Model save
Browse files- README.md +61 -0
- all_results.json +9 -0
- runs/Sep19_11-56-10_COE-CS-sv003/events.out.tfevents.1726747489.COE-CS-sv003.383402.0 +2 -2
- train_results.json +9 -0
- trainer_state.json +3115 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: gemma
|
| 3 |
+
library_name: peft
|
| 4 |
+
tags:
|
| 5 |
+
- trl
|
| 6 |
+
- sft
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
base_model: google/gemma-7b-it
|
| 9 |
+
model-index:
|
| 10 |
+
- name: gemma_preference
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# gemma_preference
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [google/gemma-7b-it](https://huggingface.co/google/gemma-7b-it) on the None dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 1e-05
|
| 39 |
+
- train_batch_size: 4
|
| 40 |
+
- eval_batch_size: 1
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 4
|
| 44 |
+
- total_train_batch_size: 16
|
| 45 |
+
- total_eval_batch_size: 4
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.11.1
|
| 58 |
+
- Transformers 4.44.2
|
| 59 |
+
- Pytorch 2.3.1+cu121
|
| 60 |
+
- Datasets 2.19.1
|
| 61 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 9703359095242752.0,
|
| 4 |
+
"train_loss": 1.3940648635773556,
|
| 5 |
+
"train_runtime": 19118.6692,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.83,
|
| 8 |
+
"train_steps_per_second": 0.114
|
| 9 |
+
}
|
runs/Sep19_11-56-10_COE-CS-sv003/events.out.tfevents.1726747489.COE-CS-sv003.383402.0
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 sha256:87f7924fcaf5c650b7f9cc75a796e53a708b38168e726ed462627912a9baa4ff
|
| 3 |
+
size 98067
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 9703359095242752.0,
|
| 4 |
+
"train_loss": 1.3940648635773556,
|
| 5 |
+
"train_runtime": 19118.6692,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.83,
|
| 8 |
+
"train_steps_per_second": 0.114
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2187,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004572473708276177,
|
| 13 |
+
"grad_norm": 8.096893602247533,
|
| 14 |
+
"learning_rate": 4.5662100456621004e-08,
|
| 15 |
+
"loss": 4.0564,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.002286236854138089,
|
| 20 |
+
"grad_norm": 9.362922497604572,
|
| 21 |
+
"learning_rate": 2.2831050228310502e-07,
|
| 22 |
+
"loss": 4.0286,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004572473708276178,
|
| 27 |
+
"grad_norm": 8.436200820951127,
|
| 28 |
+
"learning_rate": 4.5662100456621004e-07,
|
| 29 |
+
"loss": 4.0215,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006858710562414266,
|
| 34 |
+
"grad_norm": 8.311166463949595,
|
| 35 |
+
"learning_rate": 6.849315068493151e-07,
|
| 36 |
+
"loss": 3.995,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.009144947416552356,
|
| 41 |
+
"grad_norm": 8.314737539002055,
|
| 42 |
+
"learning_rate": 9.132420091324201e-07,
|
| 43 |
+
"loss": 4.0089,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.011431184270690443,
|
| 48 |
+
"grad_norm": 8.207038819761589,
|
| 49 |
+
"learning_rate": 1.1415525114155251e-06,
|
| 50 |
+
"loss": 4.0363,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.013717421124828532,
|
| 55 |
+
"grad_norm": 7.559088257570073,
|
| 56 |
+
"learning_rate": 1.3698630136986302e-06,
|
| 57 |
+
"loss": 3.9854,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.01600365797896662,
|
| 62 |
+
"grad_norm": 7.1469550333759315,
|
| 63 |
+
"learning_rate": 1.5981735159817353e-06,
|
| 64 |
+
"loss": 3.9172,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.01828989483310471,
|
| 69 |
+
"grad_norm": 5.971779564925808,
|
| 70 |
+
"learning_rate": 1.8264840182648401e-06,
|
| 71 |
+
"loss": 3.7833,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0205761316872428,
|
| 76 |
+
"grad_norm": 5.297232176162252,
|
| 77 |
+
"learning_rate": 2.0547945205479454e-06,
|
| 78 |
+
"loss": 3.7258,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.022862368541380886,
|
| 83 |
+
"grad_norm": 4.217560181118984,
|
| 84 |
+
"learning_rate": 2.2831050228310503e-06,
|
| 85 |
+
"loss": 3.6069,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.025148605395518976,
|
| 90 |
+
"grad_norm": 3.786598130029432,
|
| 91 |
+
"learning_rate": 2.511415525114155e-06,
|
| 92 |
+
"loss": 3.553,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.027434842249657063,
|
| 97 |
+
"grad_norm": 2.6863763381878782,
|
| 98 |
+
"learning_rate": 2.7397260273972604e-06,
|
| 99 |
+
"loss": 3.4564,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.029721079103795154,
|
| 104 |
+
"grad_norm": 2.322206779034821,
|
| 105 |
+
"learning_rate": 2.9680365296803653e-06,
|
| 106 |
+
"loss": 3.331,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03200731595793324,
|
| 111 |
+
"grad_norm": 1.7328570105860337,
|
| 112 |
+
"learning_rate": 3.1963470319634706e-06,
|
| 113 |
+
"loss": 3.2806,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03429355281207133,
|
| 118 |
+
"grad_norm": 1.4846651112766411,
|
| 119 |
+
"learning_rate": 3.4246575342465754e-06,
|
| 120 |
+
"loss": 3.2356,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03657978966620942,
|
| 125 |
+
"grad_norm": 1.1848731903024705,
|
| 126 |
+
"learning_rate": 3.6529680365296803e-06,
|
| 127 |
+
"loss": 3.1934,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.038866026520347506,
|
| 132 |
+
"grad_norm": 1.0381920877926754,
|
| 133 |
+
"learning_rate": 3.881278538812785e-06,
|
| 134 |
+
"loss": 3.1245,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.0411522633744856,
|
| 139 |
+
"grad_norm": 0.9141218079482407,
|
| 140 |
+
"learning_rate": 4.109589041095891e-06,
|
| 141 |
+
"loss": 3.0468,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04343850022862369,
|
| 146 |
+
"grad_norm": 1.0161753499176187,
|
| 147 |
+
"learning_rate": 4.337899543378996e-06,
|
| 148 |
+
"loss": 3.0329,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04572473708276177,
|
| 153 |
+
"grad_norm": 0.8996453649527762,
|
| 154 |
+
"learning_rate": 4.566210045662101e-06,
|
| 155 |
+
"loss": 2.9958,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.04801097393689986,
|
| 160 |
+
"grad_norm": 0.8082797335102323,
|
| 161 |
+
"learning_rate": 4.7945205479452054e-06,
|
| 162 |
+
"loss": 2.9651,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.05029721079103795,
|
| 167 |
+
"grad_norm": 0.7090431405422901,
|
| 168 |
+
"learning_rate": 5.02283105022831e-06,
|
| 169 |
+
"loss": 2.927,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.05258344764517604,
|
| 174 |
+
"grad_norm": 1.2265537925663061,
|
| 175 |
+
"learning_rate": 5.251141552511416e-06,
|
| 176 |
+
"loss": 2.862,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05486968449931413,
|
| 181 |
+
"grad_norm": 0.650571444620453,
|
| 182 |
+
"learning_rate": 5.479452054794521e-06,
|
| 183 |
+
"loss": 2.857,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05715592135345222,
|
| 188 |
+
"grad_norm": 0.7089485111846239,
|
| 189 |
+
"learning_rate": 5.7077625570776266e-06,
|
| 190 |
+
"loss": 2.8209,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.05944215820759031,
|
| 195 |
+
"grad_norm": 0.5922494361050838,
|
| 196 |
+
"learning_rate": 5.936073059360731e-06,
|
| 197 |
+
"loss": 2.8037,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06172839506172839,
|
| 202 |
+
"grad_norm": 0.5597217919230902,
|
| 203 |
+
"learning_rate": 6.164383561643836e-06,
|
| 204 |
+
"loss": 2.7487,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06401463191586648,
|
| 209 |
+
"grad_norm": 0.6045746583730743,
|
| 210 |
+
"learning_rate": 6.392694063926941e-06,
|
| 211 |
+
"loss": 2.6981,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06630086877000457,
|
| 216 |
+
"grad_norm": 0.6479924774135967,
|
| 217 |
+
"learning_rate": 6.621004566210046e-06,
|
| 218 |
+
"loss": 2.7036,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06858710562414266,
|
| 223 |
+
"grad_norm": 0.760723993748018,
|
| 224 |
+
"learning_rate": 6.849315068493151e-06,
|
| 225 |
+
"loss": 2.6821,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07087334247828075,
|
| 230 |
+
"grad_norm": 0.5889973577684341,
|
| 231 |
+
"learning_rate": 7.077625570776257e-06,
|
| 232 |
+
"loss": 2.6882,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07315957933241884,
|
| 237 |
+
"grad_norm": 0.6201384588278992,
|
| 238 |
+
"learning_rate": 7.305936073059361e-06,
|
| 239 |
+
"loss": 2.6441,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.07544581618655692,
|
| 244 |
+
"grad_norm": 0.645862573214957,
|
| 245 |
+
"learning_rate": 7.534246575342466e-06,
|
| 246 |
+
"loss": 2.5878,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07773205304069501,
|
| 251 |
+
"grad_norm": 0.880791499233313,
|
| 252 |
+
"learning_rate": 7.76255707762557e-06,
|
| 253 |
+
"loss": 2.5665,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.0800182898948331,
|
| 258 |
+
"grad_norm": 0.7581098091472079,
|
| 259 |
+
"learning_rate": 7.990867579908676e-06,
|
| 260 |
+
"loss": 2.5423,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.0823045267489712,
|
| 265 |
+
"grad_norm": 0.7502504535360037,
|
| 266 |
+
"learning_rate": 8.219178082191782e-06,
|
| 267 |
+
"loss": 2.5348,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08459076360310928,
|
| 272 |
+
"grad_norm": 0.9587325899501735,
|
| 273 |
+
"learning_rate": 8.447488584474887e-06,
|
| 274 |
+
"loss": 2.4652,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.08687700045724737,
|
| 279 |
+
"grad_norm": 1.0327228370595574,
|
| 280 |
+
"learning_rate": 8.675799086757991e-06,
|
| 281 |
+
"loss": 2.4066,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08916323731138547,
|
| 286 |
+
"grad_norm": 0.8853835960264104,
|
| 287 |
+
"learning_rate": 8.904109589041097e-06,
|
| 288 |
+
"loss": 2.3642,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09144947416552354,
|
| 293 |
+
"grad_norm": 1.0446953486337078,
|
| 294 |
+
"learning_rate": 9.132420091324201e-06,
|
| 295 |
+
"loss": 2.3237,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09373571101966163,
|
| 300 |
+
"grad_norm": 1.1013758488210148,
|
| 301 |
+
"learning_rate": 9.360730593607307e-06,
|
| 302 |
+
"loss": 2.2331,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.09602194787379972,
|
| 307 |
+
"grad_norm": 1.2192543249794923,
|
| 308 |
+
"learning_rate": 9.589041095890411e-06,
|
| 309 |
+
"loss": 2.1264,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.09830818472793781,
|
| 314 |
+
"grad_norm": 1.3533953895273099,
|
| 315 |
+
"learning_rate": 9.817351598173517e-06,
|
| 316 |
+
"loss": 2.0554,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.1005944215820759,
|
| 321 |
+
"grad_norm": 1.1876482609404326,
|
| 322 |
+
"learning_rate": 9.999993629265979e-06,
|
| 323 |
+
"loss": 1.9859,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.102880658436214,
|
| 328 |
+
"grad_norm": 1.1847416528253172,
|
| 329 |
+
"learning_rate": 9.999770655279843e-06,
|
| 330 |
+
"loss": 1.8986,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.10516689529035209,
|
| 335 |
+
"grad_norm": 1.3137466650624998,
|
| 336 |
+
"learning_rate": 9.999229160826947e-06,
|
| 337 |
+
"loss": 1.8,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.10745313214449016,
|
| 342 |
+
"grad_norm": 1.830150495140023,
|
| 343 |
+
"learning_rate": 9.998369180404283e-06,
|
| 344 |
+
"loss": 1.7138,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.10973936899862825,
|
| 349 |
+
"grad_norm": 1.1159850299398295,
|
| 350 |
+
"learning_rate": 9.997190768798639e-06,
|
| 351 |
+
"loss": 1.6867,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.11202560585276634,
|
| 356 |
+
"grad_norm": 0.9727694366367986,
|
| 357 |
+
"learning_rate": 9.995694001083103e-06,
|
| 358 |
+
"loss": 1.6469,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11431184270690443,
|
| 363 |
+
"grad_norm": 1.135743426814773,
|
| 364 |
+
"learning_rate": 9.993878972612276e-06,
|
| 365 |
+
"loss": 1.5607,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.11659807956104253,
|
| 370 |
+
"grad_norm": 1.0363437963731608,
|
| 371 |
+
"learning_rate": 9.991745799016206e-06,
|
| 372 |
+
"loss": 1.5332,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.11888431641518062,
|
| 377 |
+
"grad_norm": 1.018006180331875,
|
| 378 |
+
"learning_rate": 9.989294616193018e-06,
|
| 379 |
+
"loss": 1.4962,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.1211705532693187,
|
| 384 |
+
"grad_norm": 0.9493951106581935,
|
| 385 |
+
"learning_rate": 9.986525580300253e-06,
|
| 386 |
+
"loss": 1.4403,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12345679012345678,
|
| 391 |
+
"grad_norm": 4.150830186272059,
|
| 392 |
+
"learning_rate": 9.983438867744923e-06,
|
| 393 |
+
"loss": 1.4382,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.12574302697759487,
|
| 398 |
+
"grad_norm": 0.8458476848705546,
|
| 399 |
+
"learning_rate": 9.980034675172274e-06,
|
| 400 |
+
"loss": 1.4248,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.12802926383173296,
|
| 405 |
+
"grad_norm": 1.8171861028727991,
|
| 406 |
+
"learning_rate": 9.976313219453255e-06,
|
| 407 |
+
"loss": 1.4055,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13031550068587106,
|
| 412 |
+
"grad_norm": 0.7389926811741014,
|
| 413 |
+
"learning_rate": 9.972274737670702e-06,
|
| 414 |
+
"loss": 1.4033,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.13260173754000915,
|
| 419 |
+
"grad_norm": 0.8834746515415843,
|
| 420 |
+
"learning_rate": 9.967919487104237e-06,
|
| 421 |
+
"loss": 1.3724,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.13488797439414724,
|
| 426 |
+
"grad_norm": 0.8166186304734012,
|
| 427 |
+
"learning_rate": 9.963247745213876e-06,
|
| 428 |
+
"loss": 1.3721,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.13717421124828533,
|
| 433 |
+
"grad_norm": 0.6771475216933378,
|
| 434 |
+
"learning_rate": 9.958259809622353e-06,
|
| 435 |
+
"loss": 1.3555,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.13946044810242342,
|
| 440 |
+
"grad_norm": 0.60525762012324,
|
| 441 |
+
"learning_rate": 9.952955998096155e-06,
|
| 442 |
+
"loss": 1.36,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.1417466849565615,
|
| 447 |
+
"grad_norm": 0.6126617626167846,
|
| 448 |
+
"learning_rate": 9.94733664852529e-06,
|
| 449 |
+
"loss": 1.353,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.1440329218106996,
|
| 454 |
+
"grad_norm": 0.6630794657190928,
|
| 455 |
+
"learning_rate": 9.941402118901743e-06,
|
| 456 |
+
"loss": 1.3359,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.1463191586648377,
|
| 461 |
+
"grad_norm": 0.6758533351396738,
|
| 462 |
+
"learning_rate": 9.935152787296689e-06,
|
| 463 |
+
"loss": 1.3402,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.14860539551897575,
|
| 468 |
+
"grad_norm": 0.739719330356037,
|
| 469 |
+
"learning_rate": 9.928589051836392e-06,
|
| 470 |
+
"loss": 1.3346,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.15089163237311384,
|
| 475 |
+
"grad_norm": 0.7258290118963521,
|
| 476 |
+
"learning_rate": 9.921711330676848e-06,
|
| 477 |
+
"loss": 1.3356,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.15317786922725193,
|
| 482 |
+
"grad_norm": 0.6274092924270468,
|
| 483 |
+
"learning_rate": 9.91452006197715e-06,
|
| 484 |
+
"loss": 1.3362,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.15546410608139002,
|
| 489 |
+
"grad_norm": 0.768028072114212,
|
| 490 |
+
"learning_rate": 9.907015703871558e-06,
|
| 491 |
+
"loss": 1.3214,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.15775034293552812,
|
| 496 |
+
"grad_norm": 0.7738373400419118,
|
| 497 |
+
"learning_rate": 9.899198734440335e-06,
|
| 498 |
+
"loss": 1.331,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.1600365797896662,
|
| 503 |
+
"grad_norm": 0.6855410863811031,
|
| 504 |
+
"learning_rate": 9.891069651679273e-06,
|
| 505 |
+
"loss": 1.3142,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.1623228166438043,
|
| 510 |
+
"grad_norm": 0.6405023247699122,
|
| 511 |
+
"learning_rate": 9.882628973467972e-06,
|
| 512 |
+
"loss": 1.3171,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.1646090534979424,
|
| 517 |
+
"grad_norm": 0.6764400756880153,
|
| 518 |
+
"learning_rate": 9.873877237536854e-06,
|
| 519 |
+
"loss": 1.3189,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.16689529035208048,
|
| 524 |
+
"grad_norm": 0.6298462983903607,
|
| 525 |
+
"learning_rate": 9.86481500143289e-06,
|
| 526 |
+
"loss": 1.3059,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.16918152720621857,
|
| 531 |
+
"grad_norm": 0.6606697771559132,
|
| 532 |
+
"learning_rate": 9.855442842484101e-06,
|
| 533 |
+
"loss": 1.3267,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.17146776406035666,
|
| 538 |
+
"grad_norm": 0.5895037669135822,
|
| 539 |
+
"learning_rate": 9.84576135776276e-06,
|
| 540 |
+
"loss": 1.3057,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.17375400091449475,
|
| 545 |
+
"grad_norm": 0.5762405642901876,
|
| 546 |
+
"learning_rate": 9.835771164047365e-06,
|
| 547 |
+
"loss": 1.3016,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.17604023776863284,
|
| 552 |
+
"grad_norm": 0.6301891918568133,
|
| 553 |
+
"learning_rate": 9.825472897783344e-06,
|
| 554 |
+
"loss": 1.3046,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.17832647462277093,
|
| 559 |
+
"grad_norm": 0.6189017845225122,
|
| 560 |
+
"learning_rate": 9.814867215042503e-06,
|
| 561 |
+
"loss": 1.3089,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.18061271147690902,
|
| 566 |
+
"grad_norm": 0.6279515665165573,
|
| 567 |
+
"learning_rate": 9.803954791481239e-06,
|
| 568 |
+
"loss": 1.3011,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.18289894833104708,
|
| 573 |
+
"grad_norm": 0.6380039476156935,
|
| 574 |
+
"learning_rate": 9.792736322297489e-06,
|
| 575 |
+
"loss": 1.2758,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.18518518518518517,
|
| 580 |
+
"grad_norm": 0.7506004279154695,
|
| 581 |
+
"learning_rate": 9.781212522186442e-06,
|
| 582 |
+
"loss": 1.312,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.18747142203932327,
|
| 587 |
+
"grad_norm": 0.7054181242720778,
|
| 588 |
+
"learning_rate": 9.769384125295012e-06,
|
| 589 |
+
"loss": 1.3112,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.18975765889346136,
|
| 594 |
+
"grad_norm": 0.5797880483237029,
|
| 595 |
+
"learning_rate": 9.757251885175063e-06,
|
| 596 |
+
"loss": 1.2998,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.19204389574759945,
|
| 601 |
+
"grad_norm": 0.6040659600524477,
|
| 602 |
+
"learning_rate": 9.744816574735405e-06,
|
| 603 |
+
"loss": 1.3018,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.19433013260173754,
|
| 608 |
+
"grad_norm": 0.7044299546094256,
|
| 609 |
+
"learning_rate": 9.732078986192552e-06,
|
| 610 |
+
"loss": 1.2818,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.19661636945587563,
|
| 615 |
+
"grad_norm": 0.567841572649114,
|
| 616 |
+
"learning_rate": 9.719039931020258e-06,
|
| 617 |
+
"loss": 1.2733,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.19890260631001372,
|
| 622 |
+
"grad_norm": 0.5378351616772565,
|
| 623 |
+
"learning_rate": 9.705700239897809e-06,
|
| 624 |
+
"loss": 1.2861,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.2011888431641518,
|
| 629 |
+
"grad_norm": 0.5372339490006793,
|
| 630 |
+
"learning_rate": 9.692060762657118e-06,
|
| 631 |
+
"loss": 1.2821,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.2034750800182899,
|
| 636 |
+
"grad_norm": 0.6353680076674888,
|
| 637 |
+
"learning_rate": 9.678122368228571e-06,
|
| 638 |
+
"loss": 1.2643,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.205761316872428,
|
| 643 |
+
"grad_norm": 0.6263499547366734,
|
| 644 |
+
"learning_rate": 9.66388594458568e-06,
|
| 645 |
+
"loss": 1.2826,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.20804755372656608,
|
| 650 |
+
"grad_norm": 0.6119180746423146,
|
| 651 |
+
"learning_rate": 9.649352398688506e-06,
|
| 652 |
+
"loss": 1.2856,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.21033379058070417,
|
| 657 |
+
"grad_norm": 0.6640618234127624,
|
| 658 |
+
"learning_rate": 9.634522656425885e-06,
|
| 659 |
+
"loss": 1.2765,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.21262002743484226,
|
| 664 |
+
"grad_norm": 0.6253602428713037,
|
| 665 |
+
"learning_rate": 9.619397662556434e-06,
|
| 666 |
+
"loss": 1.2661,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.21490626428898033,
|
| 671 |
+
"grad_norm": 0.6463257272674591,
|
| 672 |
+
"learning_rate": 9.603978380648375e-06,
|
| 673 |
+
"loss": 1.2838,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.21719250114311842,
|
| 678 |
+
"grad_norm": 0.6916869993480118,
|
| 679 |
+
"learning_rate": 9.588265793018141e-06,
|
| 680 |
+
"loss": 1.2785,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.2194787379972565,
|
| 685 |
+
"grad_norm": 0.578420093141111,
|
| 686 |
+
"learning_rate": 9.572260900667794e-06,
|
| 687 |
+
"loss": 1.2627,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.2217649748513946,
|
| 692 |
+
"grad_norm": 0.6016744117162259,
|
| 693 |
+
"learning_rate": 9.555964723221258e-06,
|
| 694 |
+
"loss": 1.2672,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.2240512117055327,
|
| 699 |
+
"grad_norm": 0.6325422647436533,
|
| 700 |
+
"learning_rate": 9.539378298859365e-06,
|
| 701 |
+
"loss": 1.2667,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.22633744855967078,
|
| 706 |
+
"grad_norm": 0.674420764332063,
|
| 707 |
+
"learning_rate": 9.522502684253709e-06,
|
| 708 |
+
"loss": 1.2601,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.22862368541380887,
|
| 713 |
+
"grad_norm": 0.6942742236531446,
|
| 714 |
+
"learning_rate": 9.505338954499332e-06,
|
| 715 |
+
"loss": 1.275,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.23090992226794696,
|
| 720 |
+
"grad_norm": 0.5661617220667517,
|
| 721 |
+
"learning_rate": 9.487888203046232e-06,
|
| 722 |
+
"loss": 1.2683,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.23319615912208505,
|
| 727 |
+
"grad_norm": 0.6389133947347537,
|
| 728 |
+
"learning_rate": 9.4701515416297e-06,
|
| 729 |
+
"loss": 1.2659,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.23548239597622314,
|
| 734 |
+
"grad_norm": 0.561786602813537,
|
| 735 |
+
"learning_rate": 9.452130100199504e-06,
|
| 736 |
+
"loss": 1.2664,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.23776863283036123,
|
| 741 |
+
"grad_norm": 0.5666699221383189,
|
| 742 |
+
"learning_rate": 9.433825026847891e-06,
|
| 743 |
+
"loss": 1.2573,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.24005486968449932,
|
| 748 |
+
"grad_norm": 0.6718711112993888,
|
| 749 |
+
"learning_rate": 9.415237487736452e-06,
|
| 750 |
+
"loss": 1.2545,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.2423411065386374,
|
| 755 |
+
"grad_norm": 0.5637527283960878,
|
| 756 |
+
"learning_rate": 9.396368667021835e-06,
|
| 757 |
+
"loss": 1.2723,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.2446273433927755,
|
| 762 |
+
"grad_norm": 0.583426898925874,
|
| 763 |
+
"learning_rate": 9.377219766780288e-06,
|
| 764 |
+
"loss": 1.2473,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.24691358024691357,
|
| 769 |
+
"grad_norm": 0.7422622561747031,
|
| 770 |
+
"learning_rate": 9.3577920069311e-06,
|
| 771 |
+
"loss": 1.2609,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.24919981710105166,
|
| 776 |
+
"grad_norm": 0.7536416453907702,
|
| 777 |
+
"learning_rate": 9.338086625158867e-06,
|
| 778 |
+
"loss": 1.2655,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.25148605395518975,
|
| 783 |
+
"grad_norm": 0.5911621999933799,
|
| 784 |
+
"learning_rate": 9.318104876834652e-06,
|
| 785 |
+
"loss": 1.2652,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.25377229080932784,
|
| 790 |
+
"grad_norm": 0.6482915887304207,
|
| 791 |
+
"learning_rate": 9.297848034936007e-06,
|
| 792 |
+
"loss": 1.2488,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.25605852766346593,
|
| 797 |
+
"grad_norm": 0.7813862221549358,
|
| 798 |
+
"learning_rate": 9.277317389965871e-06,
|
| 799 |
+
"loss": 1.2678,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.258344764517604,
|
| 804 |
+
"grad_norm": 0.601959447185496,
|
| 805 |
+
"learning_rate": 9.256514249870366e-06,
|
| 806 |
+
"loss": 1.2549,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2606310013717421,
|
| 811 |
+
"grad_norm": 0.5439593292691556,
|
| 812 |
+
"learning_rate": 9.235439939955458e-06,
|
| 813 |
+
"loss": 1.2311,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.2629172382258802,
|
| 818 |
+
"grad_norm": 0.6462948109732727,
|
| 819 |
+
"learning_rate": 9.214095802802533e-06,
|
| 820 |
+
"loss": 1.2605,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2652034750800183,
|
| 825 |
+
"grad_norm": 0.6523908850821281,
|
| 826 |
+
"learning_rate": 9.192483198182876e-06,
|
| 827 |
+
"loss": 1.2577,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.2674897119341564,
|
| 832 |
+
"grad_norm": 0.6285230592028435,
|
| 833 |
+
"learning_rate": 9.170603502971017e-06,
|
| 834 |
+
"loss": 1.233,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2697759487882945,
|
| 839 |
+
"grad_norm": 0.5990676661488948,
|
| 840 |
+
"learning_rate": 9.148458111057043e-06,
|
| 841 |
+
"loss": 1.2444,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.27206218564243256,
|
| 846 |
+
"grad_norm": 0.5443537881683997,
|
| 847 |
+
"learning_rate": 9.12604843325778e-06,
|
| 848 |
+
"loss": 1.2282,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.27434842249657065,
|
| 853 |
+
"grad_norm": 0.5804764131758829,
|
| 854 |
+
"learning_rate": 9.103375897226919e-06,
|
| 855 |
+
"loss": 1.253,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.27663465935070874,
|
| 860 |
+
"grad_norm": 0.5905170219986889,
|
| 861 |
+
"learning_rate": 9.080441947364065e-06,
|
| 862 |
+
"loss": 1.2472,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.27892089620484684,
|
| 867 |
+
"grad_norm": 0.6003218456115103,
|
| 868 |
+
"learning_rate": 9.057248044722718e-06,
|
| 869 |
+
"loss": 1.2421,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2812071330589849,
|
| 874 |
+
"grad_norm": 0.5683857920528798,
|
| 875 |
+
"learning_rate": 9.033795666917191e-06,
|
| 876 |
+
"loss": 1.2551,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.283493369913123,
|
| 881 |
+
"grad_norm": 0.5908776822300396,
|
| 882 |
+
"learning_rate": 9.010086308028487e-06,
|
| 883 |
+
"loss": 1.2375,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.2857796067672611,
|
| 888 |
+
"grad_norm": 0.6118010788168986,
|
| 889 |
+
"learning_rate": 8.986121478509096e-06,
|
| 890 |
+
"loss": 1.2347,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.2880658436213992,
|
| 895 |
+
"grad_norm": 0.5787813457678733,
|
| 896 |
+
"learning_rate": 8.961902705086785e-06,
|
| 897 |
+
"loss": 1.2395,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2903520804755373,
|
| 902 |
+
"grad_norm": 0.6290839595278495,
|
| 903 |
+
"learning_rate": 8.937431530667329e-06,
|
| 904 |
+
"loss": 1.2263,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.2926383173296754,
|
| 909 |
+
"grad_norm": 0.5459763353494508,
|
| 910 |
+
"learning_rate": 8.912709514236218e-06,
|
| 911 |
+
"loss": 1.2285,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.29492455418381347,
|
| 916 |
+
"grad_norm": 0.6301840515917086,
|
| 917 |
+
"learning_rate": 8.887738230759334e-06,
|
| 918 |
+
"loss": 1.2374,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.2972107910379515,
|
| 923 |
+
"grad_norm": 0.5413584040020849,
|
| 924 |
+
"learning_rate": 8.862519271082624e-06,
|
| 925 |
+
"loss": 1.2505,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.2994970278920896,
|
| 930 |
+
"grad_norm": 0.5979355091788396,
|
| 931 |
+
"learning_rate": 8.83705424183074e-06,
|
| 932 |
+
"loss": 1.2238,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3017832647462277,
|
| 937 |
+
"grad_norm": 0.6873493941298675,
|
| 938 |
+
"learning_rate": 8.811344765304698e-06,
|
| 939 |
+
"loss": 1.2262,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.3040695016003658,
|
| 944 |
+
"grad_norm": 0.6699975954695512,
|
| 945 |
+
"learning_rate": 8.785392479378522e-06,
|
| 946 |
+
"loss": 1.23,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.30635573845450387,
|
| 951 |
+
"grad_norm": 0.6860546025784545,
|
| 952 |
+
"learning_rate": 8.759199037394888e-06,
|
| 953 |
+
"loss": 1.2424,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.30864197530864196,
|
| 958 |
+
"grad_norm": 0.7598573834174616,
|
| 959 |
+
"learning_rate": 8.732766108059814e-06,
|
| 960 |
+
"loss": 1.2138,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.31092821216278005,
|
| 965 |
+
"grad_norm": 0.723323270057115,
|
| 966 |
+
"learning_rate": 8.70609537533634e-06,
|
| 967 |
+
"loss": 1.2373,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.31321444901691814,
|
| 972 |
+
"grad_norm": 0.6170455054157933,
|
| 973 |
+
"learning_rate": 8.679188538337248e-06,
|
| 974 |
+
"loss": 1.2257,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.31550068587105623,
|
| 979 |
+
"grad_norm": 0.7413957440287698,
|
| 980 |
+
"learning_rate": 8.652047311216823e-06,
|
| 981 |
+
"loss": 1.2075,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.3177869227251943,
|
| 986 |
+
"grad_norm": 0.7424365012242525,
|
| 987 |
+
"learning_rate": 8.62467342306164e-06,
|
| 988 |
+
"loss": 1.2238,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.3200731595793324,
|
| 993 |
+
"grad_norm": 0.8566227798899636,
|
| 994 |
+
"learning_rate": 8.597068617780419e-06,
|
| 995 |
+
"loss": 1.2278,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3223593964334705,
|
| 1000 |
+
"grad_norm": 0.647075376724737,
|
| 1001 |
+
"learning_rate": 8.569234653992916e-06,
|
| 1002 |
+
"loss": 1.2407,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.3246456332876086,
|
| 1007 |
+
"grad_norm": 0.6249088936722902,
|
| 1008 |
+
"learning_rate": 8.541173304917895e-06,
|
| 1009 |
+
"loss": 1.2231,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3269318701417467,
|
| 1014 |
+
"grad_norm": 0.70817264277616,
|
| 1015 |
+
"learning_rate": 8.512886358260162e-06,
|
| 1016 |
+
"loss": 1.2345,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.3292181069958848,
|
| 1021 |
+
"grad_norm": 0.5956107721750036,
|
| 1022 |
+
"learning_rate": 8.484375616096658e-06,
|
| 1023 |
+
"loss": 1.225,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.33150434385002286,
|
| 1028 |
+
"grad_norm": 0.6062042871270218,
|
| 1029 |
+
"learning_rate": 8.455642894761684e-06,
|
| 1030 |
+
"loss": 1.2185,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.33379058070416096,
|
| 1035 |
+
"grad_norm": 0.66611343630398,
|
| 1036 |
+
"learning_rate": 8.426690024731161e-06,
|
| 1037 |
+
"loss": 1.2171,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.33607681755829905,
|
| 1042 |
+
"grad_norm": 0.6006939272932527,
|
| 1043 |
+
"learning_rate": 8.39751885050603e-06,
|
| 1044 |
+
"loss": 1.2168,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.33836305441243714,
|
| 1049 |
+
"grad_norm": 0.5888998376074026,
|
| 1050 |
+
"learning_rate": 8.36813123049474e-06,
|
| 1051 |
+
"loss": 1.2447,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3406492912665752,
|
| 1056 |
+
"grad_norm": 0.6170255283448466,
|
| 1057 |
+
"learning_rate": 8.338529036894855e-06,
|
| 1058 |
+
"loss": 1.2386,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3429355281207133,
|
| 1063 |
+
"grad_norm": 0.6592250171561639,
|
| 1064 |
+
"learning_rate": 8.308714155573785e-06,
|
| 1065 |
+
"loss": 1.2095,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.3452217649748514,
|
| 1070 |
+
"grad_norm": 0.5948350472440084,
|
| 1071 |
+
"learning_rate": 8.278688485948634e-06,
|
| 1072 |
+
"loss": 1.2204,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3475080018289895,
|
| 1077 |
+
"grad_norm": 0.6884759018973265,
|
| 1078 |
+
"learning_rate": 8.248453940865204e-06,
|
| 1079 |
+
"loss": 1.2205,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.3497942386831276,
|
| 1084 |
+
"grad_norm": 0.5629453296642776,
|
| 1085 |
+
"learning_rate": 8.218012446476128e-06,
|
| 1086 |
+
"loss": 1.2087,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.3520804755372657,
|
| 1091 |
+
"grad_norm": 0.5703699859674032,
|
| 1092 |
+
"learning_rate": 8.187365942118162e-06,
|
| 1093 |
+
"loss": 1.2038,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.35436671239140377,
|
| 1098 |
+
"grad_norm": 0.5758055939006159,
|
| 1099 |
+
"learning_rate": 8.156516380188635e-06,
|
| 1100 |
+
"loss": 1.2015,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.35665294924554186,
|
| 1105 |
+
"grad_norm": 0.6814380489670292,
|
| 1106 |
+
"learning_rate": 8.125465726021068e-06,
|
| 1107 |
+
"loss": 1.2267,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.35893918609967995,
|
| 1112 |
+
"grad_norm": 0.58819101648096,
|
| 1113 |
+
"learning_rate": 8.09421595775997e-06,
|
| 1114 |
+
"loss": 1.2065,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.36122542295381804,
|
| 1119 |
+
"grad_norm": 0.599220106737159,
|
| 1120 |
+
"learning_rate": 8.062769066234807e-06,
|
| 1121 |
+
"loss": 1.2084,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3635116598079561,
|
| 1126 |
+
"grad_norm": 0.5687079813226833,
|
| 1127 |
+
"learning_rate": 8.031127054833192e-06,
|
| 1128 |
+
"loss": 1.2311,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.36579789666209417,
|
| 1133 |
+
"grad_norm": 0.6076443328436887,
|
| 1134 |
+
"learning_rate": 7.999291939373232e-06,
|
| 1135 |
+
"loss": 1.209,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.36808413351623226,
|
| 1140 |
+
"grad_norm": 0.5767468288489239,
|
| 1141 |
+
"learning_rate": 7.967265747975124e-06,
|
| 1142 |
+
"loss": 1.2153,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.37037037037037035,
|
| 1147 |
+
"grad_norm": 0.6275130557605428,
|
| 1148 |
+
"learning_rate": 7.93505052093194e-06,
|
| 1149 |
+
"loss": 1.2206,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.37265660722450844,
|
| 1154 |
+
"grad_norm": 0.5920904031157348,
|
| 1155 |
+
"learning_rate": 7.90264831057965e-06,
|
| 1156 |
+
"loss": 1.2149,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.37494284407864653,
|
| 1161 |
+
"grad_norm": 0.5841477404583847,
|
| 1162 |
+
"learning_rate": 7.870061181166372e-06,
|
| 1163 |
+
"loss": 1.2134,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3772290809327846,
|
| 1168 |
+
"grad_norm": 0.545565275285448,
|
| 1169 |
+
"learning_rate": 7.837291208720867e-06,
|
| 1170 |
+
"loss": 1.2185,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3795153177869227,
|
| 1175 |
+
"grad_norm": 0.6183231148929101,
|
| 1176 |
+
"learning_rate": 7.804340480920274e-06,
|
| 1177 |
+
"loss": 1.2064,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.3818015546410608,
|
| 1182 |
+
"grad_norm": 0.5801259298558049,
|
| 1183 |
+
"learning_rate": 7.771211096957125e-06,
|
| 1184 |
+
"loss": 1.2049,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3840877914951989,
|
| 1189 |
+
"grad_norm": 0.579347207611424,
|
| 1190 |
+
"learning_rate": 7.737905167405596e-06,
|
| 1191 |
+
"loss": 1.2185,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.386374028349337,
|
| 1196 |
+
"grad_norm": 0.6262921976973932,
|
| 1197 |
+
"learning_rate": 7.704424814087056e-06,
|
| 1198 |
+
"loss": 1.2137,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.3886602652034751,
|
| 1203 |
+
"grad_norm": 0.6070706881138944,
|
| 1204 |
+
"learning_rate": 7.670772169934902e-06,
|
| 1205 |
+
"loss": 1.2177,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.39094650205761317,
|
| 1210 |
+
"grad_norm": 0.5688216055326876,
|
| 1211 |
+
"learning_rate": 7.636949378858647e-06,
|
| 1212 |
+
"loss": 1.2016,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.39323273891175126,
|
| 1217 |
+
"grad_norm": 0.6166249078020826,
|
| 1218 |
+
"learning_rate": 7.602958595607375e-06,
|
| 1219 |
+
"loss": 1.1957,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.39551897576588935,
|
| 1224 |
+
"grad_norm": 0.5778886288472463,
|
| 1225 |
+
"learning_rate": 7.568801985632439e-06,
|
| 1226 |
+
"loss": 1.2105,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.39780521262002744,
|
| 1231 |
+
"grad_norm": 0.6732218435967291,
|
| 1232 |
+
"learning_rate": 7.5344817249495195e-06,
|
| 1233 |
+
"loss": 1.2047,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.40009144947416553,
|
| 1238 |
+
"grad_norm": 0.672208759556888,
|
| 1239 |
+
"learning_rate": 7.500000000000001e-06,
|
| 1240 |
+
"loss": 1.1854,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.4023776863283036,
|
| 1245 |
+
"grad_norm": 0.6180565492464766,
|
| 1246 |
+
"learning_rate": 7.465359007511667e-06,
|
| 1247 |
+
"loss": 1.185,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.4046639231824417,
|
| 1252 |
+
"grad_norm": 0.6266745151721254,
|
| 1253 |
+
"learning_rate": 7.430560954358764e-06,
|
| 1254 |
+
"loss": 1.2082,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.4069501600365798,
|
| 1259 |
+
"grad_norm": 0.6163182978581346,
|
| 1260 |
+
"learning_rate": 7.395608057421406e-06,
|
| 1261 |
+
"loss": 1.2194,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.4092363968907179,
|
| 1266 |
+
"grad_norm": 0.6262674693601461,
|
| 1267 |
+
"learning_rate": 7.360502543444339e-06,
|
| 1268 |
+
"loss": 1.2188,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.411522633744856,
|
| 1273 |
+
"grad_norm": 0.5549642780561265,
|
| 1274 |
+
"learning_rate": 7.325246648895089e-06,
|
| 1275 |
+
"loss": 1.1986,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.41380887059899407,
|
| 1280 |
+
"grad_norm": 0.5540368046559051,
|
| 1281 |
+
"learning_rate": 7.289842619821475e-06,
|
| 1282 |
+
"loss": 1.2175,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.41609510745313216,
|
| 1287 |
+
"grad_norm": 0.587023330497459,
|
| 1288 |
+
"learning_rate": 7.254292711708529e-06,
|
| 1289 |
+
"loss": 1.2029,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.41838134430727025,
|
| 1294 |
+
"grad_norm": 0.5513581130094706,
|
| 1295 |
+
"learning_rate": 7.218599189334799e-06,
|
| 1296 |
+
"loss": 1.2009,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.42066758116140834,
|
| 1301 |
+
"grad_norm": 0.7237520794327035,
|
| 1302 |
+
"learning_rate": 7.182764326628068e-06,
|
| 1303 |
+
"loss": 1.2063,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.42295381801554643,
|
| 1308 |
+
"grad_norm": 0.5476819110298711,
|
| 1309 |
+
"learning_rate": 7.146790406520491e-06,
|
| 1310 |
+
"loss": 1.2107,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.4252400548696845,
|
| 1315 |
+
"grad_norm": 0.5753924094787153,
|
| 1316 |
+
"learning_rate": 7.1106797208031554e-06,
|
| 1317 |
+
"loss": 1.2133,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.4275262917238226,
|
| 1322 |
+
"grad_norm": 0.6489054914059448,
|
| 1323 |
+
"learning_rate": 7.0744345699800755e-06,
|
| 1324 |
+
"loss": 1.1991,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.42981252857796065,
|
| 1329 |
+
"grad_norm": 0.6239602498665449,
|
| 1330 |
+
"learning_rate": 7.038057263121639e-06,
|
| 1331 |
+
"loss": 1.1937,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.43209876543209874,
|
| 1336 |
+
"grad_norm": 0.5954140813357963,
|
| 1337 |
+
"learning_rate": 7.001550117717499e-06,
|
| 1338 |
+
"loss": 1.2092,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.43438500228623683,
|
| 1343 |
+
"grad_norm": 0.5953175778315464,
|
| 1344 |
+
"learning_rate": 6.9649154595289326e-06,
|
| 1345 |
+
"loss": 1.1957,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.4366712391403749,
|
| 1350 |
+
"grad_norm": 0.6030938627687562,
|
| 1351 |
+
"learning_rate": 6.92815562244068e-06,
|
| 1352 |
+
"loss": 1.1827,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.438957475994513,
|
| 1357 |
+
"grad_norm": 0.6882999466791362,
|
| 1358 |
+
"learning_rate": 6.891272948312251e-06,
|
| 1359 |
+
"loss": 1.2102,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.4412437128486511,
|
| 1364 |
+
"grad_norm": 0.6080281045836577,
|
| 1365 |
+
"learning_rate": 6.854269786828741e-06,
|
| 1366 |
+
"loss": 1.2093,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.4435299497027892,
|
| 1371 |
+
"grad_norm": 0.756192409869553,
|
| 1372 |
+
"learning_rate": 6.817148495351131e-06,
|
| 1373 |
+
"loss": 1.2159,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.4458161865569273,
|
| 1378 |
+
"grad_norm": 0.5892520162590819,
|
| 1379 |
+
"learning_rate": 6.779911438766117e-06,
|
| 1380 |
+
"loss": 1.193,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4481024234110654,
|
| 1385 |
+
"grad_norm": 0.6265917897470434,
|
| 1386 |
+
"learning_rate": 6.742560989335438e-06,
|
| 1387 |
+
"loss": 1.1951,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.45038866026520347,
|
| 1392 |
+
"grad_norm": 0.5927415516536023,
|
| 1393 |
+
"learning_rate": 6.705099526544757e-06,
|
| 1394 |
+
"loss": 1.1973,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.45267489711934156,
|
| 1399 |
+
"grad_norm": 0.5602604942191215,
|
| 1400 |
+
"learning_rate": 6.667529436952064e-06,
|
| 1401 |
+
"loss": 1.1945,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.45496113397347965,
|
| 1406 |
+
"grad_norm": 0.751574883051813,
|
| 1407 |
+
"learning_rate": 6.629853114035643e-06,
|
| 1408 |
+
"loss": 1.2134,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.45724737082761774,
|
| 1413 |
+
"grad_norm": 0.6000318274839507,
|
| 1414 |
+
"learning_rate": 6.5920729580415795e-06,
|
| 1415 |
+
"loss": 1.2104,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.45953360768175583,
|
| 1420 |
+
"grad_norm": 0.5783065549399249,
|
| 1421 |
+
"learning_rate": 6.554191375830861e-06,
|
| 1422 |
+
"loss": 1.2016,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4618198445358939,
|
| 1427 |
+
"grad_norm": 0.5751980188798808,
|
| 1428 |
+
"learning_rate": 6.516210780726032e-06,
|
| 1429 |
+
"loss": 1.1794,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.464106081390032,
|
| 1434 |
+
"grad_norm": 0.6096335885035103,
|
| 1435 |
+
"learning_rate": 6.478133592357455e-06,
|
| 1436 |
+
"loss": 1.1816,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.4663923182441701,
|
| 1441 |
+
"grad_norm": 0.5848690144740822,
|
| 1442 |
+
"learning_rate": 6.43996223650916e-06,
|
| 1443 |
+
"loss": 1.1735,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4686785550983082,
|
| 1448 |
+
"grad_norm": 0.6273777569367492,
|
| 1449 |
+
"learning_rate": 6.401699144964306e-06,
|
| 1450 |
+
"loss": 1.1864,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.4709647919524463,
|
| 1455 |
+
"grad_norm": 0.5772389229176554,
|
| 1456 |
+
"learning_rate": 6.3633467553502625e-06,
|
| 1457 |
+
"loss": 1.1953,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.4732510288065844,
|
| 1462 |
+
"grad_norm": 0.6320660706578101,
|
| 1463 |
+
"learning_rate": 6.32490751098331e-06,
|
| 1464 |
+
"loss": 1.1778,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.47553726566072246,
|
| 1469 |
+
"grad_norm": 0.628014857385664,
|
| 1470 |
+
"learning_rate": 6.286383860712982e-06,
|
| 1471 |
+
"loss": 1.1978,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.47782350251486055,
|
| 1476 |
+
"grad_norm": 0.6165011857453245,
|
| 1477 |
+
"learning_rate": 6.247778258766069e-06,
|
| 1478 |
+
"loss": 1.1783,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.48010973936899864,
|
| 1483 |
+
"grad_norm": 0.6680859473813631,
|
| 1484 |
+
"learning_rate": 6.209093164590253e-06,
|
| 1485 |
+
"loss": 1.1883,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.48239597622313674,
|
| 1490 |
+
"grad_norm": 0.6230269069079273,
|
| 1491 |
+
"learning_rate": 6.170331042697425e-06,
|
| 1492 |
+
"loss": 1.1923,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.4846822130772748,
|
| 1497 |
+
"grad_norm": 0.6472681484163015,
|
| 1498 |
+
"learning_rate": 6.131494362506693e-06,
|
| 1499 |
+
"loss": 1.1826,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4869684499314129,
|
| 1504 |
+
"grad_norm": 0.6799978087591872,
|
| 1505 |
+
"learning_rate": 6.09258559818704e-06,
|
| 1506 |
+
"loss": 1.1829,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.489254686785551,
|
| 1511 |
+
"grad_norm": 0.5617426984448537,
|
| 1512 |
+
"learning_rate": 6.053607228499719e-06,
|
| 1513 |
+
"loss": 1.1941,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.4915409236396891,
|
| 1518 |
+
"grad_norm": 0.6444058153599652,
|
| 1519 |
+
"learning_rate": 6.014561736640334e-06,
|
| 1520 |
+
"loss": 1.2,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.49382716049382713,
|
| 1525 |
+
"grad_norm": 0.6016265988080601,
|
| 1526 |
+
"learning_rate": 5.975451610080643e-06,
|
| 1527 |
+
"loss": 1.1655,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.4961133973479652,
|
| 1532 |
+
"grad_norm": 0.7053148286233416,
|
| 1533 |
+
"learning_rate": 5.936279340410082e-06,
|
| 1534 |
+
"loss": 1.172,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.4983996342021033,
|
| 1539 |
+
"grad_norm": 0.5586357561653685,
|
| 1540 |
+
"learning_rate": 5.8970474231770445e-06,
|
| 1541 |
+
"loss": 1.1922,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5006858710562414,
|
| 1546 |
+
"grad_norm": 0.7895760074140119,
|
| 1547 |
+
"learning_rate": 5.857758357729892e-06,
|
| 1548 |
+
"loss": 1.1839,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5029721079103795,
|
| 1553 |
+
"grad_norm": 0.7313666592611404,
|
| 1554 |
+
"learning_rate": 5.8184146470577265e-06,
|
| 1555 |
+
"loss": 1.1813,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.5052583447645176,
|
| 1560 |
+
"grad_norm": 0.6067591576327228,
|
| 1561 |
+
"learning_rate": 5.779018797630934e-06,
|
| 1562 |
+
"loss": 1.1855,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5075445816186557,
|
| 1567 |
+
"grad_norm": 0.6144330199450508,
|
| 1568 |
+
"learning_rate": 5.739573319241505e-06,
|
| 1569 |
+
"loss": 1.1924,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5098308184727938,
|
| 1574 |
+
"grad_norm": 0.6075048668745815,
|
| 1575 |
+
"learning_rate": 5.7000807248431466e-06,
|
| 1576 |
+
"loss": 1.1783,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5121170553269319,
|
| 1581 |
+
"grad_norm": 0.6763365315316732,
|
| 1582 |
+
"learning_rate": 5.66054353039118e-06,
|
| 1583 |
+
"loss": 1.1873,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.51440329218107,
|
| 1588 |
+
"grad_norm": 0.652936999197392,
|
| 1589 |
+
"learning_rate": 5.620964254682267e-06,
|
| 1590 |
+
"loss": 1.2019,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.516689529035208,
|
| 1595 |
+
"grad_norm": 0.7510930690144121,
|
| 1596 |
+
"learning_rate": 5.58134541919394e-06,
|
| 1597 |
+
"loss": 1.1863,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5189757658893461,
|
| 1602 |
+
"grad_norm": 0.7485282723991191,
|
| 1603 |
+
"learning_rate": 5.5416895479239665e-06,
|
| 1604 |
+
"loss": 1.1878,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5212620027434842,
|
| 1609 |
+
"grad_norm": 0.6650793765929232,
|
| 1610 |
+
"learning_rate": 5.501999167229554e-06,
|
| 1611 |
+
"loss": 1.1844,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.5235482395976223,
|
| 1616 |
+
"grad_norm": 0.6617004106280673,
|
| 1617 |
+
"learning_rate": 5.4622768056664e-06,
|
| 1618 |
+
"loss": 1.1819,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.5258344764517604,
|
| 1623 |
+
"grad_norm": 0.639306148093516,
|
| 1624 |
+
"learning_rate": 5.42252499382761e-06,
|
| 1625 |
+
"loss": 1.1844,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5281207133058985,
|
| 1630 |
+
"grad_norm": 0.590573720499581,
|
| 1631 |
+
"learning_rate": 5.38274626418248e-06,
|
| 1632 |
+
"loss": 1.1848,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5304069501600366,
|
| 1637 |
+
"grad_norm": 0.625235396788826,
|
| 1638 |
+
"learning_rate": 5.3429431509151515e-06,
|
| 1639 |
+
"loss": 1.1904,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.5326931870141747,
|
| 1644 |
+
"grad_norm": 0.5840052674712635,
|
| 1645 |
+
"learning_rate": 5.303118189763187e-06,
|
| 1646 |
+
"loss": 1.1829,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5349794238683128,
|
| 1651 |
+
"grad_norm": 0.5940842973816081,
|
| 1652 |
+
"learning_rate": 5.263273917856e-06,
|
| 1653 |
+
"loss": 1.1774,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.5372656607224509,
|
| 1658 |
+
"grad_norm": 0.5991239115995499,
|
| 1659 |
+
"learning_rate": 5.22341287355324e-06,
|
| 1660 |
+
"loss": 1.1857,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.539551897576589,
|
| 1665 |
+
"grad_norm": 0.6248756548437343,
|
| 1666 |
+
"learning_rate": 5.183537596283075e-06,
|
| 1667 |
+
"loss": 1.1799,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.541838134430727,
|
| 1672 |
+
"grad_norm": 0.6023807247895316,
|
| 1673 |
+
"learning_rate": 5.143650626380417e-06,
|
| 1674 |
+
"loss": 1.1858,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.5441243712848651,
|
| 1679 |
+
"grad_norm": 0.6101959497751839,
|
| 1680 |
+
"learning_rate": 5.103754504925071e-06,
|
| 1681 |
+
"loss": 1.1961,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5464106081390032,
|
| 1686 |
+
"grad_norm": 0.569676114190435,
|
| 1687 |
+
"learning_rate": 5.06385177357987e-06,
|
| 1688 |
+
"loss": 1.1766,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5486968449931413,
|
| 1693 |
+
"grad_norm": 0.5819652008689743,
|
| 1694 |
+
"learning_rate": 5.023944974428739e-06,
|
| 1695 |
+
"loss": 1.1734,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5509830818472794,
|
| 1700 |
+
"grad_norm": 0.5661449507234365,
|
| 1701 |
+
"learning_rate": 4.9840366498147495e-06,
|
| 1702 |
+
"loss": 1.1908,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5532693187014175,
|
| 1707 |
+
"grad_norm": 0.6109491726102372,
|
| 1708 |
+
"learning_rate": 4.944129342178156e-06,
|
| 1709 |
+
"loss": 1.1784,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5555555555555556,
|
| 1714 |
+
"grad_norm": 0.5811074689104263,
|
| 1715 |
+
"learning_rate": 4.90422559389443e-06,
|
| 1716 |
+
"loss": 1.1746,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5578417924096937,
|
| 1721 |
+
"grad_norm": 0.6060458081756667,
|
| 1722 |
+
"learning_rate": 4.864327947112281e-06,
|
| 1723 |
+
"loss": 1.195,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5601280292638318,
|
| 1728 |
+
"grad_norm": 0.6226718536570417,
|
| 1729 |
+
"learning_rate": 4.82443894359171e-06,
|
| 1730 |
+
"loss": 1.1786,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5624142661179699,
|
| 1735 |
+
"grad_norm": 0.5995864510713481,
|
| 1736 |
+
"learning_rate": 4.784561124542088e-06,
|
| 1737 |
+
"loss": 1.1791,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5647005029721079,
|
| 1742 |
+
"grad_norm": 0.5701958838449743,
|
| 1743 |
+
"learning_rate": 4.744697030460248e-06,
|
| 1744 |
+
"loss": 1.1647,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.566986739826246,
|
| 1749 |
+
"grad_norm": 0.6293939505655973,
|
| 1750 |
+
"learning_rate": 4.7048492009686525e-06,
|
| 1751 |
+
"loss": 1.1692,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5692729766803841,
|
| 1756 |
+
"grad_norm": 0.6850447194966206,
|
| 1757 |
+
"learning_rate": 4.6650201746535926e-06,
|
| 1758 |
+
"loss": 1.1673,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5715592135345222,
|
| 1763 |
+
"grad_norm": 0.6040120516739561,
|
| 1764 |
+
"learning_rate": 4.625212488903467e-06,
|
| 1765 |
+
"loss": 1.1834,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5738454503886603,
|
| 1770 |
+
"grad_norm": 0.5686706476550618,
|
| 1771 |
+
"learning_rate": 4.585428679747133e-06,
|
| 1772 |
+
"loss": 1.1716,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5761316872427984,
|
| 1777 |
+
"grad_norm": 0.5946931657837966,
|
| 1778 |
+
"learning_rate": 4.545671281692331e-06,
|
| 1779 |
+
"loss": 1.1705,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5784179240969365,
|
| 1784 |
+
"grad_norm": 0.6120143356512502,
|
| 1785 |
+
"learning_rate": 4.505942827564242e-06,
|
| 1786 |
+
"loss": 1.1807,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5807041609510746,
|
| 1791 |
+
"grad_norm": 0.6341171747185648,
|
| 1792 |
+
"learning_rate": 4.466245848344106e-06,
|
| 1793 |
+
"loss": 1.1839,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5829903978052127,
|
| 1798 |
+
"grad_norm": 0.6494090868678567,
|
| 1799 |
+
"learning_rate": 4.426582873007999e-06,
|
| 1800 |
+
"loss": 1.1684,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5852766346593508,
|
| 1805 |
+
"grad_norm": 0.6252524175950205,
|
| 1806 |
+
"learning_rate": 4.386956428365701e-06,
|
| 1807 |
+
"loss": 1.1878,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5875628715134888,
|
| 1812 |
+
"grad_norm": 0.5911175497758677,
|
| 1813 |
+
"learning_rate": 4.347369038899744e-06,
|
| 1814 |
+
"loss": 1.1828,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5898491083676269,
|
| 1819 |
+
"grad_norm": 0.5988939599453593,
|
| 1820 |
+
"learning_rate": 4.307823226604555e-06,
|
| 1821 |
+
"loss": 1.1735,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5921353452217649,
|
| 1826 |
+
"grad_norm": 0.5813355536422021,
|
| 1827 |
+
"learning_rate": 4.2683215108258145e-06,
|
| 1828 |
+
"loss": 1.1706,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.594421582075903,
|
| 1833 |
+
"grad_norm": 0.6208043705991068,
|
| 1834 |
+
"learning_rate": 4.228866408099945e-06,
|
| 1835 |
+
"loss": 1.1907,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5967078189300411,
|
| 1840 |
+
"grad_norm": 0.6512006631857741,
|
| 1841 |
+
"learning_rate": 4.189460431993788e-06,
|
| 1842 |
+
"loss": 1.1951,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.5989940557841792,
|
| 1847 |
+
"grad_norm": 0.5845471180993255,
|
| 1848 |
+
"learning_rate": 4.150106092944475e-06,
|
| 1849 |
+
"loss": 1.1717,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6012802926383173,
|
| 1854 |
+
"grad_norm": 0.5949045334275538,
|
| 1855 |
+
"learning_rate": 4.110805898099492e-06,
|
| 1856 |
+
"loss": 1.1833,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6035665294924554,
|
| 1861 |
+
"grad_norm": 0.5971913414181261,
|
| 1862 |
+
"learning_rate": 4.071562351156966e-06,
|
| 1863 |
+
"loss": 1.1786,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6058527663465935,
|
| 1868 |
+
"grad_norm": 0.6178601149254982,
|
| 1869 |
+
"learning_rate": 4.032377952206148e-06,
|
| 1870 |
+
"loss": 1.1793,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6081390032007316,
|
| 1875 |
+
"grad_norm": 0.6046188006147395,
|
| 1876 |
+
"learning_rate": 3.993255197568154e-06,
|
| 1877 |
+
"loss": 1.169,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.6104252400548696,
|
| 1882 |
+
"grad_norm": 0.5919458656130715,
|
| 1883 |
+
"learning_rate": 3.954196579636918e-06,
|
| 1884 |
+
"loss": 1.1692,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.6127114769090077,
|
| 1889 |
+
"grad_norm": 0.5727049539306068,
|
| 1890 |
+
"learning_rate": 3.91520458672042e-06,
|
| 1891 |
+
"loss": 1.1747,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6149977137631458,
|
| 1896 |
+
"grad_norm": 0.6040809405921704,
|
| 1897 |
+
"learning_rate": 3.876281702882156e-06,
|
| 1898 |
+
"loss": 1.1935,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6172839506172839,
|
| 1903 |
+
"grad_norm": 0.5747789602798682,
|
| 1904 |
+
"learning_rate": 3.837430407782896e-06,
|
| 1905 |
+
"loss": 1.175,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.619570187471422,
|
| 1910 |
+
"grad_norm": 0.6001909994942644,
|
| 1911 |
+
"learning_rate": 3.7986531765226965e-06,
|
| 1912 |
+
"loss": 1.1718,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6218564243255601,
|
| 1917 |
+
"grad_norm": 0.5499338552551708,
|
| 1918 |
+
"learning_rate": 3.759952479483232e-06,
|
| 1919 |
+
"loss": 1.1615,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6241426611796982,
|
| 1924 |
+
"grad_norm": 0.62697610396954,
|
| 1925 |
+
"learning_rate": 3.7213307821704115e-06,
|
| 1926 |
+
"loss": 1.1616,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6264288980338363,
|
| 1931 |
+
"grad_norm": 0.637904015143814,
|
| 1932 |
+
"learning_rate": 3.6827905450573022e-06,
|
| 1933 |
+
"loss": 1.1784,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6287151348879744,
|
| 1938 |
+
"grad_norm": 0.6235229612947039,
|
| 1939 |
+
"learning_rate": 3.6443342234273905e-06,
|
| 1940 |
+
"loss": 1.1674,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6310013717421125,
|
| 1945 |
+
"grad_norm": 0.744429415227132,
|
| 1946 |
+
"learning_rate": 3.6059642672181537e-06,
|
| 1947 |
+
"loss": 1.1678,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.6332876085962506,
|
| 1952 |
+
"grad_norm": 0.5903117671660288,
|
| 1953 |
+
"learning_rate": 3.5676831208649887e-06,
|
| 1954 |
+
"loss": 1.1661,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6355738454503886,
|
| 1959 |
+
"grad_norm": 0.5977435348831742,
|
| 1960 |
+
"learning_rate": 3.5294932231454838e-06,
|
| 1961 |
+
"loss": 1.1655,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6378600823045267,
|
| 1966 |
+
"grad_norm": 0.6262251229258455,
|
| 1967 |
+
"learning_rate": 3.4913970070240388e-06,
|
| 1968 |
+
"loss": 1.1827,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6401463191586648,
|
| 1973 |
+
"grad_norm": 0.6039362156672261,
|
| 1974 |
+
"learning_rate": 3.4533968994968913e-06,
|
| 1975 |
+
"loss": 1.162,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6424325560128029,
|
| 1980 |
+
"grad_norm": 0.610471777862986,
|
| 1981 |
+
"learning_rate": 3.41549532143748e-06,
|
| 1982 |
+
"loss": 1.1719,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.644718792866941,
|
| 1987 |
+
"grad_norm": 0.6124948412563855,
|
| 1988 |
+
"learning_rate": 3.3776946874422268e-06,
|
| 1989 |
+
"loss": 1.161,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6470050297210791,
|
| 1994 |
+
"grad_norm": 0.596054515528405,
|
| 1995 |
+
"learning_rate": 3.3399974056767095e-06,
|
| 1996 |
+
"loss": 1.1677,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6492912665752172,
|
| 2001 |
+
"grad_norm": 0.6199519548446956,
|
| 2002 |
+
"learning_rate": 3.30240587772224e-06,
|
| 2003 |
+
"loss": 1.1731,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6515775034293553,
|
| 2008 |
+
"grad_norm": 0.6123382818220521,
|
| 2009 |
+
"learning_rate": 3.2649224984228756e-06,
|
| 2010 |
+
"loss": 1.1751,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6538637402834934,
|
| 2015 |
+
"grad_norm": 0.6521756883889377,
|
| 2016 |
+
"learning_rate": 3.227549655732843e-06,
|
| 2017 |
+
"loss": 1.1746,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6561499771376315,
|
| 2022 |
+
"grad_norm": 0.6292502440238857,
|
| 2023 |
+
"learning_rate": 3.19028973056441e-06,
|
| 2024 |
+
"loss": 1.1796,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6584362139917695,
|
| 2029 |
+
"grad_norm": 0.7223300006546375,
|
| 2030 |
+
"learning_rate": 3.153145096636211e-06,
|
| 2031 |
+
"loss": 1.1769,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6607224508459076,
|
| 2036 |
+
"grad_norm": 0.6123252900962536,
|
| 2037 |
+
"learning_rate": 3.1161181203220146e-06,
|
| 2038 |
+
"loss": 1.1798,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6630086877000457,
|
| 2043 |
+
"grad_norm": 0.6176590524451245,
|
| 2044 |
+
"learning_rate": 3.079211160499975e-06,
|
| 2045 |
+
"loss": 1.1628,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6652949245541838,
|
| 2050 |
+
"grad_norm": 0.6851380779593121,
|
| 2051 |
+
"learning_rate": 3.0424265684023556e-06,
|
| 2052 |
+
"loss": 1.1621,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6675811614083219,
|
| 2057 |
+
"grad_norm": 0.6135186798564677,
|
| 2058 |
+
"learning_rate": 3.0057666874657365e-06,
|
| 2059 |
+
"loss": 1.1817,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.66986739826246,
|
| 2064 |
+
"grad_norm": 0.6162664151552476,
|
| 2065 |
+
"learning_rate": 2.9692338531817205e-06,
|
| 2066 |
+
"loss": 1.1621,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6721536351165981,
|
| 2071 |
+
"grad_norm": 0.6209879083469707,
|
| 2072 |
+
"learning_rate": 2.9328303929481507e-06,
|
| 2073 |
+
"loss": 1.1788,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6744398719707362,
|
| 2078 |
+
"grad_norm": 0.6564960801220917,
|
| 2079 |
+
"learning_rate": 2.8965586259208295e-06,
|
| 2080 |
+
"loss": 1.1497,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6767261088248743,
|
| 2085 |
+
"grad_norm": 0.6100366044161921,
|
| 2086 |
+
"learning_rate": 2.860420862865787e-06,
|
| 2087 |
+
"loss": 1.1641,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6790123456790124,
|
| 2092 |
+
"grad_norm": 0.6401282278697755,
|
| 2093 |
+
"learning_rate": 2.82441940601205e-06,
|
| 2094 |
+
"loss": 1.1647,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6812985825331505,
|
| 2099 |
+
"grad_norm": 0.5948814066139619,
|
| 2100 |
+
"learning_rate": 2.7885565489049948e-06,
|
| 2101 |
+
"loss": 1.1862,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6835848193872885,
|
| 2106 |
+
"grad_norm": 0.575891260626997,
|
| 2107 |
+
"learning_rate": 2.7528345762602125e-06,
|
| 2108 |
+
"loss": 1.149,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6858710562414266,
|
| 2113 |
+
"grad_norm": 0.6321328549868929,
|
| 2114 |
+
"learning_rate": 2.7172557638179674e-06,
|
| 2115 |
+
"loss": 1.1722,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6881572930955647,
|
| 2120 |
+
"grad_norm": 0.620537429422375,
|
| 2121 |
+
"learning_rate": 2.681822378198221e-06,
|
| 2122 |
+
"loss": 1.1667,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6904435299497028,
|
| 2127 |
+
"grad_norm": 0.5916688359774108,
|
| 2128 |
+
"learning_rate": 2.6465366767562162e-06,
|
| 2129 |
+
"loss": 1.1742,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6927297668038409,
|
| 2134 |
+
"grad_norm": 0.649532932905328,
|
| 2135 |
+
"learning_rate": 2.611400907438685e-06,
|
| 2136 |
+
"loss": 1.1664,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.695016003657979,
|
| 2141 |
+
"grad_norm": 0.5887639490410209,
|
| 2142 |
+
"learning_rate": 2.5764173086406306e-06,
|
| 2143 |
+
"loss": 1.1684,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6973022405121171,
|
| 2148 |
+
"grad_norm": 0.5909674256777088,
|
| 2149 |
+
"learning_rate": 2.5415881090627227e-06,
|
| 2150 |
+
"loss": 1.1681,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6995884773662552,
|
| 2155 |
+
"grad_norm": 0.6669572713903603,
|
| 2156 |
+
"learning_rate": 2.506915527569318e-06,
|
| 2157 |
+
"loss": 1.1692,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.7018747142203933,
|
| 2162 |
+
"grad_norm": 0.6291006193664693,
|
| 2163 |
+
"learning_rate": 2.472401773047107e-06,
|
| 2164 |
+
"loss": 1.1707,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7041609510745314,
|
| 2169 |
+
"grad_norm": 0.6241336853751712,
|
| 2170 |
+
"learning_rate": 2.438049044264382e-06,
|
| 2171 |
+
"loss": 1.1763,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7064471879286695,
|
| 2176 |
+
"grad_norm": 0.6233093811845397,
|
| 2177 |
+
"learning_rate": 2.4038595297309712e-06,
|
| 2178 |
+
"loss": 1.1595,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7087334247828075,
|
| 2183 |
+
"grad_norm": 0.6099376654855213,
|
| 2184 |
+
"learning_rate": 2.3698354075588105e-06,
|
| 2185 |
+
"loss": 1.1815,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7110196616369456,
|
| 2190 |
+
"grad_norm": 0.608739940642273,
|
| 2191 |
+
"learning_rate": 2.3359788453231723e-06,
|
| 2192 |
+
"loss": 1.1558,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7133058984910837,
|
| 2197 |
+
"grad_norm": 0.6060804682823651,
|
| 2198 |
+
"learning_rate": 2.3022919999245964e-06,
|
| 2199 |
+
"loss": 1.1737,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7155921353452218,
|
| 2204 |
+
"grad_norm": 0.6554029837627439,
|
| 2205 |
+
"learning_rate": 2.2687770174514674e-06,
|
| 2206 |
+
"loss": 1.1763,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7178783721993599,
|
| 2211 |
+
"grad_norm": 0.6199763037940721,
|
| 2212 |
+
"learning_rate": 2.23543603304329e-06,
|
| 2213 |
+
"loss": 1.1668,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.720164609053498,
|
| 2218 |
+
"grad_norm": 0.7002533112076955,
|
| 2219 |
+
"learning_rate": 2.20227117075468e-06,
|
| 2220 |
+
"loss": 1.1717,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7224508459076361,
|
| 2225 |
+
"grad_norm": 0.5685258465602809,
|
| 2226 |
+
"learning_rate": 2.1692845434200323e-06,
|
| 2227 |
+
"loss": 1.1793,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7247370827617741,
|
| 2232 |
+
"grad_norm": 0.5988803647429354,
|
| 2233 |
+
"learning_rate": 2.136478252518924e-06,
|
| 2234 |
+
"loss": 1.1762,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7270233196159122,
|
| 2239 |
+
"grad_norm": 0.6220944262982843,
|
| 2240 |
+
"learning_rate": 2.103854388042243e-06,
|
| 2241 |
+
"loss": 1.1732,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7293095564700502,
|
| 2246 |
+
"grad_norm": 0.5872374752551915,
|
| 2247 |
+
"learning_rate": 2.071415028359026e-06,
|
| 2248 |
+
"loss": 1.1653,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7315957933241883,
|
| 2253 |
+
"grad_norm": 0.6315378201627972,
|
| 2254 |
+
"learning_rate": 2.0391622400840665e-06,
|
| 2255 |
+
"loss": 1.1631,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7338820301783264,
|
| 2260 |
+
"grad_norm": 0.6166479295990325,
|
| 2261 |
+
"learning_rate": 2.0070980779462513e-06,
|
| 2262 |
+
"loss": 1.1632,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7361682670324645,
|
| 2267 |
+
"grad_norm": 0.6082820756952414,
|
| 2268 |
+
"learning_rate": 1.975224584657648e-06,
|
| 2269 |
+
"loss": 1.1609,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7384545038866026,
|
| 2274 |
+
"grad_norm": 0.5711567863660318,
|
| 2275 |
+
"learning_rate": 1.943543790783392e-06,
|
| 2276 |
+
"loss": 1.1629,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.7407407407407407,
|
| 2281 |
+
"grad_norm": 0.5934876997772376,
|
| 2282 |
+
"learning_rate": 1.9120577146123125e-06,
|
| 2283 |
+
"loss": 1.1711,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7430269775948788,
|
| 2288 |
+
"grad_norm": 0.6044258229955937,
|
| 2289 |
+
"learning_rate": 1.8807683620283496e-06,
|
| 2290 |
+
"loss": 1.1792,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7453132144490169,
|
| 2295 |
+
"grad_norm": 0.6414108282805848,
|
| 2296 |
+
"learning_rate": 1.8496777263827775e-06,
|
| 2297 |
+
"loss": 1.1909,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.747599451303155,
|
| 2302 |
+
"grad_norm": 0.5928077840962543,
|
| 2303 |
+
"learning_rate": 1.8187877883672024e-06,
|
| 2304 |
+
"loss": 1.177,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7498856881572931,
|
| 2309 |
+
"grad_norm": 0.5674967348667851,
|
| 2310 |
+
"learning_rate": 1.7881005158873826e-06,
|
| 2311 |
+
"loss": 1.1698,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7521719250114312,
|
| 2316 |
+
"grad_norm": 0.6190325214784786,
|
| 2317 |
+
"learning_rate": 1.757617863937865e-06,
|
| 2318 |
+
"loss": 1.1564,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7544581618655692,
|
| 2323 |
+
"grad_norm": 0.5994621485851359,
|
| 2324 |
+
"learning_rate": 1.7273417744774323e-06,
|
| 2325 |
+
"loss": 1.1682,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7567443987197073,
|
| 2330 |
+
"grad_norm": 0.6486512119864596,
|
| 2331 |
+
"learning_rate": 1.6972741763053835e-06,
|
| 2332 |
+
"loss": 1.1695,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7590306355738454,
|
| 2337 |
+
"grad_norm": 0.6124244446703457,
|
| 2338 |
+
"learning_rate": 1.6674169849386606e-06,
|
| 2339 |
+
"loss": 1.1735,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.7613168724279835,
|
| 2344 |
+
"grad_norm": 0.6215393083401685,
|
| 2345 |
+
"learning_rate": 1.6377721024898214e-06,
|
| 2346 |
+
"loss": 1.1611,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.7636031092821216,
|
| 2351 |
+
"grad_norm": 0.6379465283211975,
|
| 2352 |
+
"learning_rate": 1.608341417545849e-06,
|
| 2353 |
+
"loss": 1.1481,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7658893461362597,
|
| 2358 |
+
"grad_norm": 0.5646658898706897,
|
| 2359 |
+
"learning_rate": 1.5791268050478487e-06,
|
| 2360 |
+
"loss": 1.1732,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7681755829903978,
|
| 2365 |
+
"grad_norm": 0.6028441016085894,
|
| 2366 |
+
"learning_rate": 1.5501301261715896e-06,
|
| 2367 |
+
"loss": 1.1703,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7704618198445359,
|
| 2372 |
+
"grad_norm": 0.6313316478647917,
|
| 2373 |
+
"learning_rate": 1.5213532282089466e-06,
|
| 2374 |
+
"loss": 1.1631,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.772748056698674,
|
| 2379 |
+
"grad_norm": 0.600237347487572,
|
| 2380 |
+
"learning_rate": 1.4927979444502028e-06,
|
| 2381 |
+
"loss": 1.1642,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7750342935528121,
|
| 2386 |
+
"grad_norm": 0.5957448361281138,
|
| 2387 |
+
"learning_rate": 1.4644660940672628e-06,
|
| 2388 |
+
"loss": 1.1668,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7773205304069501,
|
| 2393 |
+
"grad_norm": 0.5872437663700951,
|
| 2394 |
+
"learning_rate": 1.4363594819977606e-06,
|
| 2395 |
+
"loss": 1.1707,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7796067672610882,
|
| 2400 |
+
"grad_norm": 0.7075549655922131,
|
| 2401 |
+
"learning_rate": 1.4084798988300684e-06,
|
| 2402 |
+
"loss": 1.1723,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7818930041152263,
|
| 2407 |
+
"grad_norm": 0.6203199463017092,
|
| 2408 |
+
"learning_rate": 1.3808291206892232e-06,
|
| 2409 |
+
"loss": 1.1668,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7841792409693644,
|
| 2414 |
+
"grad_norm": 0.5759538308213393,
|
| 2415 |
+
"learning_rate": 1.3534089091237757e-06,
|
| 2416 |
+
"loss": 1.1598,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7864654778235025,
|
| 2421 |
+
"grad_norm": 0.5942123152988342,
|
| 2422 |
+
"learning_rate": 1.3262210109935719e-06,
|
| 2423 |
+
"loss": 1.1699,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7887517146776406,
|
| 2428 |
+
"grad_norm": 0.6597153339968819,
|
| 2429 |
+
"learning_rate": 1.2992671583584587e-06,
|
| 2430 |
+
"loss": 1.163,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7910379515317787,
|
| 2435 |
+
"grad_norm": 0.5994756887911626,
|
| 2436 |
+
"learning_rate": 1.2725490683679458e-06,
|
| 2437 |
+
"loss": 1.1797,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7933241883859168,
|
| 2442 |
+
"grad_norm": 0.5942174681280669,
|
| 2443 |
+
"learning_rate": 1.2460684431518055e-06,
|
| 2444 |
+
"loss": 1.1649,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7956104252400549,
|
| 2449 |
+
"grad_norm": 0.5884403788886147,
|
| 2450 |
+
"learning_rate": 1.2198269697116416e-06,
|
| 2451 |
+
"loss": 1.1627,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.797896662094193,
|
| 2456 |
+
"grad_norm": 0.5917506875732326,
|
| 2457 |
+
"learning_rate": 1.1938263198134087e-06,
|
| 2458 |
+
"loss": 1.1729,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8001828989483311,
|
| 2463 |
+
"grad_norm": 0.5689945244963683,
|
| 2464 |
+
"learning_rate": 1.168068149880912e-06,
|
| 2465 |
+
"loss": 1.1639,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8024691358024691,
|
| 2470 |
+
"grad_norm": 0.5945700377730089,
|
| 2471 |
+
"learning_rate": 1.1425541008902852e-06,
|
| 2472 |
+
"loss": 1.1616,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8047553726566072,
|
| 2477 |
+
"grad_norm": 0.5960318855848052,
|
| 2478 |
+
"learning_rate": 1.1172857982654445e-06,
|
| 2479 |
+
"loss": 1.1796,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8070416095107453,
|
| 2484 |
+
"grad_norm": 0.606906781862042,
|
| 2485 |
+
"learning_rate": 1.092264851774536e-06,
|
| 2486 |
+
"loss": 1.1524,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8093278463648834,
|
| 2491 |
+
"grad_norm": 0.6686014083887466,
|
| 2492 |
+
"learning_rate": 1.067492855427385e-06,
|
| 2493 |
+
"loss": 1.1681,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8116140832190215,
|
| 2498 |
+
"grad_norm": 0.6637295349703526,
|
| 2499 |
+
"learning_rate": 1.0429713873739505e-06,
|
| 2500 |
+
"loss": 1.1603,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8139003200731596,
|
| 2505 |
+
"grad_norm": 0.5937746781646984,
|
| 2506 |
+
"learning_rate": 1.0187020098037759e-06,
|
| 2507 |
+
"loss": 1.1577,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8161865569272977,
|
| 2512 |
+
"grad_norm": 0.6154438358761861,
|
| 2513 |
+
"learning_rate": 9.946862688464753e-07,
|
| 2514 |
+
"loss": 1.1596,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8184727937814358,
|
| 2519 |
+
"grad_norm": 0.6511739287376433,
|
| 2520 |
+
"learning_rate": 9.709256944732343e-07,
|
| 2521 |
+
"loss": 1.1707,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8207590306355739,
|
| 2526 |
+
"grad_norm": 0.6174881374069865,
|
| 2527 |
+
"learning_rate": 9.474218003993275e-07,
|
| 2528 |
+
"loss": 1.1775,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.823045267489712,
|
| 2533 |
+
"grad_norm": 0.5791204684491382,
|
| 2534 |
+
"learning_rate": 9.241760839877023e-07,
|
| 2535 |
+
"loss": 1.1571,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.82533150434385,
|
| 2540 |
+
"grad_norm": 0.6464260391976697,
|
| 2541 |
+
"learning_rate": 9.011900261535767e-07,
|
| 2542 |
+
"loss": 1.1713,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8276177411979881,
|
| 2547 |
+
"grad_norm": 0.6102288143326278,
|
| 2548 |
+
"learning_rate": 8.784650912700909e-07,
|
| 2549 |
+
"loss": 1.1654,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8299039780521262,
|
| 2554 |
+
"grad_norm": 0.6226743471510658,
|
| 2555 |
+
"learning_rate": 8.560027270750276e-07,
|
| 2556 |
+
"loss": 1.1655,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8321902149062643,
|
| 2561 |
+
"grad_norm": 0.6079710775307922,
|
| 2562 |
+
"learning_rate": 8.338043645785698e-07,
|
| 2563 |
+
"loss": 1.1669,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8344764517604024,
|
| 2568 |
+
"grad_norm": 0.6077180347148399,
|
| 2569 |
+
"learning_rate": 8.118714179721404e-07,
|
| 2570 |
+
"loss": 1.1529,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8367626886145405,
|
| 2575 |
+
"grad_norm": 0.6420590181680129,
|
| 2576 |
+
"learning_rate": 7.902052845383112e-07,
|
| 2577 |
+
"loss": 1.1662,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8390489254686786,
|
| 2582 |
+
"grad_norm": 0.5675937752707487,
|
| 2583 |
+
"learning_rate": 7.6880734456178e-07,
|
| 2584 |
+
"loss": 1.1638,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.8413351623228167,
|
| 2589 |
+
"grad_norm": 0.5963600943686237,
|
| 2590 |
+
"learning_rate": 7.476789612414414e-07,
|
| 2591 |
+
"loss": 1.1648,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8436213991769548,
|
| 2596 |
+
"grad_norm": 0.6248451529177521,
|
| 2597 |
+
"learning_rate": 7.268214806035423e-07,
|
| 2598 |
+
"loss": 1.1704,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8459076360310929,
|
| 2603 |
+
"grad_norm": 0.6582130785897107,
|
| 2604 |
+
"learning_rate": 7.062362314159211e-07,
|
| 2605 |
+
"loss": 1.1716,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.848193872885231,
|
| 2610 |
+
"grad_norm": 0.6104979563533071,
|
| 2611 |
+
"learning_rate": 6.859245251033697e-07,
|
| 2612 |
+
"loss": 1.1551,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.850480109739369,
|
| 2617 |
+
"grad_norm": 0.6291505363028616,
|
| 2618 |
+
"learning_rate": 6.658876556640781e-07,
|
| 2619 |
+
"loss": 1.1606,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8527663465935071,
|
| 2624 |
+
"grad_norm": 0.626351910055198,
|
| 2625 |
+
"learning_rate": 6.461268995871967e-07,
|
| 2626 |
+
"loss": 1.1648,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8550525834476452,
|
| 2631 |
+
"grad_norm": 0.5991977091276379,
|
| 2632 |
+
"learning_rate": 6.266435157715222e-07,
|
| 2633 |
+
"loss": 1.1403,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8573388203017832,
|
| 2638 |
+
"grad_norm": 0.6133109082285381,
|
| 2639 |
+
"learning_rate": 6.074387454452891e-07,
|
| 2640 |
+
"loss": 1.1578,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8596250571559213,
|
| 2645 |
+
"grad_norm": 0.6062420232877472,
|
| 2646 |
+
"learning_rate": 5.885138120870965e-07,
|
| 2647 |
+
"loss": 1.1422,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8619112940100594,
|
| 2652 |
+
"grad_norm": 0.5920619164293491,
|
| 2653 |
+
"learning_rate": 5.698699213479697e-07,
|
| 2654 |
+
"loss": 1.1503,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8641975308641975,
|
| 2659 |
+
"grad_norm": 0.6179934405963249,
|
| 2660 |
+
"learning_rate": 5.515082609745465e-07,
|
| 2661 |
+
"loss": 1.1728,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8664837677183356,
|
| 2666 |
+
"grad_norm": 0.6191884681224713,
|
| 2667 |
+
"learning_rate": 5.334300007334065e-07,
|
| 2668 |
+
"loss": 1.1514,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8687700045724737,
|
| 2673 |
+
"grad_norm": 0.6148818189812965,
|
| 2674 |
+
"learning_rate": 5.156362923365587e-07,
|
| 2675 |
+
"loss": 1.1772,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8710562414266118,
|
| 2680 |
+
"grad_norm": 0.5927964681781609,
|
| 2681 |
+
"learning_rate": 4.981282693680584e-07,
|
| 2682 |
+
"loss": 1.1747,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8733424782807498,
|
| 2687 |
+
"grad_norm": 0.630038523819453,
|
| 2688 |
+
"learning_rate": 4.80907047211796e-07,
|
| 2689 |
+
"loss": 1.1638,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.8756287151348879,
|
| 2694 |
+
"grad_norm": 0.5822419290829026,
|
| 2695 |
+
"learning_rate": 4.639737229804403e-07,
|
| 2696 |
+
"loss": 1.1667,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.877914951989026,
|
| 2701 |
+
"grad_norm": 0.6169634205827448,
|
| 2702 |
+
"learning_rate": 4.473293754455399e-07,
|
| 2703 |
+
"loss": 1.1695,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.8802011888431641,
|
| 2708 |
+
"grad_norm": 0.5892947845386679,
|
| 2709 |
+
"learning_rate": 4.3097506496880325e-07,
|
| 2710 |
+
"loss": 1.1684,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.8824874256973022,
|
| 2715 |
+
"grad_norm": 0.6796811793089527,
|
| 2716 |
+
"learning_rate": 4.149118334345403e-07,
|
| 2717 |
+
"loss": 1.1604,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8847736625514403,
|
| 2722 |
+
"grad_norm": 0.5951100132603444,
|
| 2723 |
+
"learning_rate": 3.9914070418329123e-07,
|
| 2724 |
+
"loss": 1.1632,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8870598994055784,
|
| 2729 |
+
"grad_norm": 0.6710610553022762,
|
| 2730 |
+
"learning_rate": 3.836626819466338e-07,
|
| 2731 |
+
"loss": 1.1455,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8893461362597165,
|
| 2736 |
+
"grad_norm": 0.6128779790737046,
|
| 2737 |
+
"learning_rate": 3.684787527831707e-07,
|
| 2738 |
+
"loss": 1.1609,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8916323731138546,
|
| 2743 |
+
"grad_norm": 0.5800567298586133,
|
| 2744 |
+
"learning_rate": 3.53589884015712e-07,
|
| 2745 |
+
"loss": 1.1636,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8939186099679927,
|
| 2750 |
+
"grad_norm": 0.5600191099569565,
|
| 2751 |
+
"learning_rate": 3.3899702416965166e-07,
|
| 2752 |
+
"loss": 1.1721,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8962048468221308,
|
| 2757 |
+
"grad_norm": 0.5964683215562515,
|
| 2758 |
+
"learning_rate": 3.247011029125391e-07,
|
| 2759 |
+
"loss": 1.1508,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8984910836762688,
|
| 2764 |
+
"grad_norm": 0.6125213377358303,
|
| 2765 |
+
"learning_rate": 3.1070303099485055e-07,
|
| 2766 |
+
"loss": 1.1716,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.9007773205304069,
|
| 2771 |
+
"grad_norm": 0.5812964318078312,
|
| 2772 |
+
"learning_rate": 2.9700370019197287e-07,
|
| 2773 |
+
"loss": 1.1495,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.903063557384545,
|
| 2778 |
+
"grad_norm": 0.5947330421470328,
|
| 2779 |
+
"learning_rate": 2.8360398324738415e-07,
|
| 2780 |
+
"loss": 1.1446,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9053497942386831,
|
| 2785 |
+
"grad_norm": 0.5936630268160432,
|
| 2786 |
+
"learning_rate": 2.7050473381706186e-07,
|
| 2787 |
+
"loss": 1.1519,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9076360310928212,
|
| 2792 |
+
"grad_norm": 0.6228979256825669,
|
| 2793 |
+
"learning_rate": 2.577067864150906e-07,
|
| 2794 |
+
"loss": 1.1688,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9099222679469593,
|
| 2799 |
+
"grad_norm": 0.6500515468078818,
|
| 2800 |
+
"learning_rate": 2.452109563605065e-07,
|
| 2801 |
+
"loss": 1.1718,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9122085048010974,
|
| 2806 |
+
"grad_norm": 0.568112374463465,
|
| 2807 |
+
"learning_rate": 2.330180397253473e-07,
|
| 2808 |
+
"loss": 1.169,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9144947416552355,
|
| 2813 |
+
"grad_norm": 0.6014335143268985,
|
| 2814 |
+
"learning_rate": 2.2112881328394287e-07,
|
| 2815 |
+
"loss": 1.1556,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9167809785093736,
|
| 2820 |
+
"grad_norm": 0.5814781144236604,
|
| 2821 |
+
"learning_rate": 2.0954403446342753e-07,
|
| 2822 |
+
"loss": 1.1688,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9190672153635117,
|
| 2827 |
+
"grad_norm": 0.6269697024329176,
|
| 2828 |
+
"learning_rate": 1.9826444129548317e-07,
|
| 2829 |
+
"loss": 1.1791,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9213534522176497,
|
| 2834 |
+
"grad_norm": 0.5793724546294099,
|
| 2835 |
+
"learning_rate": 1.8729075236932903e-07,
|
| 2836 |
+
"loss": 1.1736,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9236396890717878,
|
| 2841 |
+
"grad_norm": 0.5757028817840649,
|
| 2842 |
+
"learning_rate": 1.7662366678593502e-07,
|
| 2843 |
+
"loss": 1.1674,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9259259259259259,
|
| 2848 |
+
"grad_norm": 0.6383512892284545,
|
| 2849 |
+
"learning_rate": 1.6626386411348783e-07,
|
| 2850 |
+
"loss": 1.1725,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.928212162780064,
|
| 2855 |
+
"grad_norm": 0.6064267969457637,
|
| 2856 |
+
"learning_rate": 1.56212004344099e-07,
|
| 2857 |
+
"loss": 1.1596,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.9304983996342021,
|
| 2862 |
+
"grad_norm": 0.6046327277263103,
|
| 2863 |
+
"learning_rate": 1.4646872785175182e-07,
|
| 2864 |
+
"loss": 1.1616,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9327846364883402,
|
| 2869 |
+
"grad_norm": 0.611959733363112,
|
| 2870 |
+
"learning_rate": 1.3703465535151505e-07,
|
| 2871 |
+
"loss": 1.1614,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9350708733424783,
|
| 2876 |
+
"grad_norm": 0.6153837948383357,
|
| 2877 |
+
"learning_rate": 1.2791038785999243e-07,
|
| 2878 |
+
"loss": 1.1494,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9373571101966164,
|
| 2883 |
+
"grad_norm": 0.5507733416769363,
|
| 2884 |
+
"learning_rate": 1.1909650665703265e-07,
|
| 2885 |
+
"loss": 1.1331,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.9396433470507545,
|
| 2890 |
+
"grad_norm": 0.5787602661155832,
|
| 2891 |
+
"learning_rate": 1.1059357324870456e-07,
|
| 2892 |
+
"loss": 1.1548,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.9419295839048926,
|
| 2897 |
+
"grad_norm": 0.5848374134615248,
|
| 2898 |
+
"learning_rate": 1.024021293315175e-07,
|
| 2899 |
+
"loss": 1.1628,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9442158207590307,
|
| 2904 |
+
"grad_norm": 0.585861722501522,
|
| 2905 |
+
"learning_rate": 9.452269675791603e-08,
|
| 2906 |
+
"loss": 1.1424,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9465020576131687,
|
| 2911 |
+
"grad_norm": 0.5870866242087308,
|
| 2912 |
+
"learning_rate": 8.69557775030344e-08,
|
| 2913 |
+
"loss": 1.181,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.9487882944673068,
|
| 2918 |
+
"grad_norm": 0.5917858310575264,
|
| 2919 |
+
"learning_rate": 7.970185363271432e-08,
|
| 2920 |
+
"loss": 1.1564,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9510745313214449,
|
| 2925 |
+
"grad_norm": 0.6272259568011471,
|
| 2926 |
+
"learning_rate": 7.276138727279669e-08,
|
| 2927 |
+
"loss": 1.1659,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.953360768175583,
|
| 2932 |
+
"grad_norm": 0.607366888512829,
|
| 2933 |
+
"learning_rate": 6.613482057968023e-08,
|
| 2934 |
+
"loss": 1.1612,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9556470050297211,
|
| 2939 |
+
"grad_norm": 0.61579614820576,
|
| 2940 |
+
"learning_rate": 5.982257571215178e-08,
|
| 2941 |
+
"loss": 1.1644,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9579332418838592,
|
| 2946 |
+
"grad_norm": 0.6162342496797737,
|
| 2947 |
+
"learning_rate": 5.382505480449274e-08,
|
| 2948 |
+
"loss": 1.1439,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.9602194787379973,
|
| 2953 |
+
"grad_norm": 0.5880335959078453,
|
| 2954 |
+
"learning_rate": 4.814263994086077e-08,
|
| 2955 |
+
"loss": 1.1405,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9625057155921354,
|
| 2960 |
+
"grad_norm": 0.5978901392727579,
|
| 2961 |
+
"learning_rate": 4.2775693130948094e-08,
|
| 2962 |
+
"loss": 1.1792,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9647919524462735,
|
| 2967 |
+
"grad_norm": 0.5725207858399001,
|
| 2968 |
+
"learning_rate": 3.772455628691829e-08,
|
| 2969 |
+
"loss": 1.1679,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9670781893004116,
|
| 2974 |
+
"grad_norm": 0.6126681514493614,
|
| 2975 |
+
"learning_rate": 3.2989551201624836e-08,
|
| 2976 |
+
"loss": 1.1621,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9693644261545497,
|
| 2981 |
+
"grad_norm": 0.6026354249744876,
|
| 2982 |
+
"learning_rate": 2.857097952810972e-08,
|
| 2983 |
+
"loss": 1.1728,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9716506630086877,
|
| 2988 |
+
"grad_norm": 0.5876159431495082,
|
| 2989 |
+
"learning_rate": 2.4469122760388264e-08,
|
| 2990 |
+
"loss": 1.1552,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9739368998628258,
|
| 2995 |
+
"grad_norm": 0.5795939734314318,
|
| 2996 |
+
"learning_rate": 2.0684242215511797e-08,
|
| 2997 |
+
"loss": 1.1586,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9762231367169639,
|
| 3002 |
+
"grad_norm": 0.6100064497073957,
|
| 3003 |
+
"learning_rate": 1.7216579016925415e-08,
|
| 3004 |
+
"loss": 1.1585,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.978509373571102,
|
| 3009 |
+
"grad_norm": 0.6410024148442394,
|
| 3010 |
+
"learning_rate": 1.4066354079101396e-08,
|
| 3011 |
+
"loss": 1.1576,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.9807956104252401,
|
| 3016 |
+
"grad_norm": 0.5946394925998356,
|
| 3017 |
+
"learning_rate": 1.1233768093468766e-08,
|
| 3018 |
+
"loss": 1.1565,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.9830818472793782,
|
| 3023 |
+
"grad_norm": 0.5993080705042445,
|
| 3024 |
+
"learning_rate": 8.719001515627434e-09,
|
| 3025 |
+
"loss": 1.1649,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.9853680841335163,
|
| 3030 |
+
"grad_norm": 0.5857680491868433,
|
| 3031 |
+
"learning_rate": 6.5222145538501595e-09,
|
| 3032 |
+
"loss": 1.176,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.9876543209876543,
|
| 3037 |
+
"grad_norm": 0.6157142971328977,
|
| 3038 |
+
"learning_rate": 4.643547158878492e-09,
|
| 3039 |
+
"loss": 1.146,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9899405578417924,
|
| 3044 |
+
"grad_norm": 0.6005659801135901,
|
| 3045 |
+
"learning_rate": 3.0831190150054646e-09,
|
| 3046 |
+
"loss": 1.1607,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.9922267946959304,
|
| 3051 |
+
"grad_norm": 0.5963682235084494,
|
| 3052 |
+
"learning_rate": 1.8410295324505778e-09,
|
| 3053 |
+
"loss": 1.1668,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9945130315500685,
|
| 3058 |
+
"grad_norm": 0.649218390898171,
|
| 3059 |
+
"learning_rate": 9.173578410281992e-10,
|
| 3060 |
+
"loss": 1.1602,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9967992684042066,
|
| 3065 |
+
"grad_norm": 0.612662110275474,
|
| 3066 |
+
"learning_rate": 3.1216278510493027e-10,
|
| 3067 |
+
"loss": 1.1596,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9990855052583447,
|
| 3072 |
+
"grad_norm": 0.6025732837303296,
|
| 3073 |
+
"learning_rate": 2.548291985149387e-11,
|
| 3074 |
+
"loss": 1.147,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 1.0,
|
| 3079 |
+
"eval_runtime": 4.0833,
|
| 3080 |
+
"eval_samples_per_second": 2.449,
|
| 3081 |
+
"eval_steps_per_second": 0.735,
|
| 3082 |
+
"step": 2187
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 1.0,
|
| 3086 |
+
"step": 2187,
|
| 3087 |
+
"total_flos": 9703359095242752.0,
|
| 3088 |
+
"train_loss": 1.3940648635773556,
|
| 3089 |
+
"train_runtime": 19118.6692,
|
| 3090 |
+
"train_samples_per_second": 1.83,
|
| 3091 |
+
"train_steps_per_second": 0.114
|
| 3092 |
+
}
|
| 3093 |
+
],
|
| 3094 |
+
"logging_steps": 5,
|
| 3095 |
+
"max_steps": 2187,
|
| 3096 |
+
"num_input_tokens_seen": 0,
|
| 3097 |
+
"num_train_epochs": 1,
|
| 3098 |
+
"save_steps": 100,
|
| 3099 |
+
"stateful_callbacks": {
|
| 3100 |
+
"TrainerControl": {
|
| 3101 |
+
"args": {
|
| 3102 |
+
"should_epoch_stop": false,
|
| 3103 |
+
"should_evaluate": false,
|
| 3104 |
+
"should_log": false,
|
| 3105 |
+
"should_save": true,
|
| 3106 |
+
"should_training_stop": true
|
| 3107 |
+
},
|
| 3108 |
+
"attributes": {}
|
| 3109 |
+
}
|
| 3110 |
+
},
|
| 3111 |
+
"total_flos": 9703359095242752.0,
|
| 3112 |
+
"train_batch_size": 4,
|
| 3113 |
+
"trial_name": null,
|
| 3114 |
+
"trial_params": null
|
| 3115 |
+
}
|