Upload folder using huggingface_hub
Browse files- README.md +27 -0
- checkpoint-6135/config.json +36 -0
- checkpoint-6135/model.safetensors +3 -0
- checkpoint-6135/optimizer.pt +3 -0
- checkpoint-6135/rng_state.pth +3 -0
- checkpoint-6135/scheduler.pt +3 -0
- checkpoint-6135/trainer_state.json +1796 -0
- checkpoint-6135/training_args.bin +3 -0
- config.json +36 -0
- model.safetensors +3 -0
- runs/May29_13-39-01_r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz/events.out.tfevents.1748525943.r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz.76.0 +2 -2
- runs/May29_13-39-01_r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz/events.out.tfevents.1748526547.r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz.76.1 +3 -0
- special_tokens_map.json +37 -0
- tokenizer.json +0 -0
- tokenizer_config.json +63 -0
- training_args.bin +3 -0
- training_params.json +30 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- autotrain
|
| 6 |
+
- text-classification
|
| 7 |
+
base_model: thenlper/gte-base
|
| 8 |
+
widget:
|
| 9 |
+
- text: "I love AutoTrain"
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Trained Using AutoTrain
|
| 13 |
+
|
| 14 |
+
- Problem type: Text Classification
|
| 15 |
+
|
| 16 |
+
## Validation Metrics
|
| 17 |
+
loss: 0.03548985347151756
|
| 18 |
+
|
| 19 |
+
f1: 0.9950522264980759
|
| 20 |
+
|
| 21 |
+
precision: 0.9945054945054945
|
| 22 |
+
|
| 23 |
+
recall: 0.9955995599559956
|
| 24 |
+
|
| 25 |
+
auc: 0.9997361672360855
|
| 26 |
+
|
| 27 |
+
accuracy: 0.995049504950495
|
checkpoint-6135/config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "thenlper/gte-base",
|
| 3 |
+
"_num_labels": 2,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "human",
|
| 15 |
+
"1": "llm"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 3072,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"human": 0,
|
| 21 |
+
"llm": 1
|
| 22 |
+
},
|
| 23 |
+
"layer_norm_eps": 1e-12,
|
| 24 |
+
"max_position_embeddings": 512,
|
| 25 |
+
"model_type": "bert",
|
| 26 |
+
"num_attention_heads": 12,
|
| 27 |
+
"num_hidden_layers": 12,
|
| 28 |
+
"pad_token_id": 0,
|
| 29 |
+
"position_embedding_type": "absolute",
|
| 30 |
+
"problem_type": "single_label_classification",
|
| 31 |
+
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.48.0",
|
| 33 |
+
"type_vocab_size": 2,
|
| 34 |
+
"use_cache": true,
|
| 35 |
+
"vocab_size": 30522
|
| 36 |
+
}
|
checkpoint-6135/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ba5d6da12fadf4d20c73020e0c623a4371f36c06756ed7a0d90544a103ad255
|
| 3 |
+
size 437958648
|
checkpoint-6135/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8f8da023fdfa0a34c172f01f4b5ca3e3fe9e72a2f2a11ceb83e70924221f092
|
| 3 |
+
size 876038394
|
checkpoint-6135/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c25038c1a620bb7db5d4ff57e7991e4748ddd19926112a9062532c61306695f9
|
| 3 |
+
size 14244
|
checkpoint-6135/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4bfcd9ab4dc405b5b3aae9c30493a08cd5e9184c679c0c4de000fa18b67acb42
|
| 3 |
+
size 1064
|
checkpoint-6135/trainer_state.json
ADDED
|
@@ -0,0 +1,1796 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.03548985347151756,
|
| 3 |
+
"best_model_checkpoint": "slop-detector/checkpoint-6135",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 6135,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.012224938875305624,
|
| 13 |
+
"grad_norm": 5.421860694885254,
|
| 14 |
+
"learning_rate": 1.8729641693811077e-06,
|
| 15 |
+
"loss": 0.7509,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.02444987775061125,
|
| 20 |
+
"grad_norm": 8.084574699401855,
|
| 21 |
+
"learning_rate": 3.908794788273615e-06,
|
| 22 |
+
"loss": 0.7689,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.03667481662591687,
|
| 27 |
+
"grad_norm": 4.011590480804443,
|
| 28 |
+
"learning_rate": 5.944625407166124e-06,
|
| 29 |
+
"loss": 0.7339,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0488997555012225,
|
| 34 |
+
"grad_norm": 6.365348815917969,
|
| 35 |
+
"learning_rate": 7.980456026058632e-06,
|
| 36 |
+
"loss": 0.6892,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.061124694376528114,
|
| 41 |
+
"grad_norm": 5.623918056488037,
|
| 42 |
+
"learning_rate": 1.0016286644951141e-05,
|
| 43 |
+
"loss": 0.6427,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.07334963325183375,
|
| 48 |
+
"grad_norm": 15.913942337036133,
|
| 49 |
+
"learning_rate": 1.1970684039087948e-05,
|
| 50 |
+
"loss": 0.6024,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.08557457212713937,
|
| 55 |
+
"grad_norm": 4.529036998748779,
|
| 56 |
+
"learning_rate": 1.4006514657980457e-05,
|
| 57 |
+
"loss": 0.4899,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.097799511002445,
|
| 62 |
+
"grad_norm": 6.691816806793213,
|
| 63 |
+
"learning_rate": 1.6042345276872963e-05,
|
| 64 |
+
"loss": 0.3174,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.1100244498777506,
|
| 69 |
+
"grad_norm": 2.052150011062622,
|
| 70 |
+
"learning_rate": 1.8078175895765474e-05,
|
| 71 |
+
"loss": 0.1372,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.12224938875305623,
|
| 76 |
+
"grad_norm": 0.288004994392395,
|
| 77 |
+
"learning_rate": 2.0114006514657982e-05,
|
| 78 |
+
"loss": 0.058,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.13447432762836187,
|
| 83 |
+
"grad_norm": 0.1465359628200531,
|
| 84 |
+
"learning_rate": 2.214983713355049e-05,
|
| 85 |
+
"loss": 0.0221,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1466992665036675,
|
| 90 |
+
"grad_norm": 0.12296482175588608,
|
| 91 |
+
"learning_rate": 2.4185667752442997e-05,
|
| 92 |
+
"loss": 0.0489,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.15892420537897312,
|
| 97 |
+
"grad_norm": 0.10336871445178986,
|
| 98 |
+
"learning_rate": 2.6140065146579805e-05,
|
| 99 |
+
"loss": 0.0513,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.17114914425427874,
|
| 104 |
+
"grad_norm": 0.10855236649513245,
|
| 105 |
+
"learning_rate": 2.8175895765472316e-05,
|
| 106 |
+
"loss": 0.0287,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.18337408312958436,
|
| 111 |
+
"grad_norm": 0.060183338820934296,
|
| 112 |
+
"learning_rate": 3.0211726384364824e-05,
|
| 113 |
+
"loss": 0.0712,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.19559902200489,
|
| 118 |
+
"grad_norm": 0.08147333562374115,
|
| 119 |
+
"learning_rate": 3.224755700325733e-05,
|
| 120 |
+
"loss": 0.1589,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.2078239608801956,
|
| 125 |
+
"grad_norm": 14.867992401123047,
|
| 126 |
+
"learning_rate": 3.4283387622149836e-05,
|
| 127 |
+
"loss": 0.0422,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.2200488997555012,
|
| 132 |
+
"grad_norm": 0.045378606766462326,
|
| 133 |
+
"learning_rate": 3.6319218241042346e-05,
|
| 134 |
+
"loss": 0.0879,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.23227383863080683,
|
| 139 |
+
"grad_norm": 0.05560455471277237,
|
| 140 |
+
"learning_rate": 3.835504885993486e-05,
|
| 141 |
+
"loss": 0.0732,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.24449877750611246,
|
| 146 |
+
"grad_norm": 0.05216033384203911,
|
| 147 |
+
"learning_rate": 4.039087947882736e-05,
|
| 148 |
+
"loss": 0.1231,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.2567237163814181,
|
| 153 |
+
"grad_norm": 24.66847038269043,
|
| 154 |
+
"learning_rate": 4.242671009771987e-05,
|
| 155 |
+
"loss": 0.0768,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.26894865525672373,
|
| 160 |
+
"grad_norm": 10.290877342224121,
|
| 161 |
+
"learning_rate": 4.446254071661238e-05,
|
| 162 |
+
"loss": 0.1635,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.28117359413202936,
|
| 167 |
+
"grad_norm": 0.16483809053897858,
|
| 168 |
+
"learning_rate": 4.649837133550489e-05,
|
| 169 |
+
"loss": 0.1273,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.293398533007335,
|
| 174 |
+
"grad_norm": 0.1313319206237793,
|
| 175 |
+
"learning_rate": 4.85342019543974e-05,
|
| 176 |
+
"loss": 0.0316,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.3056234718826406,
|
| 181 |
+
"grad_norm": 0.05546192079782486,
|
| 182 |
+
"learning_rate": 4.9936605687375476e-05,
|
| 183 |
+
"loss": 0.0012,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.31784841075794623,
|
| 188 |
+
"grad_norm": 0.06735104322433472,
|
| 189 |
+
"learning_rate": 4.971019742800217e-05,
|
| 190 |
+
"loss": 0.0536,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.33007334963325186,
|
| 195 |
+
"grad_norm": 0.01918862573802471,
|
| 196 |
+
"learning_rate": 4.9483789168628876e-05,
|
| 197 |
+
"loss": 0.0644,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.3422982885085575,
|
| 202 |
+
"grad_norm": 0.02696409448981285,
|
| 203 |
+
"learning_rate": 4.925738090925557e-05,
|
| 204 |
+
"loss": 0.0915,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.3545232273838631,
|
| 209 |
+
"grad_norm": 0.026020344346761703,
|
| 210 |
+
"learning_rate": 4.903097264988227e-05,
|
| 211 |
+
"loss": 0.0289,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.36674816625916873,
|
| 216 |
+
"grad_norm": 0.030367497354745865,
|
| 217 |
+
"learning_rate": 4.880456439050897e-05,
|
| 218 |
+
"loss": 0.0009,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.37897310513447435,
|
| 223 |
+
"grad_norm": 0.026138590648770332,
|
| 224 |
+
"learning_rate": 4.8578156131135664e-05,
|
| 225 |
+
"loss": 0.0904,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.39119804400978,
|
| 230 |
+
"grad_norm": 0.019473830237984657,
|
| 231 |
+
"learning_rate": 4.835174787176237e-05,
|
| 232 |
+
"loss": 0.0136,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.4034229828850856,
|
| 237 |
+
"grad_norm": 0.020081259310245514,
|
| 238 |
+
"learning_rate": 4.8125339612389064e-05,
|
| 239 |
+
"loss": 0.0083,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.4156479217603912,
|
| 244 |
+
"grad_norm": 0.006070868577808142,
|
| 245 |
+
"learning_rate": 4.789893135301576e-05,
|
| 246 |
+
"loss": 0.0231,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.4278728606356968,
|
| 251 |
+
"grad_norm": 0.005923378746956587,
|
| 252 |
+
"learning_rate": 4.767252309364246e-05,
|
| 253 |
+
"loss": 0.0672,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.4400977995110024,
|
| 258 |
+
"grad_norm": 5.1542463302612305,
|
| 259 |
+
"learning_rate": 4.7446114834269155e-05,
|
| 260 |
+
"loss": 0.0373,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.45232273838630804,
|
| 265 |
+
"grad_norm": 0.020631220191717148,
|
| 266 |
+
"learning_rate": 4.721970657489586e-05,
|
| 267 |
+
"loss": 0.1001,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.46454767726161367,
|
| 272 |
+
"grad_norm": 0.024150336161255836,
|
| 273 |
+
"learning_rate": 4.6993298315522555e-05,
|
| 274 |
+
"loss": 0.0381,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.4767726161369193,
|
| 279 |
+
"grad_norm": 0.009151147678494453,
|
| 280 |
+
"learning_rate": 4.6766890056149245e-05,
|
| 281 |
+
"loss": 0.0004,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.4889975550122249,
|
| 286 |
+
"grad_norm": 1.288160800933838,
|
| 287 |
+
"learning_rate": 4.654048179677595e-05,
|
| 288 |
+
"loss": 0.0773,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.5012224938875306,
|
| 293 |
+
"grad_norm": 0.02753949724137783,
|
| 294 |
+
"learning_rate": 4.6314073537402646e-05,
|
| 295 |
+
"loss": 0.0723,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.5134474327628362,
|
| 300 |
+
"grad_norm": 0.015499632805585861,
|
| 301 |
+
"learning_rate": 4.608766527802935e-05,
|
| 302 |
+
"loss": 0.0448,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.5256723716381418,
|
| 307 |
+
"grad_norm": 6.1461663246154785,
|
| 308 |
+
"learning_rate": 4.5861257018656046e-05,
|
| 309 |
+
"loss": 0.0421,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.5378973105134475,
|
| 314 |
+
"grad_norm": 0.017640691250562668,
|
| 315 |
+
"learning_rate": 4.5634848759282736e-05,
|
| 316 |
+
"loss": 0.0523,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.5501222493887531,
|
| 321 |
+
"grad_norm": 0.021943895146250725,
|
| 322 |
+
"learning_rate": 4.540844049990944e-05,
|
| 323 |
+
"loss": 0.0009,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.5623471882640587,
|
| 328 |
+
"grad_norm": 0.016655191779136658,
|
| 329 |
+
"learning_rate": 4.518203224053614e-05,
|
| 330 |
+
"loss": 0.0407,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.5745721271393643,
|
| 335 |
+
"grad_norm": 8.750455856323242,
|
| 336 |
+
"learning_rate": 4.4955623981162834e-05,
|
| 337 |
+
"loss": 0.0909,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.58679706601467,
|
| 342 |
+
"grad_norm": 0.19946545362472534,
|
| 343 |
+
"learning_rate": 4.472921572178954e-05,
|
| 344 |
+
"loss": 0.0574,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.5990220048899756,
|
| 349 |
+
"grad_norm": 0.017342837527394295,
|
| 350 |
+
"learning_rate": 4.450280746241623e-05,
|
| 351 |
+
"loss": 0.0345,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.6112469437652812,
|
| 356 |
+
"grad_norm": 0.021776428446173668,
|
| 357 |
+
"learning_rate": 4.427639920304293e-05,
|
| 358 |
+
"loss": 0.0524,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.6234718826405868,
|
| 363 |
+
"grad_norm": 0.024793416261672974,
|
| 364 |
+
"learning_rate": 4.404999094366963e-05,
|
| 365 |
+
"loss": 0.0341,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.6356968215158925,
|
| 370 |
+
"grad_norm": 0.013999883085489273,
|
| 371 |
+
"learning_rate": 4.3823582684296325e-05,
|
| 372 |
+
"loss": 0.1097,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.6479217603911981,
|
| 377 |
+
"grad_norm": 0.1016734391450882,
|
| 378 |
+
"learning_rate": 4.359717442492302e-05,
|
| 379 |
+
"loss": 0.0334,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.6601466992665037,
|
| 384 |
+
"grad_norm": 0.015132768079638481,
|
| 385 |
+
"learning_rate": 4.337076616554972e-05,
|
| 386 |
+
"loss": 0.0042,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.6723716381418093,
|
| 391 |
+
"grad_norm": 0.014992584474384785,
|
| 392 |
+
"learning_rate": 4.314435790617642e-05,
|
| 393 |
+
"loss": 0.0348,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.684596577017115,
|
| 398 |
+
"grad_norm": 0.03999190032482147,
|
| 399 |
+
"learning_rate": 4.291794964680312e-05,
|
| 400 |
+
"loss": 0.0091,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.6968215158924206,
|
| 405 |
+
"grad_norm": 0.014326846227049828,
|
| 406 |
+
"learning_rate": 4.2691541387429816e-05,
|
| 407 |
+
"loss": 0.1053,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.7090464547677262,
|
| 412 |
+
"grad_norm": 7.060765266418457,
|
| 413 |
+
"learning_rate": 4.246513312805651e-05,
|
| 414 |
+
"loss": 0.0711,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.7212713936430318,
|
| 419 |
+
"grad_norm": 57.96589279174805,
|
| 420 |
+
"learning_rate": 4.223872486868321e-05,
|
| 421 |
+
"loss": 0.0546,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.7334963325183375,
|
| 426 |
+
"grad_norm": 100.44673919677734,
|
| 427 |
+
"learning_rate": 4.201231660930991e-05,
|
| 428 |
+
"loss": 0.0127,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.7457212713936431,
|
| 433 |
+
"grad_norm": 0.015451326034963131,
|
| 434 |
+
"learning_rate": 4.178590834993661e-05,
|
| 435 |
+
"loss": 0.0008,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.7579462102689487,
|
| 440 |
+
"grad_norm": 0.021829571574926376,
|
| 441 |
+
"learning_rate": 4.155950009056331e-05,
|
| 442 |
+
"loss": 0.0817,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.7701711491442543,
|
| 447 |
+
"grad_norm": 0.03365129604935646,
|
| 448 |
+
"learning_rate": 4.1333091831190003e-05,
|
| 449 |
+
"loss": 0.1365,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.78239608801956,
|
| 454 |
+
"grad_norm": 0.02983665280044079,
|
| 455 |
+
"learning_rate": 4.11066835718167e-05,
|
| 456 |
+
"loss": 0.0298,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.7946210268948656,
|
| 461 |
+
"grad_norm": 8.518478393554688,
|
| 462 |
+
"learning_rate": 4.08802753124434e-05,
|
| 463 |
+
"loss": 0.1176,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.8068459657701712,
|
| 468 |
+
"grad_norm": 0.1485838145017624,
|
| 469 |
+
"learning_rate": 4.06538670530701e-05,
|
| 470 |
+
"loss": 0.0512,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.8190709046454768,
|
| 475 |
+
"grad_norm": 5.974130153656006,
|
| 476 |
+
"learning_rate": 4.04274587936968e-05,
|
| 477 |
+
"loss": 0.0809,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.8312958435207825,
|
| 482 |
+
"grad_norm": 0.029043443500995636,
|
| 483 |
+
"learning_rate": 4.0201050534323494e-05,
|
| 484 |
+
"loss": 0.03,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.843520782396088,
|
| 489 |
+
"grad_norm": 0.016874758526682854,
|
| 490 |
+
"learning_rate": 3.997464227495019e-05,
|
| 491 |
+
"loss": 0.0305,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.8557457212713936,
|
| 496 |
+
"grad_norm": 0.010613476857542992,
|
| 497 |
+
"learning_rate": 3.974823401557689e-05,
|
| 498 |
+
"loss": 0.0048,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.8679706601466992,
|
| 503 |
+
"grad_norm": 0.041933540254831314,
|
| 504 |
+
"learning_rate": 3.952182575620359e-05,
|
| 505 |
+
"loss": 0.141,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.8801955990220048,
|
| 510 |
+
"grad_norm": 0.01465628296136856,
|
| 511 |
+
"learning_rate": 3.929541749683028e-05,
|
| 512 |
+
"loss": 0.0348,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.8924205378973105,
|
| 517 |
+
"grad_norm": 143.03744506835938,
|
| 518 |
+
"learning_rate": 3.9069009237456985e-05,
|
| 519 |
+
"loss": 0.0474,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.9046454767726161,
|
| 524 |
+
"grad_norm": 0.010270562954246998,
|
| 525 |
+
"learning_rate": 3.884260097808368e-05,
|
| 526 |
+
"loss": 0.0283,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.9168704156479217,
|
| 531 |
+
"grad_norm": 0.00653125811368227,
|
| 532 |
+
"learning_rate": 3.861619271871038e-05,
|
| 533 |
+
"loss": 0.0033,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.9290953545232273,
|
| 538 |
+
"grad_norm": 0.005124003626406193,
|
| 539 |
+
"learning_rate": 3.838978445933708e-05,
|
| 540 |
+
"loss": 0.0027,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.941320293398533,
|
| 545 |
+
"grad_norm": 0.010172693058848381,
|
| 546 |
+
"learning_rate": 3.816337619996377e-05,
|
| 547 |
+
"loss": 0.0695,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.9535452322738386,
|
| 552 |
+
"grad_norm": 0.010921374894678593,
|
| 553 |
+
"learning_rate": 3.7936967940590476e-05,
|
| 554 |
+
"loss": 0.0879,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.9657701711491442,
|
| 559 |
+
"grad_norm": 0.07822031527757645,
|
| 560 |
+
"learning_rate": 3.771055968121717e-05,
|
| 561 |
+
"loss": 0.0517,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.9779951100244498,
|
| 566 |
+
"grad_norm": 0.03018801286816597,
|
| 567 |
+
"learning_rate": 3.748415142184387e-05,
|
| 568 |
+
"loss": 0.0595,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.9902200488997555,
|
| 573 |
+
"grad_norm": 0.559971034526825,
|
| 574 |
+
"learning_rate": 3.7257743162470574e-05,
|
| 575 |
+
"loss": 0.0476,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 1.0,
|
| 580 |
+
"eval_accuracy": 0.9796479647964796,
|
| 581 |
+
"eval_auc": 0.9993337617590141,
|
| 582 |
+
"eval_f1": 0.9800108049702864,
|
| 583 |
+
"eval_loss": 0.11735348403453827,
|
| 584 |
+
"eval_precision": 0.9628450106157113,
|
| 585 |
+
"eval_recall": 0.9977997799779978,
|
| 586 |
+
"eval_runtime": 4.416,
|
| 587 |
+
"eval_samples_per_second": 411.687,
|
| 588 |
+
"eval_steps_per_second": 25.815,
|
| 589 |
+
"step": 2045
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.0024449877750612,
|
| 593 |
+
"grad_norm": 0.013088320381939411,
|
| 594 |
+
"learning_rate": 3.7031334903097264e-05,
|
| 595 |
+
"loss": 0.0341,
|
| 596 |
+
"step": 2050
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.0146699266503667,
|
| 600 |
+
"grad_norm": 0.012992248870432377,
|
| 601 |
+
"learning_rate": 3.680492664372396e-05,
|
| 602 |
+
"loss": 0.0311,
|
| 603 |
+
"step": 2075
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 1.0268948655256724,
|
| 607 |
+
"grad_norm": 0.02359613962471485,
|
| 608 |
+
"learning_rate": 3.6578518384350664e-05,
|
| 609 |
+
"loss": 0.0432,
|
| 610 |
+
"step": 2100
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.039119804400978,
|
| 614 |
+
"grad_norm": 0.00870076846331358,
|
| 615 |
+
"learning_rate": 3.635211012497736e-05,
|
| 616 |
+
"loss": 0.0205,
|
| 617 |
+
"step": 2125
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.0513447432762837,
|
| 621 |
+
"grad_norm": 0.00848635658621788,
|
| 622 |
+
"learning_rate": 3.612570186560406e-05,
|
| 623 |
+
"loss": 0.0327,
|
| 624 |
+
"step": 2150
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 1.0635696821515892,
|
| 628 |
+
"grad_norm": 0.01124533824622631,
|
| 629 |
+
"learning_rate": 3.5899293606230755e-05,
|
| 630 |
+
"loss": 0.0121,
|
| 631 |
+
"step": 2175
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.075794621026895,
|
| 635 |
+
"grad_norm": 0.005016707349568605,
|
| 636 |
+
"learning_rate": 3.567288534685745e-05,
|
| 637 |
+
"loss": 0.0334,
|
| 638 |
+
"step": 2200
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.0880195599022005,
|
| 642 |
+
"grad_norm": 0.028895337134599686,
|
| 643 |
+
"learning_rate": 3.5446477087484155e-05,
|
| 644 |
+
"loss": 0.0095,
|
| 645 |
+
"step": 2225
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 1.1002444987775062,
|
| 649 |
+
"grad_norm": 0.009108040481805801,
|
| 650 |
+
"learning_rate": 3.522006882811085e-05,
|
| 651 |
+
"loss": 0.094,
|
| 652 |
+
"step": 2250
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 1.1124694376528117,
|
| 656 |
+
"grad_norm": 0.023064499720931053,
|
| 657 |
+
"learning_rate": 3.499366056873755e-05,
|
| 658 |
+
"loss": 0.0244,
|
| 659 |
+
"step": 2275
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 1.1246943765281174,
|
| 663 |
+
"grad_norm": 0.013536947779357433,
|
| 664 |
+
"learning_rate": 3.4767252309364246e-05,
|
| 665 |
+
"loss": 0.0353,
|
| 666 |
+
"step": 2300
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 1.136919315403423,
|
| 670 |
+
"grad_norm": 0.007536923512816429,
|
| 671 |
+
"learning_rate": 3.454084404999094e-05,
|
| 672 |
+
"loss": 0.0347,
|
| 673 |
+
"step": 2325
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.1491442542787287,
|
| 677 |
+
"grad_norm": 0.003116949461400509,
|
| 678 |
+
"learning_rate": 3.4314435790617646e-05,
|
| 679 |
+
"loss": 0.0006,
|
| 680 |
+
"step": 2350
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.1613691931540342,
|
| 684 |
+
"grad_norm": 0.009431233629584312,
|
| 685 |
+
"learning_rate": 3.408802753124434e-05,
|
| 686 |
+
"loss": 0.0403,
|
| 687 |
+
"step": 2375
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 1.17359413202934,
|
| 691 |
+
"grad_norm": 0.007902787998318672,
|
| 692 |
+
"learning_rate": 3.386161927187104e-05,
|
| 693 |
+
"loss": 0.0002,
|
| 694 |
+
"step": 2400
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 1.1858190709046454,
|
| 698 |
+
"grad_norm": 0.010425157845020294,
|
| 699 |
+
"learning_rate": 3.363521101249774e-05,
|
| 700 |
+
"loss": 0.044,
|
| 701 |
+
"step": 2425
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 1.1980440097799512,
|
| 705 |
+
"grad_norm": 0.0714653730392456,
|
| 706 |
+
"learning_rate": 3.3408802753124434e-05,
|
| 707 |
+
"loss": 0.0943,
|
| 708 |
+
"step": 2450
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 1.2102689486552567,
|
| 712 |
+
"grad_norm": 0.028381148353219032,
|
| 713 |
+
"learning_rate": 3.318239449375114e-05,
|
| 714 |
+
"loss": 0.0014,
|
| 715 |
+
"step": 2475
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.2224938875305624,
|
| 719 |
+
"grad_norm": 0.007024334277957678,
|
| 720 |
+
"learning_rate": 3.2955986234377834e-05,
|
| 721 |
+
"loss": 0.0005,
|
| 722 |
+
"step": 2500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.234718826405868,
|
| 726 |
+
"grad_norm": 0.0050694290548563,
|
| 727 |
+
"learning_rate": 3.2729577975004524e-05,
|
| 728 |
+
"loss": 0.0003,
|
| 729 |
+
"step": 2525
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 1.2469437652811737,
|
| 733 |
+
"grad_norm": 53.81669235229492,
|
| 734 |
+
"learning_rate": 3.250316971563123e-05,
|
| 735 |
+
"loss": 0.0597,
|
| 736 |
+
"step": 2550
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 1.2591687041564792,
|
| 740 |
+
"grad_norm": 0.003106011776253581,
|
| 741 |
+
"learning_rate": 3.2276761456257925e-05,
|
| 742 |
+
"loss": 0.0002,
|
| 743 |
+
"step": 2575
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 1.271393643031785,
|
| 747 |
+
"grad_norm": 0.007865176536142826,
|
| 748 |
+
"learning_rate": 3.205035319688462e-05,
|
| 749 |
+
"loss": 0.0374,
|
| 750 |
+
"step": 2600
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 1.2836185819070904,
|
| 754 |
+
"grad_norm": 0.004462777636945248,
|
| 755 |
+
"learning_rate": 3.1823944937511325e-05,
|
| 756 |
+
"loss": 0.0046,
|
| 757 |
+
"step": 2625
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.295843520782396,
|
| 761 |
+
"grad_norm": 0.0036718258634209633,
|
| 762 |
+
"learning_rate": 3.1597536678138015e-05,
|
| 763 |
+
"loss": 0.013,
|
| 764 |
+
"step": 2650
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.3080684596577017,
|
| 768 |
+
"grad_norm": 0.004185810219496489,
|
| 769 |
+
"learning_rate": 3.137112841876472e-05,
|
| 770 |
+
"loss": 0.0001,
|
| 771 |
+
"step": 2675
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 1.3202933985330074,
|
| 775 |
+
"grad_norm": 0.003247720655053854,
|
| 776 |
+
"learning_rate": 3.1144720159391416e-05,
|
| 777 |
+
"loss": 0.0343,
|
| 778 |
+
"step": 2700
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.332518337408313,
|
| 782 |
+
"grad_norm": 0.0022379523143172264,
|
| 783 |
+
"learning_rate": 3.091831190001811e-05,
|
| 784 |
+
"loss": 0.045,
|
| 785 |
+
"step": 2725
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 1.3447432762836184,
|
| 789 |
+
"grad_norm": 0.008663604035973549,
|
| 790 |
+
"learning_rate": 3.0691903640644816e-05,
|
| 791 |
+
"loss": 0.0002,
|
| 792 |
+
"step": 2750
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 1.3569682151589242,
|
| 796 |
+
"grad_norm": 0.002595770638436079,
|
| 797 |
+
"learning_rate": 3.046549538127151e-05,
|
| 798 |
+
"loss": 0.0309,
|
| 799 |
+
"step": 2775
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.36919315403423,
|
| 803 |
+
"grad_norm": 16.00531578063965,
|
| 804 |
+
"learning_rate": 3.0239087121898207e-05,
|
| 805 |
+
"loss": 0.094,
|
| 806 |
+
"step": 2800
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.3814180929095354,
|
| 810 |
+
"grad_norm": 0.0017313804710283875,
|
| 811 |
+
"learning_rate": 3.0012678862524907e-05,
|
| 812 |
+
"loss": 0.0006,
|
| 813 |
+
"step": 2825
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 1.393643031784841,
|
| 817 |
+
"grad_norm": 0.004820260684937239,
|
| 818 |
+
"learning_rate": 2.9795326933526534e-05,
|
| 819 |
+
"loss": 0.0875,
|
| 820 |
+
"step": 2850
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.4058679706601467,
|
| 824 |
+
"grad_norm": 0.005086742341518402,
|
| 825 |
+
"learning_rate": 2.9568918674153235e-05,
|
| 826 |
+
"loss": 0.0002,
|
| 827 |
+
"step": 2875
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 1.4180929095354524,
|
| 831 |
+
"grad_norm": 0.0022867824882268906,
|
| 832 |
+
"learning_rate": 2.9342510414779935e-05,
|
| 833 |
+
"loss": 0.0488,
|
| 834 |
+
"step": 2900
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 1.430317848410758,
|
| 838 |
+
"grad_norm": 0.0028263565618544817,
|
| 839 |
+
"learning_rate": 2.9116102155406632e-05,
|
| 840 |
+
"loss": 0.0184,
|
| 841 |
+
"step": 2925
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.4425427872860634,
|
| 845 |
+
"grad_norm": 0.006596155930310488,
|
| 846 |
+
"learning_rate": 2.8889693896033325e-05,
|
| 847 |
+
"loss": 0.0091,
|
| 848 |
+
"step": 2950
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.4547677261613692,
|
| 852 |
+
"grad_norm": 0.008638501167297363,
|
| 853 |
+
"learning_rate": 2.8663285636660025e-05,
|
| 854 |
+
"loss": 0.0248,
|
| 855 |
+
"step": 2975
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.466992665036675,
|
| 859 |
+
"grad_norm": 0.009834705851972103,
|
| 860 |
+
"learning_rate": 2.8436877377286726e-05,
|
| 861 |
+
"loss": 0.0019,
|
| 862 |
+
"step": 3000
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.4792176039119804,
|
| 866 |
+
"grad_norm": 0.003901288378983736,
|
| 867 |
+
"learning_rate": 2.8210469117913423e-05,
|
| 868 |
+
"loss": 0.0785,
|
| 869 |
+
"step": 3025
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 1.491442542787286,
|
| 873 |
+
"grad_norm": 0.002671342808753252,
|
| 874 |
+
"learning_rate": 2.7984060858540123e-05,
|
| 875 |
+
"loss": 0.0003,
|
| 876 |
+
"step": 3050
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.5036674816625917,
|
| 880 |
+
"grad_norm": 0.010172242298722267,
|
| 881 |
+
"learning_rate": 2.7757652599166816e-05,
|
| 882 |
+
"loss": 0.0001,
|
| 883 |
+
"step": 3075
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.5158924205378974,
|
| 887 |
+
"grad_norm": 0.0012791448971256614,
|
| 888 |
+
"learning_rate": 2.7531244339793517e-05,
|
| 889 |
+
"loss": 0.0002,
|
| 890 |
+
"step": 3100
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.528117359413203,
|
| 894 |
+
"grad_norm": 0.019383808597922325,
|
| 895 |
+
"learning_rate": 2.7304836080420217e-05,
|
| 896 |
+
"loss": 0.048,
|
| 897 |
+
"step": 3125
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.5403422982885084,
|
| 901 |
+
"grad_norm": 4.787984848022461,
|
| 902 |
+
"learning_rate": 2.7078427821046914e-05,
|
| 903 |
+
"loss": 0.0008,
|
| 904 |
+
"step": 3150
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.5525672371638142,
|
| 908 |
+
"grad_norm": 0.005636611022055149,
|
| 909 |
+
"learning_rate": 2.6852019561673614e-05,
|
| 910 |
+
"loss": 0.0226,
|
| 911 |
+
"step": 3175
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.56479217603912,
|
| 915 |
+
"grad_norm": 0.004684858489781618,
|
| 916 |
+
"learning_rate": 2.6625611302300307e-05,
|
| 917 |
+
"loss": 0.0002,
|
| 918 |
+
"step": 3200
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.5770171149144254,
|
| 922 |
+
"grad_norm": 0.008136129938066006,
|
| 923 |
+
"learning_rate": 2.6399203042927008e-05,
|
| 924 |
+
"loss": 0.0509,
|
| 925 |
+
"step": 3225
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.589242053789731,
|
| 929 |
+
"grad_norm": 0.011331688612699509,
|
| 930 |
+
"learning_rate": 2.6172794783553704e-05,
|
| 931 |
+
"loss": 0.0744,
|
| 932 |
+
"step": 3250
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.6014669926650367,
|
| 936 |
+
"grad_norm": 0.015354806557297707,
|
| 937 |
+
"learning_rate": 2.5946386524180405e-05,
|
| 938 |
+
"loss": 0.0221,
|
| 939 |
+
"step": 3275
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.6136919315403424,
|
| 943 |
+
"grad_norm": 0.021820085123181343,
|
| 944 |
+
"learning_rate": 2.5719978264807105e-05,
|
| 945 |
+
"loss": 0.0006,
|
| 946 |
+
"step": 3300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.625916870415648,
|
| 950 |
+
"grad_norm": 0.0048668403178453445,
|
| 951 |
+
"learning_rate": 2.5493570005433798e-05,
|
| 952 |
+
"loss": 0.0006,
|
| 953 |
+
"step": 3325
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.6381418092909534,
|
| 957 |
+
"grad_norm": 0.004569260869175196,
|
| 958 |
+
"learning_rate": 2.52671617460605e-05,
|
| 959 |
+
"loss": 0.0784,
|
| 960 |
+
"step": 3350
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.6503667481662592,
|
| 964 |
+
"grad_norm": 0.010390234179794788,
|
| 965 |
+
"learning_rate": 2.5040753486687195e-05,
|
| 966 |
+
"loss": 0.0005,
|
| 967 |
+
"step": 3375
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.662591687041565,
|
| 971 |
+
"grad_norm": 0.0035980944521725178,
|
| 972 |
+
"learning_rate": 2.4814345227313892e-05,
|
| 973 |
+
"loss": 0.0002,
|
| 974 |
+
"step": 3400
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.6748166259168704,
|
| 978 |
+
"grad_norm": 0.01261225901544094,
|
| 979 |
+
"learning_rate": 2.4587936967940592e-05,
|
| 980 |
+
"loss": 0.0002,
|
| 981 |
+
"step": 3425
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.687041564792176,
|
| 985 |
+
"grad_norm": 0.014144369401037693,
|
| 986 |
+
"learning_rate": 2.4361528708567293e-05,
|
| 987 |
+
"loss": 0.0753,
|
| 988 |
+
"step": 3450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 1.6992665036674817,
|
| 992 |
+
"grad_norm": 0.05743572860956192,
|
| 993 |
+
"learning_rate": 2.4135120449193986e-05,
|
| 994 |
+
"loss": 0.0005,
|
| 995 |
+
"step": 3475
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.7114914425427874,
|
| 999 |
+
"grad_norm": 0.011294255964457989,
|
| 1000 |
+
"learning_rate": 2.3908712189820686e-05,
|
| 1001 |
+
"loss": 0.0002,
|
| 1002 |
+
"step": 3500
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.723716381418093,
|
| 1006 |
+
"grad_norm": 0.014550820924341679,
|
| 1007 |
+
"learning_rate": 2.3682303930447383e-05,
|
| 1008 |
+
"loss": 0.0309,
|
| 1009 |
+
"step": 3525
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.7359413202933984,
|
| 1013 |
+
"grad_norm": 0.002219605026766658,
|
| 1014 |
+
"learning_rate": 2.3455895671074083e-05,
|
| 1015 |
+
"loss": 0.0002,
|
| 1016 |
+
"step": 3550
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.7481662591687042,
|
| 1020 |
+
"grad_norm": 0.0027680303901433945,
|
| 1021 |
+
"learning_rate": 2.322948741170078e-05,
|
| 1022 |
+
"loss": 0.0168,
|
| 1023 |
+
"step": 3575
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.76039119804401,
|
| 1027 |
+
"grad_norm": 0.0014964850852265954,
|
| 1028 |
+
"learning_rate": 2.3003079152327477e-05,
|
| 1029 |
+
"loss": 0.0002,
|
| 1030 |
+
"step": 3600
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.7726161369193154,
|
| 1034 |
+
"grad_norm": 0.007225463632494211,
|
| 1035 |
+
"learning_rate": 2.2776670892954177e-05,
|
| 1036 |
+
"loss": 0.0756,
|
| 1037 |
+
"step": 3625
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.784841075794621,
|
| 1041 |
+
"grad_norm": 0.016129551455378532,
|
| 1042 |
+
"learning_rate": 2.2550262633580874e-05,
|
| 1043 |
+
"loss": 0.0003,
|
| 1044 |
+
"step": 3650
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.7970660146699267,
|
| 1048 |
+
"grad_norm": 0.004128915257751942,
|
| 1049 |
+
"learning_rate": 2.2323854374207574e-05,
|
| 1050 |
+
"loss": 0.0032,
|
| 1051 |
+
"step": 3675
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.8092909535452324,
|
| 1055 |
+
"grad_norm": 0.002796179847791791,
|
| 1056 |
+
"learning_rate": 2.2097446114834268e-05,
|
| 1057 |
+
"loss": 0.0071,
|
| 1058 |
+
"step": 3700
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.821515892420538,
|
| 1062 |
+
"grad_norm": 0.03467123582959175,
|
| 1063 |
+
"learning_rate": 2.1871037855460968e-05,
|
| 1064 |
+
"loss": 0.0325,
|
| 1065 |
+
"step": 3725
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.8337408312958434,
|
| 1069 |
+
"grad_norm": 0.0022633096668869257,
|
| 1070 |
+
"learning_rate": 2.1644629596087665e-05,
|
| 1071 |
+
"loss": 0.0398,
|
| 1072 |
+
"step": 3750
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.8459657701711492,
|
| 1076 |
+
"grad_norm": 0.012029076926410198,
|
| 1077 |
+
"learning_rate": 2.1418221336714365e-05,
|
| 1078 |
+
"loss": 0.0002,
|
| 1079 |
+
"step": 3775
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.858190709046455,
|
| 1083 |
+
"grad_norm": 0.010079102590680122,
|
| 1084 |
+
"learning_rate": 2.1191813077341062e-05,
|
| 1085 |
+
"loss": 0.0089,
|
| 1086 |
+
"step": 3800
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.8704156479217604,
|
| 1090 |
+
"grad_norm": 0.002109576016664505,
|
| 1091 |
+
"learning_rate": 2.096540481796776e-05,
|
| 1092 |
+
"loss": 0.0101,
|
| 1093 |
+
"step": 3825
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.882640586797066,
|
| 1097 |
+
"grad_norm": 0.0009537243167869747,
|
| 1098 |
+
"learning_rate": 2.073899655859446e-05,
|
| 1099 |
+
"loss": 0.026,
|
| 1100 |
+
"step": 3850
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.8948655256723717,
|
| 1104 |
+
"grad_norm": 0.003006375627592206,
|
| 1105 |
+
"learning_rate": 2.0512588299221156e-05,
|
| 1106 |
+
"loss": 0.0001,
|
| 1107 |
+
"step": 3875
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.9070904645476774,
|
| 1111 |
+
"grad_norm": 103.76818084716797,
|
| 1112 |
+
"learning_rate": 2.0286180039847856e-05,
|
| 1113 |
+
"loss": 0.014,
|
| 1114 |
+
"step": 3900
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.919315403422983,
|
| 1118 |
+
"grad_norm": 0.01744658872485161,
|
| 1119 |
+
"learning_rate": 2.0059771780474553e-05,
|
| 1120 |
+
"loss": 0.0001,
|
| 1121 |
+
"step": 3925
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.9315403422982884,
|
| 1125 |
+
"grad_norm": 0.0038125617429614067,
|
| 1126 |
+
"learning_rate": 1.983336352110125e-05,
|
| 1127 |
+
"loss": 0.0058,
|
| 1128 |
+
"step": 3950
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.9437652811735942,
|
| 1132 |
+
"grad_norm": 0.0009853680385276675,
|
| 1133 |
+
"learning_rate": 1.960695526172795e-05,
|
| 1134 |
+
"loss": 0.01,
|
| 1135 |
+
"step": 3975
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.9559902200488999,
|
| 1139 |
+
"grad_norm": 0.0039232331328094006,
|
| 1140 |
+
"learning_rate": 1.9380547002354647e-05,
|
| 1141 |
+
"loss": 0.0092,
|
| 1142 |
+
"step": 4000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.9682151589242054,
|
| 1146 |
+
"grad_norm": 0.004057803191244602,
|
| 1147 |
+
"learning_rate": 1.9154138742981344e-05,
|
| 1148 |
+
"loss": 0.0315,
|
| 1149 |
+
"step": 4025
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.980440097799511,
|
| 1153 |
+
"grad_norm": 0.010765478946268559,
|
| 1154 |
+
"learning_rate": 1.892773048360804e-05,
|
| 1155 |
+
"loss": 0.0575,
|
| 1156 |
+
"step": 4050
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.9926650366748166,
|
| 1160 |
+
"grad_norm": 0.0014557191170752048,
|
| 1161 |
+
"learning_rate": 1.870132222423474e-05,
|
| 1162 |
+
"loss": 0.0002,
|
| 1163 |
+
"step": 4075
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 2.0,
|
| 1167 |
+
"eval_accuracy": 0.9911991199119912,
|
| 1168 |
+
"eval_auc": 0.9996677885610342,
|
| 1169 |
+
"eval_f1": 0.9912472647702407,
|
| 1170 |
+
"eval_loss": 0.041126370429992676,
|
| 1171 |
+
"eval_precision": 0.985854189336235,
|
| 1172 |
+
"eval_recall": 0.9966996699669967,
|
| 1173 |
+
"eval_runtime": 4.3541,
|
| 1174 |
+
"eval_samples_per_second": 417.541,
|
| 1175 |
+
"eval_steps_per_second": 26.182,
|
| 1176 |
+
"step": 4090
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 2.0048899755501224,
|
| 1180 |
+
"grad_norm": 0.014019358903169632,
|
| 1181 |
+
"learning_rate": 1.847491396486144e-05,
|
| 1182 |
+
"loss": 0.0639,
|
| 1183 |
+
"step": 4100
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 2.0171149144254277,
|
| 1187 |
+
"grad_norm": 0.002681875368580222,
|
| 1188 |
+
"learning_rate": 1.8248505705488138e-05,
|
| 1189 |
+
"loss": 0.0002,
|
| 1190 |
+
"step": 4125
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.0293398533007334,
|
| 1194 |
+
"grad_norm": 0.004261409398168325,
|
| 1195 |
+
"learning_rate": 1.8022097446114835e-05,
|
| 1196 |
+
"loss": 0.0003,
|
| 1197 |
+
"step": 4150
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 2.041564792176039,
|
| 1201 |
+
"grad_norm": 0.0013389644445851445,
|
| 1202 |
+
"learning_rate": 1.7795689186741532e-05,
|
| 1203 |
+
"loss": 0.0001,
|
| 1204 |
+
"step": 4175
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 2.053789731051345,
|
| 1208 |
+
"grad_norm": 0.013343879021704197,
|
| 1209 |
+
"learning_rate": 1.7569280927368232e-05,
|
| 1210 |
+
"loss": 0.0131,
|
| 1211 |
+
"step": 4200
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 2.06601466992665,
|
| 1215 |
+
"grad_norm": 0.0007498391787521541,
|
| 1216 |
+
"learning_rate": 1.7342872667994932e-05,
|
| 1217 |
+
"loss": 0.0001,
|
| 1218 |
+
"step": 4225
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 2.078239608801956,
|
| 1222 |
+
"grad_norm": 0.0014528041938319802,
|
| 1223 |
+
"learning_rate": 1.7116464408621626e-05,
|
| 1224 |
+
"loss": 0.0113,
|
| 1225 |
+
"step": 4250
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 2.0904645476772616,
|
| 1229 |
+
"grad_norm": 0.0008523334981873631,
|
| 1230 |
+
"learning_rate": 1.6890056149248326e-05,
|
| 1231 |
+
"loss": 0.0001,
|
| 1232 |
+
"step": 4275
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.1026894865525674,
|
| 1236 |
+
"grad_norm": 0.003291267203167081,
|
| 1237 |
+
"learning_rate": 1.6663647889875023e-05,
|
| 1238 |
+
"loss": 0.0,
|
| 1239 |
+
"step": 4300
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 2.1149144254278727,
|
| 1243 |
+
"grad_norm": 0.000696261296980083,
|
| 1244 |
+
"learning_rate": 1.6437239630501723e-05,
|
| 1245 |
+
"loss": 0.0001,
|
| 1246 |
+
"step": 4325
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 2.1271393643031784,
|
| 1250 |
+
"grad_norm": 0.0020809646230190992,
|
| 1251 |
+
"learning_rate": 1.621083137112842e-05,
|
| 1252 |
+
"loss": 0.0059,
|
| 1253 |
+
"step": 4350
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 2.139364303178484,
|
| 1257 |
+
"grad_norm": 0.000507664808537811,
|
| 1258 |
+
"learning_rate": 1.5984423111755117e-05,
|
| 1259 |
+
"loss": 0.0001,
|
| 1260 |
+
"step": 4375
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 2.15158924205379,
|
| 1264 |
+
"grad_norm": 0.0004800551978405565,
|
| 1265 |
+
"learning_rate": 1.5758014852381817e-05,
|
| 1266 |
+
"loss": 0.013,
|
| 1267 |
+
"step": 4400
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 2.163814180929095,
|
| 1271 |
+
"grad_norm": 0.0005327569670043886,
|
| 1272 |
+
"learning_rate": 1.5531606593008514e-05,
|
| 1273 |
+
"loss": 0.0428,
|
| 1274 |
+
"step": 4425
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.176039119804401,
|
| 1278 |
+
"grad_norm": 0.0005905398284085095,
|
| 1279 |
+
"learning_rate": 1.5305198333635214e-05,
|
| 1280 |
+
"loss": 0.0,
|
| 1281 |
+
"step": 4450
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 2.1882640586797066,
|
| 1285 |
+
"grad_norm": 0.0005032393964938819,
|
| 1286 |
+
"learning_rate": 1.5078790074261909e-05,
|
| 1287 |
+
"loss": 0.0,
|
| 1288 |
+
"step": 4475
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 2.2004889975550124,
|
| 1292 |
+
"grad_norm": 0.0011241784086450934,
|
| 1293 |
+
"learning_rate": 1.4852381814888608e-05,
|
| 1294 |
+
"loss": 0.0277,
|
| 1295 |
+
"step": 4500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 2.2127139364303177,
|
| 1299 |
+
"grad_norm": 0.0005586715415120125,
|
| 1300 |
+
"learning_rate": 1.4625973555515308e-05,
|
| 1301 |
+
"loss": 0.0001,
|
| 1302 |
+
"step": 4525
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 2.2249388753056234,
|
| 1306 |
+
"grad_norm": 0.0011677349684759974,
|
| 1307 |
+
"learning_rate": 1.4399565296142003e-05,
|
| 1308 |
+
"loss": 0.0021,
|
| 1309 |
+
"step": 4550
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 2.237163814180929,
|
| 1313 |
+
"grad_norm": 0.0006688361754640937,
|
| 1314 |
+
"learning_rate": 1.4173157036768703e-05,
|
| 1315 |
+
"loss": 0.0001,
|
| 1316 |
+
"step": 4575
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 2.249388753056235,
|
| 1320 |
+
"grad_norm": 0.0008152456721290946,
|
| 1321 |
+
"learning_rate": 1.3946748777395398e-05,
|
| 1322 |
+
"loss": 0.0286,
|
| 1323 |
+
"step": 4600
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 2.26161369193154,
|
| 1327 |
+
"grad_norm": 0.001650724676437676,
|
| 1328 |
+
"learning_rate": 1.3720340518022099e-05,
|
| 1329 |
+
"loss": 0.0,
|
| 1330 |
+
"step": 4625
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 2.273838630806846,
|
| 1334 |
+
"grad_norm": 0.0006924684857949615,
|
| 1335 |
+
"learning_rate": 1.3493932258648795e-05,
|
| 1336 |
+
"loss": 0.0,
|
| 1337 |
+
"step": 4650
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 2.2860635696821516,
|
| 1341 |
+
"grad_norm": 0.0366230309009552,
|
| 1342 |
+
"learning_rate": 1.3267523999275494e-05,
|
| 1343 |
+
"loss": 0.0,
|
| 1344 |
+
"step": 4675
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 2.2982885085574574,
|
| 1348 |
+
"grad_norm": 0.0006909920484758914,
|
| 1349 |
+
"learning_rate": 1.3041115739902193e-05,
|
| 1350 |
+
"loss": 0.0185,
|
| 1351 |
+
"step": 4700
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 2.3105134474327627,
|
| 1355 |
+
"grad_norm": 0.0007503129891119897,
|
| 1356 |
+
"learning_rate": 1.281470748052889e-05,
|
| 1357 |
+
"loss": 0.0001,
|
| 1358 |
+
"step": 4725
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 2.3227383863080684,
|
| 1362 |
+
"grad_norm": 0.0023324170615524054,
|
| 1363 |
+
"learning_rate": 1.258829922115559e-05,
|
| 1364 |
+
"loss": 0.0314,
|
| 1365 |
+
"step": 4750
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 2.334963325183374,
|
| 1369 |
+
"grad_norm": 0.0006564876530319452,
|
| 1370 |
+
"learning_rate": 1.2361890961782286e-05,
|
| 1371 |
+
"loss": 0.0001,
|
| 1372 |
+
"step": 4775
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 2.34718826405868,
|
| 1376 |
+
"grad_norm": 0.000901265419088304,
|
| 1377 |
+
"learning_rate": 1.2135482702408985e-05,
|
| 1378 |
+
"loss": 0.0001,
|
| 1379 |
+
"step": 4800
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 2.359413202933985,
|
| 1383 |
+
"grad_norm": 0.0041041006334125996,
|
| 1384 |
+
"learning_rate": 1.1909074443035682e-05,
|
| 1385 |
+
"loss": 0.0,
|
| 1386 |
+
"step": 4825
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 2.371638141809291,
|
| 1390 |
+
"grad_norm": 0.00100336829200387,
|
| 1391 |
+
"learning_rate": 1.168266618366238e-05,
|
| 1392 |
+
"loss": 0.0001,
|
| 1393 |
+
"step": 4850
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 2.3838630806845966,
|
| 1397 |
+
"grad_norm": 0.0007710576755926013,
|
| 1398 |
+
"learning_rate": 1.1456257924289077e-05,
|
| 1399 |
+
"loss": 0.0264,
|
| 1400 |
+
"step": 4875
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 2.3960880195599024,
|
| 1404 |
+
"grad_norm": 0.0004823113267775625,
|
| 1405 |
+
"learning_rate": 1.1229849664915777e-05,
|
| 1406 |
+
"loss": 0.0001,
|
| 1407 |
+
"step": 4900
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 2.4083129584352077,
|
| 1411 |
+
"grad_norm": 0.0010028666583821177,
|
| 1412 |
+
"learning_rate": 1.1003441405542474e-05,
|
| 1413 |
+
"loss": 0.0,
|
| 1414 |
+
"step": 4925
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 2.4205378973105134,
|
| 1418 |
+
"grad_norm": 0.013664372265338898,
|
| 1419 |
+
"learning_rate": 1.0777033146169173e-05,
|
| 1420 |
+
"loss": 0.0001,
|
| 1421 |
+
"step": 4950
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 2.432762836185819,
|
| 1425 |
+
"grad_norm": 0.003788791596889496,
|
| 1426 |
+
"learning_rate": 1.0550624886795871e-05,
|
| 1427 |
+
"loss": 0.0396,
|
| 1428 |
+
"step": 4975
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 2.444987775061125,
|
| 1432 |
+
"grad_norm": 0.0058593484573066235,
|
| 1433 |
+
"learning_rate": 1.0324216627422568e-05,
|
| 1434 |
+
"loss": 0.0176,
|
| 1435 |
+
"step": 5000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 2.45721271393643,
|
| 1439 |
+
"grad_norm": 0.007754920981824398,
|
| 1440 |
+
"learning_rate": 1.0097808368049267e-05,
|
| 1441 |
+
"loss": 0.0,
|
| 1442 |
+
"step": 5025
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 2.469437652811736,
|
| 1446 |
+
"grad_norm": 0.013962284661829472,
|
| 1447 |
+
"learning_rate": 9.871400108675965e-06,
|
| 1448 |
+
"loss": 0.0001,
|
| 1449 |
+
"step": 5050
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 2.4816625916870416,
|
| 1453 |
+
"grad_norm": 0.0006092582480050623,
|
| 1454 |
+
"learning_rate": 9.644991849302664e-06,
|
| 1455 |
+
"loss": 0.0,
|
| 1456 |
+
"step": 5075
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 2.4938875305623474,
|
| 1460 |
+
"grad_norm": 0.0022751325741410255,
|
| 1461 |
+
"learning_rate": 9.41858358992936e-06,
|
| 1462 |
+
"loss": 0.0002,
|
| 1463 |
+
"step": 5100
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.5061124694376526,
|
| 1467 |
+
"grad_norm": 0.018367614597082138,
|
| 1468 |
+
"learning_rate": 9.19217533055606e-06,
|
| 1469 |
+
"loss": 0.0,
|
| 1470 |
+
"step": 5125
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 2.5183374083129584,
|
| 1474 |
+
"grad_norm": 0.0010836137225851417,
|
| 1475 |
+
"learning_rate": 8.965767071182756e-06,
|
| 1476 |
+
"loss": 0.0001,
|
| 1477 |
+
"step": 5150
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 2.530562347188264,
|
| 1481 |
+
"grad_norm": 0.0121123306453228,
|
| 1482 |
+
"learning_rate": 8.739358811809455e-06,
|
| 1483 |
+
"loss": 0.0,
|
| 1484 |
+
"step": 5175
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 2.54278728606357,
|
| 1488 |
+
"grad_norm": 0.0004910171264782548,
|
| 1489 |
+
"learning_rate": 8.512950552436153e-06,
|
| 1490 |
+
"loss": 0.0,
|
| 1491 |
+
"step": 5200
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 2.555012224938875,
|
| 1495 |
+
"grad_norm": 0.0020201788283884525,
|
| 1496 |
+
"learning_rate": 8.286542293062852e-06,
|
| 1497 |
+
"loss": 0.0,
|
| 1498 |
+
"step": 5225
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 2.567237163814181,
|
| 1502 |
+
"grad_norm": 0.0015963187906891108,
|
| 1503 |
+
"learning_rate": 8.06013403368955e-06,
|
| 1504 |
+
"loss": 0.0,
|
| 1505 |
+
"step": 5250
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 2.5794621026894866,
|
| 1509 |
+
"grad_norm": 0.0004962123930454254,
|
| 1510 |
+
"learning_rate": 7.833725774316247e-06,
|
| 1511 |
+
"loss": 0.0014,
|
| 1512 |
+
"step": 5275
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 2.591687041564792,
|
| 1516 |
+
"grad_norm": 0.003313443623483181,
|
| 1517 |
+
"learning_rate": 7.607317514942945e-06,
|
| 1518 |
+
"loss": 0.0,
|
| 1519 |
+
"step": 5300
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 2.6039119804400976,
|
| 1523 |
+
"grad_norm": 0.0003791134222410619,
|
| 1524 |
+
"learning_rate": 7.380909255569643e-06,
|
| 1525 |
+
"loss": 0.003,
|
| 1526 |
+
"step": 5325
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 2.6161369193154034,
|
| 1530 |
+
"grad_norm": 0.0009152629063464701,
|
| 1531 |
+
"learning_rate": 7.154500996196342e-06,
|
| 1532 |
+
"loss": 0.0001,
|
| 1533 |
+
"step": 5350
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 2.628361858190709,
|
| 1537 |
+
"grad_norm": 0.0006914679543115199,
|
| 1538 |
+
"learning_rate": 6.92809273682304e-06,
|
| 1539 |
+
"loss": 0.0001,
|
| 1540 |
+
"step": 5375
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 2.640586797066015,
|
| 1544 |
+
"grad_norm": 0.03224007785320282,
|
| 1545 |
+
"learning_rate": 6.701684477449738e-06,
|
| 1546 |
+
"loss": 0.0,
|
| 1547 |
+
"step": 5400
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.65281173594132,
|
| 1551 |
+
"grad_norm": 0.0006668714340776205,
|
| 1552 |
+
"learning_rate": 6.475276218076436e-06,
|
| 1553 |
+
"loss": 0.0,
|
| 1554 |
+
"step": 5425
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 2.665036674816626,
|
| 1558 |
+
"grad_norm": 0.0003864137106575072,
|
| 1559 |
+
"learning_rate": 6.2488679587031335e-06,
|
| 1560 |
+
"loss": 0.0,
|
| 1561 |
+
"step": 5450
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 2.6772616136919316,
|
| 1565 |
+
"grad_norm": 0.0004962776438333094,
|
| 1566 |
+
"learning_rate": 6.022459699329832e-06,
|
| 1567 |
+
"loss": 0.0002,
|
| 1568 |
+
"step": 5475
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 2.689486552567237,
|
| 1572 |
+
"grad_norm": 0.008872649632394314,
|
| 1573 |
+
"learning_rate": 5.79605143995653e-06,
|
| 1574 |
+
"loss": 0.013,
|
| 1575 |
+
"step": 5500
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 2.7017114914425426,
|
| 1579 |
+
"grad_norm": 0.00046311679761856794,
|
| 1580 |
+
"learning_rate": 5.5696431805832274e-06,
|
| 1581 |
+
"loss": 0.0,
|
| 1582 |
+
"step": 5525
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 2.7139364303178484,
|
| 1586 |
+
"grad_norm": 0.00033471471397206187,
|
| 1587 |
+
"learning_rate": 5.343234921209926e-06,
|
| 1588 |
+
"loss": 0.0009,
|
| 1589 |
+
"step": 5550
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 2.726161369193154,
|
| 1593 |
+
"grad_norm": 0.0003606221580412239,
|
| 1594 |
+
"learning_rate": 5.116826661836624e-06,
|
| 1595 |
+
"loss": 0.0,
|
| 1596 |
+
"step": 5575
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 2.73838630806846,
|
| 1600 |
+
"grad_norm": 0.04737100005149841,
|
| 1601 |
+
"learning_rate": 4.890418402463321e-06,
|
| 1602 |
+
"loss": 0.0,
|
| 1603 |
+
"step": 5600
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 2.750611246943765,
|
| 1607 |
+
"grad_norm": 0.00031916104489937425,
|
| 1608 |
+
"learning_rate": 4.66401014309002e-06,
|
| 1609 |
+
"loss": 0.0188,
|
| 1610 |
+
"step": 5625
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 2.762836185819071,
|
| 1614 |
+
"grad_norm": 0.000310692674247548,
|
| 1615 |
+
"learning_rate": 4.4376018837167184e-06,
|
| 1616 |
+
"loss": 0.0184,
|
| 1617 |
+
"step": 5650
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 2.7750611246943766,
|
| 1621 |
+
"grad_norm": 0.00045588318607769907,
|
| 1622 |
+
"learning_rate": 4.211193624343416e-06,
|
| 1623 |
+
"loss": 0.0059,
|
| 1624 |
+
"step": 5675
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 2.787286063569682,
|
| 1628 |
+
"grad_norm": 0.0010751900263130665,
|
| 1629 |
+
"learning_rate": 3.984785364970115e-06,
|
| 1630 |
+
"loss": 0.0,
|
| 1631 |
+
"step": 5700
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 2.7995110024449876,
|
| 1635 |
+
"grad_norm": 0.0003197513287886977,
|
| 1636 |
+
"learning_rate": 3.7583771055968124e-06,
|
| 1637 |
+
"loss": 0.0102,
|
| 1638 |
+
"step": 5725
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 2.8117359413202934,
|
| 1642 |
+
"grad_norm": 0.0019770630169659853,
|
| 1643 |
+
"learning_rate": 3.53196884622351e-06,
|
| 1644 |
+
"loss": 0.0,
|
| 1645 |
+
"step": 5750
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 2.823960880195599,
|
| 1649 |
+
"grad_norm": 0.00035567875602282584,
|
| 1650 |
+
"learning_rate": 3.3055605868502086e-06,
|
| 1651 |
+
"loss": 0.0,
|
| 1652 |
+
"step": 5775
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 2.836185819070905,
|
| 1656 |
+
"grad_norm": 0.011646813713014126,
|
| 1657 |
+
"learning_rate": 3.0791523274769063e-06,
|
| 1658 |
+
"loss": 0.0012,
|
| 1659 |
+
"step": 5800
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 2.84841075794621,
|
| 1663 |
+
"grad_norm": 0.0006059286533854902,
|
| 1664 |
+
"learning_rate": 2.8527440681036044e-06,
|
| 1665 |
+
"loss": 0.0,
|
| 1666 |
+
"step": 5825
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 2.860635696821516,
|
| 1670 |
+
"grad_norm": 0.000731644919142127,
|
| 1671 |
+
"learning_rate": 2.626335808730303e-06,
|
| 1672 |
+
"loss": 0.0,
|
| 1673 |
+
"step": 5850
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 2.8728606356968216,
|
| 1677 |
+
"grad_norm": 0.00040226284181699157,
|
| 1678 |
+
"learning_rate": 2.408983879731933e-06,
|
| 1679 |
+
"loss": 0.0064,
|
| 1680 |
+
"step": 5875
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 2.885085574572127,
|
| 1684 |
+
"grad_norm": 0.0016484076622873545,
|
| 1685 |
+
"learning_rate": 2.1825756203586305e-06,
|
| 1686 |
+
"loss": 0.0,
|
| 1687 |
+
"step": 5900
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 2.8973105134474326,
|
| 1691 |
+
"grad_norm": 0.0003576139861252159,
|
| 1692 |
+
"learning_rate": 1.956167360985329e-06,
|
| 1693 |
+
"loss": 0.0,
|
| 1694 |
+
"step": 5925
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 2.9095354523227384,
|
| 1698 |
+
"grad_norm": 0.000344920001225546,
|
| 1699 |
+
"learning_rate": 1.729759101612027e-06,
|
| 1700 |
+
"loss": 0.0,
|
| 1701 |
+
"step": 5950
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 2.921760391198044,
|
| 1705 |
+
"grad_norm": 0.0004281815781723708,
|
| 1706 |
+
"learning_rate": 1.503350842238725e-06,
|
| 1707 |
+
"loss": 0.0,
|
| 1708 |
+
"step": 5975
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 2.93398533007335,
|
| 1712 |
+
"grad_norm": 0.0010791511740535498,
|
| 1713 |
+
"learning_rate": 1.276942582865423e-06,
|
| 1714 |
+
"loss": 0.0001,
|
| 1715 |
+
"step": 6000
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 2.946210268948655,
|
| 1719 |
+
"grad_norm": 0.007071612868458033,
|
| 1720 |
+
"learning_rate": 1.050534323492121e-06,
|
| 1721 |
+
"loss": 0.0,
|
| 1722 |
+
"step": 6025
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 2.958435207823961,
|
| 1726 |
+
"grad_norm": 0.0007760756998322904,
|
| 1727 |
+
"learning_rate": 8.241260641188192e-07,
|
| 1728 |
+
"loss": 0.0,
|
| 1729 |
+
"step": 6050
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 2.9706601466992666,
|
| 1733 |
+
"grad_norm": 0.00026461490779183805,
|
| 1734 |
+
"learning_rate": 5.977178047455171e-07,
|
| 1735 |
+
"loss": 0.0,
|
| 1736 |
+
"step": 6075
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 2.982885085574572,
|
| 1740 |
+
"grad_norm": 0.00028767192270606756,
|
| 1741 |
+
"learning_rate": 3.713095453722152e-07,
|
| 1742 |
+
"loss": 0.0,
|
| 1743 |
+
"step": 6100
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.9951100244498776,
|
| 1747 |
+
"grad_norm": 0.00045831213355995715,
|
| 1748 |
+
"learning_rate": 1.4490128599891324e-07,
|
| 1749 |
+
"loss": 0.0,
|
| 1750 |
+
"step": 6125
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 3.0,
|
| 1754 |
+
"eval_accuracy": 0.995049504950495,
|
| 1755 |
+
"eval_auc": 0.9997361672360855,
|
| 1756 |
+
"eval_f1": 0.9950522264980759,
|
| 1757 |
+
"eval_loss": 0.03548985347151756,
|
| 1758 |
+
"eval_precision": 0.9945054945054945,
|
| 1759 |
+
"eval_recall": 0.9955995599559956,
|
| 1760 |
+
"eval_runtime": 4.3475,
|
| 1761 |
+
"eval_samples_per_second": 418.174,
|
| 1762 |
+
"eval_steps_per_second": 26.222,
|
| 1763 |
+
"step": 6135
|
| 1764 |
+
}
|
| 1765 |
+
],
|
| 1766 |
+
"logging_steps": 25,
|
| 1767 |
+
"max_steps": 6135,
|
| 1768 |
+
"num_input_tokens_seen": 0,
|
| 1769 |
+
"num_train_epochs": 3,
|
| 1770 |
+
"save_steps": 500,
|
| 1771 |
+
"stateful_callbacks": {
|
| 1772 |
+
"EarlyStoppingCallback": {
|
| 1773 |
+
"args": {
|
| 1774 |
+
"early_stopping_patience": 5,
|
| 1775 |
+
"early_stopping_threshold": 0.01
|
| 1776 |
+
},
|
| 1777 |
+
"attributes": {
|
| 1778 |
+
"early_stopping_patience_counter": 1
|
| 1779 |
+
}
|
| 1780 |
+
},
|
| 1781 |
+
"TrainerControl": {
|
| 1782 |
+
"args": {
|
| 1783 |
+
"should_epoch_stop": false,
|
| 1784 |
+
"should_evaluate": false,
|
| 1785 |
+
"should_log": false,
|
| 1786 |
+
"should_save": true,
|
| 1787 |
+
"should_training_stop": true
|
| 1788 |
+
},
|
| 1789 |
+
"attributes": {}
|
| 1790 |
+
}
|
| 1791 |
+
},
|
| 1792 |
+
"total_flos": 3227977982684160.0,
|
| 1793 |
+
"train_batch_size": 8,
|
| 1794 |
+
"trial_name": null,
|
| 1795 |
+
"trial_params": null
|
| 1796 |
+
}
|
checkpoint-6135/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7cb80b880983dad90e2001410afc3b3aad5340c8f3467534951dc6ceb9eca388
|
| 3 |
+
size 5368
|
config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "thenlper/gte-base",
|
| 3 |
+
"_num_labels": 2,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "human",
|
| 15 |
+
"1": "llm"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 3072,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"human": 0,
|
| 21 |
+
"llm": 1
|
| 22 |
+
},
|
| 23 |
+
"layer_norm_eps": 1e-12,
|
| 24 |
+
"max_position_embeddings": 512,
|
| 25 |
+
"model_type": "bert",
|
| 26 |
+
"num_attention_heads": 12,
|
| 27 |
+
"num_hidden_layers": 12,
|
| 28 |
+
"pad_token_id": 0,
|
| 29 |
+
"position_embedding_type": "absolute",
|
| 30 |
+
"problem_type": "single_label_classification",
|
| 31 |
+
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.48.0",
|
| 33 |
+
"type_vocab_size": 2,
|
| 34 |
+
"use_cache": true,
|
| 35 |
+
"vocab_size": 30522
|
| 36 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ba5d6da12fadf4d20c73020e0c623a4371f36c06756ed7a0d90544a103ad255
|
| 3 |
+
size 437958648
|
runs/May29_13-39-01_r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz/events.out.tfevents.1748525943.r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz.76.0
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:ff76292c2578b6353c51a5af18bbe5bb29c045df97e53e965b0600700e20d643
|
| 3 |
+
size 58794
|
runs/May29_13-39-01_r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz/events.out.tfevents.1748526547.r-trentmkelly-autotrain-advanced-frsdremx-47e49-hivzz.76.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f469f7caa9f22ad18743f8c14b9797f124abb520a0ee823245b52e400562ffdf
|
| 3 |
+
size 607
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"extra_special_tokens": {},
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"max_length": 128,
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"pad_to_multiple_of": null,
|
| 52 |
+
"pad_token": "[PAD]",
|
| 53 |
+
"pad_token_type_id": 0,
|
| 54 |
+
"padding_side": "right",
|
| 55 |
+
"sep_token": "[SEP]",
|
| 56 |
+
"stride": 0,
|
| 57 |
+
"strip_accents": null,
|
| 58 |
+
"tokenize_chinese_chars": true,
|
| 59 |
+
"tokenizer_class": "BertTokenizer",
|
| 60 |
+
"truncation_side": "right",
|
| 61 |
+
"truncation_strategy": "longest_first",
|
| 62 |
+
"unk_token": "[UNK]"
|
| 63 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7cb80b880983dad90e2001410afc3b3aad5340c8f3467534951dc6ceb9eca388
|
| 3 |
+
size 5368
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "slop-detector/autotrain-data",
|
| 3 |
+
"model": "thenlper/gte-base",
|
| 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": "slop-detector",
|
| 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": "trentmkelly",
|
| 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
|
|
|