Training in progress, epoch 0
Browse files- README.md +70 -3
- all_results.json +11 -0
- config.json +57 -0
- model.safetensors +3 -0
- preprocessor_config.json +42 -0
- runs/May02_21-05-14_blackhorse/events.out.tfevents.1714664128.blackhorse.7281.1 +3 -0
- runs/May02_21-05-14_blackhorse/events.out.tfevents.1714683647.blackhorse.7281.2 +3 -0
- runs/May03_18-51-39_blackhorse/events.out.tfevents.1714742520.blackhorse.6448.0 +3 -0
- test_results.json +11 -0
- trainer_state.json +2069 -0
- training_args.bin +3 -0
README.md
CHANGED
|
@@ -1,3 +1,70 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-nc-4.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
base_model: MCG-NJU/videomae-base
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
metrics:
|
| 7 |
+
- accuracy
|
| 8 |
+
- f1
|
| 9 |
+
- precision
|
| 10 |
+
- recall
|
| 11 |
+
model-index:
|
| 12 |
+
- name: videomae-base-finetuned-numbers-augmented2
|
| 13 |
+
results: []
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 17 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 18 |
+
|
| 19 |
+
# videomae-base-finetuned-numbers-augmented2
|
| 20 |
+
|
| 21 |
+
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 1.9722
|
| 24 |
+
- Accuracy: 0.3269
|
| 25 |
+
- F1: 0.2716
|
| 26 |
+
- Precision: 0.3970
|
| 27 |
+
- Recall: 0.3277
|
| 28 |
+
|
| 29 |
+
## Model description
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Intended uses & limitations
|
| 34 |
+
|
| 35 |
+
More information needed
|
| 36 |
+
|
| 37 |
+
## Training and evaluation data
|
| 38 |
+
|
| 39 |
+
More information needed
|
| 40 |
+
|
| 41 |
+
## Training procedure
|
| 42 |
+
|
| 43 |
+
### Training hyperparameters
|
| 44 |
+
|
| 45 |
+
The following hyperparameters were used during training:
|
| 46 |
+
- learning_rate: 5e-06
|
| 47 |
+
- train_batch_size: 4
|
| 48 |
+
- eval_batch_size: 4
|
| 49 |
+
- seed: 42
|
| 50 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 51 |
+
- lr_scheduler_type: linear
|
| 52 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 53 |
+
- training_steps: 2816
|
| 54 |
+
|
| 55 |
+
### Training results
|
| 56 |
+
|
| 57 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|
| 58 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
|
| 59 |
+
| 2.2279 | 0.2504 | 705 | 2.2645 | 0.1824 | 0.1262 | 0.2559 | 0.1792 |
|
| 60 |
+
| 1.7024 | 1.25 | 1409 | 2.0462 | 0.3167 | 0.2828 | 0.3354 | 0.3152 |
|
| 61 |
+
| 1.3164 | 2.25 | 2113 | 1.9759 | 0.3081 | 0.2568 | 0.3022 | 0.3085 |
|
| 62 |
+
| 1.3877 | 3.2496 | 2816 | 1.9641 | 0.3373 | 0.2839 | 0.3031 | 0.3367 |
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
### Framework versions
|
| 66 |
+
|
| 67 |
+
- Transformers 4.40.0
|
| 68 |
+
- Pytorch 2.1.0+cu121
|
| 69 |
+
- Datasets 2.18.0
|
| 70 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.2496448863636362,
|
| 3 |
+
"eval_accuracy": 0.3268698060941828,
|
| 4 |
+
"eval_f1": 0.2716015567288163,
|
| 5 |
+
"eval_loss": 1.972209095954895,
|
| 6 |
+
"eval_precision": 0.3970093449590282,
|
| 7 |
+
"eval_recall": 0.32766441258833473,
|
| 8 |
+
"eval_runtime": 909.7137,
|
| 9 |
+
"eval_samples_per_second": 0.794,
|
| 10 |
+
"eval_steps_per_second": 0.199
|
| 11 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "MCG-NJU/videomae-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"VideoMAEForVideoClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"decoder_hidden_size": 384,
|
| 8 |
+
"decoder_intermediate_size": 1536,
|
| 9 |
+
"decoder_num_attention_heads": 6,
|
| 10 |
+
"decoder_num_hidden_layers": 4,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.0,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"id2label": {
|
| 15 |
+
"0": "0",
|
| 16 |
+
"1": "1",
|
| 17 |
+
"2": "2",
|
| 18 |
+
"3": "3",
|
| 19 |
+
"4": "4",
|
| 20 |
+
"5": "5",
|
| 21 |
+
"6": "6",
|
| 22 |
+
"7": "7",
|
| 23 |
+
"8": "8",
|
| 24 |
+
"9": "9",
|
| 25 |
+
"10": "10"
|
| 26 |
+
},
|
| 27 |
+
"image_size": 224,
|
| 28 |
+
"initializer_range": 0.02,
|
| 29 |
+
"intermediate_size": 3072,
|
| 30 |
+
"label2id": {
|
| 31 |
+
"0": 0,
|
| 32 |
+
"1": 1,
|
| 33 |
+
"10": 10,
|
| 34 |
+
"2": 2,
|
| 35 |
+
"3": 3,
|
| 36 |
+
"4": 4,
|
| 37 |
+
"5": 5,
|
| 38 |
+
"6": 6,
|
| 39 |
+
"7": 7,
|
| 40 |
+
"8": 8,
|
| 41 |
+
"9": 9
|
| 42 |
+
},
|
| 43 |
+
"layer_norm_eps": 1e-12,
|
| 44 |
+
"model_type": "videomae",
|
| 45 |
+
"norm_pix_loss": true,
|
| 46 |
+
"num_attention_heads": 12,
|
| 47 |
+
"num_channels": 3,
|
| 48 |
+
"num_frames": 16,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"patch_size": 16,
|
| 51 |
+
"problem_type": "single_label_classification",
|
| 52 |
+
"qkv_bias": true,
|
| 53 |
+
"torch_dtype": "float32",
|
| 54 |
+
"transformers_version": "4.40.0",
|
| 55 |
+
"tubelet_size": 2,
|
| 56 |
+
"use_mean_pooling": false
|
| 57 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21f12a8fe6956ff50249be1c6395871fcc700d4f6411a2319ee61046969c2231
|
| 3 |
+
size 344965060
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_valid_processor_keys": [
|
| 3 |
+
"videos",
|
| 4 |
+
"do_resize",
|
| 5 |
+
"size",
|
| 6 |
+
"resample",
|
| 7 |
+
"do_center_crop",
|
| 8 |
+
"crop_size",
|
| 9 |
+
"do_rescale",
|
| 10 |
+
"rescale_factor",
|
| 11 |
+
"do_normalize",
|
| 12 |
+
"image_mean",
|
| 13 |
+
"image_std",
|
| 14 |
+
"return_tensors",
|
| 15 |
+
"data_format",
|
| 16 |
+
"input_data_format"
|
| 17 |
+
],
|
| 18 |
+
"crop_size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
},
|
| 22 |
+
"do_center_crop": true,
|
| 23 |
+
"do_normalize": true,
|
| 24 |
+
"do_rescale": true,
|
| 25 |
+
"do_resize": true,
|
| 26 |
+
"image_mean": [
|
| 27 |
+
0.485,
|
| 28 |
+
0.456,
|
| 29 |
+
0.406
|
| 30 |
+
],
|
| 31 |
+
"image_processor_type": "VideoMAEImageProcessor",
|
| 32 |
+
"image_std": [
|
| 33 |
+
0.229,
|
| 34 |
+
0.224,
|
| 35 |
+
0.225
|
| 36 |
+
],
|
| 37 |
+
"resample": 2,
|
| 38 |
+
"rescale_factor": 0.00392156862745098,
|
| 39 |
+
"size": {
|
| 40 |
+
"shortest_edge": 224
|
| 41 |
+
}
|
| 42 |
+
}
|
runs/May02_21-05-14_blackhorse/events.out.tfevents.1714664128.blackhorse.7281.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0e72900212fbff68ab9a17ecee8f582719f4d013eb97ca18ab92e73566830127
|
| 3 |
+
size 66687
|
runs/May02_21-05-14_blackhorse/events.out.tfevents.1714683647.blackhorse.7281.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48097a0c18abcf4cfe6acf16350c8ad02ff0f03de80bf40e098d0c7f693136c7
|
| 3 |
+
size 1032
|
runs/May03_18-51-39_blackhorse/events.out.tfevents.1714742520.blackhorse.6448.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e3d1545bd2f786ff45a1263aa92838f522c1a0e4f268d9f1463bce5390862c4
|
| 3 |
+
size 17440
|
test_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.2496448863636362,
|
| 3 |
+
"eval_accuracy": 0.3268698060941828,
|
| 4 |
+
"eval_f1": 0.2716015567288163,
|
| 5 |
+
"eval_loss": 1.972209095954895,
|
| 6 |
+
"eval_precision": 0.3970093449590282,
|
| 7 |
+
"eval_recall": 0.32766441258833473,
|
| 8 |
+
"eval_runtime": 909.7137,
|
| 9 |
+
"eval_samples_per_second": 0.794,
|
| 10 |
+
"eval_steps_per_second": 0.199
|
| 11 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2069 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3373493975903614,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-finetuned-numbers-augmented2/checkpoint-2816",
|
| 4 |
+
"epoch": 3.2496448863636362,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2816,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0035511363636363635,
|
| 13 |
+
"grad_norm": 11.042091369628906,
|
| 14 |
+
"learning_rate": 1.7730496453900713e-07,
|
| 15 |
+
"loss": 2.6254,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.007102272727272727,
|
| 20 |
+
"grad_norm": 15.068171501159668,
|
| 21 |
+
"learning_rate": 3.5460992907801425e-07,
|
| 22 |
+
"loss": 2.6399,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01065340909090909,
|
| 27 |
+
"grad_norm": 12.725520133972168,
|
| 28 |
+
"learning_rate": 5.319148936170213e-07,
|
| 29 |
+
"loss": 2.5084,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.014204545454545454,
|
| 34 |
+
"grad_norm": 12.825174331665039,
|
| 35 |
+
"learning_rate": 7.092198581560285e-07,
|
| 36 |
+
"loss": 2.5423,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01775568181818182,
|
| 41 |
+
"grad_norm": 19.975149154663086,
|
| 42 |
+
"learning_rate": 8.865248226950356e-07,
|
| 43 |
+
"loss": 2.6042,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02130681818181818,
|
| 48 |
+
"grad_norm": 13.08024787902832,
|
| 49 |
+
"learning_rate": 1.0638297872340427e-06,
|
| 50 |
+
"loss": 2.4857,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.024857954545454544,
|
| 55 |
+
"grad_norm": 13.69168758392334,
|
| 56 |
+
"learning_rate": 1.2411347517730497e-06,
|
| 57 |
+
"loss": 2.5421,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.028409090909090908,
|
| 62 |
+
"grad_norm": 21.916378021240234,
|
| 63 |
+
"learning_rate": 1.418439716312057e-06,
|
| 64 |
+
"loss": 2.5581,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03196022727272727,
|
| 69 |
+
"grad_norm": 44.70994567871094,
|
| 70 |
+
"learning_rate": 1.595744680851064e-06,
|
| 71 |
+
"loss": 2.5354,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.03551136363636364,
|
| 76 |
+
"grad_norm": 16.238237380981445,
|
| 77 |
+
"learning_rate": 1.7730496453900712e-06,
|
| 78 |
+
"loss": 2.4371,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0390625,
|
| 83 |
+
"grad_norm": 11.415323257446289,
|
| 84 |
+
"learning_rate": 1.9503546099290782e-06,
|
| 85 |
+
"loss": 2.4871,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.04261363636363636,
|
| 90 |
+
"grad_norm": 8.98475456237793,
|
| 91 |
+
"learning_rate": 2.1276595744680853e-06,
|
| 92 |
+
"loss": 2.3905,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04616477272727273,
|
| 97 |
+
"grad_norm": 11.144891738891602,
|
| 98 |
+
"learning_rate": 2.3049645390070924e-06,
|
| 99 |
+
"loss": 2.4819,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04971590909090909,
|
| 104 |
+
"grad_norm": 15.341207504272461,
|
| 105 |
+
"learning_rate": 2.4822695035460995e-06,
|
| 106 |
+
"loss": 2.4592,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.053267045454545456,
|
| 111 |
+
"grad_norm": 12.760394096374512,
|
| 112 |
+
"learning_rate": 2.6595744680851065e-06,
|
| 113 |
+
"loss": 2.4855,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.056818181818181816,
|
| 118 |
+
"grad_norm": 13.143548965454102,
|
| 119 |
+
"learning_rate": 2.836879432624114e-06,
|
| 120 |
+
"loss": 2.4556,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.060369318181818184,
|
| 125 |
+
"grad_norm": 11.302803993225098,
|
| 126 |
+
"learning_rate": 3.0141843971631207e-06,
|
| 127 |
+
"loss": 2.4462,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.06392045454545454,
|
| 132 |
+
"grad_norm": 11.462811470031738,
|
| 133 |
+
"learning_rate": 3.191489361702128e-06,
|
| 134 |
+
"loss": 2.3793,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.06747159090909091,
|
| 139 |
+
"grad_norm": 14.533432960510254,
|
| 140 |
+
"learning_rate": 3.368794326241135e-06,
|
| 141 |
+
"loss": 2.3991,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.07102272727272728,
|
| 146 |
+
"grad_norm": 16.54037094116211,
|
| 147 |
+
"learning_rate": 3.5460992907801423e-06,
|
| 148 |
+
"loss": 2.4178,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.07457386363636363,
|
| 153 |
+
"grad_norm": 25.168344497680664,
|
| 154 |
+
"learning_rate": 3.723404255319149e-06,
|
| 155 |
+
"loss": 2.3658,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.078125,
|
| 160 |
+
"grad_norm": 14.53443431854248,
|
| 161 |
+
"learning_rate": 3.9007092198581565e-06,
|
| 162 |
+
"loss": 2.4974,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.08167613636363637,
|
| 167 |
+
"grad_norm": 13.964099884033203,
|
| 168 |
+
"learning_rate": 4.078014184397163e-06,
|
| 169 |
+
"loss": 2.3467,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.08522727272727272,
|
| 174 |
+
"grad_norm": 14.819562911987305,
|
| 175 |
+
"learning_rate": 4.255319148936171e-06,
|
| 176 |
+
"loss": 2.3952,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.08877840909090909,
|
| 181 |
+
"grad_norm": 9.89671802520752,
|
| 182 |
+
"learning_rate": 4.432624113475177e-06,
|
| 183 |
+
"loss": 2.4154,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.09232954545454546,
|
| 188 |
+
"grad_norm": 16.636743545532227,
|
| 189 |
+
"learning_rate": 4.609929078014185e-06,
|
| 190 |
+
"loss": 2.4225,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.09588068181818182,
|
| 195 |
+
"grad_norm": 12.833456993103027,
|
| 196 |
+
"learning_rate": 4.787234042553192e-06,
|
| 197 |
+
"loss": 2.4442,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.09943181818181818,
|
| 202 |
+
"grad_norm": 11.386968612670898,
|
| 203 |
+
"learning_rate": 4.964539007092199e-06,
|
| 204 |
+
"loss": 2.3345,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.10298295454545454,
|
| 209 |
+
"grad_norm": 15.00284481048584,
|
| 210 |
+
"learning_rate": 4.984214680347277e-06,
|
| 211 |
+
"loss": 2.3783,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.10653409090909091,
|
| 216 |
+
"grad_norm": 12.9210786819458,
|
| 217 |
+
"learning_rate": 4.964483030781374e-06,
|
| 218 |
+
"loss": 2.3678,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.11008522727272728,
|
| 223 |
+
"grad_norm": 46.81148910522461,
|
| 224 |
+
"learning_rate": 4.94475138121547e-06,
|
| 225 |
+
"loss": 2.3202,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.11363636363636363,
|
| 230 |
+
"grad_norm": 11.623534202575684,
|
| 231 |
+
"learning_rate": 4.925019731649567e-06,
|
| 232 |
+
"loss": 2.4116,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1171875,
|
| 237 |
+
"grad_norm": 12.292633056640625,
|
| 238 |
+
"learning_rate": 4.905288082083662e-06,
|
| 239 |
+
"loss": 2.3704,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.12073863636363637,
|
| 244 |
+
"grad_norm": 13.08430004119873,
|
| 245 |
+
"learning_rate": 4.885556432517759e-06,
|
| 246 |
+
"loss": 2.351,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.12428977272727272,
|
| 251 |
+
"grad_norm": 63.07197570800781,
|
| 252 |
+
"learning_rate": 4.865824782951855e-06,
|
| 253 |
+
"loss": 2.3944,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.1278409090909091,
|
| 258 |
+
"grad_norm": 20.782880783081055,
|
| 259 |
+
"learning_rate": 4.846093133385951e-06,
|
| 260 |
+
"loss": 2.3507,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.13139204545454544,
|
| 265 |
+
"grad_norm": 11.26919174194336,
|
| 266 |
+
"learning_rate": 4.8263614838200475e-06,
|
| 267 |
+
"loss": 2.316,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.13494318181818182,
|
| 272 |
+
"grad_norm": 13.572308540344238,
|
| 273 |
+
"learning_rate": 4.806629834254144e-06,
|
| 274 |
+
"loss": 2.3005,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.13849431818181818,
|
| 279 |
+
"grad_norm": 17.582345962524414,
|
| 280 |
+
"learning_rate": 4.78689818468824e-06,
|
| 281 |
+
"loss": 2.2751,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.14204545454545456,
|
| 286 |
+
"grad_norm": 12.714279174804688,
|
| 287 |
+
"learning_rate": 4.767166535122337e-06,
|
| 288 |
+
"loss": 2.3287,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1455965909090909,
|
| 293 |
+
"grad_norm": 11.136545181274414,
|
| 294 |
+
"learning_rate": 4.747434885556433e-06,
|
| 295 |
+
"loss": 2.3605,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.14914772727272727,
|
| 300 |
+
"grad_norm": 26.191139221191406,
|
| 301 |
+
"learning_rate": 4.727703235990529e-06,
|
| 302 |
+
"loss": 2.3669,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.15269886363636365,
|
| 307 |
+
"grad_norm": 15.782291412353516,
|
| 308 |
+
"learning_rate": 4.707971586424626e-06,
|
| 309 |
+
"loss": 2.3671,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.15625,
|
| 314 |
+
"grad_norm": 13.54359245300293,
|
| 315 |
+
"learning_rate": 4.6882399368587215e-06,
|
| 316 |
+
"loss": 2.3991,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.15980113636363635,
|
| 321 |
+
"grad_norm": 12.135380744934082,
|
| 322 |
+
"learning_rate": 4.668508287292818e-06,
|
| 323 |
+
"loss": 2.3314,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.16335227272727273,
|
| 328 |
+
"grad_norm": 12.856097221374512,
|
| 329 |
+
"learning_rate": 4.6487766377269145e-06,
|
| 330 |
+
"loss": 2.2192,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1669034090909091,
|
| 335 |
+
"grad_norm": 15.03661823272705,
|
| 336 |
+
"learning_rate": 4.62904498816101e-06,
|
| 337 |
+
"loss": 2.3,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.17045454545454544,
|
| 342 |
+
"grad_norm": 11.673720359802246,
|
| 343 |
+
"learning_rate": 4.609313338595107e-06,
|
| 344 |
+
"loss": 2.2383,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.17400568181818182,
|
| 349 |
+
"grad_norm": 14.064128875732422,
|
| 350 |
+
"learning_rate": 4.589581689029203e-06,
|
| 351 |
+
"loss": 2.2481,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.17755681818181818,
|
| 356 |
+
"grad_norm": 15.709892272949219,
|
| 357 |
+
"learning_rate": 4.5698500394633e-06,
|
| 358 |
+
"loss": 2.3622,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.18110795454545456,
|
| 363 |
+
"grad_norm": 14.75145149230957,
|
| 364 |
+
"learning_rate": 4.550118389897396e-06,
|
| 365 |
+
"loss": 2.2797,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.1846590909090909,
|
| 370 |
+
"grad_norm": 12.432939529418945,
|
| 371 |
+
"learning_rate": 4.530386740331492e-06,
|
| 372 |
+
"loss": 2.2493,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.18821022727272727,
|
| 377 |
+
"grad_norm": 13.656457901000977,
|
| 378 |
+
"learning_rate": 4.510655090765588e-06,
|
| 379 |
+
"loss": 2.2605,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.19176136363636365,
|
| 384 |
+
"grad_norm": 16.014392852783203,
|
| 385 |
+
"learning_rate": 4.490923441199685e-06,
|
| 386 |
+
"loss": 2.2669,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.1953125,
|
| 391 |
+
"grad_norm": 17.114728927612305,
|
| 392 |
+
"learning_rate": 4.4711917916337814e-06,
|
| 393 |
+
"loss": 2.3496,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.19886363636363635,
|
| 398 |
+
"grad_norm": 13.276822090148926,
|
| 399 |
+
"learning_rate": 4.451460142067877e-06,
|
| 400 |
+
"loss": 2.1851,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.20241477272727273,
|
| 405 |
+
"grad_norm": 14.958836555480957,
|
| 406 |
+
"learning_rate": 4.431728492501974e-06,
|
| 407 |
+
"loss": 2.2624,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.2059659090909091,
|
| 412 |
+
"grad_norm": 17.03455924987793,
|
| 413 |
+
"learning_rate": 4.41199684293607e-06,
|
| 414 |
+
"loss": 2.1772,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.20951704545454544,
|
| 419 |
+
"grad_norm": 14.998839378356934,
|
| 420 |
+
"learning_rate": 4.392265193370166e-06,
|
| 421 |
+
"loss": 2.2558,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.21306818181818182,
|
| 426 |
+
"grad_norm": 16.826061248779297,
|
| 427 |
+
"learning_rate": 4.372533543804262e-06,
|
| 428 |
+
"loss": 2.2345,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.21661931818181818,
|
| 433 |
+
"grad_norm": 12.38198471069336,
|
| 434 |
+
"learning_rate": 4.352801894238359e-06,
|
| 435 |
+
"loss": 2.2144,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.22017045454545456,
|
| 440 |
+
"grad_norm": 12.347530364990234,
|
| 441 |
+
"learning_rate": 4.3330702446724545e-06,
|
| 442 |
+
"loss": 2.1875,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2237215909090909,
|
| 447 |
+
"grad_norm": 17.95762825012207,
|
| 448 |
+
"learning_rate": 4.313338595106552e-06,
|
| 449 |
+
"loss": 2.1382,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.22727272727272727,
|
| 454 |
+
"grad_norm": 20.329545974731445,
|
| 455 |
+
"learning_rate": 4.2936069455406475e-06,
|
| 456 |
+
"loss": 2.1519,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.23082386363636365,
|
| 461 |
+
"grad_norm": 14.157989501953125,
|
| 462 |
+
"learning_rate": 4.273875295974744e-06,
|
| 463 |
+
"loss": 2.1994,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.234375,
|
| 468 |
+
"grad_norm": 17.65989112854004,
|
| 469 |
+
"learning_rate": 4.2541436464088406e-06,
|
| 470 |
+
"loss": 2.2021,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.23792613636363635,
|
| 475 |
+
"grad_norm": 19.491857528686523,
|
| 476 |
+
"learning_rate": 4.234411996842936e-06,
|
| 477 |
+
"loss": 2.25,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.24147727272727273,
|
| 482 |
+
"grad_norm": 14.291483879089355,
|
| 483 |
+
"learning_rate": 4.214680347277033e-06,
|
| 484 |
+
"loss": 2.1732,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.2450284090909091,
|
| 489 |
+
"grad_norm": 15.794211387634277,
|
| 490 |
+
"learning_rate": 4.194948697711129e-06,
|
| 491 |
+
"loss": 2.2288,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.24857954545454544,
|
| 496 |
+
"grad_norm": 13.537517547607422,
|
| 497 |
+
"learning_rate": 4.175217048145225e-06,
|
| 498 |
+
"loss": 2.2279,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.25035511363636365,
|
| 503 |
+
"eval_accuracy": 0.18244406196213425,
|
| 504 |
+
"eval_f1": 0.1261780590834414,
|
| 505 |
+
"eval_loss": 2.2644505500793457,
|
| 506 |
+
"eval_precision": 0.25592730406989506,
|
| 507 |
+
"eval_recall": 0.17920384277210025,
|
| 508 |
+
"eval_runtime": 691.0928,
|
| 509 |
+
"eval_samples_per_second": 0.841,
|
| 510 |
+
"eval_steps_per_second": 0.211,
|
| 511 |
+
"step": 705
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 1.0017755681818181,
|
| 515 |
+
"grad_norm": 16.53750228881836,
|
| 516 |
+
"learning_rate": 4.1554853985793214e-06,
|
| 517 |
+
"loss": 2.2194,
|
| 518 |
+
"step": 710
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.0053267045454546,
|
| 522 |
+
"grad_norm": 14.238625526428223,
|
| 523 |
+
"learning_rate": 4.135753749013418e-06,
|
| 524 |
+
"loss": 2.1206,
|
| 525 |
+
"step": 720
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.0088778409090908,
|
| 529 |
+
"grad_norm": 19.52684783935547,
|
| 530 |
+
"learning_rate": 4.1160220994475145e-06,
|
| 531 |
+
"loss": 2.0227,
|
| 532 |
+
"step": 730
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 1.0124289772727273,
|
| 536 |
+
"grad_norm": 21.326141357421875,
|
| 537 |
+
"learning_rate": 4.09629044988161e-06,
|
| 538 |
+
"loss": 2.2306,
|
| 539 |
+
"step": 740
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.0159801136363635,
|
| 543 |
+
"grad_norm": 18.599313735961914,
|
| 544 |
+
"learning_rate": 4.076558800315707e-06,
|
| 545 |
+
"loss": 2.0473,
|
| 546 |
+
"step": 750
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 1.01953125,
|
| 550 |
+
"grad_norm": 21.073596954345703,
|
| 551 |
+
"learning_rate": 4.056827150749803e-06,
|
| 552 |
+
"loss": 2.0669,
|
| 553 |
+
"step": 760
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.0230823863636365,
|
| 557 |
+
"grad_norm": 13.940530776977539,
|
| 558 |
+
"learning_rate": 4.037095501183899e-06,
|
| 559 |
+
"loss": 2.0881,
|
| 560 |
+
"step": 770
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.0266335227272727,
|
| 564 |
+
"grad_norm": 23.56789207458496,
|
| 565 |
+
"learning_rate": 4.017363851617996e-06,
|
| 566 |
+
"loss": 2.0495,
|
| 567 |
+
"step": 780
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.0301846590909092,
|
| 571 |
+
"grad_norm": 24.358179092407227,
|
| 572 |
+
"learning_rate": 3.997632202052092e-06,
|
| 573 |
+
"loss": 2.1131,
|
| 574 |
+
"step": 790
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 1.0337357954545454,
|
| 578 |
+
"grad_norm": 23.59551239013672,
|
| 579 |
+
"learning_rate": 3.977900552486188e-06,
|
| 580 |
+
"loss": 2.2194,
|
| 581 |
+
"step": 800
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.0372869318181819,
|
| 585 |
+
"grad_norm": 18.300479888916016,
|
| 586 |
+
"learning_rate": 3.958168902920285e-06,
|
| 587 |
+
"loss": 2.0495,
|
| 588 |
+
"step": 810
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 1.0408380681818181,
|
| 592 |
+
"grad_norm": 14.824146270751953,
|
| 593 |
+
"learning_rate": 3.9384372533543806e-06,
|
| 594 |
+
"loss": 2.0368,
|
| 595 |
+
"step": 820
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.0443892045454546,
|
| 599 |
+
"grad_norm": 30.449533462524414,
|
| 600 |
+
"learning_rate": 3.918705603788477e-06,
|
| 601 |
+
"loss": 1.9466,
|
| 602 |
+
"step": 830
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.0479403409090908,
|
| 606 |
+
"grad_norm": 30.030054092407227,
|
| 607 |
+
"learning_rate": 3.898973954222574e-06,
|
| 608 |
+
"loss": 2.037,
|
| 609 |
+
"step": 840
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.0514914772727273,
|
| 613 |
+
"grad_norm": 16.496719360351562,
|
| 614 |
+
"learning_rate": 3.879242304656669e-06,
|
| 615 |
+
"loss": 2.0618,
|
| 616 |
+
"step": 850
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 1.0550426136363635,
|
| 620 |
+
"grad_norm": 30.52495574951172,
|
| 621 |
+
"learning_rate": 3.859510655090766e-06,
|
| 622 |
+
"loss": 1.9784,
|
| 623 |
+
"step": 860
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 1.05859375,
|
| 627 |
+
"grad_norm": 16.27585792541504,
|
| 628 |
+
"learning_rate": 3.839779005524862e-06,
|
| 629 |
+
"loss": 2.0537,
|
| 630 |
+
"step": 870
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 1.0621448863636365,
|
| 634 |
+
"grad_norm": 21.442113876342773,
|
| 635 |
+
"learning_rate": 3.820047355958959e-06,
|
| 636 |
+
"loss": 2.0197,
|
| 637 |
+
"step": 880
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.0656960227272727,
|
| 641 |
+
"grad_norm": 23.343584060668945,
|
| 642 |
+
"learning_rate": 3.8003157063930545e-06,
|
| 643 |
+
"loss": 1.9195,
|
| 644 |
+
"step": 890
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.0692471590909092,
|
| 648 |
+
"grad_norm": 12.5782470703125,
|
| 649 |
+
"learning_rate": 3.7805840568271514e-06,
|
| 650 |
+
"loss": 1.9142,
|
| 651 |
+
"step": 900
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 1.0727982954545454,
|
| 655 |
+
"grad_norm": 14.708935737609863,
|
| 656 |
+
"learning_rate": 3.7608524072612475e-06,
|
| 657 |
+
"loss": 1.958,
|
| 658 |
+
"step": 910
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 1.0763494318181819,
|
| 662 |
+
"grad_norm": 14.57020378112793,
|
| 663 |
+
"learning_rate": 3.7411207576953436e-06,
|
| 664 |
+
"loss": 1.9806,
|
| 665 |
+
"step": 920
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.0799005681818181,
|
| 669 |
+
"grad_norm": 12.808838844299316,
|
| 670 |
+
"learning_rate": 3.72138910812944e-06,
|
| 671 |
+
"loss": 2.0507,
|
| 672 |
+
"step": 930
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.0834517045454546,
|
| 676 |
+
"grad_norm": 20.186933517456055,
|
| 677 |
+
"learning_rate": 3.7016574585635362e-06,
|
| 678 |
+
"loss": 2.1607,
|
| 679 |
+
"step": 940
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.0870028409090908,
|
| 683 |
+
"grad_norm": 17.26938819885254,
|
| 684 |
+
"learning_rate": 3.6819258089976323e-06,
|
| 685 |
+
"loss": 1.9388,
|
| 686 |
+
"step": 950
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.0905539772727273,
|
| 690 |
+
"grad_norm": 18.56201934814453,
|
| 691 |
+
"learning_rate": 3.662194159431729e-06,
|
| 692 |
+
"loss": 1.984,
|
| 693 |
+
"step": 960
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 1.0941051136363635,
|
| 697 |
+
"grad_norm": 13.663825035095215,
|
| 698 |
+
"learning_rate": 3.642462509865825e-06,
|
| 699 |
+
"loss": 1.9401,
|
| 700 |
+
"step": 970
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 1.09765625,
|
| 704 |
+
"grad_norm": 18.881681442260742,
|
| 705 |
+
"learning_rate": 3.622730860299921e-06,
|
| 706 |
+
"loss": 1.914,
|
| 707 |
+
"step": 980
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 1.1012073863636365,
|
| 711 |
+
"grad_norm": 25.957923889160156,
|
| 712 |
+
"learning_rate": 3.602999210734018e-06,
|
| 713 |
+
"loss": 1.9745,
|
| 714 |
+
"step": 990
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 1.1047585227272727,
|
| 718 |
+
"grad_norm": 33.24240493774414,
|
| 719 |
+
"learning_rate": 3.583267561168114e-06,
|
| 720 |
+
"loss": 1.9489,
|
| 721 |
+
"step": 1000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.1083096590909092,
|
| 725 |
+
"grad_norm": 16.282686233520508,
|
| 726 |
+
"learning_rate": 3.56353591160221e-06,
|
| 727 |
+
"loss": 1.9327,
|
| 728 |
+
"step": 1010
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.1118607954545454,
|
| 732 |
+
"grad_norm": 27.778982162475586,
|
| 733 |
+
"learning_rate": 3.5438042620363066e-06,
|
| 734 |
+
"loss": 1.801,
|
| 735 |
+
"step": 1020
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 1.1154119318181819,
|
| 739 |
+
"grad_norm": 20.463930130004883,
|
| 740 |
+
"learning_rate": 3.5240726124704027e-06,
|
| 741 |
+
"loss": 1.9934,
|
| 742 |
+
"step": 1030
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 1.1189630681818181,
|
| 746 |
+
"grad_norm": 26.343746185302734,
|
| 747 |
+
"learning_rate": 3.504340962904499e-06,
|
| 748 |
+
"loss": 1.915,
|
| 749 |
+
"step": 1040
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 1.1225142045454546,
|
| 753 |
+
"grad_norm": 21.86429214477539,
|
| 754 |
+
"learning_rate": 3.4846093133385953e-06,
|
| 755 |
+
"loss": 1.9066,
|
| 756 |
+
"step": 1050
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 1.1260653409090908,
|
| 760 |
+
"grad_norm": 21.5794620513916,
|
| 761 |
+
"learning_rate": 3.4648776637726914e-06,
|
| 762 |
+
"loss": 1.66,
|
| 763 |
+
"step": 1060
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.1296164772727273,
|
| 767 |
+
"grad_norm": 14.89384937286377,
|
| 768 |
+
"learning_rate": 3.4451460142067875e-06,
|
| 769 |
+
"loss": 1.7305,
|
| 770 |
+
"step": 1070
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.1331676136363638,
|
| 774 |
+
"grad_norm": 19.86824607849121,
|
| 775 |
+
"learning_rate": 3.4254143646408845e-06,
|
| 776 |
+
"loss": 2.0088,
|
| 777 |
+
"step": 1080
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 1.13671875,
|
| 781 |
+
"grad_norm": 19.656740188598633,
|
| 782 |
+
"learning_rate": 3.4056827150749806e-06,
|
| 783 |
+
"loss": 2.007,
|
| 784 |
+
"step": 1090
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 1.1402698863636362,
|
| 788 |
+
"grad_norm": 19.889986038208008,
|
| 789 |
+
"learning_rate": 3.3859510655090767e-06,
|
| 790 |
+
"loss": 1.9091,
|
| 791 |
+
"step": 1100
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 1.1438210227272727,
|
| 795 |
+
"grad_norm": 34.58203887939453,
|
| 796 |
+
"learning_rate": 3.366219415943173e-06,
|
| 797 |
+
"loss": 1.982,
|
| 798 |
+
"step": 1110
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 1.1473721590909092,
|
| 802 |
+
"grad_norm": 14.81122875213623,
|
| 803 |
+
"learning_rate": 3.3464877663772693e-06,
|
| 804 |
+
"loss": 1.916,
|
| 805 |
+
"step": 1120
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.1509232954545454,
|
| 809 |
+
"grad_norm": 25.33427619934082,
|
| 810 |
+
"learning_rate": 3.3267561168113654e-06,
|
| 811 |
+
"loss": 1.9497,
|
| 812 |
+
"step": 1130
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.1544744318181819,
|
| 816 |
+
"grad_norm": 18.72292709350586,
|
| 817 |
+
"learning_rate": 3.3070244672454623e-06,
|
| 818 |
+
"loss": 1.7237,
|
| 819 |
+
"step": 1140
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 1.1580255681818181,
|
| 823 |
+
"grad_norm": 20.77664566040039,
|
| 824 |
+
"learning_rate": 3.2872928176795584e-06,
|
| 825 |
+
"loss": 1.8997,
|
| 826 |
+
"step": 1150
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 1.1615767045454546,
|
| 830 |
+
"grad_norm": 23.85498809814453,
|
| 831 |
+
"learning_rate": 3.2675611681136545e-06,
|
| 832 |
+
"loss": 1.8602,
|
| 833 |
+
"step": 1160
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 1.1651278409090908,
|
| 837 |
+
"grad_norm": 22.836532592773438,
|
| 838 |
+
"learning_rate": 3.247829518547751e-06,
|
| 839 |
+
"loss": 1.8654,
|
| 840 |
+
"step": 1170
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 1.1686789772727273,
|
| 844 |
+
"grad_norm": 35.54890441894531,
|
| 845 |
+
"learning_rate": 3.228097868981847e-06,
|
| 846 |
+
"loss": 1.6553,
|
| 847 |
+
"step": 1180
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.1722301136363638,
|
| 851 |
+
"grad_norm": 19.193073272705078,
|
| 852 |
+
"learning_rate": 3.208366219415943e-06,
|
| 853 |
+
"loss": 1.8181,
|
| 854 |
+
"step": 1190
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.17578125,
|
| 858 |
+
"grad_norm": 18.36795425415039,
|
| 859 |
+
"learning_rate": 3.1886345698500397e-06,
|
| 860 |
+
"loss": 1.796,
|
| 861 |
+
"step": 1200
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 1.1793323863636362,
|
| 865 |
+
"grad_norm": 28.257104873657227,
|
| 866 |
+
"learning_rate": 3.1689029202841358e-06,
|
| 867 |
+
"loss": 1.9526,
|
| 868 |
+
"step": 1210
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.1828835227272727,
|
| 872 |
+
"grad_norm": 21.073034286499023,
|
| 873 |
+
"learning_rate": 3.149171270718232e-06,
|
| 874 |
+
"loss": 1.8972,
|
| 875 |
+
"step": 1220
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 1.1864346590909092,
|
| 879 |
+
"grad_norm": 30.784828186035156,
|
| 880 |
+
"learning_rate": 3.129439621152329e-06,
|
| 881 |
+
"loss": 1.8178,
|
| 882 |
+
"step": 1230
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 1.1899857954545454,
|
| 886 |
+
"grad_norm": 33.99855041503906,
|
| 887 |
+
"learning_rate": 3.109707971586425e-06,
|
| 888 |
+
"loss": 1.7748,
|
| 889 |
+
"step": 1240
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.1935369318181819,
|
| 893 |
+
"grad_norm": 19.716716766357422,
|
| 894 |
+
"learning_rate": 3.089976322020521e-06,
|
| 895 |
+
"loss": 1.7614,
|
| 896 |
+
"step": 1250
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.1970880681818181,
|
| 900 |
+
"grad_norm": 17.041454315185547,
|
| 901 |
+
"learning_rate": 3.0702446724546175e-06,
|
| 902 |
+
"loss": 1.7971,
|
| 903 |
+
"step": 1260
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 1.2006392045454546,
|
| 907 |
+
"grad_norm": 32.49284744262695,
|
| 908 |
+
"learning_rate": 3.0505130228887136e-06,
|
| 909 |
+
"loss": 1.8446,
|
| 910 |
+
"step": 1270
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.2041903409090908,
|
| 914 |
+
"grad_norm": 18.60474395751953,
|
| 915 |
+
"learning_rate": 3.03078137332281e-06,
|
| 916 |
+
"loss": 1.6414,
|
| 917 |
+
"step": 1280
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 1.2077414772727273,
|
| 921 |
+
"grad_norm": 16.46165657043457,
|
| 922 |
+
"learning_rate": 3.0110497237569062e-06,
|
| 923 |
+
"loss": 1.9618,
|
| 924 |
+
"step": 1290
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.2112926136363638,
|
| 928 |
+
"grad_norm": 23.773658752441406,
|
| 929 |
+
"learning_rate": 2.9913180741910023e-06,
|
| 930 |
+
"loss": 1.69,
|
| 931 |
+
"step": 1300
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.21484375,
|
| 935 |
+
"grad_norm": 26.765705108642578,
|
| 936 |
+
"learning_rate": 2.9715864246250992e-06,
|
| 937 |
+
"loss": 1.8167,
|
| 938 |
+
"step": 1310
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.2183948863636362,
|
| 942 |
+
"grad_norm": 28.61939811706543,
|
| 943 |
+
"learning_rate": 2.9518547750591953e-06,
|
| 944 |
+
"loss": 1.7087,
|
| 945 |
+
"step": 1320
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 1.2219460227272727,
|
| 949 |
+
"grad_norm": 12.1895112991333,
|
| 950 |
+
"learning_rate": 2.9321231254932914e-06,
|
| 951 |
+
"loss": 1.6608,
|
| 952 |
+
"step": 1330
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 1.2254971590909092,
|
| 956 |
+
"grad_norm": 36.88252258300781,
|
| 957 |
+
"learning_rate": 2.912391475927388e-06,
|
| 958 |
+
"loss": 1.8907,
|
| 959 |
+
"step": 1340
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.2290482954545454,
|
| 963 |
+
"grad_norm": 12.692719459533691,
|
| 964 |
+
"learning_rate": 2.892659826361484e-06,
|
| 965 |
+
"loss": 1.7991,
|
| 966 |
+
"step": 1350
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 1.2325994318181819,
|
| 970 |
+
"grad_norm": 17.56617546081543,
|
| 971 |
+
"learning_rate": 2.87292817679558e-06,
|
| 972 |
+
"loss": 1.7126,
|
| 973 |
+
"step": 1360
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.2361505681818181,
|
| 977 |
+
"grad_norm": 26.137487411499023,
|
| 978 |
+
"learning_rate": 2.853196527229677e-06,
|
| 979 |
+
"loss": 1.6474,
|
| 980 |
+
"step": 1370
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 1.2397017045454546,
|
| 984 |
+
"grad_norm": 18.798303604125977,
|
| 985 |
+
"learning_rate": 2.833464877663773e-06,
|
| 986 |
+
"loss": 1.5864,
|
| 987 |
+
"step": 1380
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.2432528409090908,
|
| 991 |
+
"grad_norm": 29.38694953918457,
|
| 992 |
+
"learning_rate": 2.8137332280978693e-06,
|
| 993 |
+
"loss": 1.8485,
|
| 994 |
+
"step": 1390
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.2468039772727273,
|
| 998 |
+
"grad_norm": 25.83952522277832,
|
| 999 |
+
"learning_rate": 2.7940015785319658e-06,
|
| 1000 |
+
"loss": 1.7024,
|
| 1001 |
+
"step": 1400
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.25,
|
| 1005 |
+
"eval_accuracy": 0.31669535283993117,
|
| 1006 |
+
"eval_f1": 0.2827610289664265,
|
| 1007 |
+
"eval_loss": 2.046189308166504,
|
| 1008 |
+
"eval_precision": 0.3353755833663839,
|
| 1009 |
+
"eval_recall": 0.31518980464596447,
|
| 1010 |
+
"eval_runtime": 747.7036,
|
| 1011 |
+
"eval_samples_per_second": 0.777,
|
| 1012 |
+
"eval_steps_per_second": 0.195,
|
| 1013 |
+
"step": 1409
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 2.0003551136363638,
|
| 1017 |
+
"grad_norm": 20.87928009033203,
|
| 1018 |
+
"learning_rate": 2.774269928966062e-06,
|
| 1019 |
+
"loss": 1.6345,
|
| 1020 |
+
"step": 1410
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 2.00390625,
|
| 1024 |
+
"grad_norm": 28.964290618896484,
|
| 1025 |
+
"learning_rate": 2.754538279400158e-06,
|
| 1026 |
+
"loss": 1.7143,
|
| 1027 |
+
"step": 1420
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 2.0074573863636362,
|
| 1031 |
+
"grad_norm": 17.409208297729492,
|
| 1032 |
+
"learning_rate": 2.7348066298342545e-06,
|
| 1033 |
+
"loss": 1.6937,
|
| 1034 |
+
"step": 1430
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.011008522727273,
|
| 1038 |
+
"grad_norm": 24.774322509765625,
|
| 1039 |
+
"learning_rate": 2.7150749802683506e-06,
|
| 1040 |
+
"loss": 1.6225,
|
| 1041 |
+
"step": 1440
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 2.014559659090909,
|
| 1045 |
+
"grad_norm": 33.20758819580078,
|
| 1046 |
+
"learning_rate": 2.6953433307024466e-06,
|
| 1047 |
+
"loss": 1.6123,
|
| 1048 |
+
"step": 1450
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 2.0181107954545454,
|
| 1052 |
+
"grad_norm": 24.023727416992188,
|
| 1053 |
+
"learning_rate": 2.6756116811365436e-06,
|
| 1054 |
+
"loss": 1.6963,
|
| 1055 |
+
"step": 1460
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 2.0216619318181817,
|
| 1059 |
+
"grad_norm": 26.691741943359375,
|
| 1060 |
+
"learning_rate": 2.6558800315706397e-06,
|
| 1061 |
+
"loss": 1.7662,
|
| 1062 |
+
"step": 1470
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 2.0252130681818183,
|
| 1066 |
+
"grad_norm": 31.09050178527832,
|
| 1067 |
+
"learning_rate": 2.6361483820047358e-06,
|
| 1068 |
+
"loss": 1.6466,
|
| 1069 |
+
"step": 1480
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 2.0287642045454546,
|
| 1073 |
+
"grad_norm": 34.407222747802734,
|
| 1074 |
+
"learning_rate": 2.6164167324388323e-06,
|
| 1075 |
+
"loss": 1.577,
|
| 1076 |
+
"step": 1490
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.032315340909091,
|
| 1080 |
+
"grad_norm": 21.151485443115234,
|
| 1081 |
+
"learning_rate": 2.5966850828729284e-06,
|
| 1082 |
+
"loss": 1.4755,
|
| 1083 |
+
"step": 1500
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 2.035866477272727,
|
| 1087 |
+
"grad_norm": 33.150962829589844,
|
| 1088 |
+
"learning_rate": 2.5769534333070245e-06,
|
| 1089 |
+
"loss": 1.7999,
|
| 1090 |
+
"step": 1510
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 2.0394176136363638,
|
| 1094 |
+
"grad_norm": 37.68964385986328,
|
| 1095 |
+
"learning_rate": 2.557221783741121e-06,
|
| 1096 |
+
"loss": 1.7355,
|
| 1097 |
+
"step": 1520
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.04296875,
|
| 1101 |
+
"grad_norm": 21.464839935302734,
|
| 1102 |
+
"learning_rate": 2.537490134175217e-06,
|
| 1103 |
+
"loss": 1.5964,
|
| 1104 |
+
"step": 1530
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 2.0465198863636362,
|
| 1108 |
+
"grad_norm": 33.7103157043457,
|
| 1109 |
+
"learning_rate": 2.517758484609313e-06,
|
| 1110 |
+
"loss": 1.6579,
|
| 1111 |
+
"step": 1540
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 2.050071022727273,
|
| 1115 |
+
"grad_norm": 34.679019927978516,
|
| 1116 |
+
"learning_rate": 2.4980268350434097e-06,
|
| 1117 |
+
"loss": 1.5309,
|
| 1118 |
+
"step": 1550
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.053622159090909,
|
| 1122 |
+
"grad_norm": 25.11505889892578,
|
| 1123 |
+
"learning_rate": 2.478295185477506e-06,
|
| 1124 |
+
"loss": 1.605,
|
| 1125 |
+
"step": 1560
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 2.0571732954545454,
|
| 1129 |
+
"grad_norm": 14.84094524383545,
|
| 1130 |
+
"learning_rate": 2.4585635359116027e-06,
|
| 1131 |
+
"loss": 1.5956,
|
| 1132 |
+
"step": 1570
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 2.0607244318181817,
|
| 1136 |
+
"grad_norm": 22.252281188964844,
|
| 1137 |
+
"learning_rate": 2.438831886345699e-06,
|
| 1138 |
+
"loss": 1.5511,
|
| 1139 |
+
"step": 1580
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.0642755681818183,
|
| 1143 |
+
"grad_norm": 50.657798767089844,
|
| 1144 |
+
"learning_rate": 2.419100236779795e-06,
|
| 1145 |
+
"loss": 1.5066,
|
| 1146 |
+
"step": 1590
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 2.0678267045454546,
|
| 1150 |
+
"grad_norm": 32.89131546020508,
|
| 1151 |
+
"learning_rate": 2.3993685872138914e-06,
|
| 1152 |
+
"loss": 1.6128,
|
| 1153 |
+
"step": 1600
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 2.071377840909091,
|
| 1157 |
+
"grad_norm": 32.423648834228516,
|
| 1158 |
+
"learning_rate": 2.3796369376479875e-06,
|
| 1159 |
+
"loss": 1.6446,
|
| 1160 |
+
"step": 1610
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.074928977272727,
|
| 1164 |
+
"grad_norm": 30.093395233154297,
|
| 1165 |
+
"learning_rate": 2.359905288082084e-06,
|
| 1166 |
+
"loss": 1.5277,
|
| 1167 |
+
"step": 1620
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 2.0784801136363638,
|
| 1171 |
+
"grad_norm": 40.93949508666992,
|
| 1172 |
+
"learning_rate": 2.34017363851618e-06,
|
| 1173 |
+
"loss": 1.5847,
|
| 1174 |
+
"step": 1630
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 2.08203125,
|
| 1178 |
+
"grad_norm": 21.456939697265625,
|
| 1179 |
+
"learning_rate": 2.320441988950276e-06,
|
| 1180 |
+
"loss": 1.5051,
|
| 1181 |
+
"step": 1640
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 2.0855823863636362,
|
| 1185 |
+
"grad_norm": 13.460980415344238,
|
| 1186 |
+
"learning_rate": 2.3007103393843727e-06,
|
| 1187 |
+
"loss": 1.6683,
|
| 1188 |
+
"step": 1650
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 2.089133522727273,
|
| 1192 |
+
"grad_norm": 25.33333969116211,
|
| 1193 |
+
"learning_rate": 2.2809786898184692e-06,
|
| 1194 |
+
"loss": 1.5145,
|
| 1195 |
+
"step": 1660
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 2.092684659090909,
|
| 1199 |
+
"grad_norm": 35.350223541259766,
|
| 1200 |
+
"learning_rate": 2.2612470402525653e-06,
|
| 1201 |
+
"loss": 1.5161,
|
| 1202 |
+
"step": 1670
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.0962357954545454,
|
| 1206 |
+
"grad_norm": 11.545449256896973,
|
| 1207 |
+
"learning_rate": 2.2415153906866614e-06,
|
| 1208 |
+
"loss": 1.5556,
|
| 1209 |
+
"step": 1680
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 2.0997869318181817,
|
| 1213 |
+
"grad_norm": 24.93936538696289,
|
| 1214 |
+
"learning_rate": 2.221783741120758e-06,
|
| 1215 |
+
"loss": 1.6311,
|
| 1216 |
+
"step": 1690
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 2.1033380681818183,
|
| 1220 |
+
"grad_norm": 17.35971450805664,
|
| 1221 |
+
"learning_rate": 2.202052091554854e-06,
|
| 1222 |
+
"loss": 1.6324,
|
| 1223 |
+
"step": 1700
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 2.1068892045454546,
|
| 1227 |
+
"grad_norm": 30.070161819458008,
|
| 1228 |
+
"learning_rate": 2.1823204419889505e-06,
|
| 1229 |
+
"loss": 1.6366,
|
| 1230 |
+
"step": 1710
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 2.110440340909091,
|
| 1234 |
+
"grad_norm": 30.212112426757812,
|
| 1235 |
+
"learning_rate": 2.1625887924230466e-06,
|
| 1236 |
+
"loss": 1.5943,
|
| 1237 |
+
"step": 1720
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 2.113991477272727,
|
| 1241 |
+
"grad_norm": 19.462881088256836,
|
| 1242 |
+
"learning_rate": 2.1428571428571427e-06,
|
| 1243 |
+
"loss": 1.5983,
|
| 1244 |
+
"step": 1730
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.1175426136363638,
|
| 1248 |
+
"grad_norm": 17.48008155822754,
|
| 1249 |
+
"learning_rate": 2.1231254932912392e-06,
|
| 1250 |
+
"loss": 1.5725,
|
| 1251 |
+
"step": 1740
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 2.12109375,
|
| 1255 |
+
"grad_norm": 18.313573837280273,
|
| 1256 |
+
"learning_rate": 2.1033938437253358e-06,
|
| 1257 |
+
"loss": 1.6039,
|
| 1258 |
+
"step": 1750
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 2.1246448863636362,
|
| 1262 |
+
"grad_norm": 12.866426467895508,
|
| 1263 |
+
"learning_rate": 2.083662194159432e-06,
|
| 1264 |
+
"loss": 1.4405,
|
| 1265 |
+
"step": 1760
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 2.128196022727273,
|
| 1269 |
+
"grad_norm": 25.331958770751953,
|
| 1270 |
+
"learning_rate": 2.063930544593528e-06,
|
| 1271 |
+
"loss": 1.728,
|
| 1272 |
+
"step": 1770
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 2.131747159090909,
|
| 1276 |
+
"grad_norm": 31.10738182067871,
|
| 1277 |
+
"learning_rate": 2.0441988950276245e-06,
|
| 1278 |
+
"loss": 1.6007,
|
| 1279 |
+
"step": 1780
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 2.1352982954545454,
|
| 1283 |
+
"grad_norm": 27.759248733520508,
|
| 1284 |
+
"learning_rate": 2.0244672454617206e-06,
|
| 1285 |
+
"loss": 1.5672,
|
| 1286 |
+
"step": 1790
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.1388494318181817,
|
| 1290 |
+
"grad_norm": 16.519384384155273,
|
| 1291 |
+
"learning_rate": 2.004735595895817e-06,
|
| 1292 |
+
"loss": 1.3412,
|
| 1293 |
+
"step": 1800
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 2.1424005681818183,
|
| 1297 |
+
"grad_norm": 19.062610626220703,
|
| 1298 |
+
"learning_rate": 1.9850039463299136e-06,
|
| 1299 |
+
"loss": 1.6176,
|
| 1300 |
+
"step": 1810
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 2.1459517045454546,
|
| 1304 |
+
"grad_norm": 48.56722640991211,
|
| 1305 |
+
"learning_rate": 1.9652722967640097e-06,
|
| 1306 |
+
"loss": 1.6479,
|
| 1307 |
+
"step": 1820
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 2.149502840909091,
|
| 1311 |
+
"grad_norm": 22.719276428222656,
|
| 1312 |
+
"learning_rate": 1.9455406471981058e-06,
|
| 1313 |
+
"loss": 1.4365,
|
| 1314 |
+
"step": 1830
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 2.153053977272727,
|
| 1318 |
+
"grad_norm": 12.384623527526855,
|
| 1319 |
+
"learning_rate": 1.9258089976322023e-06,
|
| 1320 |
+
"loss": 1.4842,
|
| 1321 |
+
"step": 1840
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 2.1566051136363638,
|
| 1325 |
+
"grad_norm": 29.254623413085938,
|
| 1326 |
+
"learning_rate": 1.9060773480662986e-06,
|
| 1327 |
+
"loss": 1.4222,
|
| 1328 |
+
"step": 1850
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 2.16015625,
|
| 1332 |
+
"grad_norm": 32.79444885253906,
|
| 1333 |
+
"learning_rate": 1.8863456985003947e-06,
|
| 1334 |
+
"loss": 1.6447,
|
| 1335 |
+
"step": 1860
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 2.1637073863636362,
|
| 1339 |
+
"grad_norm": 21.711156845092773,
|
| 1340 |
+
"learning_rate": 1.8666140489344912e-06,
|
| 1341 |
+
"loss": 1.4027,
|
| 1342 |
+
"step": 1870
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 2.167258522727273,
|
| 1346 |
+
"grad_norm": 29.14606285095215,
|
| 1347 |
+
"learning_rate": 1.8468823993685875e-06,
|
| 1348 |
+
"loss": 1.5075,
|
| 1349 |
+
"step": 1880
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 2.170809659090909,
|
| 1353 |
+
"grad_norm": 35.72248077392578,
|
| 1354 |
+
"learning_rate": 1.8271507498026836e-06,
|
| 1355 |
+
"loss": 1.6284,
|
| 1356 |
+
"step": 1890
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 2.1743607954545454,
|
| 1360 |
+
"grad_norm": 28.522951126098633,
|
| 1361 |
+
"learning_rate": 1.8074191002367799e-06,
|
| 1362 |
+
"loss": 1.5546,
|
| 1363 |
+
"step": 1900
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 2.1779119318181817,
|
| 1367 |
+
"grad_norm": 14.657508850097656,
|
| 1368 |
+
"learning_rate": 1.7876874506708764e-06,
|
| 1369 |
+
"loss": 1.5282,
|
| 1370 |
+
"step": 1910
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 2.1814630681818183,
|
| 1374 |
+
"grad_norm": 23.2537899017334,
|
| 1375 |
+
"learning_rate": 1.7679558011049725e-06,
|
| 1376 |
+
"loss": 1.4651,
|
| 1377 |
+
"step": 1920
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 2.1850142045454546,
|
| 1381 |
+
"grad_norm": 21.244699478149414,
|
| 1382 |
+
"learning_rate": 1.7482241515390688e-06,
|
| 1383 |
+
"loss": 1.4957,
|
| 1384 |
+
"step": 1930
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 2.188565340909091,
|
| 1388 |
+
"grad_norm": 42.69889831542969,
|
| 1389 |
+
"learning_rate": 1.7284925019731651e-06,
|
| 1390 |
+
"loss": 1.6205,
|
| 1391 |
+
"step": 1940
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 2.192116477272727,
|
| 1395 |
+
"grad_norm": 32.654991149902344,
|
| 1396 |
+
"learning_rate": 1.7087608524072614e-06,
|
| 1397 |
+
"loss": 1.5725,
|
| 1398 |
+
"step": 1950
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 2.1956676136363638,
|
| 1402 |
+
"grad_norm": 60.657169342041016,
|
| 1403 |
+
"learning_rate": 1.6890292028413577e-06,
|
| 1404 |
+
"loss": 1.5037,
|
| 1405 |
+
"step": 1960
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 2.19921875,
|
| 1409 |
+
"grad_norm": 39.7910041809082,
|
| 1410 |
+
"learning_rate": 1.669297553275454e-06,
|
| 1411 |
+
"loss": 1.4801,
|
| 1412 |
+
"step": 1970
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 2.2027698863636362,
|
| 1416 |
+
"grad_norm": 28.17043113708496,
|
| 1417 |
+
"learning_rate": 1.6495659037095501e-06,
|
| 1418 |
+
"loss": 1.4794,
|
| 1419 |
+
"step": 1980
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 2.206321022727273,
|
| 1423 |
+
"grad_norm": 22.981891632080078,
|
| 1424 |
+
"learning_rate": 1.6298342541436466e-06,
|
| 1425 |
+
"loss": 1.5583,
|
| 1426 |
+
"step": 1990
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 2.209872159090909,
|
| 1430 |
+
"grad_norm": 24.155668258666992,
|
| 1431 |
+
"learning_rate": 1.610102604577743e-06,
|
| 1432 |
+
"loss": 1.4282,
|
| 1433 |
+
"step": 2000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 2.2134232954545454,
|
| 1437 |
+
"grad_norm": 54.06626892089844,
|
| 1438 |
+
"learning_rate": 1.590370955011839e-06,
|
| 1439 |
+
"loss": 1.5627,
|
| 1440 |
+
"step": 2010
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 2.2169744318181817,
|
| 1444 |
+
"grad_norm": 22.495920181274414,
|
| 1445 |
+
"learning_rate": 1.5706393054459353e-06,
|
| 1446 |
+
"loss": 1.4818,
|
| 1447 |
+
"step": 2020
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 2.2205255681818183,
|
| 1451 |
+
"grad_norm": 36.265045166015625,
|
| 1452 |
+
"learning_rate": 1.5509076558800318e-06,
|
| 1453 |
+
"loss": 1.5929,
|
| 1454 |
+
"step": 2030
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 2.2240767045454546,
|
| 1458 |
+
"grad_norm": 15.528562545776367,
|
| 1459 |
+
"learning_rate": 1.531176006314128e-06,
|
| 1460 |
+
"loss": 1.4649,
|
| 1461 |
+
"step": 2040
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 2.227627840909091,
|
| 1465 |
+
"grad_norm": 26.963422775268555,
|
| 1466 |
+
"learning_rate": 1.5114443567482242e-06,
|
| 1467 |
+
"loss": 1.4863,
|
| 1468 |
+
"step": 2050
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 2.231178977272727,
|
| 1472 |
+
"grad_norm": 16.62386131286621,
|
| 1473 |
+
"learning_rate": 1.4917127071823205e-06,
|
| 1474 |
+
"loss": 1.2891,
|
| 1475 |
+
"step": 2060
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.2347301136363638,
|
| 1479 |
+
"grad_norm": 22.612010955810547,
|
| 1480 |
+
"learning_rate": 1.4719810576164168e-06,
|
| 1481 |
+
"loss": 1.4939,
|
| 1482 |
+
"step": 2070
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 2.23828125,
|
| 1486 |
+
"grad_norm": 59.00920104980469,
|
| 1487 |
+
"learning_rate": 1.4522494080505131e-06,
|
| 1488 |
+
"loss": 1.471,
|
| 1489 |
+
"step": 2080
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 2.2418323863636362,
|
| 1493 |
+
"grad_norm": 22.010278701782227,
|
| 1494 |
+
"learning_rate": 1.4325177584846095e-06,
|
| 1495 |
+
"loss": 1.471,
|
| 1496 |
+
"step": 2090
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 2.245383522727273,
|
| 1500 |
+
"grad_norm": 25.769332885742188,
|
| 1501 |
+
"learning_rate": 1.4127861089187055e-06,
|
| 1502 |
+
"loss": 1.3952,
|
| 1503 |
+
"step": 2100
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 2.248934659090909,
|
| 1507 |
+
"grad_norm": 31.15215301513672,
|
| 1508 |
+
"learning_rate": 1.393054459352802e-06,
|
| 1509 |
+
"loss": 1.3164,
|
| 1510 |
+
"step": 2110
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 2.25,
|
| 1514 |
+
"eval_accuracy": 0.3080895008605852,
|
| 1515 |
+
"eval_f1": 0.25675363952610347,
|
| 1516 |
+
"eval_loss": 1.9758973121643066,
|
| 1517 |
+
"eval_precision": 0.30216996841958826,
|
| 1518 |
+
"eval_recall": 0.3084819823166105,
|
| 1519 |
+
"eval_runtime": 716.7915,
|
| 1520 |
+
"eval_samples_per_second": 0.811,
|
| 1521 |
+
"eval_steps_per_second": 0.204,
|
| 1522 |
+
"step": 2113
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 3.0024857954545454,
|
| 1526 |
+
"grad_norm": 25.143760681152344,
|
| 1527 |
+
"learning_rate": 1.3733228097868984e-06,
|
| 1528 |
+
"loss": 1.4265,
|
| 1529 |
+
"step": 2120
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 3.0060369318181817,
|
| 1533 |
+
"grad_norm": 33.473609924316406,
|
| 1534 |
+
"learning_rate": 1.3535911602209945e-06,
|
| 1535 |
+
"loss": 1.3055,
|
| 1536 |
+
"step": 2130
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 3.0095880681818183,
|
| 1540 |
+
"grad_norm": 15.528383255004883,
|
| 1541 |
+
"learning_rate": 1.3338595106550908e-06,
|
| 1542 |
+
"loss": 1.332,
|
| 1543 |
+
"step": 2140
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 3.0131392045454546,
|
| 1547 |
+
"grad_norm": 17.708980560302734,
|
| 1548 |
+
"learning_rate": 1.3141278610891873e-06,
|
| 1549 |
+
"loss": 1.3542,
|
| 1550 |
+
"step": 2150
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 3.016690340909091,
|
| 1554 |
+
"grad_norm": 90.82689666748047,
|
| 1555 |
+
"learning_rate": 1.2943962115232834e-06,
|
| 1556 |
+
"loss": 1.551,
|
| 1557 |
+
"step": 2160
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 3.020241477272727,
|
| 1561 |
+
"grad_norm": 45.7540397644043,
|
| 1562 |
+
"learning_rate": 1.2746645619573797e-06,
|
| 1563 |
+
"loss": 1.4612,
|
| 1564 |
+
"step": 2170
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 3.0237926136363638,
|
| 1568 |
+
"grad_norm": 16.101266860961914,
|
| 1569 |
+
"learning_rate": 1.254932912391476e-06,
|
| 1570 |
+
"loss": 1.3709,
|
| 1571 |
+
"step": 2180
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.02734375,
|
| 1575 |
+
"grad_norm": 13.82054615020752,
|
| 1576 |
+
"learning_rate": 1.2352012628255723e-06,
|
| 1577 |
+
"loss": 1.3762,
|
| 1578 |
+
"step": 2190
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 3.0308948863636362,
|
| 1582 |
+
"grad_norm": 33.30424880981445,
|
| 1583 |
+
"learning_rate": 1.2154696132596686e-06,
|
| 1584 |
+
"loss": 1.4682,
|
| 1585 |
+
"step": 2200
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 3.034446022727273,
|
| 1589 |
+
"grad_norm": 30.845436096191406,
|
| 1590 |
+
"learning_rate": 1.1957379636937649e-06,
|
| 1591 |
+
"loss": 1.4023,
|
| 1592 |
+
"step": 2210
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 3.037997159090909,
|
| 1596 |
+
"grad_norm": 16.949199676513672,
|
| 1597 |
+
"learning_rate": 1.1760063141278612e-06,
|
| 1598 |
+
"loss": 1.3019,
|
| 1599 |
+
"step": 2220
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 3.0415482954545454,
|
| 1603 |
+
"grad_norm": 19.528390884399414,
|
| 1604 |
+
"learning_rate": 1.1562746645619575e-06,
|
| 1605 |
+
"loss": 1.4701,
|
| 1606 |
+
"step": 2230
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 3.0450994318181817,
|
| 1610 |
+
"grad_norm": 53.9114990234375,
|
| 1611 |
+
"learning_rate": 1.1365430149960538e-06,
|
| 1612 |
+
"loss": 1.3315,
|
| 1613 |
+
"step": 2240
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 3.0486505681818183,
|
| 1617 |
+
"grad_norm": 32.42067337036133,
|
| 1618 |
+
"learning_rate": 1.11681136543015e-06,
|
| 1619 |
+
"loss": 1.5252,
|
| 1620 |
+
"step": 2250
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 3.0522017045454546,
|
| 1624 |
+
"grad_norm": 16.656024932861328,
|
| 1625 |
+
"learning_rate": 1.0970797158642462e-06,
|
| 1626 |
+
"loss": 1.5387,
|
| 1627 |
+
"step": 2260
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 3.055752840909091,
|
| 1631 |
+
"grad_norm": 22.413053512573242,
|
| 1632 |
+
"learning_rate": 1.0773480662983427e-06,
|
| 1633 |
+
"loss": 1.3136,
|
| 1634 |
+
"step": 2270
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 3.059303977272727,
|
| 1638 |
+
"grad_norm": 11.095012664794922,
|
| 1639 |
+
"learning_rate": 1.0576164167324388e-06,
|
| 1640 |
+
"loss": 1.4976,
|
| 1641 |
+
"step": 2280
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 3.0628551136363638,
|
| 1645 |
+
"grad_norm": 18.41863250732422,
|
| 1646 |
+
"learning_rate": 1.0378847671665353e-06,
|
| 1647 |
+
"loss": 1.383,
|
| 1648 |
+
"step": 2290
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 3.06640625,
|
| 1652 |
+
"grad_norm": 26.215972900390625,
|
| 1653 |
+
"learning_rate": 1.0181531176006316e-06,
|
| 1654 |
+
"loss": 1.5059,
|
| 1655 |
+
"step": 2300
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 3.0699573863636362,
|
| 1659 |
+
"grad_norm": 28.589454650878906,
|
| 1660 |
+
"learning_rate": 9.984214680347277e-07,
|
| 1661 |
+
"loss": 1.4089,
|
| 1662 |
+
"step": 2310
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 3.073508522727273,
|
| 1666 |
+
"grad_norm": 50.552528381347656,
|
| 1667 |
+
"learning_rate": 9.786898184688242e-07,
|
| 1668 |
+
"loss": 1.4042,
|
| 1669 |
+
"step": 2320
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 3.077059659090909,
|
| 1673 |
+
"grad_norm": 13.100271224975586,
|
| 1674 |
+
"learning_rate": 9.589581689029203e-07,
|
| 1675 |
+
"loss": 1.2665,
|
| 1676 |
+
"step": 2330
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 3.0806107954545454,
|
| 1680 |
+
"grad_norm": 24.274110794067383,
|
| 1681 |
+
"learning_rate": 9.392265193370166e-07,
|
| 1682 |
+
"loss": 1.3727,
|
| 1683 |
+
"step": 2340
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 3.0841619318181817,
|
| 1687 |
+
"grad_norm": 37.16689682006836,
|
| 1688 |
+
"learning_rate": 9.19494869771113e-07,
|
| 1689 |
+
"loss": 1.4553,
|
| 1690 |
+
"step": 2350
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 3.0877130681818183,
|
| 1694 |
+
"grad_norm": 26.41681480407715,
|
| 1695 |
+
"learning_rate": 8.997632202052092e-07,
|
| 1696 |
+
"loss": 1.4915,
|
| 1697 |
+
"step": 2360
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 3.0912642045454546,
|
| 1701 |
+
"grad_norm": 39.80120849609375,
|
| 1702 |
+
"learning_rate": 8.800315706393054e-07,
|
| 1703 |
+
"loss": 1.5277,
|
| 1704 |
+
"step": 2370
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 3.094815340909091,
|
| 1708 |
+
"grad_norm": 20.090126037597656,
|
| 1709 |
+
"learning_rate": 8.602999210734018e-07,
|
| 1710 |
+
"loss": 1.3832,
|
| 1711 |
+
"step": 2380
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 3.098366477272727,
|
| 1715 |
+
"grad_norm": 35.57830047607422,
|
| 1716 |
+
"learning_rate": 8.405682715074981e-07,
|
| 1717 |
+
"loss": 1.358,
|
| 1718 |
+
"step": 2390
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 3.1019176136363638,
|
| 1722 |
+
"grad_norm": 15.675562858581543,
|
| 1723 |
+
"learning_rate": 8.208366219415943e-07,
|
| 1724 |
+
"loss": 1.3206,
|
| 1725 |
+
"step": 2400
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 3.10546875,
|
| 1729 |
+
"grad_norm": 40.23310470581055,
|
| 1730 |
+
"learning_rate": 8.011049723756907e-07,
|
| 1731 |
+
"loss": 1.3537,
|
| 1732 |
+
"step": 2410
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 3.1090198863636362,
|
| 1736 |
+
"grad_norm": 26.265541076660156,
|
| 1737 |
+
"learning_rate": 7.813733228097869e-07,
|
| 1738 |
+
"loss": 1.3955,
|
| 1739 |
+
"step": 2420
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 3.112571022727273,
|
| 1743 |
+
"grad_norm": 25.306949615478516,
|
| 1744 |
+
"learning_rate": 7.616416732438831e-07,
|
| 1745 |
+
"loss": 1.3381,
|
| 1746 |
+
"step": 2430
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 3.116122159090909,
|
| 1750 |
+
"grad_norm": 18.622577667236328,
|
| 1751 |
+
"learning_rate": 7.419100236779796e-07,
|
| 1752 |
+
"loss": 1.3047,
|
| 1753 |
+
"step": 2440
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 3.1196732954545454,
|
| 1757 |
+
"grad_norm": 14.70700454711914,
|
| 1758 |
+
"learning_rate": 7.221783741120759e-07,
|
| 1759 |
+
"loss": 1.3387,
|
| 1760 |
+
"step": 2450
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 3.1232244318181817,
|
| 1764 |
+
"grad_norm": 30.96035385131836,
|
| 1765 |
+
"learning_rate": 7.024467245461721e-07,
|
| 1766 |
+
"loss": 1.3884,
|
| 1767 |
+
"step": 2460
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 3.1267755681818183,
|
| 1771 |
+
"grad_norm": 12.719250679016113,
|
| 1772 |
+
"learning_rate": 6.827150749802685e-07,
|
| 1773 |
+
"loss": 1.4549,
|
| 1774 |
+
"step": 2470
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 3.1303267045454546,
|
| 1778 |
+
"grad_norm": 16.468534469604492,
|
| 1779 |
+
"learning_rate": 6.629834254143647e-07,
|
| 1780 |
+
"loss": 1.1182,
|
| 1781 |
+
"step": 2480
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 3.133877840909091,
|
| 1785 |
+
"grad_norm": 21.350452423095703,
|
| 1786 |
+
"learning_rate": 6.432517758484611e-07,
|
| 1787 |
+
"loss": 1.318,
|
| 1788 |
+
"step": 2490
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 3.137428977272727,
|
| 1792 |
+
"grad_norm": 33.729400634765625,
|
| 1793 |
+
"learning_rate": 6.235201262825573e-07,
|
| 1794 |
+
"loss": 1.3263,
|
| 1795 |
+
"step": 2500
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 3.1409801136363638,
|
| 1799 |
+
"grad_norm": 42.90283966064453,
|
| 1800 |
+
"learning_rate": 6.037884767166536e-07,
|
| 1801 |
+
"loss": 1.3509,
|
| 1802 |
+
"step": 2510
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 3.14453125,
|
| 1806 |
+
"grad_norm": 45.704078674316406,
|
| 1807 |
+
"learning_rate": 5.840568271507499e-07,
|
| 1808 |
+
"loss": 1.5069,
|
| 1809 |
+
"step": 2520
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 3.1480823863636362,
|
| 1813 |
+
"grad_norm": 41.951148986816406,
|
| 1814 |
+
"learning_rate": 5.643251775848462e-07,
|
| 1815 |
+
"loss": 1.3066,
|
| 1816 |
+
"step": 2530
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 3.151633522727273,
|
| 1820 |
+
"grad_norm": 12.798999786376953,
|
| 1821 |
+
"learning_rate": 5.445935280189425e-07,
|
| 1822 |
+
"loss": 1.3401,
|
| 1823 |
+
"step": 2540
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 3.155184659090909,
|
| 1827 |
+
"grad_norm": 21.00889778137207,
|
| 1828 |
+
"learning_rate": 5.248618784530387e-07,
|
| 1829 |
+
"loss": 1.2965,
|
| 1830 |
+
"step": 2550
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 3.1587357954545454,
|
| 1834 |
+
"grad_norm": 17.473670959472656,
|
| 1835 |
+
"learning_rate": 5.05130228887135e-07,
|
| 1836 |
+
"loss": 1.2885,
|
| 1837 |
+
"step": 2560
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 3.1622869318181817,
|
| 1841 |
+
"grad_norm": 28.33868980407715,
|
| 1842 |
+
"learning_rate": 4.853985793212313e-07,
|
| 1843 |
+
"loss": 1.3802,
|
| 1844 |
+
"step": 2570
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 3.1658380681818183,
|
| 1848 |
+
"grad_norm": 37.80378723144531,
|
| 1849 |
+
"learning_rate": 4.6566692975532754e-07,
|
| 1850 |
+
"loss": 1.3751,
|
| 1851 |
+
"step": 2580
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 3.1693892045454546,
|
| 1855 |
+
"grad_norm": 37.315486907958984,
|
| 1856 |
+
"learning_rate": 4.4593528018942384e-07,
|
| 1857 |
+
"loss": 1.4067,
|
| 1858 |
+
"step": 2590
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 3.172940340909091,
|
| 1862 |
+
"grad_norm": 26.541744232177734,
|
| 1863 |
+
"learning_rate": 4.2620363062352015e-07,
|
| 1864 |
+
"loss": 1.3521,
|
| 1865 |
+
"step": 2600
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 3.176491477272727,
|
| 1869 |
+
"grad_norm": 13.96971321105957,
|
| 1870 |
+
"learning_rate": 4.064719810576165e-07,
|
| 1871 |
+
"loss": 1.4583,
|
| 1872 |
+
"step": 2610
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 3.1800426136363638,
|
| 1876 |
+
"grad_norm": 19.711917877197266,
|
| 1877 |
+
"learning_rate": 3.867403314917127e-07,
|
| 1878 |
+
"loss": 1.3062,
|
| 1879 |
+
"step": 2620
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 3.18359375,
|
| 1883 |
+
"grad_norm": 28.168289184570312,
|
| 1884 |
+
"learning_rate": 3.6700868192580906e-07,
|
| 1885 |
+
"loss": 1.4479,
|
| 1886 |
+
"step": 2630
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 3.1871448863636362,
|
| 1890 |
+
"grad_norm": 16.958044052124023,
|
| 1891 |
+
"learning_rate": 3.4727703235990536e-07,
|
| 1892 |
+
"loss": 1.2349,
|
| 1893 |
+
"step": 2640
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.190696022727273,
|
| 1897 |
+
"grad_norm": 14.27488899230957,
|
| 1898 |
+
"learning_rate": 3.2754538279400156e-07,
|
| 1899 |
+
"loss": 1.3124,
|
| 1900 |
+
"step": 2650
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 3.194247159090909,
|
| 1904 |
+
"grad_norm": 39.53180694580078,
|
| 1905 |
+
"learning_rate": 3.078137332280979e-07,
|
| 1906 |
+
"loss": 1.3704,
|
| 1907 |
+
"step": 2660
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 3.1977982954545454,
|
| 1911 |
+
"grad_norm": 30.55574607849121,
|
| 1912 |
+
"learning_rate": 2.8808208366219417e-07,
|
| 1913 |
+
"loss": 1.4764,
|
| 1914 |
+
"step": 2670
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 3.2013494318181817,
|
| 1918 |
+
"grad_norm": 32.31404113769531,
|
| 1919 |
+
"learning_rate": 2.6835043409629047e-07,
|
| 1920 |
+
"loss": 1.3712,
|
| 1921 |
+
"step": 2680
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 3.2049005681818183,
|
| 1925 |
+
"grad_norm": 17.91683578491211,
|
| 1926 |
+
"learning_rate": 2.486187845303868e-07,
|
| 1927 |
+
"loss": 1.3872,
|
| 1928 |
+
"step": 2690
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 3.2084517045454546,
|
| 1932 |
+
"grad_norm": 27.588768005371094,
|
| 1933 |
+
"learning_rate": 2.2888713496448305e-07,
|
| 1934 |
+
"loss": 1.3406,
|
| 1935 |
+
"step": 2700
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.212002840909091,
|
| 1939 |
+
"grad_norm": 16.939321517944336,
|
| 1940 |
+
"learning_rate": 2.0915548539857933e-07,
|
| 1941 |
+
"loss": 1.4795,
|
| 1942 |
+
"step": 2710
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 3.215553977272727,
|
| 1946 |
+
"grad_norm": 30.011600494384766,
|
| 1947 |
+
"learning_rate": 1.8942383583267563e-07,
|
| 1948 |
+
"loss": 1.2874,
|
| 1949 |
+
"step": 2720
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 3.2191051136363638,
|
| 1953 |
+
"grad_norm": 38.46843719482422,
|
| 1954 |
+
"learning_rate": 1.696921862667719e-07,
|
| 1955 |
+
"loss": 1.4501,
|
| 1956 |
+
"step": 2730
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 3.22265625,
|
| 1960 |
+
"grad_norm": 28.307205200195312,
|
| 1961 |
+
"learning_rate": 1.4996053670086821e-07,
|
| 1962 |
+
"loss": 1.3977,
|
| 1963 |
+
"step": 2740
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 3.2262073863636362,
|
| 1967 |
+
"grad_norm": 36.628013610839844,
|
| 1968 |
+
"learning_rate": 1.302288871349645e-07,
|
| 1969 |
+
"loss": 1.2878,
|
| 1970 |
+
"step": 2750
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 3.229758522727273,
|
| 1974 |
+
"grad_norm": 38.92404556274414,
|
| 1975 |
+
"learning_rate": 1.1049723756906078e-07,
|
| 1976 |
+
"loss": 1.3168,
|
| 1977 |
+
"step": 2760
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 3.233309659090909,
|
| 1981 |
+
"grad_norm": 30.143110275268555,
|
| 1982 |
+
"learning_rate": 9.076558800315707e-08,
|
| 1983 |
+
"loss": 1.3171,
|
| 1984 |
+
"step": 2770
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 3.2368607954545454,
|
| 1988 |
+
"grad_norm": 21.9947509765625,
|
| 1989 |
+
"learning_rate": 7.103393843725336e-08,
|
| 1990 |
+
"loss": 1.3034,
|
| 1991 |
+
"step": 2780
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 3.2404119318181817,
|
| 1995 |
+
"grad_norm": 29.996990203857422,
|
| 1996 |
+
"learning_rate": 5.1302288871349646e-08,
|
| 1997 |
+
"loss": 1.3983,
|
| 1998 |
+
"step": 2790
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 3.2439630681818183,
|
| 2002 |
+
"grad_norm": 67.70350646972656,
|
| 2003 |
+
"learning_rate": 3.1570639305445937e-08,
|
| 2004 |
+
"loss": 1.5066,
|
| 2005 |
+
"step": 2800
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 3.2475142045454546,
|
| 2009 |
+
"grad_norm": 15.010125160217285,
|
| 2010 |
+
"learning_rate": 1.1838989739542227e-08,
|
| 2011 |
+
"loss": 1.3877,
|
| 2012 |
+
"step": 2810
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 3.2496448863636362,
|
| 2016 |
+
"eval_accuracy": 0.3373493975903614,
|
| 2017 |
+
"eval_f1": 0.2839461670482319,
|
| 2018 |
+
"eval_loss": 1.9640586376190186,
|
| 2019 |
+
"eval_precision": 0.30306634843723573,
|
| 2020 |
+
"eval_recall": 0.3366896237317991,
|
| 2021 |
+
"eval_runtime": 724.6553,
|
| 2022 |
+
"eval_samples_per_second": 0.802,
|
| 2023 |
+
"eval_steps_per_second": 0.201,
|
| 2024 |
+
"step": 2816
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 3.2496448863636362,
|
| 2028 |
+
"step": 2816,
|
| 2029 |
+
"total_flos": 1.4033059590817677e+19,
|
| 2030 |
+
"train_loss": 1.800246679986065,
|
| 2031 |
+
"train_runtime": 18585.0507,
|
| 2032 |
+
"train_samples_per_second": 0.606,
|
| 2033 |
+
"train_steps_per_second": 0.152
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 3.2496448863636362,
|
| 2037 |
+
"eval_accuracy": 0.3268698060941828,
|
| 2038 |
+
"eval_f1": 0.2716015567288163,
|
| 2039 |
+
"eval_loss": 1.972208857536316,
|
| 2040 |
+
"eval_precision": 0.3970093449590282,
|
| 2041 |
+
"eval_recall": 0.32766441258833473,
|
| 2042 |
+
"eval_runtime": 907.789,
|
| 2043 |
+
"eval_samples_per_second": 0.795,
|
| 2044 |
+
"eval_steps_per_second": 0.199,
|
| 2045 |
+
"step": 2816
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.2496448863636362,
|
| 2049 |
+
"eval_accuracy": 0.3268698060941828,
|
| 2050 |
+
"eval_f1": 0.2716015567288163,
|
| 2051 |
+
"eval_loss": 1.972209095954895,
|
| 2052 |
+
"eval_precision": 0.3970093449590282,
|
| 2053 |
+
"eval_recall": 0.32766441258833473,
|
| 2054 |
+
"eval_runtime": 909.7137,
|
| 2055 |
+
"eval_samples_per_second": 0.794,
|
| 2056 |
+
"eval_steps_per_second": 0.199,
|
| 2057 |
+
"step": 2816
|
| 2058 |
+
}
|
| 2059 |
+
],
|
| 2060 |
+
"logging_steps": 10,
|
| 2061 |
+
"max_steps": 2816,
|
| 2062 |
+
"num_input_tokens_seen": 0,
|
| 2063 |
+
"num_train_epochs": 9223372036854775807,
|
| 2064 |
+
"save_steps": 500,
|
| 2065 |
+
"total_flos": 1.4033059590817677e+19,
|
| 2066 |
+
"train_batch_size": 4,
|
| 2067 |
+
"trial_name": null,
|
| 2068 |
+
"trial_params": null
|
| 2069 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73521adfb0d97de93eed7bdbcd69698c1161e6efa9b954a4a4db0593f2284f73
|
| 3 |
+
size 5048
|