm": 6.218778610229492,
|
| 378 |
+
"learning_rate": 4.685199789695058e-05,
|
| 379 |
+
"loss": 1.0242,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.47906316536550747,
|
| 384 |
+
"grad_norm": 7.256498336791992,
|
| 385 |
+
"learning_rate": 4.668769716088328e-05,
|
| 386 |
+
"loss": 0.8731,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.4879347054648687,
|
| 391 |
+
"grad_norm": 5.647130966186523,
|
| 392 |
+
"learning_rate": 4.652339642481598e-05,
|
| 393 |
+
"loss": 0.9373,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.49680624556423,
|
| 398 |
+
"grad_norm": 8.66212272644043,
|
| 399 |
+
"learning_rate": 4.635909568874869e-05,
|
| 400 |
+
"loss": 0.9476,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.5056777856635912,
|
| 405 |
+
"grad_norm": 5.111779689788818,
|
| 406 |
+
"learning_rate": 4.619479495268139e-05,
|
| 407 |
+
"loss": 0.9185,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.5145493257629524,
|
| 412 |
+
"grad_norm": 4.9285502433776855,
|
| 413 |
+
"learning_rate": 4.6030494216614097e-05,
|
| 414 |
+
"loss": 1.0187,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.5234208658623137,
|
| 419 |
+
"grad_norm": 7.633114814758301,
|
| 420 |
+
"learning_rate": 4.5866193480546794e-05,
|
| 421 |
+
"loss": 0.9257,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.5322924059616749,
|
| 426 |
+
"grad_norm": 5.585442543029785,
|
| 427 |
+
"learning_rate": 4.57018927444795e-05,
|
| 428 |
+
"loss": 1.0399,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5411639460610362,
|
| 433 |
+
"grad_norm": 4.591919898986816,
|
| 434 |
+
"learning_rate": 4.5537592008412197e-05,
|
| 435 |
+
"loss": 0.9151,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5500354861603974,
|
| 440 |
+
"grad_norm": 6.754642486572266,
|
| 441 |
+
"learning_rate": 4.53732912723449e-05,
|
| 442 |
+
"loss": 1.0153,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5589070262597587,
|
| 447 |
+
"grad_norm": 6.780356407165527,
|
| 448 |
+
"learning_rate": 4.5208990536277606e-05,
|
| 449 |
+
"loss": 0.9998,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.5677785663591199,
|
| 454 |
+
"grad_norm": 7.279979228973389,
|
| 455 |
+
"learning_rate": 4.50446898002103e-05,
|
| 456 |
+
"loss": 0.998,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.5766501064584811,
|
| 461 |
+
"grad_norm": 18.642566680908203,
|
| 462 |
+
"learning_rate": 4.488038906414301e-05,
|
| 463 |
+
"loss": 0.8556,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5855216465578424,
|
| 468 |
+
"grad_norm": 5.179675579071045,
|
| 469 |
+
"learning_rate": 4.471608832807571e-05,
|
| 470 |
+
"loss": 0.9329,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5943931866572036,
|
| 475 |
+
"grad_norm": 11.334458351135254,
|
| 476 |
+
"learning_rate": 4.455178759200842e-05,
|
| 477 |
+
"loss": 0.9082,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.603264726756565,
|
| 482 |
+
"grad_norm": 7.290953159332275,
|
| 483 |
+
"learning_rate": 4.438748685594112e-05,
|
| 484 |
+
"loss": 0.8716,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.6121362668559261,
|
| 489 |
+
"grad_norm": 3.186927556991577,
|
| 490 |
+
"learning_rate": 4.422318611987382e-05,
|
| 491 |
+
"loss": 0.9728,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.6210078069552875,
|
| 496 |
+
"grad_norm": 6.01223087310791,
|
| 497 |
+
"learning_rate": 4.405888538380652e-05,
|
| 498 |
+
"loss": 0.9188,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.6298793470546487,
|
| 503 |
+
"grad_norm": 7.13861608505249,
|
| 504 |
+
"learning_rate": 4.389458464773922e-05,
|
| 505 |
+
"loss": 0.907,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.63875088715401,
|
| 510 |
+
"grad_norm": 5.580389499664307,
|
| 511 |
+
"learning_rate": 4.3730283911671926e-05,
|
| 512 |
+
"loss": 1.0589,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6476224272533712,
|
| 517 |
+
"grad_norm": 7.246824264526367,
|
| 518 |
+
"learning_rate": 4.356598317560463e-05,
|
| 519 |
+
"loss": 0.9974,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.6564939673527325,
|
| 524 |
+
"grad_norm": 4.785336017608643,
|
| 525 |
+
"learning_rate": 4.3401682439537335e-05,
|
| 526 |
+
"loss": 0.9484,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.6653655074520937,
|
| 531 |
+
"grad_norm": 6.392651081085205,
|
| 532 |
+
"learning_rate": 4.323738170347003e-05,
|
| 533 |
+
"loss": 1.0558,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.674237047551455,
|
| 538 |
+
"grad_norm": 5.654791831970215,
|
| 539 |
+
"learning_rate": 4.307308096740274e-05,
|
| 540 |
+
"loss": 0.904,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.6831085876508162,
|
| 545 |
+
"grad_norm": 8.35096263885498,
|
| 546 |
+
"learning_rate": 4.2908780231335435e-05,
|
| 547 |
+
"loss": 0.9147,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.6919801277501775,
|
| 552 |
+
"grad_norm": 6.575385570526123,
|
| 553 |
+
"learning_rate": 4.274447949526814e-05,
|
| 554 |
+
"loss": 0.9009,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.7008516678495387,
|
| 559 |
+
"grad_norm": 5.978839874267578,
|
| 560 |
+
"learning_rate": 4.2580178759200844e-05,
|
| 561 |
+
"loss": 0.9722,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.7097232079489,
|
| 566 |
+
"grad_norm": 5.449729919433594,
|
| 567 |
+
"learning_rate": 4.241587802313355e-05,
|
| 568 |
+
"loss": 0.8602,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.7185947480482612,
|
| 573 |
+
"grad_norm": 4.398986339569092,
|
| 574 |
+
"learning_rate": 4.2251577287066246e-05,
|
| 575 |
+
"loss": 0.9699,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.7274662881476224,
|
| 580 |
+
"grad_norm": 7.566812515258789,
|
| 581 |
+
"learning_rate": 4.208727655099895e-05,
|
| 582 |
+
"loss": 0.8767,
|
| 583 |
+
"step": 2050
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.7363378282469837,
|
| 587 |
+
"grad_norm": 4.754166603088379,
|
| 588 |
+
"learning_rate": 4.1922975814931655e-05,
|
| 589 |
+
"loss": 0.9233,
|
| 590 |
+
"step": 2075
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.7452093683463449,
|
| 594 |
+
"grad_norm": 8.481706619262695,
|
| 595 |
+
"learning_rate": 4.175867507886436e-05,
|
| 596 |
+
"loss": 0.7744,
|
| 597 |
+
"step": 2100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7540809084457062,
|
| 601 |
+
"grad_norm": 5.274106502532959,
|
| 602 |
+
"learning_rate": 4.159437434279706e-05,
|
| 603 |
+
"loss": 0.9135,
|
| 604 |
+
"step": 2125
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.7629524485450674,
|
| 608 |
+
"grad_norm": 5.358260631561279,
|
| 609 |
+
"learning_rate": 4.1430073606729755e-05,
|
| 610 |
+
"loss": 0.8705,
|
| 611 |
+
"step": 2150
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.7718239886444287,
|
| 615 |
+
"grad_norm": 7.29016637802124,
|
| 616 |
+
"learning_rate": 4.126577287066246e-05,
|
| 617 |
+
"loss": 0.9884,
|
| 618 |
+
"step": 2175
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.7806955287437899,
|
| 622 |
+
"grad_norm": 4.230587005615234,
|
| 623 |
+
"learning_rate": 4.1101472134595165e-05,
|
| 624 |
+
"loss": 1.0909,
|
| 625 |
+
"step": 2200
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7895670688431512,
|
| 629 |
+
"grad_norm": 3.3235785961151123,
|
| 630 |
+
"learning_rate": 4.093717139852787e-05,
|
| 631 |
+
"loss": 0.9675,
|
| 632 |
+
"step": 2225
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.7984386089425124,
|
| 636 |
+
"grad_norm": 6.612360954284668,
|
| 637 |
+
"learning_rate": 4.0772870662460574e-05,
|
| 638 |
+
"loss": 0.8865,
|
| 639 |
+
"step": 2250
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.8073101490418737,
|
| 643 |
+
"grad_norm": 3.8408925533294678,
|
| 644 |
+
"learning_rate": 4.060856992639327e-05,
|
| 645 |
+
"loss": 0.9141,
|
| 646 |
+
"step": 2275
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.8161816891412349,
|
| 650 |
+
"grad_norm": 8.467307090759277,
|
| 651 |
+
"learning_rate": 4.0444269190325976e-05,
|
| 652 |
+
"loss": 0.9242,
|
| 653 |
+
"step": 2300
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.8250532292405962,
|
| 657 |
+
"grad_norm": 6.088973522186279,
|
| 658 |
+
"learning_rate": 4.0279968454258674e-05,
|
| 659 |
+
"loss": 0.9406,
|
| 660 |
+
"step": 2325
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.8339247693399574,
|
| 664 |
+
"grad_norm": 5.104999542236328,
|
| 665 |
+
"learning_rate": 4.011566771819138e-05,
|
| 666 |
+
"loss": 0.9163,
|
| 667 |
+
"step": 2350
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.8427963094393187,
|
| 671 |
+
"grad_norm": 5.350615501403809,
|
| 672 |
+
"learning_rate": 3.995136698212408e-05,
|
| 673 |
+
"loss": 0.892,
|
| 674 |
+
"step": 2375
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.8516678495386799,
|
| 678 |
+
"grad_norm": 4.470179557800293,
|
| 679 |
+
"learning_rate": 3.978706624605679e-05,
|
| 680 |
+
"loss": 0.9243,
|
| 681 |
+
"step": 2400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8605393896380412,
|
| 685 |
+
"grad_norm": 7.627182483673096,
|
| 686 |
+
"learning_rate": 3.9622765509989485e-05,
|
| 687 |
+
"loss": 0.9249,
|
| 688 |
+
"step": 2425
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.8694109297374024,
|
| 692 |
+
"grad_norm": 3.69128155708313,
|
| 693 |
+
"learning_rate": 3.945846477392219e-05,
|
| 694 |
+
"loss": 0.9161,
|
| 695 |
+
"step": 2450
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.8782824698367636,
|
| 699 |
+
"grad_norm": 4.180213928222656,
|
| 700 |
+
"learning_rate": 3.9294164037854894e-05,
|
| 701 |
+
"loss": 0.8598,
|
| 702 |
+
"step": 2475
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.8871540099361249,
|
| 706 |
+
"grad_norm": 5.930824279785156,
|
| 707 |
+
"learning_rate": 3.912986330178759e-05,
|
| 708 |
+
"loss": 0.9908,
|
| 709 |
+
"step": 2500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8960255500354861,
|
| 713 |
+
"grad_norm": 5.566065311431885,
|
| 714 |
+
"learning_rate": 3.8965562565720296e-05,
|
| 715 |
+
"loss": 0.9378,
|
| 716 |
+
"step": 2525
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.9048970901348474,
|
| 720 |
+
"grad_norm": 5.347686767578125,
|
| 721 |
+
"learning_rate": 3.8801261829652994e-05,
|
| 722 |
+
"loss": 0.9627,
|
| 723 |
+
"step": 2550
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.9137686302342086,
|
| 727 |
+
"grad_norm": 4.1419525146484375,
|
| 728 |
+
"learning_rate": 3.86369610935857e-05,
|
| 729 |
+
"loss": 0.7985,
|
| 730 |
+
"step": 2575
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.9226401703335699,
|
| 734 |
+
"grad_norm": 4.296880722045898,
|
| 735 |
+
"learning_rate": 3.84726603575184e-05,
|
| 736 |
+
"loss": 0.8847,
|
| 737 |
+
"step": 2600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.9315117104329311,
|
| 741 |
+
"grad_norm": 5.749390602111816,
|
| 742 |
+
"learning_rate": 3.830835962145111e-05,
|
| 743 |
+
"loss": 0.8881,
|
| 744 |
+
"step": 2625
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.9403832505322924,
|
| 748 |
+
"grad_norm": 4.67301607131958,
|
| 749 |
+
"learning_rate": 3.814405888538381e-05,
|
| 750 |
+
"loss": 0.9763,
|
| 751 |
+
"step": 2650
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.9492547906316536,
|
| 755 |
+
"grad_norm": 7.461897373199463,
|
| 756 |
+
"learning_rate": 3.7979758149316517e-05,
|
| 757 |
+
"loss": 0.9073,
|
| 758 |
+
"step": 2675
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.9581263307310149,
|
| 762 |
+
"grad_norm": 5.127830982208252,
|
| 763 |
+
"learning_rate": 3.7815457413249214e-05,
|
| 764 |
+
"loss": 0.8915,
|
| 765 |
+
"step": 2700
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9669978708303761,
|
| 769 |
+
"grad_norm": 3.610241413116455,
|
| 770 |
+
"learning_rate": 3.765115667718191e-05,
|
| 771 |
+
"loss": 0.8039,
|
| 772 |
+
"step": 2725
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.9758694109297374,
|
| 776 |
+
"grad_norm": 7.237741470336914,
|
| 777 |
+
"learning_rate": 3.7486855941114617e-05,
|
| 778 |
+
"loss": 0.8631,
|
| 779 |
+
"step": 2750
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.9847409510290986,
|
| 783 |
+
"grad_norm": 5.400920867919922,
|
| 784 |
+
"learning_rate": 3.732255520504732e-05,
|
| 785 |
+
"loss": 0.9269,
|
| 786 |
+
"step": 2775
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.99361249112846,
|
| 790 |
+
"grad_norm": 7.000077247619629,
|
| 791 |
+
"learning_rate": 3.7158254468980026e-05,
|
| 792 |
+
"loss": 0.8502,
|
| 793 |
+
"step": 2800
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.0,
|
| 797 |
+
"eval_accuracy": 0.6351375332741792,
|
| 798 |
+
"eval_f1_macro": 0.646122256511988,
|
| 799 |
+
"eval_f1_micro": 0.6351375332741792,
|
| 800 |
+
"eval_f1_weighted": 0.64322619559571,
|
| 801 |
+
"eval_loss": 0.8854864835739136,
|
| 802 |
+
"eval_precision_macro": 0.7018139832423136,
|
| 803 |
+
"eval_precision_micro": 0.6351375332741792,
|
| 804 |
+
"eval_precision_weighted": 0.6951112379231563,
|
| 805 |
+
"eval_recall_macro": 0.6347383469069486,
|
| 806 |
+
"eval_recall_micro": 0.6351375332741792,
|
| 807 |
+
"eval_recall_weighted": 0.6351375332741792,
|
| 808 |
+
"eval_runtime": 1148.734,
|
| 809 |
+
"eval_samples_per_second": 4.905,
|
| 810 |
+
"eval_steps_per_second": 0.307,
|
| 811 |
+
"step": 2818
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.0024840312278211,
|
| 815 |
+
"grad_norm": 2.3094370365142822,
|
| 816 |
+
"learning_rate": 3.6993953732912723e-05,
|
| 817 |
+
"loss": 0.7619,
|
| 818 |
+
"step": 2825
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.0113555713271825,
|
| 822 |
+
"grad_norm": 6.9726386070251465,
|
| 823 |
+
"learning_rate": 3.682965299684543e-05,
|
| 824 |
+
"loss": 0.8545,
|
| 825 |
+
"step": 2850
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 1.0202271114265435,
|
| 829 |
+
"grad_norm": 6.961376667022705,
|
| 830 |
+
"learning_rate": 3.666535226077813e-05,
|
| 831 |
+
"loss": 0.8677,
|
| 832 |
+
"step": 2875
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.0290986515259049,
|
| 836 |
+
"grad_norm": 5.254246234893799,
|
| 837 |
+
"learning_rate": 3.650105152471083e-05,
|
| 838 |
+
"loss": 0.9094,
|
| 839 |
+
"step": 2900
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.0379701916252662,
|
| 843 |
+
"grad_norm": 4.578902244567871,
|
| 844 |
+
"learning_rate": 3.6336750788643535e-05,
|
| 845 |
+
"loss": 0.8537,
|
| 846 |
+
"step": 2925
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.0468417317246275,
|
| 850 |
+
"grad_norm": 3.2812981605529785,
|
| 851 |
+
"learning_rate": 3.617245005257624e-05,
|
| 852 |
+
"loss": 0.9155,
|
| 853 |
+
"step": 2950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.0557132718239886,
|
| 857 |
+
"grad_norm": 4.157433032989502,
|
| 858 |
+
"learning_rate": 3.600814931650894e-05,
|
| 859 |
+
"loss": 0.8686,
|
| 860 |
+
"step": 2975
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.0645848119233499,
|
| 864 |
+
"grad_norm": 8.107535362243652,
|
| 865 |
+
"learning_rate": 3.584384858044164e-05,
|
| 866 |
+
"loss": 0.8142,
|
| 867 |
+
"step": 3000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.0734563520227112,
|
| 871 |
+
"grad_norm": 4.729946613311768,
|
| 872 |
+
"learning_rate": 3.5679547844374346e-05,
|
| 873 |
+
"loss": 0.8848,
|
| 874 |
+
"step": 3025
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.0823278921220725,
|
| 878 |
+
"grad_norm": 4.649665832519531,
|
| 879 |
+
"learning_rate": 3.551524710830705e-05,
|
| 880 |
+
"loss": 0.82,
|
| 881 |
+
"step": 3050
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.0911994322214336,
|
| 885 |
+
"grad_norm": 6.006737232208252,
|
| 886 |
+
"learning_rate": 3.535094637223975e-05,
|
| 887 |
+
"loss": 0.928,
|
| 888 |
+
"step": 3075
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 1.1000709723207949,
|
| 892 |
+
"grad_norm": 6.667446613311768,
|
| 893 |
+
"learning_rate": 3.518664563617245e-05,
|
| 894 |
+
"loss": 0.8371,
|
| 895 |
+
"step": 3100
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.1089425124201562,
|
| 899 |
+
"grad_norm": 5.0947418212890625,
|
| 900 |
+
"learning_rate": 3.502234490010515e-05,
|
| 901 |
+
"loss": 0.807,
|
| 902 |
+
"step": 3125
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.1178140525195175,
|
| 906 |
+
"grad_norm": 7.8628129959106445,
|
| 907 |
+
"learning_rate": 3.4858044164037855e-05,
|
| 908 |
+
"loss": 0.8762,
|
| 909 |
+
"step": 3150
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.1266855926188786,
|
| 913 |
+
"grad_norm": 7.366974830627441,
|
| 914 |
+
"learning_rate": 3.469374342797056e-05,
|
| 915 |
+
"loss": 0.9567,
|
| 916 |
+
"step": 3175
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.1355571327182399,
|
| 920 |
+
"grad_norm": 7.182910442352295,
|
| 921 |
+
"learning_rate": 3.4529442691903264e-05,
|
| 922 |
+
"loss": 0.8733,
|
| 923 |
+
"step": 3200
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 1.1444286728176012,
|
| 927 |
+
"grad_norm": 4.816949844360352,
|
| 928 |
+
"learning_rate": 3.436514195583596e-05,
|
| 929 |
+
"loss": 0.9105,
|
| 930 |
+
"step": 3225
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 1.1533002129169625,
|
| 934 |
+
"grad_norm": 5.3113813400268555,
|
| 935 |
+
"learning_rate": 3.4200841219768666e-05,
|
| 936 |
+
"loss": 0.9543,
|
| 937 |
+
"step": 3250
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.1621717530163236,
|
| 941 |
+
"grad_norm": 8.444610595703125,
|
| 942 |
+
"learning_rate": 3.403654048370137e-05,
|
| 943 |
+
"loss": 0.8729,
|
| 944 |
+
"step": 3275
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.171043293115685,
|
| 948 |
+
"grad_norm": 13.175244331359863,
|
| 949 |
+
"learning_rate": 3.387223974763407e-05,
|
| 950 |
+
"loss": 0.84,
|
| 951 |
+
"step": 3300
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 1.1799148332150462,
|
| 955 |
+
"grad_norm": 4.161907196044922,
|
| 956 |
+
"learning_rate": 3.370793901156677e-05,
|
| 957 |
+
"loss": 0.8264,
|
| 958 |
+
"step": 3325
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.1887863733144073,
|
| 962 |
+
"grad_norm": 5.216485977172852,
|
| 963 |
+
"learning_rate": 3.354363827549948e-05,
|
| 964 |
+
"loss": 0.8398,
|
| 965 |
+
"step": 3350
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 1.1976579134137686,
|
| 969 |
+
"grad_norm": 5.230435371398926,
|
| 970 |
+
"learning_rate": 3.3379337539432175e-05,
|
| 971 |
+
"loss": 0.9218,
|
| 972 |
+
"step": 3375
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 1.20652945351313,
|
| 976 |
+
"grad_norm": 6.927298069000244,
|
| 977 |
+
"learning_rate": 3.321503680336488e-05,
|
| 978 |
+
"loss": 0.8361,
|
| 979 |
+
"step": 3400
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.2154009936124912,
|
| 983 |
+
"grad_norm": 8.193685531616211,
|
| 984 |
+
"learning_rate": 3.3050736067297585e-05,
|
| 985 |
+
"loss": 0.8125,
|
| 986 |
+
"step": 3425
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 1.2242725337118523,
|
| 990 |
+
"grad_norm": 4.92389440536499,
|
| 991 |
+
"learning_rate": 3.288643533123029e-05,
|
| 992 |
+
"loss": 0.9335,
|
| 993 |
+
"step": 3450
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 1.2331440738112136,
|
| 997 |
+
"grad_norm": 7.39652156829834,
|
| 998 |
+
"learning_rate": 3.272213459516299e-05,
|
| 999 |
+
"loss": 0.8435,
|
| 1000 |
+
"step": 3475
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 1.242015613910575,
|
| 1004 |
+
"grad_norm": 5.768881320953369,
|
| 1005 |
+
"learning_rate": 3.2557833859095685e-05,
|
| 1006 |
+
"loss": 0.8422,
|
| 1007 |
+
"step": 3500
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.250887154009936,
|
| 1011 |
+
"grad_norm": 6.411141872406006,
|
| 1012 |
+
"learning_rate": 3.239353312302839e-05,
|
| 1013 |
+
"loss": 0.8596,
|
| 1014 |
+
"step": 3525
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 1.2597586941092973,
|
| 1018 |
+
"grad_norm": 2.415675163269043,
|
| 1019 |
+
"learning_rate": 3.2229232386961094e-05,
|
| 1020 |
+
"loss": 0.761,
|
| 1021 |
+
"step": 3550
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.2686302342086586,
|
| 1025 |
+
"grad_norm": 4.490350246429443,
|
| 1026 |
+
"learning_rate": 3.20649316508938e-05,
|
| 1027 |
+
"loss": 0.8352,
|
| 1028 |
+
"step": 3575
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 1.27750177430802,
|
| 1032 |
+
"grad_norm": 6.4845871925354,
|
| 1033 |
+
"learning_rate": 3.19006309148265e-05,
|
| 1034 |
+
"loss": 0.9389,
|
| 1035 |
+
"step": 3600
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 1.2863733144073812,
|
| 1039 |
+
"grad_norm": 9.419644355773926,
|
| 1040 |
+
"learning_rate": 3.173633017875921e-05,
|
| 1041 |
+
"loss": 0.8249,
|
| 1042 |
+
"step": 3625
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 1.2952448545067423,
|
| 1046 |
+
"grad_norm": 4.427757740020752,
|
| 1047 |
+
"learning_rate": 3.1572029442691905e-05,
|
| 1048 |
+
"loss": 0.8947,
|
| 1049 |
+
"step": 3650
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.3041163946061036,
|
| 1053 |
+
"grad_norm": 5.791045665740967,
|
| 1054 |
+
"learning_rate": 3.140772870662461e-05,
|
| 1055 |
+
"loss": 0.9288,
|
| 1056 |
+
"step": 3675
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 1.312987934705465,
|
| 1060 |
+
"grad_norm": 6.885122299194336,
|
| 1061 |
+
"learning_rate": 3.124342797055731e-05,
|
| 1062 |
+
"loss": 0.8869,
|
| 1063 |
+
"step": 3700
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.321859474804826,
|
| 1067 |
+
"grad_norm": 4.827605247497559,
|
| 1068 |
+
"learning_rate": 3.107912723449001e-05,
|
| 1069 |
+
"loss": 0.8798,
|
| 1070 |
+
"step": 3725
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 1.3307310149041873,
|
| 1074 |
+
"grad_norm": 5.087188243865967,
|
| 1075 |
+
"learning_rate": 3.0914826498422716e-05,
|
| 1076 |
+
"loss": 0.8865,
|
| 1077 |
+
"step": 3750
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 1.3396025550035486,
|
| 1081 |
+
"grad_norm": 6.508386135101318,
|
| 1082 |
+
"learning_rate": 3.0750525762355414e-05,
|
| 1083 |
+
"loss": 0.769,
|
| 1084 |
+
"step": 3775
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 1.34847409510291,
|
| 1088 |
+
"grad_norm": 3.5530879497528076,
|
| 1089 |
+
"learning_rate": 3.058622502628812e-05,
|
| 1090 |
+
"loss": 0.8959,
|
| 1091 |
+
"step": 3800
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.3573456352022713,
|
| 1095 |
+
"grad_norm": 5.191397666931152,
|
| 1096 |
+
"learning_rate": 3.042192429022082e-05,
|
| 1097 |
+
"loss": 0.9568,
|
| 1098 |
+
"step": 3825
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 1.3662171753016323,
|
| 1102 |
+
"grad_norm": 6.676424980163574,
|
| 1103 |
+
"learning_rate": 3.0257623554153524e-05,
|
| 1104 |
+
"loss": 0.9181,
|
| 1105 |
+
"step": 3850
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.3750887154009936,
|
| 1109 |
+
"grad_norm": 3.397516965866089,
|
| 1110 |
+
"learning_rate": 3.009332281808623e-05,
|
| 1111 |
+
"loss": 0.7627,
|
| 1112 |
+
"step": 3875
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 1.3839602555003547,
|
| 1116 |
+
"grad_norm": 6.955685615539551,
|
| 1117 |
+
"learning_rate": 2.9929022082018933e-05,
|
| 1118 |
+
"loss": 0.8065,
|
| 1119 |
+
"step": 3900
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 1.392831795599716,
|
| 1123 |
+
"grad_norm": 4.397638320922852,
|
| 1124 |
+
"learning_rate": 2.976472134595163e-05,
|
| 1125 |
+
"loss": 0.9026,
|
| 1126 |
+
"step": 3925
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 1.4017033356990773,
|
| 1130 |
+
"grad_norm": 4.824989318847656,
|
| 1131 |
+
"learning_rate": 2.9600420609884332e-05,
|
| 1132 |
+
"loss": 0.8852,
|
| 1133 |
+
"step": 3950
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 1.4105748757984387,
|
| 1137 |
+
"grad_norm": 6.728146076202393,
|
| 1138 |
+
"learning_rate": 2.9436119873817037e-05,
|
| 1139 |
+
"loss": 0.8408,
|
| 1140 |
+
"step": 3975
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 1.4194464158978,
|
| 1144 |
+
"grad_norm": 16.979982376098633,
|
| 1145 |
+
"learning_rate": 2.927181913774974e-05,
|
| 1146 |
+
"loss": 0.9352,
|
| 1147 |
+
"step": 4000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.428317955997161,
|
| 1151 |
+
"grad_norm": 6.757816791534424,
|
| 1152 |
+
"learning_rate": 2.9107518401682442e-05,
|
| 1153 |
+
"loss": 0.89,
|
| 1154 |
+
"step": 4025
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 1.4371894960965224,
|
| 1158 |
+
"grad_norm": 9.163891792297363,
|
| 1159 |
+
"learning_rate": 2.894321766561514e-05,
|
| 1160 |
+
"loss": 0.9316,
|
| 1161 |
+
"step": 4050
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.4460610361958837,
|
| 1165 |
+
"grad_norm": 9.104488372802734,
|
| 1166 |
+
"learning_rate": 2.8778916929547845e-05,
|
| 1167 |
+
"loss": 0.9046,
|
| 1168 |
+
"step": 4075
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 1.4549325762952448,
|
| 1172 |
+
"grad_norm": 6.510610580444336,
|
| 1173 |
+
"learning_rate": 2.861461619348055e-05,
|
| 1174 |
+
"loss": 0.7608,
|
| 1175 |
+
"step": 4100
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 1.463804116394606,
|
| 1179 |
+
"grad_norm": 6.328439235687256,
|
| 1180 |
+
"learning_rate": 2.845031545741325e-05,
|
| 1181 |
+
"loss": 1.0008,
|
| 1182 |
+
"step": 4125
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 1.4726756564939674,
|
| 1186 |
+
"grad_norm": 6.498648166656494,
|
| 1187 |
+
"learning_rate": 2.8286014721345955e-05,
|
| 1188 |
+
"loss": 0.918,
|
| 1189 |
+
"step": 4150
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.4815471965933287,
|
| 1193 |
+
"grad_norm": 6.506234169006348,
|
| 1194 |
+
"learning_rate": 2.8121713985278653e-05,
|
| 1195 |
+
"loss": 0.8831,
|
| 1196 |
+
"step": 4175
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 1.49041873669269,
|
| 1200 |
+
"grad_norm": 3.214972734451294,
|
| 1201 |
+
"learning_rate": 2.7957413249211357e-05,
|
| 1202 |
+
"loss": 0.8107,
|
| 1203 |
+
"step": 4200
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 1.499290276792051,
|
| 1207 |
+
"grad_norm": 4.053207874298096,
|
| 1208 |
+
"learning_rate": 2.7793112513144058e-05,
|
| 1209 |
+
"loss": 0.7369,
|
| 1210 |
+
"step": 4225
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 1.5081618168914124,
|
| 1214 |
+
"grad_norm": 7.692103385925293,
|
| 1215 |
+
"learning_rate": 2.7628811777076763e-05,
|
| 1216 |
+
"loss": 0.8413,
|
| 1217 |
+
"step": 4250
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.5170333569907735,
|
| 1221 |
+
"grad_norm": 7.168026447296143,
|
| 1222 |
+
"learning_rate": 2.7464511041009467e-05,
|
| 1223 |
+
"loss": 0.7727,
|
| 1224 |
+
"step": 4275
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 1.5259048970901348,
|
| 1228 |
+
"grad_norm": 37.76559066772461,
|
| 1229 |
+
"learning_rate": 2.730021030494217e-05,
|
| 1230 |
+
"loss": 0.896,
|
| 1231 |
+
"step": 4300
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.534776437189496,
|
| 1235 |
+
"grad_norm": 7.820963382720947,
|
| 1236 |
+
"learning_rate": 2.7135909568874866e-05,
|
| 1237 |
+
"loss": 0.9137,
|
| 1238 |
+
"step": 4325
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 1.5436479772888574,
|
| 1242 |
+
"grad_norm": 5.012437343597412,
|
| 1243 |
+
"learning_rate": 2.697160883280757e-05,
|
| 1244 |
+
"loss": 0.8227,
|
| 1245 |
+
"step": 4350
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 1.5525195173882187,
|
| 1249 |
+
"grad_norm": 4.973189353942871,
|
| 1250 |
+
"learning_rate": 2.6807308096740275e-05,
|
| 1251 |
+
"loss": 0.7805,
|
| 1252 |
+
"step": 4375
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.56139105748758,
|
| 1256 |
+
"grad_norm": 5.134377956390381,
|
| 1257 |
+
"learning_rate": 2.664300736067298e-05,
|
| 1258 |
+
"loss": 0.8911,
|
| 1259 |
+
"step": 4400
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 1.570262597586941,
|
| 1263 |
+
"grad_norm": 4.107070446014404,
|
| 1264 |
+
"learning_rate": 2.647870662460568e-05,
|
| 1265 |
+
"loss": 0.9355,
|
| 1266 |
+
"step": 4425
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 1.5791341376863024,
|
| 1270 |
+
"grad_norm": 4.052985191345215,
|
| 1271 |
+
"learning_rate": 2.631440588853838e-05,
|
| 1272 |
+
"loss": 0.9195,
|
| 1273 |
+
"step": 4450
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.5880056777856635,
|
| 1277 |
+
"grad_norm": 6.549808025360107,
|
| 1278 |
+
"learning_rate": 2.6150105152471083e-05,
|
| 1279 |
+
"loss": 0.8856,
|
| 1280 |
+
"step": 4475
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 1.5968772178850248,
|
| 1284 |
+
"grad_norm": 4.466334819793701,
|
| 1285 |
+
"learning_rate": 2.5985804416403788e-05,
|
| 1286 |
+
"loss": 0.9777,
|
| 1287 |
+
"step": 4500
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 1.605748757984386,
|
| 1291 |
+
"grad_norm": 4.111423015594482,
|
| 1292 |
+
"learning_rate": 2.582150368033649e-05,
|
| 1293 |
+
"loss": 0.8266,
|
| 1294 |
+
"step": 4525
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 1.6146202980837474,
|
| 1298 |
+
"grad_norm": 5.717258453369141,
|
| 1299 |
+
"learning_rate": 2.5657202944269193e-05,
|
| 1300 |
+
"loss": 0.773,
|
| 1301 |
+
"step": 4550
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 1.6234918381831087,
|
| 1305 |
+
"grad_norm": 5.022154331207275,
|
| 1306 |
+
"learning_rate": 2.5492902208201898e-05,
|
| 1307 |
+
"loss": 0.8747,
|
| 1308 |
+
"step": 4575
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 1.6323633782824698,
|
| 1312 |
+
"grad_norm": 3.223011016845703,
|
| 1313 |
+
"learning_rate": 2.5328601472134596e-05,
|
| 1314 |
+
"loss": 0.8246,
|
| 1315 |
+
"step": 4600
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.641234918381831,
|
| 1319 |
+
"grad_norm": 5.4209303855896,
|
| 1320 |
+
"learning_rate": 2.5164300736067297e-05,
|
| 1321 |
+
"loss": 0.8347,
|
| 1322 |
+
"step": 4625
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.6501064584811922,
|
| 1326 |
+
"grad_norm": 6.209916114807129,
|
| 1327 |
+
"learning_rate": 2.5e-05,
|
| 1328 |
+
"loss": 0.946,
|
| 1329 |
+
"step": 4650
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.6589779985805535,
|
| 1333 |
+
"grad_norm": 4.774755477905273,
|
| 1334 |
+
"learning_rate": 2.4835699263932706e-05,
|
| 1335 |
+
"loss": 0.8054,
|
| 1336 |
+
"step": 4675
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 1.6678495386799148,
|
| 1340 |
+
"grad_norm": 7.4865193367004395,
|
| 1341 |
+
"learning_rate": 2.4671398527865407e-05,
|
| 1342 |
+
"loss": 0.7584,
|
| 1343 |
+
"step": 4700
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.6767210787792761,
|
| 1347 |
+
"grad_norm": 6.235969543457031,
|
| 1348 |
+
"learning_rate": 2.4507097791798108e-05,
|
| 1349 |
+
"loss": 0.7856,
|
| 1350 |
+
"step": 4725
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 1.6855926188786374,
|
| 1354 |
+
"grad_norm": 5.785863399505615,
|
| 1355 |
+
"learning_rate": 2.4342797055730813e-05,
|
| 1356 |
+
"loss": 0.8978,
|
| 1357 |
+
"step": 4750
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.6944641589779987,
|
| 1361 |
+
"grad_norm": 4.585093021392822,
|
| 1362 |
+
"learning_rate": 2.4178496319663514e-05,
|
| 1363 |
+
"loss": 0.6833,
|
| 1364 |
+
"step": 4775
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.7033356990773598,
|
| 1368 |
+
"grad_norm": 5.399811744689941,
|
| 1369 |
+
"learning_rate": 2.4014195583596215e-05,
|
| 1370 |
+
"loss": 0.7674,
|
| 1371 |
+
"step": 4800
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 1.7122072391767211,
|
| 1375 |
+
"grad_norm": 5.478691101074219,
|
| 1376 |
+
"learning_rate": 2.3849894847528916e-05,
|
| 1377 |
+
"loss": 0.8765,
|
| 1378 |
+
"step": 4825
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 1.7210787792760822,
|
| 1382 |
+
"grad_norm": 5.945518970489502,
|
| 1383 |
+
"learning_rate": 2.368559411146162e-05,
|
| 1384 |
+
"loss": 0.8706,
|
| 1385 |
+
"step": 4850
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 1.7299503193754435,
|
| 1389 |
+
"grad_norm": 4.882993698120117,
|
| 1390 |
+
"learning_rate": 2.3521293375394325e-05,
|
| 1391 |
+
"loss": 0.7293,
|
| 1392 |
+
"step": 4875
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 1.7388218594748048,
|
| 1396 |
+
"grad_norm": 6.987505912780762,
|
| 1397 |
+
"learning_rate": 2.3356992639327023e-05,
|
| 1398 |
+
"loss": 0.7372,
|
| 1399 |
+
"step": 4900
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.7476933995741661,
|
| 1403 |
+
"grad_norm": 4.7848286628723145,
|
| 1404 |
+
"learning_rate": 2.3192691903259727e-05,
|
| 1405 |
+
"loss": 0.9114,
|
| 1406 |
+
"step": 4925
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.7565649396735274,
|
| 1410 |
+
"grad_norm": 6.290490627288818,
|
| 1411 |
+
"learning_rate": 2.3028391167192432e-05,
|
| 1412 |
+
"loss": 0.8896,
|
| 1413 |
+
"step": 4950
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.7654364797728885,
|
| 1417 |
+
"grad_norm": 6.204107284545898,
|
| 1418 |
+
"learning_rate": 2.2864090431125133e-05,
|
| 1419 |
+
"loss": 0.7874,
|
| 1420 |
+
"step": 4975
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.7743080198722498,
|
| 1424 |
+
"grad_norm": 5.911350250244141,
|
| 1425 |
+
"learning_rate": 2.2699789695057834e-05,
|
| 1426 |
+
"loss": 0.8717,
|
| 1427 |
+
"step": 5000
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.783179559971611,
|
| 1431 |
+
"grad_norm": 6.586339473724365,
|
| 1432 |
+
"learning_rate": 2.2535488958990535e-05,
|
| 1433 |
+
"loss": 0.924,
|
| 1434 |
+
"step": 5025
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.7920511000709722,
|
| 1438 |
+
"grad_norm": 7.093676567077637,
|
| 1439 |
+
"learning_rate": 2.237118822292324e-05,
|
| 1440 |
+
"loss": 0.8317,
|
| 1441 |
+
"step": 5050
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.8009226401703335,
|
| 1445 |
+
"grad_norm": 7.202934265136719,
|
| 1446 |
+
"learning_rate": 2.2206887486855944e-05,
|
| 1447 |
+
"loss": 0.7207,
|
| 1448 |
+
"step": 5075
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.8097941802696949,
|
| 1452 |
+
"grad_norm": 9.095272064208984,
|
| 1453 |
+
"learning_rate": 2.2042586750788642e-05,
|
| 1454 |
+
"loss": 0.7929,
|
| 1455 |
+
"step": 5100
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 1.8186657203690562,
|
| 1459 |
+
"grad_norm": 6.040229320526123,
|
| 1460 |
+
"learning_rate": 2.1878286014721347e-05,
|
| 1461 |
+
"loss": 0.9413,
|
| 1462 |
+
"step": 5125
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.8275372604684175,
|
| 1466 |
+
"grad_norm": 4.86398983001709,
|
| 1467 |
+
"learning_rate": 2.171398527865405e-05,
|
| 1468 |
+
"loss": 0.8813,
|
| 1469 |
+
"step": 5150
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.8364088005677786,
|
| 1473 |
+
"grad_norm": 4.922494888305664,
|
| 1474 |
+
"learning_rate": 2.1549684542586752e-05,
|
| 1475 |
+
"loss": 0.8695,
|
| 1476 |
+
"step": 5175
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 1.8452803406671399,
|
| 1480 |
+
"grad_norm": 6.124248504638672,
|
| 1481 |
+
"learning_rate": 2.1385383806519453e-05,
|
| 1482 |
+
"loss": 0.7623,
|
| 1483 |
+
"step": 5200
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.854151880766501,
|
| 1487 |
+
"grad_norm": 5.949371814727783,
|
| 1488 |
+
"learning_rate": 2.1221083070452158e-05,
|
| 1489 |
+
"loss": 0.8198,
|
| 1490 |
+
"step": 5225
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.8630234208658623,
|
| 1494 |
+
"grad_norm": 5.420801162719727,
|
| 1495 |
+
"learning_rate": 2.105678233438486e-05,
|
| 1496 |
+
"loss": 0.8091,
|
| 1497 |
+
"step": 5250
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 1.8718949609652236,
|
| 1501 |
+
"grad_norm": 4.861514091491699,
|
| 1502 |
+
"learning_rate": 2.0892481598317563e-05,
|
| 1503 |
+
"loss": 0.7612,
|
| 1504 |
+
"step": 5275
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 1.8807665010645849,
|
| 1508 |
+
"grad_norm": 7.055742263793945,
|
| 1509 |
+
"learning_rate": 2.072818086225026e-05,
|
| 1510 |
+
"loss": 0.9407,
|
| 1511 |
+
"step": 5300
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 1.8896380411639462,
|
| 1515 |
+
"grad_norm": 5.109190940856934,
|
| 1516 |
+
"learning_rate": 2.0563880126182966e-05,
|
| 1517 |
+
"loss": 0.835,
|
| 1518 |
+
"step": 5325
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 1.8985095812633073,
|
| 1522 |
+
"grad_norm": 7.075034141540527,
|
| 1523 |
+
"learning_rate": 2.039957939011567e-05,
|
| 1524 |
+
"loss": 0.7638,
|
| 1525 |
+
"step": 5350
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.9073811213626686,
|
| 1529 |
+
"grad_norm": 10.149177551269531,
|
| 1530 |
+
"learning_rate": 2.023527865404837e-05,
|
| 1531 |
+
"loss": 0.7629,
|
| 1532 |
+
"step": 5375
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 1.9162526614620297,
|
| 1536 |
+
"grad_norm": 5.505281448364258,
|
| 1537 |
+
"learning_rate": 2.0070977917981073e-05,
|
| 1538 |
+
"loss": 0.8724,
|
| 1539 |
+
"step": 5400
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 1.925124201561391,
|
| 1543 |
+
"grad_norm": 4.993834495544434,
|
| 1544 |
+
"learning_rate": 1.9906677181913777e-05,
|
| 1545 |
+
"loss": 0.8732,
|
| 1546 |
+
"step": 5425
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 1.9339957416607523,
|
| 1550 |
+
"grad_norm": 8.533299446105957,
|
| 1551 |
+
"learning_rate": 1.9742376445846478e-05,
|
| 1552 |
+
"loss": 0.7291,
|
| 1553 |
+
"step": 5450
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 1.9428672817601136,
|
| 1557 |
+
"grad_norm": 5.450367450714111,
|
| 1558 |
+
"learning_rate": 1.9578075709779183e-05,
|
| 1559 |
+
"loss": 0.8014,
|
| 1560 |
+
"step": 5475
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 1.951738821859475,
|
| 1564 |
+
"grad_norm": 6.007541179656982,
|
| 1565 |
+
"learning_rate": 1.941377497371188e-05,
|
| 1566 |
+
"loss": 0.7869,
|
| 1567 |
+
"step": 5500
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.9606103619588362,
|
| 1571 |
+
"grad_norm": 4.755675792694092,
|
| 1572 |
+
"learning_rate": 1.9249474237644585e-05,
|
| 1573 |
+
"loss": 0.7357,
|
| 1574 |
+
"step": 5525
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 1.9694819020581973,
|
| 1578 |
+
"grad_norm": 6.058708667755127,
|
| 1579 |
+
"learning_rate": 1.908517350157729e-05,
|
| 1580 |
+
"loss": 0.8856,
|
| 1581 |
+
"step": 5550
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 1.9783534421575586,
|
| 1585 |
+
"grad_norm": 5.893248558044434,
|
| 1586 |
+
"learning_rate": 1.892087276550999e-05,
|
| 1587 |
+
"loss": 0.7709,
|
| 1588 |
+
"step": 5575
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 1.9872249822569197,
|
| 1592 |
+
"grad_norm": 6.941923141479492,
|
| 1593 |
+
"learning_rate": 1.8756572029442692e-05,
|
| 1594 |
+
"loss": 0.8291,
|
| 1595 |
+
"step": 5600
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.996096522356281,
|
| 1599 |
+
"grad_norm": 5.693519115447998,
|
| 1600 |
+
"learning_rate": 1.8592271293375396e-05,
|
| 1601 |
+
"loss": 0.7554,
|
| 1602 |
+
"step": 5625
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 2.0,
|
| 1606 |
+
"eval_accuracy": 0.6713398402839397,
|
| 1607 |
+
"eval_f1_macro": 0.6784994509782545,
|
| 1608 |
+
"eval_f1_micro": 0.6713398402839397,
|
| 1609 |
+
"eval_f1_weighted": 0.6752966842367127,
|
| 1610 |
+
"eval_loss": 0.7940883040428162,
|
| 1611 |
+
"eval_precision_macro": 0.6994216245852686,
|
| 1612 |
+
"eval_precision_micro": 0.6713398402839397,
|
| 1613 |
+
"eval_precision_weighted": 0.6949112315057265,
|
| 1614 |
+
"eval_recall_macro": 0.6730557962224918,
|
| 1615 |
+
"eval_recall_micro": 0.6713398402839397,
|
| 1616 |
+
"eval_recall_weighted": 0.6713398402839397,
|
| 1617 |
+
"eval_runtime": 1183.1474,
|
| 1618 |
+
"eval_samples_per_second": 4.763,
|
| 1619 |
+
"eval_steps_per_second": 0.298,
|
| 1620 |
+
"step": 5636
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 2.0049680624556423,
|
| 1624 |
+
"grad_norm": 6.007699489593506,
|
| 1625 |
+
"learning_rate": 1.8427970557308097e-05,
|
| 1626 |
+
"loss": 0.7755,
|
| 1627 |
+
"step": 5650
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 2.0138396025550036,
|
| 1631 |
+
"grad_norm": 6.981593132019043,
|
| 1632 |
+
"learning_rate": 1.82636698212408e-05,
|
| 1633 |
+
"loss": 0.8676,
|
| 1634 |
+
"step": 5675
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 2.022711142654365,
|
| 1638 |
+
"grad_norm": 4.4236040115356445,
|
| 1639 |
+
"learning_rate": 1.8099369085173503e-05,
|
| 1640 |
+
"loss": 0.7124,
|
| 1641 |
+
"step": 5700
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 2.031582682753726,
|
| 1645 |
+
"grad_norm": 6.685351371765137,
|
| 1646 |
+
"learning_rate": 1.7935068349106204e-05,
|
| 1647 |
+
"loss": 0.864,
|
| 1648 |
+
"step": 5725
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 2.040454222853087,
|
| 1652 |
+
"grad_norm": 6.873079776763916,
|
| 1653 |
+
"learning_rate": 1.777076761303891e-05,
|
| 1654 |
+
"loss": 0.8414,
|
| 1655 |
+
"step": 5750
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 2.0493257629524484,
|
| 1659 |
+
"grad_norm": 7.3745317459106445,
|
| 1660 |
+
"learning_rate": 1.760646687697161e-05,
|
| 1661 |
+
"loss": 0.7731,
|
| 1662 |
+
"step": 5775
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 2.0581973030518097,
|
| 1666 |
+
"grad_norm": 4.843000411987305,
|
| 1667 |
+
"learning_rate": 1.744216614090431e-05,
|
| 1668 |
+
"loss": 0.6793,
|
| 1669 |
+
"step": 5800
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 2.067068843151171,
|
| 1673 |
+
"grad_norm": 8.414238929748535,
|
| 1674 |
+
"learning_rate": 1.7277865404837016e-05,
|
| 1675 |
+
"loss": 0.808,
|
| 1676 |
+
"step": 5825
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 2.0759403832505323,
|
| 1680 |
+
"grad_norm": 4.896043300628662,
|
| 1681 |
+
"learning_rate": 1.7113564668769717e-05,
|
| 1682 |
+
"loss": 0.7711,
|
| 1683 |
+
"step": 5850
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.0848119233498936,
|
| 1687 |
+
"grad_norm": 5.479866981506348,
|
| 1688 |
+
"learning_rate": 1.6949263932702418e-05,
|
| 1689 |
+
"loss": 0.8178,
|
| 1690 |
+
"step": 5875
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 2.093683463449255,
|
| 1694 |
+
"grad_norm": 5.603784561157227,
|
| 1695 |
+
"learning_rate": 1.6784963196635122e-05,
|
| 1696 |
+
"loss": 0.647,
|
| 1697 |
+
"step": 5900
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 2.1025550035486162,
|
| 1701 |
+
"grad_norm": 4.867652893066406,
|
| 1702 |
+
"learning_rate": 1.6620662460567824e-05,
|
| 1703 |
+
"loss": 0.795,
|
| 1704 |
+
"step": 5925
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 2.111426543647977,
|
| 1708 |
+
"grad_norm": 5.600367069244385,
|
| 1709 |
+
"learning_rate": 1.6456361724500528e-05,
|
| 1710 |
+
"loss": 0.8532,
|
| 1711 |
+
"step": 5950
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.1202980837473384,
|
| 1715 |
+
"grad_norm": 7.279195308685303,
|
| 1716 |
+
"learning_rate": 1.629206098843323e-05,
|
| 1717 |
+
"loss": 0.6361,
|
| 1718 |
+
"step": 5975
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.1291696238466997,
|
| 1722 |
+
"grad_norm": 11.310209274291992,
|
| 1723 |
+
"learning_rate": 1.612776025236593e-05,
|
| 1724 |
+
"loss": 0.7322,
|
| 1725 |
+
"step": 6000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 2.138041163946061,
|
| 1729 |
+
"grad_norm": 8.624905586242676,
|
| 1730 |
+
"learning_rate": 1.5963459516298635e-05,
|
| 1731 |
+
"loss": 0.8561,
|
| 1732 |
+
"step": 6025
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 2.1469127040454223,
|
| 1736 |
+
"grad_norm": 7.717365741729736,
|
| 1737 |
+
"learning_rate": 1.5799158780231336e-05,
|
| 1738 |
+
"loss": 0.7117,
|
| 1739 |
+
"step": 6050
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.1557842441447836,
|
| 1743 |
+
"grad_norm": 5.242558479309082,
|
| 1744 |
+
"learning_rate": 1.5634858044164037e-05,
|
| 1745 |
+
"loss": 0.8348,
|
| 1746 |
+
"step": 6075
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 2.164655784244145,
|
| 1750 |
+
"grad_norm": 8.414390563964844,
|
| 1751 |
+
"learning_rate": 1.547055730809674e-05,
|
| 1752 |
+
"loss": 0.8712,
|
| 1753 |
+
"step": 6100
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 2.173527324343506,
|
| 1757 |
+
"grad_norm": 4.2956461906433105,
|
| 1758 |
+
"learning_rate": 1.5306256572029443e-05,
|
| 1759 |
+
"loss": 0.8593,
|
| 1760 |
+
"step": 6125
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.182398864442867,
|
| 1764 |
+
"grad_norm": 5.302322864532471,
|
| 1765 |
+
"learning_rate": 1.5141955835962146e-05,
|
| 1766 |
+
"loss": 0.8543,
|
| 1767 |
+
"step": 6150
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 2.1912704045422284,
|
| 1771 |
+
"grad_norm": 8.463561058044434,
|
| 1772 |
+
"learning_rate": 1.497765509989485e-05,
|
| 1773 |
+
"loss": 0.7953,
|
| 1774 |
+
"step": 6175
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 2.2001419446415897,
|
| 1778 |
+
"grad_norm": 7.6317853927612305,
|
| 1779 |
+
"learning_rate": 1.481335436382755e-05,
|
| 1780 |
+
"loss": 0.8429,
|
| 1781 |
+
"step": 6200
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.209013484740951,
|
| 1785 |
+
"grad_norm": 8.177587509155273,
|
| 1786 |
+
"learning_rate": 1.4649053627760254e-05,
|
| 1787 |
+
"loss": 0.8472,
|
| 1788 |
+
"step": 6225
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.2178850248403124,
|
| 1792 |
+
"grad_norm": 9.943741798400879,
|
| 1793 |
+
"learning_rate": 1.4484752891692954e-05,
|
| 1794 |
+
"loss": 0.7034,
|
| 1795 |
+
"step": 6250
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 2.2267565649396737,
|
| 1799 |
+
"grad_norm": 5.518933296203613,
|
| 1800 |
+
"learning_rate": 1.4320452155625658e-05,
|
| 1801 |
+
"loss": 0.8279,
|
| 1802 |
+
"step": 6275
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.235628105039035,
|
| 1806 |
+
"grad_norm": 5.986506462097168,
|
| 1807 |
+
"learning_rate": 1.4156151419558361e-05,
|
| 1808 |
+
"loss": 0.7338,
|
| 1809 |
+
"step": 6300
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 2.244499645138396,
|
| 1813 |
+
"grad_norm": 6.913864612579346,
|
| 1814 |
+
"learning_rate": 1.3991850683491062e-05,
|
| 1815 |
+
"loss": 0.75,
|
| 1816 |
+
"step": 6325
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 2.253371185237757,
|
| 1820 |
+
"grad_norm": 14.090570449829102,
|
| 1821 |
+
"learning_rate": 1.3827549947423765e-05,
|
| 1822 |
+
"loss": 0.8477,
|
| 1823 |
+
"step": 6350
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.2622427253371185,
|
| 1827 |
+
"grad_norm": 6.265283584594727,
|
| 1828 |
+
"learning_rate": 1.366324921135647e-05,
|
| 1829 |
+
"loss": 0.8558,
|
| 1830 |
+
"step": 6375
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 2.2711142654364798,
|
| 1834 |
+
"grad_norm": 5.962440013885498,
|
| 1835 |
+
"learning_rate": 1.3498948475289169e-05,
|
| 1836 |
+
"loss": 0.7053,
|
| 1837 |
+
"step": 6400
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 2.279985805535841,
|
| 1841 |
+
"grad_norm": 6.866328239440918,
|
| 1842 |
+
"learning_rate": 1.3334647739221873e-05,
|
| 1843 |
+
"loss": 0.7223,
|
| 1844 |
+
"step": 6425
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 2.2888573456352024,
|
| 1848 |
+
"grad_norm": 9.418810844421387,
|
| 1849 |
+
"learning_rate": 1.3170347003154576e-05,
|
| 1850 |
+
"loss": 0.8068,
|
| 1851 |
+
"step": 6450
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 2.2977288857345637,
|
| 1855 |
+
"grad_norm": 10.400421142578125,
|
| 1856 |
+
"learning_rate": 1.3006046267087277e-05,
|
| 1857 |
+
"loss": 0.7997,
|
| 1858 |
+
"step": 6475
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 2.306600425833925,
|
| 1862 |
+
"grad_norm": 6.021897792816162,
|
| 1863 |
+
"learning_rate": 1.284174553101998e-05,
|
| 1864 |
+
"loss": 0.7601,
|
| 1865 |
+
"step": 6500
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.315471965933286,
|
| 1869 |
+
"grad_norm": 4.0054521560668945,
|
| 1870 |
+
"learning_rate": 1.2677444794952681e-05,
|
| 1871 |
+
"loss": 0.8377,
|
| 1872 |
+
"step": 6525
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 2.324343506032647,
|
| 1876 |
+
"grad_norm": 4.744009971618652,
|
| 1877 |
+
"learning_rate": 1.2513144058885384e-05,
|
| 1878 |
+
"loss": 0.8302,
|
| 1879 |
+
"step": 6550
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 2.3332150461320085,
|
| 1883 |
+
"grad_norm": 12.244515419006348,
|
| 1884 |
+
"learning_rate": 1.2348843322818087e-05,
|
| 1885 |
+
"loss": 0.797,
|
| 1886 |
+
"step": 6575
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.34208658623137,
|
| 1890 |
+
"grad_norm": 5.934150218963623,
|
| 1891 |
+
"learning_rate": 1.218454258675079e-05,
|
| 1892 |
+
"loss": 0.7326,
|
| 1893 |
+
"step": 6600
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 2.350958126330731,
|
| 1897 |
+
"grad_norm": 7.615633010864258,
|
| 1898 |
+
"learning_rate": 1.2020241850683493e-05,
|
| 1899 |
+
"loss": 0.6852,
|
| 1900 |
+
"step": 6625
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 2.3598296664300924,
|
| 1904 |
+
"grad_norm": 2.2285540103912354,
|
| 1905 |
+
"learning_rate": 1.1855941114616194e-05,
|
| 1906 |
+
"loss": 0.6879,
|
| 1907 |
+
"step": 6650
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.3687012065294537,
|
| 1911 |
+
"grad_norm": 9.67764949798584,
|
| 1912 |
+
"learning_rate": 1.1691640378548897e-05,
|
| 1913 |
+
"loss": 0.7886,
|
| 1914 |
+
"step": 6675
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 2.3775727466288146,
|
| 1918 |
+
"grad_norm": 12.79229736328125,
|
| 1919 |
+
"learning_rate": 1.15273396424816e-05,
|
| 1920 |
+
"loss": 0.8171,
|
| 1921 |
+
"step": 6700
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 2.386444286728176,
|
| 1925 |
+
"grad_norm": 8.917281150817871,
|
| 1926 |
+
"learning_rate": 1.1363038906414302e-05,
|
| 1927 |
+
"loss": 0.7607,
|
| 1928 |
+
"step": 6725
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.395315826827537,
|
| 1932 |
+
"grad_norm": 6.359137535095215,
|
| 1933 |
+
"learning_rate": 1.1198738170347003e-05,
|
| 1934 |
+
"loss": 0.7509,
|
| 1935 |
+
"step": 6750
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 2.4041873669268985,
|
| 1939 |
+
"grad_norm": 5.579113960266113,
|
| 1940 |
+
"learning_rate": 1.1034437434279706e-05,
|
| 1941 |
+
"loss": 0.7344,
|
| 1942 |
+
"step": 6775
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 2.41305890702626,
|
| 1946 |
+
"grad_norm": 5.791996002197266,
|
| 1947 |
+
"learning_rate": 1.0870136698212409e-05,
|
| 1948 |
+
"loss": 0.7222,
|
| 1949 |
+
"step": 6800
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.421930447125621,
|
| 1953 |
+
"grad_norm": 4.481837272644043,
|
| 1954 |
+
"learning_rate": 1.070583596214511e-05,
|
| 1955 |
+
"loss": 0.6455,
|
| 1956 |
+
"step": 6825
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 2.4308019872249824,
|
| 1960 |
+
"grad_norm": 5.7892069816589355,
|
| 1961 |
+
"learning_rate": 1.0541535226077813e-05,
|
| 1962 |
+
"loss": 0.6221,
|
| 1963 |
+
"step": 6850
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 2.4396735273243433,
|
| 1967 |
+
"grad_norm": 4.64011287689209,
|
| 1968 |
+
"learning_rate": 1.0377234490010516e-05,
|
| 1969 |
+
"loss": 0.6901,
|
| 1970 |
+
"step": 6875
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.4485450674237046,
|
| 1974 |
+
"grad_norm": 12.223133087158203,
|
| 1975 |
+
"learning_rate": 1.0212933753943219e-05,
|
| 1976 |
+
"loss": 0.7315,
|
| 1977 |
+
"step": 6900
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 2.457416607523066,
|
| 1981 |
+
"grad_norm": 7.667706489562988,
|
| 1982 |
+
"learning_rate": 1.004863301787592e-05,
|
| 1983 |
+
"loss": 0.7421,
|
| 1984 |
+
"step": 6925
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 2.466288147622427,
|
| 1988 |
+
"grad_norm": 6.26972770690918,
|
| 1989 |
+
"learning_rate": 9.884332281808623e-06,
|
| 1990 |
+
"loss": 0.8319,
|
| 1991 |
+
"step": 6950
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.4751596877217885,
|
| 1995 |
+
"grad_norm": 4.876287937164307,
|
| 1996 |
+
"learning_rate": 9.720031545741325e-06,
|
| 1997 |
+
"loss": 0.7305,
|
| 1998 |
+
"step": 6975
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 2.48403122782115,
|
| 2002 |
+
"grad_norm": 8.227882385253906,
|
| 2003 |
+
"learning_rate": 9.555730809674028e-06,
|
| 2004 |
+
"loss": 0.8635,
|
| 2005 |
+
"step": 7000
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 2.492902767920511,
|
| 2009 |
+
"grad_norm": 4.739542007446289,
|
| 2010 |
+
"learning_rate": 9.39143007360673e-06,
|
| 2011 |
+
"loss": 0.7291,
|
| 2012 |
+
"step": 7025
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.501774308019872,
|
| 2016 |
+
"grad_norm": 1.6183698177337646,
|
| 2017 |
+
"learning_rate": 9.227129337539432e-06,
|
| 2018 |
+
"loss": 0.7295,
|
| 2019 |
+
"step": 7050
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 2.5106458481192337,
|
| 2023 |
+
"grad_norm": 8.615501403808594,
|
| 2024 |
+
"learning_rate": 9.062828601472135e-06,
|
| 2025 |
+
"loss": 0.7208,
|
| 2026 |
+
"step": 7075
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 2.5195173882185946,
|
| 2030 |
+
"grad_norm": 6.233531951904297,
|
| 2031 |
+
"learning_rate": 8.898527865404838e-06,
|
| 2032 |
+
"loss": 0.7005,
|
| 2033 |
+
"step": 7100
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.528388928317956,
|
| 2037 |
+
"grad_norm": 5.194847106933594,
|
| 2038 |
+
"learning_rate": 8.734227129337539e-06,
|
| 2039 |
+
"loss": 0.8115,
|
| 2040 |
+
"step": 7125
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 2.5372604684173172,
|
| 2044 |
+
"grad_norm": 5.162277698516846,
|
| 2045 |
+
"learning_rate": 8.569926393270242e-06,
|
| 2046 |
+
"loss": 0.6818,
|
| 2047 |
+
"step": 7150
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 2.5461320085166785,
|
| 2051 |
+
"grad_norm": 10.57759952545166,
|
| 2052 |
+
"learning_rate": 8.405625657202945e-06,
|
| 2053 |
+
"loss": 0.7705,
|
| 2054 |
+
"step": 7175
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.55500354861604,
|
| 2058 |
+
"grad_norm": 6.097165584564209,
|
| 2059 |
+
"learning_rate": 8.241324921135648e-06,
|
| 2060 |
+
"loss": 0.7589,
|
| 2061 |
+
"step": 7200
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 2.563875088715401,
|
| 2065 |
+
"grad_norm": 2.8841516971588135,
|
| 2066 |
+
"learning_rate": 8.077024185068349e-06,
|
| 2067 |
+
"loss": 0.7758,
|
| 2068 |
+
"step": 7225
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 2.5727466288147625,
|
| 2072 |
+
"grad_norm": 8.131023406982422,
|
| 2073 |
+
"learning_rate": 7.912723449001051e-06,
|
| 2074 |
+
"loss": 0.7349,
|
| 2075 |
+
"step": 7250
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.5816181689141233,
|
| 2079 |
+
"grad_norm": 9.21007251739502,
|
| 2080 |
+
"learning_rate": 7.748422712933754e-06,
|
| 2081 |
+
"loss": 0.7964,
|
| 2082 |
+
"step": 7275
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 2.5904897090134846,
|
| 2086 |
+
"grad_norm": 4.194328784942627,
|
| 2087 |
+
"learning_rate": 7.584121976866457e-06,
|
| 2088 |
+
"loss": 0.8019,
|
| 2089 |
+
"step": 7300
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 2.599361249112846,
|
| 2093 |
+
"grad_norm": 7.350373268127441,
|
| 2094 |
+
"learning_rate": 7.419821240799159e-06,
|
| 2095 |
+
"loss": 0.7725,
|
| 2096 |
+
"step": 7325
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.6082327892122072,
|
| 2100 |
+
"grad_norm": 9.057943344116211,
|
| 2101 |
+
"learning_rate": 7.255520504731862e-06,
|
| 2102 |
+
"loss": 0.8212,
|
| 2103 |
+
"step": 7350
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 2.6171043293115686,
|
| 2107 |
+
"grad_norm": 4.8207478523254395,
|
| 2108 |
+
"learning_rate": 7.091219768664564e-06,
|
| 2109 |
+
"loss": 0.7288,
|
| 2110 |
+
"step": 7375
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 2.62597586941093,
|
| 2114 |
+
"grad_norm": 5.385066032409668,
|
| 2115 |
+
"learning_rate": 6.926919032597266e-06,
|
| 2116 |
+
"loss": 0.7013,
|
| 2117 |
+
"step": 7400
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.634847409510291,
|
| 2121 |
+
"grad_norm": 7.983443737030029,
|
| 2122 |
+
"learning_rate": 6.762618296529969e-06,
|
| 2123 |
+
"loss": 0.6656,
|
| 2124 |
+
"step": 7425
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 2.643718949609652,
|
| 2128 |
+
"grad_norm": 9.354117393493652,
|
| 2129 |
+
"learning_rate": 6.598317560462672e-06,
|
| 2130 |
+
"loss": 0.7363,
|
| 2131 |
+
"step": 7450
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 2.6525904897090133,
|
| 2135 |
+
"grad_norm": 6.475661754608154,
|
| 2136 |
+
"learning_rate": 6.4340168243953736e-06,
|
| 2137 |
+
"loss": 0.7651,
|
| 2138 |
+
"step": 7475
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.6614620298083747,
|
| 2142 |
+
"grad_norm": 5.2187676429748535,
|
| 2143 |
+
"learning_rate": 6.2697160883280756e-06,
|
| 2144 |
+
"loss": 0.7259,
|
| 2145 |
+
"step": 7500
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.670333569907736,
|
| 2149 |
+
"grad_norm": 10.267646789550781,
|
| 2150 |
+
"learning_rate": 6.105415352260778e-06,
|
| 2151 |
+
"loss": 0.7818,
|
| 2152 |
+
"step": 7525
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 2.6792051100070973,
|
| 2156 |
+
"grad_norm": 8.05725383758545,
|
| 2157 |
+
"learning_rate": 5.94111461619348e-06,
|
| 2158 |
+
"loss": 0.7006,
|
| 2159 |
+
"step": 7550
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.6880766501064586,
|
| 2163 |
+
"grad_norm": 6.517486095428467,
|
| 2164 |
+
"learning_rate": 5.776813880126183e-06,
|
| 2165 |
+
"loss": 0.7342,
|
| 2166 |
+
"step": 7575
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 2.69694819020582,
|
| 2170 |
+
"grad_norm": 5.03231954574585,
|
| 2171 |
+
"learning_rate": 5.612513144058885e-06,
|
| 2172 |
+
"loss": 0.7959,
|
| 2173 |
+
"step": 7600
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 2.7058197303051807,
|
| 2177 |
+
"grad_norm": 3.084221124649048,
|
| 2178 |
+
"learning_rate": 5.448212407991588e-06,
|
| 2179 |
+
"loss": 0.7391,
|
| 2180 |
+
"step": 7625
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.7146912704045425,
|
| 2184 |
+
"grad_norm": 6.011415004730225,
|
| 2185 |
+
"learning_rate": 5.28391167192429e-06,
|
| 2186 |
+
"loss": 0.7073,
|
| 2187 |
+
"step": 7650
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.7235628105039034,
|
| 2191 |
+
"grad_norm": 1.8170585632324219,
|
| 2192 |
+
"learning_rate": 5.119610935856993e-06,
|
| 2193 |
+
"loss": 0.6981,
|
| 2194 |
+
"step": 7675
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 2.7324343506032647,
|
| 2198 |
+
"grad_norm": 8.507579803466797,
|
| 2199 |
+
"learning_rate": 4.955310199789695e-06,
|
| 2200 |
+
"loss": 0.8989,
|
| 2201 |
+
"step": 7700
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.741305890702626,
|
| 2205 |
+
"grad_norm": 4.660305976867676,
|
| 2206 |
+
"learning_rate": 4.791009463722398e-06,
|
| 2207 |
+
"loss": 0.7881,
|
| 2208 |
+
"step": 7725
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 2.7501774308019873,
|
| 2212 |
+
"grad_norm": 5.1059088706970215,
|
| 2213 |
+
"learning_rate": 4.6267087276551e-06,
|
| 2214 |
+
"loss": 0.6829,
|
| 2215 |
+
"step": 7750
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 2.7590489709013486,
|
| 2219 |
+
"grad_norm": 6.890865802764893,
|
| 2220 |
+
"learning_rate": 4.4624079915878024e-06,
|
| 2221 |
+
"loss": 0.7441,
|
| 2222 |
+
"step": 7775
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 2.7679205110007095,
|
| 2226 |
+
"grad_norm": 6.044835090637207,
|
| 2227 |
+
"learning_rate": 4.298107255520505e-06,
|
| 2228 |
+
"loss": 0.7416,
|
| 2229 |
+
"step": 7800
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 2.776792051100071,
|
| 2233 |
+
"grad_norm": 9.076371192932129,
|
| 2234 |
+
"learning_rate": 4.133806519453207e-06,
|
| 2235 |
+
"loss": 0.7572,
|
| 2236 |
+
"step": 7825
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 2.785663591199432,
|
| 2240 |
+
"grad_norm": 8.244555473327637,
|
| 2241 |
+
"learning_rate": 3.96950578338591e-06,
|
| 2242 |
+
"loss": 0.7547,
|
| 2243 |
+
"step": 7850
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.7945351312987934,
|
| 2247 |
+
"grad_norm": 10.965091705322266,
|
| 2248 |
+
"learning_rate": 3.805205047318612e-06,
|
| 2249 |
+
"loss": 0.7462,
|
| 2250 |
+
"step": 7875
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 2.8034066713981547,
|
| 2254 |
+
"grad_norm": 10.940464973449707,
|
| 2255 |
+
"learning_rate": 3.640904311251315e-06,
|
| 2256 |
+
"loss": 0.8307,
|
| 2257 |
+
"step": 7900
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 2.812278211497516,
|
| 2261 |
+
"grad_norm": 6.724726676940918,
|
| 2262 |
+
"learning_rate": 3.476603575184017e-06,
|
| 2263 |
+
"loss": 0.7771,
|
| 2264 |
+
"step": 7925
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 2.8211497515968773,
|
| 2268 |
+
"grad_norm": 8.394052505493164,
|
| 2269 |
+
"learning_rate": 3.3123028391167197e-06,
|
| 2270 |
+
"loss": 0.7948,
|
| 2271 |
+
"step": 7950
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 2.8300212916962386,
|
| 2275 |
+
"grad_norm": 3.6263391971588135,
|
| 2276 |
+
"learning_rate": 3.1480021030494217e-06,
|
| 2277 |
+
"loss": 0.6567,
|
| 2278 |
+
"step": 7975
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 2.8388928317956,
|
| 2282 |
+
"grad_norm": 7.270437717437744,
|
| 2283 |
+
"learning_rate": 2.983701366982124e-06,
|
| 2284 |
+
"loss": 0.7087,
|
| 2285 |
+
"step": 8000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.847764371894961,
|
| 2289 |
+
"grad_norm": 6.618618965148926,
|
| 2290 |
+
"learning_rate": 2.819400630914827e-06,
|
| 2291 |
+
"loss": 0.6796,
|
| 2292 |
+
"step": 8025
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 2.856635911994322,
|
| 2296 |
+
"grad_norm": 7.489841938018799,
|
| 2297 |
+
"learning_rate": 2.6550998948475293e-06,
|
| 2298 |
+
"loss": 0.8368,
|
| 2299 |
+
"step": 8050
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 2.8655074520936834,
|
| 2303 |
+
"grad_norm": 3.919420003890991,
|
| 2304 |
+
"learning_rate": 2.4907991587802318e-06,
|
| 2305 |
+
"loss": 0.7596,
|
| 2306 |
+
"step": 8075
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 2.8743789921930447,
|
| 2310 |
+
"grad_norm": 7.457404136657715,
|
| 2311 |
+
"learning_rate": 2.326498422712934e-06,
|
| 2312 |
+
"loss": 0.7399,
|
| 2313 |
+
"step": 8100
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 2.883250532292406,
|
| 2317 |
+
"grad_norm": 11.167478561401367,
|
| 2318 |
+
"learning_rate": 2.1621976866456366e-06,
|
| 2319 |
+
"loss": 0.8489,
|
| 2320 |
+
"step": 8125
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 2.8921220723917673,
|
| 2324 |
+
"grad_norm": 11.763769149780273,
|
| 2325 |
+
"learning_rate": 1.9978969505783385e-06,
|
| 2326 |
+
"loss": 0.6573,
|
| 2327 |
+
"step": 8150
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.9009936124911286,
|
| 2331 |
+
"grad_norm": 5.394194602966309,
|
| 2332 |
+
"learning_rate": 1.8335962145110412e-06,
|
| 2333 |
+
"loss": 0.6636,
|
| 2334 |
+
"step": 8175
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 2.9098651525904895,
|
| 2338 |
+
"grad_norm": 7.921031475067139,
|
| 2339 |
+
"learning_rate": 1.6692954784437436e-06,
|
| 2340 |
+
"loss": 0.7437,
|
| 2341 |
+
"step": 8200
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 2.918736692689851,
|
| 2345 |
+
"grad_norm": 5.872561454772949,
|
| 2346 |
+
"learning_rate": 1.504994742376446e-06,
|
| 2347 |
+
"loss": 0.7926,
|
| 2348 |
+
"step": 8225
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 2.927608232789212,
|
| 2352 |
+
"grad_norm": 4.337924480438232,
|
| 2353 |
+
"learning_rate": 1.3406940063091484e-06,
|
| 2354 |
+
"loss": 0.7415,
|
| 2355 |
+
"step": 8250
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 2.9364797728885734,
|
| 2359 |
+
"grad_norm": 6.103896617889404,
|
| 2360 |
+
"learning_rate": 1.1763932702418508e-06,
|
| 2361 |
+
"loss": 0.7814,
|
| 2362 |
+
"step": 8275
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 2.9453513129879347,
|
| 2366 |
+
"grad_norm": 3.75592041015625,
|
| 2367 |
+
"learning_rate": 1.0120925341745532e-06,
|
| 2368 |
+
"loss": 0.7166,
|
| 2369 |
+
"step": 8300
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.954222853087296,
|
| 2373 |
+
"grad_norm": 4.261561870574951,
|
| 2374 |
+
"learning_rate": 8.477917981072555e-07,
|
| 2375 |
+
"loss": 0.735,
|
| 2376 |
+
"step": 8325
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 2.9630943931866573,
|
| 2380 |
+
"grad_norm": 6.112985134124756,
|
| 2381 |
+
"learning_rate": 6.83491062039958e-07,
|
| 2382 |
+
"loss": 0.7205,
|
| 2383 |
+
"step": 8350
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 2.971965933286018,
|
| 2387 |
+
"grad_norm": 4.848855018615723,
|
| 2388 |
+
"learning_rate": 5.191903259726603e-07,
|
| 2389 |
+
"loss": 0.7653,
|
| 2390 |
+
"step": 8375
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 2.98083747338538,
|
| 2394 |
+
"grad_norm": 7.813061714172363,
|
| 2395 |
+
"learning_rate": 3.548895899053628e-07,
|
| 2396 |
+
"loss": 0.7253,
|
| 2397 |
+
"step": 8400
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 2.989709013484741,
|
| 2401 |
+
"grad_norm": 6.210537433624268,
|
| 2402 |
+
"learning_rate": 1.905888538380652e-07,
|
| 2403 |
+
"loss": 0.8043,
|
| 2404 |
+
"step": 8425
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 2.998580553584102,
|
| 2408 |
+
"grad_norm": 4.216333389282227,
|
| 2409 |
+
"learning_rate": 2.6288117770767614e-08,
|
| 2410 |
+
"loss": 0.7481,
|
| 2411 |
+
"step": 8450
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 3.0,
|
| 2415 |
+
"eval_accuracy": 0.6885536823425022,
|
| 2416 |
+
"eval_f1_macro": 0.6934027713968594,
|
| 2417 |
+
"eval_f1_micro": 0.6885536823425022,
|
| 2418 |
+
"eval_f1_weighted": 0.6903326985699842,
|
| 2419 |
+
"eval_loss": 0.7668559551239014,
|
| 2420 |
+
"eval_precision_macro": 0.7104396368026238,
|
| 2421 |
+
"eval_precision_micro": 0.6885536823425022,
|
| 2422 |
+
"eval_precision_weighted": 0.705612321743312,
|
| 2423 |
+
"eval_recall_macro": 0.6898371166175705,
|
| 2424 |
+
"eval_recall_micro": 0.6885536823425022,
|
| 2425 |
+
"eval_recall_weighted": 0.6885536823425022,
|
| 2426 |
+
"eval_runtime": 1239.6725,
|
| 2427 |
+
"eval_samples_per_second": 4.546,
|
| 2428 |
+
"eval_steps_per_second": 0.285,
|
| 2429 |
+
"step": 8454
|
| 2430 |
+
}
|
| 2431 |
+
],
|
| 2432 |
+
"logging_steps": 25,
|
| 2433 |
+
"max_steps": 8454,
|
| 2434 |
+
"num_input_tokens_seen": 0,
|
| 2435 |
+
"num_train_epochs": 3,
|
| 2436 |
+
"save_steps": 500,
|
| 2437 |
+
"stateful_callbacks": {
|
| 2438 |
+
"EarlyStoppingCallback": {
|
| 2439 |
+
"args": {
|
| 2440 |
+
"early_stopping_patience": 5,
|
| 2441 |
+
"early_stopping_threshold": 0.01
|
| 2442 |
+
},
|
| 2443 |
+
"attributes": {
|
| 2444 |
+
"early_stopping_patience_counter": 0
|
| 2445 |
+
}
|
| 2446 |
+
},
|
| 2447 |
+
"TrainerControl": {
|
| 2448 |
+
"args": {
|
| 2449 |
+
"should_epoch_stop": false,
|
| 2450 |
+
"should_evaluate": false,
|
| 2451 |
+
"should_log": false,
|
| 2452 |
+
"should_save": true,
|
| 2453 |
+
"should_training_stop": true
|
| 2454 |
+
},
|
| 2455 |
+
"attributes": {}
|
| 2456 |
+
}
|
| 2457 |
+
},
|
| 2458 |
+
"total_flos": 4447420182533376.0,
|
| 2459 |
+
"train_batch_size": 8,
|
| 2460 |
+
"trial_name": null,
|
| 2461 |
+
"trial_params": null
|
| 2462 |
+
}
|
checkpoint-8454/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:53979d3a7c4167b7fc8acf3d54bf0328b9ab43910addc7c9b0949c14f05c4483
|
| 3 |
+
size 5240
|
config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "neuralmind/bert-base-portuguese-cased",
|
| 3 |
+
"_num_labels": 5,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "1 - INTIMA\u00c7\u00c3O INICIAL",
|
| 15 |
+
"1": "2 - INTIMA\u00c7\u00c3O INSTRU\u00c7\u00c3O",
|
| 16 |
+
"2": "3 - INTIMA\u00c7\u00c3O RECURSAL",
|
| 17 |
+
"3": "4 - INTIMA\u00c7\u00c3O CUMPRIMENTO DE SENTEN\u00c7A",
|
| 18 |
+
"4": "5 - INTIMA\u00c7\u00c3O TR\u00c2NSITO"
|
| 19 |
+
},
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 3072,
|
| 22 |
+
"label2id": {
|
| 23 |
+
"1 - INTIMA\u00c7\u00c3O INICIAL": 0,
|
| 24 |
+
"2 - INTIMA\u00c7\u00c3O INSTRU\u00c7\u00c3O": 1,
|
| 25 |
+
"3 - INTIMA\u00c7\u00c3O RECURSAL": 2,
|
| 26 |
+
"4 - INTIMA\u00c7\u00c3O CUMPRIMENTO DE SENTEN\u00c7A": 3,
|
| 27 |
+
"5 - INTIMA\u00c7\u00c3O TR\u00c2NSITO": 4
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-12,
|
| 30 |
+
"max_position_embeddings": 512,
|
| 31 |
+
"model_type": "bert",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"output_past": true,
|
| 35 |
+
"pad_token_id": 0,
|
| 36 |
+
"pooler_fc_size": 768,
|
| 37 |
+
"pooler_num_attention_heads": 12,
|
| 38 |
+
"pooler_num_fc_layers": 3,
|
| 39 |
+
"pooler_size_per_head": 128,
|
| 40 |
+
"pooler_type": "first_token_transform",
|
| 41 |
+
"position_embedding_type": "absolute",
|
| 42 |
+
"problem_type": "single_label_classification",
|
| 43 |
+
"torch_dtype": "float32",
|
| 44 |
+
"transformers_version": "4.45.0",
|
| 45 |
+
"type_vocab_size": 2,
|
| 46 |
+
"use_cache": true,
|
| 47 |
+
"vocab_size": 29794
|
| 48 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7b5aea858f25e24827711cfbc02c35b160928b6c19e405df1c24d90f5be4a7d
|
| 3 |
+
size 435731452
|
runs/Oct30_19-52-50_r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva/events.out.tfevents.1730317972.r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva.112.0
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:d7d32a5bd25a1376d5e195451aebc68b8991c1afd7eb19cb027903b565e887c0
|
| 3 |
+
size 79890
|
runs/Oct30_19-52-50_r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva/events.out.tfevents.1730377445.r-ulisesbravo-treino-teste-boqirknt-9f37e-oreva.112.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e4db22b0bcb567a589fb70930710cc3490d846471da2098519c2d172761ec4aa
|
| 3 |
+
size 921
|
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 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": null,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:53979d3a7c4167b7fc8acf3d54bf0328b9ab43910addc7c9b0949c14f05c4483
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "autotrain-k9ag8-b7nm0/autotrain-data",
|
| 3 |
+
"model": "neuralmind/bert-base-portuguese-cased",
|
| 4 |
+
"lr": 5e-05,
|
| 5 |
+
"epochs": 3,
|
| 6 |
+
"max_seq_length": 128,
|
| 7 |
+
"batch_size": 8,
|
| 8 |
+
"warmup_ratio": 0.1,
|
| 9 |
+
"gradient_accumulation": 1,
|
| 10 |
+
"optimizer": "adamw_torch",
|
| 11 |
+
"scheduler": "linear",
|
| 12 |
+
"weight_decay": 0.0,
|
| 13 |
+
"max_grad_norm": 1.0,
|
| 14 |
+
"seed": 42,
|
| 15 |
+
"train_split": "train",
|
| 16 |
+
"valid_split": "validation",
|
| 17 |
+
"text_column": "autotrain_text",
|
| 18 |
+
"target_column": "autotrain_label",
|
| 19 |
+
"logging_steps": -1,
|
| 20 |
+
"project_name": "autotrain-k9ag8-b7nm0",
|
| 21 |
+
"auto_find_batch_size": false,
|
| 22 |
+
"mixed_precision": "fp16",
|
| 23 |
+
"save_total_limit": 1,
|
| 24 |
+
"push_to_hub": true,
|
| 25 |
+
"eval_strategy": "epoch",
|
| 26 |
+
"username": "ulisesbravo",
|
| 27 |
+
"log": "tensorboard",
|
| 28 |
+
"early_stopping_patience": 5,
|
| 29 |
+
"early_stopping_threshold": 0.01
|
| 30 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|