Model save
Browse files- README.md +66 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- runs/Sep12_16-26-43_COE-CS-sv003/events.out.tfevents.1726158429.COE-CS-sv003.3137015.0 +2 -2
- train_results.json +9 -0
- trainer_state.json +3431 -0
README.md
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- sft
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
datasets:
|
| 9 |
+
- generator
|
| 10 |
+
model-index:
|
| 11 |
+
- name: preference_p0.2_seed42_level2_raremixbatch16
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 16 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 17 |
+
|
| 18 |
+
# preference_p0.2_seed42_level2_raremixbatch16
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the generator dataset.
|
| 21 |
+
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.2978
|
| 23 |
+
|
| 24 |
+
## Model description
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Intended uses & limitations
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training and evaluation data
|
| 33 |
+
|
| 34 |
+
More information needed
|
| 35 |
+
|
| 36 |
+
## Training procedure
|
| 37 |
+
|
| 38 |
+
### Training hyperparameters
|
| 39 |
+
|
| 40 |
+
The following hyperparameters were used during training:
|
| 41 |
+
- learning_rate: 1e-05
|
| 42 |
+
- train_batch_size: 4
|
| 43 |
+
- eval_batch_size: 1
|
| 44 |
+
- seed: 42
|
| 45 |
+
- distributed_type: multi-GPU
|
| 46 |
+
- num_devices: 4
|
| 47 |
+
- total_train_batch_size: 16
|
| 48 |
+
- total_eval_batch_size: 4
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 1
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
| 58 |
+
| 0.3188 | 1.0 | 2413 | 0.2978 |
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
### Framework versions
|
| 62 |
+
|
| 63 |
+
- Transformers 4.43.4
|
| 64 |
+
- Pytorch 2.3.1+cu121
|
| 65 |
+
- Datasets 2.19.1
|
| 66 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 252616554577920.0,
|
| 4 |
+
"train_loss": 0.5160677275912573,
|
| 5 |
+
"train_runtime": 24380.0574,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.584,
|
| 8 |
+
"train_steps_per_second": 0.099
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.43.4"
|
| 6 |
+
}
|
runs/Sep12_16-26-43_COE-CS-sv003/events.out.tfevents.1726158429.COE-CS-sv003.3137015.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1963124f4b2110f30de4a818ab3245c7c0c396ff3a439f00f97430a8814bd604
|
| 3 |
+
size 107702
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 252616554577920.0,
|
| 4 |
+
"train_loss": 0.5160677275912573,
|
| 5 |
+
"train_runtime": 24380.0574,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.584,
|
| 8 |
+
"train_steps_per_second": 0.099
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3431 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2413,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004144218814753419,
|
| 13 |
+
"grad_norm": 24.514733364424554,
|
| 14 |
+
"learning_rate": 4.132231404958678e-08,
|
| 15 |
+
"loss": 1.4169,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0020721094073767096,
|
| 20 |
+
"grad_norm": 23.537184095238235,
|
| 21 |
+
"learning_rate": 2.066115702479339e-07,
|
| 22 |
+
"loss": 1.3953,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004144218814753419,
|
| 27 |
+
"grad_norm": 15.367263243187544,
|
| 28 |
+
"learning_rate": 4.132231404958678e-07,
|
| 29 |
+
"loss": 1.3778,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006216328222130129,
|
| 34 |
+
"grad_norm": 8.898381617118186,
|
| 35 |
+
"learning_rate": 6.198347107438018e-07,
|
| 36 |
+
"loss": 1.2602,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.008288437629506838,
|
| 41 |
+
"grad_norm": 10.642672694622197,
|
| 42 |
+
"learning_rate": 8.264462809917356e-07,
|
| 43 |
+
"loss": 1.1548,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.010360547036883548,
|
| 48 |
+
"grad_norm": 4.617567060806481,
|
| 49 |
+
"learning_rate": 1.0330578512396695e-06,
|
| 50 |
+
"loss": 1.0324,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.012432656444260257,
|
| 55 |
+
"grad_norm": 3.417020439965166,
|
| 56 |
+
"learning_rate": 1.2396694214876035e-06,
|
| 57 |
+
"loss": 0.9692,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.014504765851636967,
|
| 62 |
+
"grad_norm": 3.395692919461883,
|
| 63 |
+
"learning_rate": 1.4462809917355372e-06,
|
| 64 |
+
"loss": 0.9523,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.016576875259013676,
|
| 69 |
+
"grad_norm": 3.151546235627503,
|
| 70 |
+
"learning_rate": 1.6528925619834712e-06,
|
| 71 |
+
"loss": 0.9274,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.018648984666390384,
|
| 76 |
+
"grad_norm": 2.9582253473643485,
|
| 77 |
+
"learning_rate": 1.859504132231405e-06,
|
| 78 |
+
"loss": 0.9323,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.020721094073767096,
|
| 83 |
+
"grad_norm": 3.0341795582664997,
|
| 84 |
+
"learning_rate": 2.066115702479339e-06,
|
| 85 |
+
"loss": 0.8943,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.022793203481143803,
|
| 90 |
+
"grad_norm": 3.0341130670598164,
|
| 91 |
+
"learning_rate": 2.2727272727272728e-06,
|
| 92 |
+
"loss": 0.8975,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.024865312888520515,
|
| 97 |
+
"grad_norm": 3.1718207048254627,
|
| 98 |
+
"learning_rate": 2.479338842975207e-06,
|
| 99 |
+
"loss": 0.8814,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.026937422295897222,
|
| 104 |
+
"grad_norm": 3.0690467751730495,
|
| 105 |
+
"learning_rate": 2.6859504132231405e-06,
|
| 106 |
+
"loss": 0.8886,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.029009531703273934,
|
| 111 |
+
"grad_norm": 3.1173422324045084,
|
| 112 |
+
"learning_rate": 2.8925619834710743e-06,
|
| 113 |
+
"loss": 0.8779,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03108164111065064,
|
| 118 |
+
"grad_norm": 3.138742895479937,
|
| 119 |
+
"learning_rate": 3.0991735537190086e-06,
|
| 120 |
+
"loss": 0.8896,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03315375051802735,
|
| 125 |
+
"grad_norm": 3.172114081672577,
|
| 126 |
+
"learning_rate": 3.3057851239669424e-06,
|
| 127 |
+
"loss": 0.852,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.035225859925404064,
|
| 132 |
+
"grad_norm": 3.0136111662154126,
|
| 133 |
+
"learning_rate": 3.5123966942148763e-06,
|
| 134 |
+
"loss": 0.8718,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.03729796933278077,
|
| 139 |
+
"grad_norm": 3.258761073776976,
|
| 140 |
+
"learning_rate": 3.71900826446281e-06,
|
| 141 |
+
"loss": 0.8515,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.03937007874015748,
|
| 146 |
+
"grad_norm": 3.1414312111507323,
|
| 147 |
+
"learning_rate": 3.925619834710744e-06,
|
| 148 |
+
"loss": 0.8653,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04144218814753419,
|
| 153 |
+
"grad_norm": 3.207392945640886,
|
| 154 |
+
"learning_rate": 4.132231404958678e-06,
|
| 155 |
+
"loss": 0.8688,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.0435142975549109,
|
| 160 |
+
"grad_norm": 3.2052306999586695,
|
| 161 |
+
"learning_rate": 4.338842975206612e-06,
|
| 162 |
+
"loss": 0.8662,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.04558640696228761,
|
| 167 |
+
"grad_norm": 3.061390265998909,
|
| 168 |
+
"learning_rate": 4.5454545454545455e-06,
|
| 169 |
+
"loss": 0.8397,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.04765851636966432,
|
| 174 |
+
"grad_norm": 3.0154741328483023,
|
| 175 |
+
"learning_rate": 4.75206611570248e-06,
|
| 176 |
+
"loss": 0.8583,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.04973062577704103,
|
| 181 |
+
"grad_norm": 2.9687539148068796,
|
| 182 |
+
"learning_rate": 4.958677685950414e-06,
|
| 183 |
+
"loss": 0.8542,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05180273518441774,
|
| 188 |
+
"grad_norm": 3.0956570078052925,
|
| 189 |
+
"learning_rate": 5.165289256198347e-06,
|
| 190 |
+
"loss": 0.8343,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.053874844591794445,
|
| 195 |
+
"grad_norm": 2.939810124211356,
|
| 196 |
+
"learning_rate": 5.371900826446281e-06,
|
| 197 |
+
"loss": 0.8457,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.055946953999171156,
|
| 202 |
+
"grad_norm": 3.2352916246928203,
|
| 203 |
+
"learning_rate": 5.578512396694216e-06,
|
| 204 |
+
"loss": 0.8169,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.05801906340654787,
|
| 209 |
+
"grad_norm": 3.1434079053605273,
|
| 210 |
+
"learning_rate": 5.785123966942149e-06,
|
| 211 |
+
"loss": 0.8257,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06009117281392457,
|
| 216 |
+
"grad_norm": 3.1532690460920803,
|
| 217 |
+
"learning_rate": 5.991735537190083e-06,
|
| 218 |
+
"loss": 0.8392,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06216328222130128,
|
| 223 |
+
"grad_norm": 3.0343466546862037,
|
| 224 |
+
"learning_rate": 6.198347107438017e-06,
|
| 225 |
+
"loss": 0.8442,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.06423539162867799,
|
| 230 |
+
"grad_norm": 2.9508007571282318,
|
| 231 |
+
"learning_rate": 6.404958677685951e-06,
|
| 232 |
+
"loss": 0.8218,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.0663075010360547,
|
| 237 |
+
"grad_norm": 3.0527743250572033,
|
| 238 |
+
"learning_rate": 6.611570247933885e-06,
|
| 239 |
+
"loss": 0.8294,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.06837961044343141,
|
| 244 |
+
"grad_norm": 3.075132609753167,
|
| 245 |
+
"learning_rate": 6.818181818181818e-06,
|
| 246 |
+
"loss": 0.8266,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07045171985080813,
|
| 251 |
+
"grad_norm": 2.972687092057336,
|
| 252 |
+
"learning_rate": 7.0247933884297525e-06,
|
| 253 |
+
"loss": 0.8139,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.07252382925818483,
|
| 258 |
+
"grad_norm": 3.068027468513739,
|
| 259 |
+
"learning_rate": 7.231404958677687e-06,
|
| 260 |
+
"loss": 0.7992,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.07459593866556154,
|
| 265 |
+
"grad_norm": 3.158357724987422,
|
| 266 |
+
"learning_rate": 7.43801652892562e-06,
|
| 267 |
+
"loss": 0.8134,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.07666804807293826,
|
| 272 |
+
"grad_norm": 3.051134008290911,
|
| 273 |
+
"learning_rate": 7.644628099173555e-06,
|
| 274 |
+
"loss": 0.8344,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.07874015748031496,
|
| 279 |
+
"grad_norm": 3.3235580452981175,
|
| 280 |
+
"learning_rate": 7.851239669421489e-06,
|
| 281 |
+
"loss": 0.8041,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08081226688769166,
|
| 286 |
+
"grad_norm": 3.0678026193628805,
|
| 287 |
+
"learning_rate": 8.057851239669421e-06,
|
| 288 |
+
"loss": 0.8045,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.08288437629506838,
|
| 293 |
+
"grad_norm": 3.415940865122294,
|
| 294 |
+
"learning_rate": 8.264462809917356e-06,
|
| 295 |
+
"loss": 0.7953,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.08495648570244509,
|
| 300 |
+
"grad_norm": 3.815817395153579,
|
| 301 |
+
"learning_rate": 8.47107438016529e-06,
|
| 302 |
+
"loss": 0.8047,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.0870285951098218,
|
| 307 |
+
"grad_norm": 3.2984513943211153,
|
| 308 |
+
"learning_rate": 8.677685950413224e-06,
|
| 309 |
+
"loss": 0.8056,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.08910070451719851,
|
| 314 |
+
"grad_norm": 2.8880894973116793,
|
| 315 |
+
"learning_rate": 8.884297520661158e-06,
|
| 316 |
+
"loss": 0.795,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.09117281392457521,
|
| 321 |
+
"grad_norm": 2.974477767781155,
|
| 322 |
+
"learning_rate": 9.090909090909091e-06,
|
| 323 |
+
"loss": 0.8014,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.09324492333195193,
|
| 328 |
+
"grad_norm": 3.0809082608070444,
|
| 329 |
+
"learning_rate": 9.297520661157025e-06,
|
| 330 |
+
"loss": 0.812,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.09531703273932864,
|
| 335 |
+
"grad_norm": 3.11471679186006,
|
| 336 |
+
"learning_rate": 9.50413223140496e-06,
|
| 337 |
+
"loss": 0.815,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.09738914214670534,
|
| 342 |
+
"grad_norm": 3.3405073257172333,
|
| 343 |
+
"learning_rate": 9.710743801652894e-06,
|
| 344 |
+
"loss": 0.7991,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.09946125155408206,
|
| 349 |
+
"grad_norm": 3.1375906477529485,
|
| 350 |
+
"learning_rate": 9.917355371900828e-06,
|
| 351 |
+
"loss": 0.7902,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.10153336096145876,
|
| 356 |
+
"grad_norm": 2.8343248219683823,
|
| 357 |
+
"learning_rate": 9.999952884702848e-06,
|
| 358 |
+
"loss": 0.8215,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.10360547036883548,
|
| 363 |
+
"grad_norm": 2.9925055543397563,
|
| 364 |
+
"learning_rate": 9.999664961102495e-06,
|
| 365 |
+
"loss": 0.8084,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.10567757977621219,
|
| 370 |
+
"grad_norm": 3.0441664134730275,
|
| 371 |
+
"learning_rate": 9.999115304121459e-06,
|
| 372 |
+
"loss": 0.8085,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.10774968918358889,
|
| 377 |
+
"grad_norm": 3.039997210231939,
|
| 378 |
+
"learning_rate": 9.998303942534383e-06,
|
| 379 |
+
"loss": 0.7938,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.10982179859096561,
|
| 384 |
+
"grad_norm": 3.0431241796795696,
|
| 385 |
+
"learning_rate": 9.997230918816193e-06,
|
| 386 |
+
"loss": 0.7985,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.11189390799834231,
|
| 391 |
+
"grad_norm": 2.85027037020918,
|
| 392 |
+
"learning_rate": 9.99589628913988e-06,
|
| 393 |
+
"loss": 0.7861,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.11396601740571902,
|
| 398 |
+
"grad_norm": 3.051497861941328,
|
| 399 |
+
"learning_rate": 9.994300123373554e-06,
|
| 400 |
+
"loss": 0.7716,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.11603812681309573,
|
| 405 |
+
"grad_norm": 2.895125032094719,
|
| 406 |
+
"learning_rate": 9.992442505076788e-06,
|
| 407 |
+
"loss": 0.7834,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.11811023622047244,
|
| 412 |
+
"grad_norm": 3.0061812704069886,
|
| 413 |
+
"learning_rate": 9.990323531496235e-06,
|
| 414 |
+
"loss": 0.7756,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.12018234562784914,
|
| 419 |
+
"grad_norm": 3.0347710812056525,
|
| 420 |
+
"learning_rate": 9.98794331356056e-06,
|
| 421 |
+
"loss": 0.7846,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.12225445503522586,
|
| 426 |
+
"grad_norm": 2.805184541271578,
|
| 427 |
+
"learning_rate": 9.985301975874604e-06,
|
| 428 |
+
"loss": 0.7731,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.12432656444260257,
|
| 433 |
+
"grad_norm": 3.0842132440120147,
|
| 434 |
+
"learning_rate": 9.982399656712884e-06,
|
| 435 |
+
"loss": 0.8042,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.12639867384997927,
|
| 440 |
+
"grad_norm": 2.9008174183880233,
|
| 441 |
+
"learning_rate": 9.979236508012341e-06,
|
| 442 |
+
"loss": 0.7681,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.12847078325735597,
|
| 447 |
+
"grad_norm": 2.8236286325458244,
|
| 448 |
+
"learning_rate": 9.975812695364391e-06,
|
| 449 |
+
"loss": 0.7891,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.1305428926647327,
|
| 454 |
+
"grad_norm": 2.8468044382747384,
|
| 455 |
+
"learning_rate": 9.97212839800626e-06,
|
| 456 |
+
"loss": 0.7681,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.1326150020721094,
|
| 461 |
+
"grad_norm": 2.9534507280362003,
|
| 462 |
+
"learning_rate": 9.968183808811586e-06,
|
| 463 |
+
"loss": 0.7564,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.13468711147948612,
|
| 468 |
+
"grad_norm": 2.965082743686242,
|
| 469 |
+
"learning_rate": 9.963979134280344e-06,
|
| 470 |
+
"loss": 0.7529,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.13675922088686282,
|
| 475 |
+
"grad_norm": 2.847254353400155,
|
| 476 |
+
"learning_rate": 9.959514594528018e-06,
|
| 477 |
+
"loss": 0.7438,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.13883133029423952,
|
| 482 |
+
"grad_norm": 2.7966052001739206,
|
| 483 |
+
"learning_rate": 9.954790423274086e-06,
|
| 484 |
+
"loss": 0.7591,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.14090343970161626,
|
| 489 |
+
"grad_norm": 2.7319025763391576,
|
| 490 |
+
"learning_rate": 9.94980686782978e-06,
|
| 491 |
+
"loss": 0.7406,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.14297554910899296,
|
| 496 |
+
"grad_norm": 2.870893839424329,
|
| 497 |
+
"learning_rate": 9.94456418908515e-06,
|
| 498 |
+
"loss": 0.7541,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.14504765851636967,
|
| 503 |
+
"grad_norm": 3.009522094390881,
|
| 504 |
+
"learning_rate": 9.939062661495387e-06,
|
| 505 |
+
"loss": 0.7511,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.14711976792374637,
|
| 510 |
+
"grad_norm": 2.7182810127315573,
|
| 511 |
+
"learning_rate": 9.933302573066477e-06,
|
| 512 |
+
"loss": 0.7688,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.14919187733112307,
|
| 517 |
+
"grad_norm": 3.012302144551778,
|
| 518 |
+
"learning_rate": 9.927284225340105e-06,
|
| 519 |
+
"loss": 0.7341,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.1512639867384998,
|
| 524 |
+
"grad_norm": 3.337745192726695,
|
| 525 |
+
"learning_rate": 9.921007933377886e-06,
|
| 526 |
+
"loss": 0.7539,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.1533360961458765,
|
| 531 |
+
"grad_norm": 3.17882772460756,
|
| 532 |
+
"learning_rate": 9.914474025744855e-06,
|
| 533 |
+
"loss": 0.7506,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.15540820555325321,
|
| 538 |
+
"grad_norm": 3.2860227088727383,
|
| 539 |
+
"learning_rate": 9.907682844492283e-06,
|
| 540 |
+
"loss": 0.7514,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.15748031496062992,
|
| 545 |
+
"grad_norm": 3.1971805376225517,
|
| 546 |
+
"learning_rate": 9.900634745139759e-06,
|
| 547 |
+
"loss": 0.7475,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.15955242436800662,
|
| 552 |
+
"grad_norm": 2.8028862036100843,
|
| 553 |
+
"learning_rate": 9.893330096656576e-06,
|
| 554 |
+
"loss": 0.7285,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.16162453377538333,
|
| 559 |
+
"grad_norm": 2.708278578400365,
|
| 560 |
+
"learning_rate": 9.885769281442426e-06,
|
| 561 |
+
"loss": 0.7224,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.16369664318276006,
|
| 566 |
+
"grad_norm": 2.771966186486157,
|
| 567 |
+
"learning_rate": 9.877952695307382e-06,
|
| 568 |
+
"loss": 0.7287,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.16576875259013676,
|
| 573 |
+
"grad_norm": 2.645488728201908,
|
| 574 |
+
"learning_rate": 9.869880747451164e-06,
|
| 575 |
+
"loss": 0.7389,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.16784086199751347,
|
| 580 |
+
"grad_norm": 2.9190913357286306,
|
| 581 |
+
"learning_rate": 9.861553860441726e-06,
|
| 582 |
+
"loss": 0.7414,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.16991297140489017,
|
| 587 |
+
"grad_norm": 2.6920496051405998,
|
| 588 |
+
"learning_rate": 9.852972470193136e-06,
|
| 589 |
+
"loss": 0.7259,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.17198508081226688,
|
| 594 |
+
"grad_norm": 2.9415197302471663,
|
| 595 |
+
"learning_rate": 9.844137025942755e-06,
|
| 596 |
+
"loss": 0.7266,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.1740571902196436,
|
| 601 |
+
"grad_norm": 2.7808373109476294,
|
| 602 |
+
"learning_rate": 9.835047990227713e-06,
|
| 603 |
+
"loss": 0.7119,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.1761292996270203,
|
| 608 |
+
"grad_norm": 2.810707277824952,
|
| 609 |
+
"learning_rate": 9.825705838860699e-06,
|
| 610 |
+
"loss": 0.7361,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.17820140903439702,
|
| 615 |
+
"grad_norm": 3.0525714644872117,
|
| 616 |
+
"learning_rate": 9.816111060905063e-06,
|
| 617 |
+
"loss": 0.7146,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.18027351844177372,
|
| 622 |
+
"grad_norm": 2.769816103147695,
|
| 623 |
+
"learning_rate": 9.806264158649193e-06,
|
| 624 |
+
"loss": 0.7104,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.18234562784915043,
|
| 629 |
+
"grad_norm": 2.8698277588389516,
|
| 630 |
+
"learning_rate": 9.796165647580233e-06,
|
| 631 |
+
"loss": 0.7015,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.18441773725652713,
|
| 636 |
+
"grad_norm": 2.632541484736724,
|
| 637 |
+
"learning_rate": 9.785816056357096e-06,
|
| 638 |
+
"loss": 0.7148,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.18648984666390386,
|
| 643 |
+
"grad_norm": 2.6937205983501276,
|
| 644 |
+
"learning_rate": 9.775215926782788e-06,
|
| 645 |
+
"loss": 0.7203,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.18856195607128057,
|
| 650 |
+
"grad_norm": 2.7810565830560354,
|
| 651 |
+
"learning_rate": 9.764365813776042e-06,
|
| 652 |
+
"loss": 0.7068,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.19063406547865727,
|
| 657 |
+
"grad_norm": 2.8012616966262134,
|
| 658 |
+
"learning_rate": 9.753266285342271e-06,
|
| 659 |
+
"loss": 0.7104,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.19270617488603398,
|
| 664 |
+
"grad_norm": 2.970139751785694,
|
| 665 |
+
"learning_rate": 9.741917922543831e-06,
|
| 666 |
+
"loss": 0.6881,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.19477828429341068,
|
| 671 |
+
"grad_norm": 2.6941473339327824,
|
| 672 |
+
"learning_rate": 9.7303213194696e-06,
|
| 673 |
+
"loss": 0.6996,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.1968503937007874,
|
| 678 |
+
"grad_norm": 3.0409394398260554,
|
| 679 |
+
"learning_rate": 9.718477083203888e-06,
|
| 680 |
+
"loss": 0.6933,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.19892250310816412,
|
| 685 |
+
"grad_norm": 2.8048791766929146,
|
| 686 |
+
"learning_rate": 9.706385833794639e-06,
|
| 687 |
+
"loss": 0.6748,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.20099461251554082,
|
| 692 |
+
"grad_norm": 2.7366882872916887,
|
| 693 |
+
"learning_rate": 9.694048204220986e-06,
|
| 694 |
+
"loss": 0.7044,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.20306672192291753,
|
| 699 |
+
"grad_norm": 2.7267358802595374,
|
| 700 |
+
"learning_rate": 9.681464840360105e-06,
|
| 701 |
+
"loss": 0.6912,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.20513883133029423,
|
| 706 |
+
"grad_norm": 2.754099178327694,
|
| 707 |
+
"learning_rate": 9.668636400953411e-06,
|
| 708 |
+
"loss": 0.6731,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.20721094073767096,
|
| 713 |
+
"grad_norm": 3.11521121490046,
|
| 714 |
+
"learning_rate": 9.655563557572068e-06,
|
| 715 |
+
"loss": 0.7018,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.20928305014504767,
|
| 720 |
+
"grad_norm": 2.729648091432038,
|
| 721 |
+
"learning_rate": 9.642246994581833e-06,
|
| 722 |
+
"loss": 0.6919,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.21135515955242437,
|
| 727 |
+
"grad_norm": 2.788565650588865,
|
| 728 |
+
"learning_rate": 9.62868740910723e-06,
|
| 729 |
+
"loss": 0.6718,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.21342726895980108,
|
| 734 |
+
"grad_norm": 2.6821583529570527,
|
| 735 |
+
"learning_rate": 9.614885510995047e-06,
|
| 736 |
+
"loss": 0.6696,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.21549937836717778,
|
| 741 |
+
"grad_norm": 2.669430236726845,
|
| 742 |
+
"learning_rate": 9.600842022777198e-06,
|
| 743 |
+
"loss": 0.686,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.21757148777455448,
|
| 748 |
+
"grad_norm": 2.6329122210908613,
|
| 749 |
+
"learning_rate": 9.58655767963287e-06,
|
| 750 |
+
"loss": 0.6649,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.21964359718193122,
|
| 755 |
+
"grad_norm": 2.712511744397499,
|
| 756 |
+
"learning_rate": 9.57203322935006e-06,
|
| 757 |
+
"loss": 0.6691,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.22171570658930792,
|
| 762 |
+
"grad_norm": 2.971121715072785,
|
| 763 |
+
"learning_rate": 9.557269432286406e-06,
|
| 764 |
+
"loss": 0.6568,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.22378781599668462,
|
| 769 |
+
"grad_norm": 3.5850165463286587,
|
| 770 |
+
"learning_rate": 9.542267061329407e-06,
|
| 771 |
+
"loss": 0.6535,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.22585992540406133,
|
| 776 |
+
"grad_norm": 2.6756133538389593,
|
| 777 |
+
"learning_rate": 9.52702690185594e-06,
|
| 778 |
+
"loss": 0.6578,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.22793203481143803,
|
| 783 |
+
"grad_norm": 2.7169962232785947,
|
| 784 |
+
"learning_rate": 9.511549751691159e-06,
|
| 785 |
+
"loss": 0.6696,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.23000414421881477,
|
| 790 |
+
"grad_norm": 2.66297533869507,
|
| 791 |
+
"learning_rate": 9.495836421066722e-06,
|
| 792 |
+
"loss": 0.6594,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.23207625362619147,
|
| 797 |
+
"grad_norm": 2.974286886191662,
|
| 798 |
+
"learning_rate": 9.47988773257838e-06,
|
| 799 |
+
"loss": 0.6784,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.23414836303356817,
|
| 804 |
+
"grad_norm": 2.7185575923361243,
|
| 805 |
+
"learning_rate": 9.46370452114291e-06,
|
| 806 |
+
"loss": 0.658,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.23622047244094488,
|
| 811 |
+
"grad_norm": 2.829971742294898,
|
| 812 |
+
"learning_rate": 9.447287633954406e-06,
|
| 813 |
+
"loss": 0.6593,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.23829258184832158,
|
| 818 |
+
"grad_norm": 2.620399538248119,
|
| 819 |
+
"learning_rate": 9.430637930439933e-06,
|
| 820 |
+
"loss": 0.6641,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2403646912556983,
|
| 825 |
+
"grad_norm": 2.8089738469657286,
|
| 826 |
+
"learning_rate": 9.413756282214538e-06,
|
| 827 |
+
"loss": 0.6443,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.24243680066307502,
|
| 832 |
+
"grad_norm": 2.6610020148736657,
|
| 833 |
+
"learning_rate": 9.396643573035609e-06,
|
| 834 |
+
"loss": 0.6619,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.24450891007045172,
|
| 839 |
+
"grad_norm": 2.693009848715325,
|
| 840 |
+
"learning_rate": 9.37930069875662e-06,
|
| 841 |
+
"loss": 0.6469,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.24658101947782843,
|
| 846 |
+
"grad_norm": 2.8338530690883657,
|
| 847 |
+
"learning_rate": 9.36172856728023e-06,
|
| 848 |
+
"loss": 0.6571,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.24865312888520513,
|
| 853 |
+
"grad_norm": 2.676634556349702,
|
| 854 |
+
"learning_rate": 9.343928098510759e-06,
|
| 855 |
+
"loss": 0.6358,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.25072523829258186,
|
| 860 |
+
"grad_norm": 2.6815297814713674,
|
| 861 |
+
"learning_rate": 9.325900224306019e-06,
|
| 862 |
+
"loss": 0.6366,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.25279734769995854,
|
| 867 |
+
"grad_norm": 2.693794960810057,
|
| 868 |
+
"learning_rate": 9.307645888428542e-06,
|
| 869 |
+
"loss": 0.6441,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2548694571073353,
|
| 874 |
+
"grad_norm": 2.6733329911215757,
|
| 875 |
+
"learning_rate": 9.289166046496172e-06,
|
| 876 |
+
"loss": 0.6284,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.25694156651471195,
|
| 881 |
+
"grad_norm": 2.735971515551987,
|
| 882 |
+
"learning_rate": 9.270461665932035e-06,
|
| 883 |
+
"loss": 0.6394,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.2590136759220887,
|
| 888 |
+
"grad_norm": 2.7764753108215623,
|
| 889 |
+
"learning_rate": 9.251533725913893e-06,
|
| 890 |
+
"loss": 0.6308,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.2610857853294654,
|
| 895 |
+
"grad_norm": 2.9846661778059596,
|
| 896 |
+
"learning_rate": 9.23238321732289e-06,
|
| 897 |
+
"loss": 0.6381,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2631578947368421,
|
| 902 |
+
"grad_norm": 2.639475488829137,
|
| 903 |
+
"learning_rate": 9.213011142691672e-06,
|
| 904 |
+
"loss": 0.6298,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.2652300041442188,
|
| 909 |
+
"grad_norm": 2.8031268295287037,
|
| 910 |
+
"learning_rate": 9.193418516151913e-06,
|
| 911 |
+
"loss": 0.6314,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.2673021135515955,
|
| 916 |
+
"grad_norm": 2.751071861402199,
|
| 917 |
+
"learning_rate": 9.173606363381218e-06,
|
| 918 |
+
"loss": 0.6243,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.26937422295897223,
|
| 923 |
+
"grad_norm": 2.8637192589002507,
|
| 924 |
+
"learning_rate": 9.15357572154943e-06,
|
| 925 |
+
"loss": 0.6226,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.27144633236634896,
|
| 930 |
+
"grad_norm": 2.757948609951417,
|
| 931 |
+
"learning_rate": 9.133327639264334e-06,
|
| 932 |
+
"loss": 0.6195,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.27351844177372564,
|
| 937 |
+
"grad_norm": 2.7078393676429724,
|
| 938 |
+
"learning_rate": 9.112863176516761e-06,
|
| 939 |
+
"loss": 0.6063,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.2755905511811024,
|
| 944 |
+
"grad_norm": 2.7174885956763517,
|
| 945 |
+
"learning_rate": 9.092183404625107e-06,
|
| 946 |
+
"loss": 0.6201,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.27766266058847905,
|
| 951 |
+
"grad_norm": 2.692439395697707,
|
| 952 |
+
"learning_rate": 9.071289406179233e-06,
|
| 953 |
+
"loss": 0.6186,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.2797347699958558,
|
| 958 |
+
"grad_norm": 2.705577759760543,
|
| 959 |
+
"learning_rate": 9.0501822749838e-06,
|
| 960 |
+
"loss": 0.6208,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.2818068794032325,
|
| 965 |
+
"grad_norm": 2.6961720306129497,
|
| 966 |
+
"learning_rate": 9.028863116001013e-06,
|
| 967 |
+
"loss": 0.6217,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.2838789888106092,
|
| 972 |
+
"grad_norm": 2.823153530451939,
|
| 973 |
+
"learning_rate": 9.007333045292764e-06,
|
| 974 |
+
"loss": 0.6095,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.2859510982179859,
|
| 979 |
+
"grad_norm": 2.7424489434092885,
|
| 980 |
+
"learning_rate": 8.98559318996222e-06,
|
| 981 |
+
"loss": 0.6071,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.2880232076253626,
|
| 986 |
+
"grad_norm": 2.8108371792200724,
|
| 987 |
+
"learning_rate": 8.963644688094806e-06,
|
| 988 |
+
"loss": 0.6123,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.29009531703273933,
|
| 993 |
+
"grad_norm": 2.7296570733868046,
|
| 994 |
+
"learning_rate": 8.941488688698635e-06,
|
| 995 |
+
"loss": 0.6038,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.29216742644011606,
|
| 1000 |
+
"grad_norm": 2.5282731359900983,
|
| 1001 |
+
"learning_rate": 8.919126351644351e-06,
|
| 1002 |
+
"loss": 0.6051,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.29423953584749274,
|
| 1007 |
+
"grad_norm": 2.733801097585659,
|
| 1008 |
+
"learning_rate": 8.896558847604414e-06,
|
| 1009 |
+
"loss": 0.6169,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.29631164525486947,
|
| 1014 |
+
"grad_norm": 2.5922131116520433,
|
| 1015 |
+
"learning_rate": 8.873787357991811e-06,
|
| 1016 |
+
"loss": 0.6062,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.29838375466224615,
|
| 1021 |
+
"grad_norm": 2.6565071641175844,
|
| 1022 |
+
"learning_rate": 8.850813074898218e-06,
|
| 1023 |
+
"loss": 0.6069,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3004558640696229,
|
| 1028 |
+
"grad_norm": 3.0927970126548154,
|
| 1029 |
+
"learning_rate": 8.827637201031579e-06,
|
| 1030 |
+
"loss": 0.5879,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3025279734769996,
|
| 1035 |
+
"grad_norm": 2.8105831813533033,
|
| 1036 |
+
"learning_rate": 8.804260949653154e-06,
|
| 1037 |
+
"loss": 0.6124,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.3046000828843763,
|
| 1042 |
+
"grad_norm": 2.584377915155431,
|
| 1043 |
+
"learning_rate": 8.780685544514006e-06,
|
| 1044 |
+
"loss": 0.6073,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.306672192291753,
|
| 1049 |
+
"grad_norm": 2.773926231343915,
|
| 1050 |
+
"learning_rate": 8.756912219790933e-06,
|
| 1051 |
+
"loss": 0.5999,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3087443016991297,
|
| 1056 |
+
"grad_norm": 2.732022843788419,
|
| 1057 |
+
"learning_rate": 8.732942220021859e-06,
|
| 1058 |
+
"loss": 0.5762,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.31081641110650643,
|
| 1063 |
+
"grad_norm": 2.595571184744404,
|
| 1064 |
+
"learning_rate": 8.708776800040679e-06,
|
| 1065 |
+
"loss": 0.5846,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.3128885205138831,
|
| 1070 |
+
"grad_norm": 2.6942143484830985,
|
| 1071 |
+
"learning_rate": 8.684417224911579e-06,
|
| 1072 |
+
"loss": 0.6003,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.31496062992125984,
|
| 1077 |
+
"grad_norm": 2.6835198662003155,
|
| 1078 |
+
"learning_rate": 8.659864769862797e-06,
|
| 1079 |
+
"loss": 0.5838,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.31703273932863657,
|
| 1084 |
+
"grad_norm": 2.7778129772748383,
|
| 1085 |
+
"learning_rate": 8.635120720219877e-06,
|
| 1086 |
+
"loss": 0.5794,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.31910484873601325,
|
| 1091 |
+
"grad_norm": 2.6022076775767493,
|
| 1092 |
+
"learning_rate": 8.610186371338364e-06,
|
| 1093 |
+
"loss": 0.586,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.32117695814339,
|
| 1098 |
+
"grad_norm": 2.7328885442816557,
|
| 1099 |
+
"learning_rate": 8.585063028536015e-06,
|
| 1100 |
+
"loss": 0.5987,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.32324906755076666,
|
| 1105 |
+
"grad_norm": 2.635761747619665,
|
| 1106 |
+
"learning_rate": 8.559752007024449e-06,
|
| 1107 |
+
"loss": 0.5859,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3253211769581434,
|
| 1112 |
+
"grad_norm": 2.79153427275607,
|
| 1113 |
+
"learning_rate": 8.534254631840297e-06,
|
| 1114 |
+
"loss": 0.5976,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.3273932863655201,
|
| 1119 |
+
"grad_norm": 2.656496652025873,
|
| 1120 |
+
"learning_rate": 8.50857223777584e-06,
|
| 1121 |
+
"loss": 0.578,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3294653957728968,
|
| 1126 |
+
"grad_norm": 2.7338530782471655,
|
| 1127 |
+
"learning_rate": 8.482706169309139e-06,
|
| 1128 |
+
"loss": 0.5648,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.33153750518027353,
|
| 1133 |
+
"grad_norm": 2.667291889393192,
|
| 1134 |
+
"learning_rate": 8.456657780533633e-06,
|
| 1135 |
+
"loss": 0.5641,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.3336096145876502,
|
| 1140 |
+
"grad_norm": 2.9479902385492562,
|
| 1141 |
+
"learning_rate": 8.430428435087267e-06,
|
| 1142 |
+
"loss": 0.5665,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.33568172399502694,
|
| 1147 |
+
"grad_norm": 2.9408488784356748,
|
| 1148 |
+
"learning_rate": 8.404019506081103e-06,
|
| 1149 |
+
"loss": 0.5834,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.33775383340240367,
|
| 1154 |
+
"grad_norm": 2.707200722617604,
|
| 1155 |
+
"learning_rate": 8.377432376027437e-06,
|
| 1156 |
+
"loss": 0.5756,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.33982594280978035,
|
| 1161 |
+
"grad_norm": 2.70828380166259,
|
| 1162 |
+
"learning_rate": 8.350668436767413e-06,
|
| 1163 |
+
"loss": 0.5686,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3418980522171571,
|
| 1168 |
+
"grad_norm": 2.7745931767789886,
|
| 1169 |
+
"learning_rate": 8.323729089398182e-06,
|
| 1170 |
+
"loss": 0.5521,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.34397016162453375,
|
| 1175 |
+
"grad_norm": 2.685590822904809,
|
| 1176 |
+
"learning_rate": 8.296615744199533e-06,
|
| 1177 |
+
"loss": 0.5707,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.3460422710319105,
|
| 1182 |
+
"grad_norm": 2.6439480580475023,
|
| 1183 |
+
"learning_rate": 8.269329820560074e-06,
|
| 1184 |
+
"loss": 0.549,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3481143804392872,
|
| 1189 |
+
"grad_norm": 2.6684012534208845,
|
| 1190 |
+
"learning_rate": 8.241872746902934e-06,
|
| 1191 |
+
"loss": 0.5614,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.3501864898466639,
|
| 1196 |
+
"grad_norm": 2.6586973407105847,
|
| 1197 |
+
"learning_rate": 8.214245960610966e-06,
|
| 1198 |
+
"loss": 0.5596,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.3522585992540406,
|
| 1203 |
+
"grad_norm": 2.630842021442866,
|
| 1204 |
+
"learning_rate": 8.18645090795152e-06,
|
| 1205 |
+
"loss": 0.5435,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.3543307086614173,
|
| 1210 |
+
"grad_norm": 2.783114608569901,
|
| 1211 |
+
"learning_rate": 8.158489044000712e-06,
|
| 1212 |
+
"loss": 0.554,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.35640281806879404,
|
| 1217 |
+
"grad_norm": 2.6861386548441266,
|
| 1218 |
+
"learning_rate": 8.13036183256727e-06,
|
| 1219 |
+
"loss": 0.5503,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.35847492747617077,
|
| 1224 |
+
"grad_norm": 2.760451069908695,
|
| 1225 |
+
"learning_rate": 8.102070746115888e-06,
|
| 1226 |
+
"loss": 0.5504,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.36054703688354744,
|
| 1231 |
+
"grad_norm": 3.1912807546036155,
|
| 1232 |
+
"learning_rate": 8.073617265690144e-06,
|
| 1233 |
+
"loss": 0.5585,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.3626191462909242,
|
| 1238 |
+
"grad_norm": 2.7551688368780365,
|
| 1239 |
+
"learning_rate": 8.045002880834975e-06,
|
| 1240 |
+
"loss": 0.5499,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.36469125569830085,
|
| 1245 |
+
"grad_norm": 2.656529615625788,
|
| 1246 |
+
"learning_rate": 8.016229089518695e-06,
|
| 1247 |
+
"loss": 0.5472,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.3667633651056776,
|
| 1252 |
+
"grad_norm": 2.456139438562433,
|
| 1253 |
+
"learning_rate": 7.987297398054572e-06,
|
| 1254 |
+
"loss": 0.5444,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.36883547451305426,
|
| 1259 |
+
"grad_norm": 2.7021257640939678,
|
| 1260 |
+
"learning_rate": 7.95820932102198e-06,
|
| 1261 |
+
"loss": 0.5467,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.370907583920431,
|
| 1266 |
+
"grad_norm": 2.576896456231456,
|
| 1267 |
+
"learning_rate": 7.9289663811871e-06,
|
| 1268 |
+
"loss": 0.5453,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.3729796933278077,
|
| 1273 |
+
"grad_norm": 2.509602711427261,
|
| 1274 |
+
"learning_rate": 7.899570109423219e-06,
|
| 1275 |
+
"loss": 0.5315,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.3750518027351844,
|
| 1280 |
+
"grad_norm": 2.7824790919014486,
|
| 1281 |
+
"learning_rate": 7.870022044630569e-06,
|
| 1282 |
+
"loss": 0.5367,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.37712391214256114,
|
| 1287 |
+
"grad_norm": 2.635566041127458,
|
| 1288 |
+
"learning_rate": 7.84032373365578e-06,
|
| 1289 |
+
"loss": 0.5458,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.3791960215499378,
|
| 1294 |
+
"grad_norm": 2.5686297030493277,
|
| 1295 |
+
"learning_rate": 7.810476731210897e-06,
|
| 1296 |
+
"loss": 0.5538,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.38126813095731454,
|
| 1301 |
+
"grad_norm": 2.4885943009976366,
|
| 1302 |
+
"learning_rate": 7.780482599791987e-06,
|
| 1303 |
+
"loss": 0.5501,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.3833402403646913,
|
| 1308 |
+
"grad_norm": 2.6312255974545704,
|
| 1309 |
+
"learning_rate": 7.750342909597353e-06,
|
| 1310 |
+
"loss": 0.5412,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.38541234977206795,
|
| 1315 |
+
"grad_norm": 2.522165622730654,
|
| 1316 |
+
"learning_rate": 7.72005923844532e-06,
|
| 1317 |
+
"loss": 0.5313,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.3874844591794447,
|
| 1322 |
+
"grad_norm": 2.6702491559670167,
|
| 1323 |
+
"learning_rate": 7.689633171691646e-06,
|
| 1324 |
+
"loss": 0.5345,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.38955656858682136,
|
| 1329 |
+
"grad_norm": 2.6631872386085185,
|
| 1330 |
+
"learning_rate": 7.659066302146523e-06,
|
| 1331 |
+
"loss": 0.5452,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.3916286779941981,
|
| 1336 |
+
"grad_norm": 2.5785637853885195,
|
| 1337 |
+
"learning_rate": 7.628360229991198e-06,
|
| 1338 |
+
"loss": 0.5288,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.3937007874015748,
|
| 1343 |
+
"grad_norm": 2.6837857193306935,
|
| 1344 |
+
"learning_rate": 7.597516562694198e-06,
|
| 1345 |
+
"loss": 0.5306,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.3957728968089515,
|
| 1350 |
+
"grad_norm": 2.550858198260864,
|
| 1351 |
+
"learning_rate": 7.56653691492718e-06,
|
| 1352 |
+
"loss": 0.5233,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.39784500621632823,
|
| 1357 |
+
"grad_norm": 2.6115170843406132,
|
| 1358 |
+
"learning_rate": 7.535422908480408e-06,
|
| 1359 |
+
"loss": 0.5424,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.3999171156237049,
|
| 1364 |
+
"grad_norm": 2.5494802625123105,
|
| 1365 |
+
"learning_rate": 7.504176172177842e-06,
|
| 1366 |
+
"loss": 0.5171,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.40198922503108164,
|
| 1371 |
+
"grad_norm": 2.5129023648194284,
|
| 1372 |
+
"learning_rate": 7.472798341791877e-06,
|
| 1373 |
+
"loss": 0.5148,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.4040613344384584,
|
| 1378 |
+
"grad_norm": 2.6166709552589866,
|
| 1379 |
+
"learning_rate": 7.441291059957709e-06,
|
| 1380 |
+
"loss": 0.5292,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.40613344384583505,
|
| 1385 |
+
"grad_norm": 2.5745519804152073,
|
| 1386 |
+
"learning_rate": 7.409655976087338e-06,
|
| 1387 |
+
"loss": 0.5228,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.4082055532532118,
|
| 1392 |
+
"grad_norm": 2.7301378295783643,
|
| 1393 |
+
"learning_rate": 7.377894746283227e-06,
|
| 1394 |
+
"loss": 0.5343,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.41027766266058846,
|
| 1399 |
+
"grad_norm": 2.431218699982397,
|
| 1400 |
+
"learning_rate": 7.3460090332516e-06,
|
| 1401 |
+
"loss": 0.508,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4123497720679652,
|
| 1406 |
+
"grad_norm": 2.556296703014823,
|
| 1407 |
+
"learning_rate": 7.314000506215402e-06,
|
| 1408 |
+
"loss": 0.5148,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4144218814753419,
|
| 1413 |
+
"grad_norm": 2.5152139819517023,
|
| 1414 |
+
"learning_rate": 7.281870840826912e-06,
|
| 1415 |
+
"loss": 0.4999,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.4164939908827186,
|
| 1420 |
+
"grad_norm": 2.786935816025543,
|
| 1421 |
+
"learning_rate": 7.249621719080026e-06,
|
| 1422 |
+
"loss": 0.5177,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.41856610029009533,
|
| 1427 |
+
"grad_norm": 2.5491738667702672,
|
| 1428 |
+
"learning_rate": 7.217254829222201e-06,
|
| 1429 |
+
"loss": 0.5114,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.420638209697472,
|
| 1434 |
+
"grad_norm": 2.50660251762613,
|
| 1435 |
+
"learning_rate": 7.1847718656660755e-06,
|
| 1436 |
+
"loss": 0.5156,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.42271031910484874,
|
| 1441 |
+
"grad_norm": 2.5986079473520878,
|
| 1442 |
+
"learning_rate": 7.152174528900773e-06,
|
| 1443 |
+
"loss": 0.4954,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4247824285122254,
|
| 1448 |
+
"grad_norm": 2.635934213356069,
|
| 1449 |
+
"learning_rate": 7.119464525402867e-06,
|
| 1450 |
+
"loss": 0.504,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.42685453791960215,
|
| 1455 |
+
"grad_norm": 2.861696123619274,
|
| 1456 |
+
"learning_rate": 7.08664356754706e-06,
|
| 1457 |
+
"loss": 0.5063,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.4289266473269789,
|
| 1462 |
+
"grad_norm": 2.5689042563262223,
|
| 1463 |
+
"learning_rate": 7.053713373516538e-06,
|
| 1464 |
+
"loss": 0.5181,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.43099875673435556,
|
| 1469 |
+
"grad_norm": 2.5132849205491286,
|
| 1470 |
+
"learning_rate": 7.020675667213015e-06,
|
| 1471 |
+
"loss": 0.5043,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.4330708661417323,
|
| 1476 |
+
"grad_norm": 2.704397404776844,
|
| 1477 |
+
"learning_rate": 6.987532178166496e-06,
|
| 1478 |
+
"loss": 0.5022,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.43514297554910897,
|
| 1483 |
+
"grad_norm": 2.4785754646781104,
|
| 1484 |
+
"learning_rate": 6.9542846414447306e-06,
|
| 1485 |
+
"loss": 0.5027,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.4372150849564857,
|
| 1490 |
+
"grad_norm": 2.3912935006226594,
|
| 1491 |
+
"learning_rate": 6.920934797562385e-06,
|
| 1492 |
+
"loss": 0.5051,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.43928719436386243,
|
| 1497 |
+
"grad_norm": 2.3476041200753546,
|
| 1498 |
+
"learning_rate": 6.887484392389923e-06,
|
| 1499 |
+
"loss": 0.5043,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4413593037712391,
|
| 1504 |
+
"grad_norm": 2.5914055427224465,
|
| 1505 |
+
"learning_rate": 6.853935177062208e-06,
|
| 1506 |
+
"loss": 0.4974,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.44343141317861584,
|
| 1511 |
+
"grad_norm": 2.594951575952019,
|
| 1512 |
+
"learning_rate": 6.8202889078868395e-06,
|
| 1513 |
+
"loss": 0.5061,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.4455035225859925,
|
| 1518 |
+
"grad_norm": 2.6023391478211293,
|
| 1519 |
+
"learning_rate": 6.786547346252198e-06,
|
| 1520 |
+
"loss": 0.4963,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.44757563199336925,
|
| 1525 |
+
"grad_norm": 2.3528096807317995,
|
| 1526 |
+
"learning_rate": 6.7527122585352435e-06,
|
| 1527 |
+
"loss": 0.4883,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.449647741400746,
|
| 1532 |
+
"grad_norm": 2.7398459124724814,
|
| 1533 |
+
"learning_rate": 6.718785416009044e-06,
|
| 1534 |
+
"loss": 0.4968,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.45171985080812266,
|
| 1539 |
+
"grad_norm": 2.838272166911696,
|
| 1540 |
+
"learning_rate": 6.6847685947500495e-06,
|
| 1541 |
+
"loss": 0.4915,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.4537919602154994,
|
| 1546 |
+
"grad_norm": 3.0284763209341623,
|
| 1547 |
+
"learning_rate": 6.650663575545111e-06,
|
| 1548 |
+
"loss": 0.4762,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.45586406962287607,
|
| 1553 |
+
"grad_norm": 2.6967943018540517,
|
| 1554 |
+
"learning_rate": 6.61647214379826e-06,
|
| 1555 |
+
"loss": 0.4737,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.4579361790302528,
|
| 1560 |
+
"grad_norm": 2.551888108931838,
|
| 1561 |
+
"learning_rate": 6.582196089437241e-06,
|
| 1562 |
+
"loss": 0.5076,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.46000828843762953,
|
| 1567 |
+
"grad_norm": 2.514843118690436,
|
| 1568 |
+
"learning_rate": 6.547837206819804e-06,
|
| 1569 |
+
"loss": 0.4876,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.4620803978450062,
|
| 1574 |
+
"grad_norm": 2.4499641238461005,
|
| 1575 |
+
"learning_rate": 6.513397294639778e-06,
|
| 1576 |
+
"loss": 0.4785,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.46415250725238294,
|
| 1581 |
+
"grad_norm": 2.5439805360709378,
|
| 1582 |
+
"learning_rate": 6.478878155832904e-06,
|
| 1583 |
+
"loss": 0.4609,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.4662246166597596,
|
| 1588 |
+
"grad_norm": 2.567465771254044,
|
| 1589 |
+
"learning_rate": 6.444281597482449e-06,
|
| 1590 |
+
"loss": 0.4826,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.46829672606713635,
|
| 1595 |
+
"grad_norm": 2.4194190626785863,
|
| 1596 |
+
"learning_rate": 6.409609430724607e-06,
|
| 1597 |
+
"loss": 0.4639,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.4703688354745131,
|
| 1602 |
+
"grad_norm": 2.4928927116754216,
|
| 1603 |
+
"learning_rate": 6.3748634706536905e-06,
|
| 1604 |
+
"loss": 0.4755,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.47244094488188976,
|
| 1609 |
+
"grad_norm": 2.5328967096855517,
|
| 1610 |
+
"learning_rate": 6.340045536227101e-06,
|
| 1611 |
+
"loss": 0.4676,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.4745130542892665,
|
| 1616 |
+
"grad_norm": 2.580851369974543,
|
| 1617 |
+
"learning_rate": 6.305157450170112e-06,
|
| 1618 |
+
"loss": 0.4679,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.47658516369664317,
|
| 1623 |
+
"grad_norm": 2.771781788352758,
|
| 1624 |
+
"learning_rate": 6.270201038880451e-06,
|
| 1625 |
+
"loss": 0.4748,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.4786572731040199,
|
| 1630 |
+
"grad_norm": 2.3357211287270294,
|
| 1631 |
+
"learning_rate": 6.235178132332678e-06,
|
| 1632 |
+
"loss": 0.4733,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.4807293825113966,
|
| 1637 |
+
"grad_norm": 2.4376737050512993,
|
| 1638 |
+
"learning_rate": 6.200090563982397e-06,
|
| 1639 |
+
"loss": 0.4623,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.4828014919187733,
|
| 1644 |
+
"grad_norm": 2.727521949059151,
|
| 1645 |
+
"learning_rate": 6.164940170670266e-06,
|
| 1646 |
+
"loss": 0.4763,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.48487360132615004,
|
| 1651 |
+
"grad_norm": 2.485190311900293,
|
| 1652 |
+
"learning_rate": 6.129728792525847e-06,
|
| 1653 |
+
"loss": 0.4653,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.4869457107335267,
|
| 1658 |
+
"grad_norm": 2.3795187623979746,
|
| 1659 |
+
"learning_rate": 6.094458272871259e-06,
|
| 1660 |
+
"loss": 0.4576,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.48901782014090345,
|
| 1665 |
+
"grad_norm": 2.43415537867363,
|
| 1666 |
+
"learning_rate": 6.0591304581247005e-06,
|
| 1667 |
+
"loss": 0.4606,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.4910899295482801,
|
| 1672 |
+
"grad_norm": 2.5114902879972374,
|
| 1673 |
+
"learning_rate": 6.023747197703771e-06,
|
| 1674 |
+
"loss": 0.4671,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.49316203895565686,
|
| 1679 |
+
"grad_norm": 2.5122364778542225,
|
| 1680 |
+
"learning_rate": 5.988310343928665e-06,
|
| 1681 |
+
"loss": 0.4678,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.4952341483630336,
|
| 1686 |
+
"grad_norm": 2.6311201696718283,
|
| 1687 |
+
"learning_rate": 5.9528217519252e-06,
|
| 1688 |
+
"loss": 0.4653,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.49730625777041026,
|
| 1693 |
+
"grad_norm": 2.634373162867216,
|
| 1694 |
+
"learning_rate": 5.9172832795276965e-06,
|
| 1695 |
+
"loss": 0.4858,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.499378367177787,
|
| 1700 |
+
"grad_norm": 2.602845367016974,
|
| 1701 |
+
"learning_rate": 5.881696787181724e-06,
|
| 1702 |
+
"loss": 0.4646,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5014504765851637,
|
| 1707 |
+
"grad_norm": 2.4997645055614544,
|
| 1708 |
+
"learning_rate": 5.846064137846704e-06,
|
| 1709 |
+
"loss": 0.4723,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5035225859925404,
|
| 1714 |
+
"grad_norm": 2.5259484476620853,
|
| 1715 |
+
"learning_rate": 5.810387196898387e-06,
|
| 1716 |
+
"loss": 0.4592,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5055946953999171,
|
| 1721 |
+
"grad_norm": 2.460185161862939,
|
| 1722 |
+
"learning_rate": 5.7746678320311955e-06,
|
| 1723 |
+
"loss": 0.4563,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5076668048072939,
|
| 1728 |
+
"grad_norm": 2.522924782621583,
|
| 1729 |
+
"learning_rate": 5.738907913160452e-06,
|
| 1730 |
+
"loss": 0.455,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5097389142146705,
|
| 1735 |
+
"grad_norm": 2.450518901210104,
|
| 1736 |
+
"learning_rate": 5.703109312324493e-06,
|
| 1737 |
+
"loss": 0.4631,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5118110236220472,
|
| 1742 |
+
"grad_norm": 2.512677663120476,
|
| 1743 |
+
"learning_rate": 5.667273903586656e-06,
|
| 1744 |
+
"loss": 0.4636,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5138831330294239,
|
| 1749 |
+
"grad_norm": 2.508483790573203,
|
| 1750 |
+
"learning_rate": 5.6314035629371835e-06,
|
| 1751 |
+
"loss": 0.4494,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5159552424368007,
|
| 1756 |
+
"grad_norm": 2.539619573950297,
|
| 1757 |
+
"learning_rate": 5.595500168195007e-06,
|
| 1758 |
+
"loss": 0.4657,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5180273518441774,
|
| 1763 |
+
"grad_norm": 2.457760045030727,
|
| 1764 |
+
"learning_rate": 5.5595655989094525e-06,
|
| 1765 |
+
"loss": 0.4562,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.520099461251554,
|
| 1770 |
+
"grad_norm": 2.5810506702287546,
|
| 1771 |
+
"learning_rate": 5.52360173626183e-06,
|
| 1772 |
+
"loss": 0.4587,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5221715706589308,
|
| 1777 |
+
"grad_norm": 2.533803600220524,
|
| 1778 |
+
"learning_rate": 5.487610462966969e-06,
|
| 1779 |
+
"loss": 0.4473,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5242436800663075,
|
| 1784 |
+
"grad_norm": 2.408501431878237,
|
| 1785 |
+
"learning_rate": 5.451593663174647e-06,
|
| 1786 |
+
"loss": 0.4466,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5263157894736842,
|
| 1791 |
+
"grad_norm": 2.5248933210116427,
|
| 1792 |
+
"learning_rate": 5.4155532223709625e-06,
|
| 1793 |
+
"loss": 0.4427,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.528387898881061,
|
| 1798 |
+
"grad_norm": 2.5355682898170704,
|
| 1799 |
+
"learning_rate": 5.379491027279622e-06,
|
| 1800 |
+
"loss": 0.4624,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5304600082884376,
|
| 1805 |
+
"grad_norm": 2.7849780348488156,
|
| 1806 |
+
"learning_rate": 5.343408965763174e-06,
|
| 1807 |
+
"loss": 0.4487,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5325321176958143,
|
| 1812 |
+
"grad_norm": 2.401434318103116,
|
| 1813 |
+
"learning_rate": 5.3073089267241805e-06,
|
| 1814 |
+
"loss": 0.4393,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.534604227103191,
|
| 1819 |
+
"grad_norm": 2.4324353094703475,
|
| 1820 |
+
"learning_rate": 5.271192800006325e-06,
|
| 1821 |
+
"loss": 0.4405,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5366763365105678,
|
| 1826 |
+
"grad_norm": 2.5008934086234027,
|
| 1827 |
+
"learning_rate": 5.235062476295488e-06,
|
| 1828 |
+
"loss": 0.4206,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5387484459179445,
|
| 1833 |
+
"grad_norm": 2.523177697285061,
|
| 1834 |
+
"learning_rate": 5.198919847020765e-06,
|
| 1835 |
+
"loss": 0.4378,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5408205553253211,
|
| 1840 |
+
"grad_norm": 2.561504860219997,
|
| 1841 |
+
"learning_rate": 5.162766804255446e-06,
|
| 1842 |
+
"loss": 0.4369,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.5428926647326979,
|
| 1847 |
+
"grad_norm": 2.4342748058160457,
|
| 1848 |
+
"learning_rate": 5.1266052406179755e-06,
|
| 1849 |
+
"loss": 0.4429,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.5449647741400746,
|
| 1854 |
+
"grad_norm": 2.5003051526668054,
|
| 1855 |
+
"learning_rate": 5.090437049172861e-06,
|
| 1856 |
+
"loss": 0.4479,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.5470368835474513,
|
| 1861 |
+
"grad_norm": 2.41044724225261,
|
| 1862 |
+
"learning_rate": 5.054264123331583e-06,
|
| 1863 |
+
"loss": 0.4348,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.5491089929548281,
|
| 1868 |
+
"grad_norm": 2.534587174570914,
|
| 1869 |
+
"learning_rate": 5.018088356753463e-06,
|
| 1870 |
+
"loss": 0.4346,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.5511811023622047,
|
| 1875 |
+
"grad_norm": 2.3874656931988953,
|
| 1876 |
+
"learning_rate": 4.981911643246539e-06,
|
| 1877 |
+
"loss": 0.4463,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.5532532117695814,
|
| 1882 |
+
"grad_norm": 2.344775548708239,
|
| 1883 |
+
"learning_rate": 4.9457358766684175e-06,
|
| 1884 |
+
"loss": 0.4323,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.5553253211769581,
|
| 1889 |
+
"grad_norm": 2.394852642174165,
|
| 1890 |
+
"learning_rate": 4.9095629508271396e-06,
|
| 1891 |
+
"loss": 0.412,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.5573974305843349,
|
| 1896 |
+
"grad_norm": 2.4107260444211067,
|
| 1897 |
+
"learning_rate": 4.873394759382025e-06,
|
| 1898 |
+
"loss": 0.434,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.5594695399917116,
|
| 1903 |
+
"grad_norm": 2.39603695241891,
|
| 1904 |
+
"learning_rate": 4.837233195744556e-06,
|
| 1905 |
+
"loss": 0.4368,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.5615416493990882,
|
| 1910 |
+
"grad_norm": 2.484288425740336,
|
| 1911 |
+
"learning_rate": 4.8010801529792375e-06,
|
| 1912 |
+
"loss": 0.4247,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.563613758806465,
|
| 1917 |
+
"grad_norm": 2.491968721281532,
|
| 1918 |
+
"learning_rate": 4.7649375237045135e-06,
|
| 1919 |
+
"loss": 0.4239,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.5656858682138417,
|
| 1924 |
+
"grad_norm": 2.405464453263017,
|
| 1925 |
+
"learning_rate": 4.728807199993677e-06,
|
| 1926 |
+
"loss": 0.4156,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.5677579776212184,
|
| 1931 |
+
"grad_norm": 2.3649504461482405,
|
| 1932 |
+
"learning_rate": 4.692691073275822e-06,
|
| 1933 |
+
"loss": 0.4272,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.5698300870285951,
|
| 1938 |
+
"grad_norm": 2.578097130955108,
|
| 1939 |
+
"learning_rate": 4.656591034236827e-06,
|
| 1940 |
+
"loss": 0.4318,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.5719021964359718,
|
| 1945 |
+
"grad_norm": 2.3565047434019233,
|
| 1946 |
+
"learning_rate": 4.620508972720379e-06,
|
| 1947 |
+
"loss": 0.4157,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.5739743058433485,
|
| 1952 |
+
"grad_norm": 2.4914369771154776,
|
| 1953 |
+
"learning_rate": 4.584446777629038e-06,
|
| 1954 |
+
"loss": 0.4131,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.5760464152507252,
|
| 1959 |
+
"grad_norm": 2.5403254221008016,
|
| 1960 |
+
"learning_rate": 4.548406336825354e-06,
|
| 1961 |
+
"loss": 0.4209,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.578118524658102,
|
| 1966 |
+
"grad_norm": 2.4433869039162763,
|
| 1967 |
+
"learning_rate": 4.512389537033032e-06,
|
| 1968 |
+
"loss": 0.4156,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.5801906340654787,
|
| 1973 |
+
"grad_norm": 2.5532014778576606,
|
| 1974 |
+
"learning_rate": 4.476398263738171e-06,
|
| 1975 |
+
"loss": 0.4187,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.5822627434728553,
|
| 1980 |
+
"grad_norm": 2.422240802637448,
|
| 1981 |
+
"learning_rate": 4.440434401090549e-06,
|
| 1982 |
+
"loss": 0.4123,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.5843348528802321,
|
| 1987 |
+
"grad_norm": 2.327309317585392,
|
| 1988 |
+
"learning_rate": 4.404499831804993e-06,
|
| 1989 |
+
"loss": 0.4167,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.5864069622876088,
|
| 1994 |
+
"grad_norm": 2.337179254235494,
|
| 1995 |
+
"learning_rate": 4.368596437062819e-06,
|
| 1996 |
+
"loss": 0.4253,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.5884790716949855,
|
| 2001 |
+
"grad_norm": 2.6203872163413,
|
| 2002 |
+
"learning_rate": 4.332726096413346e-06,
|
| 2003 |
+
"loss": 0.4035,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.5905511811023622,
|
| 2008 |
+
"grad_norm": 2.45290146849787,
|
| 2009 |
+
"learning_rate": 4.29689068767551e-06,
|
| 2010 |
+
"loss": 0.4205,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.5926232905097389,
|
| 2015 |
+
"grad_norm": 2.442505011249494,
|
| 2016 |
+
"learning_rate": 4.261092086839549e-06,
|
| 2017 |
+
"loss": 0.4199,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.5946953999171156,
|
| 2022 |
+
"grad_norm": 2.469896647219146,
|
| 2023 |
+
"learning_rate": 4.225332167968808e-06,
|
| 2024 |
+
"loss": 0.4197,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.5967675093244923,
|
| 2029 |
+
"grad_norm": 2.525933254390451,
|
| 2030 |
+
"learning_rate": 4.189612803101614e-06,
|
| 2031 |
+
"loss": 0.4136,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.5988396187318691,
|
| 2036 |
+
"grad_norm": 2.3481926369028154,
|
| 2037 |
+
"learning_rate": 4.153935862153299e-06,
|
| 2038 |
+
"loss": 0.4098,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6009117281392458,
|
| 2043 |
+
"grad_norm": 2.4652012656554336,
|
| 2044 |
+
"learning_rate": 4.118303212818277e-06,
|
| 2045 |
+
"loss": 0.4111,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6029838375466224,
|
| 2050 |
+
"grad_norm": 2.597755453602896,
|
| 2051 |
+
"learning_rate": 4.082716720472304e-06,
|
| 2052 |
+
"loss": 0.4141,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6050559469539992,
|
| 2057 |
+
"grad_norm": 2.3332725702199912,
|
| 2058 |
+
"learning_rate": 4.0471782480748005e-06,
|
| 2059 |
+
"loss": 0.3886,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6071280563613759,
|
| 2064 |
+
"grad_norm": 2.387505345147473,
|
| 2065 |
+
"learning_rate": 4.011689656071334e-06,
|
| 2066 |
+
"loss": 0.4167,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6092001657687526,
|
| 2071 |
+
"grad_norm": 2.4113537140405388,
|
| 2072 |
+
"learning_rate": 3.97625280229623e-06,
|
| 2073 |
+
"loss": 0.4002,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6112722751761293,
|
| 2078 |
+
"grad_norm": 2.4554203658965488,
|
| 2079 |
+
"learning_rate": 3.940869541875301e-06,
|
| 2080 |
+
"loss": 0.3881,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.613344384583506,
|
| 2085 |
+
"grad_norm": 2.425914982917716,
|
| 2086 |
+
"learning_rate": 3.905541727128743e-06,
|
| 2087 |
+
"loss": 0.4069,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6154164939908827,
|
| 2092 |
+
"grad_norm": 2.5231106398066476,
|
| 2093 |
+
"learning_rate": 3.870271207474154e-06,
|
| 2094 |
+
"loss": 0.4002,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6174886033982594,
|
| 2099 |
+
"grad_norm": 2.4689338483413135,
|
| 2100 |
+
"learning_rate": 3.8350598293297345e-06,
|
| 2101 |
+
"loss": 0.4141,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6195607128056362,
|
| 2106 |
+
"grad_norm": 2.496046599900669,
|
| 2107 |
+
"learning_rate": 3.7999094360176036e-06,
|
| 2108 |
+
"loss": 0.3965,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6216328222130129,
|
| 2113 |
+
"grad_norm": 2.3726201269683553,
|
| 2114 |
+
"learning_rate": 3.7648218676673232e-06,
|
| 2115 |
+
"loss": 0.4017,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6237049316203895,
|
| 2120 |
+
"grad_norm": 2.418446591018178,
|
| 2121 |
+
"learning_rate": 3.7297989611195504e-06,
|
| 2122 |
+
"loss": 0.3938,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6257770410277662,
|
| 2127 |
+
"grad_norm": 2.4157452764623963,
|
| 2128 |
+
"learning_rate": 3.694842549829889e-06,
|
| 2129 |
+
"loss": 0.3871,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.627849150435143,
|
| 2134 |
+
"grad_norm": 2.4088293626826114,
|
| 2135 |
+
"learning_rate": 3.659954463772901e-06,
|
| 2136 |
+
"loss": 0.4002,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.6299212598425197,
|
| 2141 |
+
"grad_norm": 2.5779980646499543,
|
| 2142 |
+
"learning_rate": 3.625136529346312e-06,
|
| 2143 |
+
"loss": 0.4055,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6319933692498964,
|
| 2148 |
+
"grad_norm": 2.4110720231898806,
|
| 2149 |
+
"learning_rate": 3.590390569275395e-06,
|
| 2150 |
+
"loss": 0.3913,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6340654786572731,
|
| 2155 |
+
"grad_norm": 2.3982253161569234,
|
| 2156 |
+
"learning_rate": 3.555718402517554e-06,
|
| 2157 |
+
"loss": 0.3962,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.6361375880646498,
|
| 2162 |
+
"grad_norm": 2.392216618491618,
|
| 2163 |
+
"learning_rate": 3.521121844167098e-06,
|
| 2164 |
+
"loss": 0.399,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.6382096974720265,
|
| 2169 |
+
"grad_norm": 2.3550943907800193,
|
| 2170 |
+
"learning_rate": 3.486602705360224e-06,
|
| 2171 |
+
"loss": 0.3927,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.6402818068794033,
|
| 2176 |
+
"grad_norm": 2.4744483911955153,
|
| 2177 |
+
"learning_rate": 3.4521627931801976e-06,
|
| 2178 |
+
"loss": 0.3961,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.64235391628678,
|
| 2183 |
+
"grad_norm": 2.416762173736933,
|
| 2184 |
+
"learning_rate": 3.41780391056276e-06,
|
| 2185 |
+
"loss": 0.3959,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.6444260256941566,
|
| 2190 |
+
"grad_norm": 2.3621515478388906,
|
| 2191 |
+
"learning_rate": 3.3835278562017405e-06,
|
| 2192 |
+
"loss": 0.3889,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.6464981351015333,
|
| 2197 |
+
"grad_norm": 2.3799558311767917,
|
| 2198 |
+
"learning_rate": 3.349336424454889e-06,
|
| 2199 |
+
"loss": 0.395,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.6485702445089101,
|
| 2204 |
+
"grad_norm": 2.4290023388512143,
|
| 2205 |
+
"learning_rate": 3.3152314052499513e-06,
|
| 2206 |
+
"loss": 0.3921,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.6506423539162868,
|
| 2211 |
+
"grad_norm": 2.386844861650698,
|
| 2212 |
+
"learning_rate": 3.2812145839909566e-06,
|
| 2213 |
+
"loss": 0.382,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.6527144633236635,
|
| 2218 |
+
"grad_norm": 2.58783476073575,
|
| 2219 |
+
"learning_rate": 3.247287741464758e-06,
|
| 2220 |
+
"loss": 0.3961,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.6547865727310402,
|
| 2225 |
+
"grad_norm": 2.4711795095903906,
|
| 2226 |
+
"learning_rate": 3.2134526537478034e-06,
|
| 2227 |
+
"loss": 0.403,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.6568586821384169,
|
| 2232 |
+
"grad_norm": 2.445518124608817,
|
| 2233 |
+
"learning_rate": 3.1797110921131626e-06,
|
| 2234 |
+
"loss": 0.3949,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.6589307915457936,
|
| 2239 |
+
"grad_norm": 2.3302808051599957,
|
| 2240 |
+
"learning_rate": 3.1460648229377933e-06,
|
| 2241 |
+
"loss": 0.4003,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.6610029009531704,
|
| 2246 |
+
"grad_norm": 2.4700969540161086,
|
| 2247 |
+
"learning_rate": 3.1125156076100804e-06,
|
| 2248 |
+
"loss": 0.3845,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.6630750103605471,
|
| 2253 |
+
"grad_norm": 2.2736405570858684,
|
| 2254 |
+
"learning_rate": 3.0790652024376163e-06,
|
| 2255 |
+
"loss": 0.3755,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.6651471197679237,
|
| 2260 |
+
"grad_norm": 2.3793447678069732,
|
| 2261 |
+
"learning_rate": 3.0457153585552724e-06,
|
| 2262 |
+
"loss": 0.3764,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.6672192291753004,
|
| 2267 |
+
"grad_norm": 2.381350136652911,
|
| 2268 |
+
"learning_rate": 3.012467821833506e-06,
|
| 2269 |
+
"loss": 0.388,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.6692913385826772,
|
| 2274 |
+
"grad_norm": 2.440306240554211,
|
| 2275 |
+
"learning_rate": 2.979324332786987e-06,
|
| 2276 |
+
"loss": 0.3914,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.6713634479900539,
|
| 2281 |
+
"grad_norm": 2.473851856313499,
|
| 2282 |
+
"learning_rate": 2.946286626483463e-06,
|
| 2283 |
+
"loss": 0.3844,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.6734355573974306,
|
| 2288 |
+
"grad_norm": 2.266055102817957,
|
| 2289 |
+
"learning_rate": 2.913356432452942e-06,
|
| 2290 |
+
"loss": 0.3789,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.6755076668048073,
|
| 2295 |
+
"grad_norm": 2.4641402988030126,
|
| 2296 |
+
"learning_rate": 2.8805354745971336e-06,
|
| 2297 |
+
"loss": 0.37,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.677579776212184,
|
| 2302 |
+
"grad_norm": 2.3943294283961984,
|
| 2303 |
+
"learning_rate": 2.847825471099227e-06,
|
| 2304 |
+
"loss": 0.3777,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.6796518856195607,
|
| 2309 |
+
"grad_norm": 2.384842311722,
|
| 2310 |
+
"learning_rate": 2.815228134333925e-06,
|
| 2311 |
+
"loss": 0.382,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.6817239950269374,
|
| 2316 |
+
"grad_norm": 2.469151674287745,
|
| 2317 |
+
"learning_rate": 2.782745170777801e-06,
|
| 2318 |
+
"loss": 0.3704,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.6837961044343142,
|
| 2323 |
+
"grad_norm": 2.583143124954603,
|
| 2324 |
+
"learning_rate": 2.750378280919975e-06,
|
| 2325 |
+
"loss": 0.3736,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.6858682138416908,
|
| 2330 |
+
"grad_norm": 2.4321963843177947,
|
| 2331 |
+
"learning_rate": 2.7181291591730885e-06,
|
| 2332 |
+
"loss": 0.3782,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.6879403232490675,
|
| 2337 |
+
"grad_norm": 2.3531730582824975,
|
| 2338 |
+
"learning_rate": 2.6859994937846002e-06,
|
| 2339 |
+
"loss": 0.376,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.6900124326564443,
|
| 2344 |
+
"grad_norm": 2.4367688917720485,
|
| 2345 |
+
"learning_rate": 2.653990966748401e-06,
|
| 2346 |
+
"loss": 0.377,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.692084542063821,
|
| 2351 |
+
"grad_norm": 2.3875592466498605,
|
| 2352 |
+
"learning_rate": 2.622105253716774e-06,
|
| 2353 |
+
"loss": 0.3772,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.6941566514711976,
|
| 2358 |
+
"grad_norm": 2.424175019670591,
|
| 2359 |
+
"learning_rate": 2.5903440239126633e-06,
|
| 2360 |
+
"loss": 0.3762,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.6962287608785744,
|
| 2365 |
+
"grad_norm": 2.337415256255787,
|
| 2366 |
+
"learning_rate": 2.5587089400422936e-06,
|
| 2367 |
+
"loss": 0.3906,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.6983008702859511,
|
| 2372 |
+
"grad_norm": 2.384059902937286,
|
| 2373 |
+
"learning_rate": 2.5272016582081236e-06,
|
| 2374 |
+
"loss": 0.3661,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7003729796933278,
|
| 2379 |
+
"grad_norm": 2.346884231086367,
|
| 2380 |
+
"learning_rate": 2.4958238278221603e-06,
|
| 2381 |
+
"loss": 0.3632,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7024450891007045,
|
| 2386 |
+
"grad_norm": 2.3658711079290264,
|
| 2387 |
+
"learning_rate": 2.464577091519594e-06,
|
| 2388 |
+
"loss": 0.3695,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7045171985080813,
|
| 2393 |
+
"grad_norm": 2.3082126926984277,
|
| 2394 |
+
"learning_rate": 2.43346308507282e-06,
|
| 2395 |
+
"loss": 0.3697,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7065893079154579,
|
| 2400 |
+
"grad_norm": 2.2724590255115324,
|
| 2401 |
+
"learning_rate": 2.4024834373058024e-06,
|
| 2402 |
+
"loss": 0.3708,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7086614173228346,
|
| 2407 |
+
"grad_norm": 2.4685609897195633,
|
| 2408 |
+
"learning_rate": 2.371639770008804e-06,
|
| 2409 |
+
"loss": 0.3589,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7107335267302114,
|
| 2414 |
+
"grad_norm": 2.4411171519085704,
|
| 2415 |
+
"learning_rate": 2.3409336978534785e-06,
|
| 2416 |
+
"loss": 0.3657,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7128056361375881,
|
| 2421 |
+
"grad_norm": 2.3755249742291227,
|
| 2422 |
+
"learning_rate": 2.3103668283083563e-06,
|
| 2423 |
+
"loss": 0.3688,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7148777455449647,
|
| 2428 |
+
"grad_norm": 2.3679474209669307,
|
| 2429 |
+
"learning_rate": 2.2799407615546816e-06,
|
| 2430 |
+
"loss": 0.3738,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7169498549523415,
|
| 2435 |
+
"grad_norm": 2.351814749989449,
|
| 2436 |
+
"learning_rate": 2.2496570904026484e-06,
|
| 2437 |
+
"loss": 0.3647,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7190219643597182,
|
| 2442 |
+
"grad_norm": 2.4116114120516103,
|
| 2443 |
+
"learning_rate": 2.219517400208015e-06,
|
| 2444 |
+
"loss": 0.3736,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7210940737670949,
|
| 2449 |
+
"grad_norm": 2.3864727744957412,
|
| 2450 |
+
"learning_rate": 2.1895232687891044e-06,
|
| 2451 |
+
"loss": 0.3484,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.7231661831744716,
|
| 2456 |
+
"grad_norm": 2.2512303655303874,
|
| 2457 |
+
"learning_rate": 2.159676266344222e-06,
|
| 2458 |
+
"loss": 0.3599,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.7252382925818484,
|
| 2463 |
+
"grad_norm": 2.2576815601903872,
|
| 2464 |
+
"learning_rate": 2.1299779553694323e-06,
|
| 2465 |
+
"loss": 0.358,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.727310401989225,
|
| 2470 |
+
"grad_norm": 2.4261849484523594,
|
| 2471 |
+
"learning_rate": 2.100429890576782e-06,
|
| 2472 |
+
"loss": 0.3638,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.7293825113966017,
|
| 2477 |
+
"grad_norm": 2.3998781669660008,
|
| 2478 |
+
"learning_rate": 2.0710336188129e-06,
|
| 2479 |
+
"loss": 0.3626,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.7314546208039785,
|
| 2484 |
+
"grad_norm": 2.2949968092513697,
|
| 2485 |
+
"learning_rate": 2.0417906789780236e-06,
|
| 2486 |
+
"loss": 0.3556,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.7335267302113552,
|
| 2491 |
+
"grad_norm": 2.3439568873845222,
|
| 2492 |
+
"learning_rate": 2.0127026019454305e-06,
|
| 2493 |
+
"loss": 0.3753,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.7355988396187318,
|
| 2498 |
+
"grad_norm": 2.2377793345094896,
|
| 2499 |
+
"learning_rate": 1.9837709104813075e-06,
|
| 2500 |
+
"loss": 0.358,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.7376709490261085,
|
| 2505 |
+
"grad_norm": 2.2885196165888653,
|
| 2506 |
+
"learning_rate": 1.9549971191650263e-06,
|
| 2507 |
+
"loss": 0.3515,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.7397430584334853,
|
| 2512 |
+
"grad_norm": 2.558424767409142,
|
| 2513 |
+
"learning_rate": 1.9263827343098596e-06,
|
| 2514 |
+
"loss": 0.3657,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.741815167840862,
|
| 2519 |
+
"grad_norm": 2.44623978739584,
|
| 2520 |
+
"learning_rate": 1.8979292538841133e-06,
|
| 2521 |
+
"loss": 0.3623,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.7438872772482387,
|
| 2526 |
+
"grad_norm": 2.2614597017692692,
|
| 2527 |
+
"learning_rate": 1.8696381674327308e-06,
|
| 2528 |
+
"loss": 0.3553,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.7459593866556155,
|
| 2533 |
+
"grad_norm": 2.2881907635344576,
|
| 2534 |
+
"learning_rate": 1.8415109559992883e-06,
|
| 2535 |
+
"loss": 0.3531,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.7480314960629921,
|
| 2540 |
+
"grad_norm": 2.463206144892447,
|
| 2541 |
+
"learning_rate": 1.8135490920484832e-06,
|
| 2542 |
+
"loss": 0.3559,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.7501036054703688,
|
| 2547 |
+
"grad_norm": 2.3074379267386766,
|
| 2548 |
+
"learning_rate": 1.7857540393890337e-06,
|
| 2549 |
+
"loss": 0.3544,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.7521757148777456,
|
| 2554 |
+
"grad_norm": 2.345951065809859,
|
| 2555 |
+
"learning_rate": 1.7581272530970666e-06,
|
| 2556 |
+
"loss": 0.3495,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.7542478242851223,
|
| 2561 |
+
"grad_norm": 2.436580121963066,
|
| 2562 |
+
"learning_rate": 1.7306701794399266e-06,
|
| 2563 |
+
"loss": 0.351,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.756319933692499,
|
| 2568 |
+
"grad_norm": 2.477449269541472,
|
| 2569 |
+
"learning_rate": 1.7033842558004692e-06,
|
| 2570 |
+
"loss": 0.357,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.7583920430998756,
|
| 2575 |
+
"grad_norm": 2.336059640702188,
|
| 2576 |
+
"learning_rate": 1.6762709106018193e-06,
|
| 2577 |
+
"loss": 0.3566,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.7604641525072524,
|
| 2582 |
+
"grad_norm": 2.327604564240487,
|
| 2583 |
+
"learning_rate": 1.6493315632325873e-06,
|
| 2584 |
+
"loss": 0.3693,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.7625362619146291,
|
| 2589 |
+
"grad_norm": 2.4521027562031805,
|
| 2590 |
+
"learning_rate": 1.6225676239725663e-06,
|
| 2591 |
+
"loss": 0.3557,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7646083713220058,
|
| 2596 |
+
"grad_norm": 2.3928441636264055,
|
| 2597 |
+
"learning_rate": 1.5959804939188962e-06,
|
| 2598 |
+
"loss": 0.35,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.7666804807293826,
|
| 2603 |
+
"grad_norm": 2.3515743827298263,
|
| 2604 |
+
"learning_rate": 1.5695715649127347e-06,
|
| 2605 |
+
"loss": 0.3597,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.7687525901367592,
|
| 2610 |
+
"grad_norm": 2.4659862823505843,
|
| 2611 |
+
"learning_rate": 1.5433422194663694e-06,
|
| 2612 |
+
"loss": 0.3544,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.7708246995441359,
|
| 2617 |
+
"grad_norm": 2.3479456978647684,
|
| 2618 |
+
"learning_rate": 1.5172938306908624e-06,
|
| 2619 |
+
"loss": 0.3479,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.7728968089515127,
|
| 2624 |
+
"grad_norm": 2.491588705252418,
|
| 2625 |
+
"learning_rate": 1.4914277622241596e-06,
|
| 2626 |
+
"loss": 0.348,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.7749689183588894,
|
| 2631 |
+
"grad_norm": 2.2957936532852026,
|
| 2632 |
+
"learning_rate": 1.4657453681597056e-06,
|
| 2633 |
+
"loss": 0.3535,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.777041027766266,
|
| 2638 |
+
"grad_norm": 2.2660578633915787,
|
| 2639 |
+
"learning_rate": 1.440247992975553e-06,
|
| 2640 |
+
"loss": 0.3503,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.7791131371736427,
|
| 2645 |
+
"grad_norm": 2.4214309256458755,
|
| 2646 |
+
"learning_rate": 1.4149369714639856e-06,
|
| 2647 |
+
"loss": 0.3525,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7811852465810195,
|
| 2652 |
+
"grad_norm": 2.378191297196036,
|
| 2653 |
+
"learning_rate": 1.3898136286616364e-06,
|
| 2654 |
+
"loss": 0.3449,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.7832573559883962,
|
| 2659 |
+
"grad_norm": 2.369189696994275,
|
| 2660 |
+
"learning_rate": 1.3648792797801264e-06,
|
| 2661 |
+
"loss": 0.3411,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.7853294653957729,
|
| 2666 |
+
"grad_norm": 2.357227748595689,
|
| 2667 |
+
"learning_rate": 1.3401352301372039e-06,
|
| 2668 |
+
"loss": 0.3428,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.7874015748031497,
|
| 2673 |
+
"grad_norm": 2.283274818402107,
|
| 2674 |
+
"learning_rate": 1.315582775088421e-06,
|
| 2675 |
+
"loss": 0.3476,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.7894736842105263,
|
| 2680 |
+
"grad_norm": 2.443389423960905,
|
| 2681 |
+
"learning_rate": 1.2912231999593222e-06,
|
| 2682 |
+
"loss": 0.3456,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.791545793617903,
|
| 2687 |
+
"grad_norm": 2.3098901374263785,
|
| 2688 |
+
"learning_rate": 1.267057779978143e-06,
|
| 2689 |
+
"loss": 0.3387,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.7936179030252797,
|
| 2694 |
+
"grad_norm": 2.3606277283639403,
|
| 2695 |
+
"learning_rate": 1.2430877802090674e-06,
|
| 2696 |
+
"loss": 0.3505,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.7956900124326565,
|
| 2701 |
+
"grad_norm": 2.778705490024244,
|
| 2702 |
+
"learning_rate": 1.2193144554859938e-06,
|
| 2703 |
+
"loss": 0.334,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7977621218400331,
|
| 2708 |
+
"grad_norm": 2.3112033884634418,
|
| 2709 |
+
"learning_rate": 1.195739050346848e-06,
|
| 2710 |
+
"loss": 0.349,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.7998342312474098,
|
| 2715 |
+
"grad_norm": 2.3575714411861517,
|
| 2716 |
+
"learning_rate": 1.172362798968424e-06,
|
| 2717 |
+
"loss": 0.3449,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8019063406547866,
|
| 2722 |
+
"grad_norm": 2.491996861817634,
|
| 2723 |
+
"learning_rate": 1.1491869251017833e-06,
|
| 2724 |
+
"loss": 0.3414,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8039784500621633,
|
| 2729 |
+
"grad_norm": 2.3991288004159883,
|
| 2730 |
+
"learning_rate": 1.1262126420081887e-06,
|
| 2731 |
+
"loss": 0.3457,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.80605055946954,
|
| 2736 |
+
"grad_norm": 2.1821189723068612,
|
| 2737 |
+
"learning_rate": 1.103441152395588e-06,
|
| 2738 |
+
"loss": 0.3283,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8081226688769167,
|
| 2743 |
+
"grad_norm": 2.417668328035057,
|
| 2744 |
+
"learning_rate": 1.0808736483556486e-06,
|
| 2745 |
+
"loss": 0.3386,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8101947782842934,
|
| 2750 |
+
"grad_norm": 2.424032614605958,
|
| 2751 |
+
"learning_rate": 1.0585113113013656e-06,
|
| 2752 |
+
"loss": 0.3451,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8122668876916701,
|
| 2757 |
+
"grad_norm": 2.3278404790134823,
|
| 2758 |
+
"learning_rate": 1.036355311905194e-06,
|
| 2759 |
+
"loss": 0.3455,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8143389970990468,
|
| 2764 |
+
"grad_norm": 2.3039554290948185,
|
| 2765 |
+
"learning_rate": 1.0144068100377818e-06,
|
| 2766 |
+
"loss": 0.3381,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.8164111065064236,
|
| 2771 |
+
"grad_norm": 2.556306207112663,
|
| 2772 |
+
"learning_rate": 9.926669547072365e-07,
|
| 2773 |
+
"loss": 0.3485,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.8184832159138002,
|
| 2778 |
+
"grad_norm": 2.4582213305283775,
|
| 2779 |
+
"learning_rate": 9.711368839989904e-07,
|
| 2780 |
+
"loss": 0.3335,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.8205553253211769,
|
| 2785 |
+
"grad_norm": 2.3579094295454657,
|
| 2786 |
+
"learning_rate": 9.498177250162022e-07,
|
| 2787 |
+
"loss": 0.3346,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8226274347285537,
|
| 2792 |
+
"grad_norm": 2.3492342765735934,
|
| 2793 |
+
"learning_rate": 9.287105938207691e-07,
|
| 2794 |
+
"loss": 0.3365,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.8246995441359304,
|
| 2799 |
+
"grad_norm": 2.4946764945886017,
|
| 2800 |
+
"learning_rate": 9.078165953748936e-07,
|
| 2801 |
+
"loss": 0.336,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.8267716535433071,
|
| 2806 |
+
"grad_norm": 2.5266114489992955,
|
| 2807 |
+
"learning_rate": 8.871368234832378e-07,
|
| 2808 |
+
"loss": 0.3367,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.8288437629506838,
|
| 2813 |
+
"grad_norm": 2.473228872985738,
|
| 2814 |
+
"learning_rate": 8.66672360735668e-07,
|
| 2815 |
+
"loss": 0.3247,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.8309158723580605,
|
| 2820 |
+
"grad_norm": 2.393007820983065,
|
| 2821 |
+
"learning_rate": 8.4642427845057e-07,
|
| 2822 |
+
"loss": 0.3289,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.8329879817654372,
|
| 2827 |
+
"grad_norm": 2.492764210538724,
|
| 2828 |
+
"learning_rate": 8.263936366187825e-07,
|
| 2829 |
+
"loss": 0.3325,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.8350600911728139,
|
| 2834 |
+
"grad_norm": 2.4227393176245418,
|
| 2835 |
+
"learning_rate": 8.065814838480879e-07,
|
| 2836 |
+
"loss": 0.3288,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.8371322005801907,
|
| 2841 |
+
"grad_norm": 2.3030572780595366,
|
| 2842 |
+
"learning_rate": 7.869888573083295e-07,
|
| 2843 |
+
"loss": 0.3401,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.8392043099875673,
|
| 2848 |
+
"grad_norm": 2.3219953415669528,
|
| 2849 |
+
"learning_rate": 7.676167826771125e-07,
|
| 2850 |
+
"loss": 0.331,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.841276419394944,
|
| 2855 |
+
"grad_norm": 2.273400790612548,
|
| 2856 |
+
"learning_rate": 7.484662740861093e-07,
|
| 2857 |
+
"loss": 0.3383,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.8433485288023208,
|
| 2862 |
+
"grad_norm": 2.374061200093819,
|
| 2863 |
+
"learning_rate": 7.295383340679668e-07,
|
| 2864 |
+
"loss": 0.3343,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.8454206382096975,
|
| 2869 |
+
"grad_norm": 2.377090176007411,
|
| 2870 |
+
"learning_rate": 7.108339535038278e-07,
|
| 2871 |
+
"loss": 0.3298,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.8474927476170742,
|
| 2876 |
+
"grad_norm": 2.3279278715664016,
|
| 2877 |
+
"learning_rate": 6.923541115714577e-07,
|
| 2878 |
+
"loss": 0.3319,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.8495648570244508,
|
| 2883 |
+
"grad_norm": 2.36303398687802,
|
| 2884 |
+
"learning_rate": 6.740997756939826e-07,
|
| 2885 |
+
"loss": 0.3418,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.8516369664318276,
|
| 2890 |
+
"grad_norm": 2.415624778572703,
|
| 2891 |
+
"learning_rate": 6.560719014892425e-07,
|
| 2892 |
+
"loss": 0.3328,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.8537090758392043,
|
| 2897 |
+
"grad_norm": 2.287618281117491,
|
| 2898 |
+
"learning_rate": 6.382714327197703e-07,
|
| 2899 |
+
"loss": 0.3321,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.855781185246581,
|
| 2904 |
+
"grad_norm": 2.2595148250907036,
|
| 2905 |
+
"learning_rate": 6.206993012433815e-07,
|
| 2906 |
+
"loss": 0.3336,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.8578532946539578,
|
| 2911 |
+
"grad_norm": 2.3926760910481186,
|
| 2912 |
+
"learning_rate": 6.033564269643927e-07,
|
| 2913 |
+
"loss": 0.3342,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.8599254040613344,
|
| 2918 |
+
"grad_norm": 2.316219130405351,
|
| 2919 |
+
"learning_rate": 5.862437177854629e-07,
|
| 2920 |
+
"loss": 0.3311,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.8619975134687111,
|
| 2925 |
+
"grad_norm": 2.3766431440321707,
|
| 2926 |
+
"learning_rate": 5.693620695600671e-07,
|
| 2927 |
+
"loss": 0.3153,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.8640696228760879,
|
| 2932 |
+
"grad_norm": 2.1660555465277405,
|
| 2933 |
+
"learning_rate": 5.527123660455968e-07,
|
| 2934 |
+
"loss": 0.3268,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.8661417322834646,
|
| 2939 |
+
"grad_norm": 2.3890878403617815,
|
| 2940 |
+
"learning_rate": 5.362954788570929e-07,
|
| 2941 |
+
"loss": 0.317,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.8682138416908413,
|
| 2946 |
+
"grad_norm": 2.3910404086765515,
|
| 2947 |
+
"learning_rate": 5.201122674216208e-07,
|
| 2948 |
+
"loss": 0.329,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.8702859510982179,
|
| 2953 |
+
"grad_norm": 2.398893233495637,
|
| 2954 |
+
"learning_rate": 5.041635789332783e-07,
|
| 2955 |
+
"loss": 0.3306,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.8723580605055947,
|
| 2960 |
+
"grad_norm": 2.5664259292956495,
|
| 2961 |
+
"learning_rate": 4.884502483088421e-07,
|
| 2962 |
+
"loss": 0.3153,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.8744301699129714,
|
| 2967 |
+
"grad_norm": 2.4313343193724073,
|
| 2968 |
+
"learning_rate": 4.7297309814406113e-07,
|
| 2969 |
+
"loss": 0.3286,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.8765022793203481,
|
| 2974 |
+
"grad_norm": 2.3267523640919374,
|
| 2975 |
+
"learning_rate": 4.577329386705942e-07,
|
| 2976 |
+
"loss": 0.3321,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.8785743887277249,
|
| 2981 |
+
"grad_norm": 2.37298005463995,
|
| 2982 |
+
"learning_rate": 4.42730567713594e-07,
|
| 2983 |
+
"loss": 0.3303,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.8806464981351015,
|
| 2988 |
+
"grad_norm": 2.4400767630128084,
|
| 2989 |
+
"learning_rate": 4.2796677064994243e-07,
|
| 2990 |
+
"loss": 0.3187,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.8827186075424782,
|
| 2995 |
+
"grad_norm": 2.600660724910327,
|
| 2996 |
+
"learning_rate": 4.134423203671295e-07,
|
| 2997 |
+
"loss": 0.3347,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.884790716949855,
|
| 3002 |
+
"grad_norm": 2.416615232949827,
|
| 3003 |
+
"learning_rate": 3.9915797722280323e-07,
|
| 3004 |
+
"loss": 0.3317,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.8868628263572317,
|
| 3009 |
+
"grad_norm": 2.3982415986507344,
|
| 3010 |
+
"learning_rate": 3.851144890049535e-07,
|
| 3011 |
+
"loss": 0.326,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.8889349357646084,
|
| 3016 |
+
"grad_norm": 2.3596784596238445,
|
| 3017 |
+
"learning_rate": 3.713125908927728e-07,
|
| 3018 |
+
"loss": 0.3274,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.891007045171985,
|
| 3023 |
+
"grad_norm": 2.395731949083915,
|
| 3024 |
+
"learning_rate": 3.577530054181677e-07,
|
| 3025 |
+
"loss": 0.3337,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.8930791545793618,
|
| 3030 |
+
"grad_norm": 2.4132993194541363,
|
| 3031 |
+
"learning_rate": 3.4443644242793226e-07,
|
| 3032 |
+
"loss": 0.3205,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.8951512639867385,
|
| 3037 |
+
"grad_norm": 2.421557012145629,
|
| 3038 |
+
"learning_rate": 3.313635990465902e-07,
|
| 3039 |
+
"loss": 0.3286,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.8972233733941152,
|
| 3044 |
+
"grad_norm": 2.3514364773273053,
|
| 3045 |
+
"learning_rate": 3.1853515963989613e-07,
|
| 3046 |
+
"loss": 0.3244,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.899295482801492,
|
| 3051 |
+
"grad_norm": 2.3631578697068156,
|
| 3052 |
+
"learning_rate": 3.059517957790165e-07,
|
| 3053 |
+
"loss": 0.3224,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9013675922088686,
|
| 3058 |
+
"grad_norm": 2.3165934226078693,
|
| 3059 |
+
"learning_rate": 2.936141662053621e-07,
|
| 3060 |
+
"loss": 0.3189,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9034397016162453,
|
| 3065 |
+
"grad_norm": 2.354830595345679,
|
| 3066 |
+
"learning_rate": 2.8152291679611254e-07,
|
| 3067 |
+
"loss": 0.3164,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.905511811023622,
|
| 3072 |
+
"grad_norm": 2.4591886130861984,
|
| 3073 |
+
"learning_rate": 2.6967868053039916e-07,
|
| 3074 |
+
"loss": 0.3194,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.9075839204309988,
|
| 3079 |
+
"grad_norm": 2.4651307442780515,
|
| 3080 |
+
"learning_rate": 2.580820774561704e-07,
|
| 3081 |
+
"loss": 0.3329,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.9096560298383755,
|
| 3086 |
+
"grad_norm": 2.356380428139328,
|
| 3087 |
+
"learning_rate": 2.467337146577298e-07,
|
| 3088 |
+
"loss": 0.3125,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.9117281392457521,
|
| 3093 |
+
"grad_norm": 2.4718310107961923,
|
| 3094 |
+
"learning_rate": 2.3563418622395863e-07,
|
| 3095 |
+
"loss": 0.3267,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.9138002486531289,
|
| 3100 |
+
"grad_norm": 2.2943822251497834,
|
| 3101 |
+
"learning_rate": 2.2478407321721295e-07,
|
| 3102 |
+
"loss": 0.3124,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.9158723580605056,
|
| 3107 |
+
"grad_norm": 2.37802636947732,
|
| 3108 |
+
"learning_rate": 2.141839436429055e-07,
|
| 3109 |
+
"loss": 0.3261,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.9179444674678823,
|
| 3114 |
+
"grad_norm": 2.4929237268669504,
|
| 3115 |
+
"learning_rate": 2.038343524197689e-07,
|
| 3116 |
+
"loss": 0.3266,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.9200165768752591,
|
| 3121 |
+
"grad_norm": 2.4497686225084645,
|
| 3122 |
+
"learning_rate": 1.9373584135080893e-07,
|
| 3123 |
+
"loss": 0.3155,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.9220886862826357,
|
| 3128 |
+
"grad_norm": 2.3085410601255063,
|
| 3129 |
+
"learning_rate": 1.8388893909493776e-07,
|
| 3130 |
+
"loss": 0.3143,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.9241607956900124,
|
| 3135 |
+
"grad_norm": 2.516522380423302,
|
| 3136 |
+
"learning_rate": 1.742941611393012e-07,
|
| 3137 |
+
"loss": 0.3282,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.9262329050973891,
|
| 3142 |
+
"grad_norm": 2.250941160020719,
|
| 3143 |
+
"learning_rate": 1.6495200977228897e-07,
|
| 3144 |
+
"loss": 0.3313,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.9283050145047659,
|
| 3149 |
+
"grad_norm": 2.243976361659756,
|
| 3150 |
+
"learning_rate": 1.558629740572465e-07,
|
| 3151 |
+
"loss": 0.3123,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.9303771239121426,
|
| 3156 |
+
"grad_norm": 2.411662333101451,
|
| 3157 |
+
"learning_rate": 1.4702752980686463e-07,
|
| 3158 |
+
"loss": 0.3227,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.9324492333195192,
|
| 3163 |
+
"grad_norm": 2.4064884421540254,
|
| 3164 |
+
"learning_rate": 1.3844613955827536e-07,
|
| 3165 |
+
"loss": 0.326,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.934521342726896,
|
| 3170 |
+
"grad_norm": 2.4948851626678326,
|
| 3171 |
+
"learning_rate": 1.301192525488376e-07,
|
| 3172 |
+
"loss": 0.3177,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.9365934521342727,
|
| 3177 |
+
"grad_norm": 2.256448839133944,
|
| 3178 |
+
"learning_rate": 1.2204730469261905e-07,
|
| 3179 |
+
"loss": 0.3224,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.9386655615416494,
|
| 3184 |
+
"grad_norm": 2.386604983970806,
|
| 3185 |
+
"learning_rate": 1.1423071855757473e-07,
|
| 3186 |
+
"loss": 0.3177,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.9407376709490262,
|
| 3191 |
+
"grad_norm": 2.4955559903632087,
|
| 3192 |
+
"learning_rate": 1.0666990334342708e-07,
|
| 3193 |
+
"loss": 0.3235,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.9428097803564028,
|
| 3198 |
+
"grad_norm": 2.4680594179132953,
|
| 3199 |
+
"learning_rate": 9.936525486024362e-08,
|
| 3200 |
+
"loss": 0.3234,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.9448818897637795,
|
| 3205 |
+
"grad_norm": 2.352943986977247,
|
| 3206 |
+
"learning_rate": 9.23171555077168e-08,
|
| 3207 |
+
"loss": 0.3189,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.9469539991711562,
|
| 3212 |
+
"grad_norm": 2.3685608851078106,
|
| 3213 |
+
"learning_rate": 8.552597425514508e-08,
|
| 3214 |
+
"loss": 0.3191,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.949026108578533,
|
| 3219 |
+
"grad_norm": 2.4481455211139944,
|
| 3220 |
+
"learning_rate": 7.8992066622115e-08,
|
| 3221 |
+
"loss": 0.3283,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.9510982179859097,
|
| 3226 |
+
"grad_norm": 2.298587702409287,
|
| 3227 |
+
"learning_rate": 7.271577465989554e-08,
|
| 3228 |
+
"loss": 0.3181,
|
| 3229 |
+
"step": 2295
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.9531703273932863,
|
| 3233 |
+
"grad_norm": 2.3601024907505956,
|
| 3234 |
+
"learning_rate": 6.669742693352522e-08,
|
| 3235 |
+
"loss": 0.3183,
|
| 3236 |
+
"step": 2300
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.9552424368006631,
|
| 3240 |
+
"grad_norm": 2.3846417205933537,
|
| 3241 |
+
"learning_rate": 6.093733850461359e-08,
|
| 3242 |
+
"loss": 0.311,
|
| 3243 |
+
"step": 2305
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.9573145462080398,
|
| 3247 |
+
"grad_norm": 2.5721713736853427,
|
| 3248 |
+
"learning_rate": 5.5435810914851176e-08,
|
| 3249 |
+
"loss": 0.3225,
|
| 3250 |
+
"step": 2310
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.9593866556154165,
|
| 3254 |
+
"grad_norm": 2.4391728621326667,
|
| 3255 |
+
"learning_rate": 5.0193132170219814e-08,
|
| 3256 |
+
"loss": 0.3219,
|
| 3257 |
+
"step": 2315
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.9614587650227931,
|
| 3261 |
+
"grad_norm": 2.314964895648817,
|
| 3262 |
+
"learning_rate": 4.5209576725915305e-08,
|
| 3263 |
+
"loss": 0.3146,
|
| 3264 |
+
"step": 2320
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.9635308744301699,
|
| 3268 |
+
"grad_norm": 2.3953921094993427,
|
| 3269 |
+
"learning_rate": 4.0485405471983317e-08,
|
| 3270 |
+
"loss": 0.3099,
|
| 3271 |
+
"step": 2325
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.9656029838375466,
|
| 3275 |
+
"grad_norm": 2.4903270174654977,
|
| 3276 |
+
"learning_rate": 3.6020865719657015e-08,
|
| 3277 |
+
"loss": 0.3261,
|
| 3278 |
+
"step": 2330
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.9676750932449233,
|
| 3282 |
+
"grad_norm": 2.5600866933116153,
|
| 3283 |
+
"learning_rate": 3.181619118841517e-08,
|
| 3284 |
+
"loss": 0.3207,
|
| 3285 |
+
"step": 2335
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.9697472026523001,
|
| 3289 |
+
"grad_norm": 2.2542713880681466,
|
| 3290 |
+
"learning_rate": 2.7871601993741947e-08,
|
| 3291 |
+
"loss": 0.3154,
|
| 3292 |
+
"step": 2340
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.9718193120596768,
|
| 3296 |
+
"grad_norm": 2.3906714544505,
|
| 3297 |
+
"learning_rate": 2.4187304635608922e-08,
|
| 3298 |
+
"loss": 0.3264,
|
| 3299 |
+
"step": 2345
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.9738914214670534,
|
| 3303 |
+
"grad_norm": 2.3437139004611396,
|
| 3304 |
+
"learning_rate": 2.0763491987659812e-08,
|
| 3305 |
+
"loss": 0.3131,
|
| 3306 |
+
"step": 2350
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.9759635308744302,
|
| 3310 |
+
"grad_norm": 2.3352282804892504,
|
| 3311 |
+
"learning_rate": 1.7600343287116904e-08,
|
| 3312 |
+
"loss": 0.3236,
|
| 3313 |
+
"step": 2355
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.9780356402818069,
|
| 3317 |
+
"grad_norm": 2.2788999882928267,
|
| 3318 |
+
"learning_rate": 1.4698024125396893e-08,
|
| 3319 |
+
"loss": 0.3165,
|
| 3320 |
+
"step": 2360
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.9801077496891836,
|
| 3324 |
+
"grad_norm": 2.484272820389643,
|
| 3325 |
+
"learning_rate": 1.205668643944169e-08,
|
| 3326 |
+
"loss": 0.3265,
|
| 3327 |
+
"step": 2365
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.9821798590965602,
|
| 3331 |
+
"grad_norm": 2.3492923712340357,
|
| 3332 |
+
"learning_rate": 9.676468503765356e-09,
|
| 3333 |
+
"loss": 0.3185,
|
| 3334 |
+
"step": 2370
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.984251968503937,
|
| 3338 |
+
"grad_norm": 2.2726204174567806,
|
| 3339 |
+
"learning_rate": 7.557494923214338e-09,
|
| 3340 |
+
"loss": 0.3187,
|
| 3341 |
+
"step": 2375
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.9863240779113137,
|
| 3345 |
+
"grad_norm": 2.502123973833959,
|
| 3346 |
+
"learning_rate": 5.699876626446554e-09,
|
| 3347 |
+
"loss": 0.315,
|
| 3348 |
+
"step": 2380
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.9883961873186904,
|
| 3352 |
+
"grad_norm": 2.5216195549598046,
|
| 3353 |
+
"learning_rate": 4.103710860120513e-09,
|
| 3354 |
+
"loss": 0.3232,
|
| 3355 |
+
"step": 2385
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.9904682967260672,
|
| 3359 |
+
"grad_norm": 2.3006568402247987,
|
| 3360 |
+
"learning_rate": 2.769081183808253e-09,
|
| 3361 |
+
"loss": 0.3132,
|
| 3362 |
+
"step": 2390
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.9925404061334439,
|
| 3366 |
+
"grad_norm": 2.3057165957876977,
|
| 3367 |
+
"learning_rate": 1.69605746561885e-09,
|
| 3368 |
+
"loss": 0.3099,
|
| 3369 |
+
"step": 2395
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.9946125155408205,
|
| 3373 |
+
"grad_norm": 2.3117976401405853,
|
| 3374 |
+
"learning_rate": 8.846958785418969e-10,
|
| 3375 |
+
"loss": 0.3166,
|
| 3376 |
+
"step": 2400
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.9966846249481973,
|
| 3380 |
+
"grad_norm": 2.4474241574477613,
|
| 3381 |
+
"learning_rate": 3.3503889750485794e-10,
|
| 3382 |
+
"loss": 0.3198,
|
| 3383 |
+
"step": 2405
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.998756734355574,
|
| 3387 |
+
"grad_norm": 2.475055392564647,
|
| 3388 |
+
"learning_rate": 4.711529715262231e-11,
|
| 3389 |
+
"loss": 0.3188,
|
| 3390 |
+
"step": 2410
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 1.0,
|
| 3394 |
+
"eval_loss": 0.2978341579437256,
|
| 3395 |
+
"eval_runtime": 1.1902,
|
| 3396 |
+
"eval_samples_per_second": 2.521,
|
| 3397 |
+
"eval_steps_per_second": 0.84,
|
| 3398 |
+
"step": 2413
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 1.0,
|
| 3402 |
+
"step": 2413,
|
| 3403 |
+
"total_flos": 252616554577920.0,
|
| 3404 |
+
"train_loss": 0.5160677275912573,
|
| 3405 |
+
"train_runtime": 24380.0574,
|
| 3406 |
+
"train_samples_per_second": 1.584,
|
| 3407 |
+
"train_steps_per_second": 0.099
|
| 3408 |
+
}
|
| 3409 |
+
],
|
| 3410 |
+
"logging_steps": 5,
|
| 3411 |
+
"max_steps": 2413,
|
| 3412 |
+
"num_input_tokens_seen": 0,
|
| 3413 |
+
"num_train_epochs": 1,
|
| 3414 |
+
"save_steps": 100,
|
| 3415 |
+
"stateful_callbacks": {
|
| 3416 |
+
"TrainerControl": {
|
| 3417 |
+
"args": {
|
| 3418 |
+
"should_epoch_stop": false,
|
| 3419 |
+
"should_evaluate": false,
|
| 3420 |
+
"should_log": false,
|
| 3421 |
+
"should_save": true,
|
| 3422 |
+
"should_training_stop": true
|
| 3423 |
+
},
|
| 3424 |
+
"attributes": {}
|
| 3425 |
+
}
|
| 3426 |
+
},
|
| 3427 |
+
"total_flos": 252616554577920.0,
|
| 3428 |
+
"train_batch_size": 4,
|
| 3429 |
+
"trial_name": null,
|
| 3430 |
+
"trial_params": null
|
| 3431 |
+
}
|