Model save
Browse files- README.md +60 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- runs/Sep27_07-39-40_COE-CS-sv003/events.out.tfevents.1727423226.COE-CS-sv003.921006.0 +2 -2
- train_results.json +9 -0
- trainer_state.json +3444 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: preference_p0.1_seed42_level2_stylemixbatch16_reverse
|
| 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 |
+
# preference_p0.1_seed42_level2_stylemixbatch16_reverse
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the 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 |
+
- Transformers 4.45.1
|
| 58 |
+
- Pytorch 2.4.1+cu121
|
| 59 |
+
- Datasets 3.0.1
|
| 60 |
+
- Tokenizers 0.20.0
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 253663452856320.0,
|
| 4 |
+
"train_loss": 0.5332340838443727,
|
| 5 |
+
"train_runtime": 21510.4539,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.802,
|
| 8 |
+
"train_steps_per_second": 0.113
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.45.1"
|
| 6 |
+
}
|
runs/Sep27_07-39-40_COE-CS-sv003/events.out.tfevents.1727423226.COE-CS-sv003.921006.0
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 sha256:32c5344f899804d2dca27d90ccbcc73b368917d471e2702b30354a6862b468d4
|
| 3 |
+
size 108637
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 253663452856320.0,
|
| 4 |
+
"train_loss": 0.5332340838443727,
|
| 5 |
+
"train_runtime": 21510.4539,
|
| 6 |
+
"train_samples": 98881,
|
| 7 |
+
"train_samples_per_second": 1.802,
|
| 8 |
+
"train_steps_per_second": 0.113
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3444 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2423,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004127115146512588,
|
| 13 |
+
"grad_norm": 24.122843960524676,
|
| 14 |
+
"learning_rate": 4.1152263374485605e-08,
|
| 15 |
+
"loss": 1.4137,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0020635575732562937,
|
| 20 |
+
"grad_norm": 23.76872795123526,
|
| 21 |
+
"learning_rate": 2.05761316872428e-07,
|
| 22 |
+
"loss": 1.4306,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.004127115146512587,
|
| 27 |
+
"grad_norm": 15.835819961484214,
|
| 28 |
+
"learning_rate": 4.11522633744856e-07,
|
| 29 |
+
"loss": 1.3895,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006190672719768881,
|
| 34 |
+
"grad_norm": 8.711444647130131,
|
| 35 |
+
"learning_rate": 6.17283950617284e-07,
|
| 36 |
+
"loss": 1.2741,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.008254230293025175,
|
| 41 |
+
"grad_norm": 10.763225748048312,
|
| 42 |
+
"learning_rate": 8.23045267489712e-07,
|
| 43 |
+
"loss": 1.1629,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.010317787866281469,
|
| 48 |
+
"grad_norm": 5.160551789043553,
|
| 49 |
+
"learning_rate": 1.02880658436214e-06,
|
| 50 |
+
"loss": 1.0567,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.012381345439537762,
|
| 55 |
+
"grad_norm": 3.472599934267948,
|
| 56 |
+
"learning_rate": 1.234567901234568e-06,
|
| 57 |
+
"loss": 0.9904,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.014444903012794058,
|
| 62 |
+
"grad_norm": 3.5380974995691528,
|
| 63 |
+
"learning_rate": 1.440329218106996e-06,
|
| 64 |
+
"loss": 0.9621,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.01650846058605035,
|
| 69 |
+
"grad_norm": 3.0479724307668077,
|
| 70 |
+
"learning_rate": 1.646090534979424e-06,
|
| 71 |
+
"loss": 0.9573,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.018572018159306643,
|
| 76 |
+
"grad_norm": 2.97801536675108,
|
| 77 |
+
"learning_rate": 1.8518518518518519e-06,
|
| 78 |
+
"loss": 0.9295,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.020635575732562937,
|
| 83 |
+
"grad_norm": 3.0790095562756514,
|
| 84 |
+
"learning_rate": 2.05761316872428e-06,
|
| 85 |
+
"loss": 0.9106,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.02269913330581923,
|
| 90 |
+
"grad_norm": 3.0873399280284626,
|
| 91 |
+
"learning_rate": 2.263374485596708e-06,
|
| 92 |
+
"loss": 0.9054,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.024762690879075525,
|
| 97 |
+
"grad_norm": 3.046348701050192,
|
| 98 |
+
"learning_rate": 2.469135802469136e-06,
|
| 99 |
+
"loss": 0.9016,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.026826248452331822,
|
| 104 |
+
"grad_norm": 3.117521509733326,
|
| 105 |
+
"learning_rate": 2.674897119341564e-06,
|
| 106 |
+
"loss": 0.9005,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.028889806025588115,
|
| 111 |
+
"grad_norm": 3.046385395158289,
|
| 112 |
+
"learning_rate": 2.880658436213992e-06,
|
| 113 |
+
"loss": 0.8882,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03095336359884441,
|
| 118 |
+
"grad_norm": 3.1884168387342537,
|
| 119 |
+
"learning_rate": 3.08641975308642e-06,
|
| 120 |
+
"loss": 0.8799,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.0330169211721007,
|
| 125 |
+
"grad_norm": 3.2618306435716176,
|
| 126 |
+
"learning_rate": 3.292181069958848e-06,
|
| 127 |
+
"loss": 0.8907,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.035080478745357,
|
| 132 |
+
"grad_norm": 3.322364354858461,
|
| 133 |
+
"learning_rate": 3.4979423868312762e-06,
|
| 134 |
+
"loss": 0.8684,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.03714403631861329,
|
| 139 |
+
"grad_norm": 2.9808270129935495,
|
| 140 |
+
"learning_rate": 3.7037037037037037e-06,
|
| 141 |
+
"loss": 0.8604,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.039207593891869584,
|
| 146 |
+
"grad_norm": 3.1835415746195435,
|
| 147 |
+
"learning_rate": 3.909465020576132e-06,
|
| 148 |
+
"loss": 0.8707,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.041271151465125874,
|
| 153 |
+
"grad_norm": 3.057661991370929,
|
| 154 |
+
"learning_rate": 4.11522633744856e-06,
|
| 155 |
+
"loss": 0.8703,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.04333470903838217,
|
| 160 |
+
"grad_norm": 2.9074769839799277,
|
| 161 |
+
"learning_rate": 4.3209876543209875e-06,
|
| 162 |
+
"loss": 0.8633,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.04539826661163846,
|
| 167 |
+
"grad_norm": 2.9699831351967405,
|
| 168 |
+
"learning_rate": 4.526748971193416e-06,
|
| 169 |
+
"loss": 0.8759,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.04746182418489476,
|
| 174 |
+
"grad_norm": 3.219620261483576,
|
| 175 |
+
"learning_rate": 4.732510288065844e-06,
|
| 176 |
+
"loss": 0.8636,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.04952538175815105,
|
| 181 |
+
"grad_norm": 3.060892913438638,
|
| 182 |
+
"learning_rate": 4.938271604938272e-06,
|
| 183 |
+
"loss": 0.8512,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.051588939331407346,
|
| 188 |
+
"grad_norm": 3.0812348532752543,
|
| 189 |
+
"learning_rate": 5.1440329218107e-06,
|
| 190 |
+
"loss": 0.8412,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.053652496904663644,
|
| 195 |
+
"grad_norm": 3.208845547096666,
|
| 196 |
+
"learning_rate": 5.349794238683128e-06,
|
| 197 |
+
"loss": 0.8581,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.055716054477919934,
|
| 202 |
+
"grad_norm": 3.461069159058445,
|
| 203 |
+
"learning_rate": 5.555555555555557e-06,
|
| 204 |
+
"loss": 0.8512,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.05777961205117623,
|
| 209 |
+
"grad_norm": 3.040452060763894,
|
| 210 |
+
"learning_rate": 5.761316872427984e-06,
|
| 211 |
+
"loss": 0.857,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.05984316962443252,
|
| 216 |
+
"grad_norm": 2.9241553701138763,
|
| 217 |
+
"learning_rate": 5.967078189300412e-06,
|
| 218 |
+
"loss": 0.8556,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06190672719768882,
|
| 223 |
+
"grad_norm": 3.22517785299545,
|
| 224 |
+
"learning_rate": 6.17283950617284e-06,
|
| 225 |
+
"loss": 0.8504,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.06397028477094512,
|
| 230 |
+
"grad_norm": 3.066199171323176,
|
| 231 |
+
"learning_rate": 6.3786008230452675e-06,
|
| 232 |
+
"loss": 0.8455,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.0660338423442014,
|
| 237 |
+
"grad_norm": 3.1076485395976436,
|
| 238 |
+
"learning_rate": 6.584362139917696e-06,
|
| 239 |
+
"loss": 0.8343,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.0680973999174577,
|
| 244 |
+
"grad_norm": 3.057886812615224,
|
| 245 |
+
"learning_rate": 6.790123456790124e-06,
|
| 246 |
+
"loss": 0.833,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.070160957490714,
|
| 251 |
+
"grad_norm": 3.12354443289718,
|
| 252 |
+
"learning_rate": 6.9958847736625525e-06,
|
| 253 |
+
"loss": 0.8445,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.07222451506397029,
|
| 258 |
+
"grad_norm": 3.1566533318265515,
|
| 259 |
+
"learning_rate": 7.201646090534981e-06,
|
| 260 |
+
"loss": 0.8336,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.07428807263722657,
|
| 265 |
+
"grad_norm": 3.0910067969250288,
|
| 266 |
+
"learning_rate": 7.4074074074074075e-06,
|
| 267 |
+
"loss": 0.8335,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.07635163021048287,
|
| 272 |
+
"grad_norm": 3.2341343019009754,
|
| 273 |
+
"learning_rate": 7.613168724279836e-06,
|
| 274 |
+
"loss": 0.8313,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.07841518778373917,
|
| 279 |
+
"grad_norm": 2.9443616643904984,
|
| 280 |
+
"learning_rate": 7.818930041152263e-06,
|
| 281 |
+
"loss": 0.8271,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08047874535699547,
|
| 286 |
+
"grad_norm": 2.955202520255723,
|
| 287 |
+
"learning_rate": 8.024691358024692e-06,
|
| 288 |
+
"loss": 0.833,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.08254230293025175,
|
| 293 |
+
"grad_norm": 3.0004205685291896,
|
| 294 |
+
"learning_rate": 8.23045267489712e-06,
|
| 295 |
+
"loss": 0.8046,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.08460586050350805,
|
| 300 |
+
"grad_norm": 3.0549793246137984,
|
| 301 |
+
"learning_rate": 8.43621399176955e-06,
|
| 302 |
+
"loss": 0.8155,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.08666941807676434,
|
| 307 |
+
"grad_norm": 3.0535661505541896,
|
| 308 |
+
"learning_rate": 8.641975308641975e-06,
|
| 309 |
+
"loss": 0.8196,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.08873297565002064,
|
| 314 |
+
"grad_norm": 3.116053820452697,
|
| 315 |
+
"learning_rate": 8.847736625514404e-06,
|
| 316 |
+
"loss": 0.8347,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.09079653322327692,
|
| 321 |
+
"grad_norm": 3.022529098579132,
|
| 322 |
+
"learning_rate": 9.053497942386832e-06,
|
| 323 |
+
"loss": 0.8081,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.09286009079653322,
|
| 328 |
+
"grad_norm": 3.0433063189548792,
|
| 329 |
+
"learning_rate": 9.25925925925926e-06,
|
| 330 |
+
"loss": 0.8208,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.09492364836978952,
|
| 335 |
+
"grad_norm": 3.085322202490418,
|
| 336 |
+
"learning_rate": 9.465020576131688e-06,
|
| 337 |
+
"loss": 0.8339,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.09698720594304581,
|
| 342 |
+
"grad_norm": 3.6047031312875584,
|
| 343 |
+
"learning_rate": 9.670781893004116e-06,
|
| 344 |
+
"loss": 0.8181,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.0990507635163021,
|
| 349 |
+
"grad_norm": 3.0604827472581255,
|
| 350 |
+
"learning_rate": 9.876543209876543e-06,
|
| 351 |
+
"loss": 0.8166,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.1011143210895584,
|
| 356 |
+
"grad_norm": 3.2691636345258277,
|
| 357 |
+
"learning_rate": 9.999979232392962e-06,
|
| 358 |
+
"loss": 0.799,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.10317787866281469,
|
| 363 |
+
"grad_norm": 2.9570633048095964,
|
| 364 |
+
"learning_rate": 9.999745598795033e-06,
|
| 365 |
+
"loss": 0.8205,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.10524143623607099,
|
| 370 |
+
"grad_norm": 3.335892639195958,
|
| 371 |
+
"learning_rate": 9.999252384260794e-06,
|
| 372 |
+
"loss": 0.7989,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.10730499380932729,
|
| 377 |
+
"grad_norm": 2.9248399326152725,
|
| 378 |
+
"learning_rate": 9.998499614397364e-06,
|
| 379 |
+
"loss": 0.8063,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.10936855138258357,
|
| 384 |
+
"grad_norm": 3.0768852625514422,
|
| 385 |
+
"learning_rate": 9.997487328287675e-06,
|
| 386 |
+
"loss": 0.7968,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.11143210895583987,
|
| 391 |
+
"grad_norm": 3.002775395507999,
|
| 392 |
+
"learning_rate": 9.996215578488434e-06,
|
| 393 |
+
"loss": 0.8062,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.11349566652909616,
|
| 398 |
+
"grad_norm": 2.9801626597554174,
|
| 399 |
+
"learning_rate": 9.994684431027407e-06,
|
| 400 |
+
"loss": 0.7844,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.11555922410235246,
|
| 405 |
+
"grad_norm": 3.1190858706402604,
|
| 406 |
+
"learning_rate": 9.99289396539997e-06,
|
| 407 |
+
"loss": 0.8089,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.11762278167560875,
|
| 412 |
+
"grad_norm": 2.785366704095957,
|
| 413 |
+
"learning_rate": 9.990844274565004e-06,
|
| 414 |
+
"loss": 0.7997,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.11968633924886504,
|
| 419 |
+
"grad_norm": 3.175006190343879,
|
| 420 |
+
"learning_rate": 9.98853546494006e-06,
|
| 421 |
+
"loss": 0.8186,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.12174989682212134,
|
| 426 |
+
"grad_norm": 2.883637899845788,
|
| 427 |
+
"learning_rate": 9.985967656395823e-06,
|
| 428 |
+
"loss": 0.7889,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.12381345439537764,
|
| 433 |
+
"grad_norm": 2.991597043168701,
|
| 434 |
+
"learning_rate": 9.983140982249913e-06,
|
| 435 |
+
"loss": 0.8136,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.12587701196863393,
|
| 440 |
+
"grad_norm": 2.967721245447412,
|
| 441 |
+
"learning_rate": 9.980055589259937e-06,
|
| 442 |
+
"loss": 0.7804,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.12794056954189023,
|
| 447 |
+
"grad_norm": 2.925851320281598,
|
| 448 |
+
"learning_rate": 9.976711637615886e-06,
|
| 449 |
+
"loss": 0.7789,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.1300041271151465,
|
| 454 |
+
"grad_norm": 3.015507414806422,
|
| 455 |
+
"learning_rate": 9.973109300931813e-06,
|
| 456 |
+
"loss": 0.7821,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.1320676846884028,
|
| 461 |
+
"grad_norm": 3.076125191679388,
|
| 462 |
+
"learning_rate": 9.969248766236823e-06,
|
| 463 |
+
"loss": 0.7668,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.1341312422616591,
|
| 468 |
+
"grad_norm": 2.9816750909346137,
|
| 469 |
+
"learning_rate": 9.965130233965353e-06,
|
| 470 |
+
"loss": 0.7867,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.1361947998349154,
|
| 475 |
+
"grad_norm": 3.1216143549392514,
|
| 476 |
+
"learning_rate": 9.960753917946777e-06,
|
| 477 |
+
"loss": 0.7827,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.1382583574081717,
|
| 482 |
+
"grad_norm": 3.125152909470307,
|
| 483 |
+
"learning_rate": 9.956120045394297e-06,
|
| 484 |
+
"loss": 0.7683,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.140321914981428,
|
| 489 |
+
"grad_norm": 2.8419491540297694,
|
| 490 |
+
"learning_rate": 9.951228856893152e-06,
|
| 491 |
+
"loss": 0.7571,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.14238547255468428,
|
| 496 |
+
"grad_norm": 2.739631493281793,
|
| 497 |
+
"learning_rate": 9.946080606388115e-06,
|
| 498 |
+
"loss": 0.7743,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.14444903012794058,
|
| 503 |
+
"grad_norm": 2.888837107470469,
|
| 504 |
+
"learning_rate": 9.940675561170328e-06,
|
| 505 |
+
"loss": 0.7795,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.14651258770119685,
|
| 510 |
+
"grad_norm": 3.0512065932946886,
|
| 511 |
+
"learning_rate": 9.935014001863405e-06,
|
| 512 |
+
"loss": 0.7753,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.14857614527445315,
|
| 517 |
+
"grad_norm": 3.029181890310092,
|
| 518 |
+
"learning_rate": 9.92909622240888e-06,
|
| 519 |
+
"loss": 0.7515,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.15063970284770944,
|
| 524 |
+
"grad_norm": 2.9755768091749015,
|
| 525 |
+
"learning_rate": 9.92292253005093e-06,
|
| 526 |
+
"loss": 0.7564,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.15270326042096574,
|
| 531 |
+
"grad_norm": 2.8483284554348565,
|
| 532 |
+
"learning_rate": 9.916493245320428e-06,
|
| 533 |
+
"loss": 0.7642,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.15476681799422204,
|
| 538 |
+
"grad_norm": 3.2241540408474,
|
| 539 |
+
"learning_rate": 9.909808702018315e-06,
|
| 540 |
+
"loss": 0.758,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.15683037556747834,
|
| 545 |
+
"grad_norm": 2.871363873496698,
|
| 546 |
+
"learning_rate": 9.902869247198246e-06,
|
| 547 |
+
"loss": 0.7595,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.15889393314073463,
|
| 552 |
+
"grad_norm": 2.9539374709865607,
|
| 553 |
+
"learning_rate": 9.895675241148588e-06,
|
| 554 |
+
"loss": 0.7518,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.16095749071399093,
|
| 559 |
+
"grad_norm": 2.9560555474503043,
|
| 560 |
+
"learning_rate": 9.888227057373716e-06,
|
| 561 |
+
"loss": 0.7493,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.16302104828724723,
|
| 566 |
+
"grad_norm": 2.7957300167557926,
|
| 567 |
+
"learning_rate": 9.880525082574604e-06,
|
| 568 |
+
"loss": 0.7476,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.1650846058605035,
|
| 573 |
+
"grad_norm": 2.6981062487306606,
|
| 574 |
+
"learning_rate": 9.872569716628764e-06,
|
| 575 |
+
"loss": 0.7273,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.1671481634337598,
|
| 580 |
+
"grad_norm": 2.7734124248259366,
|
| 581 |
+
"learning_rate": 9.86436137256948e-06,
|
| 582 |
+
"loss": 0.7297,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.1692117210070161,
|
| 587 |
+
"grad_norm": 2.73464344090408,
|
| 588 |
+
"learning_rate": 9.855900476564365e-06,
|
| 589 |
+
"loss": 0.7445,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.1712752785802724,
|
| 594 |
+
"grad_norm": 2.7258763684985303,
|
| 595 |
+
"learning_rate": 9.847187467893228e-06,
|
| 596 |
+
"loss": 0.7343,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.17333883615352869,
|
| 601 |
+
"grad_norm": 2.6760448573907354,
|
| 602 |
+
"learning_rate": 9.83822279892528e-06,
|
| 603 |
+
"loss": 0.7451,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.17540239372678498,
|
| 608 |
+
"grad_norm": 2.782701927749318,
|
| 609 |
+
"learning_rate": 9.829006935095629e-06,
|
| 610 |
+
"loss": 0.7375,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.17746595130004128,
|
| 615 |
+
"grad_norm": 2.72035422100488,
|
| 616 |
+
"learning_rate": 9.819540354881136e-06,
|
| 617 |
+
"loss": 0.7385,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.17952950887329758,
|
| 622 |
+
"grad_norm": 2.928620231770996,
|
| 623 |
+
"learning_rate": 9.809823549775559e-06,
|
| 624 |
+
"loss": 0.7194,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.18159306644655385,
|
| 629 |
+
"grad_norm": 3.127420500265884,
|
| 630 |
+
"learning_rate": 9.79985702426404e-06,
|
| 631 |
+
"loss": 0.7191,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.18365662401981014,
|
| 636 |
+
"grad_norm": 3.075733323456217,
|
| 637 |
+
"learning_rate": 9.78964129579691e-06,
|
| 638 |
+
"loss": 0.7327,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.18572018159306644,
|
| 643 |
+
"grad_norm": 2.9108156590411336,
|
| 644 |
+
"learning_rate": 9.779176894762833e-06,
|
| 645 |
+
"loss": 0.7314,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.18778373916632274,
|
| 650 |
+
"grad_norm": 2.946027511832701,
|
| 651 |
+
"learning_rate": 9.768464364461248e-06,
|
| 652 |
+
"loss": 0.7154,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.18984729673957904,
|
| 657 |
+
"grad_norm": 2.9516974484587863,
|
| 658 |
+
"learning_rate": 9.75750426107419e-06,
|
| 659 |
+
"loss": 0.6993,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.19191085431283533,
|
| 664 |
+
"grad_norm": 2.8592349880443897,
|
| 665 |
+
"learning_rate": 9.746297153637386e-06,
|
| 666 |
+
"loss": 0.7116,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.19397441188609163,
|
| 671 |
+
"grad_norm": 2.9554239306486596,
|
| 672 |
+
"learning_rate": 9.73484362401073e-06,
|
| 673 |
+
"loss": 0.6998,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.19603796945934793,
|
| 678 |
+
"grad_norm": 2.780540575615214,
|
| 679 |
+
"learning_rate": 9.723144266848073e-06,
|
| 680 |
+
"loss": 0.7175,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.1981015270326042,
|
| 685 |
+
"grad_norm": 2.7790855329011657,
|
| 686 |
+
"learning_rate": 9.71119968956633e-06,
|
| 687 |
+
"loss": 0.7061,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.2001650846058605,
|
| 692 |
+
"grad_norm": 2.811087864792907,
|
| 693 |
+
"learning_rate": 9.69901051231397e-06,
|
| 694 |
+
"loss": 0.7239,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.2022286421791168,
|
| 699 |
+
"grad_norm": 2.7071720679814937,
|
| 700 |
+
"learning_rate": 9.686577367938802e-06,
|
| 701 |
+
"loss": 0.6857,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.2042921997523731,
|
| 706 |
+
"grad_norm": 2.9052904699729654,
|
| 707 |
+
"learning_rate": 9.673900901955118e-06,
|
| 708 |
+
"loss": 0.7026,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.20635575732562939,
|
| 713 |
+
"grad_norm": 2.7428004689795284,
|
| 714 |
+
"learning_rate": 9.66098177251019e-06,
|
| 715 |
+
"loss": 0.7043,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.20841931489888568,
|
| 720 |
+
"grad_norm": 2.9542233401997797,
|
| 721 |
+
"learning_rate": 9.647820650350087e-06,
|
| 722 |
+
"loss": 0.7008,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.21048287247214198,
|
| 727 |
+
"grad_norm": 2.8206780344207005,
|
| 728 |
+
"learning_rate": 9.634418218784856e-06,
|
| 729 |
+
"loss": 0.6827,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.21254643004539828,
|
| 734 |
+
"grad_norm": 2.7820042566137917,
|
| 735 |
+
"learning_rate": 9.620775173653055e-06,
|
| 736 |
+
"loss": 0.6905,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.21460998761865457,
|
| 741 |
+
"grad_norm": 2.8617753014266354,
|
| 742 |
+
"learning_rate": 9.606892223285604e-06,
|
| 743 |
+
"loss": 0.6875,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.21667354519191084,
|
| 748 |
+
"grad_norm": 2.818801617888441,
|
| 749 |
+
"learning_rate": 9.592770088469032e-06,
|
| 750 |
+
"loss": 0.7143,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.21873710276516714,
|
| 755 |
+
"grad_norm": 2.803399949113071,
|
| 756 |
+
"learning_rate": 9.578409502408037e-06,
|
| 757 |
+
"loss": 0.6596,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.22080066033842344,
|
| 762 |
+
"grad_norm": 2.832046830498785,
|
| 763 |
+
"learning_rate": 9.563811210687433e-06,
|
| 764 |
+
"loss": 0.6912,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.22286421791167974,
|
| 769 |
+
"grad_norm": 2.7793622312334842,
|
| 770 |
+
"learning_rate": 9.548975971233427e-06,
|
| 771 |
+
"loss": 0.6945,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.22492777548493603,
|
| 776 |
+
"grad_norm": 2.8345021788595073,
|
| 777 |
+
"learning_rate": 9.53390455427428e-06,
|
| 778 |
+
"loss": 0.6767,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.22699133305819233,
|
| 783 |
+
"grad_norm": 2.7737306964899284,
|
| 784 |
+
"learning_rate": 9.518597742300309e-06,
|
| 785 |
+
"loss": 0.6816,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.22905489063144863,
|
| 790 |
+
"grad_norm": 2.7181943065638388,
|
| 791 |
+
"learning_rate": 9.503056330023267e-06,
|
| 792 |
+
"loss": 0.6797,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.23111844820470492,
|
| 797 |
+
"grad_norm": 2.935392215628063,
|
| 798 |
+
"learning_rate": 9.48728112433508e-06,
|
| 799 |
+
"loss": 0.6993,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2331820057779612,
|
| 804 |
+
"grad_norm": 3.339231195371882,
|
| 805 |
+
"learning_rate": 9.471272944265948e-06,
|
| 806 |
+
"loss": 0.6565,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2352455633512175,
|
| 811 |
+
"grad_norm": 2.588226117443037,
|
| 812 |
+
"learning_rate": 9.45503262094184e-06,
|
| 813 |
+
"loss": 0.6685,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.2373091209244738,
|
| 818 |
+
"grad_norm": 2.7591510109484507,
|
| 819 |
+
"learning_rate": 9.438560997541319e-06,
|
| 820 |
+
"loss": 0.6712,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.23937267849773008,
|
| 825 |
+
"grad_norm": 2.5942405830165125,
|
| 826 |
+
"learning_rate": 9.421858929251786e-06,
|
| 827 |
+
"loss": 0.6595,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.24143623607098638,
|
| 832 |
+
"grad_norm": 3.1536570604359064,
|
| 833 |
+
"learning_rate": 9.404927283225064e-06,
|
| 834 |
+
"loss": 0.6624,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.24349979364424268,
|
| 839 |
+
"grad_norm": 2.6761790991992402,
|
| 840 |
+
"learning_rate": 9.387766938532386e-06,
|
| 841 |
+
"loss": 0.6481,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.24556335121749898,
|
| 846 |
+
"grad_norm": 2.559657538042446,
|
| 847 |
+
"learning_rate": 9.370378786118755e-06,
|
| 848 |
+
"loss": 0.6518,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.24762690879075527,
|
| 853 |
+
"grad_norm": 2.70639753951879,
|
| 854 |
+
"learning_rate": 9.352763728756677e-06,
|
| 855 |
+
"loss": 0.6704,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.24969046636401154,
|
| 860 |
+
"grad_norm": 2.7368008412010703,
|
| 861 |
+
"learning_rate": 9.334922680999304e-06,
|
| 862 |
+
"loss": 0.6621,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.25175402393726787,
|
| 867 |
+
"grad_norm": 2.852828202511116,
|
| 868 |
+
"learning_rate": 9.316856569132942e-06,
|
| 869 |
+
"loss": 0.6552,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.25381758151052414,
|
| 874 |
+
"grad_norm": 2.5928866487990057,
|
| 875 |
+
"learning_rate": 9.29856633112896e-06,
|
| 876 |
+
"loss": 0.666,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.25588113908378046,
|
| 881 |
+
"grad_norm": 2.7913637236454574,
|
| 882 |
+
"learning_rate": 9.280052916595098e-06,
|
| 883 |
+
"loss": 0.6661,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.25794469665703673,
|
| 888 |
+
"grad_norm": 2.5894434588677164,
|
| 889 |
+
"learning_rate": 9.261317286726157e-06,
|
| 890 |
+
"loss": 0.6654,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.260008254230293,
|
| 895 |
+
"grad_norm": 2.628011843241579,
|
| 896 |
+
"learning_rate": 9.242360414254098e-06,
|
| 897 |
+
"loss": 0.6649,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2620718118035493,
|
| 902 |
+
"grad_norm": 2.7172847524203636,
|
| 903 |
+
"learning_rate": 9.223183283397538e-06,
|
| 904 |
+
"loss": 0.6662,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.2641353693768056,
|
| 909 |
+
"grad_norm": 2.659395752307675,
|
| 910 |
+
"learning_rate": 9.203786889810655e-06,
|
| 911 |
+
"loss": 0.6285,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.2661989269500619,
|
| 916 |
+
"grad_norm": 2.90375149194139,
|
| 917 |
+
"learning_rate": 9.18417224053149e-06,
|
| 918 |
+
"loss": 0.6479,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.2682624845233182,
|
| 923 |
+
"grad_norm": 2.6966774373935927,
|
| 924 |
+
"learning_rate": 9.16434035392966e-06,
|
| 925 |
+
"loss": 0.6397,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.2703260420965745,
|
| 930 |
+
"grad_norm": 2.7576279699971713,
|
| 931 |
+
"learning_rate": 9.144292259653493e-06,
|
| 932 |
+
"loss": 0.651,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.2723895996698308,
|
| 937 |
+
"grad_norm": 2.8455714787136364,
|
| 938 |
+
"learning_rate": 9.124028998576568e-06,
|
| 939 |
+
"loss": 0.634,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.2744531572430871,
|
| 944 |
+
"grad_norm": 2.524900004589425,
|
| 945 |
+
"learning_rate": 9.103551622743667e-06,
|
| 946 |
+
"loss": 0.6393,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.2765167148163434,
|
| 951 |
+
"grad_norm": 2.672726083965731,
|
| 952 |
+
"learning_rate": 9.082861195316164e-06,
|
| 953 |
+
"loss": 0.6321,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.27858027238959965,
|
| 958 |
+
"grad_norm": 3.0535454448794637,
|
| 959 |
+
"learning_rate": 9.061958790516821e-06,
|
| 960 |
+
"loss": 0.6219,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.280643829962856,
|
| 965 |
+
"grad_norm": 2.961262579490319,
|
| 966 |
+
"learning_rate": 9.040845493574016e-06,
|
| 967 |
+
"loss": 0.6239,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.28270738753611224,
|
| 972 |
+
"grad_norm": 2.867116023710867,
|
| 973 |
+
"learning_rate": 9.019522400665397e-06,
|
| 974 |
+
"loss": 0.6183,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.28477094510936857,
|
| 979 |
+
"grad_norm": 2.70507865608555,
|
| 980 |
+
"learning_rate": 8.99799061886098e-06,
|
| 981 |
+
"loss": 0.6119,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.28683450268262484,
|
| 986 |
+
"grad_norm": 2.7686804191515404,
|
| 987 |
+
"learning_rate": 8.976251266065663e-06,
|
| 988 |
+
"loss": 0.6205,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.28889806025588116,
|
| 993 |
+
"grad_norm": 2.7663678432556926,
|
| 994 |
+
"learning_rate": 8.95430547096118e-06,
|
| 995 |
+
"loss": 0.6197,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.29096161782913743,
|
| 1000 |
+
"grad_norm": 2.6693543634519115,
|
| 1001 |
+
"learning_rate": 8.932154372947512e-06,
|
| 1002 |
+
"loss": 0.616,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.2930251754023937,
|
| 1007 |
+
"grad_norm": 2.606058194100552,
|
| 1008 |
+
"learning_rate": 8.90979912208373e-06,
|
| 1009 |
+
"loss": 0.5959,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.29508873297565,
|
| 1014 |
+
"grad_norm": 2.9126080020507903,
|
| 1015 |
+
"learning_rate": 8.887240879028276e-06,
|
| 1016 |
+
"loss": 0.6109,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.2971522905489063,
|
| 1021 |
+
"grad_norm": 2.712632905481665,
|
| 1022 |
+
"learning_rate": 8.864480814978715e-06,
|
| 1023 |
+
"loss": 0.6176,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.2992158481221626,
|
| 1028 |
+
"grad_norm": 2.92931205166235,
|
| 1029 |
+
"learning_rate": 8.841520111610914e-06,
|
| 1030 |
+
"loss": 0.6152,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3012794056954189,
|
| 1035 |
+
"grad_norm": 2.757497057768333,
|
| 1036 |
+
"learning_rate": 8.818359961017705e-06,
|
| 1037 |
+
"loss": 0.6269,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.3033429632686752,
|
| 1042 |
+
"grad_norm": 2.911039679371386,
|
| 1043 |
+
"learning_rate": 8.795001565646983e-06,
|
| 1044 |
+
"loss": 0.617,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3054065208419315,
|
| 1049 |
+
"grad_norm": 2.6754121417561922,
|
| 1050 |
+
"learning_rate": 8.771446138239282e-06,
|
| 1051 |
+
"loss": 0.6079,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3074700784151878,
|
| 1056 |
+
"grad_norm": 2.9820721402144947,
|
| 1057 |
+
"learning_rate": 8.747694901764807e-06,
|
| 1058 |
+
"loss": 0.6028,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3095336359884441,
|
| 1063 |
+
"grad_norm": 2.7199956091577353,
|
| 1064 |
+
"learning_rate": 8.723749089359941e-06,
|
| 1065 |
+
"loss": 0.613,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.31159719356170035,
|
| 1070 |
+
"grad_norm": 2.587929517491648,
|
| 1071 |
+
"learning_rate": 8.699609944263219e-06,
|
| 1072 |
+
"loss": 0.6113,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3136607511349567,
|
| 1077 |
+
"grad_norm": 2.7061730588333335,
|
| 1078 |
+
"learning_rate": 8.675278719750788e-06,
|
| 1079 |
+
"loss": 0.6027,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.31572430870821294,
|
| 1084 |
+
"grad_norm": 2.79901699476363,
|
| 1085 |
+
"learning_rate": 8.65075667907133e-06,
|
| 1086 |
+
"loss": 0.6025,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.31778786628146927,
|
| 1091 |
+
"grad_norm": 2.7737439842011193,
|
| 1092 |
+
"learning_rate": 8.626045095380477e-06,
|
| 1093 |
+
"loss": 0.6044,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.31985142385472554,
|
| 1098 |
+
"grad_norm": 2.6503735236627453,
|
| 1099 |
+
"learning_rate": 8.601145251674718e-06,
|
| 1100 |
+
"loss": 0.5873,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.32191498142798186,
|
| 1105 |
+
"grad_norm": 2.8347160208255078,
|
| 1106 |
+
"learning_rate": 8.576058440724777e-06,
|
| 1107 |
+
"loss": 0.5828,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.32397853900123813,
|
| 1112 |
+
"grad_norm": 2.8544698266992996,
|
| 1113 |
+
"learning_rate": 8.5507859650085e-06,
|
| 1114 |
+
"loss": 0.5939,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.32604209657449446,
|
| 1119 |
+
"grad_norm": 2.686873043228608,
|
| 1120 |
+
"learning_rate": 8.525329136643227e-06,
|
| 1121 |
+
"loss": 0.5821,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3281056541477507,
|
| 1126 |
+
"grad_norm": 2.8021270796404,
|
| 1127 |
+
"learning_rate": 8.499689277317675e-06,
|
| 1128 |
+
"loss": 0.5871,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.330169211721007,
|
| 1133 |
+
"grad_norm": 2.5870353213265402,
|
| 1134 |
+
"learning_rate": 8.473867718223317e-06,
|
| 1135 |
+
"loss": 0.5813,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.3322327692942633,
|
| 1140 |
+
"grad_norm": 2.7498697085626844,
|
| 1141 |
+
"learning_rate": 8.447865799985258e-06,
|
| 1142 |
+
"loss": 0.5937,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.3342963268675196,
|
| 1147 |
+
"grad_norm": 2.6169464644713014,
|
| 1148 |
+
"learning_rate": 8.421684872592643e-06,
|
| 1149 |
+
"loss": 0.5769,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.3363598844407759,
|
| 1154 |
+
"grad_norm": 2.7371393784664804,
|
| 1155 |
+
"learning_rate": 8.395326295328562e-06,
|
| 1156 |
+
"loss": 0.5806,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.3384234420140322,
|
| 1161 |
+
"grad_norm": 2.6465634165278575,
|
| 1162 |
+
"learning_rate": 8.368791436699482e-06,
|
| 1163 |
+
"loss": 0.5914,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3404869995872885,
|
| 1168 |
+
"grad_norm": 2.57951699963865,
|
| 1169 |
+
"learning_rate": 8.342081674364182e-06,
|
| 1170 |
+
"loss": 0.5711,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3425505571605448,
|
| 1175 |
+
"grad_norm": 2.726507330124962,
|
| 1176 |
+
"learning_rate": 8.315198395062246e-06,
|
| 1177 |
+
"loss": 0.5789,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.34461411473380105,
|
| 1182 |
+
"grad_norm": 2.6932144440323267,
|
| 1183 |
+
"learning_rate": 8.28814299454205e-06,
|
| 1184 |
+
"loss": 0.5868,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.34667767230705737,
|
| 1189 |
+
"grad_norm": 2.6545019349833052,
|
| 1190 |
+
"learning_rate": 8.260916877488308e-06,
|
| 1191 |
+
"loss": 0.585,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.34874122988031364,
|
| 1196 |
+
"grad_norm": 2.6267015172396544,
|
| 1197 |
+
"learning_rate": 8.233521457449131e-06,
|
| 1198 |
+
"loss": 0.5713,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.35080478745356997,
|
| 1203 |
+
"grad_norm": 2.446024579716919,
|
| 1204 |
+
"learning_rate": 8.205958156762647e-06,
|
| 1205 |
+
"loss": 0.5668,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.35286834502682624,
|
| 1210 |
+
"grad_norm": 2.5506551014516208,
|
| 1211 |
+
"learning_rate": 8.178228406483145e-06,
|
| 1212 |
+
"loss": 0.5683,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.35493190260008256,
|
| 1217 |
+
"grad_norm": 2.5964627157082,
|
| 1218 |
+
"learning_rate": 8.150333646306787e-06,
|
| 1219 |
+
"loss": 0.5631,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.35699546017333883,
|
| 1224 |
+
"grad_norm": 2.7366272191977212,
|
| 1225 |
+
"learning_rate": 8.122275324496855e-06,
|
| 1226 |
+
"loss": 0.5854,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.35905901774659515,
|
| 1231 |
+
"grad_norm": 2.470623506679956,
|
| 1232 |
+
"learning_rate": 8.094054897808556e-06,
|
| 1233 |
+
"loss": 0.565,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.3611225753198514,
|
| 1238 |
+
"grad_norm": 2.6937122631536496,
|
| 1239 |
+
"learning_rate": 8.065673831413396e-06,
|
| 1240 |
+
"loss": 0.5644,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.3631861328931077,
|
| 1245 |
+
"grad_norm": 3.8142387103231847,
|
| 1246 |
+
"learning_rate": 8.037133598823107e-06,
|
| 1247 |
+
"loss": 0.5799,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.365249690466364,
|
| 1252 |
+
"grad_norm": 2.760510177003008,
|
| 1253 |
+
"learning_rate": 8.008435681813139e-06,
|
| 1254 |
+
"loss": 0.566,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.3673132480396203,
|
| 1259 |
+
"grad_norm": 2.6376344821820354,
|
| 1260 |
+
"learning_rate": 7.979581570345737e-06,
|
| 1261 |
+
"loss": 0.572,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.3693768056128766,
|
| 1266 |
+
"grad_norm": 2.5348136936661043,
|
| 1267 |
+
"learning_rate": 7.950572762492577e-06,
|
| 1268 |
+
"loss": 0.5529,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.3714403631861329,
|
| 1273 |
+
"grad_norm": 2.551175872300915,
|
| 1274 |
+
"learning_rate": 7.92141076435699e-06,
|
| 1275 |
+
"loss": 0.5593,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.3735039207593892,
|
| 1280 |
+
"grad_norm": 2.609535679611351,
|
| 1281 |
+
"learning_rate": 7.892097089995767e-06,
|
| 1282 |
+
"loss": 0.5491,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.3755674783326455,
|
| 1287 |
+
"grad_norm": 2.670033948965714,
|
| 1288 |
+
"learning_rate": 7.862633261340556e-06,
|
| 1289 |
+
"loss": 0.5446,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.3776310359059018,
|
| 1294 |
+
"grad_norm": 2.406170232856436,
|
| 1295 |
+
"learning_rate": 7.83302080811883e-06,
|
| 1296 |
+
"loss": 0.535,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.37969459347915807,
|
| 1301 |
+
"grad_norm": 2.5973677086378024,
|
| 1302 |
+
"learning_rate": 7.803261267774484e-06,
|
| 1303 |
+
"loss": 0.5501,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.38175815105241434,
|
| 1308 |
+
"grad_norm": 2.7913389668469977,
|
| 1309 |
+
"learning_rate": 7.773356185388003e-06,
|
| 1310 |
+
"loss": 0.5358,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.38382170862567067,
|
| 1315 |
+
"grad_norm": 2.395012173415671,
|
| 1316 |
+
"learning_rate": 7.74330711359624e-06,
|
| 1317 |
+
"loss": 0.5492,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.38588526619892694,
|
| 1322 |
+
"grad_norm": 2.6947690399272153,
|
| 1323 |
+
"learning_rate": 7.713115612511815e-06,
|
| 1324 |
+
"loss": 0.5481,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.38794882377218326,
|
| 1329 |
+
"grad_norm": 2.711638332847109,
|
| 1330 |
+
"learning_rate": 7.682783249642104e-06,
|
| 1331 |
+
"loss": 0.5533,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.39001238134543953,
|
| 1336 |
+
"grad_norm": 2.616283183023565,
|
| 1337 |
+
"learning_rate": 7.652311599807868e-06,
|
| 1338 |
+
"loss": 0.5506,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.39207593891869585,
|
| 1343 |
+
"grad_norm": 2.4832951603969162,
|
| 1344 |
+
"learning_rate": 7.62170224506148e-06,
|
| 1345 |
+
"loss": 0.5316,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.3941394964919521,
|
| 1350 |
+
"grad_norm": 2.6721377852076804,
|
| 1351 |
+
"learning_rate": 7.590956774604791e-06,
|
| 1352 |
+
"loss": 0.5458,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.3962030540652084,
|
| 1357 |
+
"grad_norm": 2.777241124185147,
|
| 1358 |
+
"learning_rate": 7.5600767847066215e-06,
|
| 1359 |
+
"loss": 0.5527,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.3982666116384647,
|
| 1364 |
+
"grad_norm": 2.5914713737840924,
|
| 1365 |
+
"learning_rate": 7.529063878619886e-06,
|
| 1366 |
+
"loss": 0.5484,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.400330169211721,
|
| 1371 |
+
"grad_norm": 2.573321705871262,
|
| 1372 |
+
"learning_rate": 7.497919666498346e-06,
|
| 1373 |
+
"loss": 0.5435,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.4023937267849773,
|
| 1378 |
+
"grad_norm": 2.6520337314396993,
|
| 1379 |
+
"learning_rate": 7.466645765313023e-06,
|
| 1380 |
+
"loss": 0.5358,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4044572843582336,
|
| 1385 |
+
"grad_norm": 2.7056801557526486,
|
| 1386 |
+
"learning_rate": 7.435243798768242e-06,
|
| 1387 |
+
"loss": 0.5499,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.4065208419314899,
|
| 1392 |
+
"grad_norm": 2.6559172388700047,
|
| 1393 |
+
"learning_rate": 7.403715397217335e-06,
|
| 1394 |
+
"loss": 0.5241,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.4085843995047462,
|
| 1399 |
+
"grad_norm": 2.5216304625498647,
|
| 1400 |
+
"learning_rate": 7.3720621975779815e-06,
|
| 1401 |
+
"loss": 0.521,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4106479570780025,
|
| 1406 |
+
"grad_norm": 2.519874224942243,
|
| 1407 |
+
"learning_rate": 7.3402858432472416e-06,
|
| 1408 |
+
"loss": 0.5489,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.41271151465125877,
|
| 1413 |
+
"grad_norm": 2.518659981447909,
|
| 1414 |
+
"learning_rate": 7.308387984016217e-06,
|
| 1415 |
+
"loss": 0.5266,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.41477507222451504,
|
| 1420 |
+
"grad_norm": 2.721359929344381,
|
| 1421 |
+
"learning_rate": 7.276370275984408e-06,
|
| 1422 |
+
"loss": 0.5325,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.41683862979777137,
|
| 1427 |
+
"grad_norm": 2.575184642938111,
|
| 1428 |
+
"learning_rate": 7.2442343814737135e-06,
|
| 1429 |
+
"loss": 0.5357,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.41890218737102763,
|
| 1434 |
+
"grad_norm": 2.5077385160634904,
|
| 1435 |
+
"learning_rate": 7.211981968942147e-06,
|
| 1436 |
+
"loss": 0.524,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.42096574494428396,
|
| 1441 |
+
"grad_norm": 2.453550934446932,
|
| 1442 |
+
"learning_rate": 7.179614712897195e-06,
|
| 1443 |
+
"loss": 0.5366,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.42302930251754023,
|
| 1448 |
+
"grad_norm": 2.561635348596983,
|
| 1449 |
+
"learning_rate": 7.147134293808885e-06,
|
| 1450 |
+
"loss": 0.5117,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.42509286009079655,
|
| 1455 |
+
"grad_norm": 2.501066318480838,
|
| 1456 |
+
"learning_rate": 7.114542398022539e-06,
|
| 1457 |
+
"loss": 0.524,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.4271564176640528,
|
| 1462 |
+
"grad_norm": 2.7215445820801714,
|
| 1463 |
+
"learning_rate": 7.08184071767122e-06,
|
| 1464 |
+
"loss": 0.5111,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.42921997523730915,
|
| 1469 |
+
"grad_norm": 2.6088491234313604,
|
| 1470 |
+
"learning_rate": 7.049030950587874e-06,
|
| 1471 |
+
"loss": 0.5113,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.4312835328105654,
|
| 1476 |
+
"grad_norm": 2.551009519200941,
|
| 1477 |
+
"learning_rate": 7.016114800217182e-06,
|
| 1478 |
+
"loss": 0.5136,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.4333470903838217,
|
| 1483 |
+
"grad_norm": 2.457255947220057,
|
| 1484 |
+
"learning_rate": 6.983093975527128e-06,
|
| 1485 |
+
"loss": 0.5208,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.435410647957078,
|
| 1490 |
+
"grad_norm": 2.6994910463118913,
|
| 1491 |
+
"learning_rate": 6.949970190920255e-06,
|
| 1492 |
+
"loss": 0.5212,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.4374742055303343,
|
| 1497 |
+
"grad_norm": 2.597397073975052,
|
| 1498 |
+
"learning_rate": 6.916745166144671e-06,
|
| 1499 |
+
"loss": 0.5128,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4395377631035906,
|
| 1504 |
+
"grad_norm": 2.605039455033872,
|
| 1505 |
+
"learning_rate": 6.883420626204751e-06,
|
| 1506 |
+
"loss": 0.5081,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.4416013206768469,
|
| 1511 |
+
"grad_norm": 2.7509624680821,
|
| 1512 |
+
"learning_rate": 6.849998301271585e-06,
|
| 1513 |
+
"loss": 0.516,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.4436648782501032,
|
| 1518 |
+
"grad_norm": 2.4731078067929197,
|
| 1519 |
+
"learning_rate": 6.81647992659314e-06,
|
| 1520 |
+
"loss": 0.5144,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.44572843582335947,
|
| 1525 |
+
"grad_norm": 2.611162217133499,
|
| 1526 |
+
"learning_rate": 6.782867242404181e-06,
|
| 1527 |
+
"loss": 0.4979,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.44779199339661574,
|
| 1532 |
+
"grad_norm": 2.5398801893032474,
|
| 1533 |
+
"learning_rate": 6.749161993835901e-06,
|
| 1534 |
+
"loss": 0.5074,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.44985555096987206,
|
| 1539 |
+
"grad_norm": 2.5951566561368775,
|
| 1540 |
+
"learning_rate": 6.715365930825337e-06,
|
| 1541 |
+
"loss": 0.4975,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.45191910854312833,
|
| 1546 |
+
"grad_norm": 2.6092895827410034,
|
| 1547 |
+
"learning_rate": 6.681480808024503e-06,
|
| 1548 |
+
"loss": 0.5124,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.45398266611638466,
|
| 1553 |
+
"grad_norm": 2.5414548197736706,
|
| 1554 |
+
"learning_rate": 6.6475083847092895e-06,
|
| 1555 |
+
"loss": 0.4957,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.45604622368964093,
|
| 1560 |
+
"grad_norm": 2.8460005959716166,
|
| 1561 |
+
"learning_rate": 6.613450424688131e-06,
|
| 1562 |
+
"loss": 0.5028,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.45810978126289725,
|
| 1567 |
+
"grad_norm": 2.5601376301451753,
|
| 1568 |
+
"learning_rate": 6.579308696210428e-06,
|
| 1569 |
+
"loss": 0.4876,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.4601733388361535,
|
| 1574 |
+
"grad_norm": 2.6807233381503237,
|
| 1575 |
+
"learning_rate": 6.545084971874738e-06,
|
| 1576 |
+
"loss": 0.5155,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.46223689640940985,
|
| 1581 |
+
"grad_norm": 2.5044297534154336,
|
| 1582 |
+
"learning_rate": 6.51078102853675e-06,
|
| 1583 |
+
"loss": 0.5012,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.4643004539826661,
|
| 1588 |
+
"grad_norm": 2.5092897248073474,
|
| 1589 |
+
"learning_rate": 6.47639864721703e-06,
|
| 1590 |
+
"loss": 0.493,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.4663640115559224,
|
| 1595 |
+
"grad_norm": 2.4954489446094392,
|
| 1596 |
+
"learning_rate": 6.44193961300855e-06,
|
| 1597 |
+
"loss": 0.4868,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.4684275691291787,
|
| 1602 |
+
"grad_norm": 2.5451475237887666,
|
| 1603 |
+
"learning_rate": 6.407405714984011e-06,
|
| 1604 |
+
"loss": 0.4866,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.470491126702435,
|
| 1609 |
+
"grad_norm": 2.6098482720462552,
|
| 1610 |
+
"learning_rate": 6.3727987461029565e-06,
|
| 1611 |
+
"loss": 0.4956,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.4725546842756913,
|
| 1616 |
+
"grad_norm": 2.5923597566181296,
|
| 1617 |
+
"learning_rate": 6.338120503118684e-06,
|
| 1618 |
+
"loss": 0.4998,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.4746182418489476,
|
| 1623 |
+
"grad_norm": 2.7721513261613393,
|
| 1624 |
+
"learning_rate": 6.303372786484953e-06,
|
| 1625 |
+
"loss": 0.4948,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.4766817994222039,
|
| 1630 |
+
"grad_norm": 2.4572497473964012,
|
| 1631 |
+
"learning_rate": 6.2685574002625235e-06,
|
| 1632 |
+
"loss": 0.4744,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.47874535699546017,
|
| 1637 |
+
"grad_norm": 2.7248530281028653,
|
| 1638 |
+
"learning_rate": 6.233676152025474e-06,
|
| 1639 |
+
"loss": 0.4789,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.4808089145687165,
|
| 1644 |
+
"grad_norm": 2.510160739180536,
|
| 1645 |
+
"learning_rate": 6.198730852767364e-06,
|
| 1646 |
+
"loss": 0.4924,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.48287247214197276,
|
| 1651 |
+
"grad_norm": 2.5611750613730275,
|
| 1652 |
+
"learning_rate": 6.1637233168072066e-06,
|
| 1653 |
+
"loss": 0.4816,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.48493602971522903,
|
| 1658 |
+
"grad_norm": 2.580939509201322,
|
| 1659 |
+
"learning_rate": 6.1286553616952705e-06,
|
| 1660 |
+
"loss": 0.4826,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.48699958728848536,
|
| 1665 |
+
"grad_norm": 2.579364116128431,
|
| 1666 |
+
"learning_rate": 6.09352880811872e-06,
|
| 1667 |
+
"loss": 0.4926,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.4890631448617416,
|
| 1672 |
+
"grad_norm": 2.5855927528887745,
|
| 1673 |
+
"learning_rate": 6.058345479807077e-06,
|
| 1674 |
+
"loss": 0.4826,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.49112670243499795,
|
| 1679 |
+
"grad_norm": 2.468639972857448,
|
| 1680 |
+
"learning_rate": 6.023107203437547e-06,
|
| 1681 |
+
"loss": 0.4822,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.4931902600082542,
|
| 1686 |
+
"grad_norm": 2.534893122365119,
|
| 1687 |
+
"learning_rate": 5.987815808540169e-06,
|
| 1688 |
+
"loss": 0.4969,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.49525381758151055,
|
| 1693 |
+
"grad_norm": 2.4035384879566095,
|
| 1694 |
+
"learning_rate": 5.95247312740284e-06,
|
| 1695 |
+
"loss": 0.4942,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.4973173751547668,
|
| 1700 |
+
"grad_norm": 2.6544116089729552,
|
| 1701 |
+
"learning_rate": 5.917080994976172e-06,
|
| 1702 |
+
"loss": 0.4678,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.4993809327280231,
|
| 1707 |
+
"grad_norm": 2.6911232973443653,
|
| 1708 |
+
"learning_rate": 5.881641248778235e-06,
|
| 1709 |
+
"loss": 0.4796,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5014444903012794,
|
| 1714 |
+
"grad_norm": 2.9098929338992177,
|
| 1715 |
+
"learning_rate": 5.8461557287991455e-06,
|
| 1716 |
+
"loss": 0.4926,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5035080478745357,
|
| 1721 |
+
"grad_norm": 2.724813117289689,
|
| 1722 |
+
"learning_rate": 5.810626277405548e-06,
|
| 1723 |
+
"loss": 0.4837,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.505571605447792,
|
| 1728 |
+
"grad_norm": 2.5885450608927782,
|
| 1729 |
+
"learning_rate": 5.775054739244948e-06,
|
| 1730 |
+
"loss": 0.4827,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5076351630210483,
|
| 1735 |
+
"grad_norm": 2.4385273279955677,
|
| 1736 |
+
"learning_rate": 5.739442961149949e-06,
|
| 1737 |
+
"loss": 0.4777,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5096987205943045,
|
| 1742 |
+
"grad_norm": 9.983765788937596,
|
| 1743 |
+
"learning_rate": 5.703792792042363e-06,
|
| 1744 |
+
"loss": 0.4657,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5117622781675609,
|
| 1749 |
+
"grad_norm": 2.5992188623557575,
|
| 1750 |
+
"learning_rate": 5.668106082837226e-06,
|
| 1751 |
+
"loss": 0.4677,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5138258357408172,
|
| 1756 |
+
"grad_norm": 2.5248538601870916,
|
| 1757 |
+
"learning_rate": 5.632384686346682e-06,
|
| 1758 |
+
"loss": 0.4648,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5158893933140735,
|
| 1763 |
+
"grad_norm": 2.4528442128112653,
|
| 1764 |
+
"learning_rate": 5.5966304571838084e-06,
|
| 1765 |
+
"loss": 0.468,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5179529508873297,
|
| 1770 |
+
"grad_norm": 2.4985756922245375,
|
| 1771 |
+
"learning_rate": 5.560845251666307e-06,
|
| 1772 |
+
"loss": 0.4691,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.520016508460586,
|
| 1777 |
+
"grad_norm": 2.588098554859599,
|
| 1778 |
+
"learning_rate": 5.525030927720142e-06,
|
| 1779 |
+
"loss": 0.4843,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5220800660338424,
|
| 1784 |
+
"grad_norm": 2.596100924626437,
|
| 1785 |
+
"learning_rate": 5.489189344783072e-06,
|
| 1786 |
+
"loss": 0.4722,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5241436236070987,
|
| 1791 |
+
"grad_norm": 2.4880471647467526,
|
| 1792 |
+
"learning_rate": 5.4533223637081075e-06,
|
| 1793 |
+
"loss": 0.4652,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5262071811803549,
|
| 1798 |
+
"grad_norm": 2.5470383673418926,
|
| 1799 |
+
"learning_rate": 5.417431846666903e-06,
|
| 1800 |
+
"loss": 0.4657,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5282707387536112,
|
| 1805 |
+
"grad_norm": 2.32837647206492,
|
| 1806 |
+
"learning_rate": 5.381519657053072e-06,
|
| 1807 |
+
"loss": 0.4542,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5303342963268676,
|
| 1812 |
+
"grad_norm": 2.5139826780698753,
|
| 1813 |
+
"learning_rate": 5.3455876593854435e-06,
|
| 1814 |
+
"loss": 0.4665,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5323978539001238,
|
| 1819 |
+
"grad_norm": 2.4870649815664496,
|
| 1820 |
+
"learning_rate": 5.309637719211256e-06,
|
| 1821 |
+
"loss": 0.4527,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5344614114733801,
|
| 1826 |
+
"grad_norm": 2.509973264998628,
|
| 1827 |
+
"learning_rate": 5.273671703009301e-06,
|
| 1828 |
+
"loss": 0.4507,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5365249690466364,
|
| 1833 |
+
"grad_norm": 2.690719386362163,
|
| 1834 |
+
"learning_rate": 5.2376914780930225e-06,
|
| 1835 |
+
"loss": 0.4576,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5385885266198926,
|
| 1840 |
+
"grad_norm": 4.083713443737158,
|
| 1841 |
+
"learning_rate": 5.201698912513561e-06,
|
| 1842 |
+
"loss": 0.4412,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.540652084193149,
|
| 1847 |
+
"grad_norm": 2.3850290751433016,
|
| 1848 |
+
"learning_rate": 5.165695874962769e-06,
|
| 1849 |
+
"loss": 0.4633,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.5427156417664053,
|
| 1854 |
+
"grad_norm": 2.383468999919073,
|
| 1855 |
+
"learning_rate": 5.129684234676195e-06,
|
| 1856 |
+
"loss": 0.4482,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.5447791993396616,
|
| 1861 |
+
"grad_norm": 2.3602215294945568,
|
| 1862 |
+
"learning_rate": 5.0936658613360304e-06,
|
| 1863 |
+
"loss": 0.4683,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.5468427569129178,
|
| 1868 |
+
"grad_norm": 2.3769603536661035,
|
| 1869 |
+
"learning_rate": 5.057642624974036e-06,
|
| 1870 |
+
"loss": 0.4587,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.5489063144861742,
|
| 1875 |
+
"grad_norm": 2.5837187422329837,
|
| 1876 |
+
"learning_rate": 5.021616395874459e-06,
|
| 1877 |
+
"loss": 0.439,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.5509698720594305,
|
| 1882 |
+
"grad_norm": 2.587958298155525,
|
| 1883 |
+
"learning_rate": 4.9855890444769226e-06,
|
| 1884 |
+
"loss": 0.4547,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.5530334296326868,
|
| 1889 |
+
"grad_norm": 2.4489991448969795,
|
| 1890 |
+
"learning_rate": 4.94956244127932e-06,
|
| 1891 |
+
"loss": 0.4435,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.555096987205943,
|
| 1896 |
+
"grad_norm": 2.70305559316515,
|
| 1897 |
+
"learning_rate": 4.9135384567407e-06,
|
| 1898 |
+
"loss": 0.4635,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.5571605447791993,
|
| 1903 |
+
"grad_norm": 2.552605702749918,
|
| 1904 |
+
"learning_rate": 4.877518961184149e-06,
|
| 1905 |
+
"loss": 0.4395,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.5592241023524557,
|
| 1910 |
+
"grad_norm": 2.4491029644956717,
|
| 1911 |
+
"learning_rate": 4.841505824699697e-06,
|
| 1912 |
+
"loss": 0.4446,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.561287659925712,
|
| 1917 |
+
"grad_norm": 2.2903925047271003,
|
| 1918 |
+
"learning_rate": 4.805500917047217e-06,
|
| 1919 |
+
"loss": 0.4398,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.5633512174989682,
|
| 1924 |
+
"grad_norm": 2.379688312715005,
|
| 1925 |
+
"learning_rate": 4.769506107559344e-06,
|
| 1926 |
+
"loss": 0.4515,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.5654147750722245,
|
| 1931 |
+
"grad_norm": 2.881348434158553,
|
| 1932 |
+
"learning_rate": 4.7335232650444376e-06,
|
| 1933 |
+
"loss": 0.4609,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.5674783326454808,
|
| 1938 |
+
"grad_norm": 2.450954066121652,
|
| 1939 |
+
"learning_rate": 4.697554257689541e-06,
|
| 1940 |
+
"loss": 0.4556,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.5695418902187371,
|
| 1945 |
+
"grad_norm": 2.418284415029761,
|
| 1946 |
+
"learning_rate": 4.661600952963391e-06,
|
| 1947 |
+
"loss": 0.4452,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.5716054477919934,
|
| 1952 |
+
"grad_norm": 2.3480748218612457,
|
| 1953 |
+
"learning_rate": 4.6256652175194675e-06,
|
| 1954 |
+
"loss": 0.461,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.5736690053652497,
|
| 1959 |
+
"grad_norm": 2.472904201761537,
|
| 1960 |
+
"learning_rate": 4.589748917099067e-06,
|
| 1961 |
+
"loss": 0.4492,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.5757325629385059,
|
| 1966 |
+
"grad_norm": 2.468262838594691,
|
| 1967 |
+
"learning_rate": 4.553853916434448e-06,
|
| 1968 |
+
"loss": 0.4403,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.5777961205117623,
|
| 1973 |
+
"grad_norm": 2.361041942268244,
|
| 1974 |
+
"learning_rate": 4.517982079152009e-06,
|
| 1975 |
+
"loss": 0.4377,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.5798596780850186,
|
| 1980 |
+
"grad_norm": 2.5500737849526423,
|
| 1981 |
+
"learning_rate": 4.482135267675529e-06,
|
| 1982 |
+
"loss": 0.45,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.5819232356582749,
|
| 1987 |
+
"grad_norm": 2.453235877539221,
|
| 1988 |
+
"learning_rate": 4.446315343129484e-06,
|
| 1989 |
+
"loss": 0.4328,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.5839867932315311,
|
| 1994 |
+
"grad_norm": 2.3281079794605355,
|
| 1995 |
+
"learning_rate": 4.410524165242407e-06,
|
| 1996 |
+
"loss": 0.4374,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.5860503508047874,
|
| 2001 |
+
"grad_norm": 2.551505972004267,
|
| 2002 |
+
"learning_rate": 4.3747635922503364e-06,
|
| 2003 |
+
"loss": 0.4432,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.5881139083780438,
|
| 2008 |
+
"grad_norm": 2.5826844758469276,
|
| 2009 |
+
"learning_rate": 4.3390354808003464e-06,
|
| 2010 |
+
"loss": 0.4402,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.5901774659513,
|
| 2015 |
+
"grad_norm": 2.3612299259996936,
|
| 2016 |
+
"learning_rate": 4.30334168585414e-06,
|
| 2017 |
+
"loss": 0.4388,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.5922410235245563,
|
| 2022 |
+
"grad_norm": 2.4199132831049246,
|
| 2023 |
+
"learning_rate": 4.26768406059175e-06,
|
| 2024 |
+
"loss": 0.4296,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.5943045810978126,
|
| 2029 |
+
"grad_norm": 2.3293439174487274,
|
| 2030 |
+
"learning_rate": 4.2320644563153244e-06,
|
| 2031 |
+
"loss": 0.4391,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.596368138671069,
|
| 2036 |
+
"grad_norm": 2.5508830684770043,
|
| 2037 |
+
"learning_rate": 4.196484722352999e-06,
|
| 2038 |
+
"loss": 0.4341,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.5984316962443252,
|
| 2043 |
+
"grad_norm": 2.459534568261172,
|
| 2044 |
+
"learning_rate": 4.160946705962895e-06,
|
| 2045 |
+
"loss": 0.4298,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6004952538175815,
|
| 2050 |
+
"grad_norm": 2.4607489784078806,
|
| 2051 |
+
"learning_rate": 4.12545225223721e-06,
|
| 2052 |
+
"loss": 0.4238,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6025588113908378,
|
| 2057 |
+
"grad_norm": 2.5772914418218336,
|
| 2058 |
+
"learning_rate": 4.090003204006406e-06,
|
| 2059 |
+
"loss": 0.4165,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.604622368964094,
|
| 2064 |
+
"grad_norm": 2.273824183445231,
|
| 2065 |
+
"learning_rate": 4.054601401743561e-06,
|
| 2066 |
+
"loss": 0.4366,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6066859265373504,
|
| 2071 |
+
"grad_norm": 2.5922865720792183,
|
| 2072 |
+
"learning_rate": 4.019248683468781e-06,
|
| 2073 |
+
"loss": 0.4372,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6087494841106067,
|
| 2078 |
+
"grad_norm": 2.520992944241331,
|
| 2079 |
+
"learning_rate": 3.983946884653804e-06,
|
| 2080 |
+
"loss": 0.4418,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.610813041683863,
|
| 2085 |
+
"grad_norm": 2.4527707525439415,
|
| 2086 |
+
"learning_rate": 3.948697838126681e-06,
|
| 2087 |
+
"loss": 0.4186,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6128765992571192,
|
| 2092 |
+
"grad_norm": 2.3893935033867546,
|
| 2093 |
+
"learning_rate": 3.9135033739766275e-06,
|
| 2094 |
+
"loss": 0.4332,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6149401568303756,
|
| 2099 |
+
"grad_norm": 2.303193161104287,
|
| 2100 |
+
"learning_rate": 3.8783653194590055e-06,
|
| 2101 |
+
"loss": 0.4308,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6170037144036319,
|
| 2106 |
+
"grad_norm": 2.568914194114212,
|
| 2107 |
+
"learning_rate": 3.843285498900457e-06,
|
| 2108 |
+
"loss": 0.4238,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6190672719768882,
|
| 2113 |
+
"grad_norm": 2.3874626733288093,
|
| 2114 |
+
"learning_rate": 3.808265733604181e-06,
|
| 2115 |
+
"loss": 0.4211,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6211308295501444,
|
| 2120 |
+
"grad_norm": 2.397949299188709,
|
| 2121 |
+
"learning_rate": 3.7733078417553835e-06,
|
| 2122 |
+
"loss": 0.4198,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6231943871234007,
|
| 2127 |
+
"grad_norm": 2.4304137659983223,
|
| 2128 |
+
"learning_rate": 3.7384136383268673e-06,
|
| 2129 |
+
"loss": 0.4364,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6252579446966571,
|
| 2134 |
+
"grad_norm": 2.4027742214175727,
|
| 2135 |
+
"learning_rate": 3.70358493498481e-06,
|
| 2136 |
+
"loss": 0.431,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.6273215022699133,
|
| 2141 |
+
"grad_norm": 2.6288095509256255,
|
| 2142 |
+
"learning_rate": 3.668823539994706e-06,
|
| 2143 |
+
"loss": 0.4252,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6293850598431696,
|
| 2148 |
+
"grad_norm": 2.386671047852444,
|
| 2149 |
+
"learning_rate": 3.634131258127467e-06,
|
| 2150 |
+
"loss": 0.428,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6314486174164259,
|
| 2155 |
+
"grad_norm": 2.3713874902931615,
|
| 2156 |
+
"learning_rate": 3.5995098905657457e-06,
|
| 2157 |
+
"loss": 0.4286,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.6335121749896823,
|
| 2162 |
+
"grad_norm": 2.300141894247011,
|
| 2163 |
+
"learning_rate": 3.564961234810399e-06,
|
| 2164 |
+
"loss": 0.4162,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.6355757325629385,
|
| 2169 |
+
"grad_norm": 2.4635471100860933,
|
| 2170 |
+
"learning_rate": 3.530487084587173e-06,
|
| 2171 |
+
"loss": 0.4102,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.6376392901361948,
|
| 2176 |
+
"grad_norm": 2.374004467594742,
|
| 2177 |
+
"learning_rate": 3.4960892297535765e-06,
|
| 2178 |
+
"loss": 0.4144,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.6397028477094511,
|
| 2183 |
+
"grad_norm": 2.3078125248362324,
|
| 2184 |
+
"learning_rate": 3.4617694562059465e-06,
|
| 2185 |
+
"loss": 0.4076,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.6417664052827073,
|
| 2190 |
+
"grad_norm": 2.513382809869699,
|
| 2191 |
+
"learning_rate": 3.427529545786736e-06,
|
| 2192 |
+
"loss": 0.4103,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.6438299628559637,
|
| 2197 |
+
"grad_norm": 2.421843916725022,
|
| 2198 |
+
"learning_rate": 3.393371276191997e-06,
|
| 2199 |
+
"loss": 0.4151,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.64589352042922,
|
| 2204 |
+
"grad_norm": 2.469381907143848,
|
| 2205 |
+
"learning_rate": 3.3592964208790808e-06,
|
| 2206 |
+
"loss": 0.4164,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.6479570780024763,
|
| 2211 |
+
"grad_norm": 2.531906105074784,
|
| 2212 |
+
"learning_rate": 3.3253067489745716e-06,
|
| 2213 |
+
"loss": 0.4071,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.6500206355757325,
|
| 2218 |
+
"grad_norm": 2.3846085481545747,
|
| 2219 |
+
"learning_rate": 3.291404025182432e-06,
|
| 2220 |
+
"loss": 0.4075,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.6520841931489889,
|
| 2225 |
+
"grad_norm": 2.433814917027714,
|
| 2226 |
+
"learning_rate": 3.2575900096923753e-06,
|
| 2227 |
+
"loss": 0.4087,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.6541477507222452,
|
| 2232 |
+
"grad_norm": 2.4424178299468506,
|
| 2233 |
+
"learning_rate": 3.2238664580884877e-06,
|
| 2234 |
+
"loss": 0.4043,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.6562113082955014,
|
| 2239 |
+
"grad_norm": 2.3585518358327833,
|
| 2240 |
+
"learning_rate": 3.1902351212580734e-06,
|
| 2241 |
+
"loss": 0.4116,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.6582748658687577,
|
| 2246 |
+
"grad_norm": 2.4272402253853174,
|
| 2247 |
+
"learning_rate": 3.1566977453007564e-06,
|
| 2248 |
+
"loss": 0.4045,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.660338423442014,
|
| 2253 |
+
"grad_norm": 2.2735139196354184,
|
| 2254 |
+
"learning_rate": 3.1232560714378214e-06,
|
| 2255 |
+
"loss": 0.3998,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.6624019810152704,
|
| 2260 |
+
"grad_norm": 2.496149939265061,
|
| 2261 |
+
"learning_rate": 3.0899118359218104e-06,
|
| 2262 |
+
"loss": 0.3986,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.6644655385885266,
|
| 2267 |
+
"grad_norm": 2.577482137677985,
|
| 2268 |
+
"learning_rate": 3.056666769946382e-06,
|
| 2269 |
+
"loss": 0.3922,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.6665290961617829,
|
| 2274 |
+
"grad_norm": 2.432120191923341,
|
| 2275 |
+
"learning_rate": 3.0235225995564323e-06,
|
| 2276 |
+
"loss": 0.3959,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.6685926537350392,
|
| 2281 |
+
"grad_norm": 2.5315892714201507,
|
| 2282 |
+
"learning_rate": 2.9904810455584664e-06,
|
| 2283 |
+
"loss": 0.396,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.6706562113082954,
|
| 2288 |
+
"grad_norm": 2.4126722223283643,
|
| 2289 |
+
"learning_rate": 2.9575438234312807e-06,
|
| 2290 |
+
"loss": 0.4034,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.6727197688815518,
|
| 2295 |
+
"grad_norm": 2.3922710783717682,
|
| 2296 |
+
"learning_rate": 2.9247126432368685e-06,
|
| 2297 |
+
"loss": 0.3969,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.6747833264548081,
|
| 2302 |
+
"grad_norm": 2.542933247382595,
|
| 2303 |
+
"learning_rate": 2.8919892095316616e-06,
|
| 2304 |
+
"loss": 0.3914,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.6768468840280644,
|
| 2309 |
+
"grad_norm": 2.503873056742514,
|
| 2310 |
+
"learning_rate": 2.859375221278014e-06,
|
| 2311 |
+
"loss": 0.4053,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.6789104416013206,
|
| 2316 |
+
"grad_norm": 2.4704924711278275,
|
| 2317 |
+
"learning_rate": 2.826872371756e-06,
|
| 2318 |
+
"loss": 0.4171,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.680973999174577,
|
| 2323 |
+
"grad_norm": 2.2126828966128347,
|
| 2324 |
+
"learning_rate": 2.7944823484755053e-06,
|
| 2325 |
+
"loss": 0.3979,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.6830375567478333,
|
| 2330 |
+
"grad_norm": 2.452244039159894,
|
| 2331 |
+
"learning_rate": 2.762206833088608e-06,
|
| 2332 |
+
"loss": 0.3992,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.6851011143210896,
|
| 2337 |
+
"grad_norm": 2.4737732563891197,
|
| 2338 |
+
"learning_rate": 2.7300475013022666e-06,
|
| 2339 |
+
"loss": 0.4013,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.6871646718943458,
|
| 2344 |
+
"grad_norm": 2.32695849246749,
|
| 2345 |
+
"learning_rate": 2.6980060227913273e-06,
|
| 2346 |
+
"loss": 0.3984,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.6892282294676021,
|
| 2351 |
+
"grad_norm": 2.5306901436703497,
|
| 2352 |
+
"learning_rate": 2.666084061111828e-06,
|
| 2353 |
+
"loss": 0.3942,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.6912917870408585,
|
| 2358 |
+
"grad_norm": 2.40430761373431,
|
| 2359 |
+
"learning_rate": 2.6342832736146403e-06,
|
| 2360 |
+
"loss": 0.3943,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.6933553446141147,
|
| 2365 |
+
"grad_norm": 2.3692497196583955,
|
| 2366 |
+
"learning_rate": 2.6026053113594056e-06,
|
| 2367 |
+
"loss": 0.383,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.695418902187371,
|
| 2372 |
+
"grad_norm": 2.45758426227379,
|
| 2373 |
+
"learning_rate": 2.571051819028826e-06,
|
| 2374 |
+
"loss": 0.4056,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.6974824597606273,
|
| 2379 |
+
"grad_norm": 2.490132984626703,
|
| 2380 |
+
"learning_rate": 2.539624434843272e-06,
|
| 2381 |
+
"loss": 0.408,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.6995460173338837,
|
| 2386 |
+
"grad_norm": 2.316910192052694,
|
| 2387 |
+
"learning_rate": 2.508324790475731e-06,
|
| 2388 |
+
"loss": 0.3908,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7016095749071399,
|
| 2393 |
+
"grad_norm": 2.5678586000455215,
|
| 2394 |
+
"learning_rate": 2.4771545109670746e-06,
|
| 2395 |
+
"loss": 0.3886,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7036731324803962,
|
| 2400 |
+
"grad_norm": 2.333143658853672,
|
| 2401 |
+
"learning_rate": 2.446115214641719e-06,
|
| 2402 |
+
"loss": 0.3825,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7057366900536525,
|
| 2407 |
+
"grad_norm": 2.421284956930726,
|
| 2408 |
+
"learning_rate": 2.415208513023574e-06,
|
| 2409 |
+
"loss": 0.378,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7078002476269087,
|
| 2414 |
+
"grad_norm": 2.348370007461818,
|
| 2415 |
+
"learning_rate": 2.3844360107523973e-06,
|
| 2416 |
+
"loss": 0.3745,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7098638052001651,
|
| 2421 |
+
"grad_norm": 2.2561855639412047,
|
| 2422 |
+
"learning_rate": 2.353799305500467e-06,
|
| 2423 |
+
"loss": 0.3768,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7119273627734214,
|
| 2428 |
+
"grad_norm": 2.4144762778450364,
|
| 2429 |
+
"learning_rate": 2.3232999878896407e-06,
|
| 2430 |
+
"loss": 0.3941,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7139909203466777,
|
| 2435 |
+
"grad_norm": 2.510836282225756,
|
| 2436 |
+
"learning_rate": 2.2929396414087737e-06,
|
| 2437 |
+
"loss": 0.4012,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7160544779199339,
|
| 2442 |
+
"grad_norm": 2.4017944766607906,
|
| 2443 |
+
"learning_rate": 2.2627198423314988e-06,
|
| 2444 |
+
"loss": 0.3955,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7181180354931903,
|
| 2449 |
+
"grad_norm": 2.2838737017346356,
|
| 2450 |
+
"learning_rate": 2.2326421596343923e-06,
|
| 2451 |
+
"loss": 0.3833,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.7201815930664466,
|
| 2456 |
+
"grad_norm": 2.590581157502952,
|
| 2457 |
+
"learning_rate": 2.2027081549155188e-06,
|
| 2458 |
+
"loss": 0.3912,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.7222451506397028,
|
| 2463 |
+
"grad_norm": 2.3748009695139776,
|
| 2464 |
+
"learning_rate": 2.172919382313347e-06,
|
| 2465 |
+
"loss": 0.3808,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.7243087082129591,
|
| 2470 |
+
"grad_norm": 2.4065528450546445,
|
| 2471 |
+
"learning_rate": 2.1432773884260627e-06,
|
| 2472 |
+
"loss": 0.3705,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.7263722657862154,
|
| 2477 |
+
"grad_norm": 2.4049387116398564,
|
| 2478 |
+
"learning_rate": 2.1137837122312787e-06,
|
| 2479 |
+
"loss": 0.3765,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.7284358233594718,
|
| 2484 |
+
"grad_norm": 2.5287323783127174,
|
| 2485 |
+
"learning_rate": 2.0844398850061186e-06,
|
| 2486 |
+
"loss": 0.3886,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.730499380932728,
|
| 2491 |
+
"grad_norm": 2.467020632677239,
|
| 2492 |
+
"learning_rate": 2.0552474302477327e-06,
|
| 2493 |
+
"loss": 0.3875,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.7325629385059843,
|
| 2498 |
+
"grad_norm": 2.313861208086701,
|
| 2499 |
+
"learning_rate": 2.0262078635941818e-06,
|
| 2500 |
+
"loss": 0.3889,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.7346264960792406,
|
| 2505 |
+
"grad_norm": 2.3695175969299296,
|
| 2506 |
+
"learning_rate": 1.9973226927457567e-06,
|
| 2507 |
+
"loss": 0.3763,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.736690053652497,
|
| 2512 |
+
"grad_norm": 2.4901046009056884,
|
| 2513 |
+
"learning_rate": 1.968593417386705e-06,
|
| 2514 |
+
"loss": 0.3846,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.7387536112257532,
|
| 2519 |
+
"grad_norm": 2.3483856649423585,
|
| 2520 |
+
"learning_rate": 1.940021529107356e-06,
|
| 2521 |
+
"loss": 0.3856,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.7408171687990095,
|
| 2526 |
+
"grad_norm": 2.3336460330815205,
|
| 2527 |
+
"learning_rate": 1.911608511326688e-06,
|
| 2528 |
+
"loss": 0.3796,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.7428807263722658,
|
| 2533 |
+
"grad_norm": 2.3516145493656344,
|
| 2534 |
+
"learning_rate": 1.8833558392153118e-06,
|
| 2535 |
+
"loss": 0.3857,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.744944283945522,
|
| 2540 |
+
"grad_norm": 2.3634530645626164,
|
| 2541 |
+
"learning_rate": 1.8552649796188765e-06,
|
| 2542 |
+
"loss": 0.3891,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.7470078415187784,
|
| 2547 |
+
"grad_norm": 2.249211977645891,
|
| 2548 |
+
"learning_rate": 1.8273373909819132e-06,
|
| 2549 |
+
"loss": 0.3702,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.7490713990920347,
|
| 2554 |
+
"grad_norm": 2.582157058457191,
|
| 2555 |
+
"learning_rate": 1.7995745232721207e-06,
|
| 2556 |
+
"loss": 0.3849,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.751134956665291,
|
| 2561 |
+
"grad_norm": 2.3911264718615444,
|
| 2562 |
+
"learning_rate": 1.771977817905075e-06,
|
| 2563 |
+
"loss": 0.3662,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.7531985142385472,
|
| 2568 |
+
"grad_norm": 2.279432612650543,
|
| 2569 |
+
"learning_rate": 1.744548707669404e-06,
|
| 2570 |
+
"loss": 0.3697,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.7552620718118036,
|
| 2575 |
+
"grad_norm": 2.433098384561743,
|
| 2576 |
+
"learning_rate": 1.717288616652386e-06,
|
| 2577 |
+
"loss": 0.3714,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.7573256293850599,
|
| 2582 |
+
"grad_norm": 2.3966273952677764,
|
| 2583 |
+
"learning_rate": 1.6901989601660224e-06,
|
| 2584 |
+
"loss": 0.3701,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.7593891869583161,
|
| 2589 |
+
"grad_norm": 2.3793685219467697,
|
| 2590 |
+
"learning_rate": 1.6632811446735553e-06,
|
| 2591 |
+
"loss": 0.3749,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7614527445315724,
|
| 2596 |
+
"grad_norm": 2.4235058401467167,
|
| 2597 |
+
"learning_rate": 1.6365365677164413e-06,
|
| 2598 |
+
"loss": 0.3798,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.7635163021048287,
|
| 2603 |
+
"grad_norm": 2.425747353853211,
|
| 2604 |
+
"learning_rate": 1.6099666178417927e-06,
|
| 2605 |
+
"loss": 0.3754,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.7655798596780851,
|
| 2610 |
+
"grad_norm": 2.378675323045764,
|
| 2611 |
+
"learning_rate": 1.5835726745302953e-06,
|
| 2612 |
+
"loss": 0.3792,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.7676434172513413,
|
| 2617 |
+
"grad_norm": 2.372493229885198,
|
| 2618 |
+
"learning_rate": 1.5573561081245724e-06,
|
| 2619 |
+
"loss": 0.3835,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.7697069748245976,
|
| 2624 |
+
"grad_norm": 2.5336093680740026,
|
| 2625 |
+
"learning_rate": 1.531318279758055e-06,
|
| 2626 |
+
"loss": 0.3619,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.7717705323978539,
|
| 2631 |
+
"grad_norm": 2.4693265387678442,
|
| 2632 |
+
"learning_rate": 1.5054605412842915e-06,
|
| 2633 |
+
"loss": 0.3736,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.7738340899711101,
|
| 2638 |
+
"grad_norm": 2.303876812670922,
|
| 2639 |
+
"learning_rate": 1.479784235206786e-06,
|
| 2640 |
+
"loss": 0.3583,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.7758976475443665,
|
| 2645 |
+
"grad_norm": 2.590842168629549,
|
| 2646 |
+
"learning_rate": 1.4542906946092817e-06,
|
| 2647 |
+
"loss": 0.3736,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7779612051176228,
|
| 2652 |
+
"grad_norm": 2.500911682432463,
|
| 2653 |
+
"learning_rate": 1.428981243086549e-06,
|
| 2654 |
+
"loss": 0.367,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.7800247626908791,
|
| 2659 |
+
"grad_norm": 2.436965476930112,
|
| 2660 |
+
"learning_rate": 1.4038571946756724e-06,
|
| 2661 |
+
"loss": 0.3742,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.7820883202641353,
|
| 2666 |
+
"grad_norm": 2.5258796423012333,
|
| 2667 |
+
"learning_rate": 1.3789198537878202e-06,
|
| 2668 |
+
"loss": 0.3642,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.7841518778373917,
|
| 2673 |
+
"grad_norm": 2.369376183964329,
|
| 2674 |
+
"learning_rate": 1.3541705151405304e-06,
|
| 2675 |
+
"loss": 0.3746,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.786215435410648,
|
| 2680 |
+
"grad_norm": 2.3567197797039015,
|
| 2681 |
+
"learning_rate": 1.3296104636904806e-06,
|
| 2682 |
+
"loss": 0.3658,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.7882789929839042,
|
| 2687 |
+
"grad_norm": 2.4167682681302645,
|
| 2688 |
+
"learning_rate": 1.3052409745667783e-06,
|
| 2689 |
+
"loss": 0.3704,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.7903425505571605,
|
| 2694 |
+
"grad_norm": 2.3250328086103234,
|
| 2695 |
+
"learning_rate": 1.281063313004761e-06,
|
| 2696 |
+
"loss": 0.3677,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.7924061081304168,
|
| 2701 |
+
"grad_norm": 2.4369843316317,
|
| 2702 |
+
"learning_rate": 1.2570787342803064e-06,
|
| 2703 |
+
"loss": 0.3709,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7944696657036732,
|
| 2708 |
+
"grad_norm": 2.5193296767489537,
|
| 2709 |
+
"learning_rate": 1.2332884836446473e-06,
|
| 2710 |
+
"loss": 0.3719,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.7965332232769294,
|
| 2715 |
+
"grad_norm": 2.5002330071663055,
|
| 2716 |
+
"learning_rate": 1.2096937962597399e-06,
|
| 2717 |
+
"loss": 0.3677,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.7985967808501857,
|
| 2722 |
+
"grad_norm": 2.601424944526519,
|
| 2723 |
+
"learning_rate": 1.1862958971341199e-06,
|
| 2724 |
+
"loss": 0.3686,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.800660338423442,
|
| 2729 |
+
"grad_norm": 2.3874506853275363,
|
| 2730 |
+
"learning_rate": 1.16309600105931e-06,
|
| 2731 |
+
"loss": 0.3619,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8027238959966984,
|
| 2736 |
+
"grad_norm": 2.360432193713698,
|
| 2737 |
+
"learning_rate": 1.1400953125467452e-06,
|
| 2738 |
+
"loss": 0.3598,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8047874535699546,
|
| 2743 |
+
"grad_norm": 2.5865175614762053,
|
| 2744 |
+
"learning_rate": 1.117295025765235e-06,
|
| 2745 |
+
"loss": 0.3678,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8068510111432109,
|
| 2750 |
+
"grad_norm": 2.386447339664475,
|
| 2751 |
+
"learning_rate": 1.09469632447897e-06,
|
| 2752 |
+
"loss": 0.3527,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8089145687164672,
|
| 2757 |
+
"grad_norm": 2.2243473128718607,
|
| 2758 |
+
"learning_rate": 1.0723003819860555e-06,
|
| 2759 |
+
"loss": 0.3499,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8109781262897234,
|
| 2764 |
+
"grad_norm": 2.4765043887476,
|
| 2765 |
+
"learning_rate": 1.0501083610575958e-06,
|
| 2766 |
+
"loss": 0.358,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.8130416838629798,
|
| 2771 |
+
"grad_norm": 2.4123151679318093,
|
| 2772 |
+
"learning_rate": 1.0281214138773315e-06,
|
| 2773 |
+
"loss": 0.3411,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.8151052414362361,
|
| 2778 |
+
"grad_norm": 2.4030618342504892,
|
| 2779 |
+
"learning_rate": 1.0063406819818106e-06,
|
| 2780 |
+
"loss": 0.3492,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.8171687990094924,
|
| 2785 |
+
"grad_norm": 2.3438148800354988,
|
| 2786 |
+
"learning_rate": 9.847672962011233e-07,
|
| 2787 |
+
"loss": 0.3554,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.8192323565827486,
|
| 2792 |
+
"grad_norm": 2.328120737157089,
|
| 2793 |
+
"learning_rate": 9.634023766001981e-07,
|
| 2794 |
+
"loss": 0.3562,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.821295914156005,
|
| 2799 |
+
"grad_norm": 2.5107579831834816,
|
| 2800 |
+
"learning_rate": 9.422470324206385e-07,
|
| 2801 |
+
"loss": 0.3575,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.8233594717292613,
|
| 2806 |
+
"grad_norm": 2.455991687646852,
|
| 2807 |
+
"learning_rate": 9.213023620231404e-07,
|
| 2808 |
+
"loss": 0.3508,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.8254230293025175,
|
| 2813 |
+
"grad_norm": 2.3501569587377693,
|
| 2814 |
+
"learning_rate": 9.005694528304621e-07,
|
| 2815 |
+
"loss": 0.3529,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.8274865868757738,
|
| 2820 |
+
"grad_norm": 2.3892952569776527,
|
| 2821 |
+
"learning_rate": 8.800493812709677e-07,
|
| 2822 |
+
"loss": 0.3605,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.8295501444490301,
|
| 2827 |
+
"grad_norm": 2.387038888457062,
|
| 2828 |
+
"learning_rate": 8.597432127227434e-07,
|
| 2829 |
+
"loss": 0.3593,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.8316137020222865,
|
| 2834 |
+
"grad_norm": 2.3731590066808526,
|
| 2835 |
+
"learning_rate": 8.396520014582798e-07,
|
| 2836 |
+
"loss": 0.3487,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.8336772595955427,
|
| 2841 |
+
"grad_norm": 2.576723904790806,
|
| 2842 |
+
"learning_rate": 8.197767905897341e-07,
|
| 2843 |
+
"loss": 0.3541,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.835740817168799,
|
| 2848 |
+
"grad_norm": 2.2751284706261306,
|
| 2849 |
+
"learning_rate": 8.001186120147825e-07,
|
| 2850 |
+
"loss": 0.3463,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.8378043747420553,
|
| 2855 |
+
"grad_norm": 2.9241025316172804,
|
| 2856 |
+
"learning_rate": 7.806784863630324e-07,
|
| 2857 |
+
"loss": 0.3504,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.8398679323153116,
|
| 2862 |
+
"grad_norm": 2.4559935395470034,
|
| 2863 |
+
"learning_rate": 7.614574229430432e-07,
|
| 2864 |
+
"loss": 0.3538,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.8419314898885679,
|
| 2869 |
+
"grad_norm": 2.3573162952481037,
|
| 2870 |
+
"learning_rate": 7.424564196899175e-07,
|
| 2871 |
+
"loss": 0.3472,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.8439950474618242,
|
| 2876 |
+
"grad_norm": 2.3152148511829287,
|
| 2877 |
+
"learning_rate": 7.236764631134907e-07,
|
| 2878 |
+
"loss": 0.3446,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.8460586050350805,
|
| 2883 |
+
"grad_norm": 2.328674279050691,
|
| 2884 |
+
"learning_rate": 7.051185282471173e-07,
|
| 2885 |
+
"loss": 0.3494,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.8481221626083367,
|
| 2890 |
+
"grad_norm": 2.3835140443527982,
|
| 2891 |
+
"learning_rate": 6.867835785970417e-07,
|
| 2892 |
+
"loss": 0.3443,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.8501857201815931,
|
| 2897 |
+
"grad_norm": 2.366551971108674,
|
| 2898 |
+
"learning_rate": 6.68672566092376e-07,
|
| 2899 |
+
"loss": 0.3434,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.8522492777548494,
|
| 2904 |
+
"grad_norm": 2.4362523327596257,
|
| 2905 |
+
"learning_rate": 6.507864310356826e-07,
|
| 2906 |
+
"loss": 0.3361,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.8543128353281056,
|
| 2911 |
+
"grad_norm": 2.341036835199248,
|
| 2912 |
+
"learning_rate": 6.331261020541462e-07,
|
| 2913 |
+
"loss": 0.3658,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.8563763929013619,
|
| 2918 |
+
"grad_norm": 2.5075276094389753,
|
| 2919 |
+
"learning_rate": 6.156924960513638e-07,
|
| 2920 |
+
"loss": 0.3598,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.8584399504746183,
|
| 2925 |
+
"grad_norm": 2.304765407228246,
|
| 2926 |
+
"learning_rate": 5.984865181597454e-07,
|
| 2927 |
+
"loss": 0.3528,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.8605035080478746,
|
| 2932 |
+
"grad_norm": 2.334045444451915,
|
| 2933 |
+
"learning_rate": 5.815090616935115e-07,
|
| 2934 |
+
"loss": 0.3506,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.8625670656211308,
|
| 2939 |
+
"grad_norm": 2.434017607980088,
|
| 2940 |
+
"learning_rate": 5.647610081023214e-07,
|
| 2941 |
+
"loss": 0.3524,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.8646306231943871,
|
| 2946 |
+
"grad_norm": 2.400891346368746,
|
| 2947 |
+
"learning_rate": 5.482432269255011e-07,
|
| 2948 |
+
"loss": 0.3431,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.8666941807676434,
|
| 2953 |
+
"grad_norm": 2.4286183102500027,
|
| 2954 |
+
"learning_rate": 5.31956575746907e-07,
|
| 2955 |
+
"loss": 0.3464,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.8687577383408998,
|
| 2960 |
+
"grad_norm": 2.380179963058147,
|
| 2961 |
+
"learning_rate": 5.159019001503962e-07,
|
| 2962 |
+
"loss": 0.3499,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.870821295914156,
|
| 2967 |
+
"grad_norm": 2.306679882811888,
|
| 2968 |
+
"learning_rate": 5.000800336759232e-07,
|
| 2969 |
+
"loss": 0.3437,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.8728848534874123,
|
| 2974 |
+
"grad_norm": 2.289306218382388,
|
| 2975 |
+
"learning_rate": 4.844917977762653e-07,
|
| 2976 |
+
"loss": 0.3542,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.8749484110606686,
|
| 2981 |
+
"grad_norm": 2.401006920203725,
|
| 2982 |
+
"learning_rate": 4.691380017743769e-07,
|
| 2983 |
+
"loss": 0.3508,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.8770119686339248,
|
| 2988 |
+
"grad_norm": 2.294060885815769,
|
| 2989 |
+
"learning_rate": 4.5401944282136454e-07,
|
| 2990 |
+
"loss": 0.3592,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.8790755262071812,
|
| 2995 |
+
"grad_norm": 2.5092731081270654,
|
| 2996 |
+
"learning_rate": 4.3913690585510427e-07,
|
| 2997 |
+
"loss": 0.3343,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.8811390837804375,
|
| 3002 |
+
"grad_norm": 2.4672551742312523,
|
| 3003 |
+
"learning_rate": 4.244911635594856e-07,
|
| 3004 |
+
"loss": 0.3421,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.8832026413536938,
|
| 3009 |
+
"grad_norm": 2.4349291953653234,
|
| 3010 |
+
"learning_rate": 4.100829763242975e-07,
|
| 3011 |
+
"loss": 0.3482,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.88526619892695,
|
| 3016 |
+
"grad_norm": 2.532171457134297,
|
| 3017 |
+
"learning_rate": 3.959130922057497e-07,
|
| 3018 |
+
"loss": 0.3365,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.8873297565002064,
|
| 3023 |
+
"grad_norm": 2.4156983042588647,
|
| 3024 |
+
"learning_rate": 3.819822468876283e-07,
|
| 3025 |
+
"loss": 0.3424,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.8893933140734627,
|
| 3030 |
+
"grad_norm": 2.3670827733250346,
|
| 3031 |
+
"learning_rate": 3.6829116364310914e-07,
|
| 3032 |
+
"loss": 0.3489,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.8914568716467189,
|
| 3037 |
+
"grad_norm": 2.471779388721616,
|
| 3038 |
+
"learning_rate": 3.548405532972016e-07,
|
| 3039 |
+
"loss": 0.35,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.8935204292199752,
|
| 3044 |
+
"grad_norm": 2.5818181780490184,
|
| 3045 |
+
"learning_rate": 3.416311141898421e-07,
|
| 3046 |
+
"loss": 0.3468,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.8955839867932315,
|
| 3051 |
+
"grad_norm": 2.561889277902207,
|
| 3052 |
+
"learning_rate": 3.286635321396392e-07,
|
| 3053 |
+
"loss": 0.3415,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.8976475443664879,
|
| 3058 |
+
"grad_norm": 2.3816913973065654,
|
| 3059 |
+
"learning_rate": 3.159384804082666e-07,
|
| 3060 |
+
"loss": 0.335,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.8997111019397441,
|
| 3065 |
+
"grad_norm": 2.3401913676557116,
|
| 3066 |
+
"learning_rate": 3.034566196655092e-07,
|
| 3067 |
+
"loss": 0.3438,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9017746595130004,
|
| 3072 |
+
"grad_norm": 2.3557441169663598,
|
| 3073 |
+
"learning_rate": 2.912185979549587e-07,
|
| 3074 |
+
"loss": 0.3415,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.9038382170862567,
|
| 3079 |
+
"grad_norm": 2.5287391128485464,
|
| 3080 |
+
"learning_rate": 2.792250506603694e-07,
|
| 3081 |
+
"loss": 0.3433,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.905901774659513,
|
| 3086 |
+
"grad_norm": 2.454585929498319,
|
| 3087 |
+
"learning_rate": 2.67476600472672e-07,
|
| 3088 |
+
"loss": 0.3451,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.9079653322327693,
|
| 3093 |
+
"grad_norm": 2.596601343506189,
|
| 3094 |
+
"learning_rate": 2.5597385735764256e-07,
|
| 3095 |
+
"loss": 0.3583,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.9100288898060256,
|
| 3100 |
+
"grad_norm": 2.19805805166686,
|
| 3101 |
+
"learning_rate": 2.447174185242324e-07,
|
| 3102 |
+
"loss": 0.3461,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.9120924473792819,
|
| 3107 |
+
"grad_norm": 2.4174079650578624,
|
| 3108 |
+
"learning_rate": 2.3370786839356286e-07,
|
| 3109 |
+
"loss": 0.3396,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.9141560049525381,
|
| 3114 |
+
"grad_norm": 2.36584151995226,
|
| 3115 |
+
"learning_rate": 2.2294577856858236e-07,
|
| 3116 |
+
"loss": 0.3401,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.9162195625257945,
|
| 3121 |
+
"grad_norm": 2.434108457884742,
|
| 3122 |
+
"learning_rate": 2.124317078043936e-07,
|
| 3123 |
+
"loss": 0.3464,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.9182831200990508,
|
| 3128 |
+
"grad_norm": 2.2550641701066523,
|
| 3129 |
+
"learning_rate": 2.0216620197923642e-07,
|
| 3130 |
+
"loss": 0.3403,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.920346677672307,
|
| 3135 |
+
"grad_norm": 2.479423722525491,
|
| 3136 |
+
"learning_rate": 1.9214979406615097e-07,
|
| 3137 |
+
"loss": 0.34,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.9224102352455633,
|
| 3142 |
+
"grad_norm": 2.3709597200527965,
|
| 3143 |
+
"learning_rate": 1.823830041053065e-07,
|
| 3144 |
+
"loss": 0.3386,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.9244737928188197,
|
| 3149 |
+
"grad_norm": 2.4793227986038193,
|
| 3150 |
+
"learning_rate": 1.7286633917700134e-07,
|
| 3151 |
+
"loss": 0.3538,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.926537350392076,
|
| 3156 |
+
"grad_norm": 2.4372071315808896,
|
| 3157 |
+
"learning_rate": 1.6360029337533056e-07,
|
| 3158 |
+
"loss": 0.3399,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.9286009079653322,
|
| 3163 |
+
"grad_norm": 2.350437257669817,
|
| 3164 |
+
"learning_rate": 1.5458534778254319e-07,
|
| 3165 |
+
"loss": 0.3456,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.9306644655385885,
|
| 3170 |
+
"grad_norm": 2.3873927009091025,
|
| 3171 |
+
"learning_rate": 1.4582197044405556e-07,
|
| 3172 |
+
"loss": 0.3476,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.9327280231118448,
|
| 3177 |
+
"grad_norm": 2.3893579636725715,
|
| 3178 |
+
"learning_rate": 1.3731061634415787e-07,
|
| 3179 |
+
"loss": 0.3397,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.9347915806851012,
|
| 3184 |
+
"grad_norm": 2.384752010782048,
|
| 3185 |
+
"learning_rate": 1.2905172738238603e-07,
|
| 3186 |
+
"loss": 0.3487,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.9368551382583574,
|
| 3191 |
+
"grad_norm": 2.3497503513681552,
|
| 3192 |
+
"learning_rate": 1.2104573235058314e-07,
|
| 3193 |
+
"loss": 0.3383,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.9389186958316137,
|
| 3198 |
+
"grad_norm": 2.6245838898899687,
|
| 3199 |
+
"learning_rate": 1.1329304691063692e-07,
|
| 3200 |
+
"loss": 0.3414,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.94098225340487,
|
| 3205 |
+
"grad_norm": 2.3201420801643358,
|
| 3206 |
+
"learning_rate": 1.0579407357289461e-07,
|
| 3207 |
+
"loss": 0.3357,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.9430458109781263,
|
| 3212 |
+
"grad_norm": 2.390636102614023,
|
| 3213 |
+
"learning_rate": 9.854920167526983e-08,
|
| 3214 |
+
"loss": 0.3358,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.9451093685513826,
|
| 3219 |
+
"grad_norm": 2.6935442949745774,
|
| 3220 |
+
"learning_rate": 9.155880736302857e-08,
|
| 3221 |
+
"loss": 0.3516,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.9471729261246389,
|
| 3226 |
+
"grad_norm": 2.33174770965001,
|
| 3227 |
+
"learning_rate": 8.482325356925614e-08,
|
| 3228 |
+
"loss": 0.3383,
|
| 3229 |
+
"step": 2295
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.9492364836978952,
|
| 3233 |
+
"grad_norm": 2.382213172213323,
|
| 3234 |
+
"learning_rate": 7.834288999601703e-08,
|
| 3235 |
+
"loss": 0.362,
|
| 3236 |
+
"step": 2300
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.9513000412711514,
|
| 3240 |
+
"grad_norm": 2.3191616911646724,
|
| 3241 |
+
"learning_rate": 7.211805309620013e-08,
|
| 3242 |
+
"loss": 0.328,
|
| 3243 |
+
"step": 2305
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.9533635988444078,
|
| 3247 |
+
"grad_norm": 2.469419059741572,
|
| 3248 |
+
"learning_rate": 6.614906605604598e-08,
|
| 3249 |
+
"loss": 0.3331,
|
| 3250 |
+
"step": 2310
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.9554271564176641,
|
| 3254 |
+
"grad_norm": 2.4225879136955997,
|
| 3255 |
+
"learning_rate": 6.043623877837301e-08,
|
| 3256 |
+
"loss": 0.347,
|
| 3257 |
+
"step": 2315
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.9574907139909203,
|
| 3261 |
+
"grad_norm": 2.3375518129030786,
|
| 3262 |
+
"learning_rate": 5.4979867866482015e-08,
|
| 3263 |
+
"loss": 0.339,
|
| 3264 |
+
"step": 2320
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.9595542715641766,
|
| 3268 |
+
"grad_norm": 2.460240591927137,
|
| 3269 |
+
"learning_rate": 4.978023660876075e-08,
|
| 3270 |
+
"loss": 0.3409,
|
| 3271 |
+
"step": 2325
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.961617829137433,
|
| 3275 |
+
"grad_norm": 2.5514455417806388,
|
| 3276 |
+
"learning_rate": 4.4837614963976205e-08,
|
| 3277 |
+
"loss": 0.3549,
|
| 3278 |
+
"step": 2330
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.9636813867106893,
|
| 3282 |
+
"grad_norm": 2.4229916264085594,
|
| 3283 |
+
"learning_rate": 4.015225954725421e-08,
|
| 3284 |
+
"loss": 0.3462,
|
| 3285 |
+
"step": 2335
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.9657449442839455,
|
| 3289 |
+
"grad_norm": 2.392552783694151,
|
| 3290 |
+
"learning_rate": 3.572441361676171e-08,
|
| 3291 |
+
"loss": 0.3423,
|
| 3292 |
+
"step": 2340
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.9678085018572018,
|
| 3296 |
+
"grad_norm": 2.414641292682689,
|
| 3297 |
+
"learning_rate": 3.155430706107465e-08,
|
| 3298 |
+
"loss": 0.3416,
|
| 3299 |
+
"step": 2345
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.9698720594304581,
|
| 3303 |
+
"grad_norm": 2.5378189638300164,
|
| 3304 |
+
"learning_rate": 2.7642156387240882e-08,
|
| 3305 |
+
"loss": 0.3419,
|
| 3306 |
+
"step": 2350
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.9719356170037144,
|
| 3310 |
+
"grad_norm": 2.4175815882083502,
|
| 3311 |
+
"learning_rate": 2.3988164709542462e-08,
|
| 3312 |
+
"loss": 0.3321,
|
| 3313 |
+
"step": 2355
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.9739991745769707,
|
| 3317 |
+
"grad_norm": 2.2974926251848227,
|
| 3318 |
+
"learning_rate": 2.0592521738947434e-08,
|
| 3319 |
+
"loss": 0.3383,
|
| 3320 |
+
"step": 2360
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.976062732150227,
|
| 3324 |
+
"grad_norm": 2.3929562307787338,
|
| 3325 |
+
"learning_rate": 1.7455403773262713e-08,
|
| 3326 |
+
"loss": 0.3402,
|
| 3327 |
+
"step": 2365
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.9781262897234833,
|
| 3331 |
+
"grad_norm": 2.4697653358488103,
|
| 3332 |
+
"learning_rate": 1.4576973687979168e-08,
|
| 3333 |
+
"loss": 0.3245,
|
| 3334 |
+
"step": 2370
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.9801898472967395,
|
| 3338 |
+
"grad_norm": 2.4583519170902677,
|
| 3339 |
+
"learning_rate": 1.1957380927816176e-08,
|
| 3340 |
+
"loss": 0.3449,
|
| 3341 |
+
"step": 2375
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.9822534048699959,
|
| 3345 |
+
"grad_norm": 2.53558630491575,
|
| 3346 |
+
"learning_rate": 9.596761498961715e-09,
|
| 3347 |
+
"loss": 0.3378,
|
| 3348 |
+
"step": 2380
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.9843169624432522,
|
| 3352 |
+
"grad_norm": 2.4388182873283983,
|
| 3353 |
+
"learning_rate": 7.495237962011902e-09,
|
| 3354 |
+
"loss": 0.3376,
|
| 3355 |
+
"step": 2385
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.9863805200165084,
|
| 3359 |
+
"grad_norm": 2.5307806748315764,
|
| 3360 |
+
"learning_rate": 5.65291942560775e-09,
|
| 3361 |
+
"loss": 0.3437,
|
| 3362 |
+
"step": 2390
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.9884440775897647,
|
| 3366 |
+
"grad_norm": 2.494932764858625,
|
| 3367 |
+
"learning_rate": 4.0699015407702495e-09,
|
| 3368 |
+
"loss": 0.3466,
|
| 3369 |
+
"step": 2395
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.9905076351630211,
|
| 3373 |
+
"grad_norm": 2.386648335012219,
|
| 3374 |
+
"learning_rate": 2.7462664959343465e-09,
|
| 3375 |
+
"loss": 0.3384,
|
| 3376 |
+
"step": 2400
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.9925711927362774,
|
| 3380 |
+
"grad_norm": 2.7213817604169748,
|
| 3381 |
+
"learning_rate": 1.682083012681246e-09,
|
| 3382 |
+
"loss": 0.3332,
|
| 3383 |
+
"step": 2405
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.9946347503095336,
|
| 3387 |
+
"grad_norm": 2.559632513812426,
|
| 3388 |
+
"learning_rate": 8.774063421712608e-10,
|
| 3389 |
+
"loss": 0.3494,
|
| 3390 |
+
"step": 2410
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.9966983078827899,
|
| 3394 |
+
"grad_norm": 2.307214644489273,
|
| 3395 |
+
"learning_rate": 3.322782622738885e-10,
|
| 3396 |
+
"loss": 0.3417,
|
| 3397 |
+
"step": 2415
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.9987618654560462,
|
| 3401 |
+
"grad_norm": 2.3274612533228907,
|
| 3402 |
+
"learning_rate": 4.672707540231969e-11,
|
| 3403 |
+
"loss": 0.3382,
|
| 3404 |
+
"step": 2420
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 1.0,
|
| 3408 |
+
"eval_runtime": 3.3225,
|
| 3409 |
+
"eval_samples_per_second": 3.01,
|
| 3410 |
+
"eval_steps_per_second": 0.903,
|
| 3411 |
+
"step": 2423
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 1.0,
|
| 3415 |
+
"step": 2423,
|
| 3416 |
+
"total_flos": 253663452856320.0,
|
| 3417 |
+
"train_loss": 0.5332340838443727,
|
| 3418 |
+
"train_runtime": 21510.4539,
|
| 3419 |
+
"train_samples_per_second": 1.802,
|
| 3420 |
+
"train_steps_per_second": 0.113
|
| 3421 |
+
}
|
| 3422 |
+
],
|
| 3423 |
+
"logging_steps": 5,
|
| 3424 |
+
"max_steps": 2423,
|
| 3425 |
+
"num_input_tokens_seen": 0,
|
| 3426 |
+
"num_train_epochs": 1,
|
| 3427 |
+
"save_steps": 100,
|
| 3428 |
+
"stateful_callbacks": {
|
| 3429 |
+
"TrainerControl": {
|
| 3430 |
+
"args": {
|
| 3431 |
+
"should_epoch_stop": false,
|
| 3432 |
+
"should_evaluate": false,
|
| 3433 |
+
"should_log": false,
|
| 3434 |
+
"should_save": true,
|
| 3435 |
+
"should_training_stop": true
|
| 3436 |
+
},
|
| 3437 |
+
"attributes": {}
|
| 3438 |
+
}
|
| 3439 |
+
},
|
| 3440 |
+
"total_flos": 253663452856320.0,
|
| 3441 |
+
"train_batch_size": 4,
|
| 3442 |
+
"trial_name": null,
|
| 3443 |
+
"trial_params": null
|
| 3444 |
+
}
|