Upload folder using huggingface_hub
Browse files- README.md +27 -0
- checkpoint-6970/config.json +35 -0
- checkpoint-6970/model.safetensors +3 -0
- checkpoint-6970/optimizer.pt +3 -0
- checkpoint-6970/rng_state.pth +3 -0
- checkpoint-6970/scheduler.pt +3 -0
- checkpoint-6970/trainer_state.json +2014 -0
- checkpoint-6970/training_args.bin +3 -0
- config.json +35 -0
- model.safetensors +3 -0
- runs/May30_00-09-04_r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w/events.out.tfevents.1748563746.r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w.75.0 +2 -2
- runs/May30_00-09-04_r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w/events.out.tfevents.1748564133.r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w.75.1 +3 -0
- special_tokens_map.json +44 -0
- tokenizer.json +0 -0
- tokenizer_config.json +72 -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: TaylorAI/gte-tiny
|
| 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: 0.04012129828333855
|
| 18 |
+
|
| 19 |
+
f1: 0.9900353584056574
|
| 20 |
+
|
| 21 |
+
precision: 0.9859154929577465
|
| 22 |
+
|
| 23 |
+
recall: 0.9941897998708844
|
| 24 |
+
|
| 25 |
+
auc: 0.999704926354536
|
| 26 |
+
|
| 27 |
+
accuracy: 0.9899935442220787
|
checkpoint-6970/config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "TaylorAI/gte-tiny",
|
| 3 |
+
"_num_labels": 2,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 384,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "human",
|
| 14 |
+
"1": "llm"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 1536,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"human": 0,
|
| 20 |
+
"llm": 1
|
| 21 |
+
},
|
| 22 |
+
"layer_norm_eps": 1e-12,
|
| 23 |
+
"max_position_embeddings": 512,
|
| 24 |
+
"model_type": "bert",
|
| 25 |
+
"num_attention_heads": 12,
|
| 26 |
+
"num_hidden_layers": 6,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"problem_type": "single_label_classification",
|
| 30 |
+
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.48.0",
|
| 32 |
+
"type_vocab_size": 2,
|
| 33 |
+
"use_cache": true,
|
| 34 |
+
"vocab_size": 30522
|
| 35 |
+
}
|
checkpoint-6970/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1dc3603f7927d59bbb28381542e7c65f0c7d40444c17d18e2bdefc35a0d77c0
|
| 3 |
+
size 90867952
|
checkpoint-6970/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:466ebe4983d35a27315b3f0ceb0db7dc49bf07e1a58d60bc6a063c7df30bf468
|
| 3 |
+
size 181799994
|
checkpoint-6970/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31c8269223c50fe95286007f63e5c42016e01d16bf3cea35eac7ea46df2bdee1
|
| 3 |
+
size 14244
|
checkpoint-6970/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e941fa1b7d020c0e74b4c2944df4fa2d6c820122b3692afc92f145357c7443d
|
| 3 |
+
size 1064
|
checkpoint-6970/trainer_state.json
ADDED
|
@@ -0,0 +1,2014 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.04012129828333855,
|
| 3 |
+
"best_model_checkpoint": "slop-detector-mini/checkpoint-6970",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 6970,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.007173601147776184,
|
| 13 |
+
"grad_norm": 0.4645521938800812,
|
| 14 |
+
"learning_rate": 1.1950286806883365e-06,
|
| 15 |
+
"loss": 0.6918,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.014347202295552367,
|
| 20 |
+
"grad_norm": 0.9209471940994263,
|
| 21 |
+
"learning_rate": 2.390057361376673e-06,
|
| 22 |
+
"loss": 0.6919,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.021520803443328552,
|
| 27 |
+
"grad_norm": 0.5008063316345215,
|
| 28 |
+
"learning_rate": 3.58508604206501e-06,
|
| 29 |
+
"loss": 0.6918,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.028694404591104734,
|
| 34 |
+
"grad_norm": 2.1300737857818604,
|
| 35 |
+
"learning_rate": 4.780114722753346e-06,
|
| 36 |
+
"loss": 0.6872,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.035868005738880916,
|
| 41 |
+
"grad_norm": 0.9959478974342346,
|
| 42 |
+
"learning_rate": 5.9751434034416826e-06,
|
| 43 |
+
"loss": 0.6896,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.043041606886657105,
|
| 48 |
+
"grad_norm": 1.5318769216537476,
|
| 49 |
+
"learning_rate": 7.17017208413002e-06,
|
| 50 |
+
"loss": 0.6838,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05021520803443329,
|
| 55 |
+
"grad_norm": 1.0408401489257812,
|
| 56 |
+
"learning_rate": 8.365200764818356e-06,
|
| 57 |
+
"loss": 0.6665,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05738880918220947,
|
| 62 |
+
"grad_norm": 1.6900608539581299,
|
| 63 |
+
"learning_rate": 9.560229445506692e-06,
|
| 64 |
+
"loss": 0.5941,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06456241032998565,
|
| 69 |
+
"grad_norm": 1.8240679502487183,
|
| 70 |
+
"learning_rate": 1.075525812619503e-05,
|
| 71 |
+
"loss": 0.4528,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.07173601147776183,
|
| 76 |
+
"grad_norm": 9.910821914672852,
|
| 77 |
+
"learning_rate": 1.1950286806883365e-05,
|
| 78 |
+
"loss": 0.3188,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07890961262553801,
|
| 83 |
+
"grad_norm": 4.513535022735596,
|
| 84 |
+
"learning_rate": 1.3145315487571702e-05,
|
| 85 |
+
"loss": 0.292,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.08608321377331421,
|
| 90 |
+
"grad_norm": 0.8180878162384033,
|
| 91 |
+
"learning_rate": 1.434034416826004e-05,
|
| 92 |
+
"loss": 0.1973,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09325681492109039,
|
| 97 |
+
"grad_norm": 0.6453915238380432,
|
| 98 |
+
"learning_rate": 1.5535372848948377e-05,
|
| 99 |
+
"loss": 0.1999,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.10043041606886657,
|
| 104 |
+
"grad_norm": 1.2022216320037842,
|
| 105 |
+
"learning_rate": 1.6682600382409177e-05,
|
| 106 |
+
"loss": 0.155,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.10760401721664276,
|
| 111 |
+
"grad_norm": 0.4976617693901062,
|
| 112 |
+
"learning_rate": 1.7877629063097516e-05,
|
| 113 |
+
"loss": 0.1583,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.11477761836441894,
|
| 118 |
+
"grad_norm": 2.098670482635498,
|
| 119 |
+
"learning_rate": 1.9072657743785852e-05,
|
| 120 |
+
"loss": 0.1178,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.12195121951219512,
|
| 125 |
+
"grad_norm": 0.7542682886123657,
|
| 126 |
+
"learning_rate": 2.0267686424474187e-05,
|
| 127 |
+
"loss": 0.169,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1291248206599713,
|
| 132 |
+
"grad_norm": 10.018378257751465,
|
| 133 |
+
"learning_rate": 2.1414913957934994e-05,
|
| 134 |
+
"loss": 0.1356,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.13629842180774748,
|
| 139 |
+
"grad_norm": 34.80171203613281,
|
| 140 |
+
"learning_rate": 2.2609942638623326e-05,
|
| 141 |
+
"loss": 0.1258,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.14347202295552366,
|
| 146 |
+
"grad_norm": 0.1728999763727188,
|
| 147 |
+
"learning_rate": 2.3804971319311666e-05,
|
| 148 |
+
"loss": 0.0863,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15064562410329985,
|
| 153 |
+
"grad_norm": 0.1769145131111145,
|
| 154 |
+
"learning_rate": 2.5e-05,
|
| 155 |
+
"loss": 0.1017,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.15781922525107603,
|
| 160 |
+
"grad_norm": 6.390002250671387,
|
| 161 |
+
"learning_rate": 2.6195028680688337e-05,
|
| 162 |
+
"loss": 0.0829,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1649928263988522,
|
| 167 |
+
"grad_norm": 0.13392409682273865,
|
| 168 |
+
"learning_rate": 2.7390057361376676e-05,
|
| 169 |
+
"loss": 0.0991,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.17216642754662842,
|
| 174 |
+
"grad_norm": 9.020365715026855,
|
| 175 |
+
"learning_rate": 2.853728489483748e-05,
|
| 176 |
+
"loss": 0.1088,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1793400286944046,
|
| 181 |
+
"grad_norm": 0.17115327715873718,
|
| 182 |
+
"learning_rate": 2.9732313575525815e-05,
|
| 183 |
+
"loss": 0.0858,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.18651362984218078,
|
| 188 |
+
"grad_norm": 0.1999664157629013,
|
| 189 |
+
"learning_rate": 3.092734225621415e-05,
|
| 190 |
+
"loss": 0.1257,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.19368723098995697,
|
| 195 |
+
"grad_norm": 0.10073782503604889,
|
| 196 |
+
"learning_rate": 3.2122370936902486e-05,
|
| 197 |
+
"loss": 0.0434,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.20086083213773315,
|
| 202 |
+
"grad_norm": 14.117579460144043,
|
| 203 |
+
"learning_rate": 3.331739961759082e-05,
|
| 204 |
+
"loss": 0.169,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.20803443328550933,
|
| 209 |
+
"grad_norm": 48.392974853515625,
|
| 210 |
+
"learning_rate": 3.451242829827916e-05,
|
| 211 |
+
"loss": 0.0908,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2152080344332855,
|
| 216 |
+
"grad_norm": 0.10534697771072388,
|
| 217 |
+
"learning_rate": 3.570745697896749e-05,
|
| 218 |
+
"loss": 0.1517,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2223816355810617,
|
| 223 |
+
"grad_norm": 0.07163497805595398,
|
| 224 |
+
"learning_rate": 3.6902485659655836e-05,
|
| 225 |
+
"loss": 0.054,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.22955523672883787,
|
| 230 |
+
"grad_norm": 0.0702139139175415,
|
| 231 |
+
"learning_rate": 3.809751434034417e-05,
|
| 232 |
+
"loss": 0.1009,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.23672883787661406,
|
| 237 |
+
"grad_norm": 0.698357105255127,
|
| 238 |
+
"learning_rate": 3.929254302103251e-05,
|
| 239 |
+
"loss": 0.0344,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.24390243902439024,
|
| 244 |
+
"grad_norm": 0.05920593440532684,
|
| 245 |
+
"learning_rate": 4.048757170172084e-05,
|
| 246 |
+
"loss": 0.104,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.25107604017216645,
|
| 251 |
+
"grad_norm": 0.08545070886611938,
|
| 252 |
+
"learning_rate": 4.168260038240918e-05,
|
| 253 |
+
"loss": 0.1973,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2582496413199426,
|
| 258 |
+
"grad_norm": 29.820629119873047,
|
| 259 |
+
"learning_rate": 4.287762906309752e-05,
|
| 260 |
+
"loss": 0.1245,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2654232424677188,
|
| 265 |
+
"grad_norm": 18.92865562438965,
|
| 266 |
+
"learning_rate": 4.4072657743785856e-05,
|
| 267 |
+
"loss": 0.1508,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.27259684361549497,
|
| 272 |
+
"grad_norm": 3.224970579147339,
|
| 273 |
+
"learning_rate": 4.5267686424474185e-05,
|
| 274 |
+
"loss": 0.0772,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2797704447632712,
|
| 279 |
+
"grad_norm": 56.050315856933594,
|
| 280 |
+
"learning_rate": 4.646271510516252e-05,
|
| 281 |
+
"loss": 0.0839,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.28694404591104733,
|
| 286 |
+
"grad_norm": 45.18210983276367,
|
| 287 |
+
"learning_rate": 4.7657743785850863e-05,
|
| 288 |
+
"loss": 0.1926,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.29411764705882354,
|
| 293 |
+
"grad_norm": 0.06930357962846756,
|
| 294 |
+
"learning_rate": 4.88527724665392e-05,
|
| 295 |
+
"loss": 0.1326,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.3012912482065997,
|
| 300 |
+
"grad_norm": 0.05295657739043236,
|
| 301 |
+
"learning_rate": 4.9994685938994586e-05,
|
| 302 |
+
"loss": 0.0758,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3084648493543759,
|
| 307 |
+
"grad_norm": 21.23689842224121,
|
| 308 |
+
"learning_rate": 4.986183441385907e-05,
|
| 309 |
+
"loss": 0.0863,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.31563845050215206,
|
| 314 |
+
"grad_norm": 55.6703987121582,
|
| 315 |
+
"learning_rate": 4.9728982888723565e-05,
|
| 316 |
+
"loss": 0.1115,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.32281205164992827,
|
| 321 |
+
"grad_norm": 0.04365883022546768,
|
| 322 |
+
"learning_rate": 4.9596131363588055e-05,
|
| 323 |
+
"loss": 0.0404,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3299856527977044,
|
| 328 |
+
"grad_norm": 0.033937107771635056,
|
| 329 |
+
"learning_rate": 4.9463279838452545e-05,
|
| 330 |
+
"loss": 0.0963,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.33715925394548063,
|
| 335 |
+
"grad_norm": 0.049055054783821106,
|
| 336 |
+
"learning_rate": 4.933042831331704e-05,
|
| 337 |
+
"loss": 0.0539,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.34433285509325684,
|
| 342 |
+
"grad_norm": 0.038862813264131546,
|
| 343 |
+
"learning_rate": 4.919757678818153e-05,
|
| 344 |
+
"loss": 0.0102,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.351506456241033,
|
| 349 |
+
"grad_norm": 0.0405915193259716,
|
| 350 |
+
"learning_rate": 4.906472526304603e-05,
|
| 351 |
+
"loss": 0.0745,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3586800573888092,
|
| 356 |
+
"grad_norm": 0.05233052000403404,
|
| 357 |
+
"learning_rate": 4.893187373791052e-05,
|
| 358 |
+
"loss": 0.0743,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.36585365853658536,
|
| 363 |
+
"grad_norm": 0.037946607917547226,
|
| 364 |
+
"learning_rate": 4.8799022212775e-05,
|
| 365 |
+
"loss": 0.0348,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.37302725968436157,
|
| 370 |
+
"grad_norm": 0.031401779502630234,
|
| 371 |
+
"learning_rate": 4.8666170687639496e-05,
|
| 372 |
+
"loss": 0.0385,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.3802008608321377,
|
| 377 |
+
"grad_norm": 0.04547421261668205,
|
| 378 |
+
"learning_rate": 4.8533319162503986e-05,
|
| 379 |
+
"loss": 0.1143,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.38737446197991393,
|
| 384 |
+
"grad_norm": 0.05827297270298004,
|
| 385 |
+
"learning_rate": 4.840046763736848e-05,
|
| 386 |
+
"loss": 0.0772,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3945480631276901,
|
| 391 |
+
"grad_norm": 0.08565480262041092,
|
| 392 |
+
"learning_rate": 4.826761611223297e-05,
|
| 393 |
+
"loss": 0.1192,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.4017216642754663,
|
| 398 |
+
"grad_norm": 0.03007586859166622,
|
| 399 |
+
"learning_rate": 4.813476458709746e-05,
|
| 400 |
+
"loss": 0.0613,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.40889526542324245,
|
| 405 |
+
"grad_norm": 0.027408963069319725,
|
| 406 |
+
"learning_rate": 4.800191306196196e-05,
|
| 407 |
+
"loss": 0.0029,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.41606886657101866,
|
| 412 |
+
"grad_norm": 0.028056548908352852,
|
| 413 |
+
"learning_rate": 4.786906153682644e-05,
|
| 414 |
+
"loss": 0.0902,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.4232424677187948,
|
| 419 |
+
"grad_norm": 0.08005893230438232,
|
| 420 |
+
"learning_rate": 4.773621001169094e-05,
|
| 421 |
+
"loss": 0.1209,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.430416068866571,
|
| 426 |
+
"grad_norm": 0.06766340136528015,
|
| 427 |
+
"learning_rate": 4.760335848655543e-05,
|
| 428 |
+
"loss": 0.0201,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.4375896700143472,
|
| 433 |
+
"grad_norm": 0.03779875114560127,
|
| 434 |
+
"learning_rate": 4.747050696141992e-05,
|
| 435 |
+
"loss": 0.0622,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.4447632711621234,
|
| 440 |
+
"grad_norm": 0.03687926381826401,
|
| 441 |
+
"learning_rate": 4.7337655436284414e-05,
|
| 442 |
+
"loss": 0.121,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4519368723098996,
|
| 447 |
+
"grad_norm": 0.03594242036342621,
|
| 448 |
+
"learning_rate": 4.7204803911148904e-05,
|
| 449 |
+
"loss": 0.1053,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.45911047345767575,
|
| 454 |
+
"grad_norm": 0.03990463539958,
|
| 455 |
+
"learning_rate": 4.707195238601339e-05,
|
| 456 |
+
"loss": 0.0571,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.46628407460545196,
|
| 461 |
+
"grad_norm": 0.04624362289905548,
|
| 462 |
+
"learning_rate": 4.693910086087788e-05,
|
| 463 |
+
"loss": 0.1129,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4734576757532281,
|
| 468 |
+
"grad_norm": 0.035806916654109955,
|
| 469 |
+
"learning_rate": 4.680624933574237e-05,
|
| 470 |
+
"loss": 0.0464,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4806312769010043,
|
| 475 |
+
"grad_norm": 4.62434720993042,
|
| 476 |
+
"learning_rate": 4.667339781060687e-05,
|
| 477 |
+
"loss": 0.052,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.4878048780487805,
|
| 482 |
+
"grad_norm": 0.028078608214855194,
|
| 483 |
+
"learning_rate": 4.654054628547136e-05,
|
| 484 |
+
"loss": 0.1337,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.4949784791965567,
|
| 489 |
+
"grad_norm": 0.029056400060653687,
|
| 490 |
+
"learning_rate": 4.640769476033585e-05,
|
| 491 |
+
"loss": 0.0727,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.5021520803443329,
|
| 496 |
+
"grad_norm": 0.021900108084082603,
|
| 497 |
+
"learning_rate": 4.6274843235200345e-05,
|
| 498 |
+
"loss": 0.0704,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.509325681492109,
|
| 503 |
+
"grad_norm": 0.019067928194999695,
|
| 504 |
+
"learning_rate": 4.6141991710064835e-05,
|
| 505 |
+
"loss": 0.0147,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5164992826398852,
|
| 510 |
+
"grad_norm": 54.136009216308594,
|
| 511 |
+
"learning_rate": 4.6009140184929324e-05,
|
| 512 |
+
"loss": 0.1138,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5236728837876614,
|
| 517 |
+
"grad_norm": 0.030056435614824295,
|
| 518 |
+
"learning_rate": 4.5876288659793814e-05,
|
| 519 |
+
"loss": 0.0791,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.5308464849354376,
|
| 524 |
+
"grad_norm": 0.022308100014925003,
|
| 525 |
+
"learning_rate": 4.5743437134658304e-05,
|
| 526 |
+
"loss": 0.0643,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.5380200860832137,
|
| 531 |
+
"grad_norm": 0.13602350652217865,
|
| 532 |
+
"learning_rate": 4.56105856095228e-05,
|
| 533 |
+
"loss": 0.0891,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5451936872309899,
|
| 538 |
+
"grad_norm": 2.2635345458984375,
|
| 539 |
+
"learning_rate": 4.547773408438729e-05,
|
| 540 |
+
"loss": 0.0951,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.5523672883787661,
|
| 545 |
+
"grad_norm": 0.02811610698699951,
|
| 546 |
+
"learning_rate": 4.5344882559251787e-05,
|
| 547 |
+
"loss": 0.0116,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5595408895265424,
|
| 552 |
+
"grad_norm": 0.02545151859521866,
|
| 553 |
+
"learning_rate": 4.5212031034116276e-05,
|
| 554 |
+
"loss": 0.0335,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.5667144906743186,
|
| 559 |
+
"grad_norm": 0.20565590262413025,
|
| 560 |
+
"learning_rate": 4.5079179508980766e-05,
|
| 561 |
+
"loss": 0.1518,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5738880918220947,
|
| 566 |
+
"grad_norm": 26.854698181152344,
|
| 567 |
+
"learning_rate": 4.4946327983845256e-05,
|
| 568 |
+
"loss": 0.0928,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.5810616929698709,
|
| 573 |
+
"grad_norm": 0.28100213408470154,
|
| 574 |
+
"learning_rate": 4.4813476458709745e-05,
|
| 575 |
+
"loss": 0.0961,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5882352941176471,
|
| 580 |
+
"grad_norm": 0.020814087241888046,
|
| 581 |
+
"learning_rate": 4.468062493357424e-05,
|
| 582 |
+
"loss": 0.0564,
|
| 583 |
+
"step": 2050
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.5954088952654233,
|
| 587 |
+
"grad_norm": 0.019664647057652473,
|
| 588 |
+
"learning_rate": 4.454777340843873e-05,
|
| 589 |
+
"loss": 0.0344,
|
| 590 |
+
"step": 2075
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.6025824964131994,
|
| 594 |
+
"grad_norm": 0.11771227419376373,
|
| 595 |
+
"learning_rate": 4.441492188330322e-05,
|
| 596 |
+
"loss": 0.1018,
|
| 597 |
+
"step": 2100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.6097560975609756,
|
| 601 |
+
"grad_norm": 0.3748355209827423,
|
| 602 |
+
"learning_rate": 4.428207035816772e-05,
|
| 603 |
+
"loss": 0.1281,
|
| 604 |
+
"step": 2125
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.6169296987087518,
|
| 608 |
+
"grad_norm": 0.02155977115035057,
|
| 609 |
+
"learning_rate": 4.414921883303221e-05,
|
| 610 |
+
"loss": 0.0264,
|
| 611 |
+
"step": 2150
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.624103299856528,
|
| 615 |
+
"grad_norm": 0.017244333401322365,
|
| 616 |
+
"learning_rate": 4.40163673078967e-05,
|
| 617 |
+
"loss": 0.0611,
|
| 618 |
+
"step": 2175
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.6312769010043041,
|
| 622 |
+
"grad_norm": 0.014310152269899845,
|
| 623 |
+
"learning_rate": 4.388351578276119e-05,
|
| 624 |
+
"loss": 0.0444,
|
| 625 |
+
"step": 2200
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.6384505021520803,
|
| 629 |
+
"grad_norm": 0.015014859847724438,
|
| 630 |
+
"learning_rate": 4.375066425762568e-05,
|
| 631 |
+
"loss": 0.0199,
|
| 632 |
+
"step": 2225
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.6456241032998565,
|
| 636 |
+
"grad_norm": 0.028923267498612404,
|
| 637 |
+
"learning_rate": 4.361781273249017e-05,
|
| 638 |
+
"loss": 0.0668,
|
| 639 |
+
"step": 2250
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.6527977044476327,
|
| 643 |
+
"grad_norm": 0.017175152897834778,
|
| 644 |
+
"learning_rate": 4.348496120735466e-05,
|
| 645 |
+
"loss": 0.0333,
|
| 646 |
+
"step": 2275
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6599713055954088,
|
| 650 |
+
"grad_norm": 0.018409280106425285,
|
| 651 |
+
"learning_rate": 4.335210968221915e-05,
|
| 652 |
+
"loss": 0.1059,
|
| 653 |
+
"step": 2300
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.667144906743185,
|
| 657 |
+
"grad_norm": 0.019378701224923134,
|
| 658 |
+
"learning_rate": 4.321925815708365e-05,
|
| 659 |
+
"loss": 0.0403,
|
| 660 |
+
"step": 2325
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6743185078909613,
|
| 664 |
+
"grad_norm": 0.015860414132475853,
|
| 665 |
+
"learning_rate": 4.308640663194813e-05,
|
| 666 |
+
"loss": 0.0929,
|
| 667 |
+
"step": 2350
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6814921090387375,
|
| 671 |
+
"grad_norm": 0.01943328231573105,
|
| 672 |
+
"learning_rate": 4.295355510681263e-05,
|
| 673 |
+
"loss": 0.0014,
|
| 674 |
+
"step": 2375
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6886657101865137,
|
| 678 |
+
"grad_norm": 0.4460265040397644,
|
| 679 |
+
"learning_rate": 4.282070358167712e-05,
|
| 680 |
+
"loss": 0.0233,
|
| 681 |
+
"step": 2400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.6958393113342898,
|
| 685 |
+
"grad_norm": 0.013770678080618382,
|
| 686 |
+
"learning_rate": 4.268785205654161e-05,
|
| 687 |
+
"loss": 0.0438,
|
| 688 |
+
"step": 2425
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.703012912482066,
|
| 692 |
+
"grad_norm": 0.466856986284256,
|
| 693 |
+
"learning_rate": 4.2555000531406104e-05,
|
| 694 |
+
"loss": 0.0331,
|
| 695 |
+
"step": 2450
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.7101865136298422,
|
| 699 |
+
"grad_norm": 0.05924264341592789,
|
| 700 |
+
"learning_rate": 4.2422149006270594e-05,
|
| 701 |
+
"loss": 0.0502,
|
| 702 |
+
"step": 2475
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.7173601147776184,
|
| 706 |
+
"grad_norm": 2.5010907649993896,
|
| 707 |
+
"learning_rate": 4.228929748113509e-05,
|
| 708 |
+
"loss": 0.0245,
|
| 709 |
+
"step": 2500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.7245337159253945,
|
| 713 |
+
"grad_norm": 0.011777346953749657,
|
| 714 |
+
"learning_rate": 4.215644595599958e-05,
|
| 715 |
+
"loss": 0.0585,
|
| 716 |
+
"step": 2525
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.7317073170731707,
|
| 720 |
+
"grad_norm": 0.021891452372074127,
|
| 721 |
+
"learning_rate": 4.202359443086406e-05,
|
| 722 |
+
"loss": 0.1348,
|
| 723 |
+
"step": 2550
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.7388809182209469,
|
| 727 |
+
"grad_norm": 0.011559663340449333,
|
| 728 |
+
"learning_rate": 4.189074290572856e-05,
|
| 729 |
+
"loss": 0.0781,
|
| 730 |
+
"step": 2575
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.7460545193687231,
|
| 734 |
+
"grad_norm": 0.011871904134750366,
|
| 735 |
+
"learning_rate": 4.175789138059305e-05,
|
| 736 |
+
"loss": 0.1001,
|
| 737 |
+
"step": 2600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.7532281205164992,
|
| 741 |
+
"grad_norm": 0.011486608535051346,
|
| 742 |
+
"learning_rate": 4.1625039855457546e-05,
|
| 743 |
+
"loss": 0.036,
|
| 744 |
+
"step": 2625
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.7604017216642754,
|
| 748 |
+
"grad_norm": 0.03866451606154442,
|
| 749 |
+
"learning_rate": 4.1492188330322036e-05,
|
| 750 |
+
"loss": 0.0781,
|
| 751 |
+
"step": 2650
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7675753228120517,
|
| 755 |
+
"grad_norm": 0.04854358360171318,
|
| 756 |
+
"learning_rate": 4.1359336805186525e-05,
|
| 757 |
+
"loss": 0.018,
|
| 758 |
+
"step": 2675
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.7747489239598279,
|
| 762 |
+
"grad_norm": 26.049694061279297,
|
| 763 |
+
"learning_rate": 4.122648528005102e-05,
|
| 764 |
+
"loss": 0.1062,
|
| 765 |
+
"step": 2700
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7819225251076041,
|
| 769 |
+
"grad_norm": 0.09440884739160538,
|
| 770 |
+
"learning_rate": 4.1093633754915505e-05,
|
| 771 |
+
"loss": 0.0293,
|
| 772 |
+
"step": 2725
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7890961262553802,
|
| 776 |
+
"grad_norm": 0.008743391372263432,
|
| 777 |
+
"learning_rate": 4.096078222978e-05,
|
| 778 |
+
"loss": 0.1154,
|
| 779 |
+
"step": 2750
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.7962697274031564,
|
| 783 |
+
"grad_norm": 0.0563165545463562,
|
| 784 |
+
"learning_rate": 4.082793070464449e-05,
|
| 785 |
+
"loss": 0.0505,
|
| 786 |
+
"step": 2775
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.8034433285509326,
|
| 790 |
+
"grad_norm": 0.031961794942617416,
|
| 791 |
+
"learning_rate": 4.069507917950898e-05,
|
| 792 |
+
"loss": 0.0698,
|
| 793 |
+
"step": 2800
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.8106169296987088,
|
| 797 |
+
"grad_norm": 0.013981098309159279,
|
| 798 |
+
"learning_rate": 4.056222765437348e-05,
|
| 799 |
+
"loss": 0.0531,
|
| 800 |
+
"step": 2825
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.8177905308464849,
|
| 804 |
+
"grad_norm": 0.016712762415409088,
|
| 805 |
+
"learning_rate": 4.042937612923797e-05,
|
| 806 |
+
"loss": 0.0847,
|
| 807 |
+
"step": 2850
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.8249641319942611,
|
| 811 |
+
"grad_norm": 46.247108459472656,
|
| 812 |
+
"learning_rate": 4.0296524604102456e-05,
|
| 813 |
+
"loss": 0.0463,
|
| 814 |
+
"step": 2875
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.8321377331420373,
|
| 818 |
+
"grad_norm": 31.703798294067383,
|
| 819 |
+
"learning_rate": 4.0163673078966946e-05,
|
| 820 |
+
"loss": 0.0719,
|
| 821 |
+
"step": 2900
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.8393113342898135,
|
| 825 |
+
"grad_norm": 0.010652084834873676,
|
| 826 |
+
"learning_rate": 4.0030821553831436e-05,
|
| 827 |
+
"loss": 0.0805,
|
| 828 |
+
"step": 2925
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.8464849354375896,
|
| 832 |
+
"grad_norm": 0.00876002386212349,
|
| 833 |
+
"learning_rate": 3.989797002869593e-05,
|
| 834 |
+
"loss": 0.0358,
|
| 835 |
+
"step": 2950
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.8536585365853658,
|
| 839 |
+
"grad_norm": 0.042157307267189026,
|
| 840 |
+
"learning_rate": 3.976511850356042e-05,
|
| 841 |
+
"loss": 0.0322,
|
| 842 |
+
"step": 2975
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.860832137733142,
|
| 846 |
+
"grad_norm": 0.01021247822791338,
|
| 847 |
+
"learning_rate": 3.963226697842491e-05,
|
| 848 |
+
"loss": 0.0437,
|
| 849 |
+
"step": 3000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.8680057388809183,
|
| 853 |
+
"grad_norm": 0.010883198119699955,
|
| 854 |
+
"learning_rate": 3.949941545328941e-05,
|
| 855 |
+
"loss": 0.0059,
|
| 856 |
+
"step": 3025
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.8751793400286944,
|
| 860 |
+
"grad_norm": 0.006532243452966213,
|
| 861 |
+
"learning_rate": 3.93665639281539e-05,
|
| 862 |
+
"loss": 0.0149,
|
| 863 |
+
"step": 3050
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.8823529411764706,
|
| 867 |
+
"grad_norm": 0.006005235482007265,
|
| 868 |
+
"learning_rate": 3.9233712403018394e-05,
|
| 869 |
+
"loss": 0.0036,
|
| 870 |
+
"step": 3075
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.8895265423242468,
|
| 874 |
+
"grad_norm": 0.0063223764300346375,
|
| 875 |
+
"learning_rate": 3.910086087788288e-05,
|
| 876 |
+
"loss": 0.0095,
|
| 877 |
+
"step": 3100
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.896700143472023,
|
| 881 |
+
"grad_norm": 0.009072021581232548,
|
| 882 |
+
"learning_rate": 3.896800935274737e-05,
|
| 883 |
+
"loss": 0.0004,
|
| 884 |
+
"step": 3125
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.9038737446197992,
|
| 888 |
+
"grad_norm": 0.05379496142268181,
|
| 889 |
+
"learning_rate": 3.8835157827611864e-05,
|
| 890 |
+
"loss": 0.1547,
|
| 891 |
+
"step": 3150
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.9110473457675753,
|
| 895 |
+
"grad_norm": 0.013865516521036625,
|
| 896 |
+
"learning_rate": 3.870230630247635e-05,
|
| 897 |
+
"loss": 0.0036,
|
| 898 |
+
"step": 3175
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.9182209469153515,
|
| 902 |
+
"grad_norm": 0.010529842227697372,
|
| 903 |
+
"learning_rate": 3.856945477734085e-05,
|
| 904 |
+
"loss": 0.0344,
|
| 905 |
+
"step": 3200
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.9253945480631277,
|
| 909 |
+
"grad_norm": 0.014116561971604824,
|
| 910 |
+
"learning_rate": 3.843660325220534e-05,
|
| 911 |
+
"loss": 0.123,
|
| 912 |
+
"step": 3225
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.9325681492109039,
|
| 916 |
+
"grad_norm": 0.0235675647854805,
|
| 917 |
+
"learning_rate": 3.830375172706983e-05,
|
| 918 |
+
"loss": 0.0471,
|
| 919 |
+
"step": 3250
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.93974175035868,
|
| 923 |
+
"grad_norm": 0.07236867398023605,
|
| 924 |
+
"learning_rate": 3.817090020193432e-05,
|
| 925 |
+
"loss": 0.0576,
|
| 926 |
+
"step": 3275
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.9469153515064562,
|
| 930 |
+
"grad_norm": 0.008477970026433468,
|
| 931 |
+
"learning_rate": 3.803804867679881e-05,
|
| 932 |
+
"loss": 0.0656,
|
| 933 |
+
"step": 3300
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.9540889526542324,
|
| 937 |
+
"grad_norm": 0.009889235720038414,
|
| 938 |
+
"learning_rate": 3.7905197151663305e-05,
|
| 939 |
+
"loss": 0.0141,
|
| 940 |
+
"step": 3325
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.9612625538020086,
|
| 944 |
+
"grad_norm": 0.04827851429581642,
|
| 945 |
+
"learning_rate": 3.7772345626527795e-05,
|
| 946 |
+
"loss": 0.2139,
|
| 947 |
+
"step": 3350
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.9684361549497847,
|
| 951 |
+
"grad_norm": 0.022464429959654808,
|
| 952 |
+
"learning_rate": 3.7639494101392285e-05,
|
| 953 |
+
"loss": 0.0343,
|
| 954 |
+
"step": 3375
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.975609756097561,
|
| 958 |
+
"grad_norm": 0.02515058033168316,
|
| 959 |
+
"learning_rate": 3.750664257625678e-05,
|
| 960 |
+
"loss": 0.0389,
|
| 961 |
+
"step": 3400
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.9827833572453372,
|
| 965 |
+
"grad_norm": 0.04209022969007492,
|
| 966 |
+
"learning_rate": 3.737379105112127e-05,
|
| 967 |
+
"loss": 0.0556,
|
| 968 |
+
"step": 3425
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.9899569583931134,
|
| 972 |
+
"grad_norm": 0.025623561814427376,
|
| 973 |
+
"learning_rate": 3.724093952598576e-05,
|
| 974 |
+
"loss": 0.11,
|
| 975 |
+
"step": 3450
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.9971305595408895,
|
| 979 |
+
"grad_norm": 0.044107332825660706,
|
| 980 |
+
"learning_rate": 3.710808800085025e-05,
|
| 981 |
+
"loss": 0.0703,
|
| 982 |
+
"step": 3475
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.0,
|
| 986 |
+
"eval_accuracy": 0.9874112330535829,
|
| 987 |
+
"eval_auc": 0.9992416857373987,
|
| 988 |
+
"eval_f1": 0.9875120076849183,
|
| 989 |
+
"eval_loss": 0.0463985949754715,
|
| 990 |
+
"eval_precision": 0.9796696315120712,
|
| 991 |
+
"eval_recall": 0.9954809554551324,
|
| 992 |
+
"eval_runtime": 4.3248,
|
| 993 |
+
"eval_samples_per_second": 716.333,
|
| 994 |
+
"eval_steps_per_second": 44.857,
|
| 995 |
+
"step": 3485
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.0043041606886658,
|
| 999 |
+
"grad_norm": 0.006846719887107611,
|
| 1000 |
+
"learning_rate": 3.697523647571474e-05,
|
| 1001 |
+
"loss": 0.009,
|
| 1002 |
+
"step": 3500
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.011477761836442,
|
| 1006 |
+
"grad_norm": 3.4670445919036865,
|
| 1007 |
+
"learning_rate": 3.6842384950579236e-05,
|
| 1008 |
+
"loss": 0.034,
|
| 1009 |
+
"step": 3525
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.018651362984218,
|
| 1013 |
+
"grad_norm": 14.726991653442383,
|
| 1014 |
+
"learning_rate": 3.6709533425443726e-05,
|
| 1015 |
+
"loss": 0.0348,
|
| 1016 |
+
"step": 3550
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.0258249641319943,
|
| 1020 |
+
"grad_norm": 0.01886272057890892,
|
| 1021 |
+
"learning_rate": 3.6576681900308216e-05,
|
| 1022 |
+
"loss": 0.045,
|
| 1023 |
+
"step": 3575
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.0329985652797704,
|
| 1027 |
+
"grad_norm": 0.04845844581723213,
|
| 1028 |
+
"learning_rate": 3.644383037517271e-05,
|
| 1029 |
+
"loss": 0.0386,
|
| 1030 |
+
"step": 3600
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.0401721664275467,
|
| 1034 |
+
"grad_norm": 0.04974396526813507,
|
| 1035 |
+
"learning_rate": 3.63109788500372e-05,
|
| 1036 |
+
"loss": 0.0336,
|
| 1037 |
+
"step": 3625
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.0473457675753228,
|
| 1041 |
+
"grad_norm": 0.009641916491091251,
|
| 1042 |
+
"learning_rate": 3.617812732490169e-05,
|
| 1043 |
+
"loss": 0.0118,
|
| 1044 |
+
"step": 3650
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.054519368723099,
|
| 1048 |
+
"grad_norm": 0.008313027210533619,
|
| 1049 |
+
"learning_rate": 3.604527579976618e-05,
|
| 1050 |
+
"loss": 0.0005,
|
| 1051 |
+
"step": 3675
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.0616929698708752,
|
| 1055 |
+
"grad_norm": 0.6244511008262634,
|
| 1056 |
+
"learning_rate": 3.591242427463067e-05,
|
| 1057 |
+
"loss": 0.0308,
|
| 1058 |
+
"step": 3700
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.0688665710186513,
|
| 1062 |
+
"grad_norm": 0.004019048996269703,
|
| 1063 |
+
"learning_rate": 3.577957274949517e-05,
|
| 1064 |
+
"loss": 0.03,
|
| 1065 |
+
"step": 3725
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.0760401721664274,
|
| 1069 |
+
"grad_norm": 0.0038138478994369507,
|
| 1070 |
+
"learning_rate": 3.564672122435966e-05,
|
| 1071 |
+
"loss": 0.0004,
|
| 1072 |
+
"step": 3750
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.0832137733142038,
|
| 1076 |
+
"grad_norm": 0.005525535438209772,
|
| 1077 |
+
"learning_rate": 3.5513869699224154e-05,
|
| 1078 |
+
"loss": 0.0003,
|
| 1079 |
+
"step": 3775
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.0903873744619799,
|
| 1083 |
+
"grad_norm": 0.004180921241641045,
|
| 1084 |
+
"learning_rate": 3.5381018174088643e-05,
|
| 1085 |
+
"loss": 0.0003,
|
| 1086 |
+
"step": 3800
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.0975609756097562,
|
| 1090 |
+
"grad_norm": 0.004629930946975946,
|
| 1091 |
+
"learning_rate": 3.5248166648953126e-05,
|
| 1092 |
+
"loss": 0.1079,
|
| 1093 |
+
"step": 3825
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.1047345767575323,
|
| 1097 |
+
"grad_norm": 0.008479510433971882,
|
| 1098 |
+
"learning_rate": 3.511531512381762e-05,
|
| 1099 |
+
"loss": 0.0788,
|
| 1100 |
+
"step": 3850
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.1119081779053084,
|
| 1104 |
+
"grad_norm": 0.007632472552359104,
|
| 1105 |
+
"learning_rate": 3.498246359868211e-05,
|
| 1106 |
+
"loss": 0.0008,
|
| 1107 |
+
"step": 3875
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.1190817790530847,
|
| 1111 |
+
"grad_norm": 0.009593753144145012,
|
| 1112 |
+
"learning_rate": 3.484961207354661e-05,
|
| 1113 |
+
"loss": 0.0006,
|
| 1114 |
+
"step": 3900
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.1262553802008608,
|
| 1118 |
+
"grad_norm": 0.00635389843955636,
|
| 1119 |
+
"learning_rate": 3.47167605484111e-05,
|
| 1120 |
+
"loss": 0.0659,
|
| 1121 |
+
"step": 3925
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.133428981348637,
|
| 1125 |
+
"grad_norm": 0.005174864083528519,
|
| 1126 |
+
"learning_rate": 3.458390902327559e-05,
|
| 1127 |
+
"loss": 0.0007,
|
| 1128 |
+
"step": 3950
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.1406025824964132,
|
| 1132 |
+
"grad_norm": 0.0196967925876379,
|
| 1133 |
+
"learning_rate": 3.4451057498140085e-05,
|
| 1134 |
+
"loss": 0.028,
|
| 1135 |
+
"step": 3975
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.1477761836441893,
|
| 1139 |
+
"grad_norm": 0.0053730374202132225,
|
| 1140 |
+
"learning_rate": 3.431820597300457e-05,
|
| 1141 |
+
"loss": 0.0044,
|
| 1142 |
+
"step": 4000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.1549497847919656,
|
| 1146 |
+
"grad_norm": 0.016190949827432632,
|
| 1147 |
+
"learning_rate": 3.4185354447869064e-05,
|
| 1148 |
+
"loss": 0.0034,
|
| 1149 |
+
"step": 4025
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.1621233859397417,
|
| 1153 |
+
"grad_norm": 0.06172305718064308,
|
| 1154 |
+
"learning_rate": 3.4052502922733554e-05,
|
| 1155 |
+
"loss": 0.045,
|
| 1156 |
+
"step": 4050
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.169296987087518,
|
| 1160 |
+
"grad_norm": 0.005435667000710964,
|
| 1161 |
+
"learning_rate": 3.3919651397598044e-05,
|
| 1162 |
+
"loss": 0.0006,
|
| 1163 |
+
"step": 4075
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 1.1764705882352942,
|
| 1167 |
+
"grad_norm": 0.0059057194739580154,
|
| 1168 |
+
"learning_rate": 3.378679987246254e-05,
|
| 1169 |
+
"loss": 0.0006,
|
| 1170 |
+
"step": 4100
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.1836441893830703,
|
| 1174 |
+
"grad_norm": 0.003536668373271823,
|
| 1175 |
+
"learning_rate": 3.365394834732703e-05,
|
| 1176 |
+
"loss": 0.0002,
|
| 1177 |
+
"step": 4125
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.1908177905308466,
|
| 1181 |
+
"grad_norm": 0.013157457113265991,
|
| 1182 |
+
"learning_rate": 3.352109682219152e-05,
|
| 1183 |
+
"loss": 0.0712,
|
| 1184 |
+
"step": 4150
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 1.1979913916786227,
|
| 1188 |
+
"grad_norm": 0.05286416783928871,
|
| 1189 |
+
"learning_rate": 3.3388245297056016e-05,
|
| 1190 |
+
"loss": 0.0032,
|
| 1191 |
+
"step": 4175
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 1.2051649928263988,
|
| 1195 |
+
"grad_norm": 0.007331141736358404,
|
| 1196 |
+
"learning_rate": 3.32553937719205e-05,
|
| 1197 |
+
"loss": 0.1212,
|
| 1198 |
+
"step": 4200
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 1.212338593974175,
|
| 1202 |
+
"grad_norm": 0.01495648454874754,
|
| 1203 |
+
"learning_rate": 3.3122542246784996e-05,
|
| 1204 |
+
"loss": 0.0026,
|
| 1205 |
+
"step": 4225
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 1.2195121951219512,
|
| 1209 |
+
"grad_norm": 0.014520620927214622,
|
| 1210 |
+
"learning_rate": 3.2989690721649485e-05,
|
| 1211 |
+
"loss": 0.011,
|
| 1212 |
+
"step": 4250
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 1.2266857962697273,
|
| 1216 |
+
"grad_norm": 0.0068002501502633095,
|
| 1217 |
+
"learning_rate": 3.2856839196513975e-05,
|
| 1218 |
+
"loss": 0.0007,
|
| 1219 |
+
"step": 4275
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.2338593974175036,
|
| 1223 |
+
"grad_norm": 0.004121303092688322,
|
| 1224 |
+
"learning_rate": 3.272398767137847e-05,
|
| 1225 |
+
"loss": 0.0313,
|
| 1226 |
+
"step": 4300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.2410329985652797,
|
| 1230 |
+
"grad_norm": 0.005836965050548315,
|
| 1231 |
+
"learning_rate": 3.259113614624296e-05,
|
| 1232 |
+
"loss": 0.0105,
|
| 1233 |
+
"step": 4325
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.248206599713056,
|
| 1237 |
+
"grad_norm": 0.004592243582010269,
|
| 1238 |
+
"learning_rate": 3.245828462110745e-05,
|
| 1239 |
+
"loss": 0.0002,
|
| 1240 |
+
"step": 4350
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 1.2553802008608321,
|
| 1244 |
+
"grad_norm": 0.01724414713680744,
|
| 1245 |
+
"learning_rate": 3.2330747156977364e-05,
|
| 1246 |
+
"loss": 0.1373,
|
| 1247 |
+
"step": 4375
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 1.2625538020086085,
|
| 1251 |
+
"grad_norm": 0.011967150494456291,
|
| 1252 |
+
"learning_rate": 3.2197895631841854e-05,
|
| 1253 |
+
"loss": 0.0277,
|
| 1254 |
+
"step": 4400
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 1.2697274031563845,
|
| 1258 |
+
"grad_norm": 0.689711332321167,
|
| 1259 |
+
"learning_rate": 3.206504410670635e-05,
|
| 1260 |
+
"loss": 0.0309,
|
| 1261 |
+
"step": 4425
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.2769010043041606,
|
| 1265 |
+
"grad_norm": 0.003829318331554532,
|
| 1266 |
+
"learning_rate": 3.193219258157083e-05,
|
| 1267 |
+
"loss": 0.0005,
|
| 1268 |
+
"step": 4450
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.284074605451937,
|
| 1272 |
+
"grad_norm": 0.005293109454214573,
|
| 1273 |
+
"learning_rate": 3.179934105643533e-05,
|
| 1274 |
+
"loss": 0.0449,
|
| 1275 |
+
"step": 4475
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.291248206599713,
|
| 1279 |
+
"grad_norm": 0.005219063721597195,
|
| 1280 |
+
"learning_rate": 3.166648953129982e-05,
|
| 1281 |
+
"loss": 0.0398,
|
| 1282 |
+
"step": 4500
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.2984218077474892,
|
| 1286 |
+
"grad_norm": 0.00711538502946496,
|
| 1287 |
+
"learning_rate": 3.153363800616431e-05,
|
| 1288 |
+
"loss": 0.0267,
|
| 1289 |
+
"step": 4525
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.3055954088952655,
|
| 1293 |
+
"grad_norm": 0.005905709229409695,
|
| 1294 |
+
"learning_rate": 3.1400786481028806e-05,
|
| 1295 |
+
"loss": 0.0577,
|
| 1296 |
+
"step": 4550
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 1.3127690100430416,
|
| 1300 |
+
"grad_norm": 0.0065200235694646835,
|
| 1301 |
+
"learning_rate": 3.1267934955893295e-05,
|
| 1302 |
+
"loss": 0.0092,
|
| 1303 |
+
"step": 4575
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.3199426111908177,
|
| 1307 |
+
"grad_norm": 0.007951737381517887,
|
| 1308 |
+
"learning_rate": 3.113508343075779e-05,
|
| 1309 |
+
"loss": 0.0103,
|
| 1310 |
+
"step": 4600
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.327116212338594,
|
| 1314 |
+
"grad_norm": 0.004364294931292534,
|
| 1315 |
+
"learning_rate": 3.1002231905622275e-05,
|
| 1316 |
+
"loss": 0.0126,
|
| 1317 |
+
"step": 4625
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 1.33428981348637,
|
| 1321 |
+
"grad_norm": 0.004744912032037973,
|
| 1322 |
+
"learning_rate": 3.0869380380486764e-05,
|
| 1323 |
+
"loss": 0.0301,
|
| 1324 |
+
"step": 4650
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 1.3414634146341464,
|
| 1328 |
+
"grad_norm": 0.006403733976185322,
|
| 1329 |
+
"learning_rate": 3.073652885535126e-05,
|
| 1330 |
+
"loss": 0.0347,
|
| 1331 |
+
"step": 4675
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.3486370157819225,
|
| 1335 |
+
"grad_norm": 0.004806953482329845,
|
| 1336 |
+
"learning_rate": 3.060367733021575e-05,
|
| 1337 |
+
"loss": 0.0629,
|
| 1338 |
+
"step": 4700
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.3558106169296988,
|
| 1342 |
+
"grad_norm": 2.9653139114379883,
|
| 1343 |
+
"learning_rate": 3.0470825805080244e-05,
|
| 1344 |
+
"loss": 0.0087,
|
| 1345 |
+
"step": 4725
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.362984218077475,
|
| 1349 |
+
"grad_norm": 0.0028481949120759964,
|
| 1350 |
+
"learning_rate": 3.0337974279944737e-05,
|
| 1351 |
+
"loss": 0.043,
|
| 1352 |
+
"step": 4750
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.370157819225251,
|
| 1356 |
+
"grad_norm": 0.005747169256210327,
|
| 1357 |
+
"learning_rate": 3.020512275480923e-05,
|
| 1358 |
+
"loss": 0.0009,
|
| 1359 |
+
"step": 4775
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.3773314203730274,
|
| 1363 |
+
"grad_norm": 2.9883158206939697,
|
| 1364 |
+
"learning_rate": 3.007227122967372e-05,
|
| 1365 |
+
"loss": 0.0257,
|
| 1366 |
+
"step": 4800
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 1.3845050215208035,
|
| 1370 |
+
"grad_norm": 0.009053130634129047,
|
| 1371 |
+
"learning_rate": 2.9939419704538206e-05,
|
| 1372 |
+
"loss": 0.0533,
|
| 1373 |
+
"step": 4825
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.3916786226685796,
|
| 1377 |
+
"grad_norm": 0.011015841737389565,
|
| 1378 |
+
"learning_rate": 2.98065681794027e-05,
|
| 1379 |
+
"loss": 0.0388,
|
| 1380 |
+
"step": 4850
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 1.3988522238163559,
|
| 1384 |
+
"grad_norm": 0.007833955809473991,
|
| 1385 |
+
"learning_rate": 2.9673716654267192e-05,
|
| 1386 |
+
"loss": 0.001,
|
| 1387 |
+
"step": 4875
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.406025824964132,
|
| 1391 |
+
"grad_norm": 0.0066580744460225105,
|
| 1392 |
+
"learning_rate": 2.9540865129131685e-05,
|
| 1393 |
+
"loss": 0.0008,
|
| 1394 |
+
"step": 4900
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.413199426111908,
|
| 1398 |
+
"grad_norm": 0.004744118079543114,
|
| 1399 |
+
"learning_rate": 2.9408013603996175e-05,
|
| 1400 |
+
"loss": 0.0361,
|
| 1401 |
+
"step": 4925
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 1.4203730272596844,
|
| 1405 |
+
"grad_norm": 0.017294887453317642,
|
| 1406 |
+
"learning_rate": 2.9275162078860668e-05,
|
| 1407 |
+
"loss": 0.0003,
|
| 1408 |
+
"step": 4950
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.4275466284074605,
|
| 1412 |
+
"grad_norm": 0.019565369933843613,
|
| 1413 |
+
"learning_rate": 2.914231055372516e-05,
|
| 1414 |
+
"loss": 0.0066,
|
| 1415 |
+
"step": 4975
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 1.4347202295552366,
|
| 1419 |
+
"grad_norm": 0.005781451240181923,
|
| 1420 |
+
"learning_rate": 2.9009459028589647e-05,
|
| 1421 |
+
"loss": 0.0807,
|
| 1422 |
+
"step": 5000
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 1.441893830703013,
|
| 1426 |
+
"grad_norm": 0.018371548503637314,
|
| 1427 |
+
"learning_rate": 2.887660750345414e-05,
|
| 1428 |
+
"loss": 0.0383,
|
| 1429 |
+
"step": 5025
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.4490674318507892,
|
| 1433 |
+
"grad_norm": 0.019642187282443047,
|
| 1434 |
+
"learning_rate": 2.874375597831863e-05,
|
| 1435 |
+
"loss": 0.0006,
|
| 1436 |
+
"step": 5050
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.4562410329985653,
|
| 1440 |
+
"grad_norm": 0.016255564987659454,
|
| 1441 |
+
"learning_rate": 2.8610904453183123e-05,
|
| 1442 |
+
"loss": 0.0004,
|
| 1443 |
+
"step": 5075
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 1.4634146341463414,
|
| 1447 |
+
"grad_norm": 6.259925365447998,
|
| 1448 |
+
"learning_rate": 2.8478052928047616e-05,
|
| 1449 |
+
"loss": 0.1058,
|
| 1450 |
+
"step": 5100
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.4705882352941178,
|
| 1454 |
+
"grad_norm": 0.028038620948791504,
|
| 1455 |
+
"learning_rate": 2.834520140291211e-05,
|
| 1456 |
+
"loss": 0.0848,
|
| 1457 |
+
"step": 5125
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 1.4777618364418939,
|
| 1461 |
+
"grad_norm": 18.35304832458496,
|
| 1462 |
+
"learning_rate": 2.82123498777766e-05,
|
| 1463 |
+
"loss": 0.0546,
|
| 1464 |
+
"step": 5150
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 1.48493543758967,
|
| 1468 |
+
"grad_norm": 0.005022502038627863,
|
| 1469 |
+
"learning_rate": 2.8079498352641086e-05,
|
| 1470 |
+
"loss": 0.0265,
|
| 1471 |
+
"step": 5175
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.4921090387374463,
|
| 1475 |
+
"grad_norm": 0.011426527984440327,
|
| 1476 |
+
"learning_rate": 2.794664682750558e-05,
|
| 1477 |
+
"loss": 0.017,
|
| 1478 |
+
"step": 5200
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.4992826398852224,
|
| 1482 |
+
"grad_norm": 0.02600315399467945,
|
| 1483 |
+
"learning_rate": 2.7813795302370072e-05,
|
| 1484 |
+
"loss": 0.001,
|
| 1485 |
+
"step": 5225
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 1.5064562410329985,
|
| 1489 |
+
"grad_norm": 0.0031450982205569744,
|
| 1490 |
+
"learning_rate": 2.7680943777234565e-05,
|
| 1491 |
+
"loss": 0.0296,
|
| 1492 |
+
"step": 5250
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.5136298421807748,
|
| 1496 |
+
"grad_norm": 0.00543051864951849,
|
| 1497 |
+
"learning_rate": 2.7548092252099055e-05,
|
| 1498 |
+
"loss": 0.0004,
|
| 1499 |
+
"step": 5275
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 1.5208034433285509,
|
| 1503 |
+
"grad_norm": 0.007613678928464651,
|
| 1504 |
+
"learning_rate": 2.7415240726963548e-05,
|
| 1505 |
+
"loss": 0.0003,
|
| 1506 |
+
"step": 5300
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 1.527977044476327,
|
| 1510 |
+
"grad_norm": 0.00250364001840353,
|
| 1511 |
+
"learning_rate": 2.728238920182804e-05,
|
| 1512 |
+
"loss": 0.0217,
|
| 1513 |
+
"step": 5325
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.5351506456241033,
|
| 1517 |
+
"grad_norm": 0.0024825434666126966,
|
| 1518 |
+
"learning_rate": 2.7149537676692534e-05,
|
| 1519 |
+
"loss": 0.0442,
|
| 1520 |
+
"step": 5350
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.5423242467718796,
|
| 1524 |
+
"grad_norm": 0.00660035852342844,
|
| 1525 |
+
"learning_rate": 2.701668615155702e-05,
|
| 1526 |
+
"loss": 0.0002,
|
| 1527 |
+
"step": 5375
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.5494978479196555,
|
| 1531 |
+
"grad_norm": 0.0030255629681050777,
|
| 1532 |
+
"learning_rate": 2.688383462642151e-05,
|
| 1533 |
+
"loss": 0.001,
|
| 1534 |
+
"step": 5400
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.5566714490674318,
|
| 1538 |
+
"grad_norm": 56.207244873046875,
|
| 1539 |
+
"learning_rate": 2.6750983101286003e-05,
|
| 1540 |
+
"loss": 0.0662,
|
| 1541 |
+
"step": 5425
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.5638450502152081,
|
| 1545 |
+
"grad_norm": 0.005194812081754208,
|
| 1546 |
+
"learning_rate": 2.6618131576150496e-05,
|
| 1547 |
+
"loss": 0.0004,
|
| 1548 |
+
"step": 5450
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 1.5710186513629842,
|
| 1552 |
+
"grad_norm": 0.018478918820619583,
|
| 1553 |
+
"learning_rate": 2.648528005101499e-05,
|
| 1554 |
+
"loss": 0.0786,
|
| 1555 |
+
"step": 5475
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.5781922525107603,
|
| 1559 |
+
"grad_norm": 67.21288299560547,
|
| 1560 |
+
"learning_rate": 2.635242852587948e-05,
|
| 1561 |
+
"loss": 0.0224,
|
| 1562 |
+
"step": 5500
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 1.5853658536585367,
|
| 1566 |
+
"grad_norm": 0.005518095102161169,
|
| 1567 |
+
"learning_rate": 2.6219577000743972e-05,
|
| 1568 |
+
"loss": 0.0005,
|
| 1569 |
+
"step": 5525
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.5925394548063128,
|
| 1573 |
+
"grad_norm": 0.013573232106864452,
|
| 1574 |
+
"learning_rate": 2.6086725475608458e-05,
|
| 1575 |
+
"loss": 0.0703,
|
| 1576 |
+
"step": 5550
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.5997130559540889,
|
| 1580 |
+
"grad_norm": 0.022108446806669235,
|
| 1581 |
+
"learning_rate": 2.595387395047295e-05,
|
| 1582 |
+
"loss": 0.0127,
|
| 1583 |
+
"step": 5575
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.6068866571018652,
|
| 1587 |
+
"grad_norm": 0.02520466223359108,
|
| 1588 |
+
"learning_rate": 2.5821022425337444e-05,
|
| 1589 |
+
"loss": 0.0908,
|
| 1590 |
+
"step": 5600
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 1.6140602582496413,
|
| 1594 |
+
"grad_norm": 0.06917204707860947,
|
| 1595 |
+
"learning_rate": 2.5688170900201934e-05,
|
| 1596 |
+
"loss": 0.026,
|
| 1597 |
+
"step": 5625
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.6212338593974174,
|
| 1601 |
+
"grad_norm": 0.011066574603319168,
|
| 1602 |
+
"learning_rate": 2.5555319375066427e-05,
|
| 1603 |
+
"loss": 0.021,
|
| 1604 |
+
"step": 5650
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 1.6284074605451937,
|
| 1608 |
+
"grad_norm": 0.004337651655077934,
|
| 1609 |
+
"learning_rate": 2.542246784993092e-05,
|
| 1610 |
+
"loss": 0.0004,
|
| 1611 |
+
"step": 5675
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.63558106169297,
|
| 1615 |
+
"grad_norm": 0.007177389692515135,
|
| 1616 |
+
"learning_rate": 2.5289616324795413e-05,
|
| 1617 |
+
"loss": 0.0432,
|
| 1618 |
+
"step": 5700
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.642754662840746,
|
| 1622 |
+
"grad_norm": 0.004365015309303999,
|
| 1623 |
+
"learning_rate": 2.51567647996599e-05,
|
| 1624 |
+
"loss": 0.0003,
|
| 1625 |
+
"step": 5725
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.6499282639885222,
|
| 1629 |
+
"grad_norm": 0.005093116778880358,
|
| 1630 |
+
"learning_rate": 2.502391327452439e-05,
|
| 1631 |
+
"loss": 0.0006,
|
| 1632 |
+
"step": 5750
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 1.6571018651362985,
|
| 1636 |
+
"grad_norm": 0.00538448104634881,
|
| 1637 |
+
"learning_rate": 2.4891061749388883e-05,
|
| 1638 |
+
"loss": 0.0244,
|
| 1639 |
+
"step": 5775
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.6642754662840746,
|
| 1643 |
+
"grad_norm": 0.0028270191978663206,
|
| 1644 |
+
"learning_rate": 2.4758210224253376e-05,
|
| 1645 |
+
"loss": 0.0003,
|
| 1646 |
+
"step": 5800
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 1.6714490674318507,
|
| 1650 |
+
"grad_norm": 0.004735349677503109,
|
| 1651 |
+
"learning_rate": 2.462535869911787e-05,
|
| 1652 |
+
"loss": 0.0006,
|
| 1653 |
+
"step": 5825
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.678622668579627,
|
| 1657 |
+
"grad_norm": 0.002434446709230542,
|
| 1658 |
+
"learning_rate": 2.449250717398236e-05,
|
| 1659 |
+
"loss": 0.0348,
|
| 1660 |
+
"step": 5850
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.6857962697274032,
|
| 1664 |
+
"grad_norm": 9.833294868469238,
|
| 1665 |
+
"learning_rate": 2.4359655648846848e-05,
|
| 1666 |
+
"loss": 0.0005,
|
| 1667 |
+
"step": 5875
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.6929698708751793,
|
| 1671 |
+
"grad_norm": 0.002602140884846449,
|
| 1672 |
+
"learning_rate": 2.422680412371134e-05,
|
| 1673 |
+
"loss": 0.0311,
|
| 1674 |
+
"step": 5900
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 1.7001434720229556,
|
| 1678 |
+
"grad_norm": 0.006632608827203512,
|
| 1679 |
+
"learning_rate": 2.4093952598575834e-05,
|
| 1680 |
+
"loss": 0.0883,
|
| 1681 |
+
"step": 5925
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.7073170731707317,
|
| 1685 |
+
"grad_norm": 0.009323637932538986,
|
| 1686 |
+
"learning_rate": 2.3961101073440324e-05,
|
| 1687 |
+
"loss": 0.0409,
|
| 1688 |
+
"step": 5950
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 1.7144906743185078,
|
| 1692 |
+
"grad_norm": 0.012521679513156414,
|
| 1693 |
+
"learning_rate": 2.3828249548304814e-05,
|
| 1694 |
+
"loss": 0.0282,
|
| 1695 |
+
"step": 5975
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 1.721664275466284,
|
| 1699 |
+
"grad_norm": 0.008466576226055622,
|
| 1700 |
+
"learning_rate": 2.3695398023169307e-05,
|
| 1701 |
+
"loss": 0.0244,
|
| 1702 |
+
"step": 6000
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 1.7288378766140604,
|
| 1706 |
+
"grad_norm": 0.010361541993916035,
|
| 1707 |
+
"learning_rate": 2.35625464980338e-05,
|
| 1708 |
+
"loss": 0.0356,
|
| 1709 |
+
"step": 6025
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.7360114777618363,
|
| 1713 |
+
"grad_norm": 0.012751937843859196,
|
| 1714 |
+
"learning_rate": 2.342969497289829e-05,
|
| 1715 |
+
"loss": 0.0847,
|
| 1716 |
+
"step": 6050
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 1.7431850789096126,
|
| 1720 |
+
"grad_norm": 0.03529978543519974,
|
| 1721 |
+
"learning_rate": 2.3296843447762783e-05,
|
| 1722 |
+
"loss": 0.0384,
|
| 1723 |
+
"step": 6075
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 1.750358680057389,
|
| 1727 |
+
"grad_norm": 0.026526516303420067,
|
| 1728 |
+
"learning_rate": 2.3163991922627273e-05,
|
| 1729 |
+
"loss": 0.0334,
|
| 1730 |
+
"step": 6100
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 1.757532281205165,
|
| 1734 |
+
"grad_norm": 0.008869277313351631,
|
| 1735 |
+
"learning_rate": 2.3031140397491766e-05,
|
| 1736 |
+
"loss": 0.0092,
|
| 1737 |
+
"step": 6125
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 1.7647058823529411,
|
| 1741 |
+
"grad_norm": 21.071659088134766,
|
| 1742 |
+
"learning_rate": 2.2898288872356255e-05,
|
| 1743 |
+
"loss": 0.049,
|
| 1744 |
+
"step": 6150
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 1.7718794835007174,
|
| 1748 |
+
"grad_norm": 0.00842801108956337,
|
| 1749 |
+
"learning_rate": 2.276543734722075e-05,
|
| 1750 |
+
"loss": 0.0406,
|
| 1751 |
+
"step": 6175
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.7790530846484935,
|
| 1755 |
+
"grad_norm": 0.009631357155740261,
|
| 1756 |
+
"learning_rate": 2.2632585822085238e-05,
|
| 1757 |
+
"loss": 0.0033,
|
| 1758 |
+
"step": 6200
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 1.7862266857962696,
|
| 1762 |
+
"grad_norm": 0.015189352445304394,
|
| 1763 |
+
"learning_rate": 2.2499734296949728e-05,
|
| 1764 |
+
"loss": 0.0981,
|
| 1765 |
+
"step": 6225
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 1.793400286944046,
|
| 1769 |
+
"grad_norm": 0.13944695889949799,
|
| 1770 |
+
"learning_rate": 2.236688277181422e-05,
|
| 1771 |
+
"loss": 0.0003,
|
| 1772 |
+
"step": 6250
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 1.800573888091822,
|
| 1776 |
+
"grad_norm": 0.001980824628844857,
|
| 1777 |
+
"learning_rate": 2.2234031246678714e-05,
|
| 1778 |
+
"loss": 0.0004,
|
| 1779 |
+
"step": 6275
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 1.8077474892395982,
|
| 1783 |
+
"grad_norm": 0.005285199265927076,
|
| 1784 |
+
"learning_rate": 2.2101179721543204e-05,
|
| 1785 |
+
"loss": 0.0044,
|
| 1786 |
+
"step": 6300
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 1.8149210903873745,
|
| 1790 |
+
"grad_norm": 9.718914031982422,
|
| 1791 |
+
"learning_rate": 2.1968328196407693e-05,
|
| 1792 |
+
"loss": 0.0377,
|
| 1793 |
+
"step": 6325
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.8220946915351508,
|
| 1797 |
+
"grad_norm": 0.0019531676080077887,
|
| 1798 |
+
"learning_rate": 2.1835476671272187e-05,
|
| 1799 |
+
"loss": 0.0093,
|
| 1800 |
+
"step": 6350
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 1.8292682926829267,
|
| 1804 |
+
"grad_norm": 0.0016301467549055815,
|
| 1805 |
+
"learning_rate": 2.170262514613668e-05,
|
| 1806 |
+
"loss": 0.0015,
|
| 1807 |
+
"step": 6375
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 1.836441893830703,
|
| 1811 |
+
"grad_norm": 44.472381591796875,
|
| 1812 |
+
"learning_rate": 2.157508768200659e-05,
|
| 1813 |
+
"loss": 0.1747,
|
| 1814 |
+
"step": 6400
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 1.8436154949784793,
|
| 1818 |
+
"grad_norm": 0.00314929592423141,
|
| 1819 |
+
"learning_rate": 2.1442236156871082e-05,
|
| 1820 |
+
"loss": 0.0216,
|
| 1821 |
+
"step": 6425
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 1.8507890961262554,
|
| 1825 |
+
"grad_norm": 0.003403552807867527,
|
| 1826 |
+
"learning_rate": 2.1309384631735572e-05,
|
| 1827 |
+
"loss": 0.0021,
|
| 1828 |
+
"step": 6450
|
| 1829 |
+
},
|
| 1830 |
+
{
|
| 1831 |
+
"epoch": 1.8579626972740315,
|
| 1832 |
+
"grad_norm": 0.007590115536004305,
|
| 1833 |
+
"learning_rate": 2.1176533106600065e-05,
|
| 1834 |
+
"loss": 0.0012,
|
| 1835 |
+
"step": 6475
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 1.8651362984218078,
|
| 1839 |
+
"grad_norm": 0.0027949281502515078,
|
| 1840 |
+
"learning_rate": 2.1043681581464555e-05,
|
| 1841 |
+
"loss": 0.1361,
|
| 1842 |
+
"step": 6500
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 1.872309899569584,
|
| 1846 |
+
"grad_norm": 0.00557307805866003,
|
| 1847 |
+
"learning_rate": 2.0910830056329048e-05,
|
| 1848 |
+
"loss": 0.0365,
|
| 1849 |
+
"step": 6525
|
| 1850 |
+
},
|
| 1851 |
+
{
|
| 1852 |
+
"epoch": 1.87948350071736,
|
| 1853 |
+
"grad_norm": 0.010457979515194893,
|
| 1854 |
+
"learning_rate": 2.077797853119354e-05,
|
| 1855 |
+
"loss": 0.0358,
|
| 1856 |
+
"step": 6550
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 1.8866571018651364,
|
| 1860 |
+
"grad_norm": 0.003813342424109578,
|
| 1861 |
+
"learning_rate": 2.064512700605803e-05,
|
| 1862 |
+
"loss": 0.0321,
|
| 1863 |
+
"step": 6575
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 1.8938307030129125,
|
| 1867 |
+
"grad_norm": 0.00262393313460052,
|
| 1868 |
+
"learning_rate": 2.051227548092252e-05,
|
| 1869 |
+
"loss": 0.022,
|
| 1870 |
+
"step": 6600
|
| 1871 |
+
},
|
| 1872 |
+
{
|
| 1873 |
+
"epoch": 1.9010043041606886,
|
| 1874 |
+
"grad_norm": 0.004526357166469097,
|
| 1875 |
+
"learning_rate": 2.0379423955787014e-05,
|
| 1876 |
+
"loss": 0.0424,
|
| 1877 |
+
"step": 6625
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 1.9081779053084649,
|
| 1881 |
+
"grad_norm": 0.01698668673634529,
|
| 1882 |
+
"learning_rate": 2.0246572430651507e-05,
|
| 1883 |
+
"loss": 0.0009,
|
| 1884 |
+
"step": 6650
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 1.9153515064562412,
|
| 1888 |
+
"grad_norm": 0.0501057431101799,
|
| 1889 |
+
"learning_rate": 2.0113720905515997e-05,
|
| 1890 |
+
"loss": 0.0237,
|
| 1891 |
+
"step": 6675
|
| 1892 |
+
},
|
| 1893 |
+
{
|
| 1894 |
+
"epoch": 1.922525107604017,
|
| 1895 |
+
"grad_norm": 0.007618068717420101,
|
| 1896 |
+
"learning_rate": 1.9980869380380486e-05,
|
| 1897 |
+
"loss": 0.0192,
|
| 1898 |
+
"step": 6700
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 1.9296987087517934,
|
| 1902 |
+
"grad_norm": 0.01174473762512207,
|
| 1903 |
+
"learning_rate": 1.984801785524498e-05,
|
| 1904 |
+
"loss": 0.0334,
|
| 1905 |
+
"step": 6725
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 1.9368723098995697,
|
| 1909 |
+
"grad_norm": 25.125062942504883,
|
| 1910 |
+
"learning_rate": 1.9715166330109472e-05,
|
| 1911 |
+
"loss": 0.0098,
|
| 1912 |
+
"step": 6750
|
| 1913 |
+
},
|
| 1914 |
+
{
|
| 1915 |
+
"epoch": 1.9440459110473458,
|
| 1916 |
+
"grad_norm": 0.0056104715913534164,
|
| 1917 |
+
"learning_rate": 1.9582314804973962e-05,
|
| 1918 |
+
"loss": 0.1462,
|
| 1919 |
+
"step": 6775
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 1.951219512195122,
|
| 1923 |
+
"grad_norm": 0.05587226152420044,
|
| 1924 |
+
"learning_rate": 1.9449463279838452e-05,
|
| 1925 |
+
"loss": 0.081,
|
| 1926 |
+
"step": 6800
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 1.9583931133428982,
|
| 1930 |
+
"grad_norm": 0.01742021180689335,
|
| 1931 |
+
"learning_rate": 1.9316611754702945e-05,
|
| 1932 |
+
"loss": 0.0418,
|
| 1933 |
+
"step": 6825
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"epoch": 1.9655667144906743,
|
| 1937 |
+
"grad_norm": 0.45866331458091736,
|
| 1938 |
+
"learning_rate": 1.9183760229567435e-05,
|
| 1939 |
+
"loss": 0.001,
|
| 1940 |
+
"step": 6850
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 1.9727403156384504,
|
| 1944 |
+
"grad_norm": 0.04859813675284386,
|
| 1945 |
+
"learning_rate": 1.9050908704431928e-05,
|
| 1946 |
+
"loss": 0.0302,
|
| 1947 |
+
"step": 6875
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 1.9799139167862267,
|
| 1951 |
+
"grad_norm": 0.015134917572140694,
|
| 1952 |
+
"learning_rate": 1.891805717929642e-05,
|
| 1953 |
+
"loss": 0.0418,
|
| 1954 |
+
"step": 6900
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 1.9870875179340028,
|
| 1958 |
+
"grad_norm": 0.02702825888991356,
|
| 1959 |
+
"learning_rate": 1.878520565416091e-05,
|
| 1960 |
+
"loss": 0.0419,
|
| 1961 |
+
"step": 6925
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 1.994261119081779,
|
| 1965 |
+
"grad_norm": 0.023256802931427956,
|
| 1966 |
+
"learning_rate": 1.86523541290254e-05,
|
| 1967 |
+
"loss": 0.0679,
|
| 1968 |
+
"step": 6950
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 2.0,
|
| 1972 |
+
"eval_accuracy": 0.9899935442220787,
|
| 1973 |
+
"eval_auc": 0.999704926354536,
|
| 1974 |
+
"eval_f1": 0.9900353584056574,
|
| 1975 |
+
"eval_loss": 0.04012129828333855,
|
| 1976 |
+
"eval_precision": 0.9859154929577465,
|
| 1977 |
+
"eval_recall": 0.9941897998708844,
|
| 1978 |
+
"eval_runtime": 4.2712,
|
| 1979 |
+
"eval_samples_per_second": 725.331,
|
| 1980 |
+
"eval_steps_per_second": 45.421,
|
| 1981 |
+
"step": 6970
|
| 1982 |
+
}
|
| 1983 |
+
],
|
| 1984 |
+
"logging_steps": 25,
|
| 1985 |
+
"max_steps": 10455,
|
| 1986 |
+
"num_input_tokens_seen": 0,
|
| 1987 |
+
"num_train_epochs": 3,
|
| 1988 |
+
"save_steps": 500,
|
| 1989 |
+
"stateful_callbacks": {
|
| 1990 |
+
"EarlyStoppingCallback": {
|
| 1991 |
+
"args": {
|
| 1992 |
+
"early_stopping_patience": 5,
|
| 1993 |
+
"early_stopping_threshold": 0.01
|
| 1994 |
+
},
|
| 1995 |
+
"attributes": {
|
| 1996 |
+
"early_stopping_patience_counter": 1
|
| 1997 |
+
}
|
| 1998 |
+
},
|
| 1999 |
+
"TrainerControl": {
|
| 2000 |
+
"args": {
|
| 2001 |
+
"should_epoch_stop": false,
|
| 2002 |
+
"should_evaluate": false,
|
| 2003 |
+
"should_log": false,
|
| 2004 |
+
"should_save": true,
|
| 2005 |
+
"should_training_stop": false
|
| 2006 |
+
},
|
| 2007 |
+
"attributes": {}
|
| 2008 |
+
}
|
| 2009 |
+
},
|
| 2010 |
+
"total_flos": 462331386716160.0,
|
| 2011 |
+
"train_batch_size": 8,
|
| 2012 |
+
"trial_name": null,
|
| 2013 |
+
"trial_params": null
|
| 2014 |
+
}
|
checkpoint-6970/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc076067c3e486e2d8f21770bca521a80d83b3dd7aa04905e6351e987c45120c
|
| 3 |
+
size 5368
|
config.json
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "TaylorAI/gte-tiny",
|
| 3 |
+
"_num_labels": 2,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 384,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "human",
|
| 14 |
+
"1": "llm"
|
| 15 |
+
},
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 1536,
|
| 18 |
+
"label2id": {
|
| 19 |
+
"human": 0,
|
| 20 |
+
"llm": 1
|
| 21 |
+
},
|
| 22 |
+
"layer_norm_eps": 1e-12,
|
| 23 |
+
"max_position_embeddings": 512,
|
| 24 |
+
"model_type": "bert",
|
| 25 |
+
"num_attention_heads": 12,
|
| 26 |
+
"num_hidden_layers": 6,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"problem_type": "single_label_classification",
|
| 30 |
+
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.48.0",
|
| 32 |
+
"type_vocab_size": 2,
|
| 33 |
+
"use_cache": true,
|
| 34 |
+
"vocab_size": 30522
|
| 35 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1dc3603f7927d59bbb28381542e7c65f0c7d40444c17d18e2bdefc35a0d77c0
|
| 3 |
+
size 90867952
|
runs/May30_00-09-04_r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w/events.out.tfevents.1748563746.r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w.75.0
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 sha256:9ff7b0a124a2d68f7bff301a035bc548c9cfe8a0ce636fb1b5ca61415afcb921
|
| 3 |
+
size 95276
|
runs/May30_00-09-04_r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w/events.out.tfevents.1748564133.r-trentmkelly-autotrain-advanced-9f8hpchj-c1d5e-s6v1w.75.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd156ae9be79077801d670663261bd82c8f34622e6471e5d2c579df41bb32179
|
| 3 |
+
size 607
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"[PAD]",
|
| 4 |
+
"[UNK]",
|
| 5 |
+
"[CLS]",
|
| 6 |
+
"[SEP]",
|
| 7 |
+
"[MASK]"
|
| 8 |
+
],
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "[CLS]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"mask_token": {
|
| 17 |
+
"content": "[MASK]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"pad_token": {
|
| 24 |
+
"content": "[PAD]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"sep_token": {
|
| 31 |
+
"content": "[SEP]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"unk_token": {
|
| 38 |
+
"content": "[UNK]",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
}
|
| 44 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"additional_special_tokens": [
|
| 45 |
+
"[PAD]",
|
| 46 |
+
"[UNK]",
|
| 47 |
+
"[CLS]",
|
| 48 |
+
"[SEP]",
|
| 49 |
+
"[MASK]"
|
| 50 |
+
],
|
| 51 |
+
"clean_up_tokenization_spaces": true,
|
| 52 |
+
"cls_token": "[CLS]",
|
| 53 |
+
"do_basic_tokenize": true,
|
| 54 |
+
"do_lower_case": true,
|
| 55 |
+
"extra_special_tokens": {},
|
| 56 |
+
"mask_token": "[MASK]",
|
| 57 |
+
"max_length": 128,
|
| 58 |
+
"model_max_length": 512,
|
| 59 |
+
"never_split": null,
|
| 60 |
+
"pad_to_multiple_of": null,
|
| 61 |
+
"pad_token": "[PAD]",
|
| 62 |
+
"pad_token_type_id": 0,
|
| 63 |
+
"padding_side": "right",
|
| 64 |
+
"sep_token": "[SEP]",
|
| 65 |
+
"stride": 0,
|
| 66 |
+
"strip_accents": null,
|
| 67 |
+
"tokenize_chinese_chars": true,
|
| 68 |
+
"tokenizer_class": "BertTokenizer",
|
| 69 |
+
"truncation_side": "right",
|
| 70 |
+
"truncation_strategy": "longest_first",
|
| 71 |
+
"unk_token": "[UNK]"
|
| 72 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc076067c3e486e2d8f21770bca521a80d83b3dd7aa04905e6351e987c45120c
|
| 3 |
+
size 5368
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "slop-detector-mini/autotrain-data",
|
| 3 |
+
"model": "TaylorAI/gte-tiny",
|
| 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-mini",
|
| 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
|
|
|