Upload folder using huggingface_hub
Browse files- README.md +36 -0
- checkpoint-5750/config.json +78 -0
- checkpoint-5750/model.safetensors +3 -0
- checkpoint-5750/optimizer.pt +3 -0
- checkpoint-5750/rng_state.pth +3 -0
- checkpoint-5750/scheduler.pt +3 -0
- checkpoint-5750/trainer_state.json +1688 -0
- checkpoint-5750/training_args.bin +3 -0
- config.json +78 -0
- model.safetensors +3 -0
- runs/Aug16_02-58-16_r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m/events.out.tfevents.1723777098.r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m.81.0 +2 -2
- runs/Aug16_02-58-16_r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m/events.out.tfevents.1723777914.r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m.81.1 +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +55 -0
- training_args.bin +3 -0
- training_params.json +30 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
tags:
|
| 4 |
+
- autotrain
|
| 5 |
+
- text-classification
|
| 6 |
+
base_model: google-bert/bert-base-uncased
|
| 7 |
+
widget:
|
| 8 |
+
- text: "I love AutoTrain"
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Trained Using AutoTrain
|
| 12 |
+
|
| 13 |
+
- Problem type: Text Classification
|
| 14 |
+
|
| 15 |
+
## Validation Metrics
|
| 16 |
+
loss: 0.00012181053898530081
|
| 17 |
+
|
| 18 |
+
f1_macro: 1.0
|
| 19 |
+
|
| 20 |
+
f1_micro: 1.0
|
| 21 |
+
|
| 22 |
+
f1_weighted: 1.0
|
| 23 |
+
|
| 24 |
+
precision_macro: 1.0
|
| 25 |
+
|
| 26 |
+
precision_micro: 1.0
|
| 27 |
+
|
| 28 |
+
precision_weighted: 1.0
|
| 29 |
+
|
| 30 |
+
recall_macro: 1.0
|
| 31 |
+
|
| 32 |
+
recall_micro: 1.0
|
| 33 |
+
|
| 34 |
+
recall_weighted: 1.0
|
| 35 |
+
|
| 36 |
+
accuracy: 1.0
|
checkpoint-5750/config.json
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google-bert/bert-base-uncased",
|
| 3 |
+
"_num_labels": 23,
|
| 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": "BirthDate",
|
| 15 |
+
"1": "CountryCode",
|
| 16 |
+
"2": "EmailAddress",
|
| 17 |
+
"3": "FullPersonalName",
|
| 18 |
+
"4": "Gender",
|
| 19 |
+
"5": "GivenName",
|
| 20 |
+
"6": "IpAddress",
|
| 21 |
+
"7": "Latitude",
|
| 22 |
+
"8": "Locality",
|
| 23 |
+
"9": "Longitude",
|
| 24 |
+
"10": "MiddleName",
|
| 25 |
+
"11": "NationalIdentificationNumber",
|
| 26 |
+
"12": "OrganizationName",
|
| 27 |
+
"13": "PaymentCardNumber",
|
| 28 |
+
"14": "PostalCode",
|
| 29 |
+
"15": "PrimaryStreetAddress",
|
| 30 |
+
"16": "SecondaryStreetAddress",
|
| 31 |
+
"17": "Subdivision",
|
| 32 |
+
"18": "Surname",
|
| 33 |
+
"19": "TelephoneNumber",
|
| 34 |
+
"20": "Uncategorized",
|
| 35 |
+
"21": "UniqueIdentificationNumber",
|
| 36 |
+
"22": "Uuid"
|
| 37 |
+
},
|
| 38 |
+
"initializer_range": 0.02,
|
| 39 |
+
"intermediate_size": 3072,
|
| 40 |
+
"label2id": {
|
| 41 |
+
"BirthDate": 0,
|
| 42 |
+
"CountryCode": 1,
|
| 43 |
+
"EmailAddress": 2,
|
| 44 |
+
"FullPersonalName": 3,
|
| 45 |
+
"Gender": 4,
|
| 46 |
+
"GivenName": 5,
|
| 47 |
+
"IpAddress": 6,
|
| 48 |
+
"Latitude": 7,
|
| 49 |
+
"Locality": 8,
|
| 50 |
+
"Longitude": 9,
|
| 51 |
+
"MiddleName": 10,
|
| 52 |
+
"NationalIdentificationNumber": 11,
|
| 53 |
+
"OrganizationName": 12,
|
| 54 |
+
"PaymentCardNumber": 13,
|
| 55 |
+
"PostalCode": 14,
|
| 56 |
+
"PrimaryStreetAddress": 15,
|
| 57 |
+
"SecondaryStreetAddress": 16,
|
| 58 |
+
"Subdivision": 17,
|
| 59 |
+
"Surname": 18,
|
| 60 |
+
"TelephoneNumber": 19,
|
| 61 |
+
"Uncategorized": 20,
|
| 62 |
+
"UniqueIdentificationNumber": 21,
|
| 63 |
+
"Uuid": 22
|
| 64 |
+
},
|
| 65 |
+
"layer_norm_eps": 1e-12,
|
| 66 |
+
"max_position_embeddings": 512,
|
| 67 |
+
"model_type": "bert",
|
| 68 |
+
"num_attention_heads": 12,
|
| 69 |
+
"num_hidden_layers": 12,
|
| 70 |
+
"pad_token_id": 0,
|
| 71 |
+
"position_embedding_type": "absolute",
|
| 72 |
+
"problem_type": "single_label_classification",
|
| 73 |
+
"torch_dtype": "float32",
|
| 74 |
+
"transformers_version": "4.44.0",
|
| 75 |
+
"type_vocab_size": 2,
|
| 76 |
+
"use_cache": true,
|
| 77 |
+
"vocab_size": 30522
|
| 78 |
+
}
|
checkpoint-5750/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84f5a6008b9ad36981f508eb2908e9ee9a46225e32728f516f993b59622b6890
|
| 3 |
+
size 438023244
|
checkpoint-5750/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c02c8b7f9f75be5a02d0ab66505cecc797f176e2e798bd67798d1a5f19f33a54
|
| 3 |
+
size 876167546
|
checkpoint-5750/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81e0e32bf937891d9171e5c98121c0cbe0c0ec4729b0210c84f2ba1795f9141d
|
| 3 |
+
size 14244
|
checkpoint-5750/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a1768aaf9575a6165fb52e9186f755bf7074d6f8730eac54bf39bea49c06d17
|
| 3 |
+
size 1064
|
checkpoint-5750/trainer_state.json
ADDED
|
@@ -0,0 +1,1688 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.00012181053898530081,
|
| 3 |
+
"best_model_checkpoint": "autotrain-l9v4l-l3gs1/checkpoint-5750",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5750,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.008695652173913044,
|
| 13 |
+
"grad_norm": 8.19902229309082,
|
| 14 |
+
"learning_rate": 1.3904982618771727e-06,
|
| 15 |
+
"loss": 3.1857,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.017391304347826087,
|
| 20 |
+
"grad_norm": 9.214441299438477,
|
| 21 |
+
"learning_rate": 2.838933951332561e-06,
|
| 22 |
+
"loss": 3.1368,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02608695652173913,
|
| 27 |
+
"grad_norm": 9.295684814453125,
|
| 28 |
+
"learning_rate": 4.229432213209733e-06,
|
| 29 |
+
"loss": 3.0863,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.034782608695652174,
|
| 34 |
+
"grad_norm": 15.308918952941895,
|
| 35 |
+
"learning_rate": 5.677867902665122e-06,
|
| 36 |
+
"loss": 2.9409,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.043478260869565216,
|
| 41 |
+
"grad_norm": 14.55487060546875,
|
| 42 |
+
"learning_rate": 7.12630359212051e-06,
|
| 43 |
+
"loss": 2.8279,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05217391304347826,
|
| 48 |
+
"grad_norm": 9.613283157348633,
|
| 49 |
+
"learning_rate": 8.574739281575898e-06,
|
| 50 |
+
"loss": 2.6909,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06086956521739131,
|
| 55 |
+
"grad_norm": 11.267277717590332,
|
| 56 |
+
"learning_rate": 1.0023174971031286e-05,
|
| 57 |
+
"loss": 2.5096,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.06956521739130435,
|
| 62 |
+
"grad_norm": 8.382554054260254,
|
| 63 |
+
"learning_rate": 1.1471610660486674e-05,
|
| 64 |
+
"loss": 2.2595,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0782608695652174,
|
| 69 |
+
"grad_norm": 10.061994552612305,
|
| 70 |
+
"learning_rate": 1.2920046349942064e-05,
|
| 71 |
+
"loss": 2.003,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.08695652173913043,
|
| 76 |
+
"grad_norm": 10.652369499206543,
|
| 77 |
+
"learning_rate": 1.4368482039397451e-05,
|
| 78 |
+
"loss": 1.8495,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.09565217391304348,
|
| 83 |
+
"grad_norm": 7.267949104309082,
|
| 84 |
+
"learning_rate": 1.581691772885284e-05,
|
| 85 |
+
"loss": 1.5879,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.10434782608695652,
|
| 90 |
+
"grad_norm": 6.28985595703125,
|
| 91 |
+
"learning_rate": 1.7265353418308228e-05,
|
| 92 |
+
"loss": 1.3021,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.11304347826086956,
|
| 97 |
+
"grad_norm": 6.115077018737793,
|
| 98 |
+
"learning_rate": 1.8713789107763614e-05,
|
| 99 |
+
"loss": 1.0811,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.12173913043478261,
|
| 104 |
+
"grad_norm": 9.993290901184082,
|
| 105 |
+
"learning_rate": 2.0162224797219005e-05,
|
| 106 |
+
"loss": 0.9644,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.13043478260869565,
|
| 111 |
+
"grad_norm": 4.629942417144775,
|
| 112 |
+
"learning_rate": 2.161066048667439e-05,
|
| 113 |
+
"loss": 0.7733,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.1391304347826087,
|
| 118 |
+
"grad_norm": 4.087889194488525,
|
| 119 |
+
"learning_rate": 2.305909617612978e-05,
|
| 120 |
+
"loss": 0.6025,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.14782608695652175,
|
| 125 |
+
"grad_norm": 6.934291362762451,
|
| 126 |
+
"learning_rate": 2.450753186558517e-05,
|
| 127 |
+
"loss": 0.4823,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1565217391304348,
|
| 132 |
+
"grad_norm": 8.899859428405762,
|
| 133 |
+
"learning_rate": 2.5955967555040555e-05,
|
| 134 |
+
"loss": 0.3408,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16521739130434782,
|
| 139 |
+
"grad_norm": 5.459203243255615,
|
| 140 |
+
"learning_rate": 2.7404403244495948e-05,
|
| 141 |
+
"loss": 0.2558,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.17391304347826086,
|
| 146 |
+
"grad_norm": 2.3503053188323975,
|
| 147 |
+
"learning_rate": 2.885283893395133e-05,
|
| 148 |
+
"loss": 0.2116,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1826086956521739,
|
| 153 |
+
"grad_norm": 10.073762893676758,
|
| 154 |
+
"learning_rate": 3.030127462340672e-05,
|
| 155 |
+
"loss": 0.1117,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.19130434782608696,
|
| 160 |
+
"grad_norm": 0.8817861080169678,
|
| 161 |
+
"learning_rate": 3.1749710312862115e-05,
|
| 162 |
+
"loss": 0.0779,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.2,
|
| 167 |
+
"grad_norm": 0.28817218542099,
|
| 168 |
+
"learning_rate": 3.31981460023175e-05,
|
| 169 |
+
"loss": 0.0559,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.20869565217391303,
|
| 174 |
+
"grad_norm": 0.3490276038646698,
|
| 175 |
+
"learning_rate": 3.464658169177289e-05,
|
| 176 |
+
"loss": 0.0372,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.21739130434782608,
|
| 181 |
+
"grad_norm": 0.18767504394054413,
|
| 182 |
+
"learning_rate": 3.609501738122827e-05,
|
| 183 |
+
"loss": 0.0273,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.22608695652173913,
|
| 188 |
+
"grad_norm": 0.16744671761989594,
|
| 189 |
+
"learning_rate": 3.754345307068367e-05,
|
| 190 |
+
"loss": 0.0242,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.23478260869565218,
|
| 195 |
+
"grad_norm": 0.18482975661754608,
|
| 196 |
+
"learning_rate": 3.899188876013905e-05,
|
| 197 |
+
"loss": 0.0186,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.24347826086956523,
|
| 202 |
+
"grad_norm": 0.10977461189031601,
|
| 203 |
+
"learning_rate": 4.044032444959444e-05,
|
| 204 |
+
"loss": 0.0153,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.25217391304347825,
|
| 209 |
+
"grad_norm": 0.09746221452951431,
|
| 210 |
+
"learning_rate": 4.1888760139049825e-05,
|
| 211 |
+
"loss": 0.0133,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2608695652173913,
|
| 216 |
+
"grad_norm": 0.14220073819160461,
|
| 217 |
+
"learning_rate": 4.3337195828505215e-05,
|
| 218 |
+
"loss": 0.0127,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.26956521739130435,
|
| 223 |
+
"grad_norm": 0.08566069602966309,
|
| 224 |
+
"learning_rate": 4.4785631517960605e-05,
|
| 225 |
+
"loss": 0.0191,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.2782608695652174,
|
| 230 |
+
"grad_norm": 0.09003018587827682,
|
| 231 |
+
"learning_rate": 4.6234067207415995e-05,
|
| 232 |
+
"loss": 0.0114,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.28695652173913044,
|
| 237 |
+
"grad_norm": 0.07937142997980118,
|
| 238 |
+
"learning_rate": 4.7682502896871385e-05,
|
| 239 |
+
"loss": 0.0473,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.2956521739130435,
|
| 244 |
+
"grad_norm": 0.06389954686164856,
|
| 245 |
+
"learning_rate": 4.913093858632677e-05,
|
| 246 |
+
"loss": 0.0077,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.30434782608695654,
|
| 251 |
+
"grad_norm": 0.05928570777177811,
|
| 252 |
+
"learning_rate": 4.9935583612471015e-05,
|
| 253 |
+
"loss": 0.0073,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.3130434782608696,
|
| 258 |
+
"grad_norm": 0.06266701221466064,
|
| 259 |
+
"learning_rate": 4.977454264364855e-05,
|
| 260 |
+
"loss": 0.0376,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3217391304347826,
|
| 265 |
+
"grad_norm": 0.04980575665831566,
|
| 266 |
+
"learning_rate": 4.961350167482608e-05,
|
| 267 |
+
"loss": 0.0125,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.33043478260869563,
|
| 272 |
+
"grad_norm": 0.3670000731945038,
|
| 273 |
+
"learning_rate": 4.9452460706003613e-05,
|
| 274 |
+
"loss": 0.0062,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3391304347826087,
|
| 279 |
+
"grad_norm": 0.052209340035915375,
|
| 280 |
+
"learning_rate": 4.929141973718114e-05,
|
| 281 |
+
"loss": 0.0499,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.34782608695652173,
|
| 286 |
+
"grad_norm": 0.04546426981687546,
|
| 287 |
+
"learning_rate": 4.9136820407111575e-05,
|
| 288 |
+
"loss": 0.0663,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3565217391304348,
|
| 293 |
+
"grad_norm": 51.16396713256836,
|
| 294 |
+
"learning_rate": 4.89757794382891e-05,
|
| 295 |
+
"loss": 0.0721,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.3652173913043478,
|
| 300 |
+
"grad_norm": 0.06547929346561432,
|
| 301 |
+
"learning_rate": 4.881473846946664e-05,
|
| 302 |
+
"loss": 0.0303,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3739130434782609,
|
| 307 |
+
"grad_norm": 0.03635074943304062,
|
| 308 |
+
"learning_rate": 4.8653697500644166e-05,
|
| 309 |
+
"loss": 0.0405,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3826086956521739,
|
| 314 |
+
"grad_norm": 0.05633755028247833,
|
| 315 |
+
"learning_rate": 4.8492656531821695e-05,
|
| 316 |
+
"loss": 0.0113,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.391304347826087,
|
| 321 |
+
"grad_norm": 0.038711484521627426,
|
| 322 |
+
"learning_rate": 4.833161556299923e-05,
|
| 323 |
+
"loss": 0.0446,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.4,
|
| 328 |
+
"grad_norm": 0.03419716656208038,
|
| 329 |
+
"learning_rate": 4.817057459417676e-05,
|
| 330 |
+
"loss": 0.0065,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.40869565217391307,
|
| 335 |
+
"grad_norm": 0.03439817205071449,
|
| 336 |
+
"learning_rate": 4.800953362535429e-05,
|
| 337 |
+
"loss": 0.045,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.41739130434782606,
|
| 342 |
+
"grad_norm": 0.02928638830780983,
|
| 343 |
+
"learning_rate": 4.784849265653182e-05,
|
| 344 |
+
"loss": 0.0039,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4260869565217391,
|
| 349 |
+
"grad_norm": 0.03179425746202469,
|
| 350 |
+
"learning_rate": 4.7687451687709356e-05,
|
| 351 |
+
"loss": 0.003,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.43478260869565216,
|
| 356 |
+
"grad_norm": 0.022025400772690773,
|
| 357 |
+
"learning_rate": 4.7526410718886885e-05,
|
| 358 |
+
"loss": 0.0028,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4434782608695652,
|
| 363 |
+
"grad_norm": 0.027211090549826622,
|
| 364 |
+
"learning_rate": 4.736536975006442e-05,
|
| 365 |
+
"loss": 0.0025,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.45217391304347826,
|
| 370 |
+
"grad_norm": 0.019419824704527855,
|
| 371 |
+
"learning_rate": 4.7204328781241955e-05,
|
| 372 |
+
"loss": 0.0025,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4608695652173913,
|
| 377 |
+
"grad_norm": 0.019226014614105225,
|
| 378 |
+
"learning_rate": 4.704328781241948e-05,
|
| 379 |
+
"loss": 0.0024,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.46956521739130436,
|
| 384 |
+
"grad_norm": 0.01781400479376316,
|
| 385 |
+
"learning_rate": 4.688224684359701e-05,
|
| 386 |
+
"loss": 0.0022,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.4782608695652174,
|
| 391 |
+
"grad_norm": 0.0204134713858366,
|
| 392 |
+
"learning_rate": 4.672120587477454e-05,
|
| 393 |
+
"loss": 0.0025,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.48695652173913045,
|
| 398 |
+
"grad_norm": 0.022227726876735687,
|
| 399 |
+
"learning_rate": 4.6560164905952075e-05,
|
| 400 |
+
"loss": 0.0032,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.4956521739130435,
|
| 405 |
+
"grad_norm": 0.1570238471031189,
|
| 406 |
+
"learning_rate": 4.639912393712961e-05,
|
| 407 |
+
"loss": 0.0506,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.5043478260869565,
|
| 412 |
+
"grad_norm": 0.017615189775824547,
|
| 413 |
+
"learning_rate": 4.623808296830714e-05,
|
| 414 |
+
"loss": 0.0042,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.5130434782608696,
|
| 419 |
+
"grad_norm": 0.017343297600746155,
|
| 420 |
+
"learning_rate": 4.607704199948467e-05,
|
| 421 |
+
"loss": 0.0021,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.5217391304347826,
|
| 426 |
+
"grad_norm": 0.018280992284417152,
|
| 427 |
+
"learning_rate": 4.59160010306622e-05,
|
| 428 |
+
"loss": 0.0019,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5304347826086957,
|
| 433 |
+
"grad_norm": 0.017326297238469124,
|
| 434 |
+
"learning_rate": 4.5754960061839736e-05,
|
| 435 |
+
"loss": 0.0164,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5391304347826087,
|
| 440 |
+
"grad_norm": 0.014924581162631512,
|
| 441 |
+
"learning_rate": 4.5593919093017265e-05,
|
| 442 |
+
"loss": 0.0587,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5478260869565217,
|
| 447 |
+
"grad_norm": 0.023350143805146217,
|
| 448 |
+
"learning_rate": 4.54328781241948e-05,
|
| 449 |
+
"loss": 0.0207,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.5565217391304348,
|
| 454 |
+
"grad_norm": 0.014112471602857113,
|
| 455 |
+
"learning_rate": 4.527183715537233e-05,
|
| 456 |
+
"loss": 0.0038,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.5652173913043478,
|
| 461 |
+
"grad_norm": 0.016805477440357208,
|
| 462 |
+
"learning_rate": 4.5110796186549856e-05,
|
| 463 |
+
"loss": 0.0457,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5739130434782609,
|
| 468 |
+
"grad_norm": 0.018589412793517113,
|
| 469 |
+
"learning_rate": 4.494975521772739e-05,
|
| 470 |
+
"loss": 0.1005,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5826086956521739,
|
| 475 |
+
"grad_norm": 0.021146375685930252,
|
| 476 |
+
"learning_rate": 4.478871424890492e-05,
|
| 477 |
+
"loss": 0.0019,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.591304347826087,
|
| 482 |
+
"grad_norm": 0.014253458008170128,
|
| 483 |
+
"learning_rate": 4.4627673280082455e-05,
|
| 484 |
+
"loss": 0.0019,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.6,
|
| 489 |
+
"grad_norm": 0.02895612083375454,
|
| 490 |
+
"learning_rate": 4.446663231125999e-05,
|
| 491 |
+
"loss": 0.0373,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.6086956521739131,
|
| 496 |
+
"grad_norm": 0.012195841409265995,
|
| 497 |
+
"learning_rate": 4.430559134243752e-05,
|
| 498 |
+
"loss": 0.0018,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.6173913043478261,
|
| 503 |
+
"grad_norm": 0.04872201383113861,
|
| 504 |
+
"learning_rate": 4.414455037361505e-05,
|
| 505 |
+
"loss": 0.0344,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.6260869565217392,
|
| 510 |
+
"grad_norm": 0.4130329489707947,
|
| 511 |
+
"learning_rate": 4.398350940479258e-05,
|
| 512 |
+
"loss": 0.0258,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6347826086956522,
|
| 517 |
+
"grad_norm": 0.4404933750629425,
|
| 518 |
+
"learning_rate": 4.3822468435970116e-05,
|
| 519 |
+
"loss": 0.0089,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.6434782608695652,
|
| 524 |
+
"grad_norm": 0.012819499708712101,
|
| 525 |
+
"learning_rate": 4.3661427467147645e-05,
|
| 526 |
+
"loss": 0.0178,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.6521739130434783,
|
| 531 |
+
"grad_norm": 0.020058365538716316,
|
| 532 |
+
"learning_rate": 4.350038649832517e-05,
|
| 533 |
+
"loss": 0.0015,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.6608695652173913,
|
| 538 |
+
"grad_norm": 0.05424540862441063,
|
| 539 |
+
"learning_rate": 4.333934552950271e-05,
|
| 540 |
+
"loss": 0.0084,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.6695652173913044,
|
| 545 |
+
"grad_norm": 0.011423332616686821,
|
| 546 |
+
"learning_rate": 4.3178304560680236e-05,
|
| 547 |
+
"loss": 0.0014,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.6782608695652174,
|
| 552 |
+
"grad_norm": 0.03126579150557518,
|
| 553 |
+
"learning_rate": 4.301726359185777e-05,
|
| 554 |
+
"loss": 0.0012,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.6869565217391305,
|
| 559 |
+
"grad_norm": 0.010747378692030907,
|
| 560 |
+
"learning_rate": 4.28562226230353e-05,
|
| 561 |
+
"loss": 0.0011,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.6956521739130435,
|
| 566 |
+
"grad_norm": 0.010900999419391155,
|
| 567 |
+
"learning_rate": 4.2695181654212834e-05,
|
| 568 |
+
"loss": 0.001,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.7043478260869566,
|
| 573 |
+
"grad_norm": 0.10920006781816483,
|
| 574 |
+
"learning_rate": 4.253414068539037e-05,
|
| 575 |
+
"loss": 0.0371,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.7130434782608696,
|
| 580 |
+
"grad_norm": 0.008641124702990055,
|
| 581 |
+
"learning_rate": 4.2379541355320796e-05,
|
| 582 |
+
"loss": 0.009,
|
| 583 |
+
"step": 2050
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.7217391304347827,
|
| 587 |
+
"grad_norm": 0.00958116352558136,
|
| 588 |
+
"learning_rate": 4.2218500386498324e-05,
|
| 589 |
+
"loss": 0.0011,
|
| 590 |
+
"step": 2075
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.7304347826086957,
|
| 594 |
+
"grad_norm": 0.007281237281858921,
|
| 595 |
+
"learning_rate": 4.205745941767586e-05,
|
| 596 |
+
"loss": 0.001,
|
| 597 |
+
"step": 2100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7391304347826086,
|
| 601 |
+
"grad_norm": 0.00872531346976757,
|
| 602 |
+
"learning_rate": 4.1896418448853394e-05,
|
| 603 |
+
"loss": 0.0285,
|
| 604 |
+
"step": 2125
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.7478260869565218,
|
| 608 |
+
"grad_norm": 0.008862826973199844,
|
| 609 |
+
"learning_rate": 4.173537748003092e-05,
|
| 610 |
+
"loss": 0.0229,
|
| 611 |
+
"step": 2150
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.7565217391304347,
|
| 615 |
+
"grad_norm": 0.012311381287872791,
|
| 616 |
+
"learning_rate": 4.157433651120846e-05,
|
| 617 |
+
"loss": 0.0682,
|
| 618 |
+
"step": 2175
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.7652173913043478,
|
| 622 |
+
"grad_norm": 0.009278792887926102,
|
| 623 |
+
"learning_rate": 4.1413295542385986e-05,
|
| 624 |
+
"loss": 0.0517,
|
| 625 |
+
"step": 2200
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7739130434782608,
|
| 629 |
+
"grad_norm": 0.034189265221357346,
|
| 630 |
+
"learning_rate": 4.1252254573563514e-05,
|
| 631 |
+
"loss": 0.0803,
|
| 632 |
+
"step": 2225
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.782608695652174,
|
| 636 |
+
"grad_norm": 0.043665286153554916,
|
| 637 |
+
"learning_rate": 4.109121360474105e-05,
|
| 638 |
+
"loss": 0.1145,
|
| 639 |
+
"step": 2250
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.7913043478260869,
|
| 643 |
+
"grad_norm": 0.08209193497896194,
|
| 644 |
+
"learning_rate": 4.093017263591858e-05,
|
| 645 |
+
"loss": 0.0125,
|
| 646 |
+
"step": 2275
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.8,
|
| 650 |
+
"grad_norm": 0.01089281216263771,
|
| 651 |
+
"learning_rate": 4.076913166709611e-05,
|
| 652 |
+
"loss": 0.0011,
|
| 653 |
+
"step": 2300
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.808695652173913,
|
| 657 |
+
"grad_norm": 0.007122454699128866,
|
| 658 |
+
"learning_rate": 4.060809069827364e-05,
|
| 659 |
+
"loss": 0.0045,
|
| 660 |
+
"step": 2325
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.8173913043478261,
|
| 664 |
+
"grad_norm": 0.009066189639270306,
|
| 665 |
+
"learning_rate": 4.0447049729451176e-05,
|
| 666 |
+
"loss": 0.0025,
|
| 667 |
+
"step": 2350
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.8260869565217391,
|
| 671 |
+
"grad_norm": 0.00853909645229578,
|
| 672 |
+
"learning_rate": 4.0286008760628704e-05,
|
| 673 |
+
"loss": 0.0009,
|
| 674 |
+
"step": 2375
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.8347826086956521,
|
| 678 |
+
"grad_norm": 0.007053891196846962,
|
| 679 |
+
"learning_rate": 4.012496779180624e-05,
|
| 680 |
+
"loss": 0.0008,
|
| 681 |
+
"step": 2400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8434782608695652,
|
| 685 |
+
"grad_norm": 0.007313120178878307,
|
| 686 |
+
"learning_rate": 3.9963926822983774e-05,
|
| 687 |
+
"loss": 0.0013,
|
| 688 |
+
"step": 2425
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.8521739130434782,
|
| 692 |
+
"grad_norm": 0.009188150055706501,
|
| 693 |
+
"learning_rate": 3.9802885854161296e-05,
|
| 694 |
+
"loss": 0.0685,
|
| 695 |
+
"step": 2450
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.8608695652173913,
|
| 699 |
+
"grad_norm": 0.010708467103540897,
|
| 700 |
+
"learning_rate": 3.964184488533883e-05,
|
| 701 |
+
"loss": 0.0845,
|
| 702 |
+
"step": 2475
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.8695652173913043,
|
| 706 |
+
"grad_norm": 0.009248602204024792,
|
| 707 |
+
"learning_rate": 3.948080391651636e-05,
|
| 708 |
+
"loss": 0.0621,
|
| 709 |
+
"step": 2500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8782608695652174,
|
| 713 |
+
"grad_norm": 0.008605259470641613,
|
| 714 |
+
"learning_rate": 3.9319762947693894e-05,
|
| 715 |
+
"loss": 0.001,
|
| 716 |
+
"step": 2525
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.8869565217391304,
|
| 720 |
+
"grad_norm": 0.00624378165230155,
|
| 721 |
+
"learning_rate": 3.915872197887143e-05,
|
| 722 |
+
"loss": 0.026,
|
| 723 |
+
"step": 2550
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.8956521739130435,
|
| 727 |
+
"grad_norm": 0.008726169355213642,
|
| 728 |
+
"learning_rate": 3.899768101004896e-05,
|
| 729 |
+
"loss": 0.0344,
|
| 730 |
+
"step": 2575
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.9043478260869565,
|
| 734 |
+
"grad_norm": 0.007860297337174416,
|
| 735 |
+
"learning_rate": 3.883664004122649e-05,
|
| 736 |
+
"loss": 0.0321,
|
| 737 |
+
"step": 2600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.9130434782608695,
|
| 741 |
+
"grad_norm": 0.006469408981502056,
|
| 742 |
+
"learning_rate": 3.867559907240402e-05,
|
| 743 |
+
"loss": 0.0456,
|
| 744 |
+
"step": 2625
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.9217391304347826,
|
| 748 |
+
"grad_norm": 0.025073576718568802,
|
| 749 |
+
"learning_rate": 3.8514558103581556e-05,
|
| 750 |
+
"loss": 0.0607,
|
| 751 |
+
"step": 2650
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.9304347826086956,
|
| 755 |
+
"grad_norm": 0.005454166326671839,
|
| 756 |
+
"learning_rate": 3.8353517134759084e-05,
|
| 757 |
+
"loss": 0.0007,
|
| 758 |
+
"step": 2675
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.9391304347826087,
|
| 762 |
+
"grad_norm": 0.005138472653925419,
|
| 763 |
+
"learning_rate": 3.819247616593661e-05,
|
| 764 |
+
"loss": 0.0008,
|
| 765 |
+
"step": 2700
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9478260869565217,
|
| 769 |
+
"grad_norm": 0.04282465949654579,
|
| 770 |
+
"learning_rate": 3.803143519711415e-05,
|
| 771 |
+
"loss": 0.0007,
|
| 772 |
+
"step": 2725
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.9565217391304348,
|
| 776 |
+
"grad_norm": 0.00715728010982275,
|
| 777 |
+
"learning_rate": 3.7870394228291676e-05,
|
| 778 |
+
"loss": 0.0387,
|
| 779 |
+
"step": 2750
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.9652173913043478,
|
| 783 |
+
"grad_norm": 0.005337136331945658,
|
| 784 |
+
"learning_rate": 3.770935325946921e-05,
|
| 785 |
+
"loss": 0.0007,
|
| 786 |
+
"step": 2775
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.9739130434782609,
|
| 790 |
+
"grad_norm": 0.005603776779025793,
|
| 791 |
+
"learning_rate": 3.754831229064674e-05,
|
| 792 |
+
"loss": 0.0006,
|
| 793 |
+
"step": 2800
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9826086956521739,
|
| 797 |
+
"grad_norm": 0.005324610508978367,
|
| 798 |
+
"learning_rate": 3.7387271321824274e-05,
|
| 799 |
+
"loss": 0.0392,
|
| 800 |
+
"step": 2825
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.991304347826087,
|
| 804 |
+
"grad_norm": 0.004979921039193869,
|
| 805 |
+
"learning_rate": 3.722623035300181e-05,
|
| 806 |
+
"loss": 0.0204,
|
| 807 |
+
"step": 2850
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 1.0,
|
| 811 |
+
"grad_norm": 0.0061447713524103165,
|
| 812 |
+
"learning_rate": 3.706518938417934e-05,
|
| 813 |
+
"loss": 0.0007,
|
| 814 |
+
"step": 2875
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 1.0,
|
| 818 |
+
"eval_accuracy": 1.0,
|
| 819 |
+
"eval_f1_macro": 1.0,
|
| 820 |
+
"eval_f1_micro": 1.0,
|
| 821 |
+
"eval_f1_weighted": 1.0,
|
| 822 |
+
"eval_loss": 0.0005410231533460319,
|
| 823 |
+
"eval_precision_macro": 1.0,
|
| 824 |
+
"eval_precision_micro": 1.0,
|
| 825 |
+
"eval_precision_weighted": 1.0,
|
| 826 |
+
"eval_recall_macro": 1.0,
|
| 827 |
+
"eval_recall_micro": 1.0,
|
| 828 |
+
"eval_recall_weighted": 1.0,
|
| 829 |
+
"eval_runtime": 5.0042,
|
| 830 |
+
"eval_samples_per_second": 459.614,
|
| 831 |
+
"eval_steps_per_second": 28.776,
|
| 832 |
+
"step": 2875
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.008695652173913,
|
| 836 |
+
"grad_norm": 0.005072788801044226,
|
| 837 |
+
"learning_rate": 3.690414841535687e-05,
|
| 838 |
+
"loss": 0.0014,
|
| 839 |
+
"step": 2900
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.017391304347826,
|
| 843 |
+
"grad_norm": 0.004767073784023523,
|
| 844 |
+
"learning_rate": 3.67431074465344e-05,
|
| 845 |
+
"loss": 0.0039,
|
| 846 |
+
"step": 2925
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.0260869565217392,
|
| 850 |
+
"grad_norm": 0.019580593332648277,
|
| 851 |
+
"learning_rate": 3.658206647771193e-05,
|
| 852 |
+
"loss": 0.0129,
|
| 853 |
+
"step": 2950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.0347826086956522,
|
| 857 |
+
"grad_norm": 0.004164039622992277,
|
| 858 |
+
"learning_rate": 3.6421025508889464e-05,
|
| 859 |
+
"loss": 0.0005,
|
| 860 |
+
"step": 2975
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.0434782608695652,
|
| 864 |
+
"grad_norm": 0.006551030091941357,
|
| 865 |
+
"learning_rate": 3.625998454006699e-05,
|
| 866 |
+
"loss": 0.0005,
|
| 867 |
+
"step": 3000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.0521739130434782,
|
| 871 |
+
"grad_norm": 0.004048625007271767,
|
| 872 |
+
"learning_rate": 3.609894357124453e-05,
|
| 873 |
+
"loss": 0.0005,
|
| 874 |
+
"step": 3025
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.0608695652173914,
|
| 878 |
+
"grad_norm": 0.003683075774461031,
|
| 879 |
+
"learning_rate": 3.5937902602422056e-05,
|
| 880 |
+
"loss": 0.0005,
|
| 881 |
+
"step": 3050
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.0695652173913044,
|
| 885 |
+
"grad_norm": 0.003764552529901266,
|
| 886 |
+
"learning_rate": 3.577686163359959e-05,
|
| 887 |
+
"loss": 0.0005,
|
| 888 |
+
"step": 3075
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 1.0782608695652174,
|
| 892 |
+
"grad_norm": 0.014071044512093067,
|
| 893 |
+
"learning_rate": 3.561582066477712e-05,
|
| 894 |
+
"loss": 0.0005,
|
| 895 |
+
"step": 3100
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.0869565217391304,
|
| 899 |
+
"grad_norm": 0.004224707838147879,
|
| 900 |
+
"learning_rate": 3.5454779695954654e-05,
|
| 901 |
+
"loss": 0.0004,
|
| 902 |
+
"step": 3125
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.0956521739130434,
|
| 906 |
+
"grad_norm": 0.007790651638060808,
|
| 907 |
+
"learning_rate": 3.529373872713219e-05,
|
| 908 |
+
"loss": 0.0004,
|
| 909 |
+
"step": 3150
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.1043478260869566,
|
| 913 |
+
"grad_norm": 0.0033269149716943502,
|
| 914 |
+
"learning_rate": 3.513269775830972e-05,
|
| 915 |
+
"loss": 0.0004,
|
| 916 |
+
"step": 3175
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.1130434782608696,
|
| 920 |
+
"grad_norm": 0.003678582375869155,
|
| 921 |
+
"learning_rate": 3.4971656789487245e-05,
|
| 922 |
+
"loss": 0.0004,
|
| 923 |
+
"step": 3200
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 1.1217391304347826,
|
| 927 |
+
"grad_norm": 0.415933758020401,
|
| 928 |
+
"learning_rate": 3.4810615820664774e-05,
|
| 929 |
+
"loss": 0.0471,
|
| 930 |
+
"step": 3225
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 1.1304347826086956,
|
| 934 |
+
"grad_norm": 0.003935901448130608,
|
| 935 |
+
"learning_rate": 3.464957485184231e-05,
|
| 936 |
+
"loss": 0.0008,
|
| 937 |
+
"step": 3250
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.1391304347826088,
|
| 941 |
+
"grad_norm": 0.005461950786411762,
|
| 942 |
+
"learning_rate": 3.4488533883019844e-05,
|
| 943 |
+
"loss": 0.0868,
|
| 944 |
+
"step": 3275
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.1478260869565218,
|
| 948 |
+
"grad_norm": 0.00403949897736311,
|
| 949 |
+
"learning_rate": 3.432749291419737e-05,
|
| 950 |
+
"loss": 0.0314,
|
| 951 |
+
"step": 3300
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 1.1565217391304348,
|
| 955 |
+
"grad_norm": 0.0038034759927541018,
|
| 956 |
+
"learning_rate": 3.416645194537491e-05,
|
| 957 |
+
"loss": 0.0006,
|
| 958 |
+
"step": 3325
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.1652173913043478,
|
| 962 |
+
"grad_norm": 0.003490498522296548,
|
| 963 |
+
"learning_rate": 3.4005410976552435e-05,
|
| 964 |
+
"loss": 0.0004,
|
| 965 |
+
"step": 3350
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 1.1739130434782608,
|
| 969 |
+
"grad_norm": 0.002800930989906192,
|
| 970 |
+
"learning_rate": 3.384437000772997e-05,
|
| 971 |
+
"loss": 0.0004,
|
| 972 |
+
"step": 3375
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 1.182608695652174,
|
| 976 |
+
"grad_norm": 0.0032401601783931255,
|
| 977 |
+
"learning_rate": 3.36833290389075e-05,
|
| 978 |
+
"loss": 0.0004,
|
| 979 |
+
"step": 3400
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.191304347826087,
|
| 983 |
+
"grad_norm": 0.008039023727178574,
|
| 984 |
+
"learning_rate": 3.3522288070085034e-05,
|
| 985 |
+
"loss": 0.1032,
|
| 986 |
+
"step": 3425
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 1.2,
|
| 990 |
+
"grad_norm": 0.0048836818896234035,
|
| 991 |
+
"learning_rate": 3.336124710126256e-05,
|
| 992 |
+
"loss": 0.0007,
|
| 993 |
+
"step": 3450
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 1.208695652173913,
|
| 997 |
+
"grad_norm": 0.01012449711561203,
|
| 998 |
+
"learning_rate": 3.320020613244009e-05,
|
| 999 |
+
"loss": 0.0223,
|
| 1000 |
+
"step": 3475
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 1.2173913043478262,
|
| 1004 |
+
"grad_norm": 0.005898616276681423,
|
| 1005 |
+
"learning_rate": 3.3039165163617625e-05,
|
| 1006 |
+
"loss": 0.0143,
|
| 1007 |
+
"step": 3500
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.2260869565217392,
|
| 1011 |
+
"grad_norm": 7.748152256011963,
|
| 1012 |
+
"learning_rate": 3.2878124194795154e-05,
|
| 1013 |
+
"loss": 0.0928,
|
| 1014 |
+
"step": 3525
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 1.2347826086956522,
|
| 1018 |
+
"grad_norm": 0.0053124018013477325,
|
| 1019 |
+
"learning_rate": 3.271708322597269e-05,
|
| 1020 |
+
"loss": 0.0045,
|
| 1021 |
+
"step": 3550
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.2434782608695651,
|
| 1025 |
+
"grad_norm": 0.009419775567948818,
|
| 1026 |
+
"learning_rate": 3.2556042257150224e-05,
|
| 1027 |
+
"loss": 0.0008,
|
| 1028 |
+
"step": 3575
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 1.2521739130434781,
|
| 1032 |
+
"grad_norm": 0.0048626321367919445,
|
| 1033 |
+
"learning_rate": 3.239500128832775e-05,
|
| 1034 |
+
"loss": 0.0005,
|
| 1035 |
+
"step": 3600
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 1.2608695652173914,
|
| 1039 |
+
"grad_norm": 0.004047353286296129,
|
| 1040 |
+
"learning_rate": 3.223396031950529e-05,
|
| 1041 |
+
"loss": 0.0004,
|
| 1042 |
+
"step": 3625
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 1.2695652173913043,
|
| 1046 |
+
"grad_norm": 0.0036859684623777866,
|
| 1047 |
+
"learning_rate": 3.2072919350682815e-05,
|
| 1048 |
+
"loss": 0.0484,
|
| 1049 |
+
"step": 3650
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.2782608695652173,
|
| 1053 |
+
"grad_norm": 0.007856795564293861,
|
| 1054 |
+
"learning_rate": 3.191187838186035e-05,
|
| 1055 |
+
"loss": 0.022,
|
| 1056 |
+
"step": 3675
|
| 1057 |
+
},
|
| 1058 |
+
{
|
| 1059 |
+
"epoch": 1.2869565217391306,
|
| 1060 |
+
"grad_norm": 0.0035763189662247896,
|
| 1061 |
+
"learning_rate": 3.175083741303788e-05,
|
| 1062 |
+
"loss": 0.0004,
|
| 1063 |
+
"step": 3700
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.2956521739130435,
|
| 1067 |
+
"grad_norm": 0.003156292950734496,
|
| 1068 |
+
"learning_rate": 3.158979644421541e-05,
|
| 1069 |
+
"loss": 0.0004,
|
| 1070 |
+
"step": 3725
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 1.3043478260869565,
|
| 1074 |
+
"grad_norm": 0.003629323560744524,
|
| 1075 |
+
"learning_rate": 3.142875547539294e-05,
|
| 1076 |
+
"loss": 0.0004,
|
| 1077 |
+
"step": 3750
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 1.3130434782608695,
|
| 1081 |
+
"grad_norm": 68.89725494384766,
|
| 1082 |
+
"learning_rate": 3.126771450657047e-05,
|
| 1083 |
+
"loss": 0.0249,
|
| 1084 |
+
"step": 3775
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 1.3217391304347825,
|
| 1088 |
+
"grad_norm": 0.004872568417340517,
|
| 1089 |
+
"learning_rate": 3.1106673537748005e-05,
|
| 1090 |
+
"loss": 0.106,
|
| 1091 |
+
"step": 3800
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.3304347826086955,
|
| 1095 |
+
"grad_norm": 0.0038311562966555357,
|
| 1096 |
+
"learning_rate": 3.0945632568925534e-05,
|
| 1097 |
+
"loss": 0.0008,
|
| 1098 |
+
"step": 3825
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 1.3391304347826087,
|
| 1102 |
+
"grad_norm": 0.01143663190305233,
|
| 1103 |
+
"learning_rate": 3.078459160010307e-05,
|
| 1104 |
+
"loss": 0.0433,
|
| 1105 |
+
"step": 3850
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.3478260869565217,
|
| 1109 |
+
"grad_norm": 0.00519527355208993,
|
| 1110 |
+
"learning_rate": 3.0623550631280604e-05,
|
| 1111 |
+
"loss": 0.0004,
|
| 1112 |
+
"step": 3875
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 1.3565217391304347,
|
| 1116 |
+
"grad_norm": 0.0031176016200333834,
|
| 1117 |
+
"learning_rate": 3.0462509662458132e-05,
|
| 1118 |
+
"loss": 0.0005,
|
| 1119 |
+
"step": 3900
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 1.365217391304348,
|
| 1123 |
+
"grad_norm": 0.0033473202493041754,
|
| 1124 |
+
"learning_rate": 3.0301468693635664e-05,
|
| 1125 |
+
"loss": 0.0003,
|
| 1126 |
+
"step": 3925
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 1.373913043478261,
|
| 1130 |
+
"grad_norm": 0.0031863965559750795,
|
| 1131 |
+
"learning_rate": 3.0140427724813192e-05,
|
| 1132 |
+
"loss": 0.0003,
|
| 1133 |
+
"step": 3950
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 1.382608695652174,
|
| 1137 |
+
"grad_norm": 0.002523756120353937,
|
| 1138 |
+
"learning_rate": 2.9979386755990724e-05,
|
| 1139 |
+
"loss": 0.0318,
|
| 1140 |
+
"step": 3975
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 1.391304347826087,
|
| 1144 |
+
"grad_norm": 0.0034575534518808126,
|
| 1145 |
+
"learning_rate": 2.9818345787168255e-05,
|
| 1146 |
+
"loss": 0.0003,
|
| 1147 |
+
"step": 4000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.4,
|
| 1151 |
+
"grad_norm": 0.004556054249405861,
|
| 1152 |
+
"learning_rate": 2.9657304818345787e-05,
|
| 1153 |
+
"loss": 0.0003,
|
| 1154 |
+
"step": 4025
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 1.4086956521739131,
|
| 1158 |
+
"grad_norm": 0.009268323890864849,
|
| 1159 |
+
"learning_rate": 2.949626384952332e-05,
|
| 1160 |
+
"loss": 0.0296,
|
| 1161 |
+
"step": 4050
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.4173913043478261,
|
| 1165 |
+
"grad_norm": 0.01629672758281231,
|
| 1166 |
+
"learning_rate": 2.9335222880700854e-05,
|
| 1167 |
+
"loss": 0.0021,
|
| 1168 |
+
"step": 4075
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 1.4260869565217391,
|
| 1172 |
+
"grad_norm": 0.0037628798745572567,
|
| 1173 |
+
"learning_rate": 2.9174181911878385e-05,
|
| 1174 |
+
"loss": 0.0015,
|
| 1175 |
+
"step": 4100
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 1.434782608695652,
|
| 1179 |
+
"grad_norm": 0.004670215770602226,
|
| 1180 |
+
"learning_rate": 2.9013140943055917e-05,
|
| 1181 |
+
"loss": 0.0329,
|
| 1182 |
+
"step": 4125
|
| 1183 |
+
},
|
| 1184 |
+
{
|
| 1185 |
+
"epoch": 1.4434782608695653,
|
| 1186 |
+
"grad_norm": 0.003128861775621772,
|
| 1187 |
+
"learning_rate": 2.885209997423345e-05,
|
| 1188 |
+
"loss": 0.0361,
|
| 1189 |
+
"step": 4150
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.4521739130434783,
|
| 1193 |
+
"grad_norm": 0.0030446811579167843,
|
| 1194 |
+
"learning_rate": 2.869105900541098e-05,
|
| 1195 |
+
"loss": 0.0003,
|
| 1196 |
+
"step": 4175
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 1.4608695652173913,
|
| 1200 |
+
"grad_norm": 0.0032489860896021128,
|
| 1201 |
+
"learning_rate": 2.853001803658851e-05,
|
| 1202 |
+
"loss": 0.0355,
|
| 1203 |
+
"step": 4200
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 1.4695652173913043,
|
| 1207 |
+
"grad_norm": 0.002284318907186389,
|
| 1208 |
+
"learning_rate": 2.836897706776604e-05,
|
| 1209 |
+
"loss": 0.0276,
|
| 1210 |
+
"step": 4225
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 1.4782608695652173,
|
| 1214 |
+
"grad_norm": 0.005029418971389532,
|
| 1215 |
+
"learning_rate": 2.8207936098943572e-05,
|
| 1216 |
+
"loss": 0.0617,
|
| 1217 |
+
"step": 4250
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.4869565217391305,
|
| 1221 |
+
"grad_norm": 0.0027885879389941692,
|
| 1222 |
+
"learning_rate": 2.8046895130121104e-05,
|
| 1223 |
+
"loss": 0.0004,
|
| 1224 |
+
"step": 4275
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 1.4956521739130435,
|
| 1228 |
+
"grad_norm": 0.0034100012853741646,
|
| 1229 |
+
"learning_rate": 2.7885854161298635e-05,
|
| 1230 |
+
"loss": 0.0003,
|
| 1231 |
+
"step": 4300
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.5043478260869565,
|
| 1235 |
+
"grad_norm": 0.0024735534097999334,
|
| 1236 |
+
"learning_rate": 2.7724813192476167e-05,
|
| 1237 |
+
"loss": 0.0003,
|
| 1238 |
+
"step": 4325
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 1.5130434782608697,
|
| 1242 |
+
"grad_norm": 0.0018064508913084865,
|
| 1243 |
+
"learning_rate": 2.75637722236537e-05,
|
| 1244 |
+
"loss": 0.0003,
|
| 1245 |
+
"step": 4350
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 1.5217391304347827,
|
| 1249 |
+
"grad_norm": 0.001821321900933981,
|
| 1250 |
+
"learning_rate": 2.7402731254831234e-05,
|
| 1251 |
+
"loss": 0.0003,
|
| 1252 |
+
"step": 4375
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.5304347826086957,
|
| 1256 |
+
"grad_norm": 0.002236420288681984,
|
| 1257 |
+
"learning_rate": 2.7241690286008765e-05,
|
| 1258 |
+
"loss": 0.0003,
|
| 1259 |
+
"step": 4400
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 1.5391304347826087,
|
| 1263 |
+
"grad_norm": 0.0018212420400232077,
|
| 1264 |
+
"learning_rate": 2.708064931718629e-05,
|
| 1265 |
+
"loss": 0.0717,
|
| 1266 |
+
"step": 4425
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 1.5478260869565217,
|
| 1270 |
+
"grad_norm": 0.0029810129199177027,
|
| 1271 |
+
"learning_rate": 2.6919608348363822e-05,
|
| 1272 |
+
"loss": 0.0003,
|
| 1273 |
+
"step": 4450
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.5565217391304347,
|
| 1277 |
+
"grad_norm": 0.002055574208498001,
|
| 1278 |
+
"learning_rate": 2.6758567379541353e-05,
|
| 1279 |
+
"loss": 0.0003,
|
| 1280 |
+
"step": 4475
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 1.5652173913043477,
|
| 1284 |
+
"grad_norm": 0.002401245990768075,
|
| 1285 |
+
"learning_rate": 2.659752641071889e-05,
|
| 1286 |
+
"loss": 0.0471,
|
| 1287 |
+
"step": 4500
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 1.5739130434782609,
|
| 1291 |
+
"grad_norm": 0.0019228613236919045,
|
| 1292 |
+
"learning_rate": 2.643648544189642e-05,
|
| 1293 |
+
"loss": 0.004,
|
| 1294 |
+
"step": 4525
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 1.5826086956521739,
|
| 1298 |
+
"grad_norm": 0.01082328986376524,
|
| 1299 |
+
"learning_rate": 2.6275444473073952e-05,
|
| 1300 |
+
"loss": 0.0004,
|
| 1301 |
+
"step": 4550
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 1.591304347826087,
|
| 1305 |
+
"grad_norm": 0.004519768990576267,
|
| 1306 |
+
"learning_rate": 2.6114403504251483e-05,
|
| 1307 |
+
"loss": 0.0003,
|
| 1308 |
+
"step": 4575
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 1.6,
|
| 1312 |
+
"grad_norm": 0.004080561455339193,
|
| 1313 |
+
"learning_rate": 2.5953362535429015e-05,
|
| 1314 |
+
"loss": 0.0003,
|
| 1315 |
+
"step": 4600
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.608695652173913,
|
| 1319 |
+
"grad_norm": 0.0020184165332466364,
|
| 1320 |
+
"learning_rate": 2.5792321566606547e-05,
|
| 1321 |
+
"loss": 0.0003,
|
| 1322 |
+
"step": 4625
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.617391304347826,
|
| 1326 |
+
"grad_norm": 0.00274313660338521,
|
| 1327 |
+
"learning_rate": 2.563128059778408e-05,
|
| 1328 |
+
"loss": 0.0002,
|
| 1329 |
+
"step": 4650
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 1.626086956521739,
|
| 1333 |
+
"grad_norm": 0.0025938046164810658,
|
| 1334 |
+
"learning_rate": 2.5470239628961607e-05,
|
| 1335 |
+
"loss": 0.0002,
|
| 1336 |
+
"step": 4675
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 1.634782608695652,
|
| 1340 |
+
"grad_norm": 0.002224980155006051,
|
| 1341 |
+
"learning_rate": 2.530919866013914e-05,
|
| 1342 |
+
"loss": 0.0002,
|
| 1343 |
+
"step": 4700
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.643478260869565,
|
| 1347 |
+
"grad_norm": 0.0018848755862563848,
|
| 1348 |
+
"learning_rate": 2.514815769131667e-05,
|
| 1349 |
+
"loss": 0.0002,
|
| 1350 |
+
"step": 4725
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 1.6521739130434783,
|
| 1354 |
+
"grad_norm": 0.0021186743397265673,
|
| 1355 |
+
"learning_rate": 2.4987116722494202e-05,
|
| 1356 |
+
"loss": 0.012,
|
| 1357 |
+
"step": 4750
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.6608695652173913,
|
| 1361 |
+
"grad_norm": 0.002839816967025399,
|
| 1362 |
+
"learning_rate": 2.4826075753671733e-05,
|
| 1363 |
+
"loss": 0.0732,
|
| 1364 |
+
"step": 4775
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.6695652173913045,
|
| 1368 |
+
"grad_norm": 0.0023581942077726126,
|
| 1369 |
+
"learning_rate": 2.466503478484927e-05,
|
| 1370 |
+
"loss": 0.0441,
|
| 1371 |
+
"step": 4800
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 1.6782608695652175,
|
| 1375 |
+
"grad_norm": 0.0025644400157034397,
|
| 1376 |
+
"learning_rate": 2.45039938160268e-05,
|
| 1377 |
+
"loss": 0.0002,
|
| 1378 |
+
"step": 4825
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 1.6869565217391305,
|
| 1382 |
+
"grad_norm": 0.0018391634803265333,
|
| 1383 |
+
"learning_rate": 2.4342952847204332e-05,
|
| 1384 |
+
"loss": 0.0003,
|
| 1385 |
+
"step": 4850
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 1.6956521739130435,
|
| 1389 |
+
"grad_norm": 0.002724371151998639,
|
| 1390 |
+
"learning_rate": 2.418191187838186e-05,
|
| 1391 |
+
"loss": 0.0002,
|
| 1392 |
+
"step": 4875
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 1.7043478260869565,
|
| 1396 |
+
"grad_norm": 0.002339495113119483,
|
| 1397 |
+
"learning_rate": 2.402087090955939e-05,
|
| 1398 |
+
"loss": 0.0002,
|
| 1399 |
+
"step": 4900
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.7130434782608694,
|
| 1403 |
+
"grad_norm": 0.0015120344469323754,
|
| 1404 |
+
"learning_rate": 2.3859829940736923e-05,
|
| 1405 |
+
"loss": 0.0059,
|
| 1406 |
+
"step": 4925
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.7217391304347827,
|
| 1410 |
+
"grad_norm": 0.0019244271097704768,
|
| 1411 |
+
"learning_rate": 2.369878897191446e-05,
|
| 1412 |
+
"loss": 0.0488,
|
| 1413 |
+
"step": 4950
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 1.7304347826086957,
|
| 1417 |
+
"grad_norm": 0.006782891228795052,
|
| 1418 |
+
"learning_rate": 2.353774800309199e-05,
|
| 1419 |
+
"loss": 0.0002,
|
| 1420 |
+
"step": 4975
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.7391304347826086,
|
| 1424 |
+
"grad_norm": 0.006890356075018644,
|
| 1425 |
+
"learning_rate": 2.337670703426952e-05,
|
| 1426 |
+
"loss": 0.0002,
|
| 1427 |
+
"step": 5000
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 1.7478260869565219,
|
| 1431 |
+
"grad_norm": 0.0016472332645207644,
|
| 1432 |
+
"learning_rate": 2.321566606544705e-05,
|
| 1433 |
+
"loss": 0.0002,
|
| 1434 |
+
"step": 5025
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.7565217391304349,
|
| 1438 |
+
"grad_norm": 0.002031004289165139,
|
| 1439 |
+
"learning_rate": 2.305462509662458e-05,
|
| 1440 |
+
"loss": 0.0002,
|
| 1441 |
+
"step": 5050
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.7652173913043478,
|
| 1445 |
+
"grad_norm": 0.020191390067338943,
|
| 1446 |
+
"learning_rate": 2.2893584127802113e-05,
|
| 1447 |
+
"loss": 0.0006,
|
| 1448 |
+
"step": 5075
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.7739130434782608,
|
| 1452 |
+
"grad_norm": 0.0014919223031029105,
|
| 1453 |
+
"learning_rate": 2.273254315897965e-05,
|
| 1454 |
+
"loss": 0.0002,
|
| 1455 |
+
"step": 5100
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 1.7826086956521738,
|
| 1459 |
+
"grad_norm": 0.0019313708180561662,
|
| 1460 |
+
"learning_rate": 2.2571502190157177e-05,
|
| 1461 |
+
"loss": 0.0002,
|
| 1462 |
+
"step": 5125
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.7913043478260868,
|
| 1466 |
+
"grad_norm": 0.0022281836718320847,
|
| 1467 |
+
"learning_rate": 2.2410461221334708e-05,
|
| 1468 |
+
"loss": 0.0002,
|
| 1469 |
+
"step": 5150
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 1.8,
|
| 1473 |
+
"grad_norm": 0.0013652993366122246,
|
| 1474 |
+
"learning_rate": 2.224942025251224e-05,
|
| 1475 |
+
"loss": 0.0002,
|
| 1476 |
+
"step": 5175
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 1.808695652173913,
|
| 1480 |
+
"grad_norm": 0.00228705327026546,
|
| 1481 |
+
"learning_rate": 2.208837928368977e-05,
|
| 1482 |
+
"loss": 0.0002,
|
| 1483 |
+
"step": 5200
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.8173913043478263,
|
| 1487 |
+
"grad_norm": 0.001371930236928165,
|
| 1488 |
+
"learning_rate": 2.1927338314867303e-05,
|
| 1489 |
+
"loss": 0.0002,
|
| 1490 |
+
"step": 5225
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.8260869565217392,
|
| 1494 |
+
"grad_norm": 0.0013612714828923345,
|
| 1495 |
+
"learning_rate": 2.1766297346044835e-05,
|
| 1496 |
+
"loss": 0.0002,
|
| 1497 |
+
"step": 5250
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 1.8347826086956522,
|
| 1501 |
+
"grad_norm": 0.0018417349783703685,
|
| 1502 |
+
"learning_rate": 2.1605256377222367e-05,
|
| 1503 |
+
"loss": 0.0002,
|
| 1504 |
+
"step": 5275
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 1.8434782608695652,
|
| 1508 |
+
"grad_norm": 0.0018068786012008786,
|
| 1509 |
+
"learning_rate": 2.1444215408399898e-05,
|
| 1510 |
+
"loss": 0.0002,
|
| 1511 |
+
"step": 5300
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 1.8521739130434782,
|
| 1515 |
+
"grad_norm": 0.0039210072718560696,
|
| 1516 |
+
"learning_rate": 2.128317443957743e-05,
|
| 1517 |
+
"loss": 0.0002,
|
| 1518 |
+
"step": 5325
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 1.8608695652173912,
|
| 1522 |
+
"grad_norm": 0.002227018354460597,
|
| 1523 |
+
"learning_rate": 2.112213347075496e-05,
|
| 1524 |
+
"loss": 0.0156,
|
| 1525 |
+
"step": 5350
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.8695652173913042,
|
| 1529 |
+
"grad_norm": 0.001460268278606236,
|
| 1530 |
+
"learning_rate": 2.0961092501932493e-05,
|
| 1531 |
+
"loss": 0.0202,
|
| 1532 |
+
"step": 5375
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 1.8782608695652174,
|
| 1536 |
+
"grad_norm": 0.001318955677561462,
|
| 1537 |
+
"learning_rate": 2.0800051533110025e-05,
|
| 1538 |
+
"loss": 0.0101,
|
| 1539 |
+
"step": 5400
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 1.8869565217391304,
|
| 1543 |
+
"grad_norm": 0.0019615204073488712,
|
| 1544 |
+
"learning_rate": 2.0639010564287557e-05,
|
| 1545 |
+
"loss": 0.0002,
|
| 1546 |
+
"step": 5425
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 1.8956521739130436,
|
| 1550 |
+
"grad_norm": 0.0014419632498174906,
|
| 1551 |
+
"learning_rate": 2.0477969595465088e-05,
|
| 1552 |
+
"loss": 0.0002,
|
| 1553 |
+
"step": 5450
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 1.9043478260869566,
|
| 1557 |
+
"grad_norm": 0.00159283762332052,
|
| 1558 |
+
"learning_rate": 2.031692862664262e-05,
|
| 1559 |
+
"loss": 0.0002,
|
| 1560 |
+
"step": 5475
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 1.9130434782608696,
|
| 1564 |
+
"grad_norm": 0.0014433007454499602,
|
| 1565 |
+
"learning_rate": 2.0155887657820148e-05,
|
| 1566 |
+
"loss": 0.0002,
|
| 1567 |
+
"step": 5500
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.9217391304347826,
|
| 1571 |
+
"grad_norm": 0.0011873416369780898,
|
| 1572 |
+
"learning_rate": 1.9994846688997683e-05,
|
| 1573 |
+
"loss": 0.0002,
|
| 1574 |
+
"step": 5525
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 1.9304347826086956,
|
| 1578 |
+
"grad_norm": 0.0015813339268788695,
|
| 1579 |
+
"learning_rate": 1.9833805720175215e-05,
|
| 1580 |
+
"loss": 0.0002,
|
| 1581 |
+
"step": 5550
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 1.9391304347826086,
|
| 1585 |
+
"grad_norm": 0.001116643426939845,
|
| 1586 |
+
"learning_rate": 1.9672764751352747e-05,
|
| 1587 |
+
"loss": 0.0003,
|
| 1588 |
+
"step": 5575
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 1.9478260869565216,
|
| 1592 |
+
"grad_norm": 0.0013602287508547306,
|
| 1593 |
+
"learning_rate": 1.9511723782530278e-05,
|
| 1594 |
+
"loss": 0.0001,
|
| 1595 |
+
"step": 5600
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.9565217391304348,
|
| 1599 |
+
"grad_norm": 0.001626876532100141,
|
| 1600 |
+
"learning_rate": 1.9350682813707806e-05,
|
| 1601 |
+
"loss": 0.0002,
|
| 1602 |
+
"step": 5625
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 1.9652173913043478,
|
| 1606 |
+
"grad_norm": 0.0015673220623284578,
|
| 1607 |
+
"learning_rate": 1.9189641844885338e-05,
|
| 1608 |
+
"loss": 0.0012,
|
| 1609 |
+
"step": 5650
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 1.973913043478261,
|
| 1613 |
+
"grad_norm": 0.0014821627410128713,
|
| 1614 |
+
"learning_rate": 1.902860087606287e-05,
|
| 1615 |
+
"loss": 0.0001,
|
| 1616 |
+
"step": 5675
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 1.982608695652174,
|
| 1620 |
+
"grad_norm": 0.02127527818083763,
|
| 1621 |
+
"learning_rate": 1.8867559907240405e-05,
|
| 1622 |
+
"loss": 0.0491,
|
| 1623 |
+
"step": 5700
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 1.991304347826087,
|
| 1627 |
+
"grad_norm": 0.0015324688283726573,
|
| 1628 |
+
"learning_rate": 1.8706518938417936e-05,
|
| 1629 |
+
"loss": 0.0002,
|
| 1630 |
+
"step": 5725
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 2.0,
|
| 1634 |
+
"grad_norm": 0.0018975608982145786,
|
| 1635 |
+
"learning_rate": 1.8545477969595465e-05,
|
| 1636 |
+
"loss": 0.0002,
|
| 1637 |
+
"step": 5750
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 2.0,
|
| 1641 |
+
"eval_accuracy": 1.0,
|
| 1642 |
+
"eval_f1_macro": 1.0,
|
| 1643 |
+
"eval_f1_micro": 1.0,
|
| 1644 |
+
"eval_f1_weighted": 1.0,
|
| 1645 |
+
"eval_loss": 0.00012181053898530081,
|
| 1646 |
+
"eval_precision_macro": 1.0,
|
| 1647 |
+
"eval_precision_micro": 1.0,
|
| 1648 |
+
"eval_precision_weighted": 1.0,
|
| 1649 |
+
"eval_recall_macro": 1.0,
|
| 1650 |
+
"eval_recall_micro": 1.0,
|
| 1651 |
+
"eval_recall_weighted": 1.0,
|
| 1652 |
+
"eval_runtime": 5.0295,
|
| 1653 |
+
"eval_samples_per_second": 457.3,
|
| 1654 |
+
"eval_steps_per_second": 28.631,
|
| 1655 |
+
"step": 5750
|
| 1656 |
+
}
|
| 1657 |
+
],
|
| 1658 |
+
"logging_steps": 25,
|
| 1659 |
+
"max_steps": 8625,
|
| 1660 |
+
"num_input_tokens_seen": 0,
|
| 1661 |
+
"num_train_epochs": 3,
|
| 1662 |
+
"save_steps": 500,
|
| 1663 |
+
"stateful_callbacks": {
|
| 1664 |
+
"EarlyStoppingCallback": {
|
| 1665 |
+
"args": {
|
| 1666 |
+
"early_stopping_patience": 5,
|
| 1667 |
+
"early_stopping_threshold": 0.01
|
| 1668 |
+
},
|
| 1669 |
+
"attributes": {
|
| 1670 |
+
"early_stopping_patience_counter": 0
|
| 1671 |
+
}
|
| 1672 |
+
},
|
| 1673 |
+
"TrainerControl": {
|
| 1674 |
+
"args": {
|
| 1675 |
+
"should_epoch_stop": false,
|
| 1676 |
+
"should_evaluate": false,
|
| 1677 |
+
"should_log": false,
|
| 1678 |
+
"should_save": true,
|
| 1679 |
+
"should_training_stop": false
|
| 1680 |
+
},
|
| 1681 |
+
"attributes": {}
|
| 1682 |
+
}
|
| 1683 |
+
},
|
| 1684 |
+
"total_flos": 3026347648512000.0,
|
| 1685 |
+
"train_batch_size": 8,
|
| 1686 |
+
"trial_name": null,
|
| 1687 |
+
"trial_params": null
|
| 1688 |
+
}
|
checkpoint-5750/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:409762279f66edf691db8b627a5b3bf24d227881c2d962830b645d3dc4818c39
|
| 3 |
+
size 5240
|
config.json
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google-bert/bert-base-uncased",
|
| 3 |
+
"_num_labels": 23,
|
| 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": "BirthDate",
|
| 15 |
+
"1": "CountryCode",
|
| 16 |
+
"2": "EmailAddress",
|
| 17 |
+
"3": "FullPersonalName",
|
| 18 |
+
"4": "Gender",
|
| 19 |
+
"5": "GivenName",
|
| 20 |
+
"6": "IpAddress",
|
| 21 |
+
"7": "Latitude",
|
| 22 |
+
"8": "Locality",
|
| 23 |
+
"9": "Longitude",
|
| 24 |
+
"10": "MiddleName",
|
| 25 |
+
"11": "NationalIdentificationNumber",
|
| 26 |
+
"12": "OrganizationName",
|
| 27 |
+
"13": "PaymentCardNumber",
|
| 28 |
+
"14": "PostalCode",
|
| 29 |
+
"15": "PrimaryStreetAddress",
|
| 30 |
+
"16": "SecondaryStreetAddress",
|
| 31 |
+
"17": "Subdivision",
|
| 32 |
+
"18": "Surname",
|
| 33 |
+
"19": "TelephoneNumber",
|
| 34 |
+
"20": "Uncategorized",
|
| 35 |
+
"21": "UniqueIdentificationNumber",
|
| 36 |
+
"22": "Uuid"
|
| 37 |
+
},
|
| 38 |
+
"initializer_range": 0.02,
|
| 39 |
+
"intermediate_size": 3072,
|
| 40 |
+
"label2id": {
|
| 41 |
+
"BirthDate": 0,
|
| 42 |
+
"CountryCode": 1,
|
| 43 |
+
"EmailAddress": 2,
|
| 44 |
+
"FullPersonalName": 3,
|
| 45 |
+
"Gender": 4,
|
| 46 |
+
"GivenName": 5,
|
| 47 |
+
"IpAddress": 6,
|
| 48 |
+
"Latitude": 7,
|
| 49 |
+
"Locality": 8,
|
| 50 |
+
"Longitude": 9,
|
| 51 |
+
"MiddleName": 10,
|
| 52 |
+
"NationalIdentificationNumber": 11,
|
| 53 |
+
"OrganizationName": 12,
|
| 54 |
+
"PaymentCardNumber": 13,
|
| 55 |
+
"PostalCode": 14,
|
| 56 |
+
"PrimaryStreetAddress": 15,
|
| 57 |
+
"SecondaryStreetAddress": 16,
|
| 58 |
+
"Subdivision": 17,
|
| 59 |
+
"Surname": 18,
|
| 60 |
+
"TelephoneNumber": 19,
|
| 61 |
+
"Uncategorized": 20,
|
| 62 |
+
"UniqueIdentificationNumber": 21,
|
| 63 |
+
"Uuid": 22
|
| 64 |
+
},
|
| 65 |
+
"layer_norm_eps": 1e-12,
|
| 66 |
+
"max_position_embeddings": 512,
|
| 67 |
+
"model_type": "bert",
|
| 68 |
+
"num_attention_heads": 12,
|
| 69 |
+
"num_hidden_layers": 12,
|
| 70 |
+
"pad_token_id": 0,
|
| 71 |
+
"position_embedding_type": "absolute",
|
| 72 |
+
"problem_type": "single_label_classification",
|
| 73 |
+
"torch_dtype": "float32",
|
| 74 |
+
"transformers_version": "4.44.0",
|
| 75 |
+
"type_vocab_size": 2,
|
| 76 |
+
"use_cache": true,
|
| 77 |
+
"vocab_size": 30522
|
| 78 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84f5a6008b9ad36981f508eb2908e9ee9a46225e32728f516f993b59622b6890
|
| 3 |
+
size 438023244
|
runs/Aug16_02-58-16_r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m/events.out.tfevents.1723777098.r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m.81.0
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 sha256:1e50f205232783fdc316c59ed1bced0f3a4f04f6547c9ddab89c8a5539264701
|
| 3 |
+
size 81806
|
runs/Aug16_02-58-16_r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m/events.out.tfevents.1723777914.r-oulmokhtar-trainer-0uacugq3-aacd9-urx0m.81.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ff1bfc8f2af24c4d61c4438c2aaaead748f7594dff0c8b0ec252933619d21d19
|
| 3 |
+
size 921
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "BertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:409762279f66edf691db8b627a5b3bf24d227881c2d962830b645d3dc4818c39
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "autotrain-l9v4l-l3gs1/autotrain-data",
|
| 3 |
+
"model": "google-bert/bert-base-uncased",
|
| 4 |
+
"lr": 5e-05,
|
| 5 |
+
"epochs": 3,
|
| 6 |
+
"max_seq_length": 128,
|
| 7 |
+
"batch_size": 8,
|
| 8 |
+
"warmup_ratio": 0.1,
|
| 9 |
+
"gradient_accumulation": 1,
|
| 10 |
+
"optimizer": "adamw_torch",
|
| 11 |
+
"scheduler": "linear",
|
| 12 |
+
"weight_decay": 0.0,
|
| 13 |
+
"max_grad_norm": 1.0,
|
| 14 |
+
"seed": 42,
|
| 15 |
+
"train_split": "train",
|
| 16 |
+
"valid_split": "validation",
|
| 17 |
+
"text_column": "autotrain_text",
|
| 18 |
+
"target_column": "autotrain_label",
|
| 19 |
+
"logging_steps": -1,
|
| 20 |
+
"project_name": "autotrain-l9v4l-l3gs1",
|
| 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": "oulmokhtar",
|
| 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
|
|
|