ilsilfverskiold/iptc-newscodes-multilingual-text-classification
Browse files- README.md +22 -21
- all_results.json +16 -16
- config.json +10 -10
- eval_results.json +16 -16
- model.safetensors +2 -2
- runs/May16_08-11-17_22e3c5cf128a/events.out.tfevents.1715847078.22e3c5cf128a.2253.0 +3 -0
- runs/May16_08-11-17_22e3c5cf128a/events.out.tfevents.1715848420.22e3c5cf128a.2253.1 +3 -0
- special_tokens_map.json +5 -13
- tokenizer.json +0 -0
- tokenizer_config.json +23 -23
- trainer_state.json +1019 -599
- training_args.bin +1 -1
- vocab.txt +0 -0
README.md
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
base_model: roberta-base
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
|
@@ -18,27 +17,27 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
# news_category_classification
|
| 20 |
|
| 21 |
-
This model is a fine-tuned version of [
|
| 22 |
It achieves the following results on the evaluation set:
|
| 23 |
-
- Loss: 0.
|
| 24 |
-
- Accuracy: 0.
|
| 25 |
-
- F1: 0.
|
| 26 |
-
- Precision: 0.
|
| 27 |
-
- Recall: 0.
|
| 28 |
- Accuracy Label Arts, culture, entertainment and media: 0.8333
|
| 29 |
-
- Accuracy Label Conflict, war and peace: 0.
|
| 30 |
-
- Accuracy Label Crime, law and justice: 0.
|
| 31 |
- Accuracy Label Disaster, accident, and emergency incident: 0.8931
|
| 32 |
-
- Accuracy Label Economy, business, and finance: 0.
|
| 33 |
-
- Accuracy Label Environment: 0.
|
| 34 |
-
- Accuracy Label Health: 0.
|
| 35 |
- Accuracy Label Human interest: 0.3333
|
| 36 |
- Accuracy Label Labour: 0.5
|
| 37 |
- Accuracy Label Lifestyle and leisure: 0.5
|
| 38 |
-
- Accuracy Label Politics: 0.
|
| 39 |
- Accuracy Label Religion: 0.0
|
| 40 |
- Accuracy Label Science and technology: 0.4167
|
| 41 |
-
- Accuracy Label Society: 0.
|
| 42 |
- Accuracy Label Sport: 0.9615
|
| 43 |
- Accuracy Label Weather: 1.0
|
| 44 |
|
|
@@ -74,12 +73,14 @@ The following hyperparameters were used during training:
|
|
| 74 |
|
| 75 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | Accuracy Label Arts, culture, entertainment and media | Accuracy Label Conflict, war and peace | Accuracy Label Crime, law and justice | Accuracy Label Disaster, accident, and emergency incident | Accuracy Label Economy, business, and finance | Accuracy Label Environment | Accuracy Label Health | Accuracy Label Human interest | Accuracy Label Labour | Accuracy Label Lifestyle and leisure | Accuracy Label Politics | Accuracy Label Religion | Accuracy Label Science and technology | Accuracy Label Society | Accuracy Label Sport | Accuracy Label Weather |
|
| 76 |
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|:-----------------------------------------------------:|:--------------------------------------:|:-------------------------------------:|:---------------------------------------------------------:|:---------------------------------------------:|:--------------------------:|:---------------------:|:-----------------------------:|:---------------------:|:------------------------------------:|:-----------------------:|:-----------------------:|:-------------------------------------:|:----------------------:|:--------------------:|:----------------------:|
|
| 77 |
-
| 1.
|
| 78 |
-
| 1.
|
| 79 |
-
| 0.
|
| 80 |
-
| 0.
|
| 81 |
-
| 0.
|
| 82 |
-
| 0.
|
|
|
|
|
|
|
| 83 |
|
| 84 |
|
| 85 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
+
base_model: KB/bert-base-swedish-cased
|
|
|
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
metrics:
|
|
|
|
| 17 |
|
| 18 |
# news_category_classification
|
| 19 |
|
| 20 |
+
This model is a fine-tuned version of [KB/bert-base-swedish-cased](https://huggingface.co/KB/bert-base-swedish-cased) on the None dataset.
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.8615
|
| 23 |
+
- Accuracy: 0.7286
|
| 24 |
+
- F1: 0.7300
|
| 25 |
+
- Precision: 0.7351
|
| 26 |
+
- Recall: 0.7286
|
| 27 |
- Accuracy Label Arts, culture, entertainment and media: 0.8333
|
| 28 |
+
- Accuracy Label Conflict, war and peace: 0.7234
|
| 29 |
+
- Accuracy Label Crime, law and justice: 0.7919
|
| 30 |
- Accuracy Label Disaster, accident, and emergency incident: 0.8931
|
| 31 |
+
- Accuracy Label Economy, business, and finance: 0.7975
|
| 32 |
+
- Accuracy Label Environment: 0.4375
|
| 33 |
+
- Accuracy Label Health: 0.7
|
| 34 |
- Accuracy Label Human interest: 0.3333
|
| 35 |
- Accuracy Label Labour: 0.5
|
| 36 |
- Accuracy Label Lifestyle and leisure: 0.5
|
| 37 |
+
- Accuracy Label Politics: 0.6331
|
| 38 |
- Accuracy Label Religion: 0.0
|
| 39 |
- Accuracy Label Science and technology: 0.4167
|
| 40 |
+
- Accuracy Label Society: 0.4561
|
| 41 |
- Accuracy Label Sport: 0.9615
|
| 42 |
- Accuracy Label Weather: 1.0
|
| 43 |
|
|
|
|
| 73 |
|
| 74 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | Accuracy Label Arts, culture, entertainment and media | Accuracy Label Conflict, war and peace | Accuracy Label Crime, law and justice | Accuracy Label Disaster, accident, and emergency incident | Accuracy Label Economy, business, and finance | Accuracy Label Environment | Accuracy Label Health | Accuracy Label Human interest | Accuracy Label Labour | Accuracy Label Lifestyle and leisure | Accuracy Label Politics | Accuracy Label Religion | Accuracy Label Science and technology | Accuracy Label Society | Accuracy Label Sport | Accuracy Label Weather |
|
| 75 |
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|:-----------------------------------------------------:|:--------------------------------------:|:-------------------------------------:|:---------------------------------------------------------:|:---------------------------------------------:|:--------------------------:|:---------------------:|:-----------------------------:|:---------------------:|:------------------------------------:|:-----------------------:|:-----------------------:|:-------------------------------------:|:----------------------:|:--------------------:|:----------------------:|
|
| 76 |
+
| 1.7671 | 0.3373 | 200 | 1.5661 | 0.5554 | 0.5206 | 0.5828 | 0.5554 | 0.5833 | 0.7553 | 0.8960 | 0.3206 | 0.6709 | 0.125 | 0.7 | 0.0 | 0.5 | 0.5 | 0.2878 | 0.0 | 0.0 | 0.0351 | 0.9615 | 1.0 |
|
| 77 |
+
| 1.0248 | 0.6745 | 400 | 1.0774 | 0.6709 | 0.6591 | 0.6984 | 0.6709 | 0.9167 | 0.7979 | 0.8150 | 0.8626 | 0.7215 | 0.375 | 0.9 | 0.25 | 1.0 | 0.5 | 0.3094 | 0.0 | 0.4167 | 0.1930 | 0.9615 | 1.0 |
|
| 78 |
+
| 0.5845 | 1.0118 | 600 | 0.9907 | 0.6536 | 0.6563 | 0.6829 | 0.6536 | 0.9167 | 0.7287 | 0.6763 | 0.8779 | 0.7215 | 0.4375 | 0.8 | 0.0 | 1.0 | 0.75 | 0.3669 | 0.0 | 0.4167 | 0.4386 | 0.9231 | 1.0 |
|
| 79 |
+
| 0.6104 | 1.3491 | 800 | 0.8674 | 0.7240 | 0.7233 | 0.7333 | 0.7240 | 0.8333 | 0.7021 | 0.8324 | 0.8779 | 0.7848 | 0.5 | 0.7 | 0.25 | 1.0 | 0.75 | 0.6331 | 0.0 | 0.25 | 0.3684 | 0.9615 | 1.0 |
|
| 80 |
+
| 0.4223 | 1.6863 | 1000 | 0.8602 | 0.7240 | 0.7250 | 0.7387 | 0.7240 | 0.75 | 0.6755 | 0.8844 | 0.8550 | 0.7342 | 0.5 | 0.9 | 0.3333 | 1.0 | 0.625 | 0.6475 | 0.0 | 0.3333 | 0.3684 | 0.9615 | 0.0 |
|
| 81 |
+
| 0.3104 | 2.0236 | 1200 | 0.8565 | 0.7263 | 0.7266 | 0.7326 | 0.7263 | 0.8333 | 0.7181 | 0.8324 | 0.9084 | 0.7722 | 0.4375 | 0.7 | 0.25 | 0.5 | 0.75 | 0.5612 | 0.0 | 0.4167 | 0.4737 | 0.9615 | 1.0 |
|
| 82 |
+
| 0.2855 | 2.3609 | 1400 | 0.8981 | 0.7240 | 0.7283 | 0.7402 | 0.7240 | 0.75 | 0.7394 | 0.8324 | 0.8550 | 0.7975 | 0.5 | 0.7 | 0.3333 | 0.5 | 0.625 | 0.5899 | 0.0 | 0.4167 | 0.3860 | 0.9615 | 1.0 |
|
| 83 |
+
| 0.217 | 2.6981 | 1600 | 0.8667 | 0.7309 | 0.7292 | 0.7358 | 0.7309 | 0.75 | 0.7447 | 0.8382 | 0.8931 | 0.8481 | 0.375 | 0.8 | 0.3333 | 0.5 | 0.5 | 0.5396 | 0.0 | 0.4167 | 0.4561 | 0.9615 | 1.0 |
|
| 84 |
|
| 85 |
|
| 86 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,28 +1,28 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 5 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 6 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 7 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 8 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 9 |
-
"eval_accuracy_label_environment": 0.
|
| 10 |
-
"eval_accuracy_label_health": 0.
|
| 11 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 12 |
"eval_accuracy_label_labour": 0.5,
|
| 13 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 14 |
-
"eval_accuracy_label_politics": 0.
|
| 15 |
"eval_accuracy_label_religion": 0.0,
|
| 16 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 17 |
-
"eval_accuracy_label_society": 0.
|
| 18 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 19 |
"eval_accuracy_label_weather": 1.0,
|
| 20 |
-
"eval_f1": 0.
|
| 21 |
-
"eval_loss": 0.
|
| 22 |
-
"eval_precision": 0.
|
| 23 |
-
"eval_recall": 0.
|
| 24 |
-
"eval_runtime": 6.
|
| 25 |
-
"eval_samples_per_second":
|
| 26 |
-
"eval_steps_per_second": 8.
|
| 27 |
-
"step":
|
| 28 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.7286374133949192,
|
| 4 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 5 |
+
"eval_accuracy_label_conflict, war and peace": 0.723404255319149,
|
| 6 |
+
"eval_accuracy_label_crime, law and justice": 0.791907514450867,
|
| 7 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 8 |
+
"eval_accuracy_label_economy, business, and finance": 0.7974683544303798,
|
| 9 |
+
"eval_accuracy_label_environment": 0.4375,
|
| 10 |
+
"eval_accuracy_label_health": 0.7,
|
| 11 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 12 |
"eval_accuracy_label_labour": 0.5,
|
| 13 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 14 |
+
"eval_accuracy_label_politics": 0.6330935251798561,
|
| 15 |
"eval_accuracy_label_religion": 0.0,
|
| 16 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 17 |
+
"eval_accuracy_label_society": 0.45614035087719296,
|
| 18 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 19 |
"eval_accuracy_label_weather": 1.0,
|
| 20 |
+
"eval_f1": 0.730012835488787,
|
| 21 |
+
"eval_loss": 0.8615460395812988,
|
| 22 |
+
"eval_precision": 0.7350865087902028,
|
| 23 |
+
"eval_recall": 0.7286374133949192,
|
| 24 |
+
"eval_runtime": 6.8008,
|
| 25 |
+
"eval_samples_per_second": 127.338,
|
| 26 |
+
"eval_steps_per_second": 8.087,
|
| 27 |
+
"step": 1779
|
| 28 |
}
|
config.json
CHANGED
|
@@ -1,12 +1,11 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"architectures": [
|
| 4 |
-
"
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
-
"bos_token_id": 0,
|
| 8 |
"classifier_dropout": null,
|
| 9 |
-
"
|
| 10 |
"hidden_act": "gelu",
|
| 11 |
"hidden_dropout_prob": 0.1,
|
| 12 |
"hidden_size": 768,
|
|
@@ -48,17 +47,18 @@
|
|
| 48 |
"sport": 14,
|
| 49 |
"weather": 15
|
| 50 |
},
|
| 51 |
-
"layer_norm_eps": 1e-
|
| 52 |
-
"max_position_embeddings":
|
| 53 |
-
"model_type": "
|
| 54 |
"num_attention_heads": 12,
|
| 55 |
"num_hidden_layers": 12,
|
| 56 |
-
"
|
|
|
|
| 57 |
"position_embedding_type": "absolute",
|
| 58 |
"problem_type": "single_label_classification",
|
| 59 |
"torch_dtype": "float32",
|
| 60 |
"transformers_version": "4.40.2",
|
| 61 |
-
"type_vocab_size":
|
| 62 |
"use_cache": true,
|
| 63 |
-
"vocab_size":
|
| 64 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "KB/bert-base-swedish-cased",
|
| 3 |
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
|
|
|
| 7 |
"classifier_dropout": null,
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
"hidden_act": "gelu",
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
"hidden_size": 768,
|
|
|
|
| 47 |
"sport": 14,
|
| 48 |
"weather": 15
|
| 49 |
},
|
| 50 |
+
"layer_norm_eps": 1e-12,
|
| 51 |
+
"max_position_embeddings": 512,
|
| 52 |
+
"model_type": "bert",
|
| 53 |
"num_attention_heads": 12,
|
| 54 |
"num_hidden_layers": 12,
|
| 55 |
+
"output_past": true,
|
| 56 |
+
"pad_token_id": 0,
|
| 57 |
"position_embedding_type": "absolute",
|
| 58 |
"problem_type": "single_label_classification",
|
| 59 |
"torch_dtype": "float32",
|
| 60 |
"transformers_version": "4.40.2",
|
| 61 |
+
"type_vocab_size": 2,
|
| 62 |
"use_cache": true,
|
| 63 |
+
"vocab_size": 50325
|
| 64 |
}
|
eval_results.json
CHANGED
|
@@ -1,28 +1,28 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 5 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 6 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 7 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 8 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 9 |
-
"eval_accuracy_label_environment": 0.
|
| 10 |
-
"eval_accuracy_label_health": 0.
|
| 11 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 12 |
"eval_accuracy_label_labour": 0.5,
|
| 13 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 14 |
-
"eval_accuracy_label_politics": 0.
|
| 15 |
"eval_accuracy_label_religion": 0.0,
|
| 16 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 17 |
-
"eval_accuracy_label_society": 0.
|
| 18 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 19 |
"eval_accuracy_label_weather": 1.0,
|
| 20 |
-
"eval_f1": 0.
|
| 21 |
-
"eval_loss": 0.
|
| 22 |
-
"eval_precision": 0.
|
| 23 |
-
"eval_recall": 0.
|
| 24 |
-
"eval_runtime": 6.
|
| 25 |
-
"eval_samples_per_second":
|
| 26 |
-
"eval_steps_per_second": 8.
|
| 27 |
-
"step":
|
| 28 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 3.0,
|
| 3 |
+
"eval_accuracy": 0.7286374133949192,
|
| 4 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 5 |
+
"eval_accuracy_label_conflict, war and peace": 0.723404255319149,
|
| 6 |
+
"eval_accuracy_label_crime, law and justice": 0.791907514450867,
|
| 7 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 8 |
+
"eval_accuracy_label_economy, business, and finance": 0.7974683544303798,
|
| 9 |
+
"eval_accuracy_label_environment": 0.4375,
|
| 10 |
+
"eval_accuracy_label_health": 0.7,
|
| 11 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 12 |
"eval_accuracy_label_labour": 0.5,
|
| 13 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 14 |
+
"eval_accuracy_label_politics": 0.6330935251798561,
|
| 15 |
"eval_accuracy_label_religion": 0.0,
|
| 16 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 17 |
+
"eval_accuracy_label_society": 0.45614035087719296,
|
| 18 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 19 |
"eval_accuracy_label_weather": 1.0,
|
| 20 |
+
"eval_f1": 0.730012835488787,
|
| 21 |
+
"eval_loss": 0.8615460395812988,
|
| 22 |
+
"eval_precision": 0.7350865087902028,
|
| 23 |
+
"eval_recall": 0.7286374133949192,
|
| 24 |
+
"eval_runtime": 6.8008,
|
| 25 |
+
"eval_samples_per_second": 127.338,
|
| 26 |
+
"eval_steps_per_second": 8.087,
|
| 27 |
+
"step": 1779
|
| 28 |
}
|
model.safetensors
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 sha256:09c99b75510478d48af78417488eddf8f3ee52da188217f9322228944ac4d7e7
|
| 3 |
+
size 498836536
|
runs/May16_08-11-17_22e3c5cf128a/events.out.tfevents.1715847078.22e3c5cf128a.2253.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2a08607f5df1f08c964ac3da4bf398059696aee29b09bdff7f1ea3824f21553
|
| 3 |
+
size 56850
|
runs/May16_08-11-17_22e3c5cf128a/events.out.tfevents.1715848420.22e3c5cf128a.2253.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a6ec24359017c59f05b627011f5e93e56df9a9a7b73f062a628826c92c1942de
|
| 3 |
+
size 1774
|
special_tokens_map.json
CHANGED
|
@@ -1,15 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
"
|
| 4 |
-
"
|
| 5 |
-
"
|
| 6 |
-
|
| 7 |
-
"lstrip": true,
|
| 8 |
-
"normalized": false,
|
| 9 |
-
"rstrip": false,
|
| 10 |
-
"single_word": false
|
| 11 |
-
},
|
| 12 |
-
"pad_token": "<pad>",
|
| 13 |
-
"sep_token": "</s>",
|
| 14 |
-
"unk_token": "<unk>"
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
}
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -1,57 +1,57 @@
|
|
| 1 |
{
|
| 2 |
-
"add_prefix_space": false,
|
| 3 |
"added_tokens_decoder": {
|
| 4 |
"0": {
|
| 5 |
-
"content": "
|
| 6 |
"lstrip": false,
|
| 7 |
-
"normalized":
|
| 8 |
"rstrip": false,
|
| 9 |
"single_word": false,
|
| 10 |
"special": true
|
| 11 |
},
|
| 12 |
"1": {
|
| 13 |
-
"content": "
|
| 14 |
"lstrip": false,
|
| 15 |
-
"normalized":
|
| 16 |
"rstrip": false,
|
| 17 |
"single_word": false,
|
| 18 |
"special": true
|
| 19 |
},
|
| 20 |
"2": {
|
| 21 |
-
"content": "
|
| 22 |
"lstrip": false,
|
| 23 |
-
"normalized":
|
| 24 |
"rstrip": false,
|
| 25 |
"single_word": false,
|
| 26 |
"special": true
|
| 27 |
},
|
| 28 |
"3": {
|
| 29 |
-
"content": "
|
| 30 |
"lstrip": false,
|
| 31 |
-
"normalized":
|
| 32 |
"rstrip": false,
|
| 33 |
"single_word": false,
|
| 34 |
"special": true
|
| 35 |
},
|
| 36 |
-
"
|
| 37 |
-
"content": "
|
| 38 |
-
"lstrip":
|
| 39 |
"normalized": false,
|
| 40 |
"rstrip": false,
|
| 41 |
"single_word": false,
|
| 42 |
"special": true
|
| 43 |
}
|
| 44 |
},
|
| 45 |
-
"bos_token": "<s>",
|
| 46 |
"clean_up_tokenization_spaces": true,
|
| 47 |
-
"cls_token": "
|
| 48 |
-
"
|
| 49 |
-
"
|
| 50 |
-
"mask_token": "
|
| 51 |
-
"model_max_length":
|
| 52 |
-
"
|
| 53 |
-
"
|
| 54 |
-
"
|
| 55 |
-
"
|
| 56 |
-
"
|
|
|
|
|
|
|
| 57 |
}
|
|
|
|
| 1 |
{
|
|
|
|
| 2 |
"added_tokens_decoder": {
|
| 3 |
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
"rstrip": false,
|
| 8 |
"single_word": false,
|
| 9 |
"special": true
|
| 10 |
},
|
| 11 |
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
"rstrip": false,
|
| 16 |
"single_word": false,
|
| 17 |
"special": true
|
| 18 |
},
|
| 19 |
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
"rstrip": false,
|
| 24 |
"single_word": false,
|
| 25 |
"special": true
|
| 26 |
},
|
| 27 |
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
"rstrip": false,
|
| 32 |
"single_word": false,
|
| 33 |
"special": true
|
| 34 |
},
|
| 35 |
+
"4": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
"normalized": false,
|
| 39 |
"rstrip": false,
|
| 40 |
"single_word": false,
|
| 41 |
"special": true
|
| 42 |
}
|
| 43 |
},
|
|
|
|
| 44 |
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"mask_token": "[MASK]",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"never_split": null,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"sep_token": "[SEP]",
|
| 53 |
+
"strip_accents": false,
|
| 54 |
+
"tokenize_chinese_chars": true,
|
| 55 |
+
"tokenizer_class": "BertTokenizer",
|
| 56 |
+
"unk_token": "[UNK]"
|
| 57 |
}
|
trainer_state.json
CHANGED
|
@@ -3,1098 +3,1518 @@
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 3.0,
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 4.0000000000000003e-07,
|
| 15 |
-
"loss": 2.
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
"learning_rate": 8.000000000000001e-07,
|
| 22 |
-
"loss": 2.
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm":
|
| 28 |
"learning_rate": 1.2000000000000002e-06,
|
| 29 |
-
"loss": 2.
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"grad_norm":
|
| 35 |
"learning_rate": 1.6000000000000001e-06,
|
| 36 |
-
"loss": 2.
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0.
|
| 41 |
-
"grad_norm":
|
| 42 |
"learning_rate": 2.0000000000000003e-06,
|
| 43 |
-
"loss": 2.
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch": 0.
|
| 48 |
-
"grad_norm":
|
| 49 |
"learning_rate": 2.4000000000000003e-06,
|
| 50 |
-
"loss": 2.
|
| 51 |
"step": 60
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"grad_norm":
|
| 56 |
"learning_rate": 2.8000000000000003e-06,
|
| 57 |
-
"loss": 2.
|
| 58 |
"step": 70
|
| 59 |
},
|
| 60 |
{
|
| 61 |
-
"epoch": 0.
|
| 62 |
-
"grad_norm":
|
| 63 |
"learning_rate": 3.2000000000000003e-06,
|
| 64 |
-
"loss": 2.
|
| 65 |
"step": 80
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch": 0.
|
| 69 |
-
"grad_norm":
|
| 70 |
"learning_rate": 3.6000000000000003e-06,
|
| 71 |
-
"loss": 2.
|
| 72 |
"step": 90
|
| 73 |
},
|
| 74 |
{
|
| 75 |
-
"epoch": 0.
|
| 76 |
-
"grad_norm":
|
| 77 |
"learning_rate": 4.000000000000001e-06,
|
| 78 |
-
"loss": 2.
|
| 79 |
"step": 100
|
| 80 |
},
|
| 81 |
{
|
| 82 |
-
"epoch": 0.
|
| 83 |
-
"grad_norm":
|
| 84 |
"learning_rate": 4.4e-06,
|
| 85 |
-
"loss": 2.
|
| 86 |
"step": 110
|
| 87 |
},
|
| 88 |
{
|
| 89 |
-
"epoch": 0.
|
| 90 |
-
"grad_norm":
|
| 91 |
"learning_rate": 4.800000000000001e-06,
|
| 92 |
-
"loss": 2.
|
| 93 |
"step": 120
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch": 0.
|
| 97 |
-
"grad_norm": 6.
|
| 98 |
"learning_rate": 5.2e-06,
|
| 99 |
-
"loss": 2.
|
| 100 |
"step": 130
|
| 101 |
},
|
| 102 |
{
|
| 103 |
-
"epoch": 0.
|
| 104 |
-
"grad_norm": 6.
|
| 105 |
"learning_rate": 5.600000000000001e-06,
|
| 106 |
-
"loss": 2.
|
| 107 |
"step": 140
|
| 108 |
},
|
| 109 |
{
|
| 110 |
-
"epoch": 0.
|
| 111 |
-
"grad_norm":
|
| 112 |
"learning_rate": 6e-06,
|
| 113 |
-
"loss": 2.
|
| 114 |
"step": 150
|
| 115 |
},
|
| 116 |
{
|
| 117 |
-
"epoch": 0.
|
| 118 |
-
"grad_norm":
|
| 119 |
"learning_rate": 6.4000000000000006e-06,
|
| 120 |
-
"loss": 2.
|
| 121 |
"step": 160
|
| 122 |
},
|
| 123 |
{
|
| 124 |
-
"epoch": 0.
|
| 125 |
-
"grad_norm":
|
| 126 |
"learning_rate": 6.800000000000001e-06,
|
| 127 |
-
"loss":
|
| 128 |
"step": 170
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch": 0.
|
| 132 |
-
"grad_norm": 6.
|
| 133 |
"learning_rate": 7.2000000000000005e-06,
|
| 134 |
-
"loss": 1.
|
| 135 |
"step": 180
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch": 0.
|
| 139 |
-
"grad_norm":
|
| 140 |
"learning_rate": 7.600000000000001e-06,
|
| 141 |
-
"loss": 1.
|
| 142 |
"step": 190
|
| 143 |
},
|
| 144 |
{
|
| 145 |
-
"epoch": 0.
|
| 146 |
-
"grad_norm":
|
| 147 |
"learning_rate": 8.000000000000001e-06,
|
| 148 |
-
"loss": 1.
|
| 149 |
"step": 200
|
| 150 |
},
|
| 151 |
{
|
| 152 |
-
"epoch": 0.
|
| 153 |
-
"eval_accuracy": 0.
|
| 154 |
-
"eval_accuracy_label_arts, culture, entertainment and media": 0.
|
| 155 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 156 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 157 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 158 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 159 |
-
"eval_accuracy_label_environment": 0.
|
| 160 |
-
"eval_accuracy_label_health": 0.
|
| 161 |
"eval_accuracy_label_human interest": 0.0,
|
| 162 |
-
"eval_accuracy_label_labour": 0.
|
| 163 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 164 |
-
"eval_accuracy_label_politics": 0.
|
| 165 |
"eval_accuracy_label_religion": 0.0,
|
| 166 |
"eval_accuracy_label_science and technology": 0.0,
|
| 167 |
-
"eval_accuracy_label_society": 0.
|
| 168 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 169 |
-
"eval_accuracy_label_weather":
|
| 170 |
-
"eval_f1": 0.
|
| 171 |
-
"eval_loss": 1.
|
| 172 |
-
"eval_precision": 0.
|
| 173 |
-
"eval_recall": 0.
|
| 174 |
-
"eval_runtime": 6.
|
| 175 |
-
"eval_samples_per_second":
|
| 176 |
-
"eval_steps_per_second":
|
| 177 |
"step": 200
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch": 0.
|
| 181 |
-
"grad_norm": 7.
|
| 182 |
"learning_rate": 8.400000000000001e-06,
|
| 183 |
-
"loss": 1.
|
| 184 |
"step": 210
|
| 185 |
},
|
| 186 |
{
|
| 187 |
-
"epoch": 0.
|
| 188 |
-
"grad_norm":
|
| 189 |
"learning_rate": 8.8e-06,
|
| 190 |
-
"loss": 1.
|
| 191 |
"step": 220
|
| 192 |
},
|
| 193 |
{
|
| 194 |
-
"epoch": 0.
|
| 195 |
-
"grad_norm":
|
| 196 |
"learning_rate": 9.200000000000002e-06,
|
| 197 |
-
"loss": 1.
|
| 198 |
"step": 230
|
| 199 |
},
|
| 200 |
{
|
| 201 |
-
"epoch": 0.
|
| 202 |
-
"grad_norm":
|
| 203 |
"learning_rate": 9.600000000000001e-06,
|
| 204 |
-
"loss": 1.
|
| 205 |
"step": 240
|
| 206 |
},
|
| 207 |
{
|
| 208 |
-
"epoch": 0.
|
| 209 |
-
"grad_norm": 13.
|
| 210 |
"learning_rate": 1e-05,
|
| 211 |
-
"loss": 1.
|
| 212 |
"step": 250
|
| 213 |
},
|
| 214 |
{
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"grad_norm":
|
| 217 |
"learning_rate": 1.04e-05,
|
| 218 |
-
"loss": 1.
|
| 219 |
"step": 260
|
| 220 |
},
|
| 221 |
{
|
| 222 |
-
"epoch": 0.
|
| 223 |
-
"grad_norm":
|
| 224 |
"learning_rate": 1.0800000000000002e-05,
|
| 225 |
-
"loss": 1.
|
| 226 |
"step": 270
|
| 227 |
},
|
| 228 |
{
|
| 229 |
-
"epoch": 0.
|
| 230 |
-
"grad_norm":
|
| 231 |
"learning_rate": 1.1200000000000001e-05,
|
| 232 |
-
"loss": 1.
|
| 233 |
"step": 280
|
| 234 |
},
|
| 235 |
{
|
| 236 |
-
"epoch": 0.
|
| 237 |
-
"grad_norm":
|
| 238 |
"learning_rate": 1.16e-05,
|
| 239 |
-
"loss": 1.
|
| 240 |
"step": 290
|
| 241 |
},
|
| 242 |
{
|
| 243 |
-
"epoch": 0.
|
| 244 |
-
"grad_norm":
|
| 245 |
"learning_rate": 1.2e-05,
|
| 246 |
-
"loss": 1.
|
| 247 |
"step": 300
|
| 248 |
},
|
| 249 |
{
|
| 250 |
-
"epoch": 0.
|
| 251 |
-
"grad_norm": 14.
|
| 252 |
"learning_rate": 1.2400000000000002e-05,
|
| 253 |
-
"loss": 1.
|
| 254 |
"step": 310
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch": 0.
|
| 258 |
-
"grad_norm":
|
| 259 |
"learning_rate": 1.2800000000000001e-05,
|
| 260 |
-
"loss": 1.
|
| 261 |
"step": 320
|
| 262 |
},
|
| 263 |
{
|
| 264 |
-
"epoch": 0.
|
| 265 |
-
"grad_norm":
|
| 266 |
"learning_rate": 1.3200000000000002e-05,
|
| 267 |
-
"loss": 1.
|
| 268 |
"step": 330
|
| 269 |
},
|
| 270 |
{
|
| 271 |
-
"epoch": 0.
|
| 272 |
-
"grad_norm":
|
| 273 |
"learning_rate": 1.3600000000000002e-05,
|
| 274 |
-
"loss": 1.
|
| 275 |
"step": 340
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"epoch": 0.
|
| 279 |
-
"grad_norm":
|
| 280 |
"learning_rate": 1.4e-05,
|
| 281 |
-
"loss": 1.
|
| 282 |
"step": 350
|
| 283 |
},
|
| 284 |
{
|
| 285 |
-
"epoch": 0.
|
| 286 |
-
"grad_norm":
|
| 287 |
"learning_rate": 1.4400000000000001e-05,
|
| 288 |
-
"loss": 1.
|
| 289 |
"step": 360
|
| 290 |
},
|
| 291 |
{
|
| 292 |
-
"epoch": 0.
|
| 293 |
-
"grad_norm":
|
| 294 |
"learning_rate": 1.48e-05,
|
| 295 |
-
"loss": 1.
|
| 296 |
"step": 370
|
| 297 |
},
|
| 298 |
{
|
| 299 |
-
"epoch": 0.
|
| 300 |
-
"grad_norm":
|
| 301 |
"learning_rate": 1.5200000000000002e-05,
|
| 302 |
-
"loss":
|
| 303 |
"step": 380
|
| 304 |
},
|
| 305 |
{
|
| 306 |
-
"epoch": 0.
|
| 307 |
-
"grad_norm":
|
| 308 |
"learning_rate": 1.5600000000000003e-05,
|
| 309 |
-
"loss":
|
| 310 |
"step": 390
|
| 311 |
},
|
| 312 |
{
|
| 313 |
-
"epoch": 0.
|
| 314 |
-
"grad_norm":
|
| 315 |
"learning_rate": 1.6000000000000003e-05,
|
| 316 |
-
"loss": 1.
|
| 317 |
"step": 400
|
| 318 |
},
|
| 319 |
{
|
| 320 |
-
"epoch": 0.
|
| 321 |
-
"eval_accuracy": 0.
|
| 322 |
-
"eval_accuracy_label_arts, culture, entertainment and media": 0.
|
| 323 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 324 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 325 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 326 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 327 |
-
"eval_accuracy_label_environment": 0.
|
| 328 |
"eval_accuracy_label_health": 0.9,
|
| 329 |
"eval_accuracy_label_human interest": 0.25,
|
| 330 |
"eval_accuracy_label_labour": 1.0,
|
| 331 |
-
"eval_accuracy_label_lifestyle and leisure": 0.
|
| 332 |
-
"eval_accuracy_label_politics": 0.
|
| 333 |
"eval_accuracy_label_religion": 0.0,
|
| 334 |
-
"eval_accuracy_label_science and technology": 0.
|
| 335 |
-
"eval_accuracy_label_society": 0.
|
| 336 |
-
"eval_accuracy_label_sport": 0.
|
| 337 |
-
"eval_accuracy_label_weather":
|
| 338 |
-
"eval_f1": 0.
|
| 339 |
-
"eval_loss": 1.
|
| 340 |
-
"eval_precision": 0.
|
| 341 |
-
"eval_recall": 0.
|
| 342 |
-
"eval_runtime": 6.
|
| 343 |
-
"eval_samples_per_second":
|
| 344 |
-
"eval_steps_per_second":
|
| 345 |
"step": 400
|
| 346 |
},
|
| 347 |
{
|
| 348 |
-
"epoch": 0.
|
| 349 |
-
"grad_norm":
|
| 350 |
"learning_rate": 1.64e-05,
|
| 351 |
-
"loss":
|
| 352 |
"step": 410
|
| 353 |
},
|
| 354 |
{
|
| 355 |
-
"epoch":
|
| 356 |
-
"grad_norm":
|
| 357 |
"learning_rate": 1.6800000000000002e-05,
|
| 358 |
-
"loss":
|
| 359 |
"step": 420
|
| 360 |
},
|
| 361 |
{
|
| 362 |
-
"epoch":
|
| 363 |
-
"grad_norm":
|
| 364 |
"learning_rate": 1.72e-05,
|
| 365 |
-
"loss":
|
| 366 |
"step": 430
|
| 367 |
},
|
| 368 |
{
|
| 369 |
-
"epoch":
|
| 370 |
-
"grad_norm":
|
| 371 |
"learning_rate": 1.76e-05,
|
| 372 |
-
"loss": 1.
|
| 373 |
"step": 440
|
| 374 |
},
|
| 375 |
{
|
| 376 |
-
"epoch":
|
| 377 |
-
"grad_norm":
|
| 378 |
"learning_rate": 1.8e-05,
|
| 379 |
-
"loss":
|
| 380 |
"step": 450
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch":
|
| 384 |
-
"grad_norm":
|
| 385 |
"learning_rate": 1.8400000000000003e-05,
|
| 386 |
-
"loss":
|
| 387 |
"step": 460
|
| 388 |
},
|
| 389 |
{
|
| 390 |
-
"epoch":
|
| 391 |
-
"grad_norm":
|
| 392 |
"learning_rate": 1.88e-05,
|
| 393 |
-
"loss": 0.
|
| 394 |
"step": 470
|
| 395 |
},
|
| 396 |
{
|
| 397 |
-
"epoch":
|
| 398 |
-
"grad_norm":
|
| 399 |
"learning_rate": 1.9200000000000003e-05,
|
| 400 |
-
"loss":
|
| 401 |
"step": 480
|
| 402 |
},
|
| 403 |
{
|
| 404 |
-
"epoch":
|
| 405 |
-
"grad_norm":
|
| 406 |
"learning_rate": 1.9600000000000002e-05,
|
| 407 |
-
"loss":
|
| 408 |
"step": 490
|
| 409 |
},
|
| 410 |
{
|
| 411 |
-
"epoch":
|
| 412 |
-
"grad_norm":
|
| 413 |
"learning_rate": 2e-05,
|
| 414 |
-
"loss":
|
| 415 |
"step": 500
|
| 416 |
},
|
| 417 |
{
|
| 418 |
-
"epoch":
|
| 419 |
-
"grad_norm":
|
| 420 |
-
"learning_rate": 1.
|
| 421 |
-
"loss": 0.
|
| 422 |
"step": 510
|
| 423 |
},
|
| 424 |
{
|
| 425 |
-
"epoch":
|
| 426 |
-
"grad_norm":
|
| 427 |
-
"learning_rate": 1.
|
| 428 |
-
"loss": 0.
|
| 429 |
"step": 520
|
| 430 |
},
|
| 431 |
{
|
| 432 |
-
"epoch":
|
| 433 |
-
"grad_norm":
|
| 434 |
-
"learning_rate": 1.
|
| 435 |
-
"loss": 0.
|
| 436 |
"step": 530
|
| 437 |
},
|
| 438 |
{
|
| 439 |
-
"epoch":
|
| 440 |
-
"grad_norm":
|
| 441 |
-
"learning_rate": 1.
|
| 442 |
-
"loss": 0.
|
| 443 |
"step": 540
|
| 444 |
},
|
| 445 |
{
|
| 446 |
-
"epoch":
|
| 447 |
-
"grad_norm":
|
| 448 |
-
"learning_rate": 1.
|
| 449 |
-
"loss": 0.
|
| 450 |
"step": 550
|
| 451 |
},
|
| 452 |
{
|
| 453 |
-
"epoch":
|
| 454 |
-
"grad_norm":
|
| 455 |
-
"learning_rate": 1.
|
| 456 |
-
"loss": 0.
|
| 457 |
"step": 560
|
| 458 |
},
|
| 459 |
{
|
| 460 |
-
"epoch":
|
| 461 |
-
"grad_norm":
|
| 462 |
-
"learning_rate": 1.
|
| 463 |
-
"loss": 0.
|
| 464 |
"step": 570
|
| 465 |
},
|
| 466 |
{
|
| 467 |
-
"epoch":
|
| 468 |
-
"grad_norm":
|
| 469 |
-
"learning_rate": 1.
|
| 470 |
-
"loss":
|
| 471 |
"step": 580
|
| 472 |
},
|
| 473 |
{
|
| 474 |
-
"epoch":
|
| 475 |
-
"grad_norm":
|
| 476 |
-
"learning_rate": 1.
|
| 477 |
-
"loss": 0.
|
| 478 |
"step": 590
|
| 479 |
},
|
| 480 |
{
|
| 481 |
-
"epoch": 1.
|
| 482 |
-
"grad_norm":
|
| 483 |
-
"learning_rate": 1.
|
| 484 |
-
"loss": 0.
|
| 485 |
"step": 600
|
| 486 |
},
|
| 487 |
{
|
| 488 |
-
"epoch": 1.
|
| 489 |
-
"eval_accuracy": 0.
|
| 490 |
-
"eval_accuracy_label_arts, culture, entertainment and media": 0.
|
| 491 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 492 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 493 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 494 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 495 |
-
"eval_accuracy_label_environment": 0.
|
| 496 |
"eval_accuracy_label_health": 0.8,
|
| 497 |
-
"eval_accuracy_label_human interest": 0.
|
| 498 |
-
"eval_accuracy_label_labour": 0
|
| 499 |
-
"eval_accuracy_label_lifestyle and leisure": 0.
|
| 500 |
-
"eval_accuracy_label_politics": 0.
|
| 501 |
"eval_accuracy_label_religion": 0.0,
|
| 502 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 503 |
-
"eval_accuracy_label_society": 0.
|
| 504 |
-
"eval_accuracy_label_sport": 0.
|
| 505 |
"eval_accuracy_label_weather": 1.0,
|
| 506 |
-
"eval_f1": 0.
|
| 507 |
-
"eval_loss": 0.
|
| 508 |
-
"eval_precision": 0.
|
| 509 |
-
"eval_recall": 0.
|
| 510 |
-
"eval_runtime": 6.
|
| 511 |
-
"eval_samples_per_second":
|
| 512 |
-
"eval_steps_per_second":
|
| 513 |
"step": 600
|
| 514 |
},
|
| 515 |
{
|
| 516 |
-
"epoch": 1.
|
| 517 |
-
"grad_norm":
|
| 518 |
-
"learning_rate": 1.
|
| 519 |
-
"loss": 0.
|
| 520 |
"step": 610
|
| 521 |
},
|
| 522 |
{
|
| 523 |
-
"epoch": 1.
|
| 524 |
-
"grad_norm":
|
| 525 |
-
"learning_rate": 1.
|
| 526 |
-
"loss": 0.
|
| 527 |
"step": 620
|
| 528 |
},
|
| 529 |
{
|
| 530 |
-
"epoch": 1.
|
| 531 |
-
"grad_norm":
|
| 532 |
-
"learning_rate": 1.
|
| 533 |
-
"loss": 0.
|
| 534 |
"step": 630
|
| 535 |
},
|
| 536 |
{
|
| 537 |
-
"epoch": 1.
|
| 538 |
-
"grad_norm":
|
| 539 |
-
"learning_rate": 1.
|
| 540 |
-
"loss": 0.
|
| 541 |
"step": 640
|
| 542 |
},
|
| 543 |
{
|
| 544 |
-
"epoch": 1.
|
| 545 |
-
"grad_norm":
|
| 546 |
-
"learning_rate": 1.
|
| 547 |
-
"loss": 0.
|
| 548 |
"step": 650
|
| 549 |
},
|
| 550 |
{
|
| 551 |
-
"epoch": 1.
|
| 552 |
-
"grad_norm":
|
| 553 |
-
"learning_rate": 1.
|
| 554 |
-
"loss": 0.
|
| 555 |
"step": 660
|
| 556 |
},
|
| 557 |
{
|
| 558 |
-
"epoch": 1.
|
| 559 |
-
"grad_norm":
|
| 560 |
-
"learning_rate": 1.
|
| 561 |
-
"loss": 0.
|
| 562 |
"step": 670
|
| 563 |
},
|
| 564 |
{
|
| 565 |
-
"epoch": 1.
|
| 566 |
-
"grad_norm":
|
| 567 |
-
"learning_rate": 1.
|
| 568 |
-
"loss": 0.
|
| 569 |
"step": 680
|
| 570 |
},
|
| 571 |
{
|
| 572 |
-
"epoch": 1.
|
| 573 |
-
"grad_norm":
|
| 574 |
-
"learning_rate": 1.
|
| 575 |
-
"loss": 0.
|
| 576 |
"step": 690
|
| 577 |
},
|
| 578 |
{
|
| 579 |
-
"epoch": 1.
|
| 580 |
-
"grad_norm":
|
| 581 |
-
"learning_rate": 1.
|
| 582 |
-
"loss": 0.
|
| 583 |
"step": 700
|
| 584 |
},
|
| 585 |
{
|
| 586 |
-
"epoch": 1.
|
| 587 |
-
"grad_norm":
|
| 588 |
-
"learning_rate": 1.
|
| 589 |
-
"loss": 0.
|
| 590 |
"step": 710
|
| 591 |
},
|
| 592 |
{
|
| 593 |
-
"epoch": 1.
|
| 594 |
-
"grad_norm":
|
| 595 |
-
"learning_rate": 1.
|
| 596 |
-
"loss": 0.
|
| 597 |
"step": 720
|
| 598 |
},
|
| 599 |
{
|
| 600 |
-
"epoch": 1.
|
| 601 |
-
"grad_norm":
|
| 602 |
-
"learning_rate": 1.
|
| 603 |
-
"loss": 0.
|
| 604 |
"step": 730
|
| 605 |
},
|
| 606 |
{
|
| 607 |
-
"epoch": 1.
|
| 608 |
-
"grad_norm":
|
| 609 |
-
"learning_rate": 1.
|
| 610 |
-
"loss": 0.
|
| 611 |
"step": 740
|
| 612 |
},
|
| 613 |
{
|
| 614 |
-
"epoch": 1.
|
| 615 |
-
"grad_norm":
|
| 616 |
-
"learning_rate": 1.
|
| 617 |
-
"loss": 0.
|
| 618 |
"step": 750
|
| 619 |
},
|
| 620 |
{
|
| 621 |
-
"epoch": 1.
|
| 622 |
-
"grad_norm":
|
| 623 |
-
"learning_rate": 1.
|
| 624 |
-
"loss": 0.
|
| 625 |
"step": 760
|
| 626 |
},
|
| 627 |
{
|
| 628 |
-
"epoch": 1.
|
| 629 |
-
"grad_norm":
|
| 630 |
-
"learning_rate": 1.
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 770
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch": 1.
|
| 636 |
-
"grad_norm":
|
| 637 |
-
"learning_rate": 1.
|
| 638 |
-
"loss": 0.
|
| 639 |
"step": 780
|
| 640 |
},
|
| 641 |
{
|
| 642 |
-
"epoch": 1.
|
| 643 |
-
"grad_norm":
|
| 644 |
-
"learning_rate": 1.
|
| 645 |
-
"loss": 0.
|
| 646 |
"step": 790
|
| 647 |
},
|
| 648 |
{
|
| 649 |
-
"epoch": 1.
|
| 650 |
-
"grad_norm":
|
| 651 |
-
"learning_rate": 1.
|
| 652 |
-
"loss": 0.
|
| 653 |
"step": 800
|
| 654 |
},
|
| 655 |
{
|
| 656 |
-
"epoch": 1.
|
| 657 |
-
"eval_accuracy": 0.
|
| 658 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 659 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 660 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 661 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 662 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 663 |
-
"eval_accuracy_label_environment": 0.
|
| 664 |
-
"eval_accuracy_label_health": 0.
|
| 665 |
-
"eval_accuracy_label_human interest": 0.
|
| 666 |
-
"eval_accuracy_label_labour": 0
|
| 667 |
-
"eval_accuracy_label_lifestyle and leisure": 0.
|
| 668 |
-
"eval_accuracy_label_politics": 0.
|
| 669 |
"eval_accuracy_label_religion": 0.0,
|
| 670 |
-
"eval_accuracy_label_science and technology": 0.
|
| 671 |
-
"eval_accuracy_label_society": 0.
|
| 672 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 673 |
"eval_accuracy_label_weather": 1.0,
|
| 674 |
-
"eval_f1": 0.
|
| 675 |
-
"eval_loss": 0.
|
| 676 |
-
"eval_precision": 0.
|
| 677 |
-
"eval_recall": 0.
|
| 678 |
-
"eval_runtime": 6.
|
| 679 |
-
"eval_samples_per_second":
|
| 680 |
-
"eval_steps_per_second":
|
| 681 |
"step": 800
|
| 682 |
},
|
| 683 |
{
|
| 684 |
-
"epoch": 1.
|
| 685 |
-
"grad_norm":
|
| 686 |
-
"learning_rate": 1.
|
| 687 |
-
"loss": 0.
|
| 688 |
"step": 810
|
| 689 |
},
|
| 690 |
{
|
| 691 |
-
"epoch": 1.
|
| 692 |
-
"grad_norm":
|
| 693 |
-
"learning_rate": 1.
|
| 694 |
-
"loss": 0.
|
| 695 |
"step": 820
|
| 696 |
},
|
| 697 |
{
|
| 698 |
-
"epoch": 1.
|
| 699 |
-
"grad_norm":
|
| 700 |
-
"learning_rate": 1.
|
| 701 |
-
"loss": 0.
|
| 702 |
"step": 830
|
| 703 |
},
|
| 704 |
{
|
| 705 |
-
"epoch":
|
| 706 |
-
"grad_norm":
|
| 707 |
-
"learning_rate": 1.
|
| 708 |
-
"loss": 0.
|
| 709 |
"step": 840
|
| 710 |
},
|
| 711 |
{
|
| 712 |
-
"epoch":
|
| 713 |
-
"grad_norm":
|
| 714 |
-
"learning_rate": 1.
|
| 715 |
-
"loss": 0.
|
| 716 |
"step": 850
|
| 717 |
},
|
| 718 |
{
|
| 719 |
-
"epoch":
|
| 720 |
-
"grad_norm":
|
| 721 |
-
"learning_rate": 1.
|
| 722 |
-
"loss": 0.
|
| 723 |
"step": 860
|
| 724 |
},
|
| 725 |
{
|
| 726 |
-
"epoch":
|
| 727 |
-
"grad_norm":
|
| 728 |
-
"learning_rate": 1.
|
| 729 |
-
"loss": 0.
|
| 730 |
"step": 870
|
| 731 |
},
|
| 732 |
{
|
| 733 |
-
"epoch":
|
| 734 |
-
"grad_norm":
|
| 735 |
-
"learning_rate":
|
| 736 |
-
"loss": 0.
|
| 737 |
"step": 880
|
| 738 |
},
|
| 739 |
{
|
| 740 |
-
"epoch":
|
| 741 |
-
"grad_norm":
|
| 742 |
-
"learning_rate":
|
| 743 |
-
"loss": 0.
|
| 744 |
"step": 890
|
| 745 |
},
|
| 746 |
{
|
| 747 |
-
"epoch":
|
| 748 |
-
"grad_norm":
|
| 749 |
-
"learning_rate":
|
| 750 |
-
"loss": 0.
|
| 751 |
"step": 900
|
| 752 |
},
|
| 753 |
{
|
| 754 |
-
"epoch":
|
| 755 |
-
"grad_norm":
|
| 756 |
-
"learning_rate":
|
| 757 |
-
"loss": 0.
|
| 758 |
"step": 910
|
| 759 |
},
|
| 760 |
{
|
| 761 |
-
"epoch":
|
| 762 |
-
"grad_norm":
|
| 763 |
-
"learning_rate":
|
| 764 |
-
"loss": 0.
|
| 765 |
"step": 920
|
| 766 |
},
|
| 767 |
{
|
| 768 |
-
"epoch":
|
| 769 |
-
"grad_norm":
|
| 770 |
-
"learning_rate":
|
| 771 |
-
"loss": 0.
|
| 772 |
"step": 930
|
| 773 |
},
|
| 774 |
{
|
| 775 |
-
"epoch":
|
| 776 |
-
"grad_norm":
|
| 777 |
-
"learning_rate":
|
| 778 |
-
"loss": 0.
|
| 779 |
"step": 940
|
| 780 |
},
|
| 781 |
{
|
| 782 |
-
"epoch":
|
| 783 |
-
"grad_norm":
|
| 784 |
-
"learning_rate":
|
| 785 |
-
"loss": 0.
|
| 786 |
"step": 950
|
| 787 |
},
|
| 788 |
{
|
| 789 |
-
"epoch":
|
| 790 |
-
"grad_norm":
|
| 791 |
-
"learning_rate":
|
| 792 |
-
"loss": 0.
|
| 793 |
"step": 960
|
| 794 |
},
|
| 795 |
{
|
| 796 |
-
"epoch":
|
| 797 |
-
"grad_norm":
|
| 798 |
-
"learning_rate":
|
| 799 |
-
"loss": 0.
|
| 800 |
"step": 970
|
| 801 |
},
|
| 802 |
{
|
| 803 |
-
"epoch":
|
| 804 |
-
"grad_norm":
|
| 805 |
-
"learning_rate":
|
| 806 |
-
"loss": 0.
|
| 807 |
"step": 980
|
| 808 |
},
|
| 809 |
{
|
| 810 |
-
"epoch":
|
| 811 |
-
"grad_norm":
|
| 812 |
-
"learning_rate":
|
| 813 |
-
"loss": 0.
|
| 814 |
"step": 990
|
| 815 |
},
|
| 816 |
{
|
| 817 |
-
"epoch":
|
| 818 |
-
"grad_norm":
|
| 819 |
-
"learning_rate":
|
| 820 |
-
"loss": 0.
|
| 821 |
"step": 1000
|
| 822 |
},
|
| 823 |
{
|
| 824 |
-
"epoch":
|
| 825 |
-
"eval_accuracy": 0.
|
| 826 |
-
"eval_accuracy_label_arts, culture, entertainment and media": 0.
|
| 827 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 828 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 829 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 830 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 831 |
-
"eval_accuracy_label_environment": 0.
|
| 832 |
-
"eval_accuracy_label_health": 0.
|
| 833 |
-
"eval_accuracy_label_human interest": 0.
|
| 834 |
-
"eval_accuracy_label_labour": 0
|
| 835 |
"eval_accuracy_label_lifestyle and leisure": 0.625,
|
| 836 |
"eval_accuracy_label_politics": 0.6474820143884892,
|
| 837 |
"eval_accuracy_label_religion": 0.0,
|
| 838 |
-
"eval_accuracy_label_science and technology": 0.
|
| 839 |
-
"eval_accuracy_label_society": 0.
|
| 840 |
-
"eval_accuracy_label_sport":
|
| 841 |
-
"eval_accuracy_label_weather":
|
| 842 |
-
"eval_f1": 0.
|
| 843 |
-
"eval_loss": 0.
|
| 844 |
-
"eval_precision": 0.
|
| 845 |
-
"eval_recall": 0.
|
| 846 |
-
"eval_runtime": 6.
|
| 847 |
-
"eval_samples_per_second":
|
| 848 |
-
"eval_steps_per_second":
|
| 849 |
"step": 1000
|
| 850 |
},
|
| 851 |
{
|
| 852 |
-
"epoch":
|
| 853 |
-
"grad_norm":
|
| 854 |
-
"learning_rate":
|
| 855 |
-
"loss": 0.
|
| 856 |
"step": 1010
|
| 857 |
},
|
| 858 |
{
|
| 859 |
-
"epoch":
|
| 860 |
-
"grad_norm":
|
| 861 |
-
"learning_rate":
|
| 862 |
-
"loss": 0.
|
| 863 |
"step": 1020
|
| 864 |
},
|
| 865 |
{
|
| 866 |
-
"epoch":
|
| 867 |
-
"grad_norm":
|
| 868 |
-
"learning_rate":
|
| 869 |
-
"loss": 0.
|
| 870 |
"step": 1030
|
| 871 |
},
|
| 872 |
{
|
| 873 |
-
"epoch":
|
| 874 |
-
"grad_norm":
|
| 875 |
-
"learning_rate":
|
| 876 |
-
"loss": 0.
|
| 877 |
"step": 1040
|
| 878 |
},
|
| 879 |
{
|
| 880 |
-
"epoch":
|
| 881 |
-
"grad_norm":
|
| 882 |
-
"learning_rate":
|
| 883 |
-
"loss": 0.
|
| 884 |
"step": 1050
|
| 885 |
},
|
| 886 |
{
|
| 887 |
-
"epoch":
|
| 888 |
-
"grad_norm":
|
| 889 |
-
"learning_rate":
|
| 890 |
-
"loss": 0.
|
| 891 |
"step": 1060
|
| 892 |
},
|
| 893 |
{
|
| 894 |
-
"epoch":
|
| 895 |
-
"grad_norm":
|
| 896 |
-
"learning_rate":
|
| 897 |
-
"loss": 0.
|
| 898 |
"step": 1070
|
| 899 |
},
|
| 900 |
{
|
| 901 |
-
"epoch":
|
| 902 |
-
"grad_norm":
|
| 903 |
-
"learning_rate":
|
| 904 |
-
"loss": 0.
|
| 905 |
"step": 1080
|
| 906 |
},
|
| 907 |
{
|
| 908 |
-
"epoch":
|
| 909 |
-
"grad_norm":
|
| 910 |
-
"learning_rate":
|
| 911 |
-
"loss": 0.
|
| 912 |
"step": 1090
|
| 913 |
},
|
| 914 |
{
|
| 915 |
-
"epoch":
|
| 916 |
-
"grad_norm":
|
| 917 |
-
"learning_rate":
|
| 918 |
-
"loss": 0.
|
| 919 |
"step": 1100
|
| 920 |
},
|
| 921 |
{
|
| 922 |
-
"epoch":
|
| 923 |
-
"grad_norm":
|
| 924 |
-
"learning_rate":
|
| 925 |
-
"loss": 0.
|
| 926 |
"step": 1110
|
| 927 |
},
|
| 928 |
{
|
| 929 |
-
"epoch":
|
| 930 |
-
"grad_norm":
|
| 931 |
-
"learning_rate":
|
| 932 |
-
"loss": 0.
|
| 933 |
"step": 1120
|
| 934 |
},
|
| 935 |
{
|
| 936 |
-
"epoch":
|
| 937 |
-
"grad_norm":
|
| 938 |
-
"learning_rate":
|
| 939 |
-
"loss": 0.
|
| 940 |
"step": 1130
|
| 941 |
},
|
| 942 |
{
|
| 943 |
-
"epoch":
|
| 944 |
-
"grad_norm":
|
| 945 |
-
"learning_rate":
|
| 946 |
-
"loss": 0.
|
| 947 |
"step": 1140
|
| 948 |
},
|
| 949 |
{
|
| 950 |
-
"epoch":
|
| 951 |
-
"grad_norm":
|
| 952 |
-
"learning_rate":
|
| 953 |
-
"loss": 0.
|
| 954 |
"step": 1150
|
| 955 |
},
|
| 956 |
{
|
| 957 |
-
"epoch":
|
| 958 |
-
"grad_norm":
|
| 959 |
-
"learning_rate":
|
| 960 |
-
"loss": 0.
|
| 961 |
"step": 1160
|
| 962 |
},
|
| 963 |
{
|
| 964 |
-
"epoch":
|
| 965 |
-
"grad_norm":
|
| 966 |
-
"learning_rate":
|
| 967 |
-
"loss": 0.
|
| 968 |
"step": 1170
|
| 969 |
},
|
| 970 |
{
|
| 971 |
-
"epoch":
|
| 972 |
-
"grad_norm":
|
| 973 |
-
"learning_rate":
|
| 974 |
-
"loss": 0.
|
| 975 |
"step": 1180
|
| 976 |
},
|
| 977 |
{
|
| 978 |
-
"epoch": 2.
|
| 979 |
-
"grad_norm":
|
| 980 |
-
"learning_rate":
|
| 981 |
-
"loss": 0.
|
| 982 |
"step": 1190
|
| 983 |
},
|
| 984 |
{
|
| 985 |
-
"epoch": 2.
|
| 986 |
-
"grad_norm":
|
| 987 |
-
"learning_rate":
|
| 988 |
-
"loss": 0.
|
| 989 |
"step": 1200
|
| 990 |
},
|
| 991 |
{
|
| 992 |
-
"epoch": 2.
|
| 993 |
-
"eval_accuracy": 0.
|
| 994 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 995 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 996 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 997 |
-
"eval_accuracy_label_disaster, accident, and emergency incident": 0.
|
| 998 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 999 |
-
"eval_accuracy_label_environment": 0.
|
| 1000 |
-
"eval_accuracy_label_health": 0.
|
| 1001 |
"eval_accuracy_label_human interest": 0.25,
|
| 1002 |
"eval_accuracy_label_labour": 0.5,
|
| 1003 |
-
"eval_accuracy_label_lifestyle and leisure": 0.
|
| 1004 |
-
"eval_accuracy_label_politics": 0.
|
| 1005 |
"eval_accuracy_label_religion": 0.0,
|
| 1006 |
-
"eval_accuracy_label_science and technology": 0.
|
| 1007 |
-
"eval_accuracy_label_society": 0.
|
| 1008 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1009 |
"eval_accuracy_label_weather": 1.0,
|
| 1010 |
-
"eval_f1": 0.
|
| 1011 |
-
"eval_loss": 0.
|
| 1012 |
-
"eval_precision": 0.
|
| 1013 |
-
"eval_recall": 0.
|
| 1014 |
-
"eval_runtime": 6.
|
| 1015 |
-
"eval_samples_per_second":
|
| 1016 |
-
"eval_steps_per_second":
|
| 1017 |
"step": 1200
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
-
"epoch": 2.
|
| 1021 |
-
"grad_norm":
|
| 1022 |
-
"learning_rate":
|
| 1023 |
-
"loss": 0.
|
| 1024 |
"step": 1210
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
-
"epoch": 2.
|
| 1028 |
-
"grad_norm":
|
| 1029 |
-
"learning_rate": 8.
|
| 1030 |
-
"loss": 0.
|
| 1031 |
"step": 1220
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
-
"epoch": 2.
|
| 1035 |
-
"grad_norm":
|
| 1036 |
-
"learning_rate":
|
| 1037 |
-
"loss": 0.
|
| 1038 |
"step": 1230
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
-
"epoch": 2.
|
| 1042 |
-
"grad_norm":
|
| 1043 |
-
"learning_rate":
|
| 1044 |
-
"loss": 0.
|
| 1045 |
"step": 1240
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
-
"epoch": 2.
|
| 1049 |
-
"grad_norm":
|
| 1050 |
-
"learning_rate":
|
| 1051 |
-
"loss": 0.
|
| 1052 |
"step": 1250
|
| 1053 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1054 |
{
|
| 1055 |
"epoch": 3.0,
|
| 1056 |
-
"step":
|
| 1057 |
-
"total_flos":
|
| 1058 |
-
"train_loss":
|
| 1059 |
-
"train_runtime":
|
| 1060 |
-
"train_samples_per_second": 43.
|
| 1061 |
-
"train_steps_per_second": 1.
|
| 1062 |
},
|
| 1063 |
{
|
| 1064 |
"epoch": 3.0,
|
| 1065 |
-
"eval_accuracy": 0.
|
| 1066 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 1067 |
-
"eval_accuracy_label_conflict, war and peace": 0.
|
| 1068 |
-
"eval_accuracy_label_crime, law and justice": 0.
|
| 1069 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 1070 |
-
"eval_accuracy_label_economy, business, and finance": 0.
|
| 1071 |
-
"eval_accuracy_label_environment": 0.
|
| 1072 |
-
"eval_accuracy_label_health": 0.
|
| 1073 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 1074 |
"eval_accuracy_label_labour": 0.5,
|
| 1075 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 1076 |
-
"eval_accuracy_label_politics": 0.
|
| 1077 |
"eval_accuracy_label_religion": 0.0,
|
| 1078 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 1079 |
-
"eval_accuracy_label_society": 0.
|
| 1080 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1081 |
"eval_accuracy_label_weather": 1.0,
|
| 1082 |
-
"eval_f1": 0.
|
| 1083 |
-
"eval_loss": 0.
|
| 1084 |
-
"eval_precision": 0.
|
| 1085 |
-
"eval_recall": 0.
|
| 1086 |
-
"eval_runtime": 6.
|
| 1087 |
-
"eval_samples_per_second":
|
| 1088 |
-
"eval_steps_per_second": 8.
|
| 1089 |
-
"step":
|
| 1090 |
}
|
| 1091 |
],
|
| 1092 |
"logging_steps": 10,
|
| 1093 |
-
"max_steps":
|
| 1094 |
"num_input_tokens_seen": 0,
|
| 1095 |
"num_train_epochs": 3,
|
| 1096 |
"save_steps": 1000,
|
| 1097 |
-
"total_flos":
|
| 1098 |
"train_batch_size": 16,
|
| 1099 |
"trial_name": null,
|
| 1100 |
"trial_params": null
|
|
|
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 3.0,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 1779,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.016863406408094434,
|
| 13 |
+
"grad_norm": 5.893406867980957,
|
| 14 |
"learning_rate": 4.0000000000000003e-07,
|
| 15 |
+
"loss": 2.8131,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.03372681281618887,
|
| 20 |
+
"grad_norm": 4.519655704498291,
|
| 21 |
"learning_rate": 8.000000000000001e-07,
|
| 22 |
+
"loss": 2.8225,
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.050590219224283306,
|
| 27 |
+
"grad_norm": 4.535500526428223,
|
| 28 |
"learning_rate": 1.2000000000000002e-06,
|
| 29 |
+
"loss": 2.8042,
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.06745362563237774,
|
| 34 |
+
"grad_norm": 4.360924243927002,
|
| 35 |
"learning_rate": 1.6000000000000001e-06,
|
| 36 |
+
"loss": 2.7337,
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.08431703204047218,
|
| 41 |
+
"grad_norm": 5.550849914550781,
|
| 42 |
"learning_rate": 2.0000000000000003e-06,
|
| 43 |
+
"loss": 2.6976,
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.10118043844856661,
|
| 48 |
+
"grad_norm": 4.623074054718018,
|
| 49 |
"learning_rate": 2.4000000000000003e-06,
|
| 50 |
+
"loss": 2.6382,
|
| 51 |
"step": 60
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.11804384485666104,
|
| 55 |
+
"grad_norm": 6.715169906616211,
|
| 56 |
"learning_rate": 2.8000000000000003e-06,
|
| 57 |
+
"loss": 2.6356,
|
| 58 |
"step": 70
|
| 59 |
},
|
| 60 |
{
|
| 61 |
+
"epoch": 0.13490725126475547,
|
| 62 |
+
"grad_norm": 5.305912017822266,
|
| 63 |
"learning_rate": 3.2000000000000003e-06,
|
| 64 |
+
"loss": 2.558,
|
| 65 |
"step": 80
|
| 66 |
},
|
| 67 |
{
|
| 68 |
+
"epoch": 0.15177065767284992,
|
| 69 |
+
"grad_norm": 5.895500183105469,
|
| 70 |
"learning_rate": 3.6000000000000003e-06,
|
| 71 |
+
"loss": 2.4923,
|
| 72 |
"step": 90
|
| 73 |
},
|
| 74 |
{
|
| 75 |
+
"epoch": 0.16863406408094436,
|
| 76 |
+
"grad_norm": 5.270618438720703,
|
| 77 |
"learning_rate": 4.000000000000001e-06,
|
| 78 |
+
"loss": 2.3955,
|
| 79 |
"step": 100
|
| 80 |
},
|
| 81 |
{
|
| 82 |
+
"epoch": 0.18549747048903878,
|
| 83 |
+
"grad_norm": 5.943336486816406,
|
| 84 |
"learning_rate": 4.4e-06,
|
| 85 |
+
"loss": 2.3467,
|
| 86 |
"step": 110
|
| 87 |
},
|
| 88 |
{
|
| 89 |
+
"epoch": 0.20236087689713322,
|
| 90 |
+
"grad_norm": 6.1944403648376465,
|
| 91 |
"learning_rate": 4.800000000000001e-06,
|
| 92 |
+
"loss": 2.3161,
|
| 93 |
"step": 120
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.21922428330522767,
|
| 97 |
+
"grad_norm": 6.080775260925293,
|
| 98 |
"learning_rate": 5.2e-06,
|
| 99 |
+
"loss": 2.2254,
|
| 100 |
"step": 130
|
| 101 |
},
|
| 102 |
{
|
| 103 |
+
"epoch": 0.23608768971332209,
|
| 104 |
+
"grad_norm": 6.028081893920898,
|
| 105 |
"learning_rate": 5.600000000000001e-06,
|
| 106 |
+
"loss": 2.205,
|
| 107 |
"step": 140
|
| 108 |
},
|
| 109 |
{
|
| 110 |
+
"epoch": 0.25295109612141653,
|
| 111 |
+
"grad_norm": 5.6627631187438965,
|
| 112 |
"learning_rate": 6e-06,
|
| 113 |
+
"loss": 2.1021,
|
| 114 |
"step": 150
|
| 115 |
},
|
| 116 |
{
|
| 117 |
+
"epoch": 0.26981450252951095,
|
| 118 |
+
"grad_norm": 7.71439266204834,
|
| 119 |
"learning_rate": 6.4000000000000006e-06,
|
| 120 |
+
"loss": 2.1383,
|
| 121 |
"step": 160
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"epoch": 0.2866779089376054,
|
| 125 |
+
"grad_norm": 8.687724113464355,
|
| 126 |
"learning_rate": 6.800000000000001e-06,
|
| 127 |
+
"loss": 1.9821,
|
| 128 |
"step": 170
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 0.30354131534569984,
|
| 132 |
+
"grad_norm": 6.372674942016602,
|
| 133 |
"learning_rate": 7.2000000000000005e-06,
|
| 134 |
+
"loss": 1.9015,
|
| 135 |
"step": 180
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.32040472175379425,
|
| 139 |
+
"grad_norm": 7.2390217781066895,
|
| 140 |
"learning_rate": 7.600000000000001e-06,
|
| 141 |
+
"loss": 1.9485,
|
| 142 |
"step": 190
|
| 143 |
},
|
| 144 |
{
|
| 145 |
+
"epoch": 0.3372681281618887,
|
| 146 |
+
"grad_norm": 7.15785026550293,
|
| 147 |
"learning_rate": 8.000000000000001e-06,
|
| 148 |
+
"loss": 1.7671,
|
| 149 |
"step": 200
|
| 150 |
},
|
| 151 |
{
|
| 152 |
+
"epoch": 0.3372681281618887,
|
| 153 |
+
"eval_accuracy": 0.5554272517321016,
|
| 154 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.5833333333333334,
|
| 155 |
+
"eval_accuracy_label_conflict, war and peace": 0.7553191489361702,
|
| 156 |
+
"eval_accuracy_label_crime, law and justice": 0.8959537572254336,
|
| 157 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.32061068702290074,
|
| 158 |
+
"eval_accuracy_label_economy, business, and finance": 0.6708860759493671,
|
| 159 |
+
"eval_accuracy_label_environment": 0.125,
|
| 160 |
+
"eval_accuracy_label_health": 0.7,
|
| 161 |
"eval_accuracy_label_human interest": 0.0,
|
| 162 |
+
"eval_accuracy_label_labour": 0.5,
|
| 163 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 164 |
+
"eval_accuracy_label_politics": 0.28776978417266186,
|
| 165 |
"eval_accuracy_label_religion": 0.0,
|
| 166 |
"eval_accuracy_label_science and technology": 0.0,
|
| 167 |
+
"eval_accuracy_label_society": 0.03508771929824561,
|
| 168 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 169 |
+
"eval_accuracy_label_weather": 1.0,
|
| 170 |
+
"eval_f1": 0.5206037462663565,
|
| 171 |
+
"eval_loss": 1.5660511255264282,
|
| 172 |
+
"eval_precision": 0.5827940940390139,
|
| 173 |
+
"eval_recall": 0.5554272517321016,
|
| 174 |
+
"eval_runtime": 6.8757,
|
| 175 |
+
"eval_samples_per_second": 125.95,
|
| 176 |
+
"eval_steps_per_second": 7.999,
|
| 177 |
"step": 200
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.35413153456998314,
|
| 181 |
+
"grad_norm": 7.729678630828857,
|
| 182 |
"learning_rate": 8.400000000000001e-06,
|
| 183 |
+
"loss": 1.7632,
|
| 184 |
"step": 210
|
| 185 |
},
|
| 186 |
{
|
| 187 |
+
"epoch": 0.37099494097807756,
|
| 188 |
+
"grad_norm": 8.155638694763184,
|
| 189 |
"learning_rate": 8.8e-06,
|
| 190 |
+
"loss": 1.6068,
|
| 191 |
"step": 220
|
| 192 |
},
|
| 193 |
{
|
| 194 |
+
"epoch": 0.38785834738617203,
|
| 195 |
+
"grad_norm": 7.142002105712891,
|
| 196 |
"learning_rate": 9.200000000000002e-06,
|
| 197 |
+
"loss": 1.4778,
|
| 198 |
"step": 230
|
| 199 |
},
|
| 200 |
{
|
| 201 |
+
"epoch": 0.40472175379426645,
|
| 202 |
+
"grad_norm": 7.332587242126465,
|
| 203 |
"learning_rate": 9.600000000000001e-06,
|
| 204 |
+
"loss": 1.5736,
|
| 205 |
"step": 240
|
| 206 |
},
|
| 207 |
{
|
| 208 |
+
"epoch": 0.42158516020236086,
|
| 209 |
+
"grad_norm": 13.657541275024414,
|
| 210 |
"learning_rate": 1e-05,
|
| 211 |
+
"loss": 1.4757,
|
| 212 |
"step": 250
|
| 213 |
},
|
| 214 |
{
|
| 215 |
+
"epoch": 0.43844856661045534,
|
| 216 |
+
"grad_norm": 14.87531566619873,
|
| 217 |
"learning_rate": 1.04e-05,
|
| 218 |
+
"loss": 1.4364,
|
| 219 |
"step": 260
|
| 220 |
},
|
| 221 |
{
|
| 222 |
+
"epoch": 0.45531197301854975,
|
| 223 |
+
"grad_norm": 15.269392967224121,
|
| 224 |
"learning_rate": 1.0800000000000002e-05,
|
| 225 |
+
"loss": 1.3215,
|
| 226 |
"step": 270
|
| 227 |
},
|
| 228 |
{
|
| 229 |
+
"epoch": 0.47217537942664417,
|
| 230 |
+
"grad_norm": 10.192849159240723,
|
| 231 |
"learning_rate": 1.1200000000000001e-05,
|
| 232 |
+
"loss": 1.3129,
|
| 233 |
"step": 280
|
| 234 |
},
|
| 235 |
{
|
| 236 |
+
"epoch": 0.48903878583473864,
|
| 237 |
+
"grad_norm": 8.036419868469238,
|
| 238 |
"learning_rate": 1.16e-05,
|
| 239 |
+
"loss": 1.2399,
|
| 240 |
"step": 290
|
| 241 |
},
|
| 242 |
{
|
| 243 |
+
"epoch": 0.5059021922428331,
|
| 244 |
+
"grad_norm": 10.625975608825684,
|
| 245 |
"learning_rate": 1.2e-05,
|
| 246 |
+
"loss": 1.3016,
|
| 247 |
"step": 300
|
| 248 |
},
|
| 249 |
{
|
| 250 |
+
"epoch": 0.5227655986509275,
|
| 251 |
+
"grad_norm": 14.144831657409668,
|
| 252 |
"learning_rate": 1.2400000000000002e-05,
|
| 253 |
+
"loss": 1.2514,
|
| 254 |
"step": 310
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 0.5396290050590219,
|
| 258 |
+
"grad_norm": 11.439717292785645,
|
| 259 |
"learning_rate": 1.2800000000000001e-05,
|
| 260 |
+
"loss": 1.1768,
|
| 261 |
"step": 320
|
| 262 |
},
|
| 263 |
{
|
| 264 |
+
"epoch": 0.5564924114671164,
|
| 265 |
+
"grad_norm": 18.485063552856445,
|
| 266 |
"learning_rate": 1.3200000000000002e-05,
|
| 267 |
+
"loss": 1.1273,
|
| 268 |
"step": 330
|
| 269 |
},
|
| 270 |
{
|
| 271 |
+
"epoch": 0.5733558178752108,
|
| 272 |
+
"grad_norm": 12.227529525756836,
|
| 273 |
"learning_rate": 1.3600000000000002e-05,
|
| 274 |
+
"loss": 1.0002,
|
| 275 |
"step": 340
|
| 276 |
},
|
| 277 |
{
|
| 278 |
+
"epoch": 0.5902192242833052,
|
| 279 |
+
"grad_norm": 10.209256172180176,
|
| 280 |
"learning_rate": 1.4e-05,
|
| 281 |
+
"loss": 1.1188,
|
| 282 |
"step": 350
|
| 283 |
},
|
| 284 |
{
|
| 285 |
+
"epoch": 0.6070826306913997,
|
| 286 |
+
"grad_norm": 8.667159080505371,
|
| 287 |
"learning_rate": 1.4400000000000001e-05,
|
| 288 |
+
"loss": 1.0151,
|
| 289 |
"step": 360
|
| 290 |
},
|
| 291 |
{
|
| 292 |
+
"epoch": 0.6239460370994941,
|
| 293 |
+
"grad_norm": 8.289810180664062,
|
| 294 |
"learning_rate": 1.48e-05,
|
| 295 |
+
"loss": 1.0148,
|
| 296 |
"step": 370
|
| 297 |
},
|
| 298 |
{
|
| 299 |
+
"epoch": 0.6408094435075885,
|
| 300 |
+
"grad_norm": 8.391195297241211,
|
| 301 |
"learning_rate": 1.5200000000000002e-05,
|
| 302 |
+
"loss": 0.9461,
|
| 303 |
"step": 380
|
| 304 |
},
|
| 305 |
{
|
| 306 |
+
"epoch": 0.657672849915683,
|
| 307 |
+
"grad_norm": 10.535934448242188,
|
| 308 |
"learning_rate": 1.5600000000000003e-05,
|
| 309 |
+
"loss": 0.9695,
|
| 310 |
"step": 390
|
| 311 |
},
|
| 312 |
{
|
| 313 |
+
"epoch": 0.6745362563237775,
|
| 314 |
+
"grad_norm": 10.750452041625977,
|
| 315 |
"learning_rate": 1.6000000000000003e-05,
|
| 316 |
+
"loss": 1.0248,
|
| 317 |
"step": 400
|
| 318 |
},
|
| 319 |
{
|
| 320 |
+
"epoch": 0.6745362563237775,
|
| 321 |
+
"eval_accuracy": 0.6709006928406467,
|
| 322 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.9166666666666666,
|
| 323 |
+
"eval_accuracy_label_conflict, war and peace": 0.7978723404255319,
|
| 324 |
+
"eval_accuracy_label_crime, law and justice": 0.815028901734104,
|
| 325 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8625954198473282,
|
| 326 |
+
"eval_accuracy_label_economy, business, and finance": 0.7215189873417721,
|
| 327 |
+
"eval_accuracy_label_environment": 0.375,
|
| 328 |
"eval_accuracy_label_health": 0.9,
|
| 329 |
"eval_accuracy_label_human interest": 0.25,
|
| 330 |
"eval_accuracy_label_labour": 1.0,
|
| 331 |
+
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 332 |
+
"eval_accuracy_label_politics": 0.30935251798561153,
|
| 333 |
"eval_accuracy_label_religion": 0.0,
|
| 334 |
+
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 335 |
+
"eval_accuracy_label_society": 0.19298245614035087,
|
| 336 |
+
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 337 |
+
"eval_accuracy_label_weather": 1.0,
|
| 338 |
+
"eval_f1": 0.6591476133220594,
|
| 339 |
+
"eval_loss": 1.0774492025375366,
|
| 340 |
+
"eval_precision": 0.6983965116011075,
|
| 341 |
+
"eval_recall": 0.6709006928406467,
|
| 342 |
+
"eval_runtime": 6.9138,
|
| 343 |
+
"eval_samples_per_second": 125.256,
|
| 344 |
+
"eval_steps_per_second": 7.955,
|
| 345 |
"step": 400
|
| 346 |
},
|
| 347 |
{
|
| 348 |
+
"epoch": 0.6913996627318718,
|
| 349 |
+
"grad_norm": 11.188483238220215,
|
| 350 |
"learning_rate": 1.64e-05,
|
| 351 |
+
"loss": 0.8975,
|
| 352 |
"step": 410
|
| 353 |
},
|
| 354 |
{
|
| 355 |
+
"epoch": 0.7082630691399663,
|
| 356 |
+
"grad_norm": 8.035601615905762,
|
| 357 |
"learning_rate": 1.6800000000000002e-05,
|
| 358 |
+
"loss": 0.9736,
|
| 359 |
"step": 420
|
| 360 |
},
|
| 361 |
{
|
| 362 |
+
"epoch": 0.7251264755480608,
|
| 363 |
+
"grad_norm": 10.406453132629395,
|
| 364 |
"learning_rate": 1.72e-05,
|
| 365 |
+
"loss": 0.8802,
|
| 366 |
"step": 430
|
| 367 |
},
|
| 368 |
{
|
| 369 |
+
"epoch": 0.7419898819561551,
|
| 370 |
+
"grad_norm": 14.611414909362793,
|
| 371 |
"learning_rate": 1.76e-05,
|
| 372 |
+
"loss": 1.0106,
|
| 373 |
"step": 440
|
| 374 |
},
|
| 375 |
{
|
| 376 |
+
"epoch": 0.7588532883642496,
|
| 377 |
+
"grad_norm": 8.613202095031738,
|
| 378 |
"learning_rate": 1.8e-05,
|
| 379 |
+
"loss": 0.8411,
|
| 380 |
"step": 450
|
| 381 |
},
|
| 382 |
{
|
| 383 |
+
"epoch": 0.7757166947723441,
|
| 384 |
+
"grad_norm": 15.136896133422852,
|
| 385 |
"learning_rate": 1.8400000000000003e-05,
|
| 386 |
+
"loss": 0.736,
|
| 387 |
"step": 460
|
| 388 |
},
|
| 389 |
{
|
| 390 |
+
"epoch": 0.7925801011804384,
|
| 391 |
+
"grad_norm": 18.369375228881836,
|
| 392 |
"learning_rate": 1.88e-05,
|
| 393 |
+
"loss": 0.8807,
|
| 394 |
"step": 470
|
| 395 |
},
|
| 396 |
{
|
| 397 |
+
"epoch": 0.8094435075885329,
|
| 398 |
+
"grad_norm": 5.840969562530518,
|
| 399 |
"learning_rate": 1.9200000000000003e-05,
|
| 400 |
+
"loss": 0.8024,
|
| 401 |
"step": 480
|
| 402 |
},
|
| 403 |
{
|
| 404 |
+
"epoch": 0.8263069139966274,
|
| 405 |
+
"grad_norm": 14.514450073242188,
|
| 406 |
"learning_rate": 1.9600000000000002e-05,
|
| 407 |
+
"loss": 0.8334,
|
| 408 |
"step": 490
|
| 409 |
},
|
| 410 |
{
|
| 411 |
+
"epoch": 0.8431703204047217,
|
| 412 |
+
"grad_norm": 20.5344295501709,
|
| 413 |
"learning_rate": 2e-05,
|
| 414 |
+
"loss": 0.917,
|
| 415 |
"step": 500
|
| 416 |
},
|
| 417 |
{
|
| 418 |
+
"epoch": 0.8600337268128162,
|
| 419 |
+
"grad_norm": 8.904403686523438,
|
| 420 |
+
"learning_rate": 1.9843627834245506e-05,
|
| 421 |
+
"loss": 0.7659,
|
| 422 |
"step": 510
|
| 423 |
},
|
| 424 |
{
|
| 425 |
+
"epoch": 0.8768971332209107,
|
| 426 |
+
"grad_norm": 8.624781608581543,
|
| 427 |
+
"learning_rate": 1.968725566849101e-05,
|
| 428 |
+
"loss": 0.8243,
|
| 429 |
"step": 520
|
| 430 |
},
|
| 431 |
{
|
| 432 |
+
"epoch": 0.893760539629005,
|
| 433 |
+
"grad_norm": 14.050222396850586,
|
| 434 |
+
"learning_rate": 1.9530883502736514e-05,
|
| 435 |
+
"loss": 0.8085,
|
| 436 |
"step": 530
|
| 437 |
},
|
| 438 |
{
|
| 439 |
+
"epoch": 0.9106239460370995,
|
| 440 |
+
"grad_norm": 8.78807258605957,
|
| 441 |
+
"learning_rate": 1.9374511336982018e-05,
|
| 442 |
+
"loss": 0.6006,
|
| 443 |
"step": 540
|
| 444 |
},
|
| 445 |
{
|
| 446 |
+
"epoch": 0.927487352445194,
|
| 447 |
+
"grad_norm": 12.83305549621582,
|
| 448 |
+
"learning_rate": 1.9218139171227522e-05,
|
| 449 |
+
"loss": 0.8491,
|
| 450 |
"step": 550
|
| 451 |
},
|
| 452 |
{
|
| 453 |
+
"epoch": 0.9443507588532883,
|
| 454 |
+
"grad_norm": 10.532402038574219,
|
| 455 |
+
"learning_rate": 1.9061767005473026e-05,
|
| 456 |
+
"loss": 0.7823,
|
| 457 |
"step": 560
|
| 458 |
},
|
| 459 |
{
|
| 460 |
+
"epoch": 0.9612141652613828,
|
| 461 |
+
"grad_norm": 21.147523880004883,
|
| 462 |
+
"learning_rate": 1.890539483971853e-05,
|
| 463 |
+
"loss": 0.8382,
|
| 464 |
"step": 570
|
| 465 |
},
|
| 466 |
{
|
| 467 |
+
"epoch": 0.9780775716694773,
|
| 468 |
+
"grad_norm": 8.727555274963379,
|
| 469 |
+
"learning_rate": 1.8749022673964038e-05,
|
| 470 |
+
"loss": 0.7173,
|
| 471 |
"step": 580
|
| 472 |
},
|
| 473 |
{
|
| 474 |
+
"epoch": 0.9949409780775716,
|
| 475 |
+
"grad_norm": 9.259013175964355,
|
| 476 |
+
"learning_rate": 1.859265050820954e-05,
|
| 477 |
+
"loss": 0.7727,
|
| 478 |
"step": 590
|
| 479 |
},
|
| 480 |
{
|
| 481 |
+
"epoch": 1.0118043844856661,
|
| 482 |
+
"grad_norm": 6.446993350982666,
|
| 483 |
+
"learning_rate": 1.8436278342455046e-05,
|
| 484 |
+
"loss": 0.5845,
|
| 485 |
"step": 600
|
| 486 |
},
|
| 487 |
{
|
| 488 |
+
"epoch": 1.0118043844856661,
|
| 489 |
+
"eval_accuracy": 0.6535796766743649,
|
| 490 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.9166666666666666,
|
| 491 |
+
"eval_accuracy_label_conflict, war and peace": 0.7287234042553191,
|
| 492 |
+
"eval_accuracy_label_crime, law and justice": 0.6763005780346821,
|
| 493 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8778625954198473,
|
| 494 |
+
"eval_accuracy_label_economy, business, and finance": 0.7215189873417721,
|
| 495 |
+
"eval_accuracy_label_environment": 0.4375,
|
| 496 |
"eval_accuracy_label_health": 0.8,
|
| 497 |
+
"eval_accuracy_label_human interest": 0.0,
|
| 498 |
+
"eval_accuracy_label_labour": 1.0,
|
| 499 |
+
"eval_accuracy_label_lifestyle and leisure": 0.75,
|
| 500 |
+
"eval_accuracy_label_politics": 0.3669064748201439,
|
| 501 |
"eval_accuracy_label_religion": 0.0,
|
| 502 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 503 |
+
"eval_accuracy_label_society": 0.43859649122807015,
|
| 504 |
+
"eval_accuracy_label_sport": 0.9230769230769231,
|
| 505 |
"eval_accuracy_label_weather": 1.0,
|
| 506 |
+
"eval_f1": 0.6563434776934847,
|
| 507 |
+
"eval_loss": 0.990666389465332,
|
| 508 |
+
"eval_precision": 0.682947679005421,
|
| 509 |
+
"eval_recall": 0.6535796766743649,
|
| 510 |
+
"eval_runtime": 6.9215,
|
| 511 |
+
"eval_samples_per_second": 125.118,
|
| 512 |
+
"eval_steps_per_second": 7.946,
|
| 513 |
"step": 600
|
| 514 |
},
|
| 515 |
{
|
| 516 |
+
"epoch": 1.0286677908937605,
|
| 517 |
+
"grad_norm": 10.464855194091797,
|
| 518 |
+
"learning_rate": 1.8279906176700547e-05,
|
| 519 |
+
"loss": 0.565,
|
| 520 |
"step": 610
|
| 521 |
},
|
| 522 |
{
|
| 523 |
+
"epoch": 1.045531197301855,
|
| 524 |
+
"grad_norm": 10.188838958740234,
|
| 525 |
+
"learning_rate": 1.8123534010946054e-05,
|
| 526 |
+
"loss": 0.5325,
|
| 527 |
"step": 620
|
| 528 |
},
|
| 529 |
{
|
| 530 |
+
"epoch": 1.0623946037099494,
|
| 531 |
+
"grad_norm": 11.83839225769043,
|
| 532 |
+
"learning_rate": 1.7967161845191555e-05,
|
| 533 |
+
"loss": 0.603,
|
| 534 |
"step": 630
|
| 535 |
},
|
| 536 |
{
|
| 537 |
+
"epoch": 1.0792580101180438,
|
| 538 |
+
"grad_norm": 12.935112953186035,
|
| 539 |
+
"learning_rate": 1.7810789679437062e-05,
|
| 540 |
+
"loss": 0.5319,
|
| 541 |
"step": 640
|
| 542 |
},
|
| 543 |
{
|
| 544 |
+
"epoch": 1.0961214165261384,
|
| 545 |
+
"grad_norm": 8.650348663330078,
|
| 546 |
+
"learning_rate": 1.7654417513682566e-05,
|
| 547 |
+
"loss": 0.4803,
|
| 548 |
"step": 650
|
| 549 |
},
|
| 550 |
{
|
| 551 |
+
"epoch": 1.1129848229342327,
|
| 552 |
+
"grad_norm": 12.213525772094727,
|
| 553 |
+
"learning_rate": 1.749804534792807e-05,
|
| 554 |
+
"loss": 0.6975,
|
| 555 |
"step": 660
|
| 556 |
},
|
| 557 |
{
|
| 558 |
+
"epoch": 1.129848229342327,
|
| 559 |
+
"grad_norm": 12.859235763549805,
|
| 560 |
+
"learning_rate": 1.7341673182173575e-05,
|
| 561 |
+
"loss": 0.6273,
|
| 562 |
"step": 670
|
| 563 |
},
|
| 564 |
{
|
| 565 |
+
"epoch": 1.1467116357504217,
|
| 566 |
+
"grad_norm": 8.133111953735352,
|
| 567 |
+
"learning_rate": 1.718530101641908e-05,
|
| 568 |
+
"loss": 0.5179,
|
| 569 |
"step": 680
|
| 570 |
},
|
| 571 |
{
|
| 572 |
+
"epoch": 1.163575042158516,
|
| 573 |
+
"grad_norm": 8.366670608520508,
|
| 574 |
+
"learning_rate": 1.7028928850664583e-05,
|
| 575 |
+
"loss": 0.5622,
|
| 576 |
"step": 690
|
| 577 |
},
|
| 578 |
{
|
| 579 |
+
"epoch": 1.1804384485666104,
|
| 580 |
+
"grad_norm": 12.844072341918945,
|
| 581 |
+
"learning_rate": 1.6872556684910087e-05,
|
| 582 |
+
"loss": 0.6347,
|
| 583 |
"step": 700
|
| 584 |
},
|
| 585 |
{
|
| 586 |
+
"epoch": 1.197301854974705,
|
| 587 |
+
"grad_norm": 3.2994637489318848,
|
| 588 |
+
"learning_rate": 1.671618451915559e-05,
|
| 589 |
+
"loss": 0.4805,
|
| 590 |
"step": 710
|
| 591 |
},
|
| 592 |
{
|
| 593 |
+
"epoch": 1.2141652613827993,
|
| 594 |
+
"grad_norm": 6.4653167724609375,
|
| 595 |
+
"learning_rate": 1.6559812353401095e-05,
|
| 596 |
+
"loss": 0.5172,
|
| 597 |
"step": 720
|
| 598 |
},
|
| 599 |
{
|
| 600 |
+
"epoch": 1.2310286677908937,
|
| 601 |
+
"grad_norm": 6.744974613189697,
|
| 602 |
+
"learning_rate": 1.64034401876466e-05,
|
| 603 |
+
"loss": 0.5601,
|
| 604 |
"step": 730
|
| 605 |
},
|
| 606 |
{
|
| 607 |
+
"epoch": 1.2478920741989883,
|
| 608 |
+
"grad_norm": 9.451184272766113,
|
| 609 |
+
"learning_rate": 1.6247068021892107e-05,
|
| 610 |
+
"loss": 0.5289,
|
| 611 |
"step": 740
|
| 612 |
},
|
| 613 |
{
|
| 614 |
+
"epoch": 1.2647554806070826,
|
| 615 |
+
"grad_norm": 5.093151092529297,
|
| 616 |
+
"learning_rate": 1.6090695856137607e-05,
|
| 617 |
+
"loss": 0.5674,
|
| 618 |
"step": 750
|
| 619 |
},
|
| 620 |
{
|
| 621 |
+
"epoch": 1.281618887015177,
|
| 622 |
+
"grad_norm": 16.99575424194336,
|
| 623 |
+
"learning_rate": 1.5934323690383115e-05,
|
| 624 |
+
"loss": 0.4681,
|
| 625 |
"step": 760
|
| 626 |
},
|
| 627 |
{
|
| 628 |
+
"epoch": 1.2984822934232714,
|
| 629 |
+
"grad_norm": 17.180509567260742,
|
| 630 |
+
"learning_rate": 1.5777951524628616e-05,
|
| 631 |
+
"loss": 0.5974,
|
| 632 |
"step": 770
|
| 633 |
},
|
| 634 |
{
|
| 635 |
+
"epoch": 1.315345699831366,
|
| 636 |
+
"grad_norm": 11.804935455322266,
|
| 637 |
+
"learning_rate": 1.5621579358874123e-05,
|
| 638 |
+
"loss": 0.5693,
|
| 639 |
"step": 780
|
| 640 |
},
|
| 641 |
{
|
| 642 |
+
"epoch": 1.3322091062394603,
|
| 643 |
+
"grad_norm": 12.154623985290527,
|
| 644 |
+
"learning_rate": 1.5465207193119627e-05,
|
| 645 |
+
"loss": 0.6001,
|
| 646 |
"step": 790
|
| 647 |
},
|
| 648 |
{
|
| 649 |
+
"epoch": 1.3490725126475547,
|
| 650 |
+
"grad_norm": 11.09798526763916,
|
| 651 |
+
"learning_rate": 1.530883502736513e-05,
|
| 652 |
+
"loss": 0.6104,
|
| 653 |
"step": 800
|
| 654 |
},
|
| 655 |
{
|
| 656 |
+
"epoch": 1.3490725126475547,
|
| 657 |
+
"eval_accuracy": 0.7240184757505773,
|
| 658 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 659 |
+
"eval_accuracy_label_conflict, war and peace": 0.7021276595744681,
|
| 660 |
+
"eval_accuracy_label_crime, law and justice": 0.8323699421965318,
|
| 661 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8778625954198473,
|
| 662 |
+
"eval_accuracy_label_economy, business, and finance": 0.7848101265822784,
|
| 663 |
+
"eval_accuracy_label_environment": 0.5,
|
| 664 |
+
"eval_accuracy_label_health": 0.7,
|
| 665 |
+
"eval_accuracy_label_human interest": 0.25,
|
| 666 |
+
"eval_accuracy_label_labour": 1.0,
|
| 667 |
+
"eval_accuracy_label_lifestyle and leisure": 0.75,
|
| 668 |
+
"eval_accuracy_label_politics": 0.6330935251798561,
|
| 669 |
"eval_accuracy_label_religion": 0.0,
|
| 670 |
+
"eval_accuracy_label_science and technology": 0.25,
|
| 671 |
+
"eval_accuracy_label_society": 0.3684210526315789,
|
| 672 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 673 |
"eval_accuracy_label_weather": 1.0,
|
| 674 |
+
"eval_f1": 0.7232835318516546,
|
| 675 |
+
"eval_loss": 0.8674274682998657,
|
| 676 |
+
"eval_precision": 0.7332860565477397,
|
| 677 |
+
"eval_recall": 0.7240184757505773,
|
| 678 |
+
"eval_runtime": 6.8885,
|
| 679 |
+
"eval_samples_per_second": 125.717,
|
| 680 |
+
"eval_steps_per_second": 7.984,
|
| 681 |
"step": 800
|
| 682 |
},
|
| 683 |
{
|
| 684 |
+
"epoch": 1.3659359190556493,
|
| 685 |
+
"grad_norm": 12.72695255279541,
|
| 686 |
+
"learning_rate": 1.5152462861610635e-05,
|
| 687 |
+
"loss": 0.5333,
|
| 688 |
"step": 810
|
| 689 |
},
|
| 690 |
{
|
| 691 |
+
"epoch": 1.3827993254637436,
|
| 692 |
+
"grad_norm": 9.709075927734375,
|
| 693 |
+
"learning_rate": 1.4996090695856138e-05,
|
| 694 |
+
"loss": 0.488,
|
| 695 |
"step": 820
|
| 696 |
},
|
| 697 |
{
|
| 698 |
+
"epoch": 1.399662731871838,
|
| 699 |
+
"grad_norm": 11.909979820251465,
|
| 700 |
+
"learning_rate": 1.4839718530101644e-05,
|
| 701 |
+
"loss": 0.6037,
|
| 702 |
"step": 830
|
| 703 |
},
|
| 704 |
{
|
| 705 |
+
"epoch": 1.4165261382799326,
|
| 706 |
+
"grad_norm": 16.916475296020508,
|
| 707 |
+
"learning_rate": 1.4683346364347146e-05,
|
| 708 |
+
"loss": 0.5647,
|
| 709 |
"step": 840
|
| 710 |
},
|
| 711 |
{
|
| 712 |
+
"epoch": 1.433389544688027,
|
| 713 |
+
"grad_norm": 13.278093338012695,
|
| 714 |
+
"learning_rate": 1.4526974198592652e-05,
|
| 715 |
+
"loss": 0.5078,
|
| 716 |
"step": 850
|
| 717 |
},
|
| 718 |
{
|
| 719 |
+
"epoch": 1.4502529510961213,
|
| 720 |
+
"grad_norm": 7.877659320831299,
|
| 721 |
+
"learning_rate": 1.4370602032838158e-05,
|
| 722 |
+
"loss": 0.5108,
|
| 723 |
"step": 860
|
| 724 |
},
|
| 725 |
{
|
| 726 |
+
"epoch": 1.4671163575042159,
|
| 727 |
+
"grad_norm": 7.198855400085449,
|
| 728 |
+
"learning_rate": 1.421422986708366e-05,
|
| 729 |
+
"loss": 0.537,
|
| 730 |
"step": 870
|
| 731 |
},
|
| 732 |
{
|
| 733 |
+
"epoch": 1.4839797639123102,
|
| 734 |
+
"grad_norm": 7.947418212890625,
|
| 735 |
+
"learning_rate": 1.4057857701329166e-05,
|
| 736 |
+
"loss": 0.4431,
|
| 737 |
"step": 880
|
| 738 |
},
|
| 739 |
{
|
| 740 |
+
"epoch": 1.5008431703204046,
|
| 741 |
+
"grad_norm": 8.317455291748047,
|
| 742 |
+
"learning_rate": 1.3901485535574668e-05,
|
| 743 |
+
"loss": 0.4532,
|
| 744 |
"step": 890
|
| 745 |
},
|
| 746 |
{
|
| 747 |
+
"epoch": 1.5177065767284992,
|
| 748 |
+
"grad_norm": 8.073487281799316,
|
| 749 |
+
"learning_rate": 1.3745113369820174e-05,
|
| 750 |
+
"loss": 0.4624,
|
| 751 |
"step": 900
|
| 752 |
},
|
| 753 |
{
|
| 754 |
+
"epoch": 1.5345699831365935,
|
| 755 |
+
"grad_norm": 9.39900016784668,
|
| 756 |
+
"learning_rate": 1.3588741204065676e-05,
|
| 757 |
+
"loss": 0.3739,
|
| 758 |
"step": 910
|
| 759 |
},
|
| 760 |
{
|
| 761 |
+
"epoch": 1.551433389544688,
|
| 762 |
+
"grad_norm": 7.912290096282959,
|
| 763 |
+
"learning_rate": 1.3432369038311182e-05,
|
| 764 |
+
"loss": 0.497,
|
| 765 |
"step": 920
|
| 766 |
},
|
| 767 |
{
|
| 768 |
+
"epoch": 1.5682967959527825,
|
| 769 |
+
"grad_norm": 12.181069374084473,
|
| 770 |
+
"learning_rate": 1.3275996872556686e-05,
|
| 771 |
+
"loss": 0.4859,
|
| 772 |
"step": 930
|
| 773 |
},
|
| 774 |
{
|
| 775 |
+
"epoch": 1.5851602023608768,
|
| 776 |
+
"grad_norm": 9.931648254394531,
|
| 777 |
+
"learning_rate": 1.311962470680219e-05,
|
| 778 |
+
"loss": 0.4945,
|
| 779 |
"step": 940
|
| 780 |
},
|
| 781 |
{
|
| 782 |
+
"epoch": 1.6020236087689712,
|
| 783 |
+
"grad_norm": 7.948305606842041,
|
| 784 |
+
"learning_rate": 1.2963252541047694e-05,
|
| 785 |
+
"loss": 0.3742,
|
| 786 |
"step": 950
|
| 787 |
},
|
| 788 |
{
|
| 789 |
+
"epoch": 1.6188870151770658,
|
| 790 |
+
"grad_norm": 9.558001518249512,
|
| 791 |
+
"learning_rate": 1.2806880375293199e-05,
|
| 792 |
+
"loss": 0.4684,
|
| 793 |
"step": 960
|
| 794 |
},
|
| 795 |
{
|
| 796 |
+
"epoch": 1.6357504215851602,
|
| 797 |
+
"grad_norm": 8.381938934326172,
|
| 798 |
+
"learning_rate": 1.2650508209538703e-05,
|
| 799 |
+
"loss": 0.4564,
|
| 800 |
"step": 970
|
| 801 |
},
|
| 802 |
{
|
| 803 |
+
"epoch": 1.6526138279932545,
|
| 804 |
+
"grad_norm": 7.728443622589111,
|
| 805 |
+
"learning_rate": 1.2494136043784208e-05,
|
| 806 |
+
"loss": 0.4493,
|
| 807 |
"step": 980
|
| 808 |
},
|
| 809 |
{
|
| 810 |
+
"epoch": 1.669477234401349,
|
| 811 |
+
"grad_norm": 14.073138236999512,
|
| 812 |
+
"learning_rate": 1.233776387802971e-05,
|
| 813 |
+
"loss": 0.4085,
|
| 814 |
"step": 990
|
| 815 |
},
|
| 816 |
{
|
| 817 |
+
"epoch": 1.6863406408094435,
|
| 818 |
+
"grad_norm": 6.431807518005371,
|
| 819 |
+
"learning_rate": 1.2181391712275217e-05,
|
| 820 |
+
"loss": 0.4223,
|
| 821 |
"step": 1000
|
| 822 |
},
|
| 823 |
{
|
| 824 |
+
"epoch": 1.6863406408094435,
|
| 825 |
+
"eval_accuracy": 0.7240184757505773,
|
| 826 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.75,
|
| 827 |
+
"eval_accuracy_label_conflict, war and peace": 0.675531914893617,
|
| 828 |
+
"eval_accuracy_label_crime, law and justice": 0.884393063583815,
|
| 829 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8549618320610687,
|
| 830 |
+
"eval_accuracy_label_economy, business, and finance": 0.7341772151898734,
|
| 831 |
+
"eval_accuracy_label_environment": 0.5,
|
| 832 |
+
"eval_accuracy_label_health": 0.9,
|
| 833 |
+
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 834 |
+
"eval_accuracy_label_labour": 1.0,
|
| 835 |
"eval_accuracy_label_lifestyle and leisure": 0.625,
|
| 836 |
"eval_accuracy_label_politics": 0.6474820143884892,
|
| 837 |
"eval_accuracy_label_religion": 0.0,
|
| 838 |
+
"eval_accuracy_label_science and technology": 0.3333333333333333,
|
| 839 |
+
"eval_accuracy_label_society": 0.3684210526315789,
|
| 840 |
+
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 841 |
+
"eval_accuracy_label_weather": 0.0,
|
| 842 |
+
"eval_f1": 0.7249701679054074,
|
| 843 |
+
"eval_loss": 0.8601514101028442,
|
| 844 |
+
"eval_precision": 0.7387214985313825,
|
| 845 |
+
"eval_recall": 0.7240184757505773,
|
| 846 |
+
"eval_runtime": 6.9046,
|
| 847 |
+
"eval_samples_per_second": 125.423,
|
| 848 |
+
"eval_steps_per_second": 7.966,
|
| 849 |
"step": 1000
|
| 850 |
},
|
| 851 |
{
|
| 852 |
+
"epoch": 1.7032040472175378,
|
| 853 |
+
"grad_norm": 5.857812404632568,
|
| 854 |
+
"learning_rate": 1.2025019546520719e-05,
|
| 855 |
+
"loss": 0.4261,
|
| 856 |
"step": 1010
|
| 857 |
},
|
| 858 |
{
|
| 859 |
+
"epoch": 1.7200674536256324,
|
| 860 |
+
"grad_norm": 6.620238304138184,
|
| 861 |
+
"learning_rate": 1.1868647380766225e-05,
|
| 862 |
+
"loss": 0.4071,
|
| 863 |
"step": 1020
|
| 864 |
},
|
| 865 |
{
|
| 866 |
+
"epoch": 1.7369308600337268,
|
| 867 |
+
"grad_norm": 13.199769973754883,
|
| 868 |
+
"learning_rate": 1.1712275215011727e-05,
|
| 869 |
+
"loss": 0.3632,
|
| 870 |
"step": 1030
|
| 871 |
},
|
| 872 |
{
|
| 873 |
+
"epoch": 1.7537942664418211,
|
| 874 |
+
"grad_norm": 8.069077491760254,
|
| 875 |
+
"learning_rate": 1.1555903049257233e-05,
|
| 876 |
+
"loss": 0.3977,
|
| 877 |
"step": 1040
|
| 878 |
},
|
| 879 |
{
|
| 880 |
+
"epoch": 1.7706576728499157,
|
| 881 |
+
"grad_norm": 12.82445240020752,
|
| 882 |
+
"learning_rate": 1.1399530883502739e-05,
|
| 883 |
+
"loss": 0.4245,
|
| 884 |
"step": 1050
|
| 885 |
},
|
| 886 |
{
|
| 887 |
+
"epoch": 1.78752107925801,
|
| 888 |
+
"grad_norm": 15.626612663269043,
|
| 889 |
+
"learning_rate": 1.1243158717748241e-05,
|
| 890 |
+
"loss": 0.3924,
|
| 891 |
"step": 1060
|
| 892 |
},
|
| 893 |
{
|
| 894 |
+
"epoch": 1.8043844856661044,
|
| 895 |
+
"grad_norm": 14.491767883300781,
|
| 896 |
+
"learning_rate": 1.1086786551993747e-05,
|
| 897 |
+
"loss": 0.427,
|
| 898 |
"step": 1070
|
| 899 |
},
|
| 900 |
{
|
| 901 |
+
"epoch": 1.821247892074199,
|
| 902 |
+
"grad_norm": 5.425060749053955,
|
| 903 |
+
"learning_rate": 1.093041438623925e-05,
|
| 904 |
+
"loss": 0.4013,
|
| 905 |
"step": 1080
|
| 906 |
},
|
| 907 |
{
|
| 908 |
+
"epoch": 1.8381112984822934,
|
| 909 |
+
"grad_norm": 10.305951118469238,
|
| 910 |
+
"learning_rate": 1.0774042220484755e-05,
|
| 911 |
+
"loss": 0.4036,
|
| 912 |
"step": 1090
|
| 913 |
},
|
| 914 |
{
|
| 915 |
+
"epoch": 1.8549747048903877,
|
| 916 |
+
"grad_norm": 10.018141746520996,
|
| 917 |
+
"learning_rate": 1.0617670054730258e-05,
|
| 918 |
+
"loss": 0.4123,
|
| 919 |
"step": 1100
|
| 920 |
},
|
| 921 |
{
|
| 922 |
+
"epoch": 1.8718381112984823,
|
| 923 |
+
"grad_norm": 13.072948455810547,
|
| 924 |
+
"learning_rate": 1.0461297888975763e-05,
|
| 925 |
+
"loss": 0.4586,
|
| 926 |
"step": 1110
|
| 927 |
},
|
| 928 |
{
|
| 929 |
+
"epoch": 1.8887015177065767,
|
| 930 |
+
"grad_norm": 7.460719585418701,
|
| 931 |
+
"learning_rate": 1.030492572322127e-05,
|
| 932 |
+
"loss": 0.3912,
|
| 933 |
"step": 1120
|
| 934 |
},
|
| 935 |
{
|
| 936 |
+
"epoch": 1.905564924114671,
|
| 937 |
+
"grad_norm": 5.454730033874512,
|
| 938 |
+
"learning_rate": 1.0148553557466772e-05,
|
| 939 |
+
"loss": 0.452,
|
| 940 |
"step": 1130
|
| 941 |
},
|
| 942 |
{
|
| 943 |
+
"epoch": 1.9224283305227656,
|
| 944 |
+
"grad_norm": 8.758441925048828,
|
| 945 |
+
"learning_rate": 9.992181391712276e-06,
|
| 946 |
+
"loss": 0.3622,
|
| 947 |
"step": 1140
|
| 948 |
},
|
| 949 |
{
|
| 950 |
+
"epoch": 1.93929173693086,
|
| 951 |
+
"grad_norm": 8.622238159179688,
|
| 952 |
+
"learning_rate": 9.835809225957781e-06,
|
| 953 |
+
"loss": 0.439,
|
| 954 |
"step": 1150
|
| 955 |
},
|
| 956 |
{
|
| 957 |
+
"epoch": 1.9561551433389543,
|
| 958 |
+
"grad_norm": 6.299844741821289,
|
| 959 |
+
"learning_rate": 9.679437060203286e-06,
|
| 960 |
+
"loss": 0.4006,
|
| 961 |
"step": 1160
|
| 962 |
},
|
| 963 |
{
|
| 964 |
+
"epoch": 1.973018549747049,
|
| 965 |
+
"grad_norm": 14.24021053314209,
|
| 966 |
+
"learning_rate": 9.52306489444879e-06,
|
| 967 |
+
"loss": 0.3616,
|
| 968 |
"step": 1170
|
| 969 |
},
|
| 970 |
{
|
| 971 |
+
"epoch": 1.9898819561551433,
|
| 972 |
+
"grad_norm": 20.004261016845703,
|
| 973 |
+
"learning_rate": 9.366692728694294e-06,
|
| 974 |
+
"loss": 0.4253,
|
| 975 |
"step": 1180
|
| 976 |
},
|
| 977 |
{
|
| 978 |
+
"epoch": 2.0067453625632377,
|
| 979 |
+
"grad_norm": 12.190637588500977,
|
| 980 |
+
"learning_rate": 9.210320562939798e-06,
|
| 981 |
+
"loss": 0.3859,
|
| 982 |
"step": 1190
|
| 983 |
},
|
| 984 |
{
|
| 985 |
+
"epoch": 2.0236087689713322,
|
| 986 |
+
"grad_norm": 3.0981788635253906,
|
| 987 |
+
"learning_rate": 9.053948397185302e-06,
|
| 988 |
+
"loss": 0.3104,
|
| 989 |
"step": 1200
|
| 990 |
},
|
| 991 |
{
|
| 992 |
+
"epoch": 2.0236087689713322,
|
| 993 |
+
"eval_accuracy": 0.7263279445727483,
|
| 994 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 995 |
+
"eval_accuracy_label_conflict, war and peace": 0.7180851063829787,
|
| 996 |
+
"eval_accuracy_label_crime, law and justice": 0.8323699421965318,
|
| 997 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.9083969465648855,
|
| 998 |
+
"eval_accuracy_label_economy, business, and finance": 0.7721518987341772,
|
| 999 |
+
"eval_accuracy_label_environment": 0.4375,
|
| 1000 |
+
"eval_accuracy_label_health": 0.7,
|
| 1001 |
"eval_accuracy_label_human interest": 0.25,
|
| 1002 |
"eval_accuracy_label_labour": 0.5,
|
| 1003 |
+
"eval_accuracy_label_lifestyle and leisure": 0.75,
|
| 1004 |
+
"eval_accuracy_label_politics": 0.5611510791366906,
|
| 1005 |
"eval_accuracy_label_religion": 0.0,
|
| 1006 |
+
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 1007 |
+
"eval_accuracy_label_society": 0.47368421052631576,
|
| 1008 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1009 |
"eval_accuracy_label_weather": 1.0,
|
| 1010 |
+
"eval_f1": 0.7266120737327485,
|
| 1011 |
+
"eval_loss": 0.856505811214447,
|
| 1012 |
+
"eval_precision": 0.7326492002435036,
|
| 1013 |
+
"eval_recall": 0.7263279445727483,
|
| 1014 |
+
"eval_runtime": 6.9128,
|
| 1015 |
+
"eval_samples_per_second": 125.275,
|
| 1016 |
+
"eval_steps_per_second": 7.956,
|
| 1017 |
"step": 1200
|
| 1018 |
},
|
| 1019 |
{
|
| 1020 |
+
"epoch": 2.040472175379427,
|
| 1021 |
+
"grad_norm": 5.126299858093262,
|
| 1022 |
+
"learning_rate": 8.897576231430806e-06,
|
| 1023 |
+
"loss": 0.3066,
|
| 1024 |
"step": 1210
|
| 1025 |
},
|
| 1026 |
{
|
| 1027 |
+
"epoch": 2.057335581787521,
|
| 1028 |
+
"grad_norm": 5.153671741485596,
|
| 1029 |
+
"learning_rate": 8.74120406567631e-06,
|
| 1030 |
+
"loss": 0.3275,
|
| 1031 |
"step": 1220
|
| 1032 |
},
|
| 1033 |
{
|
| 1034 |
+
"epoch": 2.0741989881956155,
|
| 1035 |
+
"grad_norm": 6.91109561920166,
|
| 1036 |
+
"learning_rate": 8.584831899921814e-06,
|
| 1037 |
+
"loss": 0.2581,
|
| 1038 |
"step": 1230
|
| 1039 |
},
|
| 1040 |
{
|
| 1041 |
+
"epoch": 2.09106239460371,
|
| 1042 |
+
"grad_norm": 9.57040023803711,
|
| 1043 |
+
"learning_rate": 8.428459734167318e-06,
|
| 1044 |
+
"loss": 0.2809,
|
| 1045 |
"step": 1240
|
| 1046 |
},
|
| 1047 |
{
|
| 1048 |
+
"epoch": 2.1079258010118043,
|
| 1049 |
+
"grad_norm": 3.024446964263916,
|
| 1050 |
+
"learning_rate": 8.272087568412822e-06,
|
| 1051 |
+
"loss": 0.2288,
|
| 1052 |
"step": 1250
|
| 1053 |
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.124789207419899,
|
| 1056 |
+
"grad_norm": 6.808653831481934,
|
| 1057 |
+
"learning_rate": 8.115715402658327e-06,
|
| 1058 |
+
"loss": 0.2406,
|
| 1059 |
+
"step": 1260
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.1416526138279934,
|
| 1063 |
+
"grad_norm": 15.430974960327148,
|
| 1064 |
+
"learning_rate": 7.959343236903832e-06,
|
| 1065 |
+
"loss": 0.3031,
|
| 1066 |
+
"step": 1270
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.1585160202360876,
|
| 1070 |
+
"grad_norm": 6.104337215423584,
|
| 1071 |
+
"learning_rate": 7.802971071149336e-06,
|
| 1072 |
+
"loss": 0.2575,
|
| 1073 |
+
"step": 1280
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.175379426644182,
|
| 1077 |
+
"grad_norm": 3.1404926776885986,
|
| 1078 |
+
"learning_rate": 7.64659890539484e-06,
|
| 1079 |
+
"loss": 0.264,
|
| 1080 |
+
"step": 1290
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.1922428330522767,
|
| 1084 |
+
"grad_norm": 11.821499824523926,
|
| 1085 |
+
"learning_rate": 7.490226739640345e-06,
|
| 1086 |
+
"loss": 0.2602,
|
| 1087 |
+
"step": 1300
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.209106239460371,
|
| 1091 |
+
"grad_norm": 4.936356544494629,
|
| 1092 |
+
"learning_rate": 7.333854573885849e-06,
|
| 1093 |
+
"loss": 0.3428,
|
| 1094 |
+
"step": 1310
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.2259696458684655,
|
| 1098 |
+
"grad_norm": 12.609466552734375,
|
| 1099 |
+
"learning_rate": 7.177482408131353e-06,
|
| 1100 |
+
"loss": 0.2467,
|
| 1101 |
+
"step": 1320
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.24283305227656,
|
| 1105 |
+
"grad_norm": 10.444657325744629,
|
| 1106 |
+
"learning_rate": 7.021110242376858e-06,
|
| 1107 |
+
"loss": 0.3149,
|
| 1108 |
+
"step": 1330
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.259696458684654,
|
| 1112 |
+
"grad_norm": 9.078226089477539,
|
| 1113 |
+
"learning_rate": 6.864738076622362e-06,
|
| 1114 |
+
"loss": 0.3207,
|
| 1115 |
+
"step": 1340
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.2765598650927488,
|
| 1119 |
+
"grad_norm": 7.427735328674316,
|
| 1120 |
+
"learning_rate": 6.708365910867866e-06,
|
| 1121 |
+
"loss": 0.2458,
|
| 1122 |
+
"step": 1350
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.2934232715008434,
|
| 1126 |
+
"grad_norm": 6.552321910858154,
|
| 1127 |
+
"learning_rate": 6.551993745113371e-06,
|
| 1128 |
+
"loss": 0.273,
|
| 1129 |
+
"step": 1360
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.3102866779089375,
|
| 1133 |
+
"grad_norm": 5.591136932373047,
|
| 1134 |
+
"learning_rate": 6.395621579358875e-06,
|
| 1135 |
+
"loss": 0.2345,
|
| 1136 |
+
"step": 1370
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.327150084317032,
|
| 1140 |
+
"grad_norm": 9.349672317504883,
|
| 1141 |
+
"learning_rate": 6.239249413604379e-06,
|
| 1142 |
+
"loss": 0.3254,
|
| 1143 |
+
"step": 1380
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.3440134907251267,
|
| 1147 |
+
"grad_norm": 7.311215400695801,
|
| 1148 |
+
"learning_rate": 6.082877247849883e-06,
|
| 1149 |
+
"loss": 0.2511,
|
| 1150 |
+
"step": 1390
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.360876897133221,
|
| 1154 |
+
"grad_norm": 9.266729354858398,
|
| 1155 |
+
"learning_rate": 5.926505082095387e-06,
|
| 1156 |
+
"loss": 0.2855,
|
| 1157 |
+
"step": 1400
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.360876897133221,
|
| 1161 |
+
"eval_accuracy": 0.7240184757505773,
|
| 1162 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.75,
|
| 1163 |
+
"eval_accuracy_label_conflict, war and peace": 0.7393617021276596,
|
| 1164 |
+
"eval_accuracy_label_crime, law and justice": 0.8323699421965318,
|
| 1165 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8549618320610687,
|
| 1166 |
+
"eval_accuracy_label_economy, business, and finance": 0.7974683544303798,
|
| 1167 |
+
"eval_accuracy_label_environment": 0.5,
|
| 1168 |
+
"eval_accuracy_label_health": 0.7,
|
| 1169 |
+
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 1170 |
+
"eval_accuracy_label_labour": 0.5,
|
| 1171 |
+
"eval_accuracy_label_lifestyle and leisure": 0.625,
|
| 1172 |
+
"eval_accuracy_label_politics": 0.5899280575539568,
|
| 1173 |
+
"eval_accuracy_label_religion": 0.0,
|
| 1174 |
+
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 1175 |
+
"eval_accuracy_label_society": 0.38596491228070173,
|
| 1176 |
+
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1177 |
+
"eval_accuracy_label_weather": 1.0,
|
| 1178 |
+
"eval_f1": 0.7283152161972767,
|
| 1179 |
+
"eval_loss": 0.8981179594993591,
|
| 1180 |
+
"eval_precision": 0.7401859274408267,
|
| 1181 |
+
"eval_recall": 0.7240184757505773,
|
| 1182 |
+
"eval_runtime": 6.8981,
|
| 1183 |
+
"eval_samples_per_second": 125.541,
|
| 1184 |
+
"eval_steps_per_second": 7.973,
|
| 1185 |
+
"step": 1400
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.3777403035413154,
|
| 1189 |
+
"grad_norm": 12.713851928710938,
|
| 1190 |
+
"learning_rate": 5.770132916340891e-06,
|
| 1191 |
+
"loss": 0.2885,
|
| 1192 |
+
"step": 1410
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.39460370994941,
|
| 1196 |
+
"grad_norm": 5.382158279418945,
|
| 1197 |
+
"learning_rate": 5.6137607505863955e-06,
|
| 1198 |
+
"loss": 0.3191,
|
| 1199 |
+
"step": 1420
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.411467116357504,
|
| 1203 |
+
"grad_norm": 4.609128475189209,
|
| 1204 |
+
"learning_rate": 5.4573885848319e-06,
|
| 1205 |
+
"loss": 0.273,
|
| 1206 |
+
"step": 1430
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 2.4283305227655987,
|
| 1210 |
+
"grad_norm": 5.821180820465088,
|
| 1211 |
+
"learning_rate": 5.301016419077405e-06,
|
| 1212 |
+
"loss": 0.2397,
|
| 1213 |
+
"step": 1440
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 2.4451939291736933,
|
| 1217 |
+
"grad_norm": 3.8270256519317627,
|
| 1218 |
+
"learning_rate": 5.1446442533229095e-06,
|
| 1219 |
+
"loss": 0.3575,
|
| 1220 |
+
"step": 1450
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.4620573355817874,
|
| 1224 |
+
"grad_norm": 4.910329341888428,
|
| 1225 |
+
"learning_rate": 4.9882720875684136e-06,
|
| 1226 |
+
"loss": 0.2622,
|
| 1227 |
+
"step": 1460
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 2.478920741989882,
|
| 1231 |
+
"grad_norm": 6.537694931030273,
|
| 1232 |
+
"learning_rate": 4.831899921813918e-06,
|
| 1233 |
+
"loss": 0.1981,
|
| 1234 |
+
"step": 1470
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 2.4957841483979766,
|
| 1238 |
+
"grad_norm": 9.600226402282715,
|
| 1239 |
+
"learning_rate": 4.675527756059422e-06,
|
| 1240 |
+
"loss": 0.1977,
|
| 1241 |
+
"step": 1480
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.5126475548060707,
|
| 1245 |
+
"grad_norm": 5.246051788330078,
|
| 1246 |
+
"learning_rate": 4.519155590304926e-06,
|
| 1247 |
+
"loss": 0.2452,
|
| 1248 |
+
"step": 1490
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 2.5295109612141653,
|
| 1252 |
+
"grad_norm": 5.955575942993164,
|
| 1253 |
+
"learning_rate": 4.36278342455043e-06,
|
| 1254 |
+
"loss": 0.2239,
|
| 1255 |
+
"step": 1500
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 2.54637436762226,
|
| 1259 |
+
"grad_norm": 9.812554359436035,
|
| 1260 |
+
"learning_rate": 4.206411258795935e-06,
|
| 1261 |
+
"loss": 0.2773,
|
| 1262 |
+
"step": 1510
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.563237774030354,
|
| 1266 |
+
"grad_norm": 11.0867280960083,
|
| 1267 |
+
"learning_rate": 4.050039093041439e-06,
|
| 1268 |
+
"loss": 0.1866,
|
| 1269 |
+
"step": 1520
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 2.5801011804384486,
|
| 1273 |
+
"grad_norm": 6.4792609214782715,
|
| 1274 |
+
"learning_rate": 3.893666927286943e-06,
|
| 1275 |
+
"loss": 0.2556,
|
| 1276 |
+
"step": 1530
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 2.5969645868465427,
|
| 1280 |
+
"grad_norm": 9.363513946533203,
|
| 1281 |
+
"learning_rate": 3.737294761532447e-06,
|
| 1282 |
+
"loss": 0.2593,
|
| 1283 |
+
"step": 1540
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.6138279932546373,
|
| 1287 |
+
"grad_norm": 11.11075496673584,
|
| 1288 |
+
"learning_rate": 3.580922595777952e-06,
|
| 1289 |
+
"loss": 0.3482,
|
| 1290 |
+
"step": 1550
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 2.630691399662732,
|
| 1294 |
+
"grad_norm": 2.9576575756073,
|
| 1295 |
+
"learning_rate": 3.4245504300234562e-06,
|
| 1296 |
+
"loss": 0.2412,
|
| 1297 |
+
"step": 1560
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 2.6475548060708265,
|
| 1301 |
+
"grad_norm": 9.084090232849121,
|
| 1302 |
+
"learning_rate": 3.2681782642689603e-06,
|
| 1303 |
+
"loss": 0.3092,
|
| 1304 |
+
"step": 1570
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.6644182124789206,
|
| 1308 |
+
"grad_norm": 7.185946941375732,
|
| 1309 |
+
"learning_rate": 3.111806098514465e-06,
|
| 1310 |
+
"loss": 0.2406,
|
| 1311 |
+
"step": 1580
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 2.681281618887015,
|
| 1315 |
+
"grad_norm": 5.467381000518799,
|
| 1316 |
+
"learning_rate": 2.955433932759969e-06,
|
| 1317 |
+
"loss": 0.2709,
|
| 1318 |
+
"step": 1590
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 2.6981450252951094,
|
| 1322 |
+
"grad_norm": 6.529063701629639,
|
| 1323 |
+
"learning_rate": 2.799061767005473e-06,
|
| 1324 |
+
"loss": 0.217,
|
| 1325 |
+
"step": 1600
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.6981450252951094,
|
| 1329 |
+
"eval_accuracy": 0.73094688221709,
|
| 1330 |
+
"eval_accuracy_label_arts, culture, entertainment and media": 0.75,
|
| 1331 |
+
"eval_accuracy_label_conflict, war and peace": 0.7446808510638298,
|
| 1332 |
+
"eval_accuracy_label_crime, law and justice": 0.838150289017341,
|
| 1333 |
+
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 1334 |
+
"eval_accuracy_label_economy, business, and finance": 0.8481012658227848,
|
| 1335 |
+
"eval_accuracy_label_environment": 0.375,
|
| 1336 |
+
"eval_accuracy_label_health": 0.8,
|
| 1337 |
+
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 1338 |
+
"eval_accuracy_label_labour": 0.5,
|
| 1339 |
+
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 1340 |
+
"eval_accuracy_label_politics": 0.539568345323741,
|
| 1341 |
+
"eval_accuracy_label_religion": 0.0,
|
| 1342 |
+
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 1343 |
+
"eval_accuracy_label_society": 0.45614035087719296,
|
| 1344 |
+
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1345 |
+
"eval_accuracy_label_weather": 1.0,
|
| 1346 |
+
"eval_f1": 0.7291730687142569,
|
| 1347 |
+
"eval_loss": 0.8666742444038391,
|
| 1348 |
+
"eval_precision": 0.7358354972874402,
|
| 1349 |
+
"eval_recall": 0.73094688221709,
|
| 1350 |
+
"eval_runtime": 6.9197,
|
| 1351 |
+
"eval_samples_per_second": 125.151,
|
| 1352 |
+
"eval_steps_per_second": 7.948,
|
| 1353 |
+
"step": 1600
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 2.715008431703204,
|
| 1357 |
+
"grad_norm": 5.216823101043701,
|
| 1358 |
+
"learning_rate": 2.642689601250977e-06,
|
| 1359 |
+
"loss": 0.2422,
|
| 1360 |
+
"step": 1610
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 2.7318718381112985,
|
| 1364 |
+
"grad_norm": 14.297146797180176,
|
| 1365 |
+
"learning_rate": 2.486317435496482e-06,
|
| 1366 |
+
"loss": 0.2569,
|
| 1367 |
+
"step": 1620
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.748735244519393,
|
| 1371 |
+
"grad_norm": 3.5069735050201416,
|
| 1372 |
+
"learning_rate": 2.329945269741986e-06,
|
| 1373 |
+
"loss": 0.284,
|
| 1374 |
+
"step": 1630
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 2.7655986509274872,
|
| 1378 |
+
"grad_norm": 10.029553413391113,
|
| 1379 |
+
"learning_rate": 2.1735731039874907e-06,
|
| 1380 |
+
"loss": 0.305,
|
| 1381 |
+
"step": 1640
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 2.782462057335582,
|
| 1385 |
+
"grad_norm": 13.829554557800293,
|
| 1386 |
+
"learning_rate": 2.0172009382329948e-06,
|
| 1387 |
+
"loss": 0.3018,
|
| 1388 |
+
"step": 1650
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 2.799325463743676,
|
| 1392 |
+
"grad_norm": 7.969460487365723,
|
| 1393 |
+
"learning_rate": 1.860828772478499e-06,
|
| 1394 |
+
"loss": 0.2687,
|
| 1395 |
+
"step": 1660
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 2.8161888701517706,
|
| 1399 |
+
"grad_norm": 4.048108100891113,
|
| 1400 |
+
"learning_rate": 1.7044566067240032e-06,
|
| 1401 |
+
"loss": 0.2845,
|
| 1402 |
+
"step": 1670
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 2.833052276559865,
|
| 1406 |
+
"grad_norm": 3.3585705757141113,
|
| 1407 |
+
"learning_rate": 1.5480844409695075e-06,
|
| 1408 |
+
"loss": 0.2792,
|
| 1409 |
+
"step": 1680
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 2.8499156829679597,
|
| 1413 |
+
"grad_norm": 3.1764752864837646,
|
| 1414 |
+
"learning_rate": 1.391712275215012e-06,
|
| 1415 |
+
"loss": 0.2244,
|
| 1416 |
+
"step": 1690
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 2.866779089376054,
|
| 1420 |
+
"grad_norm": 9.154349327087402,
|
| 1421 |
+
"learning_rate": 1.2353401094605161e-06,
|
| 1422 |
+
"loss": 0.2469,
|
| 1423 |
+
"step": 1700
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 2.8836424957841484,
|
| 1427 |
+
"grad_norm": 8.833173751831055,
|
| 1428 |
+
"learning_rate": 1.0789679437060204e-06,
|
| 1429 |
+
"loss": 0.2882,
|
| 1430 |
+
"step": 1710
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 2.9005059021922426,
|
| 1434 |
+
"grad_norm": 9.59762954711914,
|
| 1435 |
+
"learning_rate": 9.225957779515247e-07,
|
| 1436 |
+
"loss": 0.2442,
|
| 1437 |
+
"step": 1720
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 2.917369308600337,
|
| 1441 |
+
"grad_norm": 10.648268699645996,
|
| 1442 |
+
"learning_rate": 7.66223612197029e-07,
|
| 1443 |
+
"loss": 0.2421,
|
| 1444 |
+
"step": 1730
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 2.9342327150084317,
|
| 1448 |
+
"grad_norm": 3.897761821746826,
|
| 1449 |
+
"learning_rate": 6.098514464425332e-07,
|
| 1450 |
+
"loss": 0.2391,
|
| 1451 |
+
"step": 1740
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 2.9510961214165263,
|
| 1455 |
+
"grad_norm": 5.289120674133301,
|
| 1456 |
+
"learning_rate": 4.5347928068803755e-07,
|
| 1457 |
+
"loss": 0.2567,
|
| 1458 |
+
"step": 1750
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 2.9679595278246205,
|
| 1462 |
+
"grad_norm": 6.413928985595703,
|
| 1463 |
+
"learning_rate": 2.9710711493354186e-07,
|
| 1464 |
+
"loss": 0.2597,
|
| 1465 |
+
"step": 1760
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 2.984822934232715,
|
| 1469 |
+
"grad_norm": 3.9349653720855713,
|
| 1470 |
+
"learning_rate": 1.4073494917904614e-07,
|
| 1471 |
+
"loss": 0.2551,
|
| 1472 |
+
"step": 1770
|
| 1473 |
+
},
|
| 1474 |
{
|
| 1475 |
"epoch": 3.0,
|
| 1476 |
+
"step": 1779,
|
| 1477 |
+
"total_flos": 7485003163459584.0,
|
| 1478 |
+
"train_loss": 0.7494140876820411,
|
| 1479 |
+
"train_runtime": 1316.4618,
|
| 1480 |
+
"train_samples_per_second": 43.214,
|
| 1481 |
+
"train_steps_per_second": 1.351
|
| 1482 |
},
|
| 1483 |
{
|
| 1484 |
"epoch": 3.0,
|
| 1485 |
+
"eval_accuracy": 0.7286374133949192,
|
| 1486 |
"eval_accuracy_label_arts, culture, entertainment and media": 0.8333333333333334,
|
| 1487 |
+
"eval_accuracy_label_conflict, war and peace": 0.723404255319149,
|
| 1488 |
+
"eval_accuracy_label_crime, law and justice": 0.791907514450867,
|
| 1489 |
"eval_accuracy_label_disaster, accident, and emergency incident": 0.8931297709923665,
|
| 1490 |
+
"eval_accuracy_label_economy, business, and finance": 0.7974683544303798,
|
| 1491 |
+
"eval_accuracy_label_environment": 0.4375,
|
| 1492 |
+
"eval_accuracy_label_health": 0.7,
|
| 1493 |
"eval_accuracy_label_human interest": 0.3333333333333333,
|
| 1494 |
"eval_accuracy_label_labour": 0.5,
|
| 1495 |
"eval_accuracy_label_lifestyle and leisure": 0.5,
|
| 1496 |
+
"eval_accuracy_label_politics": 0.6330935251798561,
|
| 1497 |
"eval_accuracy_label_religion": 0.0,
|
| 1498 |
"eval_accuracy_label_science and technology": 0.4166666666666667,
|
| 1499 |
+
"eval_accuracy_label_society": 0.45614035087719296,
|
| 1500 |
"eval_accuracy_label_sport": 0.9615384615384616,
|
| 1501 |
"eval_accuracy_label_weather": 1.0,
|
| 1502 |
+
"eval_f1": 0.730012835488787,
|
| 1503 |
+
"eval_loss": 0.8615460395812988,
|
| 1504 |
+
"eval_precision": 0.7350865087902028,
|
| 1505 |
+
"eval_recall": 0.7286374133949192,
|
| 1506 |
+
"eval_runtime": 6.8008,
|
| 1507 |
+
"eval_samples_per_second": 127.338,
|
| 1508 |
+
"eval_steps_per_second": 8.087,
|
| 1509 |
+
"step": 1779
|
| 1510 |
}
|
| 1511 |
],
|
| 1512 |
"logging_steps": 10,
|
| 1513 |
+
"max_steps": 1779,
|
| 1514 |
"num_input_tokens_seen": 0,
|
| 1515 |
"num_train_epochs": 3,
|
| 1516 |
"save_steps": 1000,
|
| 1517 |
+
"total_flos": 7485003163459584.0,
|
| 1518 |
"train_batch_size": 16,
|
| 1519 |
"trial_name": null,
|
| 1520 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94e8cefbee1b3c3f3ed8fadf6bc22fcffca1738fa8db0ef041c55e8c9c273ff5
|
| 3 |
size 4984
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|